Enrico Drioli Lidietta Giorno Editors Enc Yclopedia of Membranes

Enrico Drioli
Lidietta Giorno
Editors
Encyclopedia of
Membranes
1 3Reference
Encyclopedia of Membranes
Enrico Drioli Lidietta Giorno
Editors
Encyclopedia of
Membranes
With 968 Figures and 185 Tables

Editors
Enrico Drioli Lidietta Giorno
University of Calabria Institute on Membrane Technology
Rende, Italy National Research Council of Italy
ITM-CNR
Rende, Italy
ISBN 978-3-662-44323-1 ISBN 978-3-662-44324-8 (eBook)

ISBN 978-3-662-44325-5 (print and electronic bundle)
DOI 10.1007/978-3-662-44324-8
Library of Congress Control Number: 2016944015
# Springer-Verlag Berlin Heidelberg 2016

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or
part of the material is concerned, specically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microlms or in any other physical way,
and transmission or information storage and retrieval, electronic adaptation, computer software, or
by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specic statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained
herein or for any errors or omissions that may have been made.
Printed on acid-free paper
This Springer imprint is published by Springer Nature

The registered company is Springer-Verlag GmbH Berlin Heidelberg
Preface
The Encyclopedia of Membranes is a reference work that covers all aspects of

membrane science and technology, from basic phenomena to the most
advanced applications and future perspectives. It is aimed at students, graduate
students, academic and industrial researchers, and practitioners.
The Encyclopedia of Membranes is organized around a list of searchable
Entries or Terms, in alphabetical order, giving the user fast and easy access to
terminology, description, theory, and application of membrane science.
Entries range from one to several printed pages and are written by one or
multiple authors.
Authors are both from academic institutions, research agencies, and
industries.
The development of the Encyclopedia is undertaken by the Chief Editors,
Prof. Enrico Drioli and Dr. Lidietta Giorno, responsible for the selection of the
topics, the terms, and the quality of the contributions. An Advisory Board was
identied to discuss major items and promote holistic vision of the work. An
Editorial Board assisted the Chief Editors in their activities.
The Encyclopedia of Membranes is a landmark work covering the science,
technology, and application of membranes, membrane operations, and related
elds.
A membrane is a nanostructured/functionalized thin barrier that controls the
exchange phenomena (matter and energy) between two different phases, both
through the external forces and under the effect of uid properties, and also
through the intrinsic characteristics of the membrane material itself.
Membranes are used in various applications. They differ widely in their
structure, in their function, and the way they are operated. Membranes may be
biological or synthetic, solid or liquid, homogeneous or heterogeneous, iso-
tropic or anisotropic, and it can be a fraction of a micrometer or various
millimeters thick. The application of different membrane structures and driv-
ing forces caused a number of rather different membrane processes such as
micro-, ultra-, and nanoltration, reverse osmosis, dialysis and electrodialysis,
pervaporation, gas separation, membrane contactors (membrane distillation,
membrane crystallizer, membrane strippers, membrane scrubbers, membrane
condenser, membrane emulsier), membrane reactors, catalytic membrane
reactor, etc.
In various elds, membrane operations are already dominant technologies.
Interesting examples are in the seawater desalination, in wastewater treatment
v
vi Preface
and reuse, and in biomedical applications. They are of interest in all sectors
strategic for a sustainable development. They are clean, safe, require low
energy input, and suitable to promote process intensication. Practically all
typical unit operations of process engineering might be redesigned as mem-
brane unit operations.
Membrane operations are largely present, for example, in the following
elds:
Water treatment and purication

Energy
Gas separation
Solvent treatment and purication
Petrochemical industry (where the integration
of different membrane operations has promising applications)
Environment
Agro-food and packaging
Biotechnology
Tissue engineering and regenerative medicine
Biorenery
Pharmaceutical and cosmetic
Microelectronics (OLEDS) (which is becom-
ing an area of interesting applications)
Other elds that will exploit more and more membranes unique features
include:
Sensors
Micromanufacturing
Responsive materials
Safety and security
Space
The large spectrum of challenges membrane operations contribute to face

today and their strong multidisciplinary character suggested to develop this
work on membrane science and membrane engineering.
Enrico Drioli
Lidietta Giorno
Acknowledgments
We are very pleased to acknowledge the important help and contribution

received by our colleagues Adele Brunetti, Enrica Fontananova, Francesca
Macedonio, and Rosalinda Mazzei, in the preparation of this Encyclopedia.
Their continuous collaboration in keeping all necessary contacts with hun-
dreds of colleagues from all around the world, in discussing on the various
unexpected problems for nalizing this work, has been signicant. Thank you!
We also wish to express grateful thanks to the members of the advisory
board for their kind availability as well as to all authors for their invaluable
contribution to the work.
Enrico Drioli
Lidietta Giorno
vii
Editors Biography
Enrico Drioli Emeritus Professor, University of Calabria, Rende, Italy
Career
Distinguished Visiting Professor, Faculty of Engineering and Information
Technology, University of Technology Sydney, Australia (January 2016)
Distinguished Adjunct Professor, Center of Excellence in Desalination Tech-
nology, King Abdulaziz University, Jeddah, Saudi Arabia (2012)
WCU (World Class University) Distinguished Visiting Professor, Department
of Energy Engineering, Hanyang University, Seoul Korea (2009)
Emeritus Professor, School of Engineering of the University of Calabria
(2012)
Professor of Chemistry and Electrochemistry at the School of Engineering of
the University of Naples (since 19681982)
Professor at the School of Engineering of the University of Calabria
(19812011)
Dean of the School of Engineering of the University of Calabria (19821985)
Founding Director of the Institute on Membrane Technology (ITM) of CNR
(from 2011)
Director of the Institute on Membranes and Chemical Reactors of the National
Research Council (19932001)
Director of the Institute on Membrane Technology (ITM) of CNR
(20022008)
Research Interest
Membranes in articial organs, integrated membrane processes, membrane
preparation transport phenomena, membrane distillation and membrane
contactors, catalytic membrane, and catalytic membrane reactors.
ix
x Editors Biography
Biography
Coordinator of the European Erasmus Mundus Doctorate in Membrane Engi-
neering (EUDIME).
Scientic Coordinator of various international contracts/agreements:
ITM-CNR-KACST, Hydrophobic membrane preparation for membrane
contactors application; ITM-CNR-KISR (Kuwait Institute for Scientic
Research), Evaluation of the potentialities of the membrane distillation tech-
nology for desalting de-oiled ltered high concentrated saline waters;
RHODIA; etc.
Member of the Scientic Advisory Panel (SAP) for the KAUST Water Desa-
lination and Reuse (WDR) Center from 2010
MIAC International Award for International Collaboration and Promotion
(October 2015)
Award with the Academician Semenov Medal of Russian Academy of Engi-
neering Science (2014)
Honorary Membership of the Czech Society of Chemical Engineering (2012)
2011, Richard Maling Barrer Prize of the EMS, for his outstanding contri-
butions to membrane science and technology
2009, Doctorate Honoris Causa from University of Paul Sabatier of Toulouse,
France
2005, International Cooperation Honor Award, Membrane Industry Associa-
tion of China (MIAC) for special dedication to International Cooperation
between China and Europe in the eld of membrane science and technology
2005Present, Guest Professor, Jiangsu Polytechnic University, China
1999, Honorary President, European Membrane Society
1992, Doctorate Honoris Causa in Chemistry and Chemical Technology,
Russian Academy of Science
1991, Honorary Professor, China Northwest University in Xian, Shaanxi,
China
19821998, President, European Society of Membrane Science and Technol-
ogy (European Membrane Society)
Advisory Board of Journal of Membrane Science, Polish Journal of Chem-

ical Technology; Editorial Board of various other international scientic
journals.
Executive Editor: Applied Water Science Journal, Springer Berlin Heidel-
berg (Germany) and Membrane Water Technology, Techno Press (Korea).
Editor of more of 750 scientic papers, 22 patents, and 24 books
Editors Biography xi
Lidietta Giorno is serving as the Director of the Institute on Membrane

Technology of the National Research Council of Italy, ITM-CNR, since 2009.
Her research expertise include membrane bioengineering, biocatalytic
membrane reactors, integrated membrane systems for bioseparations and bio-
conversions, downstream processing based on molecular separation, mem-
brane chirotechnology, membrane emulsier, and integrated membrane
operations for water treatment. She has been involved in membrane science
and engineering research and development for some 25 years, and in research
cooperations at European and international levels.
She was awarded the international Guido Dorso Award for research in
2011, sponsored by the Italian Senate and the University of Naples Federico
II. She was awarded the Sapio Red Carpet Award in 2016 among female
scientists of highest scientic prole who are engaged for the development of
the country.
She worked abroad in the USA at Sepracor Inc. in 1992, in the Netherlands
at ATO-DLO in 1994, and in France at The University of Compiegne in
19972000. She is Visiting Professor at Tianjin University of Science and
Technology, China, since 2008.
Lidietta Giorno is coeditor of 8 books, over 110 peer-reviewed scientic
papers in international journals, and coeditor of the Encyclopedia of Mem-
branes, Springer, 2016. She is member of the editorial board of scientic
journals, member of the referee pool of scientic journals and research agen-
cies, and member of international committees and several scientic societies.
She has served as member of the Board of Directors of the European
Membrane House (EMH, Belgium), formed to continue the activity of the
NanoMemPro Network of Excellence funded by the European Commission
within the FP6.
She has served on the European Membrane Society Council as the Presi-
dent of the EMS Council and the Editor of the EMS Membrane Newsletter.
She was nominated Honorary Member of the European Membrane Society
in 2014.
Advisory Board
Bhattacharyya Dibakar University Alumni Professor, Lexington, KY

Tony Fane Nanyang Technological University 18, Nanyang Drive,
Singapore, Singapore
Michel Y. Jaffrin Department of Biological Engineering, Technological
University of Compiegne, UMR CNRS 6600, Compiegne Cedex, France
Ora Kedem Emeritus Professor of Desalination and Water Treatment
Research at Ben, Gurion University and also professor at the Weizmann
Institute of Science, Ben-Gurion University of the Negev, Beer Sheva, Israel
William J. Koros Georgia Institute of Technology-School of Chemical and
Biomolecular Eng., Atlanta, GA, USA
Kew-Ho Lee National Research Laboratory for Functional Membranes,
Environment and Resources, Research Center, Yuseong, Daejeon, South
Korea
Young Moo Lee Hanyang University, Seoul, Republic of Korea
Donald R. Paul Department of Chemical Engineering and Texas Materials
Institute University of Texas at Austin, Austin, Texas, TX, USA
Heiner Strathmann Institut fur Chemische Verfahrenstechnik, Universitt
Stuttgart, Stuttgart, Germany
Yuri Yampolskii A.V. Topchiev Institute of Petrochemical Synthesis,
Moscow, Russia
xiii
Editorial Board
Adele Brunetti ITM-CNR c/o University of Calabria, Rende, Italy

Enrica Fontananova ITM-CNR c/o University of Calabria, Rende, Italy
Francesca Macedonio ITM-CNR c/o University of Calabria, Rende, Italy

Rosalinda Mazzei ITM-CNR c/o University of Calabria, Rende, Italy
xv
Contributors
A. Achilli Environmental Resources Engineering Department, Humboldt

State University, Arcata, CA, USA
Pejman Ahmadiannamini Biomedical Engineering, University of Arkan-
sas, Fayetteville, AR, USA
Catia Algieri Institute on Membrane Technology, National Research Council
of Italy, ITM-CNR, Rende, Italy
Aamer Ali Department of Environmental and Chemical Engineering,
University of Calabria, Rende, Italy
Institute on Membrane Technology, National Research Council of Italy,
ITM-CNR, Rende, Italy
Sacide Alsoy Altinkaya Department of Chemical Engineering, Izmir Insti-
tute of Technology, Urla, Izmir, Turkey
Vitor D. Alves LEAF Linking Landscape, Environment, Agriculture and
Food, Instituto Superior de Agronomia, Universidade de Lisboa, Lisboa,
Portugal
Shota Ando Department of Applied Chemistry, Meiji University, Tama-ku,
Kawasaki, Japan
Muhammad Waqas Anjum School of Chemical and Materials Engineering,
National University of Science and Technology, Islamabad, Pakistan
Alice Antony School of Chemical Engineering, The University of New
South Wales, UNESCO Centre for Membrane Science and Technology, Syd-
ney, NSW, Australia
Mario Aparicio Instituto de Ceramica y Vidrio (CSIC), Madrid, Spain
Adriana Arca-Ramos Department of Chemical Engineering, Institute of
Technology, University of Santiago de Compostela, Santiago de Compostela,
Spain
Pietro Argurio Department of Environmental and Chemical Engineering,
University of Calabria, Arcavacata di Rende, Italy
Antonino Salvatore Arico CNR-ITAE Institute, Messina, Italy
Marco Avataneo RD&T, Solvay Specialty Polymers, Bollate (MI), Italy
xvii
xviii Contributors
He Bai Momentive Performance Materials Inc, Friendly, WV, USA
Renbi Bai Department of Civil and Environmental Engineering, National

University of Singapore, Faculty of Engineering, Singapore, Singapore
Malini Balakrishnan The Energy and Resources Institute (TERI), New

Delhi, India
Joerg Balster Application Technology, Evonik Fibres GmbH, Schoering,

Austria
Mohamed Abou El-Fetouh Barakat Department of Environmental Sci-

ences, Faculty of Meteorology, Environment and Arid Land Agriculture,
King Abdulaziz University, Jeddah, Saudi Arabia
Central Metallurgical R & D Institute, Helwan, Cairo, Egypt
Giuseppe Barbieri Institute on Membrane Technology, National Research

Council of Italy, ITM-CNR, Rende, Italy
Onita D. Basu Department of Civil and Environmental Engineering, Carle-

ton University, Ottawa, Canada
Arnaud Baudot Physics and Analysis Division, IFP Energies Nouvelles,

Solaize, France
Laurent Bazinet Department of Food Sciences, Universit Laval, Quebec,

QC, Canada
Fabio Bazzarelli Institute on Membrane Technology, National Research

Katalin Bela-Bako Research Institute of Bioengineering, Membrane Tech-

nology and Energetics, University of Pannonia, Veszprem, Hungary
Marie-Pierre Belleville Institut Europen des Membranes (IEM): UMR

CNRS 5635, Universit de Montpellier II, Montpellier, France
Benny Danilo Belviso Istituto di Cristallograa, Consiglio Nazionale delle

Ricerche (C.N.R), Bari, Italy
Juana Benavente Departamento de Fisica Aplicada I, Universidad de

Malaga, Facultad de Ciencias, Malaga, Spain
Nilay Bereli Department of Chemistry, Biochemistry Division, Hacettepe

University, Ankara, Turkey
Samuel Bernard Institut Europen des Membranes, ENSCM-UM2-CNRS,

(UMR 5635), Universit de Montpellier 2, Montpellier, France
A. Moura Bernardes Universidade Federal do Rio Grande do Sul

(PPGE3M-UFRGS), RS, Porto Alegre, Brazil
Paola Bernardo Institute on Membrane Technology, National Research

Contributors xix
Richard Bernewitz Food Process Engineering, Karlsruhe Institute of Tech-

nology (KIT), Karlsruhe, Germany
Anna Maria Bertasa Solvay Specialty Polymers, Bollate (MI), Italy
Dmitri Bessarabov Faculty of Engineering, DST HySA Infrastructure Cen-
ter of Competence, North-West University, Potchefstroom, South Africa
Chiranjib Bhattacharjee Department of Chemical Engineering, Jadavpur
University, Kolkata, West Bengal, India
Maarten Biesheuvel Wetsus, European Centre of Excellence for Sustainable
Water Technology, Leeuwarden, The Netherlands
M. R. Bilad Centre for Surface Chemistry and Catalysis, KU Leuven, Leu-
ven, Heverlee, Belgium
ngeles Blanco Department of Chemical Engineering, Universidad
Complutense de Madrid, Madrid, Spain
Carlo Boaretti Department of Industrial Engineering, University of Padova,
Padova, Italy
Cristiana Boi Dipartimento di Ingegneria Civile, Chimica, Ambientale e dei
Materiali, Universit di Bologna, Bologna, Italy
Przemysaw Borys Department of Physical Chemistry and Technology of
Polymers, Silesian University of Technology (SUT), Gliwice, Poland
Karel Bouzek Faculty of Chemical Technology, University of Chemistry and
Technology Prague, Technick 5, Prague 6, Czech Republic
Rune Bredesen Sustainable Energy Technology, SINTEF Materials and
Chemistry, Oslo, Norway
Eugenio Bringas Department of Chemical and Biomolecular Engineering,
University of Cantabria, Santander, Cantabria, Spain
D. J. Broer Department of Chemical Engineering and Chemistry, Group
Functional Organic Materials and Devices (SFD), Eindhoven University of
Technology, Eindhoven, The Netherlands
Stephan Brosillon Institut Europen des Membranes, ENSCM-UM-CNRS,
(UMR 5635), Universit de Montpellier, Montpellier Cedex 5, France
Adele Brunetti Institute on Membrane Technology, National Research
Peter M. Budd Organic Materials Innovation Centre, School of Chemistry,
University of Manchester, Manchester, UK
Vasilis Burganos Institute of Chemical Engineering Sciences, Foundation
for Research and Technology Hellas, FORTH/ICE-HT, Patras, Hellas
Tomas Bystron Faculty of Chemical Technology, University of Chemistry
and Technology, Prague, Technick 5, Prague 6, Czech Republic
xx Contributors
Rajnish Kaur Calay Institute for Building, Energy and Material Technol-
ogy, University of Troms, Navrik, Norway
Energy Technology Research Group, Narvik University College, Lodve
Langes gt 2, Navrik, Norway
Rocco Caliandro Institute of Crystallography, CNR, Bari, Italy
Anja Car Department of Chemistry, University of Basel, Basel, Switzerland
Alessio Caravella National Institute of Advanced Industrial Science and
Technology Research Institute for Innovation in Sustainable Chemistry,
ISC-AIST, Tsukuba, Ibaraki, Japan
Juergen Caro Institute of Physical Chemistry and Electrochemistry, Leibniz
University, Hannover, Germany
Mariolino Carta School of Chemistry, The University of Edinburgh, Edin-
burgh, Scotland, UK
Alfredo Cassano Institute on Membrane Technology, National Research
Alberto Cassetta Istituto di Cristallograa CNR UOS di Trieste, Italy
Sophie Cerneaux UMR 5635 ENSCM, CNRS, UM, Institut Europen des
Membranes, Montpellier, France
S. K. Chakarvarti Manav Rachna International University, Faridabad, India
Department of Physics, National Institute of Technology, Kurukshetra, Hary-
ana, India
Catherine Charcosset Universit Lyon 1, Lyon, Villeurbanne, France
Tzyy Haur Chong Singapore Membrane Technology Centre, Nanyang
Environment and Water Research Institute; School of Civil and Environmental
Engineering, Nanyang Technological University, Singapore, Singapore
Kwang-Ho Choo Department of Environmental Engineering, Kyungpook
National University, Daegu, Republic of Korea
Benjamin Chu Department of Chemistry, Stony Brook University, Stony
Brook, NY, USA
Liang-Yin Chu Membrane Science and Functional Materials Group, School
of Chemical Engineering, Sichuan University, Chengdu, Sichuan, China
Jose Coca Department of Chemical and Environmental Engineering, Uni-
versity of Oviedo, Oviedo, Spain
Isabel Coelhoso LAQV-REQUIMTE, Departamento de Qumica, Faculdade
de Ciencias e Tecnologia, Universidade Nova de Lisboa, Caparica, Portugal
Carmela Conidi Institute on Membrane Technology, National Research
Contributors xxi
Marc Cretin Institut Europen des Membranes, ENSCM-UM-CNRS (UMR

5635), Universit de Montpellier, Montpellier Cedex 5, France
National School of Chemistry of Montpellier, University of Montpellier,
Montpellier, France
Alessandra Criscuoli Institute on Membrane Technology, National
Research Council of Italy, ITM-CNR, Rende, Italy
Zhaoliang Cui State Key Laboratory of Materials-Oriented Chemical Engi-
neering, College of Chemistry and Chemical Engineering, Nanjing Tech
University, Nanjing, China
Efrem Curcio Department of Environmental and Chemical Engineeering,
Napoleone DAgostino Institute on Membrane Technology, National
Research Council (ITM-CNR), University of Calabria, Rende, Italy
Naser Dalali Phase Separation and FIA Lab, Department of Chemistry,
Faculty of Science, University of Zanjan, Zanjan, Iran
Nguyen Phuoc Dan Faculty of Environment and Natural Resources,
University of Technology, Vietnam National University, Ho Chi Minh City,
Viet Nam
Ranjana Das Chemical Engineering Department, Jadavpur University,
Kolkata, West Bengal, India
Maria Grazia De Angelis Dipartimento di Ingegneria Civile, Chimica,
Ambientale e dei Materiali, Universit di Bologna, Bologna, Italy
Frederico de Araujo Kronemberger COPPE Chemical Engineering Pro-
gram, Federal University of Rio de Janeiro (UFRJ), Rio de Janeiro, RJ, Brazil
Loredana De Bartolo Institute on Membrane Technology, National
Johannes De Bruijn Department of Agroindustry, Food Engineering Group,
Universidad de Concepcion, Chillan, Chile
Giovanni De Filpo Dipartimento di Chimica e Tecnologie Chimiche,
Universit della Calabria, Arcavacata di Rende, Italy
Giorgio De Luca Institute on Membrane Technology, National Research
Mario De Rosa Department of Experimental Medicine, Section of Biotech-
nology, Medical Histology and Molecular Biology, Second University of
Naples, Napoli, Italy
Maria Norberta De Pinho Chemical Engineering, Instituto Superior
Tcnico, Universidade de Lisboa, Lisbon, Portugal
xxii Contributors
Winnie Dejonghe Separation and Conversion Technology, Flemish Institute

of Technological Research (VITO), Mol, Belgium
Baolin Deng Department of Civil and Environmental Engineering, Univer-
sity of Missouri-Columbia, Columbia, MO, USA
Adil Denizli Department of Chemistry, Biochemistry Division, Hacettepe
Renzo Di Felice University of Genova, Genova, Italy
Gianluca Di Proo Institute on Membrane Technology, National Research
Maria Luisa Di Vona Dip. Di Scienze Technologie Chimiche, University of
Rome Tor Vergata, Rome, Italy
Nazely Diban Department of Chemical and Biomolecular Engineering,
University of Cantabria, Santander, Cantanbria, Spain
George Dibrov A.V. Topchiev Institute of Petrochemical Synthesis, Russian
Academy of Sciences, Moscow, Russia
Laura Donato Institute on Membrane Technology, National Research Coun-
cil of Italy, ITM-CNR, Rende, Italy
Marijana M. Dragosavac Chemical Engineering Department, Loughbor-
ough University, Loughborough, Leicestershire, UK
Enrico Drioli Institute on Membrane Technology, National Research Coun-
Department of Environmental and Chemical Engineering, University of
Calabria, Rende, Italy
WCU Energy Engineering Department, Hanyang University, Seoul, S. Korea
Center of Excellence in Desalination Technology, King Abdulaziz University,
Jeddah, Saudi Arabia
Martin Drobek Institut Europen des Membranes, Universit de Montpel-
lier, Montpellier, France
Gabriela Dudek Department of Physical Chemistry and Technology of
Mikel Duke Institute for Sustainability and Innovation, College of Engineer-
ing and Science, Victoria University, Melbourne, VIC, Australia
Hoang Hanh Phuoc Duong King Abdullah University of Science and Tech-
nology (KAUST), Biological and Environmental Science and Engineering
Division (BESE), Thuwal, Saudi Arabia
Sravya Duvvuri Membrane Separations Group, Chemical Engineering Divi-
sion, CSIR-Indian Institute of Chemical Technology, Hyderabad, India
Contributors xxiii
Piotr Dydo Silesian University of Technology, Gliwice, Poland

Ehiaze Augustine Ehimen Separation and Conversion Technology, Flemish
Institute of Technological Research (VITO), Mol, Belgium
Gemma Eibes Department of Chemical Engineering, Institute of Technol-
ogy, University of Santiago de Compostela, Santiago de Compostela, Spain
M. S. Mohy Eldin Polymer Materials Research Department, Advanced
Technologies and New Materials Research Institute (ATNMRI), City for
Scientic Research and Technology Applications (CSRTA), New Borg
El-Arab City, Alexandria, Egypt
Isabel C. Escobar University of Kentucky, Lexington, KY, USA
Anthony Gordon Fane Singapore Membrane Technology Centre, Nanyang
Environment and Water Research Institute, Nanyang Technological Univer-
sity, Singapore, Singapore
Tony Fane School of Chemical Engineering, The University of New South
Wales, Sydney, NSW, Australia
Singapore Membrane Technology Centre, Nanyang Technological University,
Mnica Faria Chemical Engineering, Columbia University, New York, NY,
USA
Gumersindo Feijoo Department of Chemical Engineering, Institute of Tech-
nology, University of Santiago de Compostela, Santiago de Compostela, Spain
Ekain Fernandez Energy and Environment, TECNALIA, Donostia-San
Sebastian, Gipuzkoa, Spain
Frederico Castelo Ferreira Department of Bioengineering, Instituto Supe-
rior Tcnico, Institute for Bioengineering and Biosciences, Universidade de
Lisboa, Lisbon, Portugal
Patrick Fievet Institut UTINAM, University of Franche-Comt, Besanon,
France
Alberto Figoli Institute on Membrane Technology, National Research
Gustavo Fimbres Weihs Ctedras CONACYT, Instituto Tecnolgico de
Sonora, Cd. Obregn, Sonora, Mexico
Greg Foley School of Biotechnology, Dublin City University, Dublin,
Ireland
Enrica Fontananova Institute on Membrane Technology, National Research
Patrizia Formoso Dipartimento di Farmacia e Scienze della Salute e della
Nutrizione, Universit della Calabria, Arcavacata di Rende, Italy
xxiv Contributors
Filomena Freitas UCIBIO-REQUIMTE, Departamento de Qumica,

Faculdade de Cincias e Tecnologia, Universidade Nova de Lisboa, Caparica,
Portugal
Karel Friess Department of Physical Chemistry, University of Chemistry

and Technology Prague, Prague, Czech Republic
Francesco Galiano Institute on Membrane Technology, National Research

Claudia F. Galinha REQUIMTE-CQFB, Chemistry Dept. Faculdade de

Cincias e Tecnologia, FCT-Universidade Nova de Lisboa, Campus da
Caparica, Caparica, Portugal
Fausto Gallucci Chemical Process Intensication, Department of Chemical

Engineering and Chemistry, Eindhoven University of Technology, Eindhoven,
The Netherlands
Volker Gaukel Food Process Engineering, Karlsruhe Institute of Technology

(KIT), Karlsruhe, Germany
Abaynesh Yihdego Gebreyohannes Institute on Membrane Technology,

National Research Council of Italy, ITM-CNR, Rende, Italy
Geoffrey Geise Department of Chemical Engineering, University of

Virginia, Charlottesville, VA, USA
Lidietta Giorno Institute on Membrane Technology, National Research

Zbigniew J. Grzywna Department of Physical Chemistry and Technology of

Annarosa Gugliuzza Institute on Membrane Technology, National Research

Gemma Gutirrez Department of Chemical and Environmental Engineer-

ing, University of Oviedo, Oviedo, Spain
May-Britt Hgg Department of Chemical Engineering, Norwegian Univer-

sity of Science and Technology, Trondheim, Norway
Vu Pha Hai Greentech Environment Joint Stock Company, Ho Chi Minh

City, Vietnam
Nobuo Hara Research Institute for Chemical Process Technology, National

Institute of Advanced Industrial Science and Technology, Tsukuba, Japan
Tao He Lab for Membrane Materials and Separation Technology, Shanghai

Advanced Research Institute CAS, Shanghai, China
Xuezhong He Department of Chemical Engineering, Norwegian University

of Science and Technology, Trondheim, Sr-Trndelag, Norway
Contributors xxv
Arash Helmi Department of Chemical Engineering and Chemistry, Chemi-

cal Process Intensication, Technische Universiteit Eindhoven, Eindhoven,
The Netherlands
Sanne Hermans Centre for Surface Chemistry and Catalysis, Department of
M2S, Faculty Bioscience Engineering, KU Leuven, Leuven, Belgium
Daphne Hermosilla Department of Chemical Engineering, Universidad
Mitsuru Higa Graduate School of Science and Engineering, Yamaguchi
University, Ube, Yamaguchi, Japan
A. Higuchi Department of Chemical and Materials Engineering, National
Central University, Chung-Li, Taoyuan, Taiwan
Nidal Hilal Centre for Water Advanced Technologies and Environmental
Research (CWATER), College of Engineering, Swansea University, Swansea,
Wales, UK
W. S. Winston Ho Department of Chemical and Biomolecular Engineering,
The Ohio State University, Columbus, OH, USA
Richard G. Holdich Department of Chemical Engineering, Loughborough
University, Loughborough, Leicestershire, UK
R. W. Holloway Department of Civil and Environmental Engineering,
Colorado School of Mines, Golden, CO, USA
Benjamin S. Hsiao Department of Chemistry, Stony Brook University,
Stony Brook, NY, USA
Shu-Hau Hsu Department of Materials Science and Engineering, University
of Florida, Gainesville, FL, USA
Department of Energy Engineering, Hanyang University, Seoul, Republic of
Korea
Yan Huang State Key Laboratory of Materials-Oriented Chemical Engineer-
ing, College of Chemical Engineering, Nanjing Tech University, Nanjing,
China
Alessandra Imbrogno Institute on Membrane Technology, National
University of Calabaria, Rende, Italy
Christophe Innocent Institut Europen des Membranes, ENSCM, CNRS,
Universit de Montpellier, Montpellier, France
Angel Irabien Departamento de Ingeniera Qumica y Qumica Inorgnica,
Universidad de Cantabria, Santander, Spain
Ahmad Fauzi Ismail Faculty of Chemical and Energy Engineering,
Universiti Teknologi Malaysia, Advanced Membrane Technology Research
Centre (AMTEC), Johor Bahru, Johor, Malaysia
xxvi Contributors
Juhana Jaafar Faculty of Chemical and Energy Engineering, Universiti

Teknologi Malaysia, Advanced Membrane Technology Research Centre
(AMTEC), Johor Bahru, Johor, Malaysia
Marijke Jacobs Sustainable Materials Management, Flemish Institute for
Technological Research VITO, Mol, Belgium
Michel Jaffrin Department of Biological Engineering, Technological Uni-
versity of Compiegne, BP 20529, Compiegne Cedex, France
Johannes Carolus Jansen Institute on Membrane Technology, National
Lan Ying Jiang School of Metallurgy and Environment, Central South
University, Changsha, China
Zhongyi Jiang School of Chemical Engineering and Technology, Tianjin
University, Nankai, Tianjin, China
Wanqin Jin State Key Laboratory of Materials-Oriented Chemical Engineer-
ing, Nanjing University of Technology, Nanjing, China
Daniel Johnson Centre for Water Advanced Technologies and Environmen-
tal Research (CWATER), College of Engineering, Swansea University,
Swansea, Wales, UK
D. M. Jovanovi Institute of Chemistry, Technology and Metallurgy
(I.Ch.T.M.), Center of Catalysis and Chemical Engineering, University of
Belgrade, Belgrade, Republic of Serbia
Anne Julbe Institut Europen des Membranes, Universit de Montpellier,
Montpellier, France
Nora Jullok School of Bioprocess Engineering, Universiti Malaysia Perlis
(UniMAP), Arau, Perlis, Malaysia
Mari Kallioinen Lappeenranta University of Technology, Lappeenranta,
Finland
Shinji Kanehashi Department of Applied Chemistry, Meiji University,
Tama-ku, Kawasaki, Japan
H. Kawakita Department of Applied Chemistry, Saga University, Saga,
Japan
A. P. Kharitonov Talrose Branch of the Institute of Energy Problems of
Chemical Physics of the Russian Academy of Sciences, Moscow Region,
Russia
Jeong F. Kim Department of Energy Engineering, Hanyang University,
Seoul, South Korea
Benjamin F. K. Kingsbury Imperial College London, London, UK
Vladimir S. Kislik Campus Givat Ram, Casali Institute of Applied Chemis-
try, The Hebrew University of Jerusalem, Jerusalem, Israel
Contributors xxvii
Chalida Klaysom Centre for Surface Chemistry and Catalysis, KU Leuven,

Leuven, Heverlee, Belgium
P. Knauth Madirel UMR 7246, Aix Marseille University-CNRS, Marseille

Cedex, France
Isao Kobayashi National Food Research Institute, National Agriculture and

Food Research Organization, Tsukuba, Ibaraki, Japan
N. M. Kocherginsky Biomime, Urbana, IL, USA
G. Kolb Product Sector Energy and Chemistry, Fraunhofer ICT-IMM,

Mainz, Germany
A. B. Koltuniewicz Biotechnology and Bioprocess Engineering Division,

Faculty of Chemical and Process Engineering, Warsaw University of Tech-
nology, Warsaw, Poland
Andras Koris Corvinus University of Budapest, Budapest, Hungary
Zoltn Kovcs Department of Food Engineering, Institue of Bioengineering

and Process Engineering, Szent Istvan University, Budapest, Hungary
Boguslaw Kruczek Department of Chemical and Biological Engineering,

University of Ottawa, Ottawa, ON, Canada
Ashwani Kumar Institute of Chemical Process and Environmental Technol-

ogy, National Research Council Canada, Ottawa, ON, Canada
Kripal S. Lakhi Australian Institute for Bioengineering and Nanotechnol-

ogy, The University of Queensland, Brisbane, QLD, Australia
Doriano Lamba Institute of Crystallography, National Research Council,

Trieste, Italy
Marek Lan University of Chemistry and Technology Prague, Prague 6,

Czech Republic
W. J. Lau Advanced Membrane Technology Research Centre (AMTEC),

Universiti Teknologi Malaysia, Skudai, Johor, Malaysia
Thi Xuan Huong Le Vanlang University, Ho Chi Minh City, Vietnam
Pierre Le-Clech School of Chemical Engineering, The University of New

South Wales, UNESCO Centre for Membrane Science and Technology,
Sydney, NSW, Australia
Kew-Ho Lee Center for Membranes, Korea Research Institute of Chemical

Technology (KRICT), Yuseong-gu, Daejeon, South Korea
Cecile Legallais UMR CNRS 7338 Biomechanics and Bioengineering,

Universit de Technologie de Compiegne (UTC), Compiegne, France
Raymond L. Legge Department of Chemical Engineering, University of

Waterloo, Waterloo, ON, Canada
xxviii Contributors
Juan M. Lema Department of Chemical Engineering, Institute of Technol-

ogy, University of Santiago de Compostela, Santiago de Compostela, Spain
Gerardo Len Departamento de Ingeniera Qumica y Ambiental,
Universidad Politcnica de Cartagena, Cartagena, Spain
Geoffroy Lesage Institut Europen des membranes, IEM, UMR-5635,
Universit de Montpellier, ENSCM, CNRS, Montpellier, France
Baoqiang Liao Department of Chemical Engineering, Lakehead University,
Thunder Bay, ON, Canada
Hongyu Li UNESCO Centre for Membrane Science and Technology, The
University of New South Wales, Sydney, Australia
Jianxin Li State Key Laboratory of Separation Membranes and Membrane
Processes, School of Materials Science and Engineering, Tianjin Polytechnic
University, Tianjin, Peoples Republic of China
Na Li Department of Chemical Engineering, School of Chemical Engineer-
ing and Technology, Xian Jiaotong University, Xian City, China
Xianhui Li State Key Laboratory of Separation Membranes and Membrane
Processes, School of Materials Science and Engineering, Tianjin Polytechnic
University, Tianjin, Peoples Republic of China
Yanshuo Li State Key Laboratory of Catalysis, Dalian Institute of Chemical
Physics, CAS, Dalian, China
Yifan Li School of Chemical Engineering and Technology, Tianjin Univer-
sity, Nankai, Tianjin, China
Haiqing Lin Department of Chemical and Biological Engineering, Univer-
sity at Buffalo, The State University of New York, Buffalo, NY, USA
Frank Lipnizki Alfa Laval Copenhagen, Sborg, Copenhagen, Denmark
Fu Liu Ningbo Institute of Materials Technology and Engineering, Chinese
Academy of Sciences, Ningbo, China
Wei Liu Energy and Environment Technology Directorate, Pacic Northwest
National Laboratory, Richland, WA, USA
Jordi Llorca Institute of Energy Technologies, Universitat Politcnica de
Catalunya, Barcelona, Spain
Lucia Lloret Chemical and Environmental Engineering Department,
Federico Santa Maria Technical University, Santiago, Chile
Juan Manuel Lpez-Romero Departamento de Quimica Organica,
Universidad de Malaga, Facultad de Ciencias, Malaga, Spain
Alessandra Lorenzetti Department of Industrial Engineering, University of
Padova, Padova, Italy
Contributors xxix
Thelmo Lu-Chau Department of Chemical Engineering, Institute of Tech-

nology, University of Santiago de Compostela, Santiago de Compostela, Spain
Diego Luna Department of Organic Chemistry, University of Cordoba,
Cordoba, Spain
Herb Lutz EMD Millipore Corporation, Playa del Rey, CA, USA
Hongyang Ma Department of Chemistry, Stony Brook University, Stony
Brook, NY, USA
Angela C. Macedo LEPABE Laboratory of Engineering of Processes,
Environment, Biotechnology and Energy, University of Porto, Porto, Portugal
Francesca Macedonio Institute on Membrane Technology, National
Sayed S. Madaeni Membrane Research Center, Chemical Engineering
Department, Razi University, Kermanshah, Iran
Hacene Mahmoudi Science and Technology, Faculty of Technology,
University Hassiba Benbouali of Chlef, Chlef, Algeria
Swachchha Majumdar Ceramic Membrane Division, CSIR-Central Glass
and Ceramic Research Institute, Kolkata, West Bengal, India
F. Xavier Malcata DEQ Department of Chemical Engineering and LEPABE
Laboratory of Engineering of Processes, Environment, Biotechnology and
Energy, University of Porto, Porto, Portugal
Mario Malinconico Research Director, Institute for Polymers, Composites
and Biomaterials (IPCB-CNR), Pozzuoli, Naples, Italy
Reyes Mallada Department of Chemical and Environmental Engineering
and Aragon Nanoscience Institute, Universidad de Zaragoza, Zaragoza, Spain
Mara Matos Department of Chemical and Environmental Engineering,
University of Oviedo, Oviedo, Spain
Davide Mattia Department of Chemical Engineering, University of Bath,
Bath, UK
Rosalinda Mazzei Institute on Membrane Technology, National Research
Beatriz Mello UFSCar, Federal University of Sao Carlos, Buri, SP, Brazil
Adelio Mendes Faculty of Engineering, University of Porto, Porto, Portugal
Julie Mendret Institut Europen des Membranes, ENSCM-UM-CNRS
(UMR 5635), Universit de Montpellier, Montpellier Cedex 5, France
xxx Contributors
Miguel Menendez Department of Chemical and Environmental Engineer-

ing, University of Zaragoza, Zaragoza, Spain
Antonietta Messina Institute on Membrane Technology, National Research
Bart Michielsen Sustainable Materials Management, Flemish Institute for
Technological Research VITO, Mol, Belgium
Vesna Middelkoop Flemish Institute for Technological Research VITO,
Mol, Belgium
Martine Mietton-Peuchot Vine and Wine Sciences Institute, Universite de
BORDEAUX, Villenave dOrnon, France
Antonella Milella Department of Chemistry, University of Bari Aldo
Moro, Institute of Nanotechnology (CNR-NANOTEC), Bari, Italy
Francesca Militano Institute on Membrane Technology, National Research
Ubavka B. Mio Faculty of Physical Chemistry, University of Belgrade,
Belgrade, Republic of Serbia
Ilaria Mirabelli HyGear B.V. Westervoortsedijk, AV, Arnhem, The
Netherlands
Hossein Mirsaeedghazi Department of Food Technology, Abouraihan
College, University of Tehran, Pakdasht, Iran
Michele Modesti Department of Industrial Engineering, University of
Padova, Padova, Italy
Khayet Mohamed Department of Applied Physics, Universidad
P. K. Mohapatra Radiochemistry Division, Bhabha Atomic Research
Centre, Mumbai, Maharashtra, India
Z. D. Mojovi Institute of Chemistry, Technology and Metallurgy
(I.Ch.T.M.), Center of Catalysis and Chemical Engineering, University of
Belgrade, Belgrade, Republic of Serbia
Raffaele Molinari Dipartimento di Ingegneria per lAmbiente e il Territorio
e Ingegneria Chimica (DIATIC), Universit della Calabria, Rende, Italy
M. Concepcin Monte Department of Chemical Engineering, Universidad
Seung-Hyeon Moon School of Environmental Science and Engineering,
Gwangju Institute of Science and Technology (GIST), Gwangju, South Korea
Contributors xxxi
Mara. T. Moreira Department of Chemical Engineering, School of Engi-

neering, University of Santiago de Compostela, Santiago de Compostela,
Spain
Sabrina Morelli Institute on Membrane Technology, National Research
Carlos Moreno Department of Analytical Chemistry, Faculty of Marine and
Environmental Sciences, University of Cdiz, Puerto Real, Spain
Siddhartha Moulik Membrane Separations Group, Chemical Engineering
Division, CSIR-Indian Institute of Chemical Technology, Hyderabad, India
Philippe Moulin Laboratoire de Mcanique, Modlisation et Procds
Propres (M2P2-UMR 7340), Equipe Procds Membranaires (EPM), Aix
Marseille Universit, AIX en PROVENCE, France
Z. V. P. Murthy Department of Chemical Engineering, S.V. National Insti-
tute of Technology, Surat, Gujarat, India
Kazukiyo Nagai Department of Applied Chemistry, Meiji University, Tama-
ku, Kawasaki, Japan
Harsha Nagar Membrane Separations Group, Chemical Engineering Divi-
sion, CSIR-Indian Institute of Chemical Technology, Hyderabad, India
Hiroki Nagasawa Department of Chemical Engineering, Hiroshima Univer-
sity, Higashi-Hiroshima, Japan
Endre Nagy Research Institute of Chemical and Process Engineering,
University of Pannonia, Veszprem, Hungary
Mitsutoshi Nakajima Faculty of Life and Environmental Sciences, Univer-
sity of Tsukuba, Tsukuba, Ibaraki, Japan
Zoran P. Nedi Faculty of Physical Chemistry, University of Belgrade,
Belgrade, Republic of Serbia
Carlos Negro Department of Chemical Engineering, Universidad
Fiore Pasquale Nicoletta Dipartimento di Farmacia e Scienze della Salute e
della Nutrizione, Universit della Calabria, Arcavacata di Rende, Italy
Johanna Niemist Helsinki Region Chamber of Commerce, Enterprise
Europe Network, Helsinki, Finland
Suzana Pereira Nunes Division of Biological and Environmental Sciences
and Engineering, Nanostructured Polymeric Membranes Lab, King Abdullah
University of Science and Technology (KAUST), Thuwal, Saudi Arabia
Darren Oatley-Radcliffe College of Engineering, Swansea University,
Swansea, UK
xxxii Contributors
Osvaldo N. Oliveira Jr. Sao Carlos Institute of Physics IFSC, University of

Sao Paulo USP, Sao Paulo, Brazil
B. S. Ooi School of Chemical Engineering, Universiti Sains Malaysia,
Nibong Tebal, Pulau Pinang, Malaysia
Immaculada Ortiz Department of Chemical and Biomolecular Engineering,
University of Cantabria, Santander, Cantabria, Spain
M. H. D. Othman Advanced Membrane Technology Research Centre
(AMTEC), Universiti Teknologi Malaysia, Skudai, Johor, Malaysia
Faculty of Chemical and Energy Engineering, Universiti Teknologi Malaysia,
Skudai, Johor, Malaysia
N. Othman Faculty of Chemical and Energy Engineering, Universiti
Teknologi Malaysia, Skudai, Johor, Malaysia
Mehmet A. Oturan Laboratory of Earth Materials and Environment,
University Paris-Est, Marne La valle, Champs-sur-Marne, France
David A. Pacheco Tanaka Energy and Environment, TECNALIA,
Donostia-San Sebastian, Gipuzkoa, Spain
Leonardo Palmisano Dipartimento di Energia, Ingegneria
dellInformazione e Modelli Matematici (DEIM), Universit di Palermo,
Palermo, Italy
Fabio Palumbo Department of Chemistry, University of Bari Aldo Moro,
Institute of Nanotechnology (CNR-NANOTEC), Bari, Italy
Chi Hoon Park Gyeongnam National University of Science and Technology
(GNTECH), Jinju-si, Gyeongsangnam-do, Republic Of Korea (South Korea)
Donald R. Paul Department of Chemical Engineering, Texas Materials Insti-
tute, Austin, TX, USA
Radoslav Paulen Department of Biochemical and Chemical Engineering,
Technische Universitt Dortmund, Dortmund, Germany
Carmen Pazos Department of Chemical and Environmental Engineering,
University of Oviedo, Oviedo, Spain
Graeme K. Pearce Membrane Consultancy Associates Limited, Reading,
RG8 7EJ, UK
Ludmila Georgieva Peeva Imperial College London, London, UK
Jinming Peng Key Laboratory for Green Chemical Technology of Ministry
of Education, School of Chemical Engineering and Technology, Tianjin Uni-
versity, Tianjin, Peoples Republic of China
Francois Perreault School of Sustainable Engineering and the Built Envi-
ronment, Arizona State University, Tempe, AZ, USA
Thijs Peters SINTEF Materials and Chemistry, SINTEF, Oslo, Norway
Contributors xxxiii
Emma Piacentini Institute on Membrane Technology, National Research

Monica Pica University of Perugia, Department of Pharmaceutical Sciences,

Perugia, Italy
Antonella Piscioneri Institute on Membrane Technology, National Research

Teresa Poerio Institute on Membrane Technology, National Research Coun-

Carla A. M. Portugal REQUIMTE-LAQV, Chemistry Department, FCT-

Universidade Nova de Lisboa, Campus da Caparica, Caparica, Portugal
Grald Pourcelly Institut Europeen des Membranes, CC 047, Universite

Montpellier II, Place Eugene Bataillon, Montpellier, France
Gonzalo Prieto Max-Planck-Institut fur Kohlenforschung, Mulheim an der

Ruhr, Germany
Pushpinder S. Puri PuriMem, LLC, Emmaus, PA, USA
Saren Qi Singapore Membrane Technology Centre, Nanyang Technological

University, Singapore, Singapore
Department of Civil Engineering, The University of Hong Kong, Hong Kong,
Hong Kong SAR
Xianghong Qian Department of Biomedical Engineering, University of

Arkansas, Fayetteville, AR, USA
Damien Quemener Institut Europen des Membranes, Montpellier, France
Cejna Anna Quist-Jensen Institute on Membrane Technology, National

Mukhlis bin A. Rahman Faculty of Chemical and Energy Engineering,

Advanced Membrane Technology Research Centre (AMTEC), Universiti
Teknologi Malaysia, Skudai, Johor, Malaysia
Department of Energy Engineering, Faculty of Chemical and Energy Engi-
neering, Universiti Teknologi Malaysia, Johor Bahru, Johor Darul Takzim,
Malaysia
Fen Ran School of Material Science and Engineering, State Key Laboratory
of Advanced Processing and Recycling of Non-ferrous Metals, Lanzhou
University of Technology, Lanzhou, P. R. China
Giuseppe Ranieri Institute on Membrane Technology, National Research

xxxiv Contributors
Maria A. M. Reis UCIBIO-REQUIMTE, Departamento de Qumica,

Faculdade de Cincias e Tecnologia, Universidade Nova de Lisboa, Caparica,
Portugal
Gilbert M. Rios Place Eugene Bataillon, IEM European Membrane Insti-

tute, Montpellier, France
Tania Rodenas Max Planck Institute fur Chemische Energiekonversion,

Mulheim an der Ruhr, Germany
Denis Roizard Laboratoire Ractions et Gnie des Procds, CNRS-

Universit de Lorraine, Nancy, France
Virginia Romero Departamento de Fisica Aplicada I, Universidad de

Malaga, Facultad de Ciencias, Malaga, Spain
Martina Roso Department of Industrial Engineering, University of Padova,

Padova, Italy
Stephanie Roualdes IEM/UM II CC 047, Place Eugene Bataillon, Mont-

pellier, France
Ren Ruby-Figueroa Institute on Membrane Technology, National

Aleksandra Rybak Department of Physical Chemistry and Technology of

Jose Sabate Department dEnginyeria Agroalimentaria i Biotecnologia,

Universitat Politecnica de Catalunya, Castelldefels, Spain
Keiji Sakaki National Institute of Advanced Industrial Science and Technol-

ogy (AIST), Research Institute for Sustainable Chemistry, Tsukuba, Ibaraki,
Japan
Simona Salerno Institute on Membrane Technology, National Research

Marius Sandru Polymer Particles and Surface Chemistry, Materials and

Chemistry, SINTEF, Trondheim, Norway
Tsuneji Sano Department of Applied Chemistry, Hiroshima University,

Graduate School of Engineering, Higashi-Hiroshima, Japan
Giulio C. Sarti Department of Chemical Engineering, Mining and Environ-

mental Technology, University of Bologna, Bologna, Italy
Shuichi Sato Department of Applied Chemistry, Meiji University, Tama-ku,

Kawasaki, Japan
Yamini Satyawali Separation and Conversion Technology, Vlaamse

Instelling voor Technologisch Onderzoek (VITO), Boeretang 200, Belgium
Contributors xxxv
Thomas Schfer Institute for Polymer Materials (Polymat), University of the

Basque Country (UPV/EHU), Donostia-San Sebastin, Spain
A. P. H. J. Schenning Department of Chemical Engineering and Chemistry,
Group Functional Organic Materials and Devices (SFD), Eindhoven Univer-
sity of Technology, Eindhoven, The Netherlands
Chiara Schiraldi Department of Experimental Medicine, Section of Bio-
technology, Medical Histology and Molecular Biology, Second University of
Naples, Napoli, Italy
Colin A. Scholes Department of Chemical and Biomolecular Engineering,
University of Melbourne, Parkville, VIC, Australia
Marco Scholz Aachener Verfahrenstechnik Chemical Process Engineering,
RWTH Aachen University, Aachen, Germany
Karin Schroen Laboratory of Food Process Engineering, Wageningen
University and Research Centre, Wageningen, The Netherlands
Anna Schuch Food Process Engineering, Karlsruhe Institute of Technology
(KIT), Karlsruhe, Germany
Heike Schuchmann Food Process Engineering, Karlsruhe Institute of
Technology (KIT), Karlsruhe, Germany
Dwaipayan Sen Department of Chemical Engineering, Heritage Institute of
Technology, Kolkata, West Bengal, India
Laurentia Setiawan Singapore Membrane Technology Centre, Nanyang
Environment and Water Research Institute, Nanyang Technological Univer-
sity, Singapore, Singapore
Jiahui Shao School of Environmental Science and Engineering, Shanghai
Jiao Tong University, Shanghai, Peoples Republic of China
Cameron Shearer School of Chemical and Physical Sciences, Flinders
University, Adelaide, South Australia, Australia
Institute of Physical Chemistry, University of Muenster, Muenster, North
RhineWestphalia, Germany
Tauqir A. Sherazi Department of Chemistry, COMSATS Institute of Infor-
mation Technology, Abbottabad, Pakistan
Jiafu Shi School of Environment Science and Engineering, Tianjin Univer-
sity, Tianjin, China
Ryohei Shindo Department of Applied Chemistry, Meiji University, Tama-
ku, Kawasaki, Japan
Barbara Siembida-Lsch Centre for Alternative Wastewater Treatment,
Fleming College, Lindsay, ON, Canada
xxxvi Contributors
Wolfgang M. Sigmund Department of Materials Science and Engineering,

University of Florida, Gainesville, FL, USA
Department of Energy Engineering, Hanyang University, Seoul, Republic of
Korea
Veronica Silva Dow Water and Process Solution Division at The Dow
Chemical Company, Boulevard Cecilia Grierson 355, Buenos Aires,
Argentina
S. T. V. Sim Singapore Membrane Technology Centre, Nanyang Environ-
ment and Water Research Institute, Nanyang Technological University,
Silvia Simone Institute on Membrane Technology, National Research Coun-
Eugene Skouras Institute of Chemical Engineering Sciences, Foundation for
Research and Technology, Hellas, Patras, Hellas
Jose M. Sousa Chemistry Department, School of Life and Environment
Sciences, University of Tras-os-Montes e Alto Douro, Vila Real, Portugal
LEPABE Laboratory for Process Engineering, Environment, Biotechnology
and Energy, Chemical Engineering Department, Faculty of Engineering,
University of Porto, Rua Roberto Frias, Porto, Portugal
S. Sridhar Membrane Separations Group, Chemical Engineering Division,
CSIR-Indian Institute of Chemical Technology, Hyderabad, India
Heiner Strathmann Universitt Stuttgart, Institute of Chemical Process
Engineering Stuttgart, Baden-Wurttemberg, Germany
Anna Strzelewicz Faculty of Chemistry, Silesian University of Technology,
Gliwice, Poland
Yanlei Su School of Chemical Engineering and Technology, Tianjin Univer-
sity, Nankai, Tianjin, Peoples Republic of China
Shing-Yi Suen Department of Chemical Engineering, National Chung Hsing
University, Taichung, Taiwan
Guofei Sun Chemical and Biomolecular Engineering, Aquaporin Asia Pte.
Ltd, Singapore, Singapore
Sridhar Sundergopal Membrane Separations Group, Chemical Engineering
Division, CSIR-Indian Institute of Chemical Technology, Hyderabad, India
Stanislaus Raditya Suwarno Singapore Membrane Technology Centre,
Nanyang Technological University, Singapore, Singapore
Petr Sysel Department of Polymers, Institute of Chemical Technology,
Prague, Czech Republic
Guttorm Syvertsen-Wiig Ceramic Powder Technology AS, Tiller, Norway
Contributors xxxvii
Gyorgy Szekely School of Chemical Engineering and Analytical Science,

The University of Manchester, Manchester, UK
Roberto Taboada-Puig Department of Chemical Engineering, Institute of
Technology, University of Santiago de Compostela, Santiago de Compostela,
Spain
Chung Tai-Shung Neal National University of Singapore, Singapore,
Singapore
Yoshinobu Tanaka IEM Research, Ushiki-shi, Ibaraki, Japan
Chuyang Y. Tang Department of Civil Engineering, The University of Hong
Kong, Hong Kong, Hong Kong SAR
Department of Civil Engineering The University of Hong Kong, Hong Kong,
Hong Kong SAR
Franco Tasselli Institute on Membrane Technology, National Research
Bui Xuan Thanh Faculty of Environment and Natural Resources, University
of Technology, Vietnam National University, Ho Chi Minh City, Vietnam
Kalsang Tharpa Department of Chemistry, University of Mysore, Mysore,
India
Elena Tocci Institute on Membrane Technology, National Research Council
of Italy, ITM-CNR, Rende, Italy
Maria Tomaszewska Faculty of Chemical Technology and Engineering,
West Pomeranian University of Technology, Szczecin, Poland
Carles Torras Font Bioenergy and Biofuels, Catalonia Institute for Energy
Research (IREC), Tarragona, Catalonia, Spain
Bijay P. Tripathi Department of Nanostructured Materials, Leibniz Institute
of Polymer Research Dresden, Dresden, Germany
Francesco Trotta Department of Chemistry, University of Torino, Torino,
Italy
Toshinori Tsuru Department of Chemical Engineering, Hiroshima Univer-
sity, Higashi-Hiroshima, Japan
Marian Turek Faculty of Chemistry, Silesian University of Technology,
Gliwice, Poland
Deniz T urkmen Department of Chemistry, Biochemistry Division,
Hacettepe University, Ankara, Turkey
Tadashi Uragami Organization for Research and Development of Innova-
tive Science and Technology (ORDIST), Kansai University, Suita, Osaka,
Japan
xxxviii Contributors
Mohamad Hekarl Uzir School of Chemical Engineering, Universiti Sains

Malaysia, Penang, Malaysia
B. Aakruthi Vaishnavi Membrane Separations Group, Chemical Engineer-

ing Division, CSIR-Indian Institute of Chemical Technology, Hyderabad,
India
Csar Valderrama Department of Chemical Engineering, Universitat

Politcnica de Catalunya-Barcelona Tech (UPC), Barcelona, Catalonia, Spain
Bart Van der Bruggen Department of Chemical Engineering, Process Engi-

neering for Sustainable Systems (ProcESS), KU Leuven, Leuven, Belgium
P. C. van H. Kuringen Department of Chemical Engineering and Chemistry,

Group Functional Organic Materials and Devices (SFD), Eindhoven Univer-
sity of Technology, Eindhoven, The Netherlands
Ivo Vankelecom Centre for Surface Chemistry and Catalysis, KU Leuven,

Leuven, Heverlee, Belgium
Todor Vasiljevic Advanced Food Systems Research Unit, College of Health

and Biomedicine, Victoria University, Melbourne, VIC, Australia
C. V. Vedavyasan Department of Technology, Annamalai University,

Chidambaram, Tamil Nadu, India
Tanja Vidakovic-Koch Max Planck Institute for Dynamics of Complex

Technical Systems, Magdeburg, Germany
Loreen O. Villacorte FMC Technologies, Separation Innovation and

Research Center, Arnhem, The Netherlands
Giuseppe Vitola Institute on Membrane Technology, National Research

Goran T. Vladisavljevi Department of Chemical Engineering, Loughbor-

ough University, Loughborough, Leicestershire, UK
Alexey Volkov A.V. Topchiev Institute of Petrochemical Synthesis Russian

Academy of Sciences (TIPS RAS), Moscow, Russia
V. V. Volkov A.V. Topchiev Institute of Petrochemical Synthesis Russian

Academy of Sciences (TIPS RAS), Moscow, Russian Federation
Diogo Volpati Sao Carlos Institute of Physics IFSC, University of Sao

Paulo USP, Sao Paulo, Brazil
Bo Wang Department of Chemical Engineering, Imperial College London,

London, UK
Honglei Wang Department of Chemical and Biomolecular Engineering,

jMedtech Pte Ltd, Singapore, Singapore
Contributors xxxix
Rong Wang Singapore Membrane Technology Centre, Nanyang Environ-

ment and Water Research Institute; School of Civil and Environmental Engi-
neering, Nanyang Technological University, Singapore, Singapore
Shaofei Wang School of Chemical Engineering and Technology, Tianjin
Xiaoli Wang School of Chemical Engineering and Technology, Tianjin
Zhi Wang Chemical Engineering Research Center, School of Chemical
Engineering and Technology, Tianjin University, Tianjin, China
Zhiwei Wang Department of Environmental Engineering, College of Envi-
ronmental Science and Engineering, Tongji University, Shanghai, China
S. Ranil Wickramasinghe Ralph E Martin Department of Chemical
Engineering, University of Arkansas, Fayetteville, AR, USA
Dianne Wiley School of Chemical and Biomolecular Engineering, The
University of Sydney, Sydney, NSW, Australia
Paul Melvyn Williams College of Engineering, Swansea University,
Swansea, UK
Daniel Winter Fraunhofer Institute for Solar Energy Systems (ISE),
Freiburg, Germany
Chunrui Wu State Key Laboratory of Hollow Fiber Membrane Materials
and Membrane Processes, Institute of Biological and Chemical Engineering,
Tianjin Polytechnic University, Tianjin, China
Hong Wu School of Chemical Engineering and Technology, Tianjin Univer-
sity, Nankai, Tianjin, Peoples Republic of China
Yvan Wyart CNRS, Centrale Marseille, M2P2 UMR 7340, Equipe Procds
Membranaires (EPM), Aix Marseille Universit, Aix en Provence Cedex,
France
Tongwen Xu Collaborative Innovation Centre of Chemistry for Energy
Materials, School of Chemistry and Material Science, University of Science
and Technology of China, Hefei, P.R. China
Yuri Yampolskii A.V. Topchiev Institute of Petrochemical Synthesis,
Russian Academy of Sciences, Moscow, Russia
Dong Yang School of Chemical Engineering and Technology, Tianjin
Handan Yavuz Department of Chemistry, Biochemistry Division, Hacettepe
Jun Yin Department of Civil and Environmental Engineering, University of
Missouri-Columbia, Columbia, MO, USA
xl Contributors
Masakazu Yoshikawa Department of Biomolecular Engineering, Kyoto

Institute of Technology, Kyoto, Japan
Miki Yoshimune Research Institute for Chemical Process Technology,
National Institute of Advanced Industrial Science and Technology (AIST),
Tsukuba, Japan
S. M. Javaid Zaidi Australian Institute for Bioengineering and Nanotech-
nology, The University of Queensland, Brisbane, QLD, Australia
King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia
Grayna Zakrzewska-Kotuniewicz Institute of Nuclear Chemistry and
Technology Instytut Chemii i Techniki Jdrowej, Warszawa, Poland
Li-Zhi Zhang Key Lab of Enhanced Heat Transfer and Energy Conserva-
tion, South China University of Technology, Guangzhou, China
Chen Zhang School of Chemical and Biomolecular Engineering, Atlanta,
GA, USA
Jianhua Zhang Department of Polymer Science and Engineering, School of
Chemical Engineering and Technology, Tianjin University, Tianjin, China
Tianjin Key Laboratory of Membrane Science and Desalination Technology,
Tianjin University, Tianjin, China
Yuzhong Zhang School of Material Science and Engineering, Tianjin Poly-
technic University, Tianjin, China
Jing Zhao School of Chemical Engineering and Technology, Tianjin Uni-
versity, Nankai, Tianjin, China
Xueting Zhao School of Chemical Engineering and Technology, Tianjin
University, Nankai, Tianjin, Peoples Republic of China
Pei Shan Zhong Functionalized Materials and Nanostructures/Global
Research Centre Singapore (GMM/F), BASF South East Asia Pte Ltd.,
Li-Ping Zhu MOE Key Laboratory of Macromolecular Synthesis and
Functionalization, Department of Polymer Science and Engineering, Zhejiang
University, Hangzhou, Zhejiang, Peoples Republic of China
Xiaoying Zhu Department of Civil and Environmental Engineering, National
University of Singapore, Faculty of Engineering, Singapore, Singapore
Andrew L. Zydney Department of Chemical Engineering, The Pennsylvania
State University, University Park, PA, USA
A
Ab Initio Calculation it is an ab initio theory, but various potentials

dening the quantum system are obtained by
Giorgio De Luca using empirical data. For example, hybrid poten-
Institute on Membrane Technology, National tials are linear combinations of pure potentials.
Research Council of Italy, ITM-CNR, Rende, The coefcients of the linear combinations are
Italy obtained using experimental data. Nevertheless,
other potentials are obtained from the thermody-
namic theory of electron gas (statistical thermo-
Although the term ab initio usually indicates cal- dynamics); thus the latter can be considered ab
culations based on rst principles, either they are initio. For this reason also the density functional
related to classical or to quantum mechanics, this theory is considered an ab initio methodology.
term is generally used when it refers to calcula- Summarizing, two broad theoretical methods can
tions based on quantum mechanics rst principles be identied: those that make use of adjustable
(Szabo and Ostlund 1994). Thus, if the parameters that should be to be dened each time,
dAlembert equation is used to describe a force ad hoc, depending on the system studied and those
balance, in which the expression of the forces is that do not use ad hoc parameters which only refer
only expression of chemical-physical average to the rst principles and are known as the ab
properties (values obtained as average of micro- initio methods.
scopic properties) without any use of empirical Particular attention should be given on
adjustable parameters, the calculation can also be multiscale simulations in which paramters
dened ab initio. As mentioned, in general, the provided by sub and nanoscales are used in
term ab initio is used when the calculations are meso and macro scales without resort of empiri-
performed by using the Schrdinger equation or cal-adjustable parameters. Even these types
rather an eigenvalue equation for the quantum of calculations can be dened ab initio. (De Luca
system. However, attention must be paid to the et al. 2014) recently, have used different
quantities or potential operators that dene the approaches, ranging from quantum mechanics to
eigenvalue equations. In fact, if the operators in macroscopic models, to simulate the formation of
the Hamiltonians are dened using experimental the protein cake on a membrane surface and the
data or by means of tting procedure, then the conseguent decay of water ux. Although various
calculation should not be dened ab initio but parameters were calculated ab-initio others have
semiempirical (Clementi and Corongiu 1995). not yet been assessed in this way; however, this
The density functional theory (Parr and Yang work represents one of the rst attempts of an ab-
1989) is a controversial case, since in principle, initio simulation of a membrane process.
E. Drioli, L. Giorno (eds.), Encyclopedia of Membranes,
DOI 10.1007/978-3-662-44324-8
2 Acetic Acid Dehydration by PV
References membrane surface, (2) liquid diffusion through

the membrane matrix, (3) vaporization, (4) vapor
Clementi E, Corongiu G (eds) (1995) Methods and tech- diffusion, and (5) vapor desorption (Van der
niques in computational chemistry METECC-95.
Bruggen 2009). Several studies have been under-
STEF, Cagliari
De Luca G, Bisignano F, Paone F, Curcio S (2014) Multi- taken on dehydration of acetic acid-water mix-
scale modeling of protein fouling in ultraltration pro- tures using organic, inorganic, and composite
cess. J Membr Sci 452:400414 membranes. It has been shown that inorganic
Parr RG, Yang W (1989) Density functional theory of
membranes, in most cases, are capable to provide
atoms and molecules. Oxford University Press, New
York a very high selectivity of water, which penetrates
Szabo A, Ostlund NS (1994) Modern quantum chemistry. through the membrane, but with a very low per-
Macmillan Publishing Co., New York meate ux, e.g., using zeolites. In an acid-proof
silicalite-1 zeolite study, it was found that the
separation factor was near innite (Masuda
et al. 2003). A study using free-radical graft poly-
Acetic Acid Dehydration by PV merization of 1-vinylimidazole (VI) onto the
as-synthesized mordenite zeolite membrane in
Nora Jullok acetic acid dehydration has been conducted by
School of Bioprocess Engineering, Universiti (Chen et al. 2010). It was found that by applying
Malaysia Perlis (UniMAP), Arau, Perlis, this membrane in pervaporation at higher temper-
Malaysia ature, it yields a higher ux and also a higher
selectivity. Another interesting study was the
investigation by Teli et al. (2007) on applying a
Acetic acid is an important organic chemical used silicotungstic acid incorporated sodium alginate
in a wide range of applications worldwide. (STA-NaAlg-5). This study gave an excellent sep-
Among others, it is used as a raw material in the aration factor and ux, ranging from 0.194 to
synthesis of vinyl acetate monomer (VAM) for the 0.489 kg/m2 h. However, the membrane perfor-
production of a wide variety of polymers. These mance decreased when applied at high tempera-
include polyvinyl acetate, polyvinyl alcohol, ture and in a more dilute feed solution. As the
polyacetals, and ethylene vinyl acetate copoly- temperature and water content in the feed solution
mers (Minardi et al. 2002). A regular distillation are increased, the ux increases and selectivity
process to dewater acetic acid is known to be decreases. Other than inorganic membranes, poly-
energy intensive and to incur high operational meric membrane shows promising solution.
costs. Since the separation of acetic acid from Pervaporation up to 90 wt.% acetic acid using
water using distillation is far from being the best polyphenylsulfone membranes at temperature
process, any cost-efcient method to dehydrate range from 30 to 80 C exhibited high ux, good
acetic acid is in demand. Among these alternative separation factor, and membrane stability (Jullok
methods, pervaporation has been suggested as a et al. 2011). The plasticizing effect was postulated
promising solution to separate acetic acid aqueous to result in a special sieving property, which
solutions (Alghezawi et al. 2005; Durmaz- assisted the water molecules to be transported
Hilmioglu et al. 2001; Huang et al. 1988). across the membrane matrix.
Pervaporation is a membrane-based process that
can be potentially applied to purify acetic acid
(i.e., via acetic acid dehydration). It requires
the availability of a system with a vacuum down- References
stream to enable separation by partial vaporization
Alghezawi N, Sanli O, Aras L, Asman G (2005) Separation
through a dense, selective membrane. The perme-
of acetic acid-water mixtures through acrylonitrile
ation of a permeating component can be described grafted poly (vinyl alcohol) membranes by
in ve consecutive steps: (1) sorption on the pervaporation. Chem Eng Proc 44:5158
Acetic Acid Dehydration by Pervaporation with Charged Membranes 3
Chen Z, Yang J, Yin D, Li Y, Wu S, Lu J, Wang J (2010) as methanol carbonylation for acetic acid synthe-
Fabrication of poly (1-vinylimidazole)/mordenite sis. About 75 % of acetic acid in the chemical
grafting membrane with high pervaporation perfor-
mance for the dehydration of acetic acid. J Membr Sci industry is produced by this process, in which A
349:175182 carbon monoxide reacts with methanol under
Durmaz-Hilmioglu N, Yildirim AE, Sakaoglu AS, inuence of a rhodium complex catalyst at
Tulbentci S (2001) Acetic acid dehydration by 180 C and pressure of 3040 atm (Shen
pervaporation. Chem Eng Proc 40
Huang RYM, Moreira A, Notarfonzo R, Xu YF et al. 2011). The main challenges in acetic acid
(1988) Pervaporation separation of acetic acid water and water separation are (1) cost minimization and
mixtures using modied membranes. I. Blended (2) optimizing separation efciency. In acetic acid
polyacrylic acid (Paa) nylon 6 membranes. J Appl Pol aqueous solutions, there is a dynamic equilibrium
Sci 35:11911200
Jullok N, Darvishmanesh S, Luis P, Van der Bruggen between neutral molecules and the acetate and
B (2011) The potential of pervaporation for separation hydronium ions. It was observed that acetic acid
of acetic acid and water mixtures using dissociation is not only driven by the breaking of
polyphenylsulfone membranes. Chem Eng H-bonds connecting the ions but also triggered by
J 175:306315
Masuda T, Otani S, Tsuji T, Kitamura M, Mukai SR a solvent reorganization around the proton-
(2003) Preparation of hydrophilic and acid-proof accepting water (Park et al. 2006), as illustrated
silicalite-1 zeolite membrane and its application to in Scheme 1. Water molecule forms four hydrogen
selective separation of water from water solutions of bonds, whereas a hydronium ion only forms three.
concentrated acetic acid by pervaporation. Sep Pur
Tech 32:181189 Due to the proximity of the boiling point between
Minardi F, Belpoggi F, Soffritti M, Ciliberti A, Lauriola M, acetic acid and water, separation using conven-
Cattin E, Maltoni C (2002) Results of long term carci- tional methods may not be the best choice to
nogenicity bioassay on vinyl acetate monomer in consider. Dehydration of acetic acid by
Sprague Dawley rats. Ann N Y Acad Sci 982:106122
Teli SB, Gokavi GS, Sairam M, Aminabhavi TM pervaporation with charged membranes can
(2007) Highly water selective silicotungstic acid potentially be applied. Ion exchange membranes
(H4siw12o40) incorporated novel sodium alginate are pictured as porous with charges lining the pore
hybrid composite membranes for pervaporation dehy- walls. An example of ion exchange membrane
dration of acetic acid. Sep Pur Tech 54:178186
Van der Bruggen B. Fundamentals of membrane solvent used for this specic application is Naon
separation and pervaporation. In: I. o. M. T. I.-C. Uni- 117, which has been modied to form Naon
versity of Calabria, Via P. Bucci, 17/C, 87030 Rende (C8H17)4N+. Other ion exchange membranes
(CS), Italy and D. o. C. E. a. M. University of Calabria, which have been evaluated include sulfonated
Via P. Bucci, cubo 44/A, 87030 Rende (CS), Italy eds.,
Memb Oper. (Italy, 2009), pp 4561 polysulfone (PSF(SO3)-H+) and crossed-linked
styrene-co-butadiene base. They are mechanically
stabilized with poly(vinyl chloride) baking and
contain quaternary ammonium and sulfonic acid
moieties, respectively (AMV-CH3OO and
Acetic Acid Dehydration by CMV-H+). The interaction of polymers with
Pervaporation with Charged water is strongly inuenced by the xed ion and
Membranes counterion. Hydrogen-bonding, dipole-dipole,
and ion-dipole are the three possible interactions
Nora Jullok between water molecules and membrane surface.
School of Bioprocess Engineering, Universiti As the interaction between xed ion and counter-
Malaysia Perlis (UniMAP), Arau, Perlis, ion with water increases, the sorption selectivity
Malaysia to water increases. A polymer having H+ as coun-
terion has a lower water sorption and lower water
sorption selectivity compared to a polymer having
Acetic acid is produced both synthetically and metal counterion. Besides, H+-containing poly-
biologically. The synthetic production typically mer also has a high permeability but low separa-
uses the Monsanto process, which is also known tion factor. On the other hand, for a similar
4 Acetic Acid Dehydration by Pervaporation Zeolite Membranes
Acetic Acid Dehydration

by Pervaporation with
O O
1/2
Charged Membranes, H3C + H 2O H3C + H OH2+
Scheme 1 Dissociation of
acetic acid in water OH O
1/2
polymer, which has a metal counterion, removed continuously using a water permeable
pervaporation shows lower permeability but acid resistant membrane.
higher separation factor (Semenova et al. 1997). The most promising results concerning separa-
tions in acid medium have been obtained with zeo-
lites with a moderate Si/Al ratio such as zeolite
References T with Si/Al = 4 or mordenite with Si/Al ratio
from 7 to 20. Both zeolites have been tested in
Park JM, Laio A, Iannuzzi M, Parrinello M (2006) Disso- water/acetic acid mixture (50/50 wt.%) showing
ciation mechanism of acetic acid in water. J Am Chem
stability after the exposure to this mixture at mild
Soc 128:1131811319
Semenova SI, Ohya H, Soontarapa K (1997) Hydrophilic temperature around 70 C for several days.
membranes for pervaporation: an analytical review. Although the performance of these zeolites in
Desalination 110:251286 pervaporation for water dehydration are lower
Shen Z, Zhou J, Zhou X, Zhang Y (2011) The production
compared to zeolite A (Arruebo et al. 2008), the
of acetic acid from microalgae under hydrothermal
conditions. Appl Energy 88:34443447 acid stability of these zeolites is higher than that of
zeolite A and makes them good candidates for the
separation of acid mixtures. The researchers of
Mitsubishi studied the technical feasibility
of mordenite membranes for dehydration of
Acetic Acid Dehydration by acetic acid aqueous solution with a feed rate of
Pervaporation Zeolite Membranes 1000 kg h1. Their results suggested that the
developed mordenite membrane deserved to the
Reyes Mallada further technically developments of up-scale and
Department of Chemical and Environmental mass-production (Sato et al. 2011). The mem-
Engineering and Aragon Nanoscience Institute, brane performance in pervaporation was mea-
Universidad de Zaragoza, Zaragoza, Spain sured in a feed mixture of water (50 wt.%)/AAc
(50 wt.%) to be 10.9 kg m2/h for permeate ux
and to be 0.77 10 6 mol m2/sPa for water
The best performance in pervaporation for water permeance with separation factor of 500 at
dehydration corresponds to zeolite A, which is the 130 C.
zeolite with the lowest Si/Al ratio (equal to 1) and
therefore the highest hydrophilicity. However, the
high aluminum content of this zeolite makes it References
unstable under acidic conditions and it dissolves
at low pH. Arruebo M, Mallada R, Pina MP (2008) Zeolite membranes:
synthesis, characterization, important applications, and
The separation of water in the presence of acetic
recent advances, Chapter 10. In: Handbook of mem-
acid or other acids could be important in the case of brane separations, 1st edn. CRC Press, Boca Raton,
membrane reactors, active zeolite membrane reac- pp 233323
tors (AZMRs), for equilibrium displacement (see Sato K, Sugimoto K, Kyotani T, Shimotsuma N, Kurata
T (2011) Synthesis, reproducibility, characterization,
also Membrane reactor equilibrium conversion and
pervaporation and technical feasibility of preferentially
Pervaporation membrane reactor). The esterication b-oriented mordenite membranes for dehydration of
reaction using acid produces water that could be acetic acid solution. J Membr Sci 385386:2029
A-CNT 5
even though both H2S and CO2 have smaller

Acid Gas: Effect on Membrane kinetic diameters than CH4. In the natural gas
Properties sweetening industry, both cellulose acetate- and A
polyimide-based membranes have been
Colin A. Scholes commercialized.
Department of Chemical and Biomolecular Prolonged exposure to both H2S and CO2 can
Engineering, University of Melbourne, Parkville, lead to plasticization and degradation of the poly-
VIC, Australia meric membrane over time. This is because the
acidic nature of the gases means they will readily
interact with the polymeric chains, disrupting
Acidic gas removal is the largest application of chain packing and lubricating polymer chain
gas separation membranes and a vital process in motion, resulting in more rubbery behavior. Mem-
the natural gas processing industry. Membranes brane degradation by these gases is greatly
used in this process are designed to separate H2S enhanced if water is present, because of the
and CO2, and this process is known as natural gas weak acidic dissociation of these gases (Scholes
sweetening. This is readily required for et al. 2009).
low-quality natural gas reserves, which need to For inorganic and metallic membranes, H2S is
be processed to achieve pipeline specications of of particular concern, because of the effect sulfur
CO2 <2 mol% and H2S <4 ppm (Baker 2001). can have on the membrane material
The removal of acidic gases is also required for (Lu et al. 2007). A good example of this is in H2
biogas and gasication, where H2S and CO2 are separation with Pd-based membranes. The pres-
separated from CH4, CO, and H2. For natural gas ence of H2S leads to membrane brittlement and
sweetening, the membrane must exhibit high per- performance degradation because of lattice
meability for H2S and CO2, with a large selectiv- expansion on the formation of Pd-S. Hence, in
ity relative to methane. applications such as gasication, it is necessary
In general, nonporous polymeric membranes for the gas to have H2S removed prior to the gas
are applied in acidic gas removal, taking advan- separation membrane, use sacricial layers, or add
tage of the solution-diffusion mechanism and the sulfur-resistant materials.
high solubility of H2S and CO2, compared to CH4.
Also polymeric membranes almost always dem-
onstrate selectivity for H2S over CO2, again References
because of the higher condensability of H2S com-
pared to CO2, as demonstrated by the difference in Baker R (2001) Future directions of membrane
gas-separation technology. Membr Technol 138:510
critical temperature. Table 1 provides the H2S
Lu GQ, da Costa D, Duke JC, Giessler M, Socolow R,
permeability and H2S/CO2 selectivity for a range Williams RH, Kreutz TG (2007) Inorganic membranes
of polymeric membranes. The study of porous for hydrogen production and purication: a critical
membranes for acidic gas removal is limited, review and perspective. J Colloid Interface Sci
314:589603
Scholes CA, Kentish SE, Stevens GW (2009) Effect of
minor components in carbon dioxide capture using
Acid Gas: Effect on Membrane Properties, polymeric gas separation membranes. Sep Purif Rev
Table 1 H2S permeabilities (barrier) and H2S/CO2 selec- 38:144
tivity in a range of polymeric membranes
H2S permeability H2S/CO2
Ethyl cellulose 320 3.8
Nylon 0.34 3.86
PDMS 5,100 1.6 A-CNT
Polyethylene 36 2.84
PTMSP 21,400 1.2 Aligned Carbon Nanotube
6 Activity Coefficient
cathodic overpotential (Zhang 2008). In this case,

Activity Coefficient overpotential is dened as the extra potential over
the equilibrium value that must be applied to cause
Dmitri Bessarabov an electrode reaction at a certain rate. Thus, an
Faculty of Engineering, DST HySA Infastructure electrocatalyst with high activity is required to
Center of Competence, North-West University, minimize the overpotential.
Potchefstroom, South Africa According to the Sabatier principle (Bligaard
et al. 2004), a volcano-shaped curve is obtained
when the activity of a catalyst for ORR is plotted
Chemical Thermodynamics as a function of a variable that relates to the ability
of the electrocatalyst surface to form chemical
Depending on a specic area of fundamental sci- bonds with oxygen, reaction intermediates,
ence and technology, the term activity may have and/or products (Nrskov et al. 2004). For exam-
a different meaning. For example, in chemical ple, it was shown that the catalytic activity of a
thermodynamics, activity (a) is well known as a single Pt-based electrocatalyst toward ORR
convenient variable to represent data, obtained depends strongly on such variables as the O2
experimentally or calculated theoretically, for the adsorption energy, the dissociation energy of the
values of chemical potentials of components in OO bond, and the binding energy of OH on the
real chemical solutions. For the i th component catalyst surface. The electronic structure of the Pt
of a mixture, activity ai is dened as follows: catalyst (Pt d-band vacancy) and the PtPt
mi m0i RTlnai , where mi is the chemical interatomic distance (geometric effect) can
potential of a component and m0i is the chemical strongly affect these variables. Theoretical calcu-
potential of that component in standard state. lations on O2 and OH binding energy on several
R is the gas constant and T is the thermodynamic metals have predicted that Pt should have the
temperature. The numerical value of the highest electrocatalytic activity among other
activity depends on the choice of a standard metals with the ORR activity of Pt > Pd > Ir >
state, as it describes the difference between an Rh, which is in agreement with experimental
actual chemical potential and a standard chemical results. Calculations have also predicted that
potential. PtM (M = Fe, Co, Ni, etc.) bimetallic alloys
should have higher catalytic activity than pure
Pt, which has again been proven by experiments
Fuel Cells and Electrocatalysis (Nrskov et al. 2004).
In hydrogen fuel cell science and technology, activ-

ity is often referred to as the measure of efciency Practical Measurements
of electrocatalysts. Electrocatalytic processes in
fuel cells occur on the surface of electrodes, and Practically, electrocatalytic activity of various
one of the main challenges is to explain, predict, electrocatalysts can be compared by measuring
and improve electrocatalytic activity of various electrical currents produced at a given
electrode materials. In the area of low temperature overpotential. The electrocatalyst that generates
PEM (proton exchange membrane, also called higher current at lower overpotential is generally
polymer electrolyte membrane) fuel cells, catalytic a better electrocatalyst.
activity is often used to compare efciency of The rotating disk electrode (RDE) technique is
electrocatalysts toward kinetics of the oxygen often used to measure ORR activity of
reduction reaction (ORR) occurring at the cathode. electrocatalysts for hydrogen PEM fuel cell appli-
ORR is known to have slow kinetics due to the cations. Typically, current density j is measured at
reaction having multiple steps and difculties in 0.9 V at the positive voltage sweep at a certain
breaking the OO bond, which leads to a larger electrode rotating speed. The measured current
Acute Kidney Injury (AKI) 7
density j is expressed by the following equation: Acute kidney injury (AKI) or acute renal failure
j j j j , where jk is kinetic current density,
1 1 1 1 indicates kidney injury from less severe forms of
k d f
jd is hydrodynamic diffusion limiting current den- injury to more advanced injury when acute kidney A
sity, and jf is lm diffusion limiting current den- failure may require renal replacement therapy.
sity. RDE experiments should be designed to Acute kidney failure occurs when kidneys loss
maximize jf. In that case, 1j j1 j1 and after their functions that include excretion of wastes,
rearrangement jk j jd =jd j.
k d
acid-base homeostasis, osmolality regulation,
For ORR the value of jd is normally taken at blood pressure regulation, and hormone secretion
0.4 V during RDE experiments. (Waikar et al. 2008). When kidneys lose their
Once jk is known, activity is often reported in ltering ability, dangerous levels of wastes may
practical values as mass activity (A/mg Pt) by accumulate, resulting in an increasingly toxic
normalization to the value of Pt loading of the blood stream and uid overload. Acute kidney
sample on the disk electrode. failure requires replacement therapy that includes:
Another practically important way of reporting hemodialysis, peritoneal dialysis, hemoltration,
the electrocatalyst activity is to use area-specic hemodialtration, and renal transplantation.
activity. The area-specic activity, expressed in Dialysis treatment involves the diffusion of
the following units, uA/cm2 Pt, is reported by solutes through a semipermeable membrane
normalization of jk to the electrochemical surface driven by a concentration gradient. For the
area (ECSA) of the catalyst. treatment of uremic patients, an articial
membrane (hemodialysis) or a biological mem-
brane (peritoneal dialysis) can be utilized. The
References membrane-based treatments involve pumping
patients blood from the body through a mem-
Bligaard T, Nrskov JK, Dahl S, Matthiesen J, Christensen brane device that acts as an articial kidney by
CH, Sehested S (2004) The Brnstedevanspolanyi separating waste products including toxins and
relation and the volcano curve in heterogeneous catal-
ysis. J Catal 224:206217 excess water from the blood. The therapies for
Nrskov JK, Rossmeisl J, Logadottir A, Lindqvist L, the treatment of patients with acute renal failure
Kitchin JR, Bligaard T, Jonsson H (2004) Origin of are not limited to dialysis but rather include other
the overpotential for oxygen reduction at a fuel-cell modalities. Hemoltration has been developed
cathode. J Phys Chem B 108:1788617892
Zhang J (ed) (2008) PEM fuel cell electrocatalysts and to overcome the reduced efcacy of diffusion for
catalyst layers. Springer, London molecules with middle and large MW. In this
process the driving force is a hydrostatic pressure
Further Reading gradient that induces a large ux of water across
Vielstich W, Lamm A, Gasteiger HA (eds) (2003) Handbook the membrane from blood side to the ltrate side.
of fuel cells, vol 2, Electrocatalysis. Wiley, New York As a result the rate of solute removal is propor-
tional to the applied pressure. Hemodialtration
is the combination of hemodialysis and
hemoltration processes. The high rate of ultral-
Acute Kidney Injury (AKI)
tration is combined with the efcient diffusion
overcoming the disadvantages related to each sin-
Loredana De Bartolo
gle treatment.
Institute on Membrane Technology, National
Research Council of Italy, ITM-CNR, Rende,
Italy
References
Synonyms Waikar SS, Liu KD, Chertow GM (2008) Diagnosis,
epidemiology and outcomes of acute kidney Injury.
Membrane articial organs Clin J Am Soc Nephrol 3:38443861
8 Addition Norbornene Polymer-Based Membrane Materials
et al. 2010): permeability coefcients increase

Addition Norbornene Polymer-Based for penetrants with larger size or molecular mass
Membrane Materials (e.g., butane). This makes them important candi-
dates for membranes for separation of natural and
Yuri Yampolskii associated petroleum gas, especially because their
A.V. Topchiev Institute of Petrochemical gas permeation parameters are relatively stable in
Synthesis, Russian Academy of Sciences, time in comparison with highly permeable
Moscow, Russia polyacetylenes. These polymers are soluble in
various organic solvents, so high-ux composite
membranes (permeance of about 515 m3(STP)/
Substituted norbornenes and tricyclononenes can m2 h bar) can be prepared from them.
be polymerized according to the addition scheme,
that is, via opening of the double bonds of these
cycloolens (Finkelshtein et al. 2011; Yampolskii References
2010).
The resulting glassy polymers have very rigid E. Finkelshtein, M. Bermeshev, M. Gringolts,
L. Starannikova, Y. Yampolskii, Substituted
main chains and high glass transition tempera-
polynorbornenes as promising materials for gas sepa-
tures. The gas permeation parameters of the ration membranes. Russ Chem Rev 80, 341361 (2011)
obtained polymers strongly depend on the M. Gringolts, M. Bermeshev, Y. Yampolskii,
substituents in the monomer molecules. L. Starannikova, V. Shantarovich, E. Finkelshtein,
New high permeable addition poly(tricyclononenes)
Thus, nonsubstituted addition polynorbornene
with Si(CH3)3 side groups. Synthesis, gas permeation
(PNB) is a relatively low-permeability polymer parameters, and free volume. Macromolecules 43,
(P(O2) = 11 Barrer), though when 71657172 (2010)
substituents are introduced permeability Y. Yampolskii, Norbornene polymers as materials for
membrane gas separation, in Comprehensive mem-
increases markedly (Fig. 1). In particular, brane science and engineering, ed. by E. Drioli,
methyl-substituted PNB has P(O2) = 89 Barrer, L. Giorno, vol. 1 (Elsevier, Oxford, 2010), pp. 131146
while for the polymer with bulkier, SiMe3 substit-
uent, the permeability is further strongly enhanced
(P(O2) = 980 Barrer) (Yampolskii 2010). SiMe3-
substituted PNB and poly(tricyclononenes) reveal
large free volume according to the positron anni- Adhesion to Membrane Surfaces
hilation lifetime spectroscopy and solubility-
controlled selectivity in separation of hydrocar- Li-Ping Zhu
bons(Finkelshtein et al. 2011; Gringolts MOE Key Laboratory of Macromolecular
Synthesis and Functionalization, Department of
Polymer Science and Engineering, Zhejiang
R2 University, Hangzhou, China
n
n
R1 Adhesion to membrane surfaces generally

describes the sticking or holding fast of some
substances such as paints, cells, platelets, bacteria,
R1 R2
proteins, etc. on membrane surfaces when they are
Addition Norbornene Polymer-Based Membrane
in contact. Usually, the intermolecular interaction
Materials, Fig. 1 Addition Norbornene Polymer-based is responsible for the adhesion of substances on
Membrane Materials membrane surfaces (Huang et al. 1999). The
Adsorption onto Membrane Surfaces 9
adhesion can be divided into several categories, Cross-References

including mechanical adhesion, chemical adhe-
sion, dispersive adhesion, electrostatic adhesion, Adsorption onto Membrane Surfaces A
and diffusive adhesion according to the forces that Bioadhesion to Membrane Surface
cause adhesion. The adhesion is a stronger inter-
action between unlike substances than adsorption
References
(Jacek et al. 1999). In the ltration operation using
a membrane, the adsorption and adhesion of Huang YC, Zhang ES, Ying K (1999) Progress of bacterial
foulants such as colloids, particles, proteins, bac- adhesion on biomaterials. Foreign Med Sci (Biomed
teria, and microorganisms on membrane surfaces Eng Fascicle) 22(3):148153
result often in membrane fouling that induces the Jacek D, Robert K, Jarosaw K, Hanna MK (1999) Cell
adhesion to polymeric surfaces: experimental study and
severe ux loss and the increase of transmem- simple theoretical approach. J Biomed Mater Res
brane pressure. Membrane fouling can be divided 47(2):234242
into reversible fouling and irreversible fouling. Jonathan AB, Amy EC (2004) Colloidal adhesion to
Reversible fouling is often attributed to the hydrophilic membrane surfaces. J Membr Sci
241:235248
adsorption of foulants on membrane surface, and
it can be recovered by physical or chemical mem-
brane cleaning. While irreversible fouling is usu-
ally generated from the rm adhesion of foulants
on membrane surface, and it is difcult to be Adsorption onto Membrane Surfaces
removed even if a thorough membrane cleaning
is done (Jonathan and Amy 2004). Therefore, the Li-Ping Zhu
adhesion of foulants on membrane surface often MOE Key Laboratory of Macromolecular
brings a permanent deterioration of membrane Synthesis and Functionalization, Department of
performance. In such a case, the fouled membrane Polymer Science and Engineering, Zhejiang
modules have to be replaced to continue a normal University, Hangzhou, Zhejiang, China
ltration operation.
In blood-contact membrane separation, for the
blood incompatible membrane materials, the plate- Adsorption onto membrane surface is the accu-
lets in blood tend to accumulate and adhere onto mulation of gases, liquids, or solutes to form a thin
membrane surface due to a biological immune lm on membrane surface. In adsorption, the
reaction. As a result, thrombus formation may adsorbate is attracted to the surface of adsorbent,
form and further result in some serious complica- but does not enter its molecular spaces. As a
tions. Therefore, the blood-contact membranes surface phenomenon, adsorption is a consequence
such as hemodialysis membranes often need be of surface energy. Adsorption is similar to adhe-
modied to prevent the adhesion of platelets on sion and absorption, but there exist differences
membrane surface during blood ltration. In mem- between them. Generally, compared to adhesion,
brane science, the experiment of platelet adhesion adsorption is looser interaction between unlike
on membrane surface is usually employed to eval- substances. Adsorption is a surface-based pro-
uate the hemocompatibility of a blood-contact cess, while absorption involves the whole volume
membrane. In biomedical application, the adhesion of the material. The exact bonding nature of
of microbes on membrane surface often leads to the adsorption depends on the details of the species
bacterial infection of patients and the deterioration involved, but the adsorption process is generally
of membrane materials. Therefore, the control of classied as physical adsorption and chemical
microbial infection is a very important issue in adsorption. The physical adsorption is attributed
biomedical membrane separation. to the non-covalent interaction such as van der
10 Advanced Oxidation Process (AOP) by Membrane Reactors
Waals force, electrostatic attraction, etc. The

chemical adsorption features the formation of Advanced Oxidation Process (AOP)
covalent bonding between adsorbate and adsor- by Membrane Reactors
bent (Ferrari et al. 2010).
Commonly, in liquid ltration using a mem- Kwang-Ho Choo
brane, adsorptions onto membrane surface Department of Environmental Engineering,
describe the accumulation and attachment of sol- Kyungpook National University, Daegu,
utes in liquid on membrane surface driven by the Republic of Korea
minimization of interfacial energy. In water phase
ltration, the adsorption of foulants such as pro-
teins, colloids, microorganisms, etc., on mem- Advanced oxidation process (AOP) by membrane
brane surfaces and membrane pore walls is often reactors is a synergistic hybrid of hydroxyl radical
responsible to membrane fouling that induces a reaction and membrane ltration in a single reac-
severe loss of membrane permeability and a tor. Advanced oxidation process (AOP) is the
change in membrane separation characteristics. uncontrolled, destructive action of hydroxyl radi-
The adsorption of foulants can be removed cals generated from ozone, H2O2, ultraviolet radi-
(desorption) by chemical or physical washing if ation, Fenton reaction, electrolysis, sonolysis, and
the interaction between foulants and surfaces is photocatalysis (e.g., TiO2). AOP is most widely
weak. As a result, the membrane performances used for the decomposition of nonbiodegradable
can be recovered partially or fully. In blood- organic contaminants present in industrial and
contact membrane separation, protein municipal wastewaters. It can be also applied for
adsorption on membrane surface often leads to the degradation of micro-contaminants and taste-
the activation of complement system and blood and-odor causing compounds in drinking water
coagulation and further results in serious compli- sources, such as pesticides, pharmaceuticals,
cations of patients. Protein adsorption onto mem- endocrine disruptors, microcystins, geosmin,
brane surface is inuenced by the surface etc. Although micro-/ultraltration membranes
characteristics such as wettability, surface are almost complete barriers for particulate matter
chemistry, morphology, surface charge, and pathogens in water, they are limited in remov-
etc. Therefore, surface modication of the blood- ing dissolved organic compounds. Thus, the cou-
contact membranes to depress the adsorption of pling of membrane ltration with AOP in a reactor
proteins on membrane surface has been a provides multifunctionalities and exibilities for
pursuit of membrane researchers (Rana and water and wastewater treatment systems.
Matsuura 2010). Advanced oxidative membrane reactors are
similar to biological membrane reactors (i.e.,
membrane bioreactors) in their concept except
that biological oxidation is replaced by chemical
Cross-References oxidation. Due to the incompatibility of hydroxyl
radicals and membrane materials (e.g., polymeric
Adhesion to Membrane Surfaces membranes), however, only a few examples of
process congurations are developed and intro-
References duced, such as ozonated membrane reactor (Heng
et al. 2007), photocatalytic membrane reactor
Ferrari L, Kaufmann J, Winnefeld F, Plank J (2010) Inter- (Choo et al. 2008), and electrochemically active
action of cement model systems with superplasticizers ltration device (Vecitis et al. 2011). Figure 1
investigated by atomic force microscopy, zeta potential, shows the schematics of such advanced oxidative
and adsorption measurements. J Colloid Interface Sci
347:1524
membrane reactors. The ozone membrane reactor
Rana D, Matsuura T (2010) Surface modications for consists of ceramic tubular membranes made of
antifouling membranes. Chem Rev 110:24482471 alumina and zeolite which are chemically durable
Advanced Oxidation Processes (AOPs) and Membrane Operations 11
Advanced Oxidation Process (AOP) by Membrane 2007), (b) photocatalytic membrane reactor (Choo et al.
Reactors, Fig. 1 Examples of advanced oxidative mem- 2008), and (c) electrochemically active ltration device
brane reactors: (a) ozone membrane reactor (Heng et al. (Vecitis et al. 2011)
while degrading a fraction of total organic carbon Vecitis CD, Schnoor MH, Rahaman MS, Schiffman JD,
by ozone. The photocatalytic membrane reactor Elimelech M (2011) Electrochemical multiwalled car-
bon nanotube lter for viral and bacterial removal and
with suspended TiO2 particles in aqueous phase inactivation. Environ Sci Technol 45:36723679
enables polymeric membranes to be used because
photocatalysis only occurs on the surface of TiO2
photocatalysts where UV light bombards. The
membrane permeability is enhanced with contin-
uous photocatalysis of organic substances; other-
Advanced Oxidation Processes
wise they can foul the membrane severely. The (AOPs) and Membrane Operations
fabrication and use of TiO2 membranes for
Kwang-Ho Choo
photocatalytic membrane reactors are of great
interest, but UV light radiation is a major huddle Department of Environmental Engineering,
to be resolved. When a direct current of 23 V was Kyungpook National University, Daegu,
Republic of Korea
applied onto multiwall carbon nanotubes loaded
on a 5-mm polytetraoroethylene lter, substantial
viral and bacterial removals are achieved by elec-
Advanced oxidation processes (AOPs) and
trolysis in addition to ltration. Further process
developments on AOP membrane reactors are membrane operations are composed of
expected with novel membrane materials and pro- advanced oxidation followed by membrane ltra-
tion, where the former decomposes dissolved
cess designs.
organic matter and the latter removes particulate
matter. Advanced oxidation process (AOP) is the
uncontrolled, destructive action of hydroxyl radi-
References
cals generated from ozone, H2O2, ultraviolet radi-
Choo K-H, Tao R, Kim M-J (2008) Use of a photocatalytic ation, Fenton reaction, electrolysis, sonolysis, and
membrane reactor for the removal of natural organic photocatalysis (e.g., TiO2). AOP is most widely
matter in water: effect of photoinduced desorption and used for the decomposition of nonbiodegradable
ferrihydrite adsorption. J Membr Sci 322:368374
Heng S, Yeung KL, Djafer M, Schrotter J-C
organic contaminants present in industrial and
(2007) A novel membrane reactor for ozone water municipal wastewaters. It can be also applied
treatment. J Membr Sci 289:6775 for the degradation of micro-contaminants and
12 Advanced Oxidation Processes (AOPs) and Membrane Operations
with ozone dosage (1.7 mgO3/mg TOC) without ozone dosage

1.2
a 220 b without electrolysis
200 1.0 with electrolysis (2.0 V)
Normalized Flux (J/J0)

180 breakdown with electrolysis (4.0 V)
permeability [l/(m2 h bar)]
160 in ozone 0.8

generator
140
120 0.6
100
80 0.4
60
40 0.2
20
0 0.0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 1.0 1.2 1.4 1.6 1.8 2.0
filtration time [h] Concentration Factor
Advanced Oxidation Processes (AOPs) ozonation plus ultraltration (Schlichter et al. 2004); (b)
and Membrane Operations, Fig. 1 Membrane ux electrolysis plus microltration (Park et al. 2013)
enhancement with advanced oxidation pretreatments: (a)
taste- and odor-causing compounds in drinking and ultimately mineralizing organic matter. Mem-
water sources, such as pesticide, pharmaceuticals, brane operations after electrolysis achieve further
endocrine disruptors, microcystins, geosmin, removal of particles that can be hardly oxidized.
etc. This combined process can compensate the A process conguration that consists of advanced
weak point of each process, but the oxidants left oxidation (e.g., Fenton oxidation) followed by a
over should be quenched before their contact with membrane bioreactor is also an attractive option
polymeric membranes. This is because oxidants for the treatment of industrial wastewater
can deteriorate the membrane lifetime or integrity containing recalcitrant organic chemicals (e.g.,
signicantly due to the oxidation of the mem- dyes) (Feng et al. 2010). This combination can
brane. Polyamide membranes can be damaged be applied for the minimization of biological
even by the attack of chlorine and thereby, their sludge generation while disintegrating excessive
physical properties and ltration performance are activated sludge (He and Wei 2010). To take
signicantly impaired. Reducing agents, such as advantage of prolonged oxidation, inorganic
sodium sulte, are mostly used for oxidant membranes (e.g., ZrO2 and Al2O3 membranes),
quenching. which are tolerant to strong oxidants, can be used
A series of oxidation (e.g., ozonation) and instead of polymeric membranes. Particularly,
micro-/ultraltration processes are designed and advanced oxidation treatment prior to membrane
used for the treatment of drinking water which ltration helps to reduce membrane fouling
contains trace organic contaminants as well as caused by organic substances (Fig. 1) (Park
metal species (e.g., Fe2+, Mn2+) (Hyung et al. 2013; Park 2002; Schlichter et al. 2004).
et al. 2000; Choo et al. 2005). Ozone can facilitate
the oxidation of reduced forms of such metal ions
and so help to precipitate out. The precipitates can
References
be readily rejected by porous membranes. Elec- Choo K-H, Lee H, Choi S-J (2005) Iron and manganese
trolysis in combination with microltration is removal and membrane fouling during UF in conjunc-
developed and tested for the treatment of munic- tion with prechlorination for drinking water treatment.
ipal wastewater organic matter with simultaneous J Membr Sci 267:1826
Feng F, Xu Z, Li X, You W, Zhen Y (2010) Advanced
hydrogen production (Park et al. 2013). Electro-
treatment of dyeing wastewater towards reuse by the
chemical oxidation at a voltage of >2.3 V enables combined Fenton oxidation and membrane bioreactor
hydroxyl radicals to be generated while oxidizing process. J Environ Sci 22:16571665
Aerobic Membrane Bioreactor 13
He M-h, Wei C-h (2010) Performance of membrane biore- contaminants (carbon and ammonia) and separate
actor (MBR) system with sludge Fenton oxidation pro- solids from the treated municipal or industrial
cess for minimization of excess sludge production.
J Hazard Mater 176:597601 wastewater. Carbon is removed by microorgan- A
Hyung H, Lee S, Yoon J, Lee C-H (2000) Effect of isms that metabolize the carbon in the presence of
preozonation on ux and water quality in ozonation- dissolved oxygen for microbial growth and respi-
ultraltration hybrid system for water treatment. Ozone ration (organic carbon reduced to carbon dioxide).
Sci Eng 22:637652
Park YG (2002) Effect of ozonation for reducing Ammonia is removed through ammonia oxidation
membrane-fouling in the UF membrane. Desalination (nitrication). Nitrication is a microbially miti-
147:4348 gated reduction process that occurs in an aquatic
Park H, Choo K-H, Park H-S, Choi J, Hoffmann MR environment that contains moderate to high con-
(2013) Electrochemical oxidation and microltration
of municipal wastewater with simultaneous hydrogen centrations of ammonia and dissolved oxygen and
production: inuence of organic and particulate matter. low concentrations of organic carbon.
Chem Eng J 215216:802810 In submerged MBRs (Fig. 1), microporous
Schlichter B, Mavrov V, Chmiel H (2004) Study of a (microltration (MF) or ultraltration
hybrid process combining ozonation and
microltration/ultraltration for drinking water produc- (UF) ( Ultraltration)) membranes are
tion from surface water. Desalination 168:307317 immersed in a bioreactor, and water is ltered
( Permeate) through the membranes using vac-
uum; suspended solids are retained in the system;
and high levels of treatment (including nutrient
removal) can be achieved (Judd 2006). The MBR
Aerobic Membrane Bioreactor replaces the two-stage conventional activated
sludge process (biotreatment and clarication)
A. Achilli1 and R. W. Holloway2 with a single, integrated process ( Wastewater
1
Environmental Resources Engineering Treatment in Membrane Bioreactors). The advan-
Department, Humboldt State University, Arcata, tages of MBRs over conventional treatment
CA, USA have been thoroughly reviewed (Stephenson
2
Department of Civil and Environmental et al. 2000), and they include product consistency,
Engineering, Colorado School of Mines, Golden, reduced footprint, reduced sludge production,
CO, USA and nearly complete suspended solid separation
from the efuent. Additionally, MBR efuent
may be suitable for use as irrigation water, as
Synonyms process water, or as a pretreatment for potable
reuse applications (Membrane Bioreactors for
Cross-ow membrane bioreactor; Membrane bio- Reuse; Potable Water Production) (Lawrence
reactor; Microltration membrane bioreactor; et al. 2002).
Submerged membrane bioreactor; Ultraltration However, the establishment of membrane bio-
membrane bioreactor reactor technology (Anaerobic Membrane Biore-
actor; Attached Growth Membrane Bioreactor;
Membrane Bioreactors (MBR) for Plants; Sub-
Introduction merged Membrane Bioreactor) has been slower
than expected because decision-makers view
Aerobic membrane bioreactors (MBRs) are one of MBRs as high risk and costly compared to
the leading technologies to achieve sustainability conventional technology (Judd 2006). To date,
in wastewater treatment through reuse, decentral- MBRs have been used to treat municipal and
ization, and low energy consumption (Fane and industrial wastewater where water reuse is
Fane 2005; Fawehinmi et al. 2005). In aerobic desired, a small footprint is required, or stringent
MBRs, aerated activated sludge is coupled with discharge standards exist (Kang et al. 2007; Yang
membrane process to remove dissolved et al. 2006).
14 Aerobic Membrane Bioreactor
Aerobic Membrane Membrane

Bioreactor, Cassette
Fig. 1 Schematic Membrane Activated
representation of a Channels Sludge
submerged aerobic MBR
system
Product
Water
Membranes
Current Limitations pumped in the opposite direction through the

membrane, effectively removing most of the
One of the major limitations to widespread appli- reversible fouling ( Backwashing). The ef-
cation of MBR technologies is to control mem- ciency of backwashing has been studied in
brane fouling with modest energy and chemical detail, and the key parameters have been
input ( Irreversible Fouling; Reversible Foul- found to be frequency, duration, and intensity
ing) (Le-Clech et al. 2006). Membrane fouling ( Backwashing Frequency) (Bouhabila
(Biological Fouling; Cake Layer; Inorganic et al. 2001; Psoch and Schiewer 2005, 2006).
Scaling) has been investigated from various per- During membrane relaxation, permeate suction
spectives, including the causes, characteristics, is stopped, and the back transport of foulants is
and mechanisms of fouling and methods, to pre- naturally enhanced as reversibly attached foulants
vent or reduce membrane fouling (Achilli diffuse away from the membrane surface.
et al. 2011; Judd 2005; Le-Clech et al. 2006; Membrane backwashing and relaxation are
Meng et al. 2009; Wang and Wu 2009). Fouling regularly used for tubular and hollow ber mem-
markedly affects membrane cleaning (Membrane branes to control fouling (Bouhabila et al. 2001;
cleaning, 369690) and replacement intervals, sys- Hong et al. 2002; Psoch and Schiewer 2005,
tem productivity, and membrane integrity; all of 2006; Smith et al. 2005). This is not the case for
which are factors that affect energy requirements at-sheet membranes that cannot be backwashed
and costs (Judd 2006; Le-Clech et al. 2006). due to their inability to withstand pressure in the
opposite direction of the operating ow; for this
reason, relaxation is used to control the fouling of
Operation these membranes (Le-Clech et al. 2006). Regard-
less of the membrane conguration, chemicals
In order to operate conventional MBRs at constant must be used at regular intervals to enhance phys-
ux, physical membrane-cleaning techniques are ical cleaning (Le-Clech et al. 2005).
utilized; they include air scouring, backwashing,
relaxation, or a combination of the three,
depending on the membrane conguration Cross-References
(hollow bers, at sheet, or tubular). Air scouring
is required for submerged MBR congurations to Backwashing
gas lift fresh sludge through the membrane bundle Backwashing Frequency
or cassette and to scour solids from the membrane Industrial/Tannery Wastewater Treatment
surface. During backwashing, permeate is Using Membrane Bioreactors
Affinity Membranes 15
Inorganic Scaling Smith PJ, Vigneswaran S, Ngo HH, Ben-Aim R, Nguyen H

Irreversible Fouling (2005) Design of a generic control system for optimising
back ush durations in a submerged membrane hybrid
Permeate reactor. J Membr Sci 255:99106 A
Reversible Fouling Stephenson T, Judd S, Jefferson B, Brindle K (2000) Mem-
Submerged Membrane Bioreactor brane bioreactors for wastewater treatment. IWA Pub-
Ultraltration (UF) lishing, London
Wang ZW, Wu ZC (2009) A review of membrane fouling
in MBRs: characteristics and role of sludge cake
formed on membrane surfaces. Sep Sci Technol
References 44(15):35713596
Yang W, Cicek N, Ilg J (2006) State of the art of membrane
Achilli A, Marchand EA, Childress AE bioreactors: worldwide research and commercial appli-
(2011) A performance evaluation of three membrane cations in north America. J Membr Sci 270:201211
bioreactor systems: aerobic, anaerobic, and attached-
growth. Water Sci Technol 63:29993005
Bouhabila EH, Ben-Aim R, Buisson H (2001) Fouling
characterisation in membrane bioreactors. Sep Purif
Technol 2223:123132
Fane AG, Fane SA (2005) The role of membrane technol- Affinity Membranes
ogy in sustainable decentralized wastewater systems.
Water Sci Technol 51(10):317325
Fawehinmi F, Lens P, Stephenson T, Rogalla F, Jefferson
Francesca Militano1,2 and Lidietta Giorno1
1
B (2005) The inuence of operating conditions on Institute on Membrane Technology, National
extracellular polymeric substances (EPS), soluble Research Council of Italy, ITM-CNR, Rende,
microbial products (SMP) and bio-fouling in anaerobic Italy
membrane bioreactors. In: IWA specialized conference 2
on water environment
Univeristy of Calabria, Rende, Italy
Hong SP, Bae TH, Tak TM, Hong S, Randall A (2002)
Fouling control in activated sludge submerged
hollow ber membrane bioreactors. Desalination Introduction
143:219228
Judd S (2005) Fouling control in submerged membrane
bioreactors. Water Sci Technol 51(67):2734 Afnity separation methods rely on a molecular
Judd S (2006) The MBR book: principles and applications recognition phenomenon between species.
of membrane bioreactors in water and wastewater treat- A molecule, known as the ligand, is permanently
ment. Elsevier Science, Oxford
Kang Y-T, Cho Y-H, Chung E-H (2007) Development of
bounded onto an inert matrix and specically
the wastewater reclamation and reusing system with a recognizes the molecule of interest, known as
submerged membrane bioreactor combined the ligate, that can be separated. The ligand can
bio-ltration. Desalination 202:6876 be a naturally occurring molecule, an engineered
Lawrence P, Adham S, Barro L (2002) Ensuring water
re-use projects succeed institutional and technical
macromolecule, or a synthetic molecule linked to
issues for treated wastewater re-use. Desalination the matrix by covalent coupling. The ligand-ligate
152:291298 interaction is selective and reversible, enabling
Le-Clech P, Fane A, Leslie G, Childress A (2005) MBR the separation and ne purication of biological
focus: the operators perspective. Filtr Sep 42(5):2023
Le-Clech P, Chen V, Fane TAG (2006) Fouling in mem- substances such as proteins, peptides, and nucleic
brane bioreactors used in wastewater treatment. acids on the basis of its individual chemical struc-
J Membr Sci 284:1753 ture or biological function (Wilson and Poole
Meng FG, Chae SR, Drews A, Kraume M, Shin HS, Yang 2009).
FL (2009) Recent advances in membrane bioreactors
(MBRs): membrane fouling and membrane material.
Among the separation techniques based on the
Water Res 43(6):14891512 afnity method, afnity chromatography is the
Psoch C, Schiewer S (2005) Critical ux aspect of air most widely used. Due to the limitation associated
sparging and backushing on membrane bioreactors. to the traditional afnity chromatography with
Desalination 175:6171
Psoch C, Schiewer S (2006) Resistance analysis for
porous bead-packed columns (i.e., limited ow
enhanced wastewater membrane ltration. J Membr rate by pore diffusion), the membrane-based sep-
Sci 280:284297 aration technique is gaining an increasing
16 Affinity Membranes
importance. Indeed, afnity membranes combine density of functional groups and increase the
the specicity of afnity adsorption (of the com- mechanical strength, composite cellulose mem-
mon afnity resins) with the high productivity branes have been also prepared by chemical
associated with ltration membranes. They pro- grafting with acrylic polymers.
vide low pressure separation systems without dif- Another suitable membrane material is polyvi-
fusional limitation, as the mass transfer is mainly nyl alcohol, thanks to its hydrophilicity and bio-
governed by convection (Klein 2000). Micropo- compatibility. Like cellulose, it contains hydroxyl
rous membranes, both in at sheet and hollow groups that can be easily activated.
ber conguration, coupled with biological or Polyamide and nylon have been also used for
biomimetic ligands have been used as afnity the preparation of afnity membranes, thanks to
membrane chromatography supports (Zou their good mechanical and chemical stability.
et al. 2001). Nylon membranes have a low concentration of
amino groups to serve as functional groups for
ligand coupling. Also, due to the hydrophobic
Affinity Membrane Preparation surface, the membrane presents a high nonspecic
adsorption of biomolecules during afnity sepa-
The realization of afnity membranes usually ration process. Therefore, the hydrolysis of the
involves three steps: (1) preparation of the basic membrane is generally performed to increase the
membrane, (2) activation (functionalization) of density of reactive groups and to prevent electro-
the basic membrane, and (3) coupling of afnity static interaction with proteins. To overcome these
ligands to the activated membrane. problems, composite nylon membranes have been
also prepared.
Membrane Material Other materials that have been used for afnity
The membrane materials should possess some membranes are poly(methyl methacrylate),
characteristics: (i) hydrophilicity to minimize the polysulfone and its derivative, polycaprolactam,
nonspecic adsorption of bioactive species, polyvinylidene diuoride, poly(ether-urethane-
(ii) chemical and physical stability under harsh urea), and inorganic materials such as glass.
conditions used during ligand coupling and
ligand-ligate complex formation, (iii) large sur- Membrane Activation and Ligand Coupling
face area relative to membrane volume, If the basic membrane does not possess the func-
(iv) biocompatibility when the membrane is used tional groups for ligand coupling, it can be
for blood treatment, and (v) presence of functional activated.
groups required for the coupling of the ligands In Table 1 are reported the commonly used
(such as OH, NH, SH, COOH). activation agents with respect to the functional
The commercially available materials used for groups present on the basic membrane.
afnity membrane preparation include organic The coupling of the ligand directly on the
(natural or synthetic polymer), inorganic, and activated membranes may result in low binding
composite materials. specicity due to the low steric availability
Among the rst materials used for afnity (in particular when the ligand is a small molecule).
membrane preparation, cellulose and its deriva- This problem is generally overcome by the
tives (regenerated cellulose, cellulose acetate) introduction of a spacer molecule to the mem-
were the most common. These present a hydro- brane prior to attaching the ligand, improving
philic and biocompatible surface, low nonspecic the ligand accessibility for the molecule to
adsorption, and abundant reactive hydroxyl be separated. Spacer arms are bifunctional mole-
groups that can be easily activated by different cules able to react with both the membrane and
strategies for the ligand coupling. To improve the ligand.
Affinity Membranes 17
Affinity Membranes, Table 1 Some commonly used activation agents for affinity membrane preparation
Activation agent Structure Functional groups of the basic membrane
Cyanogen bromide
N C Br
OH A
Carbonyldiimidazole O OH, NH2
N N
N N
2-Amino-4,6-trichloro-s-triazine Cl OH
N N
Cl N NH2
Glutaraldehyde OH, SH, NH2

O O
Divinyl sulfone O OH, SH

CH2
S
H2C O
Epichlorohydrin OH, NH2

O
Cl
Affinity Ligands Although the biospecic ligands possess high

specicity for proteins, they have some limitation
Afnity ligands can be classied into biospecic for large-scale application due to their poor stabil-
and pseudo-biospecic ligands (Klein 1991). ity and high price.
Biospecic ligands are biomolecules such as The alternative approach to biospecic ligands
antibodies, antigens, and proteins A and G that involves the use of pseudo-biospecic ligands.
show specicity for only one complementary These are usually molecules with higher chemical
biomolecule. and physical stability than biomolecules. Pseudo-
Because of their selectivity, biomolecules have biospecic ligands can be distinguished by bio-
been the most widely used afnity ligands on logical (amino acids, specially histidine, lysine,
afnity membrane separation technology. One of tryptophan) or non-biological molecules (dyes,
the most common applications is the use of chelated metal ions).
immobilized monoclonal antibodies for The working principle of pseudo-biospecic
immunoafnity separation. Another important ligands relies on the complementarity of structural
example is membranes with covalently coupled features of ligand and ligate rather than on a
protein A or protein G for immunoglobulin puri- biological function. Immobilized dyes have
cation from plasma, serum, or cell culture been found to act as afnity ligands for a wide
supernatants. variety of biological molecules. For example,
18 Affinity Membranes
Affinity Membranes, Table 2 Some examples of the use of affinity membranes for isolation and purification of
biomolecules
Membrane Ligand Ligate Application
Cellulose and regenerated Cibacron Blue F3GA Alkaline phosphatase Purication
cellulose Protein A/G IgG Purication
Histidine Endotoxin Removal of
endotoxin
Poly(ethylene-co-vinyl Histidine IgG Purication
alcohol)
Nylon Cibacron Blue Glucose-6-phosphate Purication
dehydrogenase
Protein G IgG Purication
Histidine Endotoxin Removal of
endotoxin
Polyvinylidene diuoride Protein A IgG Purication
(PVDF)
Polysulfone (PS) Trypsin Trypsin inhibitor Purication
Iminodiacetic acid Histidine/tryptophan Purication
(IDA)
Glass IDA-Cu2+ Lysozyme, cytochrome c, Purication
ribonuclease A
triazine-linked dyes have been used to mimic are complementary in shape and functionality to
coenzymes that bind a number of dehydroge- the target molecule.
nases, hexokinases, and alkaline phosphatases;
the reactive triazine groups can be linked to any
matrix containing hydroxyl groups by mixing the Applications of Affinity Membranes
two together. Cibacron Blue F3GA (a textile dye)
has been employed as ligand in afnity mem- Afnity membranes are used for several
branes for the purication of over 80 enzymes different applications such as purication of bio-
and proteins. molecules, removal of unwanted substances from
Another important example of pseudo- biological uids, and also small-scale analytical
biospecic afnity ligands is chelated metal ions separations.
used for the purication of histidine-tagged fusion The most common application is the separation
proteins. and purication of biomolecules and especially
proteins for large-scale production.
In Table 2 are reported some typical examples
Molecular Imprinted Membranes of their use for the separation and purication of
biomolecules.
A different approach for the preparation of afnity
membranes is the use of molecularly imprinted
polymeric materials. These are produced by
entrapping a template molecule (the molecule to References
be separated) in a polymer matrix during poly-
Klein E (ed) (1991) Afnity membranes: their chemistry
merization and subsequent extraction. In this way, and performance in adsorptive separation processes.
binding sites are introduced in the polymer that Wiley, New York
Affinity Membranes for Adsorptive Separation Process 19
Klein E (2000) Afnity membranes: a 10-year review. Adsorptive membranes are mainly applied in
J Membr Sci 179:127 separation of biomolecules (proteins, DNA, endo-
Wilson ID, Poole CF (2009) Handbook of methods and
instrumentation in separation science, vol 1. Academic, toxins, etc.). Their application has been extended A
Boston to the removal of toxic contaminants from waste-
Zou H, Luo Q, Zhou D (2001) Afnity membrane chro- water (Wu et al. 2008). The interactions between
matography for the analysis and purication of pro- adsorbates and adsorptive membranes are identi-
teins. J Biochem Biophys Methods 49:199240
cal to those using particulate adsorbents. They
could be categorized into four modes: afnity,
ion exchange, hydrophobic interaction, and
reverse phase (Roper and Lightfoot 1995;
Charcosset 1998). Afnity mode provides strong
Affinity Membranes for Adsorptive specic binding, such as biospecic,
Separation Process immunoafnity, dye afnity, immobilized metal
ion afnity, and pseudo afnity. Ion exchange
Shing-Yi Suen adsorption, including weak and strong, cation
Department of Chemical Engineering, National and anion exchange, is based on electrostatic
Chung Hsing University, Taichung, Taiwan interaction so that fast binding and high recovery
are easily attained to increase its popularity in
commercial products. The last two modes are
Afnity membrane (also called adsorptive mem- both induced by van der Waals force and hydro-
brane, membrane adsorber, or membrane chro- phobic interaction, with a difference that the
matography) is a design using membrane for strength of reverse phase mode is larger. The
adsorptive separation. The traditional shape for applications of these two modes are limited due
adsorbents is particle. In batch process, particulate to the probable inactivation of biomolecules.
adsorbents need to be suspended uniformly in Several methods could be employed to prepare
liquid bath by agitation for good contact with the adsorptive membranes: (1) polymerization of
adsorbates. After adsorption, centrifugation/ltra- monomers containing special functional groups
tion/precipitation is required to separate solid par- (or ligands), followed by dissolving of polymer
ticles from the liquid. On the other hand, and casting it into a lm; (2) introduction of func-
particulate adsorbents are commonly packed in a tional groups into a polymer in solution and cast-
column/vessel (so-called chromatography or ing it into a lm; (3) immobilization of functional
packed bed) for ow operation, where various groups onto a preformed membrane; and
operational conditions (xed bed, moving bed, (4) blending of particulate adsorbents and poly-
uidized bed, etc.) could be applied. However, mer solution together and casting it into a lm
pressure drop and intraparticle diffusion limita- (so-called mixed matrix membrane) (Avramescu
tions usually restrict the separation effectiveness et al. 2003; Sau and Fee 2009).
of large-scale chromatographic system. An Adsorptive membrane system can be operated
advantageous alternative for adsorptive separa- using commercially available geometries such as
tion is to adopt porous membrane with suitable at sheet, tubular/hollow ber, spiral wound,
functionality (adsorptive membrane), which has etc. (Roper and Lightfoot 1995; Zou et al. 2001;
exhibited better mass transfer effects by convec- Suen et al. 2003). Figure 1 depicts the ow pro-
tive transport of solutes and minimal thickness to cess of adsorptive membrane (i.e., membrane
allow operation at low pressure (Roper and chromatographic system). Similar to a conven-
Lightfoot 1995; Charcosset 1998; Zou tional chromatographic system with particulate
et al. 2001; Ghosh 2002; Suen et al. 2003). adsorbents, a cycle of operation consists of
20 Affinity Membranes for Purification of Enzymes
Affinity Membranes
for Adsorptive
Separation Process,
Fig. 1 Schematic diagram
of ow process for
adsorptive membrane Membrane UV-vis
module detector
Peristaltic
pump
Computer
Feed Effluent
solution
loading (adsorption), washing, elution

(desorption), and/or regeneration. To achieve a Affinity Membranes for Purification
larger-scale adsorptive capacity, it is simple and of Enzymes
convenient to stack at sheet membranes in series
or wind a larger sheet around a cylindrical core. Renbi Bai and Xiaoying Zhu
Department of Civil and Environmental
Engineering, National University of Singapore,
Faculty of Engineering, Singapore, Singapore
References
Avramescu ME, Borneman Z, Wessling M (2003) Mixed- Enzymes and Their Purification
matrix membrane adsorbers for protein separation.
J Chromatogr A 1006:171183
Charcosset C (1998) Purication of proteins by membrane
Enzymes are large biological molecules that cata-
chromatography. J Chem Technol Biotechnol lyze chemical interconversions. Most enzymes are
71:95110 proteins and they are highly selective catalysts.
Ghosh R (2002) Protein separation using membrane chro- Since enzymes are selective for their substrates and
matography: opportunities and challenges.
speed up only specic reactions from among many
Roper DK, Lightfoot EN (1995) Separation of biomole- possibilities, the types of enzymes determine the
cules using adsorptive membranes. J Chromatogr pathways and therefore affect the products of reac-
702:326 tions. Hence, there is always the need for enzymes
Sau SM, Fee CJ (2009) Fractionation of beta-
lactoglobulin from whey by mixed matrix membrane
of different types to be separated and puried.
ion exchange chromatography. Biotechnol Bioeng
103:138147
Suen SY, Liu YC, Chang CS (2003) Exploiting Affinity Membranes
immobilized metal afnity membranes for the isolation
or purication of therapeutically relevant species.
J Chromatogr B 797:305319 Conventional membranes generally act like a
Wu JS, Liu CH, Chu KH, Suen SY (2008) Removal of sieve for separation, blocking substances larger
cationic dye methyl violet 2B from water by cation than the membrane pores and passing substances
exchange membranes. J Membr Sci 309:239245
Zou H, Luo Q, Zhou D (2001) Afnity membrane chro-
smaller than the membrane pores. However,
matography for the analysis and purication of pro- afnity membranes operate not only on the siev-
teins. J Biochem Biophys Methods 49:199240 ing mechanism but more importantly have
Affinity Membranes for Purification of Enzymes 21
selective binding capabilities, provided by vari- positively charged or negatively charged

ous binding sites on the external and internal enzymes, respectively. For example, porous
surfaces of the membranes. Thus, afnity mem- nylon membranes were modied with poly A
branes can separate and purify substances with (2-(methacryloyloxy)ethyl succinate) brushes to
the combined advantages of high selectivity due purify positively charged lysozyme from chicken
to the specic binding sites and the high pro- egg white (Jain et al. 2010). Negatively charged
ductivity of conventional ltration membranes. sulfonyl matrices acting like ion exchangers were
Many afnity membranes were developed by prepared on cellulose membranes to purify crude
postmodications of existing microporous ultra- cationic urokinase (plasminogen-activating
ltration and microltration membranes. Due to enzyme) from human urine at pH 4.5 (Hou and
the different membrane materials, various modi- Zaniewski 1990). These afnity membranes can
cation approaches may have to be used to provide trap all positively charged enzymes or negatively
the specic binding sites on the membranes. For charged enzymes at the same time. In other words,
example, polymeric membrane materials with if one needs to have a particular type of enzymes,
functional end-groups can directly react with reac- the concentrated enzymes from the membrane
tive ligands. On the other hand, inert membrane need further purication.
materials, such as aliphatic hydrocarbons without The second type of afnity membranes can
active side chains or end-groups, are usually mod- provide more specic enzyme purication. These
ied by radiation chemistry to incur their reactiv- afnity membranes can bind enzymes through the
ity for ligand immobilization. Premodication of interaction between a Lewis acid (electron pair
membrane materials may be another route to pre- donor), such as a chelated metal ion and an electron
pare afnity membrane. Functional monomers acceptor group, such as histidine or tryptophan
with binding sites were polymerized to have residues on His-tagged recombinant proteins.
functionalized polymers that are subsequently This purication mechanism is the same as
used to fabricate afnity membranes. immobilized metal afnity chromatography
The applications of afnity membranes are (IMAC) which is a commercial protein purication
similar to those of afnity chromatography method targeting His-tagged recombinant proteins.
developed with the bead columns, but afnity For example, poly(2-(methacryloyloxy)ethyl suc-
membranes have the advantage of shorter process cinate) brushes of a porous nylon membrane were
time, increased separation efciency at low further functionalized with nitrilotriacetate (NTA)-
pressures, and greater productivity. Ni 2+ to bind His-tagged ubiquitin (Jain
et al. 2010). These afnity membranes usually
can rapidly get highly puried His-tagged proteins
Purification of Enzymes with Affinity from cell extracts.
Membranes The third type of afnity membranes is
using immobilized ligands to bind complimen-
One of the most important applications of afnity tary enzymes. For example, heparin was cova-
membrane is in purication of proteins or lently linked to a commercial membrane to bind
enzymes. The afnity membranes will trap or antithrombin III which was recovered with 91 %
adsorb specic proteins or enzymes during the purity and 97 % yield (Lutkemeyer et al. 1993).
ltration of their solutions. The trapped proteins
or enzymes are subsequently eluted from the
membranes and concentrated to get the puried
products. In general, there are three types of inter- References
actions between the afnity membranes and the
Hou KC, Zaniewski R (1990) Purication of urokinase by
enzymes. The rst type of interaction is electro-
combined cation exchanger and anity chromato-
static interaction. Anions or cations are covalently graphic cartridges. J Chromatogr-Biomed Appl
immobilized onto the membranes to trap 525(2):297306
22 Affinity Separation and Crystallization of Fc Fragments
Jain P, Vyas MK, Geiger JH, Baker GL, Bruening ML the effector function and biodistribution of the
(2010) Protein purication with polymeric afnity antibody (Beale 1987). Proteolytic digestion pro-
membranes containing functionalized poly(acid)
brushes. Biomacromolecules 11(4):10191026 vides important information to determine an anti-
Lutkemeyer D, Bretschneider M, Buntemeyer H, Lehmann body molecules structure and function. Antibody
J (1993) Membrane chromatography for rapid purica- molecules can be cleaved into fragments and each
tion of recombinant antihrombin-III and monoclonal- of these fragments has a distinct activity.
antibiodies from cell-culture supernatant. J Chromatogr
639(1):5766 A number of enzymes like papain, pepsin, cin,
bromelain, and elastase can be used for this pur-
pose, papain being the most frequently used (Luo
et al. 2002). Because of their smaller size as func-
tional components of the whole molecule, anti-
Affinity Separation body fragments offer several advantages over
and Crystallization of Fc Fragments intact antibodies for use in certain immunochem-
ical techniques and experimental applications. For
Nilay Bereli, Handan Yavuz and Adil Denizli this purpose, researchers have attempted to use
Department of Chemistry, Biochemistry Division, smaller, more stable counterparts of antibodies
Hacettepe University, Ankara, Turkey and have also been searching for alternative
ways to obtain antibody-like receptors (Haupt
et al. 1998). Thus, bioengineered antibody frag-
Immunoglobulins are used by the immune system ments that are smaller than antibodies and have
present in serum and tissue uids to identify and the ability to bind antigen bivalently have been
neutralize foreign targets. Immunoglobulins are promoted by the researchers (Cortese 1995). IgG
Y-shaped molecules comprised of two identical molecule can be digested with papain for gaining
light chains (L) and two identical heavy chains Fab and Fc fragments.
(H) linked via disulde covalent forces and via For the separation of Fab and Fc fragments,
non-covalent interactions. Each chain has con- HiTrap_r Protein A column can be used and col-
stant (CL and CH) and variable (VL and VH) lected the fragments by the fraction unit of the fast
domains (Holt et al. 2003). These form two protein liquid chromatography (FPLC) system
antigen-binding fragments (Fab) that specically (Erturk et al. 2011; Bereli et al. 2013; Aslyuce
recognize and bind to antigen molecule and a et al. 2013). As shown in Fig. 1, a successful
constant fragment (Fc) which is responsible for separation was observed and Fab and Fc fragments
Affinity Separation
and Crystallization of Fc 450
Fragments, Fig. 1 FPLC
chromatogram of IgG after Fab
digestion with papain 350
Absorbance (m/U)
(Erturk et al. 2011)

250
Fc
150
50
50
0 5 10 15 20 25
Time (min)
Aged Carbon Membrane 23
Affinity Separation and Crystallization of Fc Frag-

ments, Table 1 Chromatographic separation data Aged Carbon Membrane
a
Fragment tR N K Rs
A
Fab 3.06 52.39 1.19 6.14 - Miki Yoshimune
Fc 11.16 1089.60 7.31 - 4.70 Research Institute for Chemical Process
Technology, National Institute of Advanced
Industrial Science and Technology (AIST),
were separately collected. The retention times of Tsukuba, Japan
the fragments were achieved at 3.06 and 11.64
min, respectively. The calculated chromato-
graphic parameters such as tR, N, k0 , a, and Rs If the use of high-performance carbon mem-
values are summarized and given in Table 1. Rs branes in gas separations is to be commercial-
value was calculated as 4.70 for Fc fragment. ized, the aging characteristics of such
Because the Rs value should be higher than 1.0 membranes will become important. Little work
for a good resolution of two peaks in such a has been done on studying this aspect of carbon
chromatography system, the results for the reso- membranes, but some researchers have pointed
lution of Fab fragment/Fc fragment can be out that there are signicant problems related to
accepted as good resolution values. the performance stability of carbon membranes
(Jones and Koros 1995a; Lagorsse et al. 2004,
2005; Sau and Ismail 2004). Although they are
chemically and thermally resistant, carbon mem-
References branes generally suffer deterioration in perfor-
mance upon exposure to substances that become
Aslyuce S, Uzun L, Say R, Denizli A (2013) Immuno- strongly adsorbed, such as water vapor or oxygen.
globulin G recognition with Fab fragments imprinted
monolithic cryogels: evaluation of the effects of metal- As carbon membranes have a sharp pore-size dis-
ion assisted-coordination of template molecule. React tribution with a critical size ranging from 0.3 to
Funct Polym 73:813820 0.7 nm, small changes in the effective size of the
Beale D (1987) Molecular fragmentations. Some pore structures can markedly affect their perme-
applications in immunology. Dev Comp Immunol
11:287296 ability toward adsorbing gas molecules. In addi-
Bereli N, Erturk G, Tumer MA, Say R, Denizli A (2013) tion, constrictions in the pore system can occur as a
Oriented immobilized anti-hIgG via Fc fragment- result of chemisorption of oxygen or by strong
imprinted PHEMA cryogel for IgG purication. adsorption of species having very low diffusivity
Biomed Chromatogr 27:599607
Cortese R (1995) Identication of biologically active pep- because of their organophilic nature. Such species
tides using random libraries displayed on phage. Curr will block the pores in the membrane and attenuate
Opin Biotechnol 6:7380 the permeation of other species.
Ert
urk G, Bereli N, Uzun L, Tumer MA (2011) In a study of the effects of air oxidation and
Afnity separation and characterization of IgG
subfragments by fast protein liquid chromatography humidity at room temperature on the aging of
with HiTrap_r Protein A column. Hacet J Biol Chem carbon membranes, Menendez and Fuertes
39:133138 (2001) found the membrane stored in dry or
Haupt K, Mosbach K (1998) Plastic antibodies: humid air underwent a rapid loss of permeability,
developments and applications. Tibtech November
16, 468475 whereas membranes stored under propene or
Holt LJ, Herring C, Jespers LS, Woolven BP, Tomlinson IA nitrogen were protected from damage. The
(2003) Domain antibodies: proteins for therapy. Trends researchers suggested that chemisorption of oxy-
Biotechnol 21:484490 gen rather than physisorption of water is princi-
Luo Q, Mao X, Kong L, Huang X, Zou H (2002) High-
performance afnity chromatography for characteriza- pally responsible for the loss of permeability that
tion of human immunoglobulin G digestion with occurs when a fresh carbon membrane is exposed
papain. J Chromatogr B 776:139147 to air.
24 Air Bubble Geometry
Jones and Koros (1994) found that carbon oxygen. This treatment can be used to restore the
membranes are highly vulnerable to adverse permeation properties to those of fresh
effects of exposure to organic contaminants, membranes.
even at concentrations as low as 0.1 ppm. These studies demonstrated that the effects of
Attempts to regenerate membranes after exposure aging are more pronounced when membranes are
to hydrocarbons by high-temperature (363 K) stored under a high-humidity atmosphere and are
treatment in a vacuum were unsuccessful. Jones exposed to air. The chemisorption of oxygen leads
and Koros (1995a) also studied the effects of to the formation of CO surface groups, thereby
humidity on the O2/N2 selectivity and permeabil- creating a more hydrophilic interface that favors
ity of a carbon membrane by using feeds with increased adsorption of water. Adsorbed species
relative humidities of 2385 %. Some losses in can reduce the pore size of the carbon over time,
performance occurred at all humidity levels, leading to decreases in the permeability and selec-
but this problem could be alleviated by rendering tivity of the membrane. Therefore, studies on the
the surface of the membranes hydrophobic by stability of carbon membranes will be an impor-
coating them with thin layers of Teon AF1600 tant aspect on research on such membranes in the
or AF2400 (Jones and Koros 1995b). The future.
resulting composite carbon membranes were sig-
nicantly protected from the adverse effects of
humidity.
Lagorsse et al. (2008) made a thorough inves- References
tigation on the effects of exposure to air and
Jones CW, Koros WJ (1994) Carbon molecular sieve gas
humidity on the performance of commercialized separation membranes. II. Regeneration following
carbon hollow ber membranes, supplied by Car- organic exposure. Carbon 32:14271432
bon Membranes Ltd. (Israel). The production of Jones CW, Koros WJ (1995a) Characterization of
these membranes involved pyrolysis followed by ultramicroporous carbon membranes with humidied
feed. Ind Eng Chem Res 34:158163
chemical vapor deposition and an activation step Jones CW, Koros WJ (1995b) Carbon composite mem-
(Lagorsse et al. 2004). The researchers found that branes: a solution to adverse humidity effects. Ind
the membrane underwent a loss of permeability Eng Chem Res 34:164167
with time when exposed to pure oxygen or to air; Lagorsse S, Magalhes FS, Mendes A (2004) Carbon
molecular sieve membranes: sorption, kinetic
however, the performance of the membrane ulti- and structural characterization. J Membr Sci
mately stabilized after prolonged exposure. 241:275287
Attempts to regenerate the membranes by expo- Lagorsse S, Campo MC, Magalhes FD, Mendes A (2005)
sure to owing nitrogen at 343 K or by treatment Water adsorption on carbon molecular sieve mem-
branes. Experimental data and Isotherm model. Carbon
with propene and propyne as cleaning agents 43:27692779
resulted in no signicant recovery of permeability. Lagorsse S, Magalhes FD, Mendes A (2008) Aging study
However, no deterioration in permeability of of carbon molecular sieve membranes. J Membr Sci
these membranes was observed when they were 310:494502
Menendez I, Fuertes AB (2001) Aging of carbon mem-
stored under nitrogen. The researchers therefore branes under different environments. Carbon
concluded that aging of carbon membranes is 39:733740
caused exclusively by chemisorption of oxygen Sau SM, Ismail AF (2004) Fabrication of carbon mem-
from air. On the other hand, humidity adversely branes for gas separationa review. Carbon
42:241259
affects the performance of membranes in
multicomponent separation processes because
water is strongly adsorbed, but diffuses relatively
easily through the membranes. Heating under
an atmosphere of hydrogen at 893 K to remove Air Bubble Geometry
surface oxygen groups had some effect on
stable surfaces protected from chemisorption of Honeycomb Membrane Structure
Air Bubbling in Membrane Distillation 25
in direct contact membrane distillation (DCMD)

Air Bubbling in Membrane and the ow sheet is shown in Fig. 1.
Distillation The TCM extract was heated to 4362 C and A
pumped to the module; another pump was used to
Chunrui Wu circulate the permeate water (maintained at 25 C)
State Key Laboratory of Hollow Fiber Membrane at the permeate side of the membrane module. The
Materials and Membrane Processes, Institute of velocity of the feed and permeate ow in the mem-
Biological and Chemical Engineering, Tianjin brane module ranged from 0.07 to 0.13 m/s. To
Polytechnic University, Tianjin, China mitigate membrane fouling, a fan was used to create
gas bubbling at the feed side of the membrane
module. This fan was also utilized to generate gas
Air bubbling is a common method that is used in sparging at the permeate side for gas backwashing
membrane separation process to alleviate mem- of the membrane. The concentration ratio (CR) was
brane fouling and improve ux. It also is a kind of dened as the ratio of the volume of initial extract to
membrane cleaning method. Till now, gas-liquid the volume of concentrated extract to indicate the
two phase ow has been well studied and widely concentrating performance of the experiments.
used in microltration, ultraltration (Cabassud The result showed that a sharp ux decline right
et al. 1997), membrane bioreactor (Psoch and after the air was introduced into the feed circulation,
Schiewer 2005), and other traditional membrane and the ux remained at much lower level when air
separation technologies. was continuously injected. The explanation authors
Based on the above advantages of air bubbling, gave was that the bubbles occupied part of the
it has been introduced into membrane distillation liquid-contacting area in the membrane, thus reduc-
(MD) process to alleviate membrane fouling and ing the effective membrane area, and when gas
improve the heat and mass transfer efciency. entered the membrane pores, this reduced the partial
Researchers have introduced air bubbling into pressure of water vapor within the pores, thus reduc-
MD process and developed several new kinds of ing the driving force. To reduce these effects, inter-
MD process. mittent gas sparging at the feed side at intervals of
The rst application of air bubbling in MD 20 min and lasting only 2 min was applied.
process was rstly reported in 2008 (Ding The effect of gas bubbling on the DCMD per-
et al. 2008) to concentrate the extract of traditional formance was shown in Fig. 2. It could be seen
Chinese medicine (TCM). Air bubbling was used that ux declined rapidly in the beginning of the
Air Bubbling
in Membrane 9
Distillation,
Fig. 1 Schematic of
DCMD apparatus. (1)
Thermostatic bath, (2) tank
of TCM extract, (3) gear
pump for extract
circulation, (4) chiller, (5)
gear pump for permeate, (6)
permeate tank, (7) electric
balance, (8) fan for gas
bubbling, and (9) at sheet
membrane module
3 4 6 8
2 5
7
1
26 Air Bubbling in Membrane Distillation
Air Bubbling 24
in Membrane C0 =1.35g/l, tf=60C, u f=u p=0.102m/s
Distillation, 22 The gas bubbling lasted for 2 min every 20 min
Fig. 2 Performance of 20 1L/min gas flowrate
DCMD using gas bubbling 2L/min gas flowrate
18
on the feed side without gas bubbling
16
N, kg/m2h
14
12
10
8
6
4
2
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
CR
Air Bubbling
in Membrane
Distillation,
Fig. 3 Theoretic diagram
of ABMD process Vacuum
Membrane Module
Hollow Fiber
Gas-Liquid two phase

flow
Hot Solution Compressed Air
experiment and then slowly as a result of the very process. Compressed air was pressed into the
low-ux level at this time. The introduction of gas lumen side of the hollow ber membranes
bubbling at the feed side resulted in a higher ux together with the hot feed solution just at the
in the beginning of the experiment and a much inlet of membrane module to form gas-liquid
slower ux decline. Very similar trends were two phase ow, as shown in Fig. 3.
observed with gas bubbling at 1 and 2 L/min The results showed that the permeate ux
which meant that increasing the gas ow rate did increased with the air ow rate and/or feed tem-
not result in further process enhancement. perature increasing. For the VMD process, the
A novel air-bubbling vacuum membrane permeate ux was 22 kg/(m2h) when tested at
distillation (AVMD) process was constructed in 75 C with a feed ow rate of 120 L/h and a
2011 (Wu et al. 2011) to investigate the enhance- vacuum pressure of 0.0085 MPa. While, the ux
ment effect of air bubbling on traditional VMD of AVMD process could reach 40 kg/(m2h)
Air Bubbling in Membrane Distillation 27
Air Bubbling 50 4500

in Membrane N
Distillation, Fig. 4 Effect Re
of feed gas ow prate on the 40
4000
A
performance of AVMD
process
N (kg/m2h)
30
Re
3500
20
3000
10
0 2500
0 10 20 30 40 50 60 70
Gas flow rate (L/h)
Air Bubbling a
in Membrane
Distillation, Fig. 5 (a) Air
inlet connected to the
membrane module; (b) Air
nozzle
Feed
Air
when gas was introduced into feed solution at the J gas

F
rate of 60 L/h, as shown in Fig. 4. For the two J nogas
processes, ux both declined gradually as the feed
concentration increased from 3.5 to 300 g/L, but it where Jgas and Jnogas are the permeate uxes in the
was much slower for AVMD process. SEM pic- DCMD process with and without bubbling,
tures showed that membranes used in AVMD respectively.
process had lighter fouling than that in VMD Gas bubbling is supplied by an air pump. The
process. air inlet at the feed side entrance of the membrane
For elevated salt concentrations in the feed module and the air nozzle for dispersing the air are
steam, gas bubbling was introduced into direct shown in Fig. 5.
contact membrane distillation (DCMD) in 2013 The results showed that the introduction of gas
(Chen et al. 2013) to examine its effect on the MD bubbling could effectively enhance MD ux, and
performance. In the paper, to characterize the per- the ux enhancement ratio could reach 1.72 at an
formance improvement achieved by gas bubbling, optimized gas ow rate. When the condition was
the ux enhancement ratio F is dened: either a high feed operating temperature, low feed
28 Air Bubbling in Membrane Distillation
Air Bubbling 1.8

in Membrane Laminar flow, Ref = 842
Distillation, Fig. 6 Effect Turbulent flow, Ref = 2808
of gas ow rate on F in
laminar and turbulent ows 1.6
1.4
F
1.2
1.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Gas flowrate (L.min1)
and permeate Reynolds numbers, or a module still relatively stable with an average value of
with 45 inclined orientation, shorter ber length, 5.16 L/(m2h).
or lower packing density, a higher ux enhance- To sum up, air bubbling was an effective
ment ratio could be achieved. Because air bub- method in membrane distillation to enhance the
bling could enhance the turbulence of feed performance of MD process. Due to the role of
solution, temperature polarization was effectively increasing turbulence of feed solution, air bub-
reduced. However it was not that the bigger the bling could promote the agitation and mix of
gas ow rate, the higher the enhancement ratio. It feed solution to decrease the temperature and con-
was observed that beyond an optimal gas ow centration polarization effect. In addition, the
rate, the enhancement ratio fell, possibly due to enhancing surface shear rate increases the remove
excessive ow bypassing. of foulant from membrane surface at the feed side
The effect of gas ow rate on F in laminar and to inhibit the inuence of membrane fouling.
turbulent ows was shown in Fig. 6. In the gure,
it was observed that these two F curves had a
similar trend. At rst, F increased with increasing References
gas ow rate in the range of 0 Qg 0.2 L min1
for the laminar ow and 0 Qg 0.5 L min1 for Cabassud C, Laborie S, Lain JM (1997) How slug-ow
can improve ultraltration ux in organic hollow bres.
the turbulent ow, while F decreased when gas J Membr Sci 128(1):93101
rate was higher. Furthermore, the value of the Chen GZ, Yang X, Wang R, Fane AG (2013) Performance
laminar condition was much higher than that of enhancement and scaling control with gas bubbling in
direct contact membrane distillation. Desalination
the turbulent ow when other conditions were the
2013(308):4755
same. Ding Z, Liu L, Yu J, Ma R, Yang Z (2008) Concentrating the
In 2009, air bubbling was used in a hybrid extract of traditional Chinese medicine by direct contact
process combining membrane distillation in a membrane distillation. J Membr Sci 310(12):539549
Jirachote P, Fane AG, Pasquier ACS, Wu B (2009) Exper-
submerged membrane bioreactor (MDBR) to
imental study and design of a submerged membrane
mitigate severe fouling and ux decline distillation bioreactor. Chem Eng Technol 32(1):3844
(Phattaranawik et al. 2009). The result showed Psoch C, Schiewer S (2005) Long-term study of an inter-
that adding two inlets for air sparging at mittent air sparged MBR for synthetic wastewater treat-
ment. J Membr Sci 2005(260):5665
2.5 L/min each near the potting at the center of
Wu CR, Yue J, HuaYan C, Xuan W, QiJun G (2011) Study
the membrane bundle induced more turbulence at on air-bubbling strengthened membrane distillation
the membrane surface. Over 15 days, uxes were process. Desalination Water Treat 34(13):25
Air Dehydration by Membrane Technology 29
feed pressure (pF) favor for deep dehumidication,

Air Dehydration by Membrane i.e., allowing low moisture content in feed air (xF).
Technology If no sweep gas is used, the permeate pressure A
has to be decreased by vacuum to obtain a positive
Wei Liu partial pressure gradient of water vapor across the
Energy and Environmental Technology membrane:
Directorate, Pacic Northwest National
Laboratory, Richland, WA, USA pP < x F pF
Membrane dehydration is commonly used in

Synonym industries for removal of moisture from pressur-
ized air. In these applications, ambient air at atmo-
Membrane air dehydration spheric pressure is often used as sweep gas.
Though moisture content (such as molar fraction)
Air dehydration refers to removal of moisture in the sweep air may be the same as in the process
from humid air and is also termed as air dehumid- air, a signicant partial pressure gradient of water
ication and air drying. Due to the ubiquitous vapor can be managed due to high pressures in
presence of moisture in air, air dehydration repre- the process air. For example, hollow ber mem-
sents one of the largest membrane applications. branes are used for drying of the air pressurized to
Air dehydration through a membrane is essen- about 10 bar (Morgan et al. 1996). Air-sweep
tially a vapor-phase separation process (Fig. 1), dehumidication can also be found in some
which can be carried out in two ways. In one way, laboratory membrane devices, such as Naon
moisture in the feed air diffuses across the mem- membrane tubes for moisture exchange (Ye and
brane and is swept out of the membrane unit with LeVan 2003).
a sweep gas stream (Fig. 1a). In another way, the When humid air is at low or atmospheric pres-
permeated moisture is pulled away from the mem- sures, using ambient air as a sweep stream is no
brane unit by vacuum (Fig. 1b). longer practical, and pulling vacuum in the per-
In the sweeping process, degree of air dehu- meate side becomes necessary. In this case, sev-
midication will be limited by moisture content in eral other pieces of equipment in addition to the
the sweep gas and operation pressures, as membrane separator are required to make an inte-
explained by the following equations: grated dehumidication system (Fig. 2). Such a
membrane dehumidier has been recently pro-
Dp xF pF xS pP > 0
x P pP posed and is still under development (Xing
xF >
pF et al. 2013). An array of membrane units can be
assembled together that the permeate sides of all
Low molar fraction of moisture in sweep the membrane units are connected to one common
gas (xP), low permeate pressure (pP), and high vacuum line. The vacuum is generated by the use
Air Dehydration by Membrane Technology, Fig. 1 Removal of moisture from humid air through a membrane
30 Air Enrichment, by Polymeric Magnetic Membranes
WATER
VAPOR
LOW
HUMID
HUMIDITY
AIR
DEHUMIDIFICATION AIR
1 UNIT
2 3 CONDENSER 4
COMPRESSOR
5 NON-CONDENSABLE
VACUUM
PUMP GAS
LIQUID LIQUID
RESERVOIR PUMP
LIQUID
Air Dehydration by Membrane Technology, Fig. 2 Process ow diagram of a membrane dehumidication system
of a vacuum pump or gas compressor. The per-

meated moisture is compressed to a pressure Air Enrichment, by Polymeric
above water dew point at environmental Magnetic Membranes
temperature so that water vapor is condensed
into liquid-phase water by rejecting heat of con- Anna Strzelewicz
densation into environment. The condensed water Faculty of Chemistry, Silesian University of
is collected in a reservoir. Water in the reservoir Technology, Gliwice, Poland
may be discharged using a liquid pump, while
residual air noncondensable gas is discharged
into environment using a secondary vacuum The idea of using magnetic membranes for enrich-
pump. ment of air by oxygen is based on the observation
that oxygen and nitrogen have quite different
magnetic properties, i.e., oxygen is paramagnetic
whereas nitrogen diamagnetic. The oxygen mole-
cule is paramagnetic with a magnetic moment of
References
mO2 2:73 1023 JT 1 : Magnetic susceptibil-
Morgan WH, Bleikamp LK, Kalthod DG (1996) Hollow ity of nitrogen is equal to w 150:8 106
ber membrane dryer with internal sweep, US patent mol1 2:5 1028 molecule1 which corre-
no. 5,525,143
sponds only to m 2:5 1028 JT 1 in a mag-
Xing R, Rao Y, TeGrotenhuis W, Caneld N, Zheng F,
Winiarski DW, Liu W (2013) Advanced thin zeolite/ netic eld of 1 T, a value ve orders of magnitude
metal at sheet membrane for energy efcient air dehu- smaller than the O2 magnetic moment (Morrish
midication and conditioning. Chem Eng Sci 1965; Borys et al. 2011).
104:596609
Magnetic membranes are polymeric mem-
Ye X, LeVan MD (2003) Water transport properties of
Naon membranes: part I. Single-tube membrane mod- branes (ethyl cellulose (EC) or poly(2,6-dimethyl-
ule for air drying. J Membr Sci 221:147161 1,4-phenylene oxide) (PPO)) with dispersed
Air Enrichment, by Polymeric Magnetic Membranes 31
Air Enrichment, by
Polymeric Magnetic
Membranes,
Fig. 1 Scheme of the A
experimental setup
magnetic powder (ferrite, praseodymium, and Integration of ux with respect to time allows to
neodymium). The membranes are made by cast- sketch a downstream permeation curve. Typical
ing of appropriate polymer solution with dis- plot of Qa(l,t) versus time is presented in the
persed magnetic powder in an external magnetic Fig. 2.
eld of a specially designed coil (stable magnetic Time-lag method and D1-D8 system allow to
eld with range of induction 040 mT). Removed get some insight into the nature of the considered
from Petri dish membranes are dried in 40 C for transport process. In papers (Strzelewicz and
at least 2 days and then are stored in an exsiccator Grzywna 2007, 2008; Rybak et al. 2009a, b,
under the vacuum conditions (p = 3 mmHg). 2012; Grzywna et al. 2010) a concept of magnetic
Collection of permeation quantities both for indi- membranes for air enrichment is explained. The
vidual gases (O2, N2) and for their mixture (air authors observed an increase in the oxygen ux
21 %O2/79 % N2) is measured in experimental with respect to the nitrogen ux and the enrich-
setup (Fig. 1). ment in the oxygen content of the permeate up to
The measurements are carried out in room 55.6 % for ethyl cellulose magnetic membranes.
temperature for membranes with dispersed mag- Further permeation measurements done in
netic powder before and after magnetization in a polyphenylene oxide (PPO) magnetic membranes
eld magnet with magnetic induction about provided higher enrichments, up to 61.9 %
2.5 T. The setup furnished with a gas chromato- (Table 1).
graph allows to measure the oxygen and the nitro- The transport of the molecules through the
gen concentration in permeate. The main part of magnetic membranes can be modeled by a diffu-
the experimental setup is diffusive chamber, sion with a position-dependent diffusion coef-
where the membrane is put in the form of disk. cient. Such a diffusion coefcient reects the
The setup is used for a low-pressure (from 0.1 to changes in the membrane composition along the
1.0 MPa) analysis of gas permeation. Transport permeation axis. One can observe on the scanning
coefcients can be calculated using ow rate data electron microscope photograph (Borys
and percentage of air enrichment. The ow rate of et al. 2011) of the magnetic membrane cross sec-
the permeate can be recorded using a owmeter. tion that the feed side of the membrane is
32 Air Enrichment, by Polymeric Magnetic Membranes
Air Enrichment, by a
Q
Polymeric Magnetic point for checking
Membranes, oxygen content
stationary state
Fig. 2 Downstream
absorption permeation
curve
a
Q (I, t)
La(L) t
C01
6
Air Enrichment, by Polymeric Magnetic Membranes, the transport of both nitrogen and oxygen by
Table 1 Air enrichment for various membranes prolonging their residence in the channel.
(Strzelewicz and Grzywna 2007)
Oxygen content in The method of air enrichment by magnetic
Membrane B[mT] permeate [%]
membranes seems to be effective and efcient.
EC + 1.38 g Nd 0.79 40.7 1.1
EC + 1.49 g Nd 1.25 43.8 1.1
PPO + 1.80 g Nd 1.70 54.1 1.4
PPO + 1.80 g Nd 2.70 61.9 1.5
References
Borys P, Pawelek K, Grzywna ZJ (2011) On the magnetic

composed of a pure polymer while the output side channels in polymer membranes. Phys Chem Chem
Phys 13:1712217129
consists of the polymer with dispersed magnetic Grzywna ZJ, Rybak A, Strzelewicz A (2010) Air enrich-
granules. When the magnetic eld is too strong, ment by polymeric magnetic membrane. In:
then magnetic aggregates are created, which inu- Yampolskii Y, Freeman B (eds) Membrane gas separa-
ence the permeation of gases. Detailed analysis of tion. Wiley, Chichester, pp 159182
Morrish AH (1965) The physical principles of magnetism,
the available data and microscopy images allowed
1st edn. Wiley, New York
to arrive four conclusions (Borys et al. 2011): Rybak A, Krasowska M, Strzelewicz A, Grzywna ZJ
(2009a) Smoluchowski type equations for modelling
1. There are magnetic channels formed around of air separation by membranes with various structure.
Acta Phys Pol B 40:14471454
the magnetic granules.
Rybak A, Grzywna ZJ, Kaszuwara W (2009b) On the air
2. The channels provide high permeability high- enrichment by polymer magnetic membranes. J Membr
ways for the diffusion of permeating Sci 336:7985
molecules. Rybak A, Strzelewicz A, Krasowska M, Dudek G,
Grzywna ZJ (2012) On the air separation process by
3. The oxygen molecules, due to their paramag- magnetic membranes inuence of various parameters.
netic properties, stick to these highways for a Sep Sci Technol 47:13951404
longer time than nitrogen, which is probably Strzelewicz A, Grzywna ZJ (2007) Studies on the air
based on the interaction with the Weiss molec- membrane separation in the presence of magnetic
eld. J Membr Sci 294:6067
ular eld of the permanent magnet.
Strzelewicz A, Grzywna ZJ (2008) On the permeation time
4. The magnetic eld induces aggregation lag for different transport equations by Frisch method.
between oxygen and nitrogen which enhances J Membr Sci 322:460465
Air Gap Membrane Distillation (AGMD) 33
may be changed from pure conduction and diffu-

Air Gap Membrane Distillation sion to natural convection. The optimum of air
(AGMD) gap width should be determined to achieve the A
best performance of the AGMD system taking
Maria Tomaszewska into account a reduction of the heat loss and the
Faculty of Chemical Technology and value of the permeate ux.
Engineering, West Pomeranian University of Since in AGMD the permeate is condensed on
Technology, Szczecin, Poland the cold surface, the liquid permeate is separated
from the membrane surface, which prevents the
membrane wetting. Thus an important advantage
AGMD is a conguration of membrane distilla- of the process is the ability to remove the organic
tion (MD) in which an air layer is interposed compounds from aqueous solutions.
between a porous hydrophobic membrane and The permeate ux in AGMD can vary from
the condensation surface (Fig. 1). In the process, ca. 0.3 to 26 kg/m2h, depending on the operating
volatile compounds (including water) present in a conditions and feed composition. The rejection
warm feed evaporate at the liquid/vapor interface coefcient of nonvolatile species, e.g., NaCl, usu-
formed at the membrane surface. The vapor is ally amounts close to 100 %. Generally, in AGMD
transferred through the membrane pores and the the nonvolatile species present in a feed are almost
gas gap and nally condenses on a cold surface completely retained, whereas volatile compounds
inside the membrane module. The driving force of are partially transferred through the membrane.
the mass transfer in AGMD is the difference in The separation of compounds present in aqueous
vapor pressure on both sides of the membrane. solutions using AGMD is based not only on the
The air gap signicantly reduces the heat loss vapor/liquid equilibrium, but also the diffusion
by the membrane material, therefore the permeate rate of the feed components through the gas lling
ux could increase. On the other hand air gap the membrane pores and the air gap between the
forms a barrier for the vapor, and due to the membrane and the condensation surface.
mass transfer resistance the permeate ux The air entrapped within the membrane pores
decreases with an increase of the air gap width and in the air gap can be considered as a stagnant
(Table 1) (Alkhudhiri et al. 2012). When the gap lm, thus the mass transfer in AGMD is generally
width is changed, the mechanism of heat transfer described according to the ordinary molecular dif-
fusion including both the air and noncondensable
Feed Tf,in gases in the membrane pores and in the air gap. The
vapor air gap width is usually 10100 times greater than
the membrane thickness; therefore the membrane
characteristics are usually neglected. In AGMD,
Tf,m Tc
Important condition: the heat transfer includes the heat transfer through
Tf,m > Tc the feed boundary layer, through the membrane
and the gas gap, condensation at the cold surface,
air gap
and heat transfer through the condensate liquid

membrane boundary layer and to the cooling water. To opti-
mize the AGMD systems, the effects of operating
Tf,in > Tf,out parameters should be taken into account (Khayet
and Matsuura 2011).
Tf,out cooling liquid The feed temperature strongly affects the per-
Liquid meate ux. Since the vapor pressure of aqueous
permeate
solutions grows exponentially with temperature,
Air Gap Membrane Distillation (AGMD), the driving force increases for both water and
Fig. 1 Schema of AGMD volatile compounds present in a feed. Thus, the
34 Air Gap Membrane Distillation (AGMD), Applications of
Air Gap Membrane Distillation (AGMD), Table 1 Air gap effect on the permeate flux (reprinted from Alkhudhiri
et al. (2012), copyright (2012) with permission from Elsevier)
Membrane Pore size Tin Air gap Permeate ux
Ref material (mm) Feed ( C) (mm) kg/m2.h
[5] PVDF 0.45 Articial seawater 60 1.99.9 52.1
[6] PTFE 0.2 Isopropanol 50 1.620.55 5.16.3
[108] PVDF 0.22 Sucrose 25.8 14 1.70.8 l/m2h
[18] PTFE 0.2 NaCl (3.8 %) 60 0.39 191.5
[120] PTFE 0.22 HNO3 (4 M) 80 0.52 5.34.25 l/m2h
[123] PTFE 0.45 HCl/water 60 47 3.72.4
[123] PTFE 0.2 Propionic acid/ 60 47 7.44.6
water
Comments the references can be found in the original paper
total permeate ux Jt (a sum of the water vapor

ux Jw and the volatile compounds ux Jv) Air Gap Membrane Distillation
increases, but the selectivity of volatile solutes (AGMD), Applications of
decreases with a feed temperature. The heat loss
by conduction through the membrane material is Maria Tomaszewska
lower at higher feed temperature which also Faculty of Chemical Technology and
explains the higher permeate ux. Engineering, West Pomeranian University of
An increase in the concentration of nonvolatile Technology, Szczecin, Poland
compounds in a feed reduces the Jt due to a
decrease in the Jw. In the case of an increase of
concentration of volatile compounds in a feed, the The separation using AGMD is based not only on
Jt and Jv increase, whereas the Jw decreases. It is vapor/liquid equilibrium but also on the diffusion
associated with changes of the partial pressure of rates of the feed components through the gas
dissolved species and water. The selectivity of trapped in the membrane pores and the air gap
volatile compounds can be different depending between the membrane and the condensation sur-
on their properties and concentration in the feed face. Nonvolatile compounds are rejected by a
and can be improved by an increase of their partial membrane in almost 100 %. Therefore, AGMD
pressure by salt addition into a feed. Generally, a can be applied for desalination and production of
higher feed ow rate diminishes the effects of high-purity water. A permeate with electrical con-
both temperature and concentration polarization ductivity as low as 4 mS/cm can be obtained from
and AGMD permeate ux increases. Different a solution with composition similar to seawater.
spacers can be applied to improve the hydrody- Moreover, AGMD can be applied for the separa-
namic conditions. The effect of coolant tempera- tion of volatile compounds. Since the vapor is
ture is not very signicant because the heat condensed on the cold surface and the liquid per-
transfer coefcient in the air gap dominates the meate does not come into contact with a hydro-
overall heat transfer coefcient (Khayet and phobic membrane, even the organic solutes can be
Matsuura 2011). separated without a risk of the membrane wetting.
The air gas reduces considerably the heat loss and
the temperature polarization which improve the
References separation efciency of AGMD; however, this
leads simultaneously to a signicant decrease in
Alkhudhiri A, Darwish N, Hilal N (2012) Membrane distil-
the permeate ux. The decrease depends strongly
lation: a comprehensive review. Desalination 287:218
Khayet M, Matsuura T (2011) Membrane distillation. Prin- on the width of the air gap. The air gap is about
ciples and applications. Elsevier, Amsterdam 10100 times thicker than the membrane, so the
Air Gap Membrane Distillation (AGMD), Applications of 35
Air Gap Membrane Distillation (AGMD), Applications of, Table 1 Application of air gap membrane distillation
Area Membrane material Application
Desalination PTFE, PVDF Water and wastewater treatment, production of pure water, A
Modied ZrO2, modied Al2O3, solar-powered desalination, geothermal wastewater
modied TiO2, modied treatment
aluminosilicate
Food industry PVDF, PTFE Fruit and vegetable juice concentration, milk concentration,
sugar and gelatine solution concentration, ethanol
concentration
Biotechnology PTFE, PVDF Ethanol separation from a broth
Others PTFE, PVDF Methanol, isopropanol concentration, separation of
azeotropic mixtures HCl/water, propionic acid/water,
formic acid/water, separation of radioisotopes 18O and 18F,
concentration of wastewater containing isotopes, HI and
H2SO4 concentration
Separation/removal of HNO3, HCl, formic acid, acetic acid,
propanone from aqueous solutions
PTFE polytetrauoroethylene, PVDF polyvinylidene uoride
effect of air inside the membrane can be feedwater, tness for intermittent energy supply,
neglected. AGMD creates the opportunity for the and the ability to use direct solar heat without heat
separation of species forming the azeotropic solu- storage, the plants have not been commercialized
tions. It is attributed to the difference in the diffu- or implemented in a large scale. This is mainly
sion coefcients of volatile compounds present in because reverse osmosis (RO) is considered as
a feed through the gas gap (Khayet and Matsuura more economical and energy-efcient desalina-
2011). Thus, AGMD can nd the application in tion process (Saffarini et al. 2012). However, the
water and wastewater treatment, biotechnology, AGMD/solar energy systems are still in progress
food industry, etc. (Table 1). and nd a growing interest.
The plate and frame, tubular, capillary, and Another eld of AGMD application is food
very frequently spiral wound modules equipped processing. Figure 2 presents the possibility of
with porous hydrophobic membranes are used juice concentration by AGMD to achieve much
in AGMD. higher feed concentration than using RO. Low
The main area of AGMD application is desali- processing temperature allows to concentrate the
nation. The desalination pilot plants combining solutions sensitive to high temperature and pro-
AGMD with solar energy collectors were devel- vides a good quality of juice concentrate (Kimura
oped, especially for operation in arid and semiarid and Nakao 1987). The concentration of milk dem-
remote regions with a lack of electricity and drink- onstrated that the membranes were easily fouled
able water but with high solar radiation (Table 2). by milk fat and the permeate ux decreased very
The production of potable water, e.g., from plants fast (Khayet and Matsuura 2011).
installed in Jordan was of 100 L/day (utilizing Ethanol preferentially vaporizes from the
brackish water) and 1 m3/day (Fig. 1) (utilizing aqueous feed and is concentrated in the permeate.
untreated seawater from the Red Sea). The instal- Since in AGMD the liquid ethanol does not come
lations were equipped with AGMD spiral wound into contact with a membrane, it can be separated
modules with area of 1040 m2, respectively. from a fermentation broth and concentrated with-
Long-term experiments have demonstrated that out a risk of the membrane wetting. It was found
the average salt rejection was 98 % (Qtaishat and that a salt addition into diluted ethanol increases
Banat 2012; Khayet and Matsuura 2011). Distil- alcohol selectivity (Khayet and Matsuura 2011).
late conductivity of 35 mS/cm can be obtained for AGMD can be useful in breaking the
brackish water feeding. Despite the advantages azeotropic mixtures of water and hydrochloric or
such as the ability to desalinate high saline propionic acids. This phenomenon was associated
36 Air Gap Membrane Distillation (AGMD), Applications of
Air Gap Membrane Distillation (AGMD), Applications of, Table 2 Timeline of solar-powered membrane distillation
systems and their general features (Adapted from Saffarini et al. 2012)
Energy system type Solar
Membrane Feedwater collector
Year Location MD type type/material type Thermal Electrical type
1991 New South DCMD Hollow ber/ N/A Solar Grid N/A
Whales, N/A collector
Australia
1995 New South AGMD Spiral wound/ N/A Solar Grid VPC
Wales, Australia PTFE collector
1997 Tokyo, Japan N/A N/A/N/A N/A Solar PV N/A
collector
1999 Texas, USA AGMD N/A/N/A N/A Solar Grid N/A
pond
1999 Irbid, Jordan DCMD N/A/PTFE Brackish Solar Grid N/A
still
2003 Freiburg, AGMD Spiral wound/ Freshwater Solar Grid FPC
Germany PTFE collector
2004 Texas, USA AGMD N/A/N/A NaCl Solar Grid N/A
solution pond
2007 Irbid, Jordan AGMD Spiral wound/ Brackish Solar PV FPC
PTFE collector
2007 Aqaba, Jordan AGMD Spiral wound/ Seawater Solar PV FPC
PTFE collector
2008 Alexandria, AGMD Spiral wound/ Brackish Solar PV FPC
Egypt PTFE collector
2008 Mexico DCMD Hollow ber/ N/A Solar Grid FPC
N/A collector
2009 Almeria, Spain AGMD Flat sheet/ Seawater Solar Grid CPC
PTFE collector
2009 Hangzhou, China VDM Hollow ber/ Brackish Solar Grid N/A
PP collector
2009 Nevada, Reno, Vacuum N/A/N/A N/A Solar N/A N/A
USA enhanced pond
DCMD
2009 Gran Canaria, AGMD Spiral wound/ Seawater Solar PV Double
Spain PTFE collector glass FPC
2011 Marina Barrage, VMEMD Flat sheet/ Seawater Solar PV N/A
Singapore PTFE collector
DCMD direct contact membrane distillation, VMD vacuum membrane distillation, PP polypropylene, N/A not available,
PV photovoltaic panels, VPC vacuum plate collector, FPC at plate collector, CPC concentrated parabolic collectors,
VMEMD vacuum multi-effect membrane distillation
with the differences in the acid-air and water-air AGMD was successfully used for the concen-
diffusion rates in the air gap. In AGMD the tration of a mixture of HI and H2SO4, compounds
hydrochloric acid/water azeotrope was shifted to of great interest for H2 generation in thermochem-
higher acid concentration, and the azeotrope was ical water-splitting cycle (Khayet and Matsuura
eliminated in the propionic acid/water system. It 2011). AGMD has also a great potential in the
was found that the heavy gases such as SF6 help application in the environment protection area.
more in eliminating the azeotropic point than the Therefore, this technique requires more studies,
lighter ones as air or He (Khayet and Matsuura devoted especially to a long-term performance
2011). and membrane life.
Air Gap Membrane Distillation (AGMD), Applications of 37
Solar
irradiation Brine
A
Collector
feild
Feed
Storage pump
tank
Heat
exchanger
Distillate MD modules
Battery PV
Expansion
vessel
Control
unit DC PV
AC
PV array
Air Gap Membrane Distillation (AGMD), Applications powdered (Reprinted from Qtaishat and Banat (2012),
of, Fig. 1 Schema of the large system (two-loop desali- copyright (2012) with permission from Elsevier)
nation system) of desalination using AGMD solar
Air Gap Membrane 400

Distillation (AGMD),
Applications of, RO, P = 5 MPa
350 AGMD, Tf = 321 K
Permeate flux, J [kg m2d1]
Fig. 2 The comparison of

the permeate ux during
concentration of the 300
mandarin orange juice
using RO and AGMD
(Adapted from Kimura and 250
Nakao 1987)
200
150
100
1 1.5 2 2.5 3
Concentration ratio, CV
References Qtaishat MR, Banat F (2012) Desalination by solar pow-

dered membrane distillation systems. Desalination.
Khayet M, Matsuura T (2011) Membrane distillation. Prin- doi:10.1016/desal2012.01.021
ciples and applications. Elsevier, Amsterdam Saffarini RB, Summers EK, Arafat HA, Lienhard VJH
Kimura S, Nakao SI (1987) Transport phenomena (2012) Technical evaluation of stand-alone solar
in membrane distillation. J Membr Sci 33: powered membrane distillation systems. Desalination
285298 286:332341
38 Air Products
The compressed air at a temperature of

Air Products 80100 C and a pressure of about 1000 kPa
(135150 psi) is puried by the use of coalescing
Pushpinder S. Puri lter and fed to the membrane module. The pro-
PuriMem, LLC, Emmaus, PA, USA ductivity and purity of the nitrogen produced is
controlled by the use of a product pressure valve.
Membrane systems produce nitrogen at a concen-
The majority of air separation is done by cryo- tration ranging from 95 % to 99.9 %. However,
genic distillation; however, for several small due to the modest selectivity (<8) of air separa-
applications a gas separation membrane is desir- tion membranes, production of high purity nitro-
able. The key commercial gaseous products made gen is not economically favorable. Key producers
from air by membranes are nitrogen, oxygen of nitrogen separation membranes are Air Prod-
enriched air, and dry air. Further, gas separation ucts, Air Liquide, and Ube. Other nitrogen mem-
membranes in hybrid system with a cryogenic brane producers are: Generon Systems, Aquillo
process is used to produce Argon (Agrawal and and Parker Henifn.
Xu 1996) and with an adsorption process to pro- Dry air with dew point up to 40 C (40 F)
duce high concentration oxygen enriched air is produced by removing moisture from the atmo-
(Choe et al. 1997). spheric air using a water permeable membrane
In membrane nitrogen production systems which are either made of glassy polymers
(Drioli and Barbieri 2011), the compressed air is (polysulfone, polyimide, etc.) or uorine
fed on one side of the membrane; oxygen, argon, containing ionomers. Schematic of a typical air
carbon dioxide, and water which have higher per- drying system is shown in Fig. 2.
meation than nitrogen, permeate through the In air drying system, clean wet air is fed under
membrane leaving nitrogen behind. A schematic pressure on one side of the membrane module and
of the membrane nitrogen generation system is dry air at slightly reduced pressure is recovered as
shown in Fig. 1. retentate. In some systems, an internal or external
VENT TO AIR
ATMOSPHERE
COMPRESSOR WASTE GAS
NITROGEN
ELECTRIC NITROGEN TO
HEATER CUSTOMER
OXYGEN NITROGEN
CONDENSATE ANALYSER BUFFER
NITROGEN VESSEL
PRODUCTION
X
MEMBRANE
Air Products, Fig. 1 Schematic of nitrogen generation system (Reproduced with permission from Air Products)
Air Separation 39
Purging Valve
Purging Air
A
Drain Filter
Compressor Cooler
Separetor Product
Dry Air
SUNSEP-WTM Dryer
Drain
Air Products, Fig. 2 Typical air drying system (Courtesy: AGC Chemicals Americas)
dry air sweep is used to prevent water buildup on

the membrane in permeate, which retards mem- Air Separation
brane performance (Morgan et al. 1996).
In some instances, membrane systems are used Ahmad Fauzi Ismail1 and M. H. D. Othman2,3
1
to produce oxygen enriched air with oxygen con- Faculty of Chemical and Energy Engineering,
centration <30 % O2. Mostly, rubbery mem- Universiti Teknologi Malaysia, Advanced
branes of modest selectivity (23) and high Membrane Technology Research Centre
oxygen ux are used for this application3. The (AMTEC), Johor Bahru, Johor, Malaysia
2
membrane system used is similar to that used for Advanced Membrane Technology Research
nitrogen production described above. Sometimes, Centre (AMTEC), Universiti Teknologi Malaysia,
use of vacuum on the permeate side is preferred Skudai, Johor, Malaysia
3
over the use of a compressor for the production of Faculty of Chemical and Energy Engineering,
oxygen enriched air. Universiti Teknologi Malaysia, Skudai, Johor,
Malaysia
Air separation technology is used for the produc-

References
tion of oxygen, nitrogen, and rare gases that are
Agrawal R, Xu J (1996) Gas-separation membrane cas- present in air such as argon and neon. There are
cades utilizing limited numbers of compressors. AIChE two fundamental approaches to air separation,
J 42:2141 which are cryogenic and non-cryogenic pro-
Choe JC et al (1997) Process for separating components of
cesses. The cryogenic process which is carried
a gas stream. US Patent 4,701,187
Drioli E, Barbieri G (eds) (2011) Membrane engineering out in distillation column has the capability to
for the treatment of gases. Gas-separation problems deliver large and high purities of products, while
with membranes, vol 1. RSC Publishing, Cambridge. the non-cryogenic which is based on absorption
Chapter 8,
and membrane technologies is more suitable for
Morgan WH, Bleikamp LK, Kalthod DG (1996) Hollow
ber membrane dryers with internal sweep. US Patent on-site production, which is most common for
5,525,143 small and medium throughputs.
40 Air Separation
Air Separation,
Pressurized Oxygen
Fig. 1 Schematic
air feed O 2 + 2e O2 depleted air
representation of the
oxygen transport in MIEC MIEC
ceramic membrane ceramic
O2 - e- O2- e-
membrane
Oxygen rich O2 O 2 + 2e Sweep gas

stream
Membrane technology for air separation has Interestingly, the oxygen separation from air in
developed rapidly in recent years. Polymeric and MIEC ceramic membrane system requires neither
ceramic membranes have been used commercially electrodes nor an external circuit to operate. As
for oxygen production. Polymeric membranes depicted in Fig. 1, the electronic conductivity
operate based on the difference in rates of diffu- itself performs as an internal short circuit that
sion of oxygen and nitrogen through a membrane. involves oxygen partial pressure gradient. The
Due to the smaller size of the oxygen molecule, oxygen molecule permeates from the high oxygen
most membrane materials are more permeable to partial pressure side to the low oxygen partial
oxygen than to nitrogen. Materials such as pressure side, while the overall charge neutrality
polysulfone or acetate membranes make it is maintained by counterbalancing the ux of
possible to permeable oxygen ve times than of electrons (Liu and Gavalas 2005). Several indus-
nitrogen. Membrane units capable of producing trial gas companies are working on developing
nearly 600 tonnes per day nitrogen have been ceramic membranes for oxygen separation from
built (Castle 2002). A major benet of polymeric air at high temperatures. Air Products and
membrane separation is the simple, continuous Chemicals has developed an ion transport mem-
nature of the process and operation at near brane (ITM) system, which is based on patented,
ambient conditions (Smith and Klosek 2001); high-temperature ceramic membranes for the pro-
however, the low separation factor of two to six duction of oxygen from air separation (Hashim
limits the polymeric membrane to produce et al. 2011). Praxair is also working on oxygen-
oxygen-enriched air rather than pure oxygen conducting ceramic membrane systems that are
(Zhu et al. 2008). specially designed to separate oxygen from air at
Production of pure oxygen from air can be elevated temperature environment (Hashim
achieved by using ceramic membrane system at et al. 2011).
elevated temperatures, typically in the range of
800900 C (Hashim et al. 2011). The oxygen
transporting through this type of membrane is in Cross-References
the form of oxygen ions instead of oxygen mole-
cules; therefore, the pure oxygen is obtained. Enor- Oxygen Separation
mous efforts have been directed to ceramic
membranes with mixed ionicelectronic References
conducting (MIEC) characteristics. Among them,
perovskite-type (ABO3) ceramic membranes Castle WF (2002) Air separation and liquefaction: recent
exhibit the highest oxygen permeability due to developments and prospects for the beginning of the
their high ionic and electronic conductivity. The new millennium. Int J Refrig 25:158172
Hashim SS, Mohamed AR, Bhatia S (2011) Oxygen sep-
perovskite oxide based on La1xAxCo1yFeyO3d aration from air using ceramic-based membrane tech-
achieved very high oxygen permeation uxes as nology for sustainable fuel production and power
reported by Teraoka et al. (1985). generation. Renew Sustain Energy Rev 15:12841293
Alamine 336 41
Liu S, Gavalas GR (2005) Oxygen selective ceramic hol- electrodeposition and alumina membrane etching
low ber membranes. J Membr Sci 246:103108 (Burggraaf and Cot 1996; Lee and Pope 1994).
Smith AR, Klosek J (2001) A review of air separation
technologies and their integration with energy conver- The ltration process and therefore the elimi- A
sion processes. Fuel Process Technol 70:115134 nation of different species from polluted water, for
Teraoka Y, Zhang HM, Furukawa S, Yamazoe N (1985) example, strongly depend on the type of mem-
Oxygen permeation through perovskite-type oxides. brane used, its porosity, and pore size. As
Chem Lett 167:17431746
Zhu X, Sun S, He Y, Cong Y, Yang W (2008) New concept recommended by IUPAC, porous membranes are
on air separation. J Membr Sci 323:221224 classied according to their connected pore size
being microporous for pore diameters lower than
2 nm, mesoporous for pore diameters in the range
250 nm, and macroporous for pore sizes superior
Al2O3 Porous Membranes than 50 nm (Koros et al. 1996). Respectively,
porous membranes will be applied to separation
Sophie Cerneaux processes such as nanoltration, ultraltration,
UMR 5635 ENSCM, CNRS, UM, Institut and microltration.
Europen des Membranes, Montpellier, France The synthesis of macroporous and mesoporous
membranes is often made by consolidation of
particles followed by thermal removal of organic
Synonyms additives. Going to nanoltration application, the
microporous network is directly prepared from a
Aluminum (III) oxide porous membranes or alu- sol containing particles lower than 10 nm
mina porous membranes followed by an appropriate sintering step.
Characteristics References
Al2O3 is an amphoteric oxide that exhibits poly- Burggraaf AJ, Cot L (1996) Fundamentals of inorganic
membrane science and technology. Elsevier, Amsterdam
morphic crystalline forms upon temperature treat-
Koros WJ, Ma YH, Shimidzu T (1996) Terminology for
ment. A wide variety of alumina powders are membranes and membrane processes. Pure Appl Chem
commercially available with different particle 68(7):14791489
sizes. Lee BI, Pope EJA (1994) Chemical processing of
ceramics. Marcel Dekker, New York
Alumina porous membranes are commercially
available inorganic ceramic membranes widely
used in membrane processes related to water treat-
ment essentially. They are shaped in monolith, Alamine 336
honey comb, at, monochannel, and multichannel
congurations or as thin lms. Karel Friess
Alumina membranes may be prepared by Department of Physical Chemistry, University of
anodic oxidation of aluminum foil to design Chemistry and Technology Prague, Prague,
well-ordered and straight pore channels in the Czech Republic
membrane or by sol-gel route using colloids or
Al molecular precursors to yield porous supported
membranes. In addition, alumina macroporous Synonyms
supports are also used for multilayered structures
of other oxide or non-oxide materials. More Alamine 3365; Alamine 336S; Farmin 08;
recently, alumina porous membranes found a Octanamine; Tricaprylamine; Tridioctylamine;
new application as hard template for the fabrica- Tri-n-octylamine; Tri-n-caprylylamine;
tion of nanotubes and nanowires through Trioctylamine
42 Alamine 336
Alamine 336,
Fig. 1 Schematic structure
of Alamine 336
N
The IUPAC systematic name is N,N-dioctyl-1- Alamine 336, Table 1 Properties of Alamine 336 (Steele
octanamine (Fig. 1). Alamine 336 is colorless et al. 1996)
(C24H51N, CAS Reg. No. 1116-76-3) or light Molar mass 353.67 gmol1

yellow liquid (mixture of Tri C8-10 Alkyl Melting point 34.6 C

Amines, C27H57N CAS Reg. No. 57176-40-6, Boiling point 363.5 C
produced by Cognis Corp., now part of BASF). If Density 0.818 gcm3 at 25 C
released to air, estimated vapor pressure is about Viscosity 7.86102 Pas at 34.6 C
3.48102 Nm1 at 34.6 C
0.007 Pa at 25 C. It is moderately toxic by inges- Surface tension
Constant pressure 750.8 Jmol1K1 at
tion and intraperitoneal routes; when heated to
heat capacity of 25 C
decomposition, it emits toxic vapors of NOx liquid
(Zhu et al. 2012). Manufacturing of Alamine Solubility in 0.1 gcm3 at 25 C
336 is possible by catalytic amination of octanol chloroform
(Li et al. 2011) or by catalytic hydrogenation of Solubility in water 0.050 mgdm3 at 25 C
caprylonitrile (Lazier 1940). It is used as the
extractant for reactor fuel processing (Moyer and
References
McDowell 1981), for dye identication
(Puttemans et al. 1982), and for metal adsorption ChemSpider (CSID). http://www.chemspider.com/
(Tasker et al. 2003) and recovery from diluted Chemical-Structure.13591.html. Accessed 1 June 2012
aqueous solutions (Kislik 2012; Coca Coca J, Dez FV, Mors MA (1990) Solvent extraction of
molybdenum and tungsten by Alamine 336 and DEHP-
et al. 1990; Sun and Lee 2011). Mixtures of
A. Hydrometallurgy 25:125135
Alamine 336 with meta-xylene can be success- Dygiel P, Wieczorek PP (2010) Supported liquid membranes
fully applied for extraction metal ions from their and their modications: denition, classication, theory,
strong acidic aqueous chloride solutions (Sayar stability, application and perspectives. In: Kislik VS
(ed) Liquid membranes. Elsevier, Amsterdam, pp 73140
et al. 2007). Alamine 336 can be also used for
Kislik VS (2012) Examples of application of solvent
specic extractions in biotechnology, e.g., separa- extraction techniques in chemical, radiochemical, bio-
tion of carboxylic acids (Tamada et al. 1990; chemical, pharmaceutical, analytical separations, and
Yordanov and Boyadzhiev 2004), and in wastewater treatment. In: Kislik VS (ed) Solvent
extraction classical and novel approaches. Elsevier,
supported liquid membranes (San Romn
Amsterdam, pp 185314
et al. 2010; Dygiel and Wieczorek 2010) Lazier WA (1940) Process for catalytic hydrogenation of
(Table 1). higher aliphatic nitriles. US Patent 2225059
Albumin Purification with Affinity Membranes 43
Li Y, Li Q, Zhi L, Zhang M (2011) Catalytic amination of

octanol for synthesis of trioctylamine and catalyst char- Alamine 336S
acterization. Catal Lett 141:16351642
Moyer BA, McDowell WJ (1981) Factors inuencing A
phase disengagement rates in solvent extraction sys- Alamine 336
tems employing tertiary amine extractants. Sep Sci
Technol 16:12611289
National Institute of Standards and Technology (NIST)
database. http://webbook.nist.gov. Accessed 1 June
2012 Albumin Purification with Affinity
Puttemans ML, Dryon L, Massart DL (1982) Evaluation of Membranes
thin layer-, paper- and high performance liquid chro-
matography for the identication of dyes extracted as
ion-pairs with tri-n-octylamine. J Assoc Off Anal Chem Adil Denizli and Handan Yavuz
65:730736 Department of Chemistry, Biochemistry Division,
San Romn MF, Bringas E, Ibaez R, Ortiz I (2010) Hacettepe University, Ankara, Turkey
Liquid membrane technology: fundamentals and
review of its applications. J Chem Technol Biotechnol
85:210
Sayar NA, Filiz M, Sayar AA (2007) Extraction of Zn Nearly 175 years ago, a very accurate description
(II) from aqueous hydrochloric acid solutions into of the properties of human serum albumin (HSA)
Alamine 336m-xylene systems. Modeling consider- was available (Ancell 1839). HSA is one of the
ations to predict optimum operational conditions.
Hydrometallurgy 86:2736 most abundant proteins in plasma and, together
Steele WV, Chirico RD, Knipmeyer SE, Nguyen A, Smith with immunoglobulin, constitutes 80 % of all
NK, Tasker IR (1996) Thermodynamic properties and plasma proteins. In addition to plasma, HSA is
ideal-gas enthalpies of formation for cyclohexene, also found in tissues and bodily secretions
phthalan (2,5-dihydrobenzo-3,4-furan), isoxazole,
octylamine, dioctylamine, trioctylamine, phenyl throughout the body. It has several physiological
isocyanate, and 1,4,5,6-tetrahydropyrimidine. J Chem functions in vivo, which contribute signicantly
Eng Data 41:1269-1284 to colloid osmotic blood pressure and aid in the
Sun PP, Lee MS (2011) Separation of Pt(IV) and Pd transport, distribution, and metabolism of many
(II) from the loaded Alamine 336 by stripping. Hydro-
metallurgy 109:181184 endogenous and exogenous substances (He and
Tamada JA, Kertes AS, King CJ (1990) Extraction of Carter 1992). The ethyl alcohol fractionation is
carboxylic acids with amine extractants. Ind Eng the oldest process of industrial fractionation of
Chem Res 29:13191326 blood proteins but it cannot produce sufciently
Tasker PA, Plieger PG, West LC (2003) Metal
complexes for hydrometallurgy and extraction. In: pure albumin (Stotz et al. 1990). Of all purica-
McCleverty JA, Meyer TJ (eds) Comprehensive coor- tion methods, chromatography may be the best
dination chemistry II, vol 9. Elsevier: Oxford, UK,. candidate for minimizing process-generated albu-
pp 759808 min heterogeneity. Afnity chromatography is a
Yordanov B, Boyadzhiev L (2004) Pertraction of citric acid
by means of emulsion liquid membranes. J Membr Sci well-established method for the purication of
238:191197 proteins, and it is based on highly specic molec-
Zhu Y, Cao M, Ma X, Xu C, Wang X, Ren L, Hu C (2012) ular recognition. Conventional packed-bed col-
Dalton Trans 41(10):293540 umn chromatography is often limited by long
processing times due to low ow-rates, which
result from the high pressure drop (Denizli
et al. 1999). In order to overcome these opera-
tional limitations, microporous membrane
adsorbers have been developed (Ghosh 2002).
Alamine 3365 Microporous membranes as a support for chroma-
tography have several potential advantages.
Alamine 336 A conguration in which the feed solution ows
44 Alcohol and Water Separation
Albumin Purification with Affinity Membranes, Table 1 Albumin adsorption capacities for various membrane
affinity adsorbents
Adsorbent Ligand Ligand loading Qmax References
Polyethylene hollow ber CB F3GA 52 mmol/mL 6.5 mg/mL (Wolman et al. 2000)
PTFE membrane CB F3GA 89.8 mmol/g 85.3 mg/g (Gu et al. 2007)
P(GMA-DMAA) hollow ber CB F3GA 117 mmol/g 15.3 mg/mL (Wolman et al. 2005)
Polyamide hollow ber R. Green HE4BD 39.4 mmol/g 86.7 mg/g (Yavuz and Denizli 2004)
PTFE membrane CB F3GA/Zn2+ not reported 198.5 mg/g (Gu et al. 2007)
Polyamide hollow ber Cu2+ 150 mmol/g 226.0 mg/g (Uzun and Denizli 2002)
Ni2+ 250 mmol/g 289.0 mg/g (Uzun and Denizli 2002)
Co2+ 250 mmol/g 195.0 mg/g (Uzun and Denizli 2002)
through the membrane provides a very short, wide (EGDMA-HEMA) microbeads. J Appl Polym Sci
bed; thus, high velocities and very short residence 74:28032810
Ghosh J (2002) Protein separation using membrane chroma-
times are attainable with modest transmembrane tography: opportunities and challenges. J Chromatogr
pressure drops. Elimination of diffusional resis- A 952:327
tance usually leaves a system controlled by much Gu J, Lei Z, Qizhi Y (2007) Novel method for human
faster binding kinetics, thereby enabling adsorp- serum albumin adsorption/separation from aqueous
solutions and human plasma with Cibacron Blue
tive separation of proteins, typically, one-tenth the F3GA-Zn(II) attached microporous afnity membra-
time common for packed columns. Adsorptive nous capillaries. J Membr Sci 287:271279
membranes exhibit high binding capacities simi- He XM, Carter DC (1992) Atomic structure and chemistry
lar in magnitude to packed columns and using of human serum albumin. Nature 358:209215
Stotz JF, Rivat C, Geschier C, Colosett P, Streiff F (1990)
stepwise elution, proteins can rapidly be concen- Chromatography purication of a high purity human
trated tenfold or more with 85100 % recovery. plasmatic albumin for clinical or biological uses. Swiss
Analytical separations equivalent to those of col- Biotechnol 8:710
umn chromatography are reached by using suf- Uzun L, Denizli A (2002) Metal-chelated polyamide hol-
low bres for human serum albumin separation. J Appl
cient numbers of membrane stacks and gradient Polym Sci 86:33463354
elution methods. The linear scalability of the Wolman FJ, Graselli M, Smolko EE, Cascone O (2000)
membrane systems further adds to the attractive- Preparation and characterisation of cibacron blue
ness of the technology. When dealing with vis- F3GA poly(ethylene) hollow-bre membranes.
Biotechnol Lett 22:14071411
cous mediums such as blood, mass transfer in a Wolman FJ, Smolko EE, Cascone O, Grasselli M (2005)
diffusive-transport mode is more efcient, where Improved hollow-bre membranes for dye-afnity
the liquid is made to ow tangentially past the chromatography. J Sep Sci 28:4551
membrane surface while the adsorbate diffuses Yavuz H, Denizli A (2004) Dye afnity hollow bers for
albumin purication. Macromol Biosci 4:8491
into the microporous membrane to meet the selec-
tive ligand immobilized there. Several ligands can
be used for the afnity separation of albumin as
summarized in Table 1. Alcohol and Water Separation
Tadashi Uragami
Organization for Research and Development of
References Innovative Science and Technology (ORDIST),
Kansai University, Suita, Osaka, Japan
Ancell H (1839) Course of lectures on the physiology and
pathology of the blood and other animal uids. Lancet
1:222230
Cross-linked poly(vinyl alcohol) (PVA) composite
Denizli A, Kkturk G, Yavuz H, Pikin E (1999) Dye
ligand column chromatography: albumin adsorption membranes have been used in commercial PV
from aqueous media and human plasma with poly plants for dehydration of ethanol beyond the
Alcohol and Water Separation 45
azeotrope. However aqueous ethanol solutions that permeate through a poly(dimethylsiloxane)

can be produced by bio-fermentation are dilute (PDMS) membrane during PV and that sorbed
(about 10 wt.%). Therefore, if ethanol/water- into a PDMS membrane. These results support A
selective membranes with high efciency can be the hypothesis that the difference in the solubility
prepared, the distillation process in the rst stage to of the permeants contributes to the ethanol/water
obtain an azeotrope can be replaced ethanol-/ selectivity. PDMS membranes show high ethanol/
water-selective membrane which is very advanta- water selectivity, but their mechanical strength is
geous for reduction of energy cost. There are fewer weak, and it is difcult to prepare thin membranes
reports on ethanol-/water-selective membranes from PDMS. In order to obtain both ethanol/water
compared with those of water-/ethanol-selective selectivity and mechanical strength, graft copoly-
membranes. One reason why the development mers composed of PDMS macromonomer and
of efcient high-performance ethanol-/water- vinyl monomers were synthesized.
selective membranes is difcult can be attributed Graft copolymer membranes, which were
to the fact that ethanol has a larger molecular size either ethanol or water selective, were prepared
than water and must be preferentially permeated by copolymerization of a dimethylsiloxane
through the membrane. In fact, permeation and (DMS) macromonomer with methyl methacrylate
separation in a pervaporation (PV) process through (MMA) (Miyata et al. 1995, 1996). Two glass
dense membranes are based on the solution- transition temperatures (Tg) were observed at
diffusion mechanism (Binding et al. 1961; Aptel about 120 C and 127 C in the graft copolymer
et al. 1974). Therefore, when it is required that membranes. Transmission electron micrograph
ethanol molecules with larger molecular size pref- (TEM) demonstrated that the PMMA-g-PDMS
erentially permeate from an aqueous ethanol solu- membranes showed microphase-separated struc-
tion, it cannot be expected to be separated by the tures. When an aqueous solution of 10 wt.% eth-
diffusion process. Consequently, only a difference anol was permeated through the PMMA-g-PDMS
of solubility selectivity in the solution process in membranes by PV, the ethanol concentration in
which both ethanol and water components are the permeate and the permeation rate increased
dissolved can contribute to the separation. drastically with the DMS content in the copoly-
Figure 1 shows the ethanol concentration in the mer. In particular, at a DMS content of less than
100 100
ve
cur
EtOH in membrane (wt%)
80 80
EtOH in permeate (wt%)
ibrium
equl
id
or-liqu
vap
60 60
40 40
20 20
0
0 20 40 60 80 100
EtOH in feed solution (wt%) Alcohol and Water Separation, Fig. 2 Effects of the
DMS content on the normalized permeation rate () and
Alcohol and Water Separation, Fig. 1 Permeation and separation factor () through the PMMA-g-PDMS mem-
separation characteristics of aqueous ethanol solutions brane during PV. Feed: aqueous solution of 10 wt.% etha-
through a PDMS membrane during PV nol. Dashed line is the feed composition
46 Alcohol and Water Separation
Alcohol and Water Separation, Table 1 Performance of ethanol-/water-selective membranes

Feed Applied NPRa
(wt temperature aEtOH/ (kgmm
Membrane %) Method ( C) H2O (m2h)1) References
PDMS 7 PV 25 11.8 2.1 Eustache and Histi (1981)
PTMSP 7 PV 25 11.2 1.1 Ishihara et al. (1986)
PTMSP 10 PV 30 12 4.5 Masuda et al. (1986)
PEA-g-PDMS/ 10 PV 40 20 24.1 Uragami et al. (2000);
PTMSPb Uragami and Shinomiya
(1991)
PPP-g-PDMS 7.28 PV 30 22.5 5.5 Nagase et al. (1989)
PSt-g-PhdFDA (7.6/ 8 PV 30 45.9 0.6 Ishihara and Matsui (1987)
12.4)
TFE/i-OcVE/C18VE 15 PV 50 7.13 5 Kashiwagi et al. (1988)
terpolymer (50/25/25)
Modied silicone 10 PV 40 3.65 11 Uragami and Shinomiya
(1991)
Modied silicone 10 TDEV 30/40 19.3 16.6 Miyata et al. (1996)
PDMS 10 PV 40 7.44 6.4 Miyata et al. (1995, 1996)
PDMS 10 TDEV 30/40 85.7 0.9 Uragami and
Shinomiya (1992)
PMMA-g-PDMS 10 PV 40 7.1 4.8 Miyata et al. (1996)
(34/66)
PMMA-g-PDMS 10 PV 40 8 5.1 Miyata et al. (1999a)
(27/73)
PMMA-g-PDMS 10 PV 40 6.8 3.5 Miyata et al. (1999b)
(38/62)c
PTMST 10 TDEV 0/40 77.5 38 Uragami (2010)
Porous PDMS 10 TDEV 0.5 23.1 1,250 Uragami (2008)
a
Normalized permeation rate
b
PFA-g-PDMS is 0.2 wt.%
c
Annealing is 120 C, 2 h
40 mol%, water permeates preferentially from an In Table 1, the performance of the ethanol-/
aqueous solution of 10 wt.% ethanol, whereas water-selective polymer membranes is compared.
membranes with more than about 40 mol% of As can be seen in this Table, the addition of
DMS are ethanol/water selective, as shown in PFA-g-PDMS to the PTMSP membrane in PV
Fig. 2. The change in the ethanol/water selectivity was very effective, the application of TDEV
of the PMMA-g-PDMS membranes can be method to the membrane separation technique
explained by a microphase-separated polymer was also very interesting for the enhancement of
structure using Maxwells model and a combined the ethanol/water selectivity for the ethanol/water
model consisting of both parallel and series expres- mixtures, and in particular the application of
sions. Furthermore, image processing of TEMs porous PDMS membrane to temperature-
allowed the determination of the percolation tran- difference controlled evapomeation (TDEV)
sition of the PDMS phase at a DMS content of (Uragami 1998, 2005, 2006a, b, 2008, 2010,
about 40 mol%. These results suggest that the con- 2011; Uragami and Shinomiya 1991, 1992;
tinuity of the PDMS phases in the microphase- Uragami and Tanaka 1991, 1993, 1994; Uragami
separated PMMA-g-PDMS membranes directly et al. 2002; Uragami and Morikawa 1989) was a
affects their ethanol/water selectivity for aqueous very excellent performance for the ethanol/water
ethanol solutions (Miyata et al. 1995, 1996). mixtures.
Alcohol and Water Separation 47
References Uragami T (1998) Structures and properties of membranes

from polysaccharide derivatives. In: Dumitriu S (ed)
Aptel P, Cuny J, Jozenfonvice J, Morel G, Neel J (1974) Polysaccharide, structural diversity and functional ver-
Liquid transport through membranes prepared by satility. Marcer Dekker, New York/Basel/Hong Kong, A
grafting of polar monomer onto poly(tetrauor- pp 887924
oethylene) lms. II. Some factors determining Uragami T (2005) Structures and functionalities of mem-
pervaporation rate and selectivity. J Appl Polym Sci branes from polysaccharide derivatives. In: Dumitriu
18:365 S (ed) Polysaccharide, structural diversity and func-
Binding RC, Lee RJ, Jennings JF, Mertic EC (1961) Sepa- tional versatility, 2nd edn. Marcer Dekker, New York,
ration of liquid mixtures by pervaporation. Ind Eng pp 10871122
Chem 53:47 Uragami T (2006a) Separation materials derived from chi-
Eustache H, Histi G (1981) Separation of aqueous organic tin and chitosan. In: Uragami T, Tokura S (eds) Material
mixtures by pervaporation and analysis by mass spec- science of chitin and chitosan. KODANSHA, Springer,
trometry or a coupled gas chromatograph-mass spec- Berlin/Heiderberg/New York/Tokyo, pp 113163
trometer. J Membr Sci 8:105 Uragami T (2006b) Polymer membranes for separation of
Ishihara K, Matsui K (1987) Ethanol permselective poly- organic liquid mixtures. In: Yampolskii Y, Pinau I,
mer membranes 3. Pervaporation of ethanol water mix- Freeman BD (eds) Materials science of membranes
ture through composite membranes composited of for gas and vapor separation. Wiley, Chichester,
styrene uoroalkyl acrylate graft copolymers and pp 355372
cross-linked polydimethylsiloxane membrane. J Appl Uragami T (2008) Structural design of polymer mem-
Polym Sci 34:437 branes for concentration of bio-ethanol. Polym
Ishihara K, Nagase Y, Matsui K (1986) Pervaporation of J 40:485
alcohol/water mixtures through poly[1-(trimethylsilyl)- Uragami T (2010) Selective membranes for purication
1-propyne] membrane. Macromol Chem Rapid and separation of organic liquid mixtures. In:
Commun 7:43 Drioli E, Giorno L (eds) Comprehensive membrane
Kashiwagi T, Okabe K, Okita K (1988) Separation of science and engineering, vol 2, Membrane operations
ethanol from ethanol/water mixtures by plasma- in molecular separations. Elsevier, Amsterdam/Boston/
polymerized membranes from silicone compounds. Heidelberg/London/New York/Oxford/Paris/San
J Membr Sci 36:353 Diego/San Francisco/Singapore/Sydney/Tokyo,
Masuda T, Tang B-Z, Higashimura T (1986) Ethanol-water pp 273324
separation by pervaporation through substituted- Uragami T (2011) Concentration of bio-ethanol through
polyacetylene membranes. Polym J 18:565 cellulose ester membranes during temperature-
Miyata T, Takagi T, Kadota T, Uragami T (1995) Charac- difference controlled evapomeation. Mater Sci Appl
teristics of permeation and separation for aqueous eth- 2:1499
anol solutions through methyl methacrylate- Uragami T (2012) Concentration of bio-ethanol through
dimethylsiloxane graft copolymer membranes. poly[1-(trimethylsilyl)-1-propyne] membranes during
Macromol Chem Phys 196:1211 temperature-difference controlled evapomeation. 11th
Miyata T, Takagi T, Uragami T (1996) Microphase sepa- World Filtration Congress & Exhibition, p 313
ration in graft copolymer membranes with pendant Uragami T, Morikawa T (1989) Permeation of ethanol
oligodimethylsiloxanes and their permselectivity for through polydimethyl siloxane membranes using
aqueous ethanol solutions. Macromolecules 29:7787 temperature-difference in membrane process of the
Miyata T, Obata S, Uragami T (1999a) Morphological evapomeation method. Makromol Chem Rapid
effects of microphase separation on the permselectivity Commun 10:287
for aqueous ethanol solutions of block and graft copol- Uragami T, Morikawa T (1992) Permeation and separation
ymer membranes containing poly(dimethylsiloxane). characteristics of alcohol-water mixtures through
Macromolecules 32:3712 dimethylsiloxane membrane by pervaporation and
Miyata T, Obata S, Uragami T (1999b) Annealing effect of evapomeation. J Appl Polym Sci 44:2009
microphase-separated membranes containing poly Uragami T, Shinomiya H (1991) Concentration of aqueous
(dimethylsiloxane) on their permselectivity for aqueous alcoholic solutions through a modied silicone rubber
ethanol solutions. Macromolecules 32:8465 membrane by pervaporation and evapomeation.
Nagase Y, Mori S, Matsui K (1989) Chemical modication Makromol Chem 192:2293
of poly(substituted-acetylene). 4. Pervaporation of Uragami T, Tanaka Y (1991) Method of separating liquid
organic liquid water mixture through poly(1-phenyl- component from a solution containing two or more
1-propylene) polydimethylsiloxane graft copolymer liquid component. European Patent 0,346,739
membrane. J Appl Polym Sci 37:1259 Uragami T, Shinomiya H (1992) Concentration of aqueous
Uragami T (1993) Separation of aqueous organic liquid dimethyl sulfoxide solutions through a chitosan mem-
solutions through polymer membranes. Desalination brane by permeation with a temperature difference.
90:325 J Membr Sci J Membr Sci 7:183
48 Alcohol Removal by Membrane Operations
Uragami T, Tanaka Y (1993) Method of separating liquid of dilute alcohol solutions (excluding methanol).
component from a solution containing two or more Especially the use of selective membranes
liquid component. U.S. Patent 5,271,846
Uragami T, Tanaka Y (1994) Separation method for mixed for ethanol or n-butanol removal from dilute fer-
solutions. Japanese Patent 1,906,854 mentation broths has gained a lot of interest
Uragami T, Doi T, Miyata T (1999) Control of recently.
permselectivity with surface modications of poly
[1-(trimethylsilyl)-1-propyne] membranes. Inter
J Adhes Adhes 19:405
Uragami T, Doi T, Miyata T (2000) Pervaporation property Membrane Operations
of surface modied poly[1-(trimethylsilyl)-1propyne]
membranes. In: Pinnau I, Freeman BD (eds) Membrane The most used membrane operations for alcohol
formation and modication, ACS Symposium Series
744. American Chemical Society, Washington, DC, removal include separation techniques like ltra-
pp 263279 tion, membrane distillation (also known as mem-
Uragami T, Tanaka Y, Nishida S (2002) Permeation and brane evaporation), pertraction, pervaporation
separation under high temperature and high pressure (PV), and vapor permeation (VP). Micro-, nano-,
for ethanol/water vapors through cross-linked
quaternized chitosan composite membranes. Desalina- and ultraltration are used in solid-liquid separa-
tion 147:449 tions mainly as a pretreatment step prior to the
alcohol removal. In pertraction, the membrane is
placed between fermentation broth and extractant.
The membrane passes through the alcohol
enabling its extraction, while other components
Alcohol Removal by Membrane such as nutrients and intermediate products remain
Operations in the feed side. In pervaporation and vapor per-
meation, desired minor compounds from a feed
Johanna Niemist solution are diffused through the hydrophobic
Helsinki Region Chamber of Commerce, membrane and desorbed in the permeate side of
Enterprise Europe Network, Helsinki, Finland the membrane as a vapor which can then be con-
densed back to the liquid state by, e.g., a condenser
(Schfer and Crespo 2005). Feed solution is in a
Synonyms liquid phase in PV applications and in a vapor
phase in VP processes. The driving force, a chem-
Alcohol separation by Membrane Distillation ical potential difference of the compounds, is
(MD); Alcohol separation by Pertraction; Alcohol obtained by applying different partial vapor pres-
separation by Pervaporation (PV); Alcohol sepa- sures of compound(s) on the opposite sides of the
ration by Vapor permeation membrane: The feed side is under atmospheric
pressure, whereas vacuum or very low pressure
Biochemical, i.e., fermentative, production of is applied in the permeate side. Because only per-
alcohols and other solvents has been increasing meated compounds are evaporated in
signicantly in the last years. Recovery and puri- pervaporation, lower energy, capital, and
cation of alcohols from dilute aqueous solutions processing costs can be obtained as compared to
obtained after fermentative solvent production is other separation techniques.
energy intensive and expensive. For example,
conventional distillation is stated to be an efcient
alcohol separation technique for the ethanol in Membranes
concentration range from 10 to 85 wt.% (Huang
et al. 2008). The use of membrane operations is Membranes used in alcohol removal are hydro-
however seen as a more cost- and energy-efcient phobic, i.e., organophilic, having preferential per-
method over conventional distillation in the case meation for organic compounds instead of water.
Alcohol Separation by Membrane Distillation (MD) 49
Nonporous membranes are used in PV and VP concentration of around 92.4 wt% (below the
applications, while membrane distillation opera- azeotropic point), and further dehydration is
tions are performed with semipermeable porous done by other separation techniques including, A
membranes. The semipermeable polymeric mem- e.g., adsorption by molecular sieves. Especially,
branes used mostly are made of polydimethyl- pervaporation and vapor permeation are more
siloxane (PDMS). Other used materials include, energy-efcient and promising techniques, but
e.g., polyether block amide (PEBA), only a few number of pilot- or industrial-scale
polyoctylmethyl siloxane (POMS), polypropyl- facilities for alcohol separation are in use, despite
ene (PP), polytetrauoroethylene (PTFE), the active research going on in laboratory scale.
polytrimethylsilylpropyne (PTMSP), Hence, there is still need for research and devel-
polyvinylidene uoride (PVDF), and liquid mem- opment activities to obtain feasible industrial-
branes. Ceramic and zeolite membranes are also scale applications.
available. Composite membranes having a porous
support layer and a thin separation layer are
widely used due to better separation efciency Cross-References
and membrane stability. Besides a good chemical,
mechanical, and thermal stability, membranes Alcohol and Water Separation
should also have high ux and selectivity toward Butanol-Water Mixtures: Separation by
alcohols. Membrane modules can be in the form Pervaporation
of at sheets or plates, spiral wounds, envelopes, Ceramic Membranes
or (hollow) tubular modules (Bruschke 2001). Fouling in Membranes
Pervaporation (PV)
Polarization in Membrane Operations
Advantages and Disadvantages
References
Advantages of membrane operations include the
possibility to be used also in the case of azeotropic Br
uschke HEA (2001) State-of-the-art of pervaporation
mixtures and for the separation of volatile organic processes in the chemical industry. In: Nunes SP,
compounds. In addition, combining the mem- Peinemann K-V (eds) Membrane technology in the
brane operation with a fermentor as a hybrid pro- chemical industry, 1st edn. Wiley VCH Verlag GmbH
& Co, Weinheim, pp 127172
cess as in situ product removal (ISPR) enables Huang H-J, Ramaswamy S, Tschirner UW, Ramarao BV
obtaining an increased yield in fermentation and (2008) A review of separation technologies in current
more efcient separation efciency. Other hybrid and future bioreneries. Sep Purif Technol 62:121
processes are typically a membrane operation Schfer T, Crespo JG (2005) Vapor permeation and
pervaporation. In: Afonso CAM, Crespo JPG (eds)
combined with distillation or with other mem- Green separation processes: fundamentals and applica-
brane technique. The main disadvantage in the tions. Wiley VCH Verlag GmbH & Co, Weinheim,
operation is the fouling tendency of the membrane pp 271289
surface or of the pores inside the membrane, in Vane LM (2005) A review of pervaporation for product
recovery from biomass fermentation processes. J Chem
addition to concentration polarization. Fouling Technol Biotechnol 80:603629
can be reduced by improved membrane module
design or operating procedures (Vane 2005).
Industrial Applications Alcohol Separation by Membrane

Distillation (MD)
In commercial scale, ethanol is usually rst
dehydrated by distillation up to the ethanol Alcohol Removal by Membrane Operations
50 Alginate
necessary to have an aqueous solvent free of

Alginate cross-linking ions to enable dissolution. The sol-
ubility of alginates in organic media requires for-
Emma Piacentini mation of a tetrabutylammonium (TBA) salt. The
Institute on Membrane Technology, National solubility of alginates depends strongly on the
Research Council of Italy, ITM-CNR, Rende, state of the backbone carboxylic acid groups.
Italy Alginic acid with its carboxylic acid groups in
their protonated form was not fully soluble in
any solvent system examined, including water.
Alginates are unbranched anionic polysaccharides
Chemical modication or derivatization of
consisting of 1!4 linked b-D-mannuronic acid
alginates is used as a tool to attain one of two
(M) and its C-5 epimer a-L-guluronic acid
ends, i.e., (a) enhance existing properties (such
(G) (Fig. 1).
as improvement of ionic gel strength by additional
The natural copolymer is an important compo-
covalent cross-linking, increase hydrophobicity
nent of algae such as kelp and is also an
of the backbone, improve biodegradation, or
exopolysaccharide of bacteria including Pseudo-
achieve greater HAP nucleation and growth) or
monas aeruginosa. The copolymer composition,
(b) introduce completely new properties other-
sequence, and molecular weights vary with the
wise not existing in unmodied alginates (such
source and species that produce the copolymer.
as afford anticoagulant properties, provide chem-
The molecular weights of commercially available
ical/biochemical anchors to interact with cell sur-
sodium alginates range between 32,000 and
faces, or bestow temperature-dependent
400,000 g/mol. The viscosity of alginate solutions
characteristics such as lower critical solution tem-
increases as pH decreases and reaches a maximum
perature). Alginate derivatization (Pawar and
around pH 33.5, as carboxylate groups in the
Edgar 2012) includes acetylation, phosphoryla-
alginate backbone become protonated and form
tion, sulfation, hydrophobic modication, attach-
hydrogen bonds. The solubility of alginates in
ment of cell signaling molecules, covalent cross-
water is governed by three parameters, i.e.,
linking, and graft copolymerization. In particular,
(a) pH of the solvent, (b) ionic strength of the
amphiphilic alginate derivatives have been syn-
medium, and (c) the presence of gelling ions in
thesized by introducing hydrophobic moieties
the solvent. To make alginates soluble, it is essen-
(e.g., alkyl chains, hydrophobic polymers) to the
tial that the pH be above a certain critical value
alginate backbone. These derivatives can form
and the carboxylic acid groups be deprotonated.
self-assembled structures such as particles and
Changing the ionic strength of the medium affects
gels in aqueous media and have potential in
solution properties such as polymer conformation,
many drug delivery applications. Alginate deriv-
chain extension, viscosity, and therefore solubil-
atives containing cell-adhesive peptides have
ity. Alginates gel in the presence of divalent cat-
been gaining signicant attraction recently.
ions such as Ca2+, Sr2+, and Ba2+. It is therefore
These derivatives are typically prepared by chem-
ically introducing peptides as side chains, using
O
carbodiimide chemistry to couple via the carbox-
ylic groups of the sugar residues. Since alginate
HO C
inherently lacks mammalian cell adhesivity,
-D-mannuronic acid (M)
O OH appropriate ligands are crucial to promote and
HO O regulate cellular interactions, especially for cell
O O
O culture and tissue engineering applications. Pep-
HO C OH
OH tides including the sequence arginine-glycine-
O
-L-guluronic acid (G)
aspartic acid (RGD) have been extensively used
as model adhesion ligands, due to the widespread
Alginate, Fig. 1 Components of alginates presence of integrin receptors for this ligand on
Alginate 51
various cell types. RGD-containing peptides can produces more uniform structures and greater
be chemically coupled to the alginate backbone mechanical integrity. One critical drawback of
using water-soluble carbodiimide chemistry. ionically cross-linked alginate gels is the limited A
Alginate is typically used in the form of a long-term stability in physiological conditions,
hydrogel. Hydrogels are three-dimensionally because these gels can be dissolved due to release
cross-linked networks composed of hydrophilic of divalent ions into the surrounding media due to
polymers with high water content. Hydrogels are exchange reactions with monovalent cations and
often biocompatible, as they are structurally sim- the hemostasis promoted by the calcium ions
ilar to the macromolecular-based components in released from the gel that serves as a matrix for
the body, and can often be delivered into the body aggregation of platelets and erythrocytes. Cova-
via minimally invasive administration. Alginate lent cross-linking has been widely investigated in
chelates with divalent cations to form hydrogels. an effort to improve the physical properties of gels
Gel formation is driven by the interactions for many applications, including tissue engineer-
between G blocks which associate to form tightly ing. However, covalent cross-linking reagents
held junctions in the presence of divalent cations. may be toxic, and the unreacted chemicals may
In addition to G blocks, MG blocks also partici- need to be removed thoroughly from gels. The
pate, forming weak junctions. Thus, alginates inability to dissociate and reform bonds leads to
with high G contents yield stronger gels. signicant elastic deformation. Thermosensitive
Chemical and/or physical cross-linking of hydrogels have been widely investigated to date
hydrophilic polymers are typical approaches to in many drug delivery applications, due to their
form hydrogels, and their physicochemical prop- adjustable swelling properties in response to tem-
erties are highly dependent on the cross-linking perature changes, leading to on-demand modula-
type and cross-linking density, in addition to the tion of drug release from the gels. When alginate
molecular weight and chemical composition of is modied with cell adhesion ligands, the ability
the polymers. The most common method to of cells to bind multiple polymer chains can lead
prepare hydrogels from an aqueous alginate to long-distance, reversible network formation
solution is to combine the solution with ionic even in the absence of chemical cross-linking
cross-linking agents, such as divalent cations. agents. Cells added to an RGD-modied alginate
The afnity of alginates toward divalent solution form a uniform dispersion within the
ions decreases in the following order: solution, and this system subsequently generates
Pb > Cu > Cd > Ba > Sr > Ca >Co, Ni, the cross-linked network structure via specic
Zn > Mn. Ca2++; however, it is the most com- receptor-ligand interactions without using any
monly used cation to induce alginate gel forma- additional cross-linking molecules. In contrast,
tion. Calcium cross-linking of alginates can be cells added to non-modied alginate solutions
performed by two methods. The rst is a diffu- aggregate and form a nonuniform structure, due
sion method, wherein cross-linking ions diffuse to the dominance of cell-cell interactions in that
into the alginate solution from an outside reser- system. This gelation behavior is shear reversible
voir. The second is the internal setting method, and can be repeated multiple times. Once the gel
where the ion source is located within the alginate structure is broken down by applying shear forces,
solution and a controlled trigger (typically pH or cross-linked structures are recovered within a few
solubility of the ion source) sets off the release of minute. This behavior is governed by the weak
cross-linking ions into solution. Diffusion set gels and reversible ligand-receptor interactions in the
are typically made by dropping a Na-alginate system. This system might be ideal for cell deliv-
solution into a CaCl2 bath. Internal set gels typi- ery in tissue engineering because a gel can ow
cally use insoluble calcium salts such as CaCO3 as like a liquid during injection into the body but
a calcium source. The gelation rate is a critical solidify once it is placed in the body.
factor in controlling gel uniformity and strength Large pellets (greater than 1.0 mm in diameter)
when using divalent cations, and slower gelation are conventionally prepared using a simple
52 Aligned Carbon Nanotube
syringe or pipette. Droplets of sodium alginate The chemical and physical properties of algi-
solution are extruded into a divalent cross-linking nates resulted in many commercial applications
solution, and pellets formed are then allowed to (Goh et al. 2012; Lee and Mooney 2012):
cure in the cross-linking solution (varying from a
few minutes to hours), rinsed with water, and (i) Food and beverage industry (stabilizers,
air-dried. Micropellets (less than 0.2 mm in diam- thickeners in the preparation of drinks, ice
eter) can be prepared using atomization, emulsi- cream, and jelly; encapsulation material of
cation, and coacervation methods. yeast cells in ethanol production)
Polysaccharides undergo hydrolytic cleavage (ii) Pharmaceutical industry (encapsulation
under acidic conditions. The mechanism of acid material in cell culture and transplantation,
hydrolysis of the glycosidic bond involves three mold in dental impression material,
steps: (1) protonation of the glycosidic oxygen to adhesive agent and sustained release in tab-
give the conjugate acid; (2) heterolysis of the lets, hemostatic and absorbent in wound
conjugate acid forming a nonreducing end group dressing
and a carbonium-oxonium ion; and (3) rapid addi- (iii) Other industries (adhesive agent and ller in
tion of water to the carbonium-oxonium ion, paper industry, stabilizer and suspending
forming a reducing end group. The enzymatic agent in paint industry)
degradation of alginates by lyase occurs by a
b-elimination mechanism resulting in unsaturated
compounds. The mechanism of b-elimination References
involves abstraction of the proton at the C-5 posi-
tion, which is enhanced by the electron- Goh CH, Heng PWS, Chan LW (2012) Alginates as a
useful natural polymer for microencapsulation and
withdrawing effect of the carbonyl group at C-6
therapeutic applications. Carbohydr Polym 88:112
position. Alginates chain degradation can occur Lee KY, Mooney DJ (2012) Alginate: properties
also at neutral pH values in the presence of reduc- and biomedical applications. Prog Polym Sci
ing compounds such as hydroquinone, sodium 37:106126
Pawar SN, Edgar KJ (2012) Alginate derivatization: a
sulte, sodium hydrogen sulde, cysteine,
review of chemistry, properties and applications. Bio-
ascorbic acid, hydrazine sulfate, and leuco- materials 33:32793305
methylene blue that also caused degradation in
alginates. The mechanism of degradation involves
the formation of a peroxide leading to free radical
creation, which eventually causes breakdown of
the alginate chain. Aligned Carbon Nanotube
Alginate is inherently nondegradable in mam-
mals, as they lack the enzyme (i.e., alginase) Cameron Shearer
which can cleave the polymer chains, but School of Chemical and Physical Sciences,
ionically cross-linked alginate gels can be Flinders University, Adelaide, South Australia,
dissolved by release of the divalent ions cross- Australia
linking the gel into the surrounding media due to Institute of Physical Chemistry, University of
exchange reactions with monovalent cations such Muenster, Muenster, North RhineWestphalia,
as sodium ions. An attractive approach to make Germany
alginate degradable in physiological conditions
includes partial oxidation of alginate chains.
Slightly oxidized alginate can degrade in aqueous
media, and these materials have demonstrated Synonyms
potential as a delivery vehicle of drugs and cells
for various applications. Alginate is typically oxi- A-CNT; Horizontally aligned carbon nanotube;
dized with sodium periodate. Vertically aligned carbon nanotube
Aligned Carbon Nanotube 53
Main Text et al. 2008). Horizontal CNTs can form a dense,

exible, stable mat (called a bucky paper) which
Carbon nanotubes (CNTs) are allotropes of car- can potentially be used as a membrane as the space A
bon in which the carbon atoms form a hexagonal between the CNTs is generally quite small
sheet of sp2 bonds (analogous to graphite) which (<50 nm) (Dume et al. 2010; Sears et al. 2010).
is seamlessly rolled up into a cylinder that can be Vertically aligned CNTs can be prepared via
microns long but is only nanometers in diameter the CVD growth of CNTs on pre-patterned sub-
(Yamamoto et al. 2008). Following typical syn- strate supported catalyst nanoparticles. Verti-
thesis procedures, CNTs exist as a tangled web of cally aligned CNTs produced via this method
randomly oriented bers. For many applications, have been grown greater than 1 mm in length
the CNTs are required to be aligned with either (Joselevich et al. 2008). Alignment is enhanced
horizontal or vertical alignment (Fig. 1). via the use of a strong electric eld such as in
Horizontally aligned CNTs are of particular plasma-enhanced CVD. Vertical alignment is of
importance in the generation of eld effect transis- importance in the elds of eld emission, chem-
tors and conductive wires based upon CNTs. CNTs ical sensing, and CNT nanoltration mem-
can be horizontally aligned via a range of processes branes (Endo et al. 2008; Shearer et al. 2012).
including spin coating, ltering, magnetic eld CNT nanoltration membranes are of particular
alignment, and drying-induced and chemical interest as molecular dynamics simulations have
vapor deposition (CVD) (Dai 2001; Joselevich shown that water can ow through CNTs at a
Aligned Carbon Nanotube, Fig. 1 SEM image CNTs produced via CVD showing (a) randomly entangled, (b)
horizontally aligned, and (c) vertically aligned
54 Alkaline Fuel Cells (AFCs)
remarkable rate (Hummer et al. 2001), while the

ow of contaminates is restricted by the nano- Alkaline Fuel Cells (AFCs)
size of the CNT inner pore ( 1 nm). The fabri-
cation of such a membrane required vertically Tanja Vidakovic-Koch
aligned CNTs, encased within an impermeable Max Planck Institute for Dynamics of Complex
matrix such that ow can only occur through the Technical Systems, Magdeburg, Germany
CNT inner pore. This is experimentally quite
difcult, but a few reports with promising
results have been published (Majumder Alkaline fuel cell (AFC) is a fuel cell type which
et al. 2005; Holt et al. 2006; Wu et al. 2010). utilizes alkaline electrolyte, usually potassium
hydroxide. It consumes hydrogen and oxygen
producing only water, heat, and electricity.
Depending on a concentration of potassium
References hydroxide, AFC can operate at temperatures
between 60 C and 250 C. The fuel cell reactions
Dai H (2001) Nanotube growth and characterization. In: are as follows:
Dresselhaus MS, Dresselhaus G, Avouris P (eds) Car-
bon nanotubes, topics in applied physics. Springer-
Verlag, Berlin, pp 2953 Anode : H2 2OH 2H2 O 2e
Dume LF, Sears K, Sch utz J et al (2010) Characterization
and evaluation of carbon nanotube bucky-paper mem-
branes for direct contact membrane distillation.
Cathode : 12 O2 2e 2H2 O 2OH
J Membr Sci 351(12):3643
Endo M, Strano M, Ajayan P (2008) Potential applications Overall reaction : H2 12 O2 H2 O
of carbon nanotubes. In: Jorio A, Dresselhaus G,
Dresselhaus MS (eds) Topics in applied physics: car-
bon nanotubes. Springer, Berlin, pp 1361 The main ionic charge carriers are OH ions
Holt J, Park H, Wang Y et al (2006) Fast mass transport which mitigate from the cathode to the anode.
through sub-2-nanometer carbon nanotubes. Science
312:1034 Water is formed at the anode side and has to be
Hummer G, Rasaiah JC, Noworyta JP (2001) Water con- removed from the system in order to prevent KOH
duction through the hydrophobic channel of a carbon dilution. The AFC has improved cathode perfor-
nanotube. Nature 414(6860):188190 mance compared to acidic fuel cells due to more
Joselevich E, Dai H, Liu J et al (2008) Carbon nanotube
synthesis and organization. In: Jorio A, Dresselhaus G, favorable oxygen reduction reaction kinetics. For
Dresselhaus MS (eds) Carbon nanotubes, topics in this reason, AFC can achieve higher efciency,
applied physics. Springer-Verlag, Berlin, pp 101164 i.e., higher voltage at comparable current densities
Majumder M, Chopra N, Andrews R, Hinds BJ than other fuel cell types. It has very long operat-
(2005) Enhanced ow in carbon nanotubes. Nature
438:44 ing life time, e.g., 15,000 h have been demon-
Sears K, Dume L, Sch utz J et al (2010) Recent develop- strated (Cifrain and Kordesch 2003).
ments in carbon nanotube membranes for water puri- Furthermore, alkaline conditions allow applica-
cation and gas separation. Mater Lett 3(1):127149 tion of nonprecious metal catalysts which could
Shearer C, Fahy A, Barr M et al (2012) Improved eld
emission stability from single-walled carbon nanotubes reduce signicantly the material costs of this fuel
chemically attached to silicon. Nanoscale Res Lett cell. One of major disadvantages of AFC is its low
7(1):432 CO2 tolerance due to formation of carbonates
Wu J, Paudel KS, Strasinger C et al (2010) Programmable according to:
transdermal drug delivery of nicotine using carbon
nanotube membranes. Proc Natl Acad Sci
107(26):1169811702
Yamamoto T, Watanabe K, Hernndez E (2008) Mechan- CO2 2OH CO3 2 H2 O
ical properties, thermal stability and heat transport in
carbon nanotubes. In: Jorio A, Dresselhaus G,
Dresselhaus MS (eds) Carbon nanotubes. Springer, The carbonates have low solubility in strong alka-
Berlin, pp 165194 line environments forming crystals, capable of
Alkane and Alkene Separation by Membrane Operations 55
blocking of electrolyte pathways. The low CO2

tolerance and the application of liquid electrolyte Alkane and Alkene Separation by
are two main hurdles for the broader commercial- Membrane Operations A
ization of these systems. However, due to its high
efciency and high power density, AFC found Chen Zhang
applications in aerospace industry, e.g., they School of Chemical and Biomolecular
were employed on the Apollo missions as well Engineering, Atlanta, GA, USA
as on the Space Shuttle orbitals. For other
examples of developed AFC systems, please
see Cifrain and Kordesch (2003) and Alkenes (e.g., ethene, propene, butadiene, and
Gulzow (2012). butenes isomers) are extremely important building
Hydrogen and oxygen gases in the AFC are blocks in petrochemical industries with many
separated by a membrane. Usually permeable derived chemicals and polymers. Traditionally,
membranes also called diaphragms have been alkenes are recovered from alkene/alkane
used (G ulzow 2012). A common diaphragm mixtures produced by cracking hydrocarbon feed-
material up to 1980s was asbestos which due to stocks (e.g., ethane, LPGs, and naphtha). Alkene/
health and environmental concerns is nowadays alkane separations (e.g., separation of ethene/eth-
abandoned. Alternative materials are different ane, propene/propane, and butadiene/crude C4) are
polymer materials like porous polyethylene traditionally performed on a massive scale by dis-
plates, nonwoven polypropylene, and similar. tillation, which are extremely energy intensive due
Potassium hydroxide electrolyte wets the dia- to these components close volatilities. For exam-
phragm pores in order to ensure its ionic conduc- ple, propene/propane separation was credited to be
tivity. For a good fuel cell performance, the the most energy intensive commercially practiced
electrolyte resistance induced by diaphragm has distillation (Jarvelin and Fair 1993).
to be minimized. This resistance is inuenced by Membrane gas separations, among other com-
the thickness of the diaphragm material, its pore peting alternatives (i.e., adsorptive separation and
size and tortuosity, and the KOH concentration. chemical absorption), can be used to either retrot
Since liquid electrolyte causes some practical or replace existing distillation columns to make
problems in AFC usage, new developments go commercial alkene/alkane separations more
into direction of alkaline polymer electrolyte energy efcient. Compared with well-developed
membranes which, in addition to separation, membrane gas separations such as air separation,
possesses ionic conductivity in the absence of membrane-based alkene/alkane separation is
liquid electrolyte. Currently there is no long-term more challenging. The feeds are highly condens-
stable membrane functioning without a liquid able and usually under high pressure (or in liquid
electrolyte phase. state), requiring that membrane materials must be
sufciently robust without being plasticized.
Among alkene/alkane separations, ethene/ethane
and propene/propane are most extensively studied
References with membranes. Conversely, membranes are less
commonly considered for separations of C4 and
Cifrain M, Kordesch K (2003) Hydrogen/oxygen (Air) fuel heavier alkene/alkane mixtures due to their much
cells with alkaline electrolytes. In: Vielstich W,
Gasteiger HA, Lamm A (eds) Handbook of fuel cells-
lower diffusivities and correspondingly unattrac-
fundamentals, technology and applications, tive productivities. Membrane separations are
vol 1, Fundamentals and survey of systems. Wiley, based on differences in how fast the penetrants
Chichester, pp 267280 diffuse through the membrane (i.e., diffusion
Gulzow E (2012) Alkaline fuel cells. In: Stolten D, Emonts
selectivity) and/or how much they sorb in the
B (eds) Fuel cell science and engineering: materials,
processes, systems and technology, vol 1. Wiley-VCH, membrane (i.e., sorption selectivity). For exam-
Weinheim, pp 97129 ple, propene molecule is slimmer than propane
56 Alkane Oxidative Dehydrogenation by Membrane Reactor
molecule. Remarkable propene/propane diffusion separations (Faiz and Li 2012). Alkene transport
selectivity is possible for materials with rigid in such membranes is assisted by carriers
structures and properly tailored ultramicropores. (e.g., Ag+ ions) that reversibly react with alkenes
Additionally, propene is slightly polar and a lim- through p-complexation. Amazing alkene perme-
ited (<10) propene/propane sorption selectivity ability and alkene/alkane selectivity can be
may be achieved with polar materials (e.g., cat- achieved in facilitated transport membranes at
ionic zeolites with high Al/Si ratio). the same time. Nevertheless, the carriers may be
Membranes for alkene/alkane separations can irreversibly deactivated by certain impurities
be categorized into two classes: membrane sepa- (e.g., hydrogen, hydrogen sulde, and acetylenes)
rations based on sorption-diffusion and others in the feed, resulting in permanent loss in both
based on facilitated transport. The majority of alkene permeability and alkene/alkane selectivity.
membranes in the rst category achieve the sepa- While in some cases membrane performance may
ration relying on diffusion selectivity. Polymer be partially recovered by periodically recharging
membranes, despite their excellent scalability, the carrier, such approach may not be practical for
are incapable of offering attractive alkene/alkane large-scale applications.
selectivity without signicantly compromising
their permeability due to the permeability-
selectivity trade-off (Burns and Koros 2003). References
Additionally, uncrosslinked polymer membranes
may be plasticized with high-pressure alkene/ Adams R, Johnson JR, Zhang C, Lively RP, Dai Y,
Esekhile O, Liu J, Koros WJ (2013) Preparation of
alkane feeds, resulting in signicant selectivity
mixed matrix membranes. In: Hoek EMV, Tarabara
loss. On the other hand, membranes derived VV (eds) Encyclopedia of membrane science and tech-
from more rigid molecular sieving materials nology. Wiley, New York, pp 398430
(e.g., zeolites, metal-organic frameworks Burns RL, Koros WJ (2003) Dening the challenges for
C3H6/C3H8 separation using polymeric membranes.
[MOFs], zeolitic imidazolate frameworks [ZIFs],
J Membr Sci 211:299309
and carbon molecular sieves [CMS]) can offer Faiz R, Li K (2012) Olen/parafn separation using mem-
simultaneously high permeability and selectivity. brane based facilitated transport/chemical absorption
However, fabrication of molecular sieve mem- techniques. Chem Eng Sci 72:261284
Jarvelin H, Fair JR (1993) Adsorptive separation of propylene-
branes is usually difcult and many technical
propane mixtures. Ind Eng Chem Res 32:22012207
challenges must be solved to deliver such perfor- Zhang C, Zhang K, Xu L, Labreche Y, Kraftschik B, Koros
mance economically on a large scale. Among WJ (2014) Highly scalable ZIF-based mixed-matrix
molecular sieve membranes, CMS membrane is hollow ber membranes for advanced hydrocarbon
separations. AIChE J 60:26252635
the most promising as it can be conveniently
realized in the scalable hollow ber geometry by
pyrolyzing polymer precursor hollow bers.
A compromise to address the limitations of both
polymer membrane and molecular sieve mem- Alkane Oxidative Dehydrogenation
brane is mixed-matrix membrane (Adams by Membrane Reactor
et al. 2013), which are formed by dispersing
molecular sieve particles/platelets in polymer Miguel Menendez
matrices. Selectivity and/or permeability of Department of Chemical and Environmental
mixed-matrix membranes may be substantially Engineering, University of Zaragoza, Zaragoza,
enhanced over the polymer matrices and in the Spain
meantime, the membrane can still be inexpen-
sively processed into hollow bers (Zhang
et al. 2014). Alkanes are converted to olens in the chemical
Facilitated transport membrane is another class industry because of the high demand of polyole-
of membrane studied for alkene/alkane ns, such as polypropylene or polybutadiene, and
Alkane Oxidative Dehydrogenation by Membrane Reactor 57
Alkane Oxidative
Dehydrogenation by
Membrane Reactor,
Fig. 1 Scheme of a A
membrane reactor for the
selective oxidation of
hydrocarbons with a
distributed feed of oxygen
also because olens are the starting materials for distributes oxygen along a packed bed of catalyst.
many other chemicals. These reactions can be This improvement is based on the favorable effect
represented as of low oxygen partial pressure on the selectivity.
In fact, for many catalysts the observed
Cn H2n2 ! Cn H2n H2 reaction order respect to oxygen for the above
reaction is lower than for the formation of CO
A common characteristic of these reactions is the and CO2. This observation implies that low partial
achievable conversion being limited by thermo- pressure of oxygen increases the selectivity to
dynamic equilibrium. Two options to increase the olens.
conversion are This general principle has been applied for the
oxidative dehydrogenation of ethane, propane,
Using a membrane reactor where H2 is and butane, with successful results in all cases
removed from the reactor. This will increase (Coronas and Santamara 1999). Details of the
the achievable conversion according to the results obtained with membrane reactors with pro-
Le Chaterliers principle. Any material which pane and butane (oxidative dehydrogenation and
can remove hydrogen selectively can be used, oxidation to maleic anhydride) are described in
but Pd and Pd alloys are the most employed. specic chapters.
The preparation of this kind of membranes A general scheme of a membrane reactor for
has been widely described (Basile and Gallucci this kind of reactions is shown in Fig. 1.
2011). The oxygen containing feed is usually intro-
Adding an oxidizing agent (e.g., oxygen) to the duced through the shell side. A porous membrane
reaction, in order to form water instead of distributes this oxygen along the catalyst bed,
hydrogen, the resulting reaction would be: keeping a low partial pressure of oxygen in all
the beds. A feed containing the hydrocarbon is fed
Cn H2n2 1=2 O2 ! Cn H2n H2 O to the tube side, i.e., to the catalyst bed. The
reaction products are removed at the exit of the
This approach has the advantage of being ther- catalyst bed.
modynamically favored but has the problem that As an alternative to porous membranes, dense
simultaneous formation of carbon oxides reduces ceramic membranes can provide an additional
the selectivity. advantage: the separation of oxygen-nitrogen
A way to improve the selectivity is to use a mixtures (Hashim et al. 2011). This would
membrane reactor, where the membrane decrease the cost of oxygen separation.
58 Alumina Membranes
References membrane cost, making it about one order of

magnitude higher than its polymeric counterparts.
Basile A, Gallucci F (eds) (2011) Membranes for mem- However, because of much longer lifetimes and
brane reactors. Preparation, optimization and selection.
higher uxes of alumina membranes, the long-
Wiley, Chichester
Coronas J, Santamara J (1999) Catalytic reactors based on term operating cost of alumina membrane plants
porous ceramic membranes. Catal Today 51:377389 is comparable with polymeric membrane plants,
Hashim SS, Mohamed AR, Bhatia S (2011) Oxygen sep- no mention there are many scenarios in which
aration from air using ceramic-based membrane tech-
only alumina/ceramic membranes can be used.
nology for sustainable fuel production and power
generation. Renew Sustain Energy Rev 15:12841293
Membrane Designs and Fabrication
Alumina Membranes Most alumina membranes use asymmetric mem-

brane architectures. Coarse alumina particles are
Bo Wang used for the supporting layer to provide mechan-
Department of Chemical Engineering, Imperial ical strength, and the transport resistance in this
College London, London, UK layer is usually low due to its big pore size; ne
alumina particles are coated on top of the support
to form a thin separating layer to reach desired
Alumina membranes are membranes made of pore sizes and high uxes; and depending on the
high-purity (mostly >97 %) aluminum oxides pore size difference between the top and the
(alumina). They belong to the categories of supporting layer, one or few transition layers
ceramic membranes and in most cases, might be used between the top and supporting
microltration and ultraltration membranes. layers to avoid coating defects and to increase
Pure alumina membranes are predominant in all adhesion between layers.
kinds of commercial ceramic membranes, and Alumina membranes may adopt a tubular
most other commercial ceramic membranes, if shape, plate shape, or monolithic multichannel
using other materials as the separation coating, design. The latest is the most commonly used
are based on alumina membrane support. design in industry because it provides higher
membrane area per unit volume, therefore reduces
the size of the separation plant and operating
Advantages and Disadvantages costs. Some commercial products can achieve an
area to volume ratio of 800 m2/m3 with the mono-
Alumina membranes are chemically inert and lithic design, which is comparable with spiral
mechanically strong, and also thermally stable, wound polymeric membrane modules.
therefore they can be used under very harsh con- Fabrication of alumina membranes usually
ditions such as aggressive chemical environments involves several forming and sintering steps due
and elevated temperatures, making them versatile to their multilayer asymmetric membrane archi-
in various industrial processes. Their strong tectures, and the total production process is often
mechanical properties allow them to be operated longer than 1 week. The supporting layer can be
under high pressures and high feed ow rates to formed by extrusion (for tubular or monolith
achieve high productivity and to tolerate high designs) or by pressing (for plate or disk designs),
solid contents in the feed. followed by partial sintering at a high temperature
The high fabrication cost is a major obstacle for (up to 1600 C) to achieve a high strength and
alumina membranes to expand its market share. porous structures. Transition layers and the top
Although alumina is widely available, the com- layer then can be coated on the support, also
plicated fabrication process and high energy con- followed by sintering after each coating.
sumption for sintering lift signicantly the total Depending on the pore size, two types of alumina
Americium Separation by Membrane Operation 59
could be used for the top layer coating: for Ebrahimi M, Willershausen D, Ashaghi KS, Engel L,
microltration membranes whose pore size is Placido L, Mund P, Bolduan P, Czermak P (2010)
Investigations on the use of different ceramic mem-
larger than 100 nm, a-alumina is used; for ultra- branes for efcient oil-eld produced water treatment. A
ltration membranes whose pore size is smaller Desalination 250(3):991996
than 100 nm, g-alumina or a mixture of a and g
types are often used.
There are also other fabrication techniques that
can produce unique alumina membranes, for
example, the anodic alumina membranes that Aluminum (III) Oxide Porous
have straight cylindrical pores and asymmetric Membranes or Alumina Porous
alumina membranes made by single-step phase Membranes
inversion/sintering methods. These membranes
have shown good potential in ltration but have Al2O3 Porous Membranes
not been widely accepted as a feasible replace-
ment of conventional alumina membranes yet.
Americium Separation by Membrane

Applications Operation
Alumina membranes are widely used in water and P. K. Mohapatra

wastewater treatment, pharmaceuticals, food Radiochemistry Division, Bhabha Atomic
processing, chemical processing, and others. Pro- Research Centre, Mumbai, Maharashtra, India
duction of drinking water and wastewater treat-
ment are the most important applications, which
occupy about half of the market share of alumina Americium (Am) is the rst among the trans-
membranes. The stable pore structure of alumina plutonium elements, produced in nuclear reactors
membranes even under high pressures and abra- in large quantities. Due to long half-lives of its
sive conditions guarantees constant quality of the isotopes (t1/2 of 241Am, 432 y; 243Am, 7370 y),
permeate and makes it preferable in the scenarios this element can have long-term hazards to the
where quality control is crucial, such as in drink- environment if not separated from the radioactive
ing water production and food industry. And the wastes before their disposal. Separation of Am
natural hydrophilic surface of alumina offers bet- mostly involves solvent extraction-based methods
ter anti-fouling property than polymeric mem- involving specially designed ligands for Am
branes; therefore its usage in wastewater uptake. Membrane-based separation has been
treatment is expanding rapidly; and in some used for Am separation using pressure-driven
other heavy-duty ltration processes such as techniques such as nanoltration, facilitated by
treating produced water from oil elds, alumina complexation (Favre-Rguillon et al. 2007).
membranes have been dominant. However, selectivity factors in such separation
systems are not encouraging. Liquid membrane-
based separation systems, with Am-specic car-
rier extractants, are reported to give better
Further Reading
decontaminations.
Benfer S, rki P, Tomandl G (2004) Ceramic membranes Some of the rst studies on Am transport
for ltration applications preparation and characteri- using supported liquid membranes include
zation. Adv Eng Mater 6(7):495500 the work of Danesi, who used CMPO
Lehman SG, Adham S, Liu L (2008) Performance of
new generation ceramic membranes using hybrid coag-
(carbamoylmethylphosphine oxide) as the carrier
ulation pretreatment. J Environ Eng Manage extractant (Danesi et al. 1983). Feed solutions
18(4):25760 containing LiNO3 resulted in highly encouraging
60 Americium Separation by Membrane Operation
Am transport rates which were signicantly TODGA-based hollow-ber supported liquid

affected in the presence of nitric acid in the feed, membrane system has been successfully used in a
attributed to the cotransport of nitric acid. The Am mini pilot-scale processing of synthetic high-level
transport rates could be improved by using a sec- waste (Ansari et al. 2011). A series of studies have
ond membrane in parallel containing Primene indicated that the diluents play a signicant role in
JM-T as the carrier extractant for selective trans- Am transport rates which are mainly due to two
port of nitric acid (Chiarizia and Danesi 1987). In factors: (i) nature of the extracted species and
another study, triphosphine trioxide extractants (ii) diffusivity of the extractable complex which
have been employed for Am(III) transport from is based on the viscosity of the medium (Panja
NaNO3 solutions with encouraging results et al. 2012).
(Cristau et al. 1999). Other extractants such as Am(III) separation of trivalent lanthanides has
tetra-alkyl malonamides and diglycolamides have also been studied using various Am(III) selective
shown comparable or even better transport ef- extractants such as Cyanex 301 (Hoshi et al. 2000),
ciency for Am as compared to CMPO. BTP (Geist et al. 2005), etc. In one of the early
Malonamides such as DMDBTDMA (N,N0 - reports on Am(III)Eu(III) separations using liquid
dimethyl-N,N0 -dibutyltetradecyl malonamide) membranes (Novikov and Myasoedov 1987),
0
and DMDOHEMA (N,N -dimethyl-N, almost no transport of Am(III) was observed
N0 -dioctylhexylethoxymalonamide) are reported while effecting complete quantitative Eu(III) trans-
to be quite efcient for Am(III) transport (Sriram port using a feed composition of 0.01 M DTPA,
et al. 2000; Patil et al. 2013). TODGA (N,N,N0 , 0.5 M citric acid, 0.5 M NaNO3 at pH 2.9, and 1 M
N0 -tetraalkyldiglycolamide) has been found to be HDEHP as the carrier extractant. Cyanex 301 has
one of the best carriers for Am(III) transport due to been used for highly selective Am(III) transport of
signicantly lower acid cotransport as compared to a feed also containing several trivalent lanthanide
CMPO or malonamides (Ansari et al. 2009). ions (Fig. 1; Bhattacharyya et al. 2006). Oxidation
Americium Separation 3000 166

Ho
by Membrane 2500
Operation,
Counts/900 s
2000
Fig. 1 Gamma ray 241
Am Feed (initial)
spectrum of the feed and 1500
152,154
receiver solutions measured 1000 Eu160 140 160
using HPGe detector Yb La Tb
500
(Bhattacharyya et al. 2006)
0
3000 166
Ho
2500
Count/900 s
2000 Feed after 4 hrs

1500 152,
154 140
1000 Eu 160 La 160
Yb Tb
500
0
3000
2500
Counts/900 s
2000 Strip after 4 hrs

1500
1000 241
Am
500
0
0 1000
Energy (keV)
Americium Separation by Membrane Operation 61
to the +4 and +6 states reportedly enhanced Am Bhattacharyya A, Mohapatra PK, Manchanda VK

transport rates due to higher extraction efciencies (2006) Separation of trivalent actinides and lanthanides
using a at sheet supported liquid membrane
of these ions. Tri-n-octylamine containing liquid containing Cyanex-301 as the carrier. Sep Purif A
emulsion membrane (LEM) systems has been Technol 50:278281
used for the transport of Am(IV) which was done Chiarizia R, Danesi PR (1987) A double liquid membrane
using phosphotungstate (Myasoedov et al. 1990). system for the removal of actinides and lanthanides
from acidic nuclear wastes. Sep Sci Technol
In another report (Mikheeva et al. 1994), the trans- 22:641659
port of hexavalent Am by electrochemical oxida- Cristau HJ, Virieux D, Dozol JF, Rouquette H (1999)
tion followed by simultaneous transport using Selective extraction of actinides by polyphosphine
HDEHP as the carrier. polyoxides through supported liquid membrane.
J Radioanal Nucl Chem 241:543547
There are several reports on the Am(III) recov- Danesi PR, Horwitz EP, Rickert PG (1983) Rate and mech-
ery from acidic feeds using liquid emulsion mem- anism of facilitated americium(III) transport through a
branes (LEM) using extractants such as HDEHP, supported liquid membrane containing a bifunctional
PC-88A, and TODGA as well. Usually, the LEM organophosphorus mobile carrier. J Phys Chem
87:47084715
methods are much faster as compared to the Favre-Rguillon A, Sorin A, Pellet-Rostaing S, Bernier G,
SLM-based separation methods which are due to Lemaire M (2007) Nanoltration assisted by complex-
highly dispersed globules containing the internal ation: a promising process for the separation of trivalent
phase as the strippants present in the emulsion long-lived minor actinides from lanthanides. C R Chim
10:9941000
which effect rapid mass transfer. In a recent Geist A, Weigl M, Gompper K (2005) Small-scale actinide
study, multifunctional ligands such as the (III) partitioning processes in miniature hollow bre
calixarenes anchoring diglycolamide functional modules. Radiochim Acta 93:197202
groups have shown improved transport behavior Hoshi H, Tsuyoshi A, Akiba K (2000) Separation of amer-
icium from europium using liquid membrane impreg-
as compared to the simple extractants (Ansari nated with organodithiophosphinic acid. J Radioanal
et al. 2013). The liquid membrane methods with Nucl Chem 243:641644
lower ux as compared to the solvent extraction Mikheeva MN, Novikov AP, Myasoedov BF, Tikhomirov
methods can be used for pre-concentration SV (1994) Kinetics of americium(VI) mass transfer
through solid supported liquid membrane with HDEH-
and separation of Am from dilute solutions P. J Radioanal Nucl Chem Art 185:265271
including environmental samples and radioactive Myasoedov BF, Novikov AP, Milyukova MS, Mikheeva
wastes. MN, Bunina TV (1990) Emulsion membrane extrac-
tion of Am(III) and Am(IV) in systems with tertiary and
secondary amines. J Radioanal Nucl Chem
142:481487
References Novikov AP, Myasoedov BF (1987) Separation of trivalent
europium, terbium, americium and curium by a
Ansari SA, Mohapatra PK, Raut DR, Kumar M, supported liquid membrane impregnated with
Rajeswari B, Manchanda VK (2009) Performance of di-(2-ethylhexyl)phosphoric acid. Solvent Extr Ion
some extractants used for actinide partitioning in a Exch 5:117127
comparative hollow bre supported liquid membrane Panja S, Mohapatra PK, Tripathi SC, Gandhi PM, Janardan
transport study using simulated high level nuclear P (2012) Role of organic diluents on Am(III)
waste. J Membr Sci 337:304309 pertraction across N, N, N0 N0 -tetraoctyl-3-
Ansari SA, Gujar RB, Mohapatra PK, Kandwal P, oxapentanediamide (TODGA) supported liquid mem-
Sengupta A, Thulasidas SK, Manchanda VK branes. J Membr Sci 403404:7177
(2011) A cold actinide partitioning run at 20 L scale Patil AB, Shinde V, Pathak PN, Mohapatra PK (2013)
with hollow ber supported liquid membranes Evaluation of DMDOHEMA based supported
containing diglycolamide extractants. Radiochim Acta liquid membrane system for high level waste remedia-
99:815821 tion under simulated conditions. J Membr Sci
Ansari SA, Mohapatra PK, Iqbal M, Kandwal P, 442:4856
Huskens J, Verboom W (2013) Novel diglycolamide- Sriram S, Mohapatra PK, Pandey AK, Manchanda VK,
functionalized calix[4]arenes for actinide extraction Badheka LP (2000) Facilitated transport of americium
and supported liquid membrane studies: part II. Role (III) across a PTFE membrane containing dimethyl
of substituents in the pendent arms and mass transfer dibutyl tetradecyl-1,3-malonamide as carrier. J Membr
modeling. J Membr Sci 430:304311 Sci 177:163175
62 Ammonia Decomposition in Catalytic Membrane Reactors
SRM reaction : CH4 2H2 O $ CO2 4H2

Ammonia Decomposition in Catalytic DH 164:5 kJ=mol
Membrane Reactors MTH cycle : C7 H14 $ C7 H8 3H2
DH 204:6 kJ=mol
Toshinori Tsuru NH3 system : NH3 ! 0:5N2 1:5H2
Department of Chemical Engineering, Hiroshima DH 46 kJ=mol
University, Higashi-Hiroshima, Japan
Membrane reactors can be categorized based on
conguration of catalysts as packed type membrane
reactor, uidized membrane reactor, and catalytic
Recently, NH3 has been proposed as a H2 car- membrane reactor. As shown in Fig. 1, packed bed
rier in a promising solution that would bypass membrane reactors consist of catalysts for reaction as
the intractable issues related to H2 storage and well as membranes for separation. Both catalyst
delivery, since NH3 has a high hydrogen stor- particles and membranes are packed into one unit,
age capacity of 17.6 wt.% and can be easily but reaction and separation occur separately. On the
stored and transported in liquid form under other hand, catalytic membranes have catalytic activ-
mild conditions. NH3 decomposes to H2, ity as well as separation ability in one membrane.
which produces only H2O after combustion, Zeolite membranes such as ZSM-5 and TS-1 are
and N2, which is not a greenhouse gas; in inherently catalytically active, while another type
other words, NH3 is a COx-free H2 carrier. of catalytic membrane is a silica hydrogen separation
Another class of a hydrogen carrier is an membrane on catalytically active support which can
organic hydride such as cyclohexane (CH) and be prepared using catalytically active powders such
methylcyclohexane (MCH). In dehydrogenation as nickel oxides or impregnation of Ni-catalyst inside
reaction, CH and MCH reversibly produce ben- porous a-alumina supports (Tsuru et al. 2004).
zene (BEN) and toluene (TOL), based on the Membrane reactors combining a catalytic reac-
equilibrium reaction, respectively, while BEN tion with the separation process into a single unit
and TOL are re-hydrogenated into CH and offer an ideal device to produce H2 from NH3
MCH, respectively. Methylcyclohexane- efciently, where enhanced NH3 conversion and
toluene-hydrogen (MTH cycle) and high-purity COx-free H2 can be simultaneously
cyclohexane-benzene-hydrogen (CBH cycle) obtained by applying a hydrogen separation mem-
need reversible and highly selective conversion brane. Ru is reported the most active catalyst for
to sustainable use. Therefore, the advantages of ammonia decomposition and is supported on
NH3 mentioned above make the NH3-mediated porous carbon such as carbon nanotubes (CNT)
hydrogen economy extremely attractive and with supporting catalysts such as potassium (K).
promising. Other H2 carriers, such as methanol, To date, only a few studies have reported mem-
dimethyl ether, etc., inevitably cause end-user brane reactors for NH3 decomposition in a packed
CO2 emissions, since they are carbonaceous bed Pd membrane reactor, and improved NH3
compounds and on-board CO2 capture is not conversion was obtained after H2 extraction.
feasible. Another class of membrane reactor is a cata-
As shown in the following equations, lytic membrane reactor. Figure 2 shows a sche-
another advantage of NH3-energy systems is matic bimodal catalytic membrane reactor
the low heat of reaction. Approximately 30 kJ (BCMR) (Li et al. 2013). This new type of mem-
is required for one mole of produced H2 in NH3 brane reactor had a conguration that consisted of
system, which is lower than in MTH cycle a silica separation layer and a bimodal catalytic
(70 kJ) and steam reforming of methane support (Ru/g-Al2O3/a-Al2O3). The impregnation
(SRM) (40 kJ). of g-Al2O3 into an a-Al2O3 membrane support to
Ammonia Decomposition in Catalytic Membrane Reactors 63
Reactants retentate
Reactants retentate
A
packed catalyst catalytic
permeate membrane permeate
membrane
Packed Bed Membrane Reactor (PBMR) Catalytic Membrane Reactor (CMR)
Ammonia Decomposition in Catalytic Membrane Reactors, Fig. 1 Conguration of membrane reactors
COx-free H2
Silica membrane
Intermediate layer
Ru catalyst
-Al2O3
Bimodal catalytic support layer
-Al2O3

NH3 1.5H2 0.5N2
NH3 N2
Ammonia Decomposition in Catalytic Membrane Reactors, Fig. 2 Schematic diagram of the proposed bimodal
catalytic membrane reactor for COx-free H2 production via NH3 decomposition (Li et al. 2013)
Ammonia
Decomposition
in Catalytic Membrane
Reactors, Fig. 3 Bimodal
catalytic membrane reactor
performance for NH3
decomposition at 450 C
(Li et al. 2013)
form a bimodal support is expected to effectively improved catalytic activities by mesopores and
improve catalytic membrane reactor performance, high gas diffusivity by macropores.
since the bimodal structure contains both Figure 3 shows an example of enhancement of
mesopores and macropores, which results in conversion for the BCMR performance for NH3
64 Amorphous Fluoropolymers
decomposition at 450 C. During the initial of the polymeric sequences and increase the
250 min without sweep gas, a NH3 conversion intermolecular distances. The complete suppres-
was 45 %. When N2 sweep gas was applied to sion of the crystallinity is, instead, more easily
the permeate stream of the membrane reactor, achieved by introduction of elevated percentage
NH3 conversion was signicantly enhanced from of cyclic monomeric units; by following this strat-
45 to 95 % after the selective H2 extraction. NH3 egy, new families of amorphous glassy peruor-
decomposition rate is strongly inhibited by the opolymers (Teon AF, Cytop , and Hyon
presence of H2, particularly at a high pressure. AD whose structures are reported in Fig. 1) were
Therefore, the highly enhanced NH3 conversion developed and commercialized starting from the
can be ascribed to the improved kinetics from H2 late of 1980s (Merkel et al. 2006).
extraction in the BCMR. The cyclic groups, present in each of these
three structures, hinder efcient chain packing,
thus preventing crystal formation and yielding
References completely amorphous polymers with a glass
transition (Tg)tunable but always sensibly above
Li G, Kanezashi M, Tsuru T (2013) Highly enhanced room temperature (from 95 C of Hyon AD40
ammonia decomposition in bimodal catalytic silica
to 240 C of Teon AF2400).
membrane reactors for COx-free hydrogen production.
AIChE J 59:168179 Teon AF, Cytop , and Hyon AD com-
Tsuru T, Yamaguchi K, Yoshioka T, Asaeda M (2004) bine the properties typical of peruorinated semi-
Methane steam reforming by microporous catalytic crystalline materials (high thermal and chemical
membrane reactors. AIChE J 50:2794
resistance, low dielectric constant, low surface
energy) with new peculiar properties imparted
by the amorphous structure such as the high sol-
ubility in uorinated solvents and excellent trans-
Amorphous Fluoropolymers parency across a broad wavelength range (from
UV to near IR). Furthermore, the high chain stiff-
Marco Avataneo ness imparted by the cyclic units leads to a dif-
RD&T, Solvay Specialty Polymers, Bollate (MI), cult packing of the chain segments with the
Italy consequent formation of nano-voids which are
responsible for the high gas diffusion of these
polymers. Gas separation membranes, pellicles
Synonyms for microlithography, protective coatings, dielec-
tric layers for organic and electronic and plastic
Fluorinated polymers; Fluoroplastics; optical bers are the most important applications
Noncrystalline of these glassy peruoropolymers.
Other two important families of amorphous
Amorphous uoropolymers are noncrystalline uoropolymers, in this case displaying a Tg
synthetic polymers where the hydrogen atoms of value below room temperature, are peruoropo-
the backbone have been partially or completely lyethers (PFPEs) and uorinated elastomers. The
replaced by uorine atoms. While common former (Fig. 2), obtained by means of different
uoroplastics are semicrystalline, the number of manufacturing technologies (Rizvi Syed 2009;
commercial amorphous peruoropolymers is rel- Marchionni et al. 1996), are low molecular weight
atively limited (Scheirs 2001). polymers containing etheric oxygen atoms in the
In uorinated polymers the crystallinity, char- backbone which make the chains extremely ex-
acteristic of peruorocarbon CF2CF2- units, can ible so preventing the formation of crystallite
be reduced by incorporation of monomers with phases.
pendant side groups, like hexauoropropene or These synthetic uids display extremely low
peruorovinyl ethers, which disrupt the regularity Tg, even below 100 C. They nd applications
Amorphous Fluoropolymers 65
Amorphous a b c
Fluoropolymers, OCF3
Fig. 1 Structure of three CF CF CF2 CF2 CF2 CF CF CF2
commercial amorphous CF C CF2 CF2 m A
glassy peruropolymers m O O O CF2
O O
C CF2
CF2 n
n CF3 CF3
n
Hyflon AD Teflon AF Cytop
Amorphous CF3CF2CF2O- (CFCF2O)nCF2CF3 Krytox, Du Pont

Fluoropolymers,
Fig. 2 Structure of the four CF3
most important commercial CF3CF2CF2O- (CF2CF2CF2O)nCF2CF3 Demnum, Daikin
peruoropolyethers
CF3O- (CF2CFO)m(CFO)n(CF2O)p-CF3 Fomblin Y, Solvay
CF3 CF3
CF3O- (CF2O)m(CF2CF2O)n-CF3 Fomblin Z, Solvay
as lubricants for extreme conditions (high temper- Cross-References

atures, high pressure, presence of aggressive
chemicals), as testing and heat transfer uids in Gas Separation
electronic, and as inert solvents. PFPEs Glass Transition Temperature (Tg)
containing functional-end groups can be used as Hyon AD
building blocks for the synthesis of two-phase
materials (partially uorinated, partially hydroge-
References
nated), surface treatment agents, and surfactants.
Few applications in the eld of membranes are Chaouk H, Wilkie JS, Meijs GF, Cheng HY (2001) New
known for PFPEs (Chaouk et al. 2001). porous peruoropolyether membranes. J Appl Polym
Fluorinated elastomers (Moore 2006) are Sci 80:17561763
cross-linked amorphous polymers with Tg Marchionni G, Ajroldi G, Pezzin G (1996)
Structureproperty relationships in peruoropo-
below room temperature. They are generally lyethers: a family of polymeric oils. In: Allen SG,
obtained by copolymerization of at least two of Aggarwal SL, Russo S (eds) Comprehensive polymer
the following monomers: tetrauoroethylene, science, 2nd suppl. Pergamon Press, Oxford, UK,
vinylidene uoride, hexauoropropene, and pp 347388
Merkel TC, Pinnau I, Prabhakar R, Freeman BD
peruorovinyl ethers. Fluorosilicone elastomers, (2006) Gas and vapour properties of peruoro-
i.e., uorochemically modied polysiloxanes, are polymers. In: Yampolskii Y, Pinnau I, Freeman BD
another important class of uorinated elastomers. (eds) Materials science of membranes for gas and
Fluoroelastomers show extreme resistance to vapour separation. Wiley, New York, pp 251270
Moore AL (2006) Fluoroelastomers Handbook, Schroeder
high temperatures and to a variety of aggressive H. In: Rubber Technology (Morton, M., Ed.), 3rd ed.
uids. They exhibit excellent resistance to oils, William Andrew Publishing, Norwich: NY, p. 105
solvents, acids, alkalis, and ozone over a wide Rizvi Syed QA (2009) A Comprehensive Review of Lubri-
temperature range (30 C to 250 C) and, there- cant Chemistry, Technology, Selection and Design,
ASTM International, West Conshohocken, PA, pp
fore, they are suitable for aggressive environ- 100-211
ments and severe conditions. No relevant Scheirs J (2001) Fluoropolymers technology, markets and
industrial applications in the eld of membranes trends a Rapra industry analysis report, Rapra Tech-
are known. nology Limited, Shawbury
66 Amorphous Perfluoropolymers
Cross-References
Amorphous Perfluoropolymers
Glass Transition Temperature (Tg)
Marco Avataneo Hyon AD
RD&T, Solvay Specialty Polymers, Bollate (MI),
Italy
References
Marchionni G, Ajroldi G, Pezzin G (1996)

Synonyms Structureproperty relationships in peruoropolyethers:
a family of polymeric oils. In: Allen SG, Aggarwal SL,
Fully uorinated polymers; Noncrystalline Russo S (eds) Comprehensive polymer science, 2nd
suppl. Pergamon Press, Oxford, pp 347388
Marshall JB (1997) In: Scheirs J (ed) Modern
Amorphous peruoropolymers are noncrystalline uoroplastics. Wiley, Chichester
fully uorinated synthetic polymers composed Merkel TC, Pinnau I, Prabhakar R, Freeman BD
exclusively by carbon, uorine, and oxygen (2006) Gas and vapour properties of peruoro-
polymers. In: Yampolskii Y, Pinnau I, Freeman BD
atoms. The most relevant product family for mem- (eds) Materials science of membranes for gas and
brane applications is represented by the glassy vapour separation. Wiley, New York, pp 251270
amorphous peruoropolymers containing dioxole Moore AL (2006) Fluoroelastomers handbook the den-
groups (Teon AF, Hyon AD, and Cytop) itive users guide and databook. William Andrew Pub-
lishing/Plastics Design Library, Norwich, NY
which display a glass transition above room tem- Sanguineti A, Guarda PA, Marchionni G, Ajrolidi G (1995)
perature (Merckel et al. 2006). Peruoropolyethers Solution properties of peruoropolyether polymers.
and peruorinated elastomers are, instead, the two Polymers 36:36973703
most important families of amorphous peruoro-
polymers with a glass transition below room tem-
peratures. Peruoropolyethers (Marchionni
et al. 2006; Sanguineti et al. 1995) are a class of
synthetic lubricants characterized by extremely low Amorphous Polymers
Tg, also below 100 C. They nd applications as and the Amorphous Region
lubricants for extreme conditions (high tempera-
tures, high pressure, presence of aggressive Petr Sysel
chemicals), as heat transfer uids and as inert sol- Department of Polymers, Institute of Chemical
vent (they are good solvent for the glassy amor- Technology, Prague, Czech Republic
phous peruoropolymers). The commercially
available PFPEs include Krytox, Fomblin, and
Demnum. Peruoroelastomers (Moore 2006), Polymers consist of the large molecules (polymer
also referred to as FFKM elastomers, are amor- chains). An arrangement of these chains in
phous cross-linked polymers obtained by copoly- the bulk polymer is given by their chemical com-
merization of tetrauoroethylene with position, spatial orientation (e.g., an isotactic vs
hexauoropropene and/or peruorovinyl ethers atactic polypropylene), molecular weight, and
which effectively disrupts the crystallinity imparted processing conditions (e.g., a cooling rate).
by the TFE segments (Marshall 1997). Peruoroe- Based on these factors, bulk polymers show an
lastomers are characterized by improved thermo- amorphous or partially crystalline
oxidative stability and oil resistance (compared (semicrystalline) phase state. In the amorphous
with VDF-based elastomers) for extreme industrial state, an ordering of the (entangled) polymer
and military applications. Unlike the glassy amor- chains is random. The amorphous polymer does
phous peruoropolymers, no relevant industrial not include a long-range order, but it is character-
membrane applications for peruoropolyethers ized by the existence of some regularity on a
and for peruoroelastomers are known. short-range order. As a consequence of it, an
Amorphous Polymers and the Amorphous Region 67
deformation
A
glassy state rubbery state plastic state
glass transition flow temperature

temperature
temperature
Amorphous Polymers and the Amorphous Region, Fig. 1 Thermomechanical curve of an amorphous polymer
X-ray diffraction provides a diffuse ring only An extent of this (segmental) chain motion and
(a set of discrete rings is typical for well-ordered free volume inuences a lot of the polymer prop-
structures). If the amorphous polymer is exposed erties, e.g., a transport of the different
to a gradually increased temperature, it passes low-molecular-weight media through the mem-
from a glass (hard and rigid) to a rubber (soft branes made of these polymers (Strathmann
and exible) and nally to a molten consistency. 2011). For example, a nonporous (dense)
Boundaries between the glass/rubber and rubber/ membrane made of a rubbery polysiloxane [(Si
molten consistency are given by a glass transition (R)2-O-)n, where R is often CH3] shows a much
temperature (Tg) and ow temperature, respec- higher gas permeation (and also diffusivity) in
tively (see a thermomechanical curve of an amor- comparison with those made of common glassy
phous polymer, Fig. 1). polymers at room temperature (Pandey 2001;
The glass transition temperature of the polymer Bernardo et al. 2009). But their selectivity is tra-
is also an important indicator of its application ditionally low (Robeson 2008).
temperature region (e.g., polydimethylsiloxane Some polymers (e.g., polyethylene, polyam-
(Tg = ca 120 C) is rubbery and polystyrene ides) include both amorphous and crystalline frac-
(Tg = ca +100 C) is a glassy polymer at room tions, and they are called semicrystalline ones.
temperature). A character of the main backbone Chains of such polymers may be organized in
(exible/rigid), its substitution (small vs bulk sub- spherulites having a well-ordered lamellar struc-
stituents), and a level of molecular (chain) inter- ture (i.e., a folded chain) with an incorporated
actions (e.g., van der Waals forces vs hydrogen disordered amorphous fraction (Fig. 3).
bonds) are important factors inuencing the posi- Amorphous polymers are usually transparent;
tion (value) of this transition. A exibility of the semicrystalline ones are opaque. A segmental
chains (or segments) of an amorphous polymer is mobility of the amorphous regions in semicrystal-
restricted at temperatures below its Tg. Under line polymers varies from that of the amorphous
heating, molecular interactions (secondary forces) polymer. These regions are affected by the pres-
are disrupted, and a free volume among more ence of crystalline phase, and this phenomenon
exible chains increases (Fig. 2). inuences also transport characteristics of the
68 Amphipathic Molecules
Amorphous Polymers
and the Amorphous
Region,
Fig. 2 Temperature specific
dependence of the specic volume
volume of an amorphous
polymer
glass transition
temperature
temperature
Amorphous Polymers
and the Amorphous
Region, Fig. 3 Schematic
amorphous region
arrangement of the
spherulite
lamellar structure
penetrating media. Traditionally, the more Colomines G, Ducruet V, Courgneau C, Guinault A,

dense crystalline phase decreases a chain mobil- Domenek S (2010) Barrier properties of poly(lactic
acid) and its morphological changes induced by
ity and prolongs a transport trajectory of penetrat- aroma compound sorption. Polym Int 59:818826
ing media, and as a consequence of this their Pandey P, Chauhan RS (2001) Membranes for gas separa-
permeability decreases (Pandey 2001). Neverthe- tion. Prog Polym Sci 26:853893
less, some exceptions were also found, e.g., an Robeson LM (2008) The upper bound revisited. J Membr
Sci 320:390400
increase of gas permeability in a polylactide mem- Strathmann H (2011) Introduction to membrane science
brane with its increasing crystallinity (Colomines and technology. Wiley-VCH, Weinheim
et al. 2010).
References
Amphipathic Molecules
Bernardo P, Drioli E, Golemme G (2009) Membrane gas
separation. Ind Eng Chem Res 48:46384663 Amphiphilic Molecules
Amphiphilic Membrane 69
Amphiphilic Membrane from

Amphiphiles Amphiphilic Molecules
A
Amphiphilic Molecules Amphiphilic membrane is a kind of planar arrays
of one or two monomolecular layers, which are
derived from self-assembly of amphiphilic mole-
cules, such as lipids, surfactants, or block copoly-
Amphiphilic Block Copolymers mer amphiphiles, through noncovalent interactions
in aqueous solution or in oilwater mixtures
Amphiphilic Molecules (Lipowsky et al. 1992; Lipowsky 1991). Amphi-
philic molecules, such as lipids or surfactants, pos-
sess a hydrophilic (lipophobic) head and a
hydrophobic (lipophilic) tail consisted of one or
Amphiphilic Conetwork Membrane two hydrocarbon chains. The hydrophilic head pre-
fers to be in contact with water in aqueous solution,
Amphiphilic Membrane
whereas the hydrophobic tail tends to minimize
contact with water. In aqueous solution or in
oilwater mixtures, such molecules can form a
Amphiphilic Film variety of structures (assemblies) due to the hydro-
phobic effect. In many cases, the molecules form
Amphiphilic Membrane amphiphilic membranes consisted of a monomo-
lecular monolayer of amphiphiles. For example,
the phospholipid molecules self-assemble verti-
cally as a monolayer lm at the interface between
Amphiphilic Lamellar Structure water and air, with their hydrophilic head groups
immersed in the water and their hydrophobic tails
Amphiphilic Membrane pointed to the air, as shown in Fig. 1a (Costa
et al. 2012). The two-dimensional membrane also
can be observed in the lamellar aggregates of sur-
factant molecules at the oilwater interface in a
Amphiphilic Membrane microemulsion. These two-dimensional monomo-
lecular insoluble membranes formed by spreading
Jianhua Zhang amphiphilic molecules on the surface of a liquid are
Department of Polymer Science and Engineering, also referred to as Langmuir monolayers (Kaganer
School of Chemical Engineering and Technology, et al. 1999; Duchs and Schmid 2001). These
Tianjin University, Tianjin, China one-molecule-thick lms display many advantages
Tianjin Key Laboratory of Membrane Science and compared to the other model membranes, which
Desalination Technology, Tianjin University, are excellent model systems for studying chemical
Tianjin, China and biological reactions in two dimensions,
dynamic adsorptions and interfacial behaviors, as
well as membrane hydrolysis processes
Synonyms (Dynarowicz-atka et al. 2001). These monolayers
also provide a way to manipulate molecules and
Amphiphilic conetwork membrane; Amphiphilic construct articial materials for optical, electronic,
lm; Amphiphilic lamellar structure; Amphiphilic or sensor applications (Kaganer et al. 1999).
monolayer/bilayer; Hydrophilic-hydrophobic Amphiphilic molecules dissolved in water at
membrane; Lipid monolayer/bilayer; Two- appropriate concentrations can also self-assemble
dimensional amphiphilic aggregates into bilayer structures with a thickness of only a
70 Amphiphilic Membrane
a Curved bilayer
Monolayer membrane Planar bilayer membrane Solid-supported membrane membrane
air
water
asymmetric symmetric asymmetric symmetric Liposome
b Block copolymeric membrane c Polymeric vesicles(Polymersomes)
Hydrophilic brush
Hydrophobic
membrane
100nm
Polymeric Bilayer Interdigitated layer
Vesicle Wall
Bilayer membrane
Amphiphilic Membrane, Fig. 1 Typical amphiphilic schematic illustrations were made according to the refer-
membranes based on the self-assembly of amphiphiles. ences (Gutsmann and Seydel 2010; Battaglia and Ryan
(a) Phospholipid molecules. (b) Amphiphilic diblock 2005; Du and OReilly 2009 with permission from
copolymers. (c) Amphiphilic diblock copolymers. The Elsevier and The Royal Society of Chemistry)
few nanometers, as shown in Fig. 1a. Such bilayer have universal application in biology and nanotech-
membranes consist of two opposing layers of nology, such as ion channel activity, lipid phase
amphiphiles. The hydrophilic heads are oriented separation, charged lipids manipulation, cell signal-
towards the surrounding water, while the hydropho- ing, cell-cell interaction, membrane base biosensor,
bic tails overlapped and packed the middle of the and protein binding (Han et al. 2013).
membrane (Lipowsky et al. 1992; Lipowsky 1991). Amphiphilic membranes prepared from the
The bilayer membranes are usually found either in self-assembly of lipids have a certain amount of
the form of closed vesicles or as part of a multilayer uidity, depending crucially on the temperature as
stack of lamellae (Chowdhury and Stauffer 1991; well as chain structures and physicochemical
Leibler and Andelman 1987). Particularly, lipid properties of amphiphiles. Therefore, their perme-
bilayers are one of most important amphiphilic ability, stability, and mechanical characteristics
membranes due to their properties and their impor- are not particularly well controlled (Illya
tance in biological systems. The lipid bilayers are a et al. 2005; Discher et al. 1999; Discher and
general structure of cell membrane of almost all Eisenberg 2002). Like natural phospholipids,
living organisms, forming the barrier between cyto- amphiphilic block copolymers in water can also
sol and extracellular environment. Many biological self-assemble into various ordered lamellar struc-
functions and cellular processes for cell-surface tures. The morphology of the resultant
recognition, signaling, and transport are performed nanostructures was controlled by many factors.
at such biological interfaces (Shillcock and The volume ratio of the hydrophilic to hydropho-
Lipowsky 2002; Han et al. 2013). The importance bic block of membrane-forming amphiphiles was
of the lipid bilayers have aroused great interest in proposed to be one of the most important param-
fabricating articial bilayer membranes, including eters in the self-assembly process. Bilayer mem-
both freestanding lipid membranes and solid- branes, as shown in Fig. 1b, can be formed if the
supported membranes. These bilayer membranes hydrophilic/hydrophobic length ratio is suf-
are popular models of cell membranes, which ciently small (i.e., the critical aggregation
Amphiphilic Membrane 71
Amphiphilic Membrane, Fig. 2 Amphiphilic mem- response to hydrophobic and hydrophilic solvents. The
branes based on amphiphilic polymer conetworks. (a) schematic illustrations were made according to the refer-
Typical structures of segmented amphiphilic polymer ences (Patrickios and Georgiou 2003; Erdodi and Kennedy
conetworks. (b) A typical synthesis method of amphiphilic 2006 with permission from Elsevier)
polymer conetworks. (c) Amphiphilic membranes
concentrations approach zero) (Battaglia and make amphiphilic conetworks exhibit amphi-
Ryan 2005; Du and OReilly 2009). Moreover, philic characters (Isayeva and Kennedy 2004;
the lamellar structures of amphiphilic diblock Bruns et al. 2005; Erdodi and Kennedy 2006).
copolymers are often thermodynamically unstable, They commonly can be synthesized by a
which tend to close to form a hollow structure, i.e., multistep condensation (Fig. 2b) or a free-radical
polymer vesicles or polymersomes (Fig. 1c). Com- copolymerization with simultaneous cross-
pared to lipid-based vesicles, the stiffness and sta- linking (Erdodi and Kennedy 2006; Patrickios
bility of polymersomes are higher because of the and Georgiou 2003). As shown in Fig. 2c, such
length and conformational freedom of polymer networks can swell in both aqueous and organic
chains, which made polymersomes have potential media without losing their dimensional stability
for more applications in materials science, biology, and without encountering macroscopic phase sep-
and biomedical science (Battaglia and Ryan 2005; aration or polymer leaching, which have a great
Du and OReilly 2009). potential for applications such as contact lenses,
pervaporation membranes, and drug delivery sys-
tems (Erdodi and Kennedy 2006; Patrickios and
Amphiphilic Membrane Based Georgiou 2003; Isayeva and Kennedy 2004).
on Amphiphilic Conetworks
The term amphiphilic membrane also refers to a Cross-References

polymeric membrane based on amphiphilic
conetworks (Fig. 2a), which consist of hydro- Amphiphilic Molecules
philic and hydrophobic chain segments covalently Cell Membrane
bonded to each other. The cocontinuous Langmuir and LangmuirBlodgett (LB) Films
interconnected hydrophilic/hydrophobic phases Plasma Membrane
72 Amphiphilic Molecules
References Patrickios CS, Georgiou TK (2003) Covalent amphiphilic

polymer networks. Curr Opin Colloid Interface Sci
Battaglia G, Ryan AJ (2005) Bilayers and interdigitation in 8:7685
block copolymer vesicles. J Am Chem Soc Shillcock JC, Lipowsky R (2002) Equilibrium structure
127:87578764 and lateral stress distribution of amphiphilic bilayers
Bruns N, Scherble J, Hartmann L, Thomann R, Ivn B, from dissipative particle dynamics simulations. J Chem
Mulhaupt R, Tiller JC (2005) Nanophase separated Phys 117:50485061
amphiphilic conetwork coatings and membranes. Mac-
romolecules 38:24312438
Chowdhury D, Stauffer D (1991) Lifetime of unstable
bilayers: a lattice model for amphiphilic membranes.
Phys Rev A 44:22472250 Amphiphilic Molecules
Costa AP, Xu X, Burgess DJ (2012) Langmuir balance
investigation of superoxide dismutase interactions with Jianhua Zhang
mixed-lipid monolayers. Langmuir 28:1005010056 Department of Polymer Science and Engineering,
Discher DE, Eisenberg A (2002) Polymer vesicles. Science
297:967973 School of Chemical Engineering and Technology,
Discher BM, Won YY, Ege DS, Lee JCM, Bates FS, Tianjin University, Tianjin, China
Discher DE, Hammer DA (1999) Polymersomes: Tianjin Key Laboratory of Membrane Science and
tough vesicles made from diblock copolymers. Science Desalination Technology, Tianjin University,
284:11431146
Du J, OReilly RK (2009) Advances and challenges in Tianjin, China
smart and functional polymer vesicles. Soft Matter
5:35443561
Duchs D, Schmid F (2001) Phase behaviour of amphiphilic Synonyms
monolayers: theory and simulation. J Phys Condens
Matter 13:48534862
Dynarowicz-atka P, Dhanabalan A, Oliveira ON Jr Amphipathic molecules; Amphiphiles; Amphi-
(2001) Modern physicochemical research on Langmuir philic block copolymers; Amphiphilic polymers;
monolayers. Adv Colloid Interface Sci 91:221293 Amphipols; Detergents; Emulsiers;
Erdodi G, Kennedy JP (2006) Amphiphilic conetworks:
denition, synthesis, applications. Prog Polym Sci Hydrotropes; Lipids/phospholipids; Soaps;
31:118 Surfactants
Gutsmann T, Seydel U (2010) Impact of the glycostructure
of amphiphilic membrane components on the function Amphiphilic molecules is a general term that
of the outer membrane of Gram-negative bacteria as a
matrix for incorporated channels and a target for anti- describes any compound that contains two distinct
microbial peptides or proteins. Eur J Cell Biol covalently bonded components with different
89:1123 afnity for the solvent in the same molecule, in
Han X, Qi G, Xu X, Wang L (2013) Lipid bilayer mem- which one part possesses a high afnity for polar
brane arrays: fabrication and applications. Adv
Biochem Eng Biotechnol 131:121152 solvents (such as water), and another part has a
Illya G, Lipowsky R, Shillcock JC (2005) Effect of chain strong afnity for nonpolar solvents, such as
length and asymmetry on material properties of bilayer hydrocarbons, ethers, and esters. Surfactants,
membranes. J Chem Phys 122:244901244906 polymer amphiphiles, and some lipid molecules,
Isayeva IS, Kennedy JP (2004) Amphiphilic membranes
crosslinked and reinforced by POSS. J Polym Sci containing both hydrophilic and hydrophobic
Polym Chem 42:43374352 components, are typical examples of amphiphilic
Kaganer VM, Mhwald H, Dutta P (1999) Structure and molecules.
phase transitions in Langmuir monolayers. Rev Mod
Phys 71:799819
Leibler S, Andelman D (1987) Ordered and curved meso-
structures in membranes and amphiphilic lms. J Phys Low Molecular Weight Amphiphiles
48:20132018
Lipowsky R (1991) The conformation of membranes. Surfactants (surface active agents), more popu-
Nature 349:475481
larly known as soaps, are among the oldest
Lipowsky R, Richter D, Kremer K (1992) The structure
and conformation of amphiphilic membranes: over- known chemicals used by human. This type of
view. Springer Proc Phys 66:16 molecules now play a crucial role in many areas
Amphiphilic Molecules 73
Amphiphilic Molecules, Fig. 1 Schematic illustrations segment. When p < 1/3, spheres are formed; when 1/3 <
of surfactants and corresponding assemblies. (a) Structures p < 1/2, cylinders will be favored; when 1/2 < p <
of surfactants, (b) Surfactant assemblies and the effect of 1, exible lamellae or vesicles are produced; nally,
the packing parameter (p). p = n/(a0lc), determining the when p = 1, planar lamellae are obtained. If p > 1, reversed
assemblies morphologies, where n is the volume of the structures can be observed (The schematic illustrations
hydrophobic segment, a0 is the contact area of the polar were made according to the Zhang et al. (2012) with
head group, and lc is the length of the hydrophobic permission from Elsevier)
such as cleaning products, food, cosmetics, phar- attracted much attention because of greater sur-
maceuticals, and oileld chemicals as well as face activity (Chevalier 2002; Fuhrhop and Wang
biological processes (Holmberg et al. 2002; 2004). Owing to their special structures and
Schramm et al. 2003; Chevalier 2002). Generally amphiphilic properties, surfactants can reduce
speaking, surfactants possess a hydrophilic head the surface tension when adsorbing at airwater
which might be either charged or uncharged and a or oilwater interface at very low concentration,
hydrophobic tail composed of one or more hydro- which can offer a number of surface chemistry
carbon, uorocarbon, or dimethylsiloxane chains functions, such as solubilizing, wetting, emulsify-
(Fig. 1a). According to the nature of their polar ing, foaming, and surface conditioning
head group, these molecules can be typically clas- (Holmberg et al. 2002; Schramm et al. 2003).
sied into four categories (Fig. 1a): (1) cationic More importantly, these amphiphilic molecules
surfactants containing a positively charged head can spontaneously segregate and self-assemble
group (e.g., fatty amine salts and quaternary into a wide variety of nanostructures through
ammoniums); (2) anionic surfactants with a neg- solvophobic interactions, when the concentration
atively charged head group such as carboxylate, of amphiphiles is above a certain critical concen-
sulfate, sulfonate, or phosphate; (3) nonionic sur- tration (i.e., the critical micelle concentration,
factants with an oligo(ethylene glycol) chain as CMC, dened as the concentration of surfactants
hydrophile; and (4) Zwitterionic surfactants com- above which micelles are spontaneously formed)
prised of a long hydrocarbon chain and a (Holmberg et al. 2002). The morphologies,
hydrophile containing both positive and negative including spherical micelles, cylindrical micelles,
charges. Some new surfactants such as lamellae, vesicles, etc. (Fig. 1b), are governed by
bolaamphiphiles and Gemini surfactants have the molecular structure of the surfactants as well
74 Amphiphilic Molecules
Amphiphilic Molecules, Fig. 2 Schematic illustrations schematic illustrations were made according to the Wang and
of typical chain structures (a) and morphologies of Grayson (2012), Mai and Eisenberg (2012) with permission
amphiphilic polymers in solution (b) and in bulk (c) (The from Elsevier and The Royal Society of Chemistry)
as by physical parameters such as concentration, (Holmberg et al. 2002; Nagarajan 2011). Due to
temperature, and ionic strength. The size and the coexistence of both hydrophobic and hydro-
shape of the aggregates were often predicted by philic moieties on the same molecule, they are
using the packing parameter (p), which is the ratio also able to self-assemble into a number of
of area occupied by the hydrophilic head and nanostructures in solution (Fig. 2b) and in bulk
hydrophobic groups of amphiphiles (Fig. 2c), which are analogous to those aggregates
(Israelachvili et al. 1976). The properties and per- of the small-molecule amphiphiles. Moreover,
formances of surfactants are closely related to their self-assembly behaviors follow the same
their self-assembly behaviors, and thus it still principle as that of surfactants (Mai and Eisenberg
requires much deeper understanding on the rela- 2012), and thus at least to some degree, the mor-
tionship among the molecular structures, aggrega- phologies of polymer assemblies can also be
tion behaviors, and application properties of these predicted by the dimensionless p, as shown in
amphiphiles. Fig. 2b. In addition, the hydrophilic/hydrophobic
ratio, copolymer concentration, solvent proper-
ties, etc. can signicantly affect the morphologies
Macromolecular Amphiphiles of the nal nanostructures (Mai and Eisenberg
2012). Due to the great versatility of polymers in
Amphiphilic polymers are another kind of typical terms of their molecular weight, chain architec-
amphiphilic molecules of high molecular weight ture, polydispersity, and reactivity as well as
Amphipols 75
synthetic diversity, amphiphilic polymers can pro- Israelachvili JN, Mitchell J, Ninham BW (1976) Theory of
vide greater opportunities for exibility, diversity, self-assembly of hydrocarbons amphiphiles into
micelles and bilayers. J Chem Soc Faraday Trans
and functionality. This is especially true in the 72:15251568 A
light of recent advances in controlled/living radi- Mai Y, Eisenberg A (2012) Self-assembly of block copol-
cal polymerization chemistry, such as ATRP and ymers. Chem Soc Rev 41:59695985
RAFT polymerization, for the preparation of Nagarajan R (2011) Amphiphilic surfactants and amphi-
philic polymers: principles of molecular assembly.
polymers of various architectures, including lin- ACS Symp Ser 1070(1):122
ear block copolymers, graft copolymers, dendritic Schacher FH, Rupar PA, Manners I (2012) Functional
polymers, star-like polymers, and cyclic poly- block copolymers: nanostructured materials
mers, as shown in Fig. 2a. Under certain condi- with emerging applications. Angew Chem Int Ed
51:225
tions, all of these types of amphiphilic polymers Schramm LL, Stasiuk EN, Marangoni DG (2003) Surfac-
can also self-organize into aggregates of diverse tants and their applications. Annu Rep Prog Chem Sect
morphologies (Mai and Eisenberg 2012; Wang C Phys Chem 99:348
and Grayson 2012). In addition, the size, mor- Wang Y, Grayson SM (2012) Approaches for the prepara-
tion of non-linear amphiphilic polymers and their appli-
phology, and function of the resulting polymeric cations to drug delivery. Adv Drug Deliv Rev
assemblies can be easily controlled by adjusting 64:852865
the structure of the amphiphilic copolymers. Zhang J, Li X, Li X (2012) Stimuli-triggered structural
Compared to small-molecule aggregates, polymer engineering of synthetic and biological polymeric
assemblies. Prog Polym Sci 37:11301176
aggregates exhibit higher stability and durability
(Mai and Eisenberg 2012). Particularly, stimuli-
sensitive polymer assemblies that can respond to
subtle changes in surrounding environment such
as pH, temperature, and solvent quality can be
feasibly realized via the self-assembly of polymer Amphiphilic Monolayer/Bilayer
amphiphiles with responsive groups (Schacher
et al. 2012; Zhang et al. 2012). Therefore, it is Amphiphilic Membrane
no wonder that amphiphilic polymers and their
assemblies have attracted considerable attention
and achieved signicant success in many elds,
such as biomedicine, biomaterials, pharmaceuti-
cals, microelectronics, and photoelectric Amphiphilic Polymers
materials.
Amphiphilic Molecules
Cross-References
Lipid Monolayer/Bilayer
Surface Tension Amphiphilic Substances:
Amphiphiles
References Encapsulation
Chevalier Y (2002) New surfactants: new chemical func-
tions and molecular architectures. Curr Opin Colloid
Interface Sci 7:311
Fuhrhop JH, Wang T (2004) Bolaamphiphiles. Chem Rev
104:29012937
Holmberg K, Jonsson B, Kronberg B, Lindman B (2002)
Amphipols
Surfactants and polymers in aqueous solution, 2nd edn.
Wiley, Chichester Amphiphilic Molecules
76 Amphoteric Ion Exchange Membrane
Amphoteric Ion Exchange Membrane
Tongwen Xu
Collaborative Innovation Centre of Chemistry for
Energy Materials, School of Chemistry and
Material Science, University of Science and
Technology of China, Hefei, China
Amphoteric ion exchange membrane is a kind

of membrane containing random arrangements of
both cation and anion exchange groups. When
rstly proposed by Sllner in 1932 (Sllner
1932), amphoteric ion exchange membranes Amphoteric Ion Exchange Membrane, Fig. 1 Transport
mainly refer to membranes having a combination number of hydrogen ions relative to sodium ions (PNaH)
and current efciency of cations to the composition of
of weak acidic (cation exchange) and weak basic
amphoteric ion exchange membranes. (,): quaternized
(anion exchange) groups. Nowadays, they are membranes (N-methyl pyridinium and sulfonic acid
expanded to include membranes containing a groups); (D, ~): tertiary amino groups membrane;
combination of weak acidic (or weak basic) with (pyridinium hydrochloride and sulfonic acid groups); (,
~): PNaH; (, D): current efciency (%). Electrodialysis
strong basic (or acidic) functionalities or a com-
was carried out at a current density of 20mAcm2 using a
bination of strong acidic and basic groups. Basi- mixed solution of 0.25 N sodium chloride and 0.25 N
cally, there are four types (Sata 2004): hydrochloric acid for 60 min at 25.0 C (Yamane
(1) membranes in which both anion and cation et al. 1965)
exchange groups are distributed at random during
the preparation procedures, such as copolymeri-
zation, (2) snake cage-type amphoteric mem- Meanwhile, some of the membranes have com-
branes, (3) polysalt membranes, and positions that show hydrogen ion
(4) membranes prepared by a layer-by-layer permselectivity relative to sodium ions in elec-
method. The amphoteric ion exchange mem- trodialysis (Yamane et al. 1965) (Fig. 1).
branes can show superior antifouling properties,
peculiar selectivity, and permeability and hence
are useful in the biomedical and industrial elds,
such as medical devices (e.g., hemodialysis References
membrane), separation of ionic drugs and pro-
teins, and depletion of electrolytes by Eguchi T, Mori S, Shimokawa M (1981) Process for pre-
nanoltration or piezodialysis as well as high paring a composite amphoteric ion exchange mem-
brane. US patent US874,687
performance gel actuator. Membrane having Sata T (2004) Ion exchange membranes: preparation, char-
properly combined anion and cation exchange acterization, modication and application. Royal soci-
groups shows selective permeability for inor- ety of Chemistry, Cambridge
ganic salts and, therefore, is able to separate a Sllner K (1932) Uber mosaikmembranen. Biochem
Z 244:390
low molecular organic substance from an inor- Yamane R, Izuo R, Mizutani Y (1965) Permselectivity
ganic salt, acid, or alkali as well as salt from acid of the amphoteric ion-exchange membranes(2).
or salt from alkali (Eguchi et al. 1981). Denki Kagaku (J Electrochem Soc Jpn) 33:589593
Analytical Pervaporation 77
Analytical Pervaporation
A
Wanqin Jin
State Key Laboratory of Materials-Oriented
Chemical Engineering, Nanjing University of
Technology, Nanjing, China
The relatively new developed pervaporation in

analytical chemistry is in principle a separation Analytical Pervaporation, Fig. 1 Conventional
technique for the removal of volatile analytes or pervaporation module. DC donor chamber, AC acceptor
their volatile derivatives from the sample matrix. chamber, MS membrane support, m membrane, AG air gap,
AS aluminum supports, DI, DO inlet/outlet of donor
Also, it can be used for sample pretreatment, e.g.,
stream, respectively, AL, AO inlet/outlet of acceptor
for solid samples where leaching and derivatiza- stream, respectively
tion of the analytes are done simultaneously. Ana-
lytical pervaporation can be dened as the
integration of evaporation and gas diffusion in a order to increase the volume of the corresponding
single module. Through a porous membrane, the chamber. The different parts of the module are
volatile substances present in a heated donor aligned by inserting metallic rods into four ori-
phase evaporate. And on the other side of the ces and a closer contact is achieved by screwing
membrane, the vapor condenses on the surface them between two aluminum supports (AS). The
of a cool acceptor stream. The driving force for whole module is normally made of methacrylate,
the separation is due to the temperature difference, which permits continuous monitoring of the sam-
which leads to a vapor pressure difference across ple level in the lower chamber during the experi-
the membrane. For analytical pervaporation, an ments Luque de Castro and Papaefstathiou
important feature is that the existence of a (1998).
constant-volume air gap between the sample in Constantly being hydrophobic, the membranes
the donor chamber and the membrane, which for analytical pervaporation are usually employed
hinders any contact between them. As a result, in ultraltration and gas-diffusion processes.
the clogging and/or deterioration of the membrane Practically, polytetrauoroethylene (PTFE) is the
can be avoided (Luque de Castro 1998). most frequently used membrane material,
The conventional version of the analytical followed by hydrophobic polyvinylidene uoride
pervaporation module is shown in Fig. 1 and (PVDF) Luque de Castro and Gdmiz-Gracia
consists of the following parts: (i) a lower donor (2000). In combination with the large surface
chamber (DC) in which the sample containing the area of both the donor and acceptor chambers,
analyte is introduced, either by injection or by the ultraltration membranes, which are very
continuous aspiration for liquid samples and by thin, lead to their easy bending. In result, the ux
direct weighing for solid samples, (ii) an upper of the permeating component changes through an
acceptor chamber (AC) in which an appropriate altered membrane area, and therefore, it leads to a
uid is circulated for receiving the pervaporated change in the efciency of the process in addition
analytes, (iii) a membrane support (MS) approxi- to reproducibility problems; consequently, the
mately 1 mm thick, and (iv) spacers of varying membranes must be changed quite often. Due to
thickness (between 2 and 10 mm) which can be their thickness, gas-diffusion membranes are not
placed above or below the membrane support in so easily bent, so the same membrane can be used
78 Anion-Exchange Membrane (AEM)
over longer periods. The efciency of the process Anion-exchange groups are positively charged
is inuenced by the pore size of the membrane groups: primary, secondary, and tertiary amino
(or its directly proportional thickness) and hence groups, quaternary ammonium groups, tertiary
the sensitivity of the method; it affords application sulfonium groups, quaternary phosphonium
to samples with variable analyte concentrations, groups, cobalticinum groups, and other groups
as a result. that provide a positive xed charge in aqueous
Analytical pervaporation can be used for the or mixed water and organic solvent solutions.
determination of a single analyte or a number of Based on the materials constituting membranes,
analytes, being of inorganic, organic, or organo- AEMs can be classied as (i) organic membranes,
metallic nature. Application of analytical (ii) inorganic membranes, and (iii) composite
pervaporation in the environmental eld has membranes of inorganic ion exchangers and
been reported especially for food analysis organic polymers. AEMs are also classied into
Amador and Luque de Castro (2000). which is two types by their microstructure: heterogeneous
the analytical eld in which pervaporation has and homogeneous. Heterogeneous AEMs consist
been most widely applied. Other elds of appli- of nely powdered anion exchanger and an inert
cation of pervaporation have been pharmaceutical binder polymer. Various methods have been reported
and clinical analyses. for preparing homogeneous AEMs. A typical exam-
ple of a hydrocarbon AEMs is a copolymer mem-
brane composed of styrene and divinylbenzene with
References benzyltrimethylammonium groups. These mem-
branes will be prepared, for example, by the reaction
Amador HJ, Luque de Castro MD (2000) Pervaporation: a of trimethylamine with a copolymer membrane pre-
useful tool in food analysis. Food Chem 68:387394
pared from chloromethylstyrene and divinylbenzene
Luque de Castro MD, Gdmiz-Gracia L (2000) Analytical
pervaporation: an advantageous alternative to head- or by alkylation with alkyl halide of a copolymer
space and purge-and-trap techniques. Chromatographia membrane prepared from vinylpyridine and
52:265272 divinylbenzene.
Luque de Castro MD, Papaefstathiou I (1998) Analytical
Though AEMs have been used in many elds,
pervaporation: a new separation technique. Trends
Anal Chem 17:4149 most are used in electrodialysis, separation of
electrolysis, and solid polymer electrolytes for
fuel cells. The properties required depend on the
intended application of AEMs. Generally required
properties are (1) low electrical resistance,
Anion-Exchange Membrane (AEM) (2) high transport number of anions, (3) low dif-
fusion coefcient of electrolytes, (4) low osmotic
Mitsuru Higa water and low electroosmotic water, (5) antifoul-
Graduate School of Science and Engineering, ing properties, (6) mechanical strength, (7) dimen-
Yamaguchi University, Ube, Yamaguchi, Japan sional stability, (8) high chemical stability and
durability, and (9) low cost (Sata 2004).
AEMs with low sulfate ion permeability have
The anion-exchange membrane (AEM) is a thin been industrially used to prevent precipitation of
lm with anion-exchange groups (positively calcium sulfate in the AEMs and electrodialyzer.
charged groups) and permeates anions selectively. A nitrate ion permselective AEM has been devel-
AEMs are classied according to the species oped and has contributed to human health because
of the ion-exchange groups and materials the concentrations of nitrate ions are greatly
constituting the membrane and microstructure. increased in groundwater. To prepare these
Anion-Exchange Resin 79
functionalized AEMs, ordinary AEMs are modi-

ed by suitable chemical or physical methods. For Anion-Exchange Resin
example, most commercially available homoge- A
neous AEMs are mainly cross-linked with Tongwen Xu
divinylbenzene. When the contents of Collaborative Innovation Centre of Chemistry for
divinylbenzene increase or compact layers are Energy Materials, School of Chemistry and
formed on the AEMs, the pore size of the mem- Material Science, University of Science and
brane decreases and the transport number of sul- Technology of China, Hefei, China
fate ions, which are bulky relative to chloride ions,
decreases.
Fouling of separation membranes is a common Anion-exchange resin contains positively
problem. AEMs are fouled by ionic materials of charged groups, such as NH3+, NRH2+,
medium molecular weight such as ionic surfac- NR2H+, NR3+, PR3+, SR2+, etc., which
tants having the charge opposite to the xed are xed to the membrane backbone and allow the
charged of the membrane. Almost all of organic exchange of anions but repulse cations. As shown
foulants in many efuent streams have negative in Fig. 1, the active groups existed in the body of
charges; hence, fouling of AEMs due to deposi- resin determine its chemical behavior. It has a
tion and/or adsorption of the foulants on/in the highly developed structure of pores on the surface,
AEM is one of the serious problems in their appli- which could easily trap and release ions.
cations. The pore size of the AEM is generally There are two main types of anion-exchange
recognized to be about 1 nm; therefore, ions of resins differing in their active groups: strongly
medium molecular weight permeate with dif- basic and weakly basic. (1) Strongly basic resins
culty through the membrane. Consequently, the are highly ionized, and their chemical behavior is
fouling can lead to an unacceptably high stack similar to that of a strong base which can be used over
resistance and replacement of membranes in an the entire pH range. They can react with anions in
electrodialysis system due to clogging of the solution and convert an acid solution to pure water.
membrane pores with the medium molecular The exhausted resin can be regenerated with concen-
weight ions. To alleviate the problem of organic trated sodium hydroxide (NaOH) and converted to
fouling, there are basically two methods on the the hydroxide form. (2) The degree of ionization of
membrane side: rst, to increase the pore size of weakly basic resin is strongly inuenced by pH
the membrane to allow easy permeation of large because it is like a weak base. The weak base resin
ionic materials and, second, to prevent penetration does not have a hydroxide ion form as does the
of the materials into the membrane at the mem- strong base resin. Consequently, its regeneration
brane surfaces. There are two methods to prevent needs only to neutralize the absorbed acid and does
penetration of large organic materials into the not need to provide hydroxide ions. Less expensive
membrane matrix: forming a thin charged layer weakly basic reagents such as ammonia (NH3) or
opposite in sign to the ion-exchange groups of the sodium carbonate can be employed.
membrane or forming a very thin and dense layer The selectivity of a resin for a given ion is
on the membrane surfaces. measured by the selectivity coefcient (K), as for
the reaction
Reference R A B R B A
Sata T (2004) Ion exchange membrane. The Royal Society The K can be expressed as K = (concentration
of Chemistry, Cambridge of B in resin/concentration of A in resin)
80 Annealing of Polymer Membranes
Anion-Exchange Resin,
Fig. 1 Diagram of anion-
backbone
exchange resin

+ +
+ + active group

+ active group (fixed group)

+
free ion
+
(concentration of A in solution/concentration of then cooled to room temperature in order to alter

B in solution). the material properties. The annealing process is
In general, introducing ionic groups could start time and temperature dependent.
with a monomer containing a moiety, a polymer Annealing of polymers is a heating of a poly-
lm, or start with polymers or polymer meric part to below its glass transition temperature
blends (Helfferich 1962; Strathmann 2004). As in order to relieve the internal stresses introduced
summarized in our recent book chapter, the during its fabrication (molding, cooling after
specic implementation methods include molding, machining, welding, etc.).
(1) amidoalkylation of cross-linking polystyrene, Annealing of conjugated polymers and poly-
(2) copolymerization of (chloromethyl) styrene, mer blend lms is a widely used process that
(3) chloroacetylation of polymers, (4) acetylation achieves optimal lm morphology and therefore
of cross-linked polystyrene, and (5) grafting improves material properties, such as electrical
vinylbenzyl ammonium chloride onto inert poly- mobility for photovoltaic devices and other appli-
mers (Xu 2009). cations. Annealing of ion-conducting polymers
reduces the excessive swelling observed in pres-
ence of water.
References
In particular, annealing is used (Young and
Helfferich FG (1962) Ion exchange. Dover Publications Lovell 1991):
Strathmann H (2004) Ion-exchange membrane separation
processes. Elsevier Science (a) To increase the melting point of some poly-
Xu T (2009) Advances in membrane science and technol-
ogy. Nova Science mers because it facilitates the melting of small
and irregular crystals in favor of larger and
better formed ones
(b) To reduce or cancel deformations previously
Annealing of Polymer Membranes provoked at higher temperature which remain
as frozen (permanent) deformations when the
Maria Luisa Di Vona material is cooled to room temperature
Dip. Di Scienze Technologie Chimiche, (c) To increase crystallization
University of Rome Tor Vergata, Rome, Italy (d) To decrease the free volume between poly-
meric chains thus inducing a mechanical sta-
bilization of polymeric materials
Annealing can be dened as a process where a
material undergoes a heat treatment to a certain Annealing of liquid crystalline polymers
temperature, is kept there for a denite time, and is between the glass transition temperature and the
Antibiological Fouling Surface 81
melting temperature leads to an increased level of the crystallization temperature (Tc) and it became
crystallinity and increased crystalline perfection clear that the annealing temperature cannot be
analogous to the behavior of conventional semi- higher than the melting temperature (Tm) of the A
crystalline polymers. Annealing at or above the semicrystalline phase.
melting point also leads to enhanced crystallinity
and crystalline perfection.
The annealing of semicrystalline polymers References
may change the crystal structure, the degree of
crystallinity, the perfection of the crystals, the Alberti G, Narducci R, Di Vona ML, Giancola S (2013)
Annealing of Naon 1100 in the presence of an
orientation of both crystalline and amorphous
annealing agent: a powerful method for increasing
phases, their contiguous structural morphology, ionomer working temperature in PEMFCs. Fuel Cells
and the number of tie chains between the crystal- 13:4247
lites. Morphological changes are also observed on Young RJ, Lovell PA (1991) Introduction to polymers,
2nd edn. CRC Press, Boca Raton
annealing bulk-crystallized samples. Polymor-
phism may result due to annealing crystalline
polymers as its one crystalline form transforms
into another. The concept that mechanical stabili-
zation was due to a generic decrease of the free Antibiological Fouling Surface
volume between ionomer chains when thermally
treated was abandoned and replaced by the con- Franois Perreault
sideration that the mechanical stabilization was School of Sustainable Engineering and the Built
induced by an increased crystallization of a Environment, Arizona State University, Tempe,
preexisting semicrystalline phase (Alberti AZ, USA
et al. 2013).
Thermal annealing is a simple route for stabi-
lizing glassy polymers via the densication of Antibiological fouling surface is a surface hav-
their polymer chains and is accomplished through ing the property to resist the attachment or prolif-
thermal treatment of the glassy thermoplastics. eration of living organisms. In membrane
For amorphous polymers, the thermal treatment systems, the main organisms of concern for
can play a very important role in changing the biological fouling are microorganisms. Anti-
morphology. If the temperature is lower than the biological fouling membranes may resist biologi-
glass transition temperature Tg, the polymer cal fouling, or biofouling, by three main
chains are mainly immobile, thereby maintaining mechanisms: reducing the attachment of microor-
the existing morphology; however, if the ganisms to the surface, killing the microorganisms
annealing temperature is greater than the Tg, the to prevent their multiplication, or inhibiting
polymer chains will relax. It has been found that their proliferation on the membrane surface
the morphology can be modied upon annealing (De Kwaadsteniet et al. 2011).
at a temperature higher than the Tg, thus changing Biological fouling can be reduced by reducing
the physical properties of the polymers. the adhesion of microorganisms and biological
A certain amount of molecular rearrangement foulants to the membrane. Adhesion of foulants
must take place during crystallization but this on a membrane depends on the physicochemical
rearrangement cannot arise through large-scale properties of the surface and is increased with
molecular diffusion as the process is occurring in higher surface roughness, hydrophobicity, and
the solid state. Thermal treatments were believed electrostatic charge. Surface roughness increases
to be more and more efcient as the temperature fouling by increasing the contact area between the
increases and therefore empirical short treatments cell and the surface (Vrijenhoek et al. 2001). Strat-
were performed at high temperatures. Thus, the egies to reduce the surface roughness include
annealing temperature (Tann) was connected with changes in the fabrication procedure or lling
82 Antibiotic Production by Membrane Operations
the valleys of the surface with polymers or Such compounds can be directly attached to the
nanomaterials (Rana and Matsuura 2010). Hydro- membrane or included in a slow-release platform
phobic surfaces increase fouling due to the hydro- such as polymeric particle (Glinel et al. 2012).
phobic nature of most biological foulants.
Increase in the hydrophilicity of a surface gener-
ates a hydrated layer that reduces the adhesion of References
hydrophobic proteins and microorganisms. To
achieve a more hydrophilic surface, the mem- Banerjee I, Pangule RC, Kane RS (2011) Antifouling coat-
ings: recent developments in the design of surfaces that
brane may be coated or functionalized with hydro-
prevent fouling by proteins, bacteria, and marine organ-
philic polymers and nanomaterials or chemically isms. Adv Mater 23:690718
modied by treating the membrane with plasma, De Kwaadsteniet M, Botes M, Cloete TE (2011) Applica-
UV light, different acids or bases, and oxidizing tion of nanotechnology in antimicrobial coatings in the
water industry. Nano 6:395407
agents (Rana and Matsuura 2010). Electrostatic
Glinel K, Thebault P, Humblot V, Pradier CM, Jouenne
interactions may increase the adhesion of micro- T (2012) Antibacterial surfaces developed from
organisms on the membrane. Proteins and micro- bio-inspired approaches. Acta Biomater 8:16701684
bial cells are generally negatively charged in Rana D, Matsuura T (2010) Surface modications for
antifouling membranes. Chem Rev 110:24482471
natural pH. Therefore, lowering the surface
Vrijenhoek EM, Hong S, Elimelech M (2001) Inuence of
charge may decrease foulant adhesion by increas- membrane surface properties on initial rate of colloidal
ing the electrostatic repulsion. The surface charge fouling of reverse osmosis and nanoltration mem-
of a membrane can be modied by changing the branes. J Membr Sci 188:115128
nature of the functional groups present on the
membrane surface (Rana and Matsuura 2010).
Finally, adhesion of biological foulants and
microorganisms may be decreased by surface Antibiotic Production by Membrane
modication with polymer brushes. The use of Operations
long-chain molecules such as polyethylene glycol
increases steric repulsion between foulants and Cristiana Boi
the membrane surface (Banerjee et al. 2011). Dipartimento di Ingegneria Civile, Chimica,
The antifouling effectiveness of the polymer Ambientale e dei Materiali, Universit di
brushes depends on the length and density of the Bologna, Bologna, Italy
polymer layer.
Biological fouling can be reduced by
inactivating microorganisms in order to prevent Synonyms
their development on the membrane surface. Anti-
microbial compounds can be integrated in a mem- Membranes in antibiotic production
brane by including them in the polymer solution
during membrane fabrication or by post- Antibiotics are medicines produced by fermenta-
fabrication membrane functionalization. These tion that ght bacterial infections; they can either
antimicrobial compounds may be biocide- kill other microorganisms or inhibit their growth.
releasing materials, contact-action antimicrobial The development of antibiotics started with the
agents, or photocatalytic materials (Banerjee discovery of penicillin by Fleming in 1928 that
et al. 2011). gave rise to many other discoveries in the subse-
Biological fouling may be reduced by quent 40 years. The number of new antibiotics on
preventing the development of microbial commu- the market had a steady increase until the 1980s,
nities without directly killing the microorganisms. but the unrestrained use of antibiotics led to the
This is accomplished by using bacteriostatic or emergence of antibiotic-resistant pathogens and
quorum-sensing inhibitors that prevent the assem- the number of approved antibiotics decreased
bly of deposited microorganisms into a biolm. constantly in the last 30 years (Coates
Antifouling Membrane Surface 83
et al. 2011). The research for therapeutics may also be used. The introduction of an ultral-
endowed with a broaden antimicrobial range tration step greatly improves the extraction pro-
brought to the development of semisynthetic anti- cess (Koltuniewicz 2010). A
biotics which are produced by chemical or enzy- Other membrane technologies like membrane
matic transformation of penicillin and contactors, liquid membranes, and membrane bio-
cephalosporins (Srirangan et al. 2013). Penicillin reactors are less established; however, they have
is still one of the antibiotics with the largest annual great potential for industrial applications in anti-
bulk production; it is part of the b-lactam antibi- biotic production.
otic class, with about 15 billion US$ annual sales
and a market share of about 65 % of the total
world antibiotic market (Elander 2003). References
The production of antibiotics typically
involves a series of process steps consisting of Coates ARM, Halls G, Hu Y (2011) Novel classes of
antibiotics or more of the same? Br J Pharmacol
fermentation, removal of biomass, purication,
163:184194
and crystallization (Koltuniewicz 2010). Mem- Elander RP (2003) Industrial production of ?-lactam anti-
brane operations can be integrated within this biotics. Appl Microbiol Biotechnol 61:385392
process sequence, both in the upstream and in Koltuniewicz A (2010) Integrated membrane operations in
various industrial sectors. In: Drioli E, Giorno L (eds)
the downstream processing, for production of
Comprehensive membrane science and engineering,
either natural or semisynthetic antibiotics. vol 4. Elsevier, Oxford, pp 109164
In upstream processing, membranes are uti- Lipnizki F (2010) Basic aspects and applications of mem-
lized for ancillary operations and, in particular, brane processes. In: Drioli E, Giorno L (eds) Compre-
hensive membrane science and engineering,
for sterile ltration of culture media and air
vol 4. Elsevier, Oxford, pp 165194
which is supplied as a source of oxygen for aero- Srirangan K, Orr V, Akawi L, Wesrbrook A,
bic fermentation, whereas in the downstream part Moo-Young M, Perry Chou C (2013) Biotechnological
of the process, they can be utilized as alternative advances on penicillin G acylase: pharmaceutical
implications, unique expression mechanism and pro-
unit operations. In the dowstream section,
duction strategies. Biotecnol Adv 31:13191332
membranes can be applied for cell harvest,
clarication, and concentration. Among the mem-
brane technologies used for recovery and puri-
cation, crossow microltration and ultraltration
can be employed for biomass separation, often in
combination with dialtration to improve product Antifouling Membrane Surface
yield and purity (Lipnizki 2010).
Antibiotic recovery from fermentation broths Zhongyi Jiang1, Jinming Peng2, Xueting Zhao1,
can be done directly if the product is extracellular, Yanlei Su1 and Hong Wu1
1
as in penicillin production, or following cell dis- School of Chemical Engineering and
ruption if the product is intracellular, as for some Technology, Tianjin University, Nankai, Tianjin,
tetracyclines. Since antibiotics are relatively small China
2
molecules, with molecular weights of less than Key Laboratory for Green Chemical Technology
2,000 amu, the MF/UF membrane retains the bio- of Ministry of Education, School of Chemical
mass or the cell debris when the antibiotic is an Engineering and Technology, Tianjin University,
intracellular products, while the antibiotics are Tianjin, China
found in the permeate side. The claried antibiotic
stream can then be concentrated with
nanoltration and/or reverse osmosis (Lipnizki The term fouling generally refers to an undesir-
2010). Purication of antibiotics is generally car- able process in which a surface becomes
ried out with solvent extraction, but depending on encrusted due to the undesirable accumulation
the target molecule, absorption or precipitation of foulants, such as biomacromolecules,
84 Antifouling Membrane Surface
microorganisms, hydrocarbons, particles, and col- (Callow and Callow 2011). The hydrophilic mem-
loids from the surrounding environment. Anti- brane surface can resist the fouling caused by
fouling surface, also referred to as a stealth biomacromolecules and natural organic matter.
surface, can minimize the intermolecular forces The hydrophobic surface is based on nonpolar,
of interactions between the foulants and the syn- low surface energy hydrophobic polymers, such
thetic surface so that the foulants can be easily as poly(dimethylsiloxane) (PDMS) elastomers
released under low shear stresses. Antifouling and uoropolymers. The mechanism of this
surfaces lie at the heart of several contemporary hydrophobic surface is that polar molecules
and advanced technologies, ranging from coat- (including adhesive protein) barely adhere to
ings for ship hulls, biomedical implants, biosen- membrane surface because of reduced opportuni-
sors, drug delivery, and membrane separations. ties for forming the hydrogen-bonding and polar
Antifouling membrane surface is one kind of anti- interactions. This hydrophobic surface aims to
fouling surfaces. The antifouling surface is often weaken the interfacial bonds so that the attached
relevant to membranes utilized in microltration, foulants are more readily removed by hydraulic
ultraltration, nanoltration, and reverse osmosis. shear forces, not prevent foulants from attaching,
The major foulants are different for different which is known as fouling release. The hydro-
membrane processes. The main foulants are phobic membrane surface is suitable for mitigat-
microorganisms and emulsied oils for ing the microorganism and hydrocarbon fouling.
microltration, biomacromolecules and colloids The amphiphilic surface comprising hydrophilic
for ultraltration, natural organic matters for domains and hydrophobic domains on membrane
nanoltration, inorganic salts and bioorganic/ surface incorporates some of the benets of both
organic substances for reverse osmosis. hydrophobic and hydrophilic surfaces. The gen-
All the strategies for construction of antifoul- eral aim is to create a dynamic and compositional
ing membrane surface are based on manipulating surface complexity that deters settlement of
the physicochemical and/or topography structure foulants and reduces the interfacial bonding with
of membrane surface to weaken the interactions membrane surface. The general aim of amphi-
between foulants and membrane surface and thus philic surface is to combine the nonpolar, low
inhibiting foulants adsorption or settlement. There surface energy properties of hydrophobic
are four major types of antifouling membrane domains to reduce polar and hydrogen-bonding
surfaces, including hydrophilic surface, hydro- interactions with foulants, with the well-known
phobic surface, amphiphilic surface, and reactive protein repellency properties of the hydrophilic
surface. Whitesides (Ostuni et al. 2001) rst pro- domains. The resulting amphiphilic surface may
posed that antifouling surface showing resistance lower the entropic and enthalpic driving forces for
to protein should have the following four attri- the adsorption of the protein and organism. The
butes: hydrophilic, hydrogen-bond acceptors but amphiphilic membrane surface can defer the foul-
not hydrogen-bond donors, and overall electri- ing caused by microorganisms, biomacro-
cally neutral. Poly(ethylene glycol) (PEG)-based molecules, natural organic matter, and
polymers and zwitterionic polymers are the most hydrocarbons. Therefore, the amphiphilic mem-
widely used materials to fabricate antifouling brane surface is the most promising antifouling
membranes. The antifouling mechanism of the membrane surface for the practical separation sys-
PEG-based polymers can be mainly ascribed to tem containing a wide range of foulants. The
the large repulsive hydration forces of tightly reactive surface immobilizes the component
bound water layer around the polymer chains via which can degrade organic composition of
hydrogen bonds or electrostatic attraction. The foulants on membrane surface. The proteases
objective of hydrophilic surface is to prevent the and titania nanoparticles are often immobilized
foulants from attaching onto the membrane sur- on membrane surface, thus preventing biomacro-
face, which is known as fouling resistance molecules and microorganism fouling via
Antimicrobial Activity 85
enzymatic or photocatalytic degradation References

(Banerjee et al. 2011).
Several modication approaches are employed Banerjee I, Pangule RC, Kane RS (2011) Antifouling coat-
ings: recent developments in the design of surface that A
to create antifouling membrane surface including
surface coating, surface grafting, and surface seg- isms. Adv Mater 23:690718
regation (Shannon et al. 2008; Rana and Matsuura Callow JA, Callow ME (2011) Trends in the development
2010; Wang et al. 2005). For surface coating, of environmentally friendly fouling-resistant marine
coatings. Nat Commun 2:244253
membranes are immersed in a polymer modier
Ostuni E, Chapman RG, Holmlin RE, Takayama S,
solution or inorganic particles suspension. The Whitesides GM (2001) A survey of structure-property
modier is settled on membrane surface to create relationships of surfaces that resist the adsorption of
antifouling coating through van der Waals force or protein. Langmuir 17:56055620
Rana D, Matsuura T (2010) Surface modication for anti-
electrostatic interactions. For surface grafting, the
fouling membranes. Chem Rev 110:24482474
modier is settled on membrane surface via graft Shannon MA, Bohn PW, Elimelech M, Georgiadis JG,
polymerization including graft-to and graft-from Marinas BJ, Mayes AM (2008) Science and technology
approach. The termed graft-to approaches consist for water purication in the coming decades. Science
452:301310
of the grafting of the pre-synthesized polymer
Wang YQ, Wang T, Su YL, Peng FB, Wu H, Jiang ZY
chains with a reactive terminal group. Alterna- (2005) Remarkable reduction of irreversible fouling
tively, graft-from approaches, in which a poly- and improvement of the permeation properties of poly
mer is grown in situ from the surface via a (ether sulfone) ultraltration membranes by blending
with Pluronic F127. Langmuir 21:185611862
surface-adsorbed initiation group, are generally
capable of producing denser polymer layers. The
surface coating and surface grafting belong to the
post-modication approach, whereas surface
segregation belongs to the in situ modication Antimicrobial Activity
approach. The polymer modiers comprising
hydrophobic backbones and hydrophilic side Franois Perreault
chains are added to casting solution, lining mem- School of Sustainable Engineering and the Built
brane surfaces, and internal pore during the con- Environment, Arizona State University, Tempe,
ventional phase inversion process. The AZ, USA
hydrophilic chains segregate to membrane sur-
face and endow membrane with surface hydro-
philicity and antifouling property, while Antimicrobial activity of a compound is dened
hydrophobic backbones anchor to membrane as its capacity to prevent the development of
matrix polymers to defer loss of modier during microorganisms. Antimicrobial compounds can
phase inversion process. be divided into two main categories according to
The fouling resistance property could be eval- their mechanisms of antimicrobial action: bacteri-
uated by ux recovery ratio (FRR), while the cidal and bacteriostatic. Bactericidal agents act by
fouling release property could be evaluated by killing bacterial cells. Common bactericidal
ux decline ratio (FDR). The ux recovery ratio mechanisms include inhibition of cell wall syn-
(FRR) is the ratio of the water ux after hydraulic thesis, disruption of the cytoplasmic membrane,
cleaning to initial water ux, while the ux and inhibition of nucleic acid synthesis. Bacterio-
decline ratio is the ratio of ux decline (the initial static agents are compounds that prevent micro-
water ux minus the ux for real feed) to initial bial development by inhibiting specic metabolic
water ux. The higher the value of FRR, the better pathways, thus preventing cell growth and multi-
the fouling resistance property of the membrane; plication. Bacteriostatic agents act by inhibiting
the lower the FDR, the better the fouling release protein synthesis, by acting on the 30S, 50S, or
property of the membranes. 70S ribosomal complexes, or by inhibiting other
86 Antimicrobial Membrane
essential metabolic pathways such as folic acid polycations, silver and metal oxide nanoparticles,
synthesis. Bacteriostatic effect is reversible and carbon nanomaterials, biocide-releasing poly-
bacteria can proliferate once the bacteriostatic meric structures, and photoactive materials
agent is removed (McCarter 2005). (Banerjee et al. 2011).
Methods to evaluate the antimicrobial activity
can be divided into two categories: diffusion
methods and dilution methods. Diffusion References
methods, also called agar diffusion methods,
depend on the diffusion of an antimicrobial com- Banerjee I, Pangule RC, Kane RS (2011) Antifouling coat-
pound in an agar plate covered by a bacterial lawn.
Antimicrobial activity is measured in terms of the isms. Adv Mater 23:690718
diameter of the zone of inhibition, where bacterial Matin A, Khan Z, Zaidi SMJ, Boyce MC (2011) Biofouling
growth is inhibited. The most common agar dif- in reverse osmosis membranes for seawater desalina-
tion: phenomena and prevention. Desalination
fusion assay is the Kirby-Bauer assay. Dilution
281:116
methods aim to determine the minimum inhibi- McCarter YS (2005) Antimicrobial modes of action. In:
tory concentration of a compound, dened as the Coyle MB (ed) Manual of antimicrobial susceptibility
minimal concentration that completely inhibits testing. American Society for Microbiology, Washing-
ton, DC
microbial growth, by incubating a dened quan-
tity of bacteria with different dilutions of an anti-
microbial compound. Dilution methods can be
performed in culture broth or in agar media.
In membrane processes, antimicrobial com- Antimicrobial Membrane
pounds are used to reduce microbial growth and
biological membrane fouling. Antimicrobial com- Jun Yin and Baolin Deng
pounds can be introduced in the feed efuent Department of Civil and Environmental
continuously or intermittently, during membrane Engineering, University of Missouri-Columbia,
cleaning stages, to kill microorganisms and Columbia, MO, USA
reduce biolm development (Matin et al. 2011).
Traditional approach using chlorine-based disin-
fectants is usually effective to control biolm Antimicrobial membrane (also called antibacterial
development; however, chlorine disinfection membrane or anti-biofouling membrane) is a
leads to the formation of undesirable and carcino- membrane capable of reducing or inhibiting
genic disinfection by-products. Additionally, microbial growth on its surface. The biocidal abil-
chlorine is not compatible with polyamide thin- ity of the antimicrobial membrane can generally
lm composite membranes due to the susceptibil- be exerted by (1) active amphiphilic polymers or
ity of the aromatic polyamide layer to chlorine synthetic mimics of naturally occurring
attack. Alternative strategies to disinfection of antibacterial peptides, (2) microbe-repelling anti-
the feed efuent include ozone, UV treatments, adhesive polymers, and (3) composite materials
or advanced oxidation processes, which produce containing various slow-releasing biocides
less disinfection by-products but are limited by (Sambhy et al. 2006; Yin and Deng 2015).
their higher cost compared to chlorine. Microbial
growth and biolm development may also be
reduced by imparting antimicrobial properties Why Antimicrobial Membrane Is
directly to the membrane by modication with Important?
antimicrobial compounds. Antimicrobial coating
of membranes can be achieved using different Membrane biofouling due to the microbial growth
types of antimicrobial compounds, such as anti- and biolm formation is one of the most challeng-
microbial polymers, antimicrobial peptides, ing issues in membrane separation for water and
Antimicrobial Membrane 87
wastewater treatment (Zhu et al. 2010). This Recently, Cu is explored as an alternative of Ag

undesirable phenomenon decreases membrane due to its relatively low cost and decent biocidal
permeability, reduces permeate quality, and property. The antimicrobial property of TiO2 is A
increases energy consumption of the separation driven by the photocatalytic reaction, where light
process. It is therefore critical to develop anti- irradiation is necessary.
microbial membranes that will most likely
increase membrane efciency and application
duration. In addition, use of antimicrobial
Fabrication and Modification Methods
membrane is important to provide pathogen-
free clean water.
The antimicrobial functionalization of mem-
branes could be implemented through procedures
including blending, surface coating, or grafting. In
Commonly Used Biocidal Agents
the blending approach, biocidal agents are mixed
for Membrane Modification
with polymer solution during the membrane fab-
rication. In the coating or grafting approach, bio-
Biocidal agents that could be used for water treat-
cidal agents are attached onto membrane surface
ment include quaternary ammonium compounds
through adsorption, deposition, electrostatic
(QACs), metal ions or nanoparticles (NPs) such as
attraction, or chemical bonding. For instance, Ag
silver (Ag) and copper (Cu), titanium dioxide
NPs or Cu NPs could be encapsulated in posi-
(TiO2), and so on.
tively charged polyethylenimine (PEI) and then
QACs possess good antibacterial properties,
attached onto a negatively charged membrane
low toxicity, environmentally friendly perfor-
surface (Mauter et al. 2011; Ben-Sasson
mance, and excellent cell membrane penetration
et al. 2014). In order to improve the durability of
properties (Song et al. 2011).
antimicrobial function, Ag NPs can also be
Silver ions/NPs and silver-based compounds
attached onto thin-lm composite (TFC) mem-
have excellent biocidal properties and are widely
branes through chemical bonding, resulting in
used to prepare antimicrobial plastics, coatings,
membranes with excellent antimicrobial proper-
and wound and burn dressing, etc. They could
ties and reduced membrane biofouling (Fig. 1, Yin
inhibit the growth of microbes through multiple
et al. 2013).
pathways. Firstly, released Ag+ ions could interact
with disulde or thiol groups of enzymes or DNA
and then disrupt microbial metabolic processes,
generate reactive oxygen species, or interrupt rep- Challenges
lication of DNA (Russell and Hugo 1994; Feng
et al. 2000; Matsumura et al. 2003), leading to the Antimicrobial membrane has attracted much
damage or even the death of bacterial cells. attention due to its demonstrated resistance to
Secondly, Ag NPs can attach to the surface of membrane biofouling. However, several chal-
the cell membrane and disturb its proper func- lenges need to be addressed in its practical appli-
tion including destabilizing the outer mem- cations. For long-term application, the loss of
brane, collapsing the plasma membrane antimicrobial activity due to the depletion of bio-
potential, and depleting the levels of intracellu- cidal agents or insufcient contact between bac-
lar ATP (Sondi and Salopek-Sondi 2004; Lok teria and biocide caused by other foulants
et al. 2006; Choi et al. 2008). Thirdly, Ag NPs necessitates frequent recharging and cleaning of
with smaller particle size (110 nm) are able to membranes. How to attach biocidal agents onto
penetrate inside the bacteria and cause further membrane and control their release in a most
damage by possibly interacting with sulfur- and effective way should be further explored. For
phosphorus-containing compounds such as commercial applications, there is a need to con-
DNA (Morones et al. 2005). sider cost-effectiveness of attaching biocidal
88 Antimicrobial Membrane
Antimicrobial Membrane, Fig. 1 SEM image of membrane surface (1) and TEM image of membrane cross section (2)
(Yin et al. 2013)
agents onto a membrane and recharging them as Matsumura Y, Yoshikata K, Kunisaki SI, Tsuchido
needed. T (2003) Mode of bactericidal action of silver zeolite
and its comparison with that of silver nitrate. Appl
Environ Microbiol 69:42784281
Mauter MS, Wang Y, Okemgbo KC, Osuji CO, Giannelis
Cross-References EP, Elimelech M (2011) Antifouling ultraltration mem-
branes via post-fabrication grafting of biocidal
Antifouling Membrane Surface nanomaterials. ACS Appl Mater Interfaces 3:28612868
Morones JR, Elechiguerra JL, Camacho A, Holt K, Kouri
Bacterial Biolm Formation JB, Ramrez JT, Yacaman MJ (2005) The bactericidal
Cross-Linked Aromatic Polyamide Membranes effect of silver nanoparticles. Nanotechnology
Inorganic Fillers 16:23462353
Mixed Matrix Membranes (MMMs) Russell AD, Hugo WB (1994) Antimicrobial activity and
action of silver. Prog Med Chem 31:351370
Nanocomposite Membranes Sambhy V, MacBride MM, Peterson BR, Sen A (2006)
Silver bromide nanoparticle/polymer composites: dual
action tunable antimicrobial materials. J Am Chem Soc
References 128(30):97989808
Sondi I, Salopek-Sondi B (2004) Silver nanoparticles as
Ben-Sasson M, Zodrow KR, Genggeng Q, Kang Y, antimicrobial agent: a case study on E. coli as a model
Giannelis EP, Elimelech M (2014) Surface functiona- for Gram-negative bacteria. J Colloid Interface Sci
lization of thin-lm composite membranes with copper 275:177182
nanoparticles for antimicrobial surface properties. Song J, Kong H, Jang J (2011) Bacterial adhesion inhibition
Environ Sci Tech 48:384393 of the quaternary ammonium functionalized silica
Choi O, Deng KK, Kim NJ, Ross L Jr, Surampalli RY, Hu nanoparticles. Colloids Surf B Biointerfaces 82:651656
Z (2008) The inhibitory effects of silver nanoparticles, Yin J, Deng B (2015) Polymer-matrix nanocomposite
silver ions, and silver chloride colloids on microbial membrane for water treatment. J Membr Sci 479:
growth. Water Res 42:30663074 256275
Feng QL, Wu J, Chen GQ, Cui FZ, Kim TN, Kim JO Yin J, Yang Y, Hu Z, Deng B (2013) Attachment of silver
(2000) A mechanistic study of the antibacterial effect nanoparticles (AgNPs) onto thin-lm composite (TFC)
of silver ions on Escherichia coli and Staphylococcus membranes through covalent bonding to reduce mem-
aureus. J Biomed Mater Res 52:662668 brane biofouling. J Membr Sci 441:7382
Lok CN, Ho CM, Chen R, He QY, Yu WY, Sun H, Tam Zhu X, Bai R, Wee KH, Liu C, Tang SL (2010) Membrane
PKH, Chiu JF, Che CM (2006) Proteomic analysis of surfaces immobilized with ionic or reduced silver and
the mode of antibacterial action of silver nanoparticles. their anti-biofouling performances. J Membr Sci
J Proteome Res 5:916924 363:278286
Antimicrobial Properties of Membranes 89
upon interaction with cell wall components or

Antimicrobial Properties penetration through the cytoplasmic membrane.
of Membranes This mechanism does not deplete over time and A
may be effective as long as a direct contact
Francois Perreault between the cell and the membrane surface is
School of Sustainable Engineering and the Built provided.
Environment, Arizona State University, Tempe, Photocatalytic materials generate reactive oxy-
AZ, USA gen species upon exposure to visible or ultraviolet
light. Their main applications in membrane pro-
cesses are contaminant degradation and disinfec-
Antimicrobial properties of membrane improve tion, where the membrane serves as a support to
the membrane resistance to biological fouling. immobilize photocatalysts. The generation of
Antimicrobial properties can be provided to mem- reactive oxygen species by these materials causes
branes by modication with different types of bacterial inactivation by inducing oxidative dam-
antimicrobial compounds, which can be divided age to cell membranes, proteins, and nucleic
into three categories based on their mechanisms of acids, leading to cell death. Integration of
bacterial inactivation: biocide-releasing com- photocatalytic materials into membranes therefore
pounds, contact-action antimicrobial agents, and improves the antimicrobial properties of the mem-
photocatalytic materials. brane (Kwak and Kim 2001). Several compounds
Biocide-releasing materials inactivate microor- may be light-activated to generate reactive oxygen
ganisms by the release of biocide in the vicinity of species. Among them are dyes likes rose bengal,
the membrane. Common examples of biocide- porphyrins, and toluidine blue and particles such
releasing compounds used for antimicrobial coat- as TiO2 and fullerene (Banerjee et al. 2011). The
ings are silver nanoparticles, metal oxide most common photocatalytic material used in
nanoparticles, and polymeric capsules (Banerjee membrane processes is TiO2 due to its low cost,
et al. 2011). Biocide-releasing materials may be durability, and self-cleaning properties.
included in the polymer solutions during fabrica-
tion or attached to the membrane via post-
fabrication membrane modication. Embedding
the biocide in the depth of the membrane may References
however limit its diffusion to the microorganisms
and therefore decrease the antimicrobial activity Banerjee I, Pangule RC, Kane RS (2011) Antifouling coat-
of the modied membrane (Zodrow et al. 2009). prevent fouling by proteins, bacteria, and marine organ-
To maximize the effectiveness of biocide release, isms. Adv Mater 23:690718
surface attachment of biocide can be preferred Kwak S-Y, Kim S (2001) Hybrid organic/inorganic reverse
(Mauter et al. 2011). A limitation to biocide- osmosis (RO) membrane for bactericidal anti-fouling.
1. Preparation and characterization of TiO2 nanoparti-
release materials is the eventual depletion of the cle self-assembled aromatic polyamide thin-lm-
biocide and loss of antimicrobial activity. composite (TFC) membrane. Environ Sci Technol
Contact-action antimicrobials are compounds 35:23882394
that require direct contact with microorganisms Mauter MS, Wang Y, Okemgbo KC, Osuji CO, Giannelis
EP, Elimelech M (2011) Antifouling ultraltration
for cell inactivation. Examples of contact-action membranes via post-fabrication grafting of biocidal
antimicrobial materials are antimicrobial nanomaterials. ACS Appl Mater Interfaces
peptides, enzymes, polycationic polymers, and 3:28612868
nanomaterials such as carbon nanomaterials Zodrow K, Brunet L, Mahendra S, Li D, Zhang A, Li Q,
Alvarez PJJ (2009) Polysulfone ultraltration mem-
(Banerjee et al. 2011). A common mechanism of branes impregnated with silver nanoparticles show
cell inactivation by contact-action antimicrobial improved biofouling resistance and virus removal.
compounds is the disruption of cell integrity Water Res 43:715723
90 Antioxidant Recovery by Membranes
improve the efciency of antioxidant production

Antioxidant Recovery by Membranes (Table 1).
Laurent Bazinet Pressure-Driven Membrane Processes

Department of Food Sciences, Universit Laval, Pressure-driven membrane technologies are often
Quebec, QC, Canada used in the food industry to isolate antioxidant
molecules such as specic peptides or peptide
fractions (Bazinet and Firdaous 2012; Pouliot
Antioxidants and Membrane Processes et al. 2000; Tessier et al. 2006a, b; Vandanjon
et al. 2009) and phenolic compounds (Nawaz
Antioxidants are natural or synthesized sub- et al. 2006; Kalbasi and Cisneros-Zevallos 2007;
stances that inhibit or retard the oxidation of Mello et al. 2010; Tylkowski et al. 2010). Pressure
other molecules (fat, oil, food, petroleum is the main driving force for these processes. The
product. . .) (Focke et al. 2012; Rozoy pressure used depends on the pore size of the
et al. 2012; Roblet et al. 2012). However, for ltration membrane and has to be adjusted as a
natural antioxidants, which are present in complex function of the concentration rate desired.
matrices, their separation and purication are Pressure-driven membrane processes comprise
required. Indeed, they must be separated from microltration (MF), ultraltration (UF),
the feedstock in order to increase their purity and nanoltration (NF), and reverse osmosis (RO).
antioxidant capacity. The solution properties, charges of ionic species
Membrane processes are the main processes and the type of membrane used, affect the behav-
used for the separation of antioxidant molecules. ior and the efciency of the separation method
Membranes are selective barriers that allow for (Martin-Orue et al. 1998). Working at low con-
the transmission of certain feed components centration is preferable to avoid the formation of a
while retaining other components. Membrane polarization layer which could affect selectivity of
technology can be used to concentrate and/or the membrane (Shahidi 2004).
selectively fractionate bioactive compounds For the separation of tocopherols and
with antioxidant activity from aqueous and alco- tocotrienols, only MF (de Souza et al. 2008) and
holic processing streams of products, NF (Subramanian et al. 1998a, b, 2003) were
by-products, and wastes from agro-food indus- reported. For phenols recovery, MF was used on
try (Diaz-Reinoso et al. 2011). It is important to vegetation waters (VW) (Russo 2007) and pine-
note that, when compared with other concentra- apple juice (Laorko et al. (2010). The authors
tion methods (evaporation, spray drying. . .), showed that MF did not affect signicantly the
during membrane processes the product is not pH, reducing sugar and acidity of claried juice
subjected to high temperatures and there is no whereas the suspended solids and microorganism
change in the physical state of the solvent, were completely removed. UF was studied on
meaning that the functional properties of the different sources such as olive mill wastewater
compounds of interest are preserved and the (Russo 2007; El-Abbassi 2012), fruit juices
process as a whole is energy saving (Negrao (Borneman et al. 2001; Cassano et al. 2008;
Murakami et al. 2011). Laorko et al. 2010; Ciss et al. 2011; Conidi
et al. 2011), grape seeds (Nawaz et al. 2006),
and green tea (Li et al. 2005). NF was tested on
Membrane Processes and Applications roselle extract (Ciss et al. 2011), on mate (Negrao
to Antioxidant Recovery Murakami et al. 2011), on aqueous propolis
extract (Mello et al. 2010), and on bergamot
On a large production scale, pressure-driven juice (Conidi et al. 2011). Organic acids such as
membrane technologies and electrically driven Ascorbic, isoascorbic, folic, and citric acids can
membrane technologies have been developed to also be recovered by ltration process. Hence,
Antioxidant Recovery by Membranes 91
Antioxidant Recovery by Membranes, Table 1 Membrane processes for separation/concentration of different types
of antioxidants
Driving force Membrane process Antioxidants References
A
Pressure-driven Microltration Tocopherols (de Souza et al. 2008)
membrane and
processes tocotrienols
Phenols (Russo 2007; Laorko et al. 2010)
Organic (Laorko et al. 2010)
acids
Ultraltration Phenols (Russo 2007; Laorko et al. 2010; El-Abbassi 2012;
Borneman et al. 2001; Cassano et al. 2008; Ciss
et al. 2011; Conidi et al. 2011; Nawaz et al. 2006; Li
et al. 2005)
Organic (Cassano et al. 2006)
acids
Peptides (Bouhallab and Touz 1995; Lajoie et al. 2001;
Vandanjon et al. 2007)
Nanoltration Tocopherols (Subramanian et al. 1998a, b, 2003)
Phenols (Ciss et al. 2011; Negrao Murakami et al. 2011; Mello
et al. 2010; Conidi et al. 2011)
Peptides (Vandanjon et al. 2007; Fenton-May et al. 1971; Tessier
et al. 2006b)
Electrically Electrodialysis Phenols (Bazinet et al. 2005)
driven membrane Organic (Vera Calle et al. 2003)
processes acids
Electrodialysis Organic (Vera Calle et al. 2002)
with bipolar acids
membranes
Electrodialysis Phenols (Labb et al. 2005; Bazinet et al. 2009, 2012)
with ltration Organic (Bazinet et al. 2012)
membranes acids
Peptides (Langevin et al. 2012)
MF was used for the recovery of ascorbic acid of molecular weight (Vandanjon et al. 2007). The
(Laorko et al. 2010), while UF was tested on use of integrated membrane system (UF/MF/RO,
kiwifruit juice to recover folic, ascorbic, and citric MF/NF, . . .) is becoming another real alternative
acids (Cassano et al. 2006). Concerning antioxi- to recover, purify, or concentrate antioxidants as it
dant peptides, they are separated into fractions is established in some recent works (Paraskeva
using membranes in the range 110 kDa et al. 2007; Russo 2007; Garcia-Castello
(Je et al. 2005; Jeon et al. 1999; Rajapakse et al. 2010; Conidi et al. 2011; Diaz-Reinoso
et al. 2005). NF can be used to concentrate hydro- et al. 2011).
lysates (Vandanjon et al. 2007; Fenton-May
et al. 1971; Tessier et al. 2006b) whereas UF Electrically Driven Membrane Processes
membranes with high MWCO (20100 kDa) are Electrically driven processes, like electrodialysis
adapted to the separation of peptides and (ED) and hybrid processes (ED with bipolar mem-
nonhydrolyzed proteins or proteolytic enzymes branes (EDBM) and ED with ltration mem-
(Bouhallab and Touz 1995; Lajoie et al. 2001). branes (EDFM)), use ion-exchange membranes
On the other hand, UF membranes with interme- and/or ltration membrane to separate molecules
diate molecular weight cutoffs (about (Bazinet 2005). The electric eld is the main
40008000 Da) allow hydrolysates to be fraction- driving force involved in these processes. Molec-
ated with the result of enrichment in some ranges ular transfer is mainly due to the charge of the
92 Antioxidant Recovery by Membranes
molecule and the ux depends on the strength of EDFM recovered larger range of peptide molec-
the electric eld (Poulin et al. 2007; Doyen ular weights and amount of polar amino acids.
et al. 2012).
Conventional electrodialysis has been tested
as a mean of selectively extracting polyphenols
from aqueous tobacco extracts (Bazinet References
et al. 2005). High polyphenol migration rates
around 31.190.8 % were reported. EDFM a Bazinet L (2005) Electrodialytic phenomena and their
new electrically driven technology has been applications in the dairy industry: a review. Crit Rev
Food Sci Nutr 44:525544
recently tested for the recovery of polyphenols.
Bazinet L, Firdaous L (2013) Separation of bioactive pep-
EDFM was used on green tea to perform selective tides by membrane processes: technologies and
extraction of catechins (Labb et al. 2005) and on devices. Recent Pat Biotechnol 7 (1): 927
cranberry juice to enrich cranberry juice in its own Bazinet L, DeGrandpr Y, Porter A (2005)
Electromigration of tobacco polyphenols. Sep Purif
natural phenolic antioxidant compounds (Bazinet
Technol 41:101107
et al. 2009, 2012). For organic acids having anti- Bazinet L, Cossec C, Gaudreau H, Desjardins Y (2009)
oxidant properties,Vera Calle et al. (2003) were Production of a phenolic antioxidant enriched cran-
indirectly precursor in the use of ED for the berry juice by electrodialysis with ltration membrane.
J Agric Food Chem 57:1024510251
recovery of citric acid. In all the deacidied juices,
Bazinet L, Brianceau S, Dub P, Desjardins Y (2012) Evo-
a decrease in the citrate and malate concentrations lution of cranberry juice physico-chemical parameters
was obtained. Vera Calle et al. (2002) also tested during phenolic antioxidant enrichment by electrodial-
the deacidication of claried passion fruit juice, ysis with ltration membrane. Sep Purif Technol
87:3139
by electrodialysis with bipolar membrane
Borneman Z, Gkmen V, Nijhuis HH (2001) Selective
(EDBM). Citric acid was formed in the concen- removal of polyphenols and brown colour in apple
trate compartment by citrate ions extracted from juices using PES/PVP membranes in a single ultral-
juice and protons provided by the BM separating tration process. Sep Purif Technol 2223:5361
Bouhallab S, Touz C (1995) Continuous hydrolysis of
the concentrate and electrode compartments. This
caseinomacropeptide in a membrane reactor: kinetic
conguration allowed the production of citric acid study and gram-scale production of antithrombotic
with 89 % purity. During EDFM of cranberry peptides. Lait 75:251258
juice, Bazinet et al. (2012) observed that enriched Cassano A, Figoli A, Tagarelli A, Sindona G, Drioli E
(2006) Integrated membrane process for the production
juice concentrations in citric and malic acids
of highly nutritional kiwifruit juice. Desalination
decreased respectively of 7.3 % and 4.7 % 189:2130
resulting in a signicant decrease of its sour taste Cassano A, Donato L, Conidi C, Drioli E (2008) Recovery
intensity, a major problem for cranberry juice of bioactive compounds in kiwifruit juice by ultraltra-
tion. Innov Food Sci Emerg Technol 9:556562
producer. However, the concentration in the raw
Ciss M, Vaillant F, Pallet D, Dornier M (2011) Selecting
juice was not signicant due to the high ratio of ultraltration and nanoltration membranes to concen-
raw juice/enriched juice. Very recently, a combi- trate anthocyanins from roselle extract (Hibiscus
nation of UF membrane stacked in an electrodial- sabdariffa L.). Food Res Int 44:26072614
Conidi C, Cassano A, Drioli E (2011) A membrane-based
ysis cell was tested successfully for the selective
study for the recovery of polyphenols from bergamot
separation of anionic and cationic antioxidant juice. J Membr Sci 375:182190
peptides (Langevin et al. 2012). The results de Souza MP, Cunha Petrus JC, Guaraldo Gonalves LA,
showed only a signicant increase in the antioxi- Viotto LA (2008) Degumming of corn oil/hexane
miscella using a ceramic membrane. J Food Eng
dant capacity for anionic peptide fractions recov-
86:557564
ered at pH 3 and pH 6 compared to the feed Daz-Reinoso B, Moure A, Domnguez H, Paraj JC
hydrolysate (Langevin et al. 2012). In the com- (2011) Membrane concentration of antioxidants from
parative study of Langevin et al. (2012), it was Castanea sativa leaves aqueous extracts. Chem Eng
J 175:95102
demonstrated that NF was more efcient in terms
Doyen A, Saucier L, Beaulieu L, Pouliot Y, Bazinet L
of mass ux than EDFM when compared on a (2012) Electroseparation of an antibacterial peptide
same basis (membrane area, process duration), but fraction from snow crab by-products hydrolysate by
Antioxidant Recovery by Membranes 93
electrodialysis with ultraltration membranes. Food Negro Murakami AN, de Mello Castanho Amboni RD,
Chem 132:11771184 Prudncio ES, Amante ER, de Moraes Zanotta L,
El-Abbassi A, Kiai H, Hadi A (2012) Phenolic prole and Maraschin M, Cunha Petrus JC, Telo RF (2011) Con-
antioxidant activities of olive mill wastewater. Food centration of phenolic compounds in aqueous mate A
Chem 132:406412 (Ilex paraguariensis A. St. Hil) extract through
Fenton-May RI, Hill CG, Amundson CH (1971) Use of nanoltration. LWT Food Sci Technol 44:22112216
ultraltration and reverse osmosis systems for the con- Paraskeva CA, Papadakis VG, Tsarouchi E, Kanellopoulou
centration and fractionation of whey. J Food Sci DG, Koutsoukos PG (2007) Membrane processing for
36:1421 olive mill wastewater fractionation. Desalination
Focke WW, van der Westhuizen I, Loft Grobler AB, 213:218229
Nshoane KT, Reddy JK, Luyt AS (2012) The effect of Poulin J-F, Amiot J, Bazinet L (2007) Improved peptide
synthetic antioxidants on the oxidative stability of bio- fractionation by electrodialysis with ultraltration
diesel. Fuel 94:227233 membrane: inuence of ultraltration membrane
Garcia-Castello E, Cassano A, Criscuoli A, Conidi C, stacking and electrical eld strength. J Membr Sci
Drioli E (2010) Recovery and concentration of poly- 299:8390
phenols from olive mill wastewaters by integrated Pouliot Y, Gauthier SF, LHeureux J (2000) Effect of pep-
membrane system. Water Res 44:38833892 tide distribution on the fractionation of whey protein
Je J-Y, Kim S-Y, Kim S-K (2005) Preparation and hydrolysates by nanoltration membranes. Lait
antioxidative activity of hoki frame protein hydrolysate 80:113120
using ultraltration membranes. Eur Food Res Technol Rajapakse N, Mendis E, Jung W-K, Je J-Y, Kim S-K
221:157162 (2005) Purication of a radical scavenging peptide
Jeon Y-J, Byun H-G, Kim S-K (1999) Improvement of from fermented mussel sauce and its antioxidant prop-
functional properties of cod frame protein hydrolysates erties. Food Res Int 38:175182
using ultraltration membranes. Process Biochem Roblet C, Amiot J, Lavigne C, Marette A, Lessard M,
35:471478 Jean J, Ramassamy C, Moresoli C, Bazinet L (2012)
Kalbasi A, Cisneros-Zevallos L (2007) Fractionation of Screening of in vitro bioactivities of a soy protein
monomeric and polymeric anthocyanins from concord hydrolysate separated by hollow ber and spiral-
grape (Vitis labrusca L.) juice by membrane ultraltra- wound ultraltration membranes. Food Res Int
tion. J Agric Food Chem 55:70367042 46:237249
Labb D, Araya-Farias M, Tremblay A, Bazinet L (2005) Rozoy E, Simard S, Liu Y, Kitts D, Lessard J, Bazinet L
Electromigration feasibility of green tea catechins. (2012) The use of cyclic voltammetry to study the
J Membr Sci 254:101109 oxidation of l-5-methyltetrahydrofolate and its preser-
Lajoie N, Gauthier SF, Pouliot Y (2001) Improved storage vation by ascorbic acid. Food Chem 132:14291435
stability of model infant formula by whey peptides Russo C (2007) A new membrane process for the selective
fractions. J Agric Food Chem 49:19992007 fractionation and total recovery of polyphenols, water
Langevin M-E, Roblet C, Moresoli C, Ramassamy C, and organic substances from vegetation waters (VW).
Bazinet L (2012) Comparative application of pressure- J Membr Sci 288:239246
and electrically-driven membrane processes for isola- Shahidi F (2004) Functional foods: their role in health
tion of bioactive peptides from soy protein hydrolysate. promotion and disease prevention. J Food Sci
J Membr Sci 403404:1524 69:146149
Laorko A, Li Z, Tongchitpakdee S, Chantachum S, Subramanian R, Nakajima M, Kawakatsu T (1998a)
Youravong W (2010) Effect of membrane property Processing of vegetable oils using polymeric composite
and operating conditions on phytochemical properties membranes. J Food Eng 38:4156
and permeate ux during clarication of pineapple Subramanian R, Nakajima M, Kimura T, Maekawa T (1998b)
juice. J Food Eng 100:514521 Membrane process for premium quality expeller-pressed
Li P, Wang Y, Ma R, Zhang X (2005) Separation of tea vegetable oils. Food Res Int 31:587593
polyphenol from green tea leaves by a combined Subramanian R, Raghavarao KSMS, Nakajima M,
CATUFM-adsorption resin process. J Food Eng Nabetani H, Yamaguchi T, Kimura T (2003) Applica-
67:253260 tion of dense membrane theory for differential perme-
Martin-Orue C, Bouhallab S, Garem A (1998) ation of vegetable oil constituents. J Food Eng
Nanoltration of amino acid and peptide solutions: 60:249256
mechanisms of separation. J Membr Sci 142:225233 Tessier B, Harscoat-Schiavo C, Marc I (2006a) Contribu-
Mello BCBS, Petrus JCC, Hubinger MD (2010) Concen- tion of electrostatic interactions during fractionation of
tration of avonoids and phenolic compounds in aque- small peptides complex mixtures by UF/NF mem-
ous and ethanolic propolis extracts through branes. Desalination 200:333334
nanoltration. J Food Eng 96:533539 Tessier B, Harscoat-Schiavo C, Marc I (2006b) Selective
Nawaz H, Shi J, Mittal GS, Kakuda Y (2006) Extraction of separation of peptides contained in a rapeseed
polyphenols from grape seeds and concentration by (Brassica campestris L.) protein hydrolysate using
ultraltration. Sep Purif Technol 48:176181 UF/NF membranes. J Agric Food Chem 54:35783584
94 Antioxidants Recovery by Integrated Membrane Operations
Tylkowski B, Trusheva B, Bankova V, Giamberini M, which should be processed by selective tech-

Peev G, Nikolova A (2010) Extraction of biologically niques able to produce concentrated streams with
active compounds from propolis and concentration of
extract by nanoltration. J Membr Sci 348:124130 enhanced antioxidant activity.
Vandanjon L, Johannsson R, Derouiniot M, Bourseau P, The recovery of antioxidants from agricultural
Jaouen P (2007) Concentration and purication of blue by-products can be conducted according to the 5-
whiting peptide hydrolysates by membrane processes. Stages Universal Recovery Processing strategy
J Food Eng 83:581589
Vandanjon L, Grignon M, Courois E, Bourseau P, Jaouen P which includes (1) a macroscopic pretreatment,
(2009) Fractionating white sh llet hydrolysates by (2) a macro- and micro-molecule separation,
ultraltration and nanoltration. J Food Eng 95:3644 (3) an extraction step, (4) an isolation and puri-
Vera Calle E, Ruales J, Dornier M, Sandeaux J, cation step, and (5) a product formation
Sandeaux R, Pourcelly G (2002) Deacidication of
the claried passion fruit juice. Desalination (Galanakis 2012).
149:357361 Conventional extraction methodologies are
Vera Calle E, Ruales J, Dornier M, Sandeaux J, Persin F, based on the use of resins, organic solvents,
Pourcelly G, Vaillant F, Reynes M (2003) Comparison enzymes, g-irradiation, and heat treatment.
of different methods for deacidication of claried
passion fruit juice. J Food Eng 59:361367 These methodologies have drawbacks to some
degree. For example, the extraction with organic
solvents is characterized by safety problems
(some of them are believed to be toxic), low
efciency, and time consumption; heat treatment
Antioxidants Recovery by Integrated results in pyrolysis; enzymes can be denatured in
Membrane Operations enzyme-assisted extraction; and g-irradiation-
assisted extraction is still unknown in terms of
Alfredo Cassano and Rosalinda Mazzei safety.
Institute on Membrane Technology, National Membrane technologies are the main processes
Research Council of Italy, ITM-CNR, Rende, used at large scale for the separation of antioxidant
Italy compounds from different sources. They offer
several advantages over conventional methods,
in terms of high separation efciency, low energy
Natural antioxidants are synthesized by plants, requirements, mild operating conditions, no addi-
microorganisms, fungi, and animals. They include tives, absence of phase transition, simple equip-
phenolics, terpenes, proteins, peptides, and carbo- ment, and easy scale-up. Pressure-driven
hydrates. Waste streams from the processing of membrane operations include microltration
agricultural feedstocks (residues from grape and (MF), ultraltration (UF), nanoltration (NF),
wine, apple pomace, peels, woods, olives, seeds, and reverse osmosis (RO). These processes
leaves) are particularly attractive as sources of cover most of the 5-Stages Universal Recovery
antioxidants. Processing strategy since they can be used as
Antioxidant compounds play an essential role macroscopic pretreatment (MF), in macro- and
in the human body to reduce oxidative processes. micro-molecule separation (UF), in purication
Therefore they offer interesting perspectives and (NF), and in concentration (RO).
opportunities in the production of dietary supple- These processes lead to the partial separation
ments and functional foods and potential utiliza- of the targeted compounds through size
tion in pharmaceutical and cosmetic industries. exclusion.
However, since these compounds are generally In addition to pressure-driven membrane oper-
present in complex matrices, extraction, separa- ations, membrane distillation (MD), osmotic dis-
tion, and purication methods are required in tillation (OD), and electrodialysis (ED) have
order to improve their purity. Extraction methods emerged as new membrane technologies with
ideally should be nondestructive, time efcient, great potential in integrated systems for the recov-
and suitable for producing high quantity of extract ery of antioxidants from vegetable sources.
Antioxidants Recovery by Integrated Membrane Operations 95
Retentate
MICROMOLECULES SEPARATION
A
Pre-
FEED MF UF/ED
treatment
MACROSCOPIC PRE-TREATMENT
NF
RO
ISOLATION AND PURIFICATION
OD/MD
CONCENTRATION
Antioxidants Recovery by Integrated Membrane microltration, UF ultraltration, NF nanoltration, RO

Operations, Fig. 1 Schematic representation of antioxi- reverse osmosis, OD osmotic distillation, MD membrane
dants recovery by membrane operations (MF distillation, ED electrodialysis)
In addition to RO, the concentration of bioac- low-molecular weight (MW) molecules cross the
tive compounds can be also performed by mem- membrane, while higher-MW molecules remain
brane distillation (MD) or osmotic distillation in the feed.
(OD). In these processes the driving force for the Specic applications of pressure-driven and
removal of water vapor across the pores of a electrically driven membranes processes for the
hydrophobic membrane is given by a vapor pres- recovery of antioxidant molecules have been
sure difference between the two sides of the mem- recently reviewed by Bazinet and Doyen (2016).
brane generated by a concentration (OD) or A schematic of integrated membrane opera-
thermal (MD) gradient. Electrically driven pro- tions, such as those previously described, for the
cesses, such as electrodialysis (ED) or a combina- recovery, separation, purication, and concentra-
tion of ED with UF membranes (EDUF), have tion of antioxidants, is depicted in Fig. 1. Such
demonstrated very high selectivity for the separa- ow sheet is an example of exible approach to
tion of organic charged biomolecules. These pro- valorize different agro-food by-products, as alter-
cesses use ion-exchange membranes and/or native to their disposal, within the logic of process
ltration membranes (such as UF) to separate intensication strategy.
molecules. The electric eld is the main driving A further improvement of the proposed inte-
force involved in these processes. Molecular grated system has been recently investigated
transfer is mainly due to the charge of the mole- through the addition of a nal step of membrane
cules and the ux depends on the strength of the emulsication (EM) or biocatalytic membrane
electric eld. In EDUF, fractionation of molecules reactor (BMR) nalized to the formulation and
is a function of their charges and molecular valorization of antioxidant compounds in olive
weights: according to the molecular weight cutoff mill wastewaters (OMWWs) (Bazzarelli
of the UF membrane located in the cell, et al. 2015; Conidi et al. 2014).
96 Antiscalant
References membrane. The selective and precise dosing of

the antisolvent, controlled by the porous mem-
Bazinet L, Doyen A (2016) Antioxidants, mechanisms, brane, allows a more ne control of the crystal-
and recovery by membrane processes. Crit Rev Food
lizing solution composition during the process
Sci Nutr. doi:10.1080/10408398.2014.912609
Bazzarelli F, Poerio T, Mazzei R, DAgostino N, Giorno and at the nucleation point, with consequent
L (2015) Study of OMWWs suspended solids destabi- advantages for the nal crystal characteristics.
lization to improve membrane processes performance. In solvent/antisolvent demixing conguration,
Sep Purif Technol 149:183189
a certain solute is dissolved in an appropriate mix
Conidi C, Mazzei R, Cassano A, Giorno L (2014) Inte-
grated membrane system for the production of of a solvent and an antisolvent whose composition
phytotherapics from olive mill wastewaters. J Membr is chosen in such a way that the solute remains
Sci 454:322329 indenitely in the solution in the original condi-
Galanakis C (2012) Recovery of high added-value compo-
tions. When a gradient of chemical potential is
nents from food wastes: conventional, emerging tech-
nologies and commercialized applications. Trends generated between the two sides of the mem-
Food Sci Technol 26:6887 branes, e.g., by a temperature difference, the sol-
vent, which is supposed to have a higher vapor
pressure than the antisolvent at the same temper-
ature, evaporates at higher ow rate thus produc-
Antiscalant ing solvent/antisolvent demixing. As the amount
of solvent in the mixture decreases, the lower
Scale Inhibitor solubility of the solute generates supersaturation
and a phase separation occurs when the
antisolvent exceeds a certain volume fraction.
The requirements for this conguration are that:
Antisolvent Membrane (i) the antisolvent and the solvent are miscible;
Crystallization (ii) the initial solvent/antisolvent balance in the
mixture guarantees that the solute is under its
Gianluca Di Proo solubility limit; and (iii) the solvent evaporates at
Institute on Membrane Technology, National higher velocity than the antisolvent. Crystalliza-
Research Council of Italy, ITM-CNR, Rende, tion of a solute from an aqueous/organic extrac-
Italy tion mixture can be successfully carried out by
this system.
In antisolvent addition conguration, a sol-
This is an extension of the solvent evaporation ute is dissolved in a solvent and then an
membrane crystallization concept, where a antisolvent is gradually evaporated from the
porous membrane is used to dose the amount other side of the membrane by applying a gra-
of antisolvent in the crystallizing solution, dient of chemical potential. As the antisolvent
which has been proposed very recently mixes with the solvent the solute dilutes but, at
(Di Proo et al. 2009). In this system, two the same time, the composition of the mixture
operative congurations have been developed: changes. At a certain point, the excess of
solvent/antisolvent demixing and antisolvent antisolvent creates supersaturation and solute
addition membrane crystallization. In both crystallization. This conguration requires that
cases, solvent/antisolvent migration occurs in the antisolvent and the solvent are miscible. In
vapor phase, according with the general con- this case, the compound to be crystallized can
cept of membrane crystallization and not by be soluble in aqueous solutions and poorly sol-
forcing it in liquid phase through the uble in organic low boiling liquids.
Antoine Equation 97
References the liquid states of the component, which depends

on the cohesive forces linking the molecules. In a
Di Proo G, Stabile C, Caridi A, Curcio E, Drioli E (2009) closed cell, the vapor pressure of a pure compo-
Antisolvent membrane crystallization of pharmaceuti- A
nent is a nonlinear relation of the temperature: the
cal compounds. J Pharm Sci 98:4902
more the component is volatile, the more the
vapor pressure is high.
Antoine Equation Antoine Equation
Denis Roizard For a given pure component in a closed cell, it

Laboratoire Ractions et Gnie des Procds, calculates the saturated vapor pressure P (mmHg)
CNRS- Universit de Lorraine, Nancy, France with the temperature T ( C) as follows:
log10 P A B=T C (1)

Synonyms
with A, B, and C being the Antoine coefcients, in
Vapor PressureTemperature Relationship mmHg and C units, which are component-
specic constants; e.g., for water : A = 8.07131;
B = 1730.63 and C = 233.426 in the tempera-
History ture range 199 C.
Other forms of the Antoine equation can be
Until the mid-nineteenth century, the prediction of deduced:
the saturated vapor pressure of a liquid mixture or
even of a pure liquid in relation with the temper- P 10AB=TC (2)
ature was not accurate despite the phenomena
being studied long before the Middle ages. The T B=A log10 P C (3)
Antoine equation, which solves this problem for
pure components, is due to a French engineer These equations t very well for experimental
(Louis Charles Antoine, 18251897) and was vapor pressure data.
rst published in Annales de Physique et de It is worth noting that the Antoine equation can
Chimie in 1891 (Antoine 1891). Antoine intro- also be used at high pressure or near the critical
duced an equation able to predict the vapor pres- point. Indeed, for each component, two sets of
sure of pure liquids (vaporization) and solids parameters can be used according to the consid-
(sublimation). It is worth noting that this equation ered temperature range, i.e., either up to the nor-
is still widely used today because of its accuracy. mal boiling point or from normal boiling point to
Wisniak (2001) has recently reviewed the histor- the critical point; in this last range, the Clausius-
ical development of the vapor pressure equations Clapeyron relation (Reid 1990) does not apply
from Dalton to Antoine. properly. In the case of pure water, in the temper-
ature range 99374 C, the dedicated Antoine
coefcients are A = 8.140191, B = 1810.94,
Phenomenon and C = 244.485.
The extension (Wagner 1973) of the Antoine
The vapor pressure over a liquid is due to the equation has been worked out to cover with a
thermodynamic equilibrium between the gas and single set of parameters the whole temperature
98 Antoine Equation
Antoine Equation, 180000

Fig. 1 Curves of P(sat) H2O
versus temperature; range 160000
50150 C calculated with
the Antoine equation 600
140000
ZOOM
120000 400 0100C
P mmHg
100000 200
80000 00
100
60000
40000
20000
0
0 100 200 300 400
T C
range going from the critical point to the triple For C coefcient, subtract 273.15 to take into
point. This expression is required when computa- account the modication of the
tional techniques must be used. temperature unit.
On the other hand, a simplied equation using
only two parameters was known to evaluate the The Antoine coefcients have been tabulated
vapor pressure; it was the August equation in for most of pure compounds. They can be
which the parameter C is set to zero, thus assum- obtained from various sources, including web
ing a temperature-independent heat of access Data banks of Antoine Coefcients
vaporization. (visited June 2013).
Antoine Coefficients Example Calculation
Two systems of units can be used, either based This calculation has been done for water with the
on temperature and pressure, respectively, in coefcients: A = 8.07131, B = 1730.63, and

C and in mmHg or on K and Pa (SI system). C = 233.426. The temperature validity range is
Note that for historic reasons, the Antoine from 1 C to100 C.
coefcients are still normally given based on
the CGS system. Conversion from the histori- At T = 50 C: log10 (P) = 8.071311730.63/
cal system unit to the SI one can be made (50 + 233.426) = 1.9652
easily: Hence, P(sat) H2O = 92 mmHg
B coefcient is the same in both systems; The results shown in Fig. 1 and Table 1 corre-
To get A coefcient in SI, add 2.124903 to the spond to water vapor pressure calculated with the
historical A; this value corresponds the respective sets of parameters of each domain with
pressure unit modication, i.e., to log10 the Eq. 1. The discontinuity at 100 C can be seen
(101325/760). in the Table 1.
Apple Juice and Membrane Operations 99
Antoine Equation, Table 1 Example calculation of P(sat) H2O; range 50150 C

T C log10 (P) P mmHg T C log10 (P) P mmHg
50 1.9652012 92.30 100 2.8832575 764.29 A
60 2.1732982 149.04 110 3.0320002 1076.47
70 2.3676788 233.17 120 3.1745848 1494.81
80 2.5496558 354.53 130 3.3093227 2038.56
90 2.7203796 525.27 140 3.4368443 2734.29
100 2.8808629 760.09 150 3.5577142 3611.72
A = 8.07131 B = 1730.63 C = 233.426 A = 8.140191 B = 1810.94 C = 244.485
References phenolic compounds. Most part of the juice is

consumed as claried product.
Antoine C (1891) Annales de Physique et de Chimie 22: Traditional methods of processing apple juice
281; ibid, Annales de Physique et de Chimie (1892) 26:
are based on the use of several batch operations
426; Comptes Rendus Acad Sci (Paris) (1888) 107:
1143 that are labor and time-consuming. Typically,
Data banks of Antoine Coefcients (visited June 2013) after preliminary operations (sorting, washing,
http://webbook.nist.gov/ http://booksite.elsevier. and peeling steps), apple fruits are crushed and
com/9780080966595/content/Appendices/Appendix%
then pressed.
20C.pdf (free download) http://www.eqi.ethz.ch/fmi/
xsl/eqi/eqi_property_details_en.xsl?node_id=983 The juice is submitted to a heat treatment for
(directory for Physical Properties Sources) Vapor- 16 s at temperature of 72 C to inhibit yeast
Liquid Equilibrium Data Collection, DECHEMA growth and avoid fermentation. During the press-
Chemistry Data Series, Jurgen Gmehling
ing step, pectinase is added in order to hydrolyze
et al. (Frankfurt)
Reid CE (1990) Chemical thermodynamics. McGraw-Hill, the pectin, reducing juice viscosity and making
New York, p 73 the product easier to lter. This process is
Wagner W (1973) New vapour pressure measurements for performed at temperatures in the range of 40 to
argon and nitrogen and an new method for establishing
50 C for 1 h. After this step, ning agents such as
rational vapour pressure equations. Cryogenics
13(8):470482 bentonite, gelatine, or silica solution are added to
Wisniak J (2001) Historical development of the vapor the juice to improve the effect of settling.
pressure equation from Dalton to Antoine. J Phase After decanting, the juice is submitted to a
Equilib 226:622630. doi:10.1007/s11669-001-
precoat ltration, using lter aids, i.e., diatoma-
0026-x1143
ceous earths and perlite, to remove suspended
solids, colloidal particles, proteins, and
polyphenols.
Apple Juice and Membrane The nal product can be submitted to a con-
Operations centration process by evaporation. It is generally
performed in multiple effect evaporators at tem-
Carmela Conidi peratures between 45 C and 90 C. During this
Institute on Membrane Technology, National step, undesirable alterations of the original aroma
Research Council of Italy, ITM-CNR, Rende, prole of the juice due to high operating temper-
Italy atures are produced.
Membrane operations are competitive and
attractive alternatives to the conventional treat-
Apple juice is one of the most popular juices ments of apple juices. The advantages of using
worldwide. It approximately accounts for a 20 % membrane technology, over conventional
of the juice market share, with a production of 1.5 methods, are related to product costs, working
million tons per year. Besides vitamins and min- conditions, environmental aspects, and product
erals, apple juice is characterized by high levels of quality.
100 Apple Juice and Membrane Operations
Apple juice clarication, stabilization, recov- RO is a pressure-driven membrane process that

ery of aroma, and concentration are typical steps can be used as an alternative process for apple
where membrane operations as microltration juice concentration, as it does not involve phase
(MF), ultraltration (UF), enzymatic membrane change or the use of high temperatures. The main
reactors (EMR), pervaporation (PV), reverse advantages of RO concentration are the attain-
osmosis (RO), osmotic distillation (OD), and ment of high-quality products due to low operat-
membrane distillation (MD) are successfully ing temperatures, resulting in the retention of
utilized. nutritional compounds and lower energy
The clarication of apple juice by membrane consumption.
processes has been used commercially for over However, the performance of RO mem-
30 years. In particular, the use of MF and UF branes is limited by the osmotic pressures of
represents a valid alternative to traditional the juice when the concentration reaches
methods in apple juice clarication and stabiliza- values of 2530 Brix; therefore, RO can be
tion, resulting in increase juice yield, improved considered an advantageous technique as a
product quality, and avoidance of ning agents pre-concentration step.
(gelatin, bentonite, and silica sol) and lter aids Polyamide and cellulose acetate RO mem-
that are costly and present disposal problems. branes in both tubular and spiral wound congu-
MF or UF permits the removal of haze-forming rations have been extensively used during the
components, such as suspended solids, colloidal concentration of apple juice (Alvarez et al. 2002).
particles, and proteins resulting in the production In order to overcome the limitation of RO in
of superior juice quality. Turbidity values less apple juice concentration , more recently, osmotic
than 0.10.2 NTU are typically obtained if com- distillation (OD) and membrane distillation
pared to 25 NTU of conventional methods with (MD) have been proposed as attractive membrane
ning agents. processes for apple juice concentration allowing
Polymeric membranes made from very high concentrations (up to 65 Brix) to be
polyethersulfone (PES), polyvinylidene uoride reached under atmospheric pressure and at room
(PVDF), cellulose acetate (AC), and polysulfone temperature, thus avoiding thermal and mechani-
(PS) with pore size of 0.10.2 mm and molecular cal damage (Onsekizoglu et al. 2010).
weight cutoff (MWCO) of 10100 kDa are used These different membrane-based concentra-
for the clarication of apple juice. tion techniques resulted very efcient in the con-
Tubular, capillary, and plate-and-frame mem- centration of apple juice, since the concentrated
brane modules are the most important congura- juice presented nutritional and sensorial quality
tions used at industrial level. similar to that of the original juice, especially
The claried juice is usually concentrated in regarding the retention of the bright color and
order to reduce storage, package, and shipping pleasant aroma, which are lost during thermal
costs. In addition, concentrated apple juice is evaporation.
more stable, presenting higher resistance to Esters, aldehydes, and alcohols such as ethyl
microbial and chemical deterioration than the butanoate, ethyl-2-methylbutanoate, and hexanal
original juice as a result of water activity are the main compounds which contribute to the
reduction. aroma prole of apple juice.
Conventional methods of apple juice con- PV constitutes a promising alternative to tradi-
centration, such as vacuum evaporation, usu- tional techniques, such as distillation and partial
ally employ high temperature to remove water. condensation, for aroma recovery from apple
However, heat determines undesirable changes juices.
of juice quality such as color changes, Alvarez et al. (2000) proposed an integrated
off-avor formation, and reduction in the membrane process for producing apple juice and
nutritional value. apple juice aroma concentrates.
Application of FEEM to Monitoring Membrane Fouling 101
Apple
References
Alvarez S, Riera FA, lvarez R, Coca J, Cuperus FP, Th

Juice extraction PV Bouwer S, Boswinkel G, Van Gemert RW, Veldsink A
JW, Giorno L, Donato L, Todisco S, Drioli E, Olsson J,
Trgrdh G, Gaeta SN, Panyor L (2000) New inte-
apple juice aroma recovery grated membrane process for producing claried
apple juice and apple juice aroma concentrate. J Food
Eng 46:109125
Alvarez S, Riera FA, lvarez R, Coca J (2002) Concentra-
tion of apple juice by reverse osmosis at laboratory
depectinization EMR and pilot-plant scales. Ind Eng Chem Res 41:
61566164
Onsekizoglu P, Savas Bahceci K, Jale Acar M (2010)
MF UF Clarication and concentration of apple juice using
membrane processes: a comparative quality assess-
ment. J Membr Sci 352:160165
Clarified juice
RO OD MD
Application of FEEM to Monitoring
Membrane Fouling
Concentrated juice Raymond L. Legge
Department of Chemical Engineering, University
Apple Juice and Membrane Operations, of Waterloo, Waterloo, ON, Canada
Fig. 1 Membrane operations for the treatment of apple
juice
FEEM refers to a uorescence excitation-

The process involves the clarication of raw emission matrix that consists of a collection of
apple juice by using an enzymatic membrane reac- intensity values for a group of excitation-emission
tor, an RO unit to pre-concentrate the claried juice pairs over a range of emission and excitation
up to 25 Brix, a PV unit to recover and concentrate wavelengths. FEEM or uorescence excitation-
aroma compounds, and a nal evaporation step to emission matrix is important in the context of
concentrate the juice up to 72 Brix. monitoring membrane fouling because NOM, or
Rejection of aroma compounds in the PV step natural organic matter, is a signicant foulant in
exceeded 90 % for most compounds considered. some cases for drinking water production, and
Organoleptic evaluation of the claried and con- some components of NOM are uorescent.
centrated juice resulted excellent in terms of odor Fluorophores can be classied into two categories
and avor. The nal products were more clear and that can be described as intrinsic or extrinsic
brilliant than apple juice produced by conven- (Lakowicz 2006). Two of the three major fractions
tional methods. of NOM are intrinsic uorophores: protein and
An economic evaluation of the integrated humic substances. Other biological constituents
membrane system indicated a reduction of the that may end up in water may also be uorescent
total capital investment of 14 % and an increase and can contribute to the FEEM, but this will
in process yield of 5 % when compared with the depend on the water source. For the protein com-
conventional process. ponent, the amino acids that can contribute to the
A conceptual process design based on the inte- uorescent signal are tryptophan, tyrosine, and
gration of membrane operations in apple juice phenylalanine. The third component, the colloid/
processing is depicted in Fig. 1. particulate matter, may be captured in the
102 Aptamer Membrane Functionalization
Rayleigh scattering portion of the FEEM and is Peiris RH, Hall C, Budman H, Moresoli C, Peldszus S,
not necessarily uorescent (Peiris et al. 2010a). Huck PM, Legge RL (2010b) Identifying fouling
events in a membrane-based drinking water treatment
This approach has the advantage that little or no process using principal component analysis of uores-
sample preparation is required so analysis is rapid cence excitation-emission matrices. Water Res
and can possibly be conducted online. Analysis of 44:185194
the FEEM data is required because it is rich in
information, but various approaches are required
to identify or correlate which components are
contributing to fouling. Among the various
approaches for analysis of the data is principle Aptamer Membrane
component analysis or PCA (Peiris et al. 2010b). Functionalization
The advantage of FEEM-based analysis is that
this approach can provide potentially online infor- Thomas Schfer
mation. Possible options that would allow data Institute for Polymer Materials (Polymat),
acquisition online would include ow through University of the Basque Country (UPV/EHU),
systems or ber optic approaches. The data Donostia-San Sebastin, Spain
acquired can then be used in concert with instan-
taneous membrane ux or transmembrane pres-
sure (TMP) measurements to garner more
information related to the important foulants. Aptamer Membrane Functionalization. Aptamers
Advantages of this approach are that water can be incorporated into adequate porous mem-
pretreatment strategies can be tuned or tailored brane structures in order to obtain stimulus-
to minimize the components that are contributing responsive membranes whose permeability is
to the NOM-based fouling or to take appropriate modulated via a molecular recognition event.
action before membrane fouling becomes signi- The concept relies on the fact that aptamers can
cant. Similarly, it may provide a useful technique recognize very specically a molecular target,
for monitoring the membrane permeate for NOM upon which a signicant conformational change
constituents that may contribute to disinfection can occur if the aptamer is designed accordingly.
by-products (DBPs) which can serve as possible In this sense, aptamer-modied membranes fol-
carcinogens. low the concept structure determines separa-
tion. The stimulus-responsive membranes so
obtained therefore do not require a bulk stimulus,
such as a change in temperature or pH, which
Cross-References makes them particularly interesting for being
employed in biomedical separations or
Fouling in Membranes DNA-based nanodevices (Bhattacharyya
Irreversible Fouling et al. 2013). Incorporating the aptamer into an
Membrane Fouling adequate porous structure, the aptamer conforma-
tional change upon target recognition can give rise
to a hindered pore ow, modulating in this way
References the overall membrane permeability. Aptamer-
Lakowicz JR (2006) Principles of uorescence spectros-
target interactions can be highly specic, target
copy, 3rd edn. Springer, New York concentration dependent, and do not involve the
Peiris RH, Budman H, Moresoli C, Legge RL (2010a) formation of chemical bonds but are physical. As a
Understanding fouling behavior of ultraltration mem- consequence, membranes functionalized with
brane processes and natural water using principal
component analysis of uorescence excitation-
aptamers change their permeability depending on
emission matrices. Water Sci Technol Water Supply the target concentration and in a reversible manner.
11(2):179185 Figure 1 depicts the concept of a membrane pore
Aptamer Membrane Functionalization 103
Aptamer Membrane Functionalization, Fig. 1 Scheme of a stimulus-responsive aptamer-functionalized membrane
c
a
3. Amino functionalized
Aptamer
100 nm 10 nm HN
b 120 O
100 2. GMBS crosslinker

O
80
Intensity
N
O
60
40
S
1. Thiol Addition
20 via silanization
0 Si
O O
150 160 170 180 190 200 210 220 O
Average Size (nm) Silica Surface
Aptamer Membrane Functionalization, Fig. 2 Mesoporous particles (a), their size distribution (b), and the
functionalization of mesoporous silica with an aptamer (c)
functionalized with a DNA-aptamer hairpin struc- open and the membrane permeability maximum.
ture which changes its conformation upon a molec- Upon interaction with AMP, the aptamer undergoes
ular stimulus such as adenosine 50 -monophosphate a conformational change which signicantly
(AMP). In the absence of the target AMP, the reduces the pore ow and, hence, overall membrane
aptamer-functionalized membrane pores remain permeability.
104 Aptamer Membrane Functionalization
Aptamer Membrane Functionalization, change upon target recognition (b), and function to release
Fig. 3 Sequence of a DNA-aptamer hairpin that binds in a reversible, specic, and concentration-dependent fash-
selectively to ATP (a), scheme of its conformational ion cargo molecules (c)
For the conformational change to take effect, conformation. It could be shown that the thickness
the pore diameter needs to be of the same order of change of a DNA-aptamer hairpin lm amounts
magnitude. Therefore, mesoporous structures are up to about 2 nm upon interaction with the target
preferably employed that furthermore possess a (Serrano-Santos et al. 2012). Hence, mesoporous
high degree of isoporosity. Aptamer-modulated structures with a pore diameter between 2 and
pore ow has been thoroughly studied for 3 nm have been found to be a suitable base mate-
mesoporous particles that are used as controlled rial for aptamer-functionalized particles and mem-
delivery devices (zalp and Schfer 2011). branes (zalp and Schfer 2011).
Figure 2 shows the immobilization strategy for
modifying mesoporous silica particles with
aptamers and Fig. 3 the concept of how such an
aptamer can serve as a reversibly opening lid in References
order to liberate cargo molecules upon target rec-
ognition (here: ATP). Bhattacharyya D, Schfer T, Wickramasinghe RR, Daunert
S (eds) (2013) Responsive membranes and materials.
A key parameter for the functioning of Wiley, Hoboken
aptamer-functionalized membranes is the ne- zalp VC, Schfer T (2011) Aptamer-based switchable
tuning of the pore size within which the aptamer nanovalves for stimuli-responsive drug delivery.
conformational change takes place. Pores too Chem Eur J 11:98939896
Serrano-Santos MB, Llobet E, zalp VC, Schfer T (2012)
large would not result in any aptamer-modulated
Characterization of structural changes in aptamer lms
permeability, while pores too narrow would hin- for controlled release nanodevices. Chem Commun
der the aptamer from freely changing its 48:1008710089
Aptamer Screening 105
against a given target (i.e., protein, small mole-

Aptamer Screening cules, or even whole cells) by an iterative process
that involves (i) library/target interaction, A
Thomas Schfer (ii) selection of the sequences that bind to the target
Institute for Polymer Materials (Polymat), and removal of no binders, (iii) enrichment of the
University of the Basque Country (UPV/EHU), selected sequences by PCR, and (iv) after several
Donostia-San Sebastin, Spain screening rounds, cloning and sequencing to iden-
tify individual sequences. Finally, bioinformatic
analysis is required to determine the most promis-
Aptamer screening or selection, also called sys- ing motifs or sequences. Next, the candidate
tematic evolution of ligands by exponential enrich- sequences are evaluated to conrm their binding
ment (SELEX), is a combinatorial chemistry capability. SELEX technology could be compared
method by which aptamers (single-stranded DNA to antibody production; however, SELEX offers
or RNA) are selected. SELEX is an evolutionary interesting features that overcome some limitations
method driven by the binding of oligonucleotide of the antibody production, such as fast in vitro
sequences to a specic target based on their partic- performance, no need for cells or animals, ease of
ular tridimentional structure, which confers high selection under physiological or nonphysiological
afnity and specicity. The selection process starts conditions, no limitation for target selection, and
with a random oligonucleotide library consisting of batch-to-batch reproducibility.
a random region anked by xed primer regions Since the rst report of SELEX in 1992 by two
required for PCR amplication. The random region independent groups (Tuerk and Gold 1990;
confers outstanding variability, derived from the Ellington and Szostak1990), several modica-
number of possible sequences in the library. tions of the standard method have been reported
The variability of the library can be calculated to improve its performance. Nowadays, a signi-
with the expression 4n, where 4 is the number of cant number of aptamers have been selected and
possibilities at each position (A, T, C, or G) and used as recognition molecules for environmental,
n is the length of the random region. SELEX food, diagnosis, and therapeutic applications
allows the screening of oligonucleotide libraries (Fig. 1).
Aptamer Screening,
Library Design by regions
Fig. 1 Aptamer selection
fixed random fixed Cloning, Sequencing
process from library
and Bioinformatics
Library
Interaction
Library/Target
Target
iteration
Selection of Target
Binders
Removal of
no binders
Enrichment of
selected binders
106 Aptamers
References intermolecular interactions creates unique binding

pockets. Hence, along with a remarkable afnity,
Ellington AD, Szostak W (1990) In vitro selection of RNA aptamers are generally characterized by high
molecules that bind specic ligands. Nature
specicity. One of the most noteworthy examples
346:818822
Tuerk C, Gold L (1990) Systematic evolution of ligands by is the anti-theophylline aptamer which can dis-
exponential enrichment: RNA ligands to bacteriophage criminate between theophylline and caffeine,
T4 DNA polymerase. Science 249:505510 two related molecules that only differ by a methyl
group of the imidazole ring, having about 11,000
times higher afnity for the former. Similarly,
aptamers have been reported to discriminate
Aptamers between enantiomers, macromolecules, proteins,
and whole cells.
Thomas Schfer Apart from their high afnity and specicity,
Institute for Polymer Materials (Polymat), aptamers possess unique chemical and biochemi-
University of the Basque Country (UPV/EHU), cal characteristics which clearly set them apart
Donostia-San Sebastin, Spain from other receptors. For example, aptamers are
exceptionally stable: they may undergo denatur-
ation, but the process is fully reversible within
Aptamers (from Latin aptus, to t) are articial minutes; hence, temperature changes or long-
ligands binding with high afnity and specicity term storage does not affect their functionality.
to their cognate target, selected through a Furthermore, the well-known chemistry of
Darwinian-like evolution method referred to as aptamers allows for site-directed chemical modi-
SELEX. Aptamers generally consist of structured cations and competitive production costs. New
single-strand nucleic acid molecules such as RNA functional groups can be introduced a
and ssDNA; however, dsDNA and peptide priori within the SELEX process using mod-
aptamers have been described as well (Colas ied nucleic acid libraries or a posteriori modify-
et al. 1996; Patel et al. 1997). ing a selected DNA/RNA through chemical
Aptamers can be selected against virtually any synthesis. While the rst approach relies on the
target and under nonphysiological conditions compatibility of the modication adopted with the
because no animal host is required. Hitherto, enzymatic amplication necessary for selection,
aptamers were selected against a broad range of the latter approach is streamlined by detailed
target molecules, such as amino acids, peptides, knowledge of the structure of both the aptamer
proteins, drugs, organic and inorganic molecules, and the aptamer-target complex. Since their rst
or even whole cells, with afnities often compa- appearance in the early 1990s, aptamers have
rable to those of monoclonal antibodies. received constantly increasing attention, reected
The binding proprieties of aptamers are due to in their use as diagnostic reagents and therapeutic
the formation of specic aptamer-target com- compounds.
plexes stabilized by non-covalent interactions.
The latter are a combination of van der Waals
forces, hydrogen bonds, and electrostatic interac-
References
tions. Binding of an aptamer to its cognate target
can trigger an adaptive folding in which the target Colas P, Cohen B, Jessen T, Grishina I, McCoy J, Brent
promotes and stabilizes the secondary and tertiary R (1996) Genetic selection of peptide aptamers that
structures of aptamers. The dynamic conforma- recognize and inhibit cyclin-dependent kinase 2. Nature
tions of free aptamers, consisting of labile 380(6574):548550
Patel DJ, Suri AK, Jiang F, Jiang L, Fan P, Kumar RA,
knots, loops, and stems, turn into stable architec- Nonin S (1997) Structure, recognition and adaptive
tures within the aptamer-target complexes binding in RNA aptamer complexes. J Mol Biol
where noncanonical base pairing along with 272(5):645664
Aquaporin-Incorporated Biomimetic Membranes 107
aquaporin-incorporated materials onto a porous

Aquaporin-Incorporated Biomimetic support, or (3) autopainting membrane (APM)
Membranes method. A
LB and LS monolayer transfer (Sun et al. 2012):
Chung Tai-Shung Neal1, Honglei Wang2, The rst monolayer of amphiphilic materials
Guofei Sun3, Hoang Hanh Phuoc Duong4 and (lipids or copolymers) is deposited onto a solid
Pei Shan Zhong5 support by the LB method. The second monolayer,
1
National University of Singapore, Singapore, which contains aquaporin-incorporated amphi-
Singapore philic materials, is subsequently deposited onto
2
Department of Chemical and Biomolecular the rst layer by the LS method.
Engineering, jMedtech Pte Ltd, Singapore, Vesicle rupture: Aquaporins are rst incorpo-
Singapore rated into liposomes (Borgnia et al. 1999; Wang
3
Chemical and Biomolecular Engineering, et al. 2011) or polymersomes (Kumar et al. 2007)
Aquaporin Asia Pte. Ltd, Singapore, Singapore through a typical transmembrane protein incorpo-
4
King Abdullah University of Science and ration method (Borgnia et al. 1999). Subse-
Technology (KAUST), Biological and quently, these aquaporin-incorporated vesicles
Environmental Science and Engineering Division are ruptured onto a porous support via charge
(BESE), Thuwal, Saudi Arabia interaction (Wang et al. 2011; Kaufman
5
Functionalized Materials and Nanostructures/ et al. 2010; Li et al. 2012) or covalent conjugation
Global Research Centre Singapore (GMM/F), between the vesicles and the support
BASF South East Asia Pte Ltd., Singapore, (Wang et al. 2012; Zhong et al. 2012; Duong
Singapore et al. 2012).
APM method: Free-standing aquaporin-
incorporated lipid/polymeric membranes can be
formed analogous to the conventional black lipid
Aquaporin, a transmembrane protein, has membrane preparation method (Hansen
attracted attention globally since its serendipitous et al. 2009; Gonzlez-Prez et al. 2009).
discovery by Peter Agre in 1992 (Agre 2006). Its A mixed solution of aquaporins and amphiphilic
facilitation in rapid and selective water transport materials is smeared across the hydrophobic
has inspired many scientists to utilize this unique arrays apertures in electrolyte conditions. The
function for water treatment applications membranes are spontaneously formed on the
(Kaufman et al. 2010; Wang et al. 2012). Much apertures at the interface of organic and aqueous
effort has been devoted to build and overcome the phases.
challenges faced in fabricating the synthetic bio-
mimetic membrane housing the water channel
(Kumar et al. 2012). It is envisioned that this Aquaporin Biomimetic Membrane
technology will bring about higher efciency for Designs
desalination (Kaufman and Freger 2011) and
hence lower cost. According to their conformations, the designs of
aquaporin-embedded biomimetic membranes can
be classied into two categories, namely, (1) free-
standing lipidic/polymeric membranes (Fig. 1)
Membrane Preparation Methods and (2) supported lipidic/polymeric membranes
(Fig. 2).
Aquaporin-incorporated membranes can be The free-standing lipidic/polymeric membrane
prepared via the (1) Langmuir-Blodgett is fabricated by assembling black lipid/block
(LB) and Langmuir-Schaefer (LS) monolayer copolymer domains across multiple apertures in
transfer techniques, (2) vesicle rupture of the ethylene tetrauoroethylene (ETFE) scaffolds
108 Aquaporin-Incorporated Biomimetic Membranes
(about 300 mm diameter each aperture) (Hansen Characterization of Water Transport

et al. 2009; Gonzlez-Prez et al. 2009; Rein Through the Aquaporin-Incorporated
et al. 2011). This membrane array can be stabi- Membranes
lized with the support of a highly permeable
hydrogel layer (Lander et al. 2011; Ibragimova The permeability of reconstituted aquaporin in ves-
et al. 2012). icles is characterized mainly by the stopped-ow
The supported lipidic/polymeric membranes spectroscopy technique. In a stopped-ow experi-
with aquaporins have been designed for ment, the osmotic permeability Pf (m3/m2/s)
several applications, such as forward osmosis can be calculated by using the equation
(FO), nanoltration (NF), and reverse osmosis (Borgnia et al. 1999)
(RO). Detailed information on the materials and
preliminary performance is summarized in Pf k=Cosm V w So =V o
Table 1.
where k is the rate constant obtained from light
scattering signal and So(m2) and Vo(m3) are the
initial surface area and volume of the vesicle,
respectively. Vw is the molar volume of water
(18 106 m3/mol) and Cosm is the osmolarity
difference that drives the vesicle shrinkage.
The hydraulic permeability coefcient Lp (m/s/
bar) represents the rate of water movement across
a given area, per unit of osmotic gradient, and can
be calculated from the equation (Hillyard 2012):
Lp Pf V w =RT
where R and T are gas constant and temperature,

respectively. When the biomimetic membrane
transforms from a sphere to a at surface, the
intrinsic water volume ux Jv (m3/s) can be
obtained by using equation (Hillyard 2012):
J v ALp slRTDC DP
Aquaporin-Incorporated Biomimetic Membranes,
Fig. 1 Schematic graph of free-standing lipidic/polymeric where A is the membrane area, in m2. s is the
membranes reection coefcient and is assumed to be 1. l is
Aquaporin-Incorporated
Biomimetic Membranes,
Fig. 2 Schematic graph of
supported lipidic/polymeric
membranes incorporated
with aquaporins
Aquaporin-Incorporated Biomimetic Membranes 109
Aquaporin-Incorporated Biomimetic Membranes, Table 1 Demonstration of aquaporin-incorporated biomimetic

membranes
Material of active layer and support Membrane performance Reference
A
PMOXA-PDMS-PMOXAa (with acrylate ends), aquaporin FO Wang et al. (2012)
Z (molar ratio 1:50) Draw solution: 2M NaCl
Methacrylate-functionalized PCTE support (average pore size Feed solution: water
50 nm) Water ux: >100 L/m2/h
Salt reverse ux: <10 g/m2/h
PMOXA-PDMS-PMOXAa (with acrylate ends), aquaporin NF Zhong et al. (2012)
Z (molar ratio 1:50) Feed: 200 ppm NaCl
Methacrylate-functionalized CA support (average pore size Pressure: 5 bar
19 nm) PWPb: 34 L/m2/h/bar
Salt rejection: 32 %
PMOXA-PDMS-PMOXAa (with disulde ends), aquaporin NF Duong et al. (2012)
Z (molar ratio 1:100) Feed: 200 ppm NaCl
Gold-coated porous alumina support (average pore size 55 nm) Pressure: 5 bar
Water ux: 8 L/m2/h/bar
DOPCc and DOTAPd RO Li et al. (2012)
Aquaporin Z (molar ratio 1:200) Feed: 1 mM NaCl
Commercially available NF-270 support Pressure: 1 bar
Water ux: 3 L/m2/h/bar
a
Poly(2-methyloxazolineb-dimethylsiloxane-b-2-methyloxazoline)
b
Pure water permeability
c
1,2-dioleoyl-sn-glycero-3- phosphocholine
d
1,2- dioleoyl-3-trimethylammonium-propane (chloride salt)
the molecularity. DP is the hydrostatic pressure establishment of functional lipid bilayer arrays.
difference across the membrane, in bar. DC is J Micromech Microeng 19:025014
Hillyard SD (2012) Sensing meets separation: water trans-
concentration difference across the membrane, in port across biological membranes. In: Hlix-Nielsen
mol/m3. C (ed) Biomimetic membranes for sensor and separa-
tion applications. Springer, Netherlands, pp 24
Ibragimova S, Stibius K, Szewczykowski P, Perry M,
Bohr H, Helix-Nielsen C (2012) Hydrogels for in situ
References encapsulation of biomimetic membrane arrays. Polym
Adv Technol 23(2):182189
Agre P (2006) The Aquaporin water channels. Proc Am Kaufman Y, Freger V (2011) Supported biomimetic mem-
Thorac Soc 3:513 branes for pressure-driven water purication. On biomi-
Borgnia MJ, Kozono D, Calamita G, Maloney PC, Agre metics. InTech. http://www.intechopen.com/books/on-
P (1999) Functional reconstitution and characterization biomimetics/supported-biomimetic-membranes-for-press-
of AqpZ, the E. coli water channel protein. J Mol Biol ure-driven-water-purication
291:11691179 Kaufman Y, Berman A, Freger V (2010) Supported lipid
Duong PHH, Chung TS, Jeyaseelan K, Armugam A, bilayer membranes for water purication by reverse
Chen Z, Yang J, Hong M (2012) Planar biomimetic osmosis. Langmuir 26:73887395
aquaporin-incorporated triblock copolymer mem- Kumar M, Grzelakowski M, Zilles J, Clark M, Meier
branes on porous alumina supports for nanoltration. W (2007) Highly permeable polymeric membranes
J Membr Sci 409410:3443 based on the incorporation of the functional water
Gonzlez-Prez A, Stibius KB, Vissing T, Nielsen CH, channel protein Aquaporin Z. Proc Natl Acad Sci U S
Mouritsen OG (2009) Biomimetic triblock copolymer A 104:2071920724
membrane arrays: a stable template for functional Kumar M, Payne MM, Poust SK, Zilles JL (2012)
membrane proteins. Langmuir 25(18):1044710450 Polymer-based biomimetic membranes for desalination.
Hansen JS, Perry M, Vogel J, Vissing T, Hansen CR, In: Hlix-Nielsen C (ed) Biomimetic membranes for
Geschke O, Emnus J, Hlix-Nielsen C (2009) Devel- sensor and separation applications. Springer, Nether-
opment of an automation technique for the lands, pp 4362
110 Aquaporins (AQPs) or Water Channels
Lander MR, Ibragimova S, Rein C, Vogel J, Stibius KB, intracellular pH. As with any membrane transport
Geschke O, Perry M, Helix-Nielsen C (2011) Biomi- facilitator, aquaporins have evolved to be highly
metic membrane arrays on cast hydrogel supports.
Langmuir 27:70027007 selective for their transported substrate without
Li X, Wang R, Tang C, Vararattanavech A, Zhao Y, binding water so strongly that transport is
Torres J, Fane T (2012) Preparation of supported lipid inhibited. On the basis of their selectivity,
membranes for Aquaporin Z incorporation. Colloids aquaporins can be divided into two groups: the
Surf B: Biointerfaces 94:333340
Rein C, Pszon-Bartosz K, Stibius K, Bjrnholm T, Helix- ordinary aquaporins, permeable to water only, and
Nielsen C (2011) Free-standing biomimetic polymer aquaglyceroporins which also permit transport of
membrane imaged with atomic force microscopy. small solutes such as glycerol and urea. A number
Langmuir 27(2):499503 of other compounds have also been reported to be
Sun G, Zhou H, Li Y, Jeyaseelan K, Armugam A, Chung
TS (2012) A novel method of Aquaporin transported through aquaporins, including CO2,
Z incorporation via binary-lipid Langmuir monolayers. NH3, and arsenite (Kreida and Tornroth-
Colloids Surface B 89:283288 Horseeld 2015). AQPs are transmembrane chan-
Wang H, Chung TS, Tong YW, Meier W, Chen Z, Hong M, nels; thus the ability of a molecule to cross an
Jeyaseelan K, Armugam A (2011) Preparation and
characterization of pore-suspending biomimetic mem- AQP channel depends on its own characteristics
branes embedded with Aquaporin Z on carboxylated (size, polarity, charge) and on the features of the
polyethylene glycol polymer cushion. Soft Matter AQP involved (Di Giorgio et al. 2014). They have
7:72747280 a similar basic structure; AQPs are monomers of
Wang H, Chung TS, Tong YW, Jeyaseelan K, Armugam A,
Chen Z, Hong M, Meier W (2012) Highly permeable about 30 kDa and, in general, contain six
and selective pore-spanning biomimetic membrane membrane-spanning helical segments and two
embedded with Aquaporin Z. Small 8:11851190 shorter helical segments that do not span the entire
Zhong PS, Chung TS, Jeyaseelan K, Armugam A (2012) membrane. The AQPs generally form stable tet-
Aquaporin-embedded biomimetic membranes for
nanoltration. J Membr Sci 407408:2733 ramers in membranes, although each monomer
contains a separate water pore. High-resolution
structural data show that the membrane-spanning
helical domains surround cytoplasmic and extra-
cellular vestibules that are connected by a narrow
Aquaporins (AQPs) or Water aqueous pore. Structural data and molecular
Channels dynamics simulations suggest that water mole-
cules move through this narrow aqueous pore
Fabio Bazzarelli and Lidietta Giorno and that steric and electrostatic factors are respon-
Institute on Membrane Technology, National sible for the water selectivity of AQPs. The pore is
Research Council of Italy, ITM-CNR, Rende, less constricted in the aquaglyceroporins than in
Italy the water-selective AQPs (diameter of 3.4 ver-
sus 2.8 , respectively) and is lined by more
hydrophobic residues (Papadopoulos and
Aquaporins (AQPs) or water channels are a Verkman 2013). An important property of
family of integral membrane proteins that form aquaporin-mediated water transport is its ability
hydrophilic pores in the cellular membrane. They to be regulated in response to cellular or environ-
are involved in the water transport through the mental signals. This is achieved by controlling
membrane. water transport at the individual protein level
All cells depend on their ability to maintain through a conformational change, so-called gat-
water homeostasis. This is achieved through the ing, or by altering the aquaporin density of a
action of aquaporins, membrane-bound water particular membrane. These proteins are present
channels that facilitate water ow across cellular in all kingdoms of life, demonstrating their central
membrane along osmotic gradients, while exclud- role in maintaining normal physiology of all
ing the passage of ions and protons. It is required organisms. The rst member of this family,
for maintenance of the membrane potential and AQP1, was identied in erythrocytes in 1991.
Aromatic and Nonaromatic Separation by Membrane Operations 111
This discovery led to homology cloning of hun- of aquaporin function by integrating evolutionary and
dreds of AQPs homologues from throughout the functional analyses. J Membr Biol 247:107125
Kreida S, Tornroth-Horseeld S (2015) Structural insights
animal and plant kingdoms, as well as from lower into aquaporin selectivity and regulation. Curr Opin A
organisms. Struct Biol 33:126134
In humans there have been identied Papadopoulos MC, Verkman AS (2013) Aquaporin water
13 aquaporins (AQP012) with specic organ, channels in the nervous system. Nat Rev Neurosci
14:265277
tissue, and cellular localization. Thus, different
members of the AQPs family are expected to
function in virtually all physiological processes
that involve water transport across the membrane.
The AQPs are expressed in many cell types
involved in uid transport, including epithelia Aromatic and Nonaromatic
and endothelia in the kidney, lung, exocrine Separation by Membrane Operations
glands, eyes, and gastrointestinal tract. However,
aquaporins are also expressed in cells that do not Lan Ying Jiang
have an obvious role in uid transport, such as School of Metallurgy and Environment, Central
erythrocytes and some leukocytes, adipocytes, South University, Changsha, China
and muscle. In addition, these are also expressed
in astrocytes throughout the central nervous sys-
tem and in supportive cells. Aquaporins have been Synonyms
linked to a number of pathological conditions,
including brain edema, renal disease, obesity, Dearomatization by membrane operations
and cancer, raising their attractiveness as drug
targets. Efcient separation of aromatics/nonaromatics
Given the key role played by aquaporins in the (generally referred to as dearomatization) is of
kidney, for recovering water permeated together great interest and signicance to the chemical
with other ions and molecules through the rst and petrochemical industries (Kung et al. 2010).
part of the glomerulus, they have been investi- Aromatic hydrocarbons such as toluene, benzene,
gated for developing biohybrid membranes able ethylbenzene, and xylenes are important feed-
to desalinate seawater. The biohybrid membranes stock in petrochemical industry; therefore, puri-
containing aquaporins showed very high water cation of aromatics is desired after they are
selectivity and permeability. Technological chal- produced through catalytic reforming, cracking,
lenges for productive application include or other processes (Sittig 1976). Nonaromatics
biohybrid membrane preparation on a large also have a wide range of applications (e.g., sol-
scale, aquaporin stability under operating condi- vents, fuels). They are generally derived from
tions or during membrane module cleaning, and aromatics/nonaromatics mixed feed, which neces-
maintenance operation. Development of synthetic sitates the separation of aromatic and nonaromatic
water channels mimicking aquaporins is a new hydrocarbons. For example, removal of aromatic
strategy under investigation. Highly selective hydrocarbons is needed to enhance the thermal
and permeable water channels are very interesting and oxidation stability, viscosity, and color of
for the development of seawater desalination plant lube oils. Aromatics such as benzene is carcino-
operating with very low energy input. genic; it is regulated by law that the levels of
aromatics in petrol and solvents must be con-
trolled (Singh et al. 2003). Removal of aromatics
References by-products will facilitate the improvement in the
production efciency of some petrochemical pro-
Di Giorgio J, Soto G, Alleva K, Jozefkowicz C, cesses, such as ethylene production through steam
Amodeo G, Muschietti JP, Ayub ND (2014) Prediction cracking (Corma et al. 2005).
112 Aromatic and Nonaromatic Separation by Membrane Operations
Presently, great interest from academia and content, while total ux is correspondingly
industry is given to studying the separation of aro- reduced. This is because PBI has stronger polarity,
matic/nonaromatic by membrane process, which is tighter and more rigid structure, as well as
pervaporation specically due to the dense structure enhanced anti-swelling property.
of its membrane capable of discriminating A pervaporation process alone might not be an
molecular-level difference as encountered in hydro- economical choice for aromatic/nonaromatic sep-
carbon mixtures. The separation in this process is aration. In the comparison among a conventional
based on the differential sorption and diffusion of extractive distillation process alone, a
permeating components inside membrane matrix, pervaporation system alone, and a hybrid system
coupled with the intrinsic volatility difference of consisting of the two processes in benzene/cyclo-
permeating components. It is the involvement of hexane separation, it is found that a hybrid system
selective membrane that makes pervaporation pro- is more economical than either of the two single
cess independent of the vaporliquid equilibrium processes (Rautenbach and Albrecht 1985). The
and therefore more energy efcient and environ- feed solution used for the evaluation is 1:1
ment friendly than classical distillation, rectica- (mol/mol) benzene/cyclohexane, and the target is
tion, extraction, and crystallization. to separate this mixture into two streams: 99.5 mol%
Numerous polymeric materials and some inor- benzene and 99.2 mol% cyclohexane. In both
ganic materials as well as their derivatives by extractive distillation process and hybrid process,
chemical (e.g., grafting, cross-linking) or physical furfural was used to break the benzene/cyclohex-
modications (e.g., polymer blending, inorganic ane azeotrope in the rst stage. The product from
lling) have been applied in preparing this stage is a cyclohexane-rich stream and a mix-
pervaporation membrane for aromatics/ ture of benzene/furfural. The benzene impurity in
nonaromatic separation (Smith et al. 2004). Aro- the cyclohexane-enriched stream is then removed
matics have double-bond-related p-electrons, by pervaporation, and the benzene/furfural mix-
hence having afnity toward membrane matrix ture is separated by distillation to get furfural and
containing polar functional groups. Size selection highly puried benzene. The advantage of hybrid
also favors the aromatics in the separation of some system is more striking when higher purity is
aromatic/nonaromatic pairs. A typical pair is ben- demanded of the products.
zene/cyclohexane wherein the former has smaller
molar volume. All of the organic and inorganic
materials utilized show preferential transport References
toward aromatics. Carbon-modied nylon
6 exhibited an extremely high separation factor Corma A, Melo FV, Sauvanaud L, Ortega F (2005) Light
cracked naphtha processing: controlling chemistry for
of more than 400. Commercialized polyvinyl
maximum propylene production. Catal Today
alcohol composite membrane also presents rec- 107108:699706
ommendable separation factor and ux. As an Kung G, Jiang LY, Wang Y, Chung TS (2010) Asymmetric
effort toward practical application, integrally hollow bers by polyimide and polybenzimidazole
blends for toluene/iso-octane separation. J Membr Sci
skinned asymmetric polymeric membranes for
360:303314
toluene/iso-octane separation were fabricated Rautenbach R, Albrecht R (1985) The separation potential
from blends of commercially available polybenzi- of pervaporation Part 2. Process design and economics.
midazole (PBI) and polyimide (PI) Matrimid J Membr Sci 25:2554
Singh AP, Mukherji S, Tewari AK, Kalsi WR, Sarpal AS
using dry-jet wet spinning followed by (2003) Determination of benzene and total aromatics in
non-solvent-induced phase inversion (Kung commercial gasolines using packed column GC and
et al. 2010). The best performance achieved is a NMR techniques. Fuel 82:2333
separation factor of 200 and a ux of 1.35 kg/m2-h Sittig M (1976) Aromatic hydrocarbons: manufacture and
technology. Noyes Data Corp, Ridge Park
in separating toluene/iso-octane (50/50 wt%).
Smith B, Suhany D, Sridhar S, Ramakrish M (2004) Sepa-
Separation factor for the selective transport of ration of organicorganic mixtures by pervaporationa
toluene increases with higher polybenzimidazole review. J Membr Sci 241:121
Artificial Blood Cell 113
They have a long shelf life. Unlike donated

Artificial Biological Membrane blood that breaks down after 1 month, articial
blood can be stored for over a year or more and A
Bioarticial Synthetic Membrane it can be stocked and easily shipped where
needed (Suman 2008; Cohn 2000; Nucci and
Abuchowski 1998; Lesley 2001).
Artificial Blood Cell A brief history. Though popular folklore sug-

gests that Maya Empire was trying to nd a blood
Antonietta Messina and Loredana De Bartolo substitute, it is only after William Harvey discov-
Institute on Membrane Technology, National ered and described the blood pathways and circu-
Research Council of Italy, ITM-CNR, Rende, lation throughout the body, in 1616, that medical
Italy practitioners tried numerous substances such as
milk, beer, plant resins, urine, and animal blood
as a substitute for blood. Therefore, the rst suc-
The expression articial blood refers to articial cessful human blood transfusion was done in
substitutes or articial surrogates purposely 1667. Only in 1868, researchers found that solu-
designed in laboratory in order to mimic and fulll tions containing hemoglobin isolated from red
two specic functions of the biological blood: the blood cells could be used as blood replacements.
transport of oxygen and carbon dioxide through- Therefore, the most important turning point in the
out the body and the maintenance of the hemosta- development of articial blood was made in 1883
sis inducing the blood clotting. Actually, the and 1890s with the development of the Ringer's
blood is a special type of connective tissue com- solution and the discovery by Karl Landsteiner of
posed of four different components in charge of a the human blood groups (A, B, AB, and O)
variety of functions. The white cells, responsible according to the capacity of their red cells to
for the immune defense; the red cells, responsible express on their membrane surface specic anti-
for the transportation of oxygen and carbon gens (Suman 2008; Nucci and Abuchowski
dioxide throughout the body; the platelets, 1998).
involved in the hemostasis process; and the As already said, there are two different prod-
plasma, an extracellular component made up of ucts under development as blood substitutes.
water, salts, and various proteins that, along with They differ primarily in their functions: they
platelets, encourages blood to clot. Actually, due carry oxygen or induce the blood clotting.
to this complex composition, it is not possible to Red blood cell substitutes. They are classied
reproduce the blood articially and only two of its in hemoglobin-based and peruorocarbon (PFC)-
components are laboratory designed: the red based substitutes. Hemoglobin-based products.
blood cells and the platelet substitutes. The arti- Derived from different sources (animal, human,
cial blood products have advantageous charac- recombinant) all hemoglobin products carry oxy-
teristics nowadays: gen by taking advantage of the natural hemoglo-
bin function. The natural hemoglobin products are
They are safe and biocompatible and lacking of obtained from human or animal blood. The human
the membrane antigens; therefore they are substitutes, in particular, are derived from blood
usable for all patient blood types. and red cells collected during plasma donations.
They are able to transport oxygen throughout The synthetic products, on the contrary, are the
the body and release it where needed. results of amino acids specic combinations or the
They possess a great level of safety. They can outcome of the recombinant biochemical technol-
be processed by ltration and pasteurization to ogy, which allow to derive hemoglobin harvested
remove all disease-causing agents such as from an E. coli bacteria strain. Independently on
viruses and microorganisms. the sources, the protein breaks down easily into
114 Artificial Blood Cell
smaller, toxic compounds. Then once obtained, it Platelet substitutes, they represent an essential
must undergo purication and partial modica- component of the thrombogenic response to
tion in order to decrease (i) the rapid formation bleeding through the adhesion to the damaged
of dimers and monomers that freely diffuse into vessel lining and releasing several coagulation
renal tubules causing renal failure and (ii) its factors. Platelet-Derived Products. Frozen plate-
strong afnity for factors (i.e., NO) causing vaso- lets and cold liquid-stored platelets have been
constriction, hypertension, and bradycardia. tested since 1950s in order to develop a treatment
Among the strategies employed to stabilize hemo- for thrombocytopenia. Despite being the only
globin the main important are the chemically alternative to fresh platelets in clinical use, disad-
cross-linking molecule processes and the encap- vantages as expensiveness with respect to the
sulation technology. Polyethylene Glycol- fresh platelets, morphological defects, and lower
Modied, Liposome-Encapsulated Hemoglobin, functionality directed the research toward new
and Polymersome-encapsulated hemoglobin discoveries. The Lyophilized platelet products
have been produced and commercialized (Enzon were developed since 1950s and are currently
and APEX Bioscience) for the treatment of cancer under animal trials. They are obtained by xing
and stroke patients and for treatment of hypoten- human platelets in paraformaldehyde prior to
sion induced by septic shock respectively (Chang freeze-drying them in an albumin solution with
1998). Moreover, hemoglobin products cova- no inuences on the adhesive properties of the
lently cross-linked with small bridges of sugar cells. The Platelet microparticles are
molecules showed not to induce vasoconstriction microvesicles formed spontaneously during the
in phase III of clinical trials (Gould and Moss storage. Deriving from the platelet membranes,
1996; Danielson et al. 1997). they showed similar hemostatic properties as
Peruorocarbon (PFC) products. PFC-based intact platelets. Cypress Bioscience developed
solutions are synthetic hydrocarbons with halide Infusive Platelet Membranes (IPM; Cyplex,
substitutions able to dissolve up to 50 times more Cypress Bioscience) using outdated human plate-
oxygen than plasma. These biologically inert let that are purposely fragmented, virally
materials allow to avoid the spreading of infec- inactivated, and lyophilized (Lesley 2001). With
tious disease via a blood transfusion. Neverthe- a shelf life of over 2 years these products are in
less, they are not soluble in water solution and phase II of clinical trials and appeared to be able to
require to be emulsied in oils or lipids before stop the bleeding in the 60 % of the patients
being injected into patients blood. Despite that a resulted refractory to platelet transfusion due to
special product, developed by the PFC therapeu- the immune response to HLA or platelet antigens,
tics, is currently in phase III of clinical trials for with no adverse effects or risks of thrombosis. In
use in cardiac and general surgical patients, cur- order to overcome the disadvantages of the plate-
rently these substitutes showed to have a short let processing, storage, and shelf life Platelet Sub-
duration of action, about 24 h, to be very expen- stitutes have been proposed. The Red cells with
sive, and to cause some adverse reactions associ- surface-bound brinogen or RGD peptides have
ated with their long-term usage. Their prolonged been subjected to numerous studies. Fibrinogen is
administration can cause several adverse reactions a free molecule able to bind activated platelets
such as the increase of the blood pressure and of together through the multiple Arginine-Glycine-
the pancreatic proteins, the gastrointestinal Aspartine (RGD) sequences. Bounding to the sur-
dysmobility, thrombocytopenia, u-like syn- face of autologous erythrocytes (red blood cells),
drome, and a quite important oxygen toxicity brinogen, or the RGD fragments, the
(Suman 2008; Cohn 2000; Lesley 2001). thromboerythrocytes obtained showed to be
Platelet substitutes. These articial products hemostatically effective in some animal
are very different from those already described. models of thrombocytopenia. Other important
Classied in Platelet-Derived Products and products are the Fibrinogen-coated albumin
Artificial Brain Model: Biohybrid Membrane System 115
microcapsules/microspheres. These are coated

with brinogen of the human protein albumin. Artificial Brain Model: Biohybrid
Two preparations have been evaluated in preclin- Membrane System A
ical trials: SynthocytesTM (Andaris Group Ltd,
Nottingham, UK) and ThrombospheresTM Sabrina Morelli
(Hemosphere, Irvine, Calif). The latter, due to Institute on Membrane Technology, National
their smaller mean diameter, showed in in vivo Research Council of Italy, ITM-CNR, Rende,
to have a long duration of action. Last, two Italy
Liposome-based agents have been studied: the
plateletsomes, which are lipid vesicles with plate-
let glycoproteins on their surface, and the factor Articial brain, in membrane operation, is a
Xa with phospholipid vesicles. Both demon- biohybrid membrane system which provides suit-
strated to be hemostatically effective in vitro and able conditions for the creation of a biomimetic
in some animal models, but problems of high microenvironment for neuronal tissue engineering
toxicity must be overcome in order to pass the resembling in vivo neuronal tissue properties and
clinical trials. functions.
Despite that numerous new discoveries and In the last decade, neuronal cell behavior on
products showed to be useful as blood substitutes, biomaterial such as membrane has become of
the current articial blood technology is limited great interest since it offers the advantage of
by the short-term shelf life, the costs, and the developing neuronal tissue that may be used for
toxicity problems clinically proved. Then new the in vitro simulation of human brain function
materials able to carry oxygen in the body and substitute. This could provide further insights not
to induce the blood clot must be found and only into the cell but also in developing therapies
longer-lasting products should be developed in neurodegenerative disorders such as
(Suman 2008; Lesley 2001; Lee and Blajchman Parkinsons or Alzheimers diseases. A biohybrid
1998). system using neurons could also represent a
useful instrument for predictive drug testing or
constitute a future model of a bioneuronal
References network device. To accomplish this task the
articial membrane must exhibit a specic micro-
Chang TMS (1998) Modied hemoglobin blood substi- architecture able to stimulate neuronal regenera-
tutes: present status and future perspectives. Biotechnol tion, therefore supporting axonal regrowth and
Annu Rev 4:75112
affecting cellular orientation and differentiation in
Cohn SM (2000) Blood substitutes in surgery. Surgery
127:599602 a positive way.
Danielson C, Ryder K, Glick M, Limiac A (1997) This biohybrid articial system consisting of
Hemoglobinbased blood substitutes in the blood bank. neurons and membranes offers a valuable tool for
Lab Med 28(2):131134
studying neuronal cell behavior and fundamental
Gould SA, Moss GS (1996) Clinical development of
human polymerized hemoglobin as a blood substitute. questions in neuroscientic research, and it also
World J Surg 20:12001207 could be used in a wide range of biotechnological
Lee DH, Blajchman MA (1998) Novel platelet applications.
products and substitutes. Transfus Med Rev
Several works in the literature show evidence
12:175187
Lesley K (2001) Articial blood: an update on current red that semipermeable polymeric membranes in at
cell and platelet substitutes. Proc (Bayl Univ Med Cent) and hollow ber congurations, thanks to their
14(2):158161 highly selective structural, physicochemical, and
Nucci ML, Abuchowski A (1998) The search for blood
transport properties, allow the successful in vitro
substitutes. Sci Am 278(2):7277
Suman S (2008) Articial blood. J Crit Care Med reconstruction of neuronal tissue reproducing a
12(3):140144 tissue model for studying metabolic diseases and
116 Artificial Brain Model: Biohybrid Membrane System
drug effects (De Bartolo et al. 2008; Zhang The importance of FC membrane culture sub-
et al. 2005 Morelli et al. 2014; Piscioneri et al. strates toward the highlighting of the effects of
2015). highly preferred GABAAR a2,5 agonists and
For typical neuronal tissue-engineered con- inverse agonists on the successful elongation of
structs, the properties of the material components neuronal processes and transcriptional activities
relevant to the tissue response include the surface of hamster hippocampal Gluergic cells was dem-
microgeometry, the MWCO, the electrical prop- onstrated for the rst time (Giusi et al. 2009).
erties, possibly the bioactive factors that are A biohybrid system composed of neuronal
loaded and sustainedly released through the chan- cells and silicon-supported nanoporous mem-
nel wall, and the degradation rate for biodegrad- branes has been designed to facilitate control of
able materials. The wettability of the membrane the biochemical environment of neuronal net-
surface has proved to be a factor that affects the works with cellular resolution. Nanoporous mem-
growth of neurons modulating the adsorption of branes may be used to interface with biological
adhesion proteins contained in the culture materials in a biohybrid system, for example, as
medium or secreted by cells that mediate the an articial chemical synapse interface (Wolfrum
adhesion of cells to the substrate. et al. 2006).
Advances in polymer chemistry have facili- Compared to traditional two-dimensional
tated the engineering of synthetic membranes (2D) culture systems, 3D in vitro cultures of neu-
that can be specically manipulated with regard ronal cells appear to better mimic in vivo neuronal
to their physical and mechanical characteristics, microenvironments representing a great potential
which may affect the interactions with cells. for building 3D neuronal circuits as well as for
The inuence of topographical features on creating tissue-based biosensors and for promot-
neuronal cell adhesion and differentiation has ing systemic restoration of severe nerve injuries.
been studied by using patterned adhesive areas In this context, several studies have been
that provide only a fraction of the surface for cell performed toward the determination of differ-
adhesion while the rest is cell repellent or by using ences between 2D and 3D environments for neu-
contact guidance cues in combination with also ronal function in order to provide great benet to
nerve growth factors or an electric eld. Topo- neuroengineering efforts. Recently, a complete
graphic guidance of neuritis outgrowth has been 3D neural tissue-like structure has been assembled
explored in vitro with a culture substrate in vitro using primary hippocampal cells and PAN
containing etches, microchannels, nanotubes, or HF membranes. As a consequence, this advanced
microgrooves. Micro-patterned biodegradable model system could allow the mimicking of hip-
poly(L-lactic acid) membranes directly improved pocampal neuronal events in order to study natu-
the guidance of neurite extension and thereby ral neurobiological properties of these cells, such
enhanced their orientation with a consequently as memory and learning processes, or the induc-
highly ordered neuronal cell matrix, which may tion of disease pathologies, allowing testing of
have strong bearings on the elucidation of regen- new therapies on the resolution of these states
eration mechanisms (Morelli et al. 2010). (Morelli et al. 2012).
Hippocampal neurons exhibited a different In the last decade, with rapid advances in bio-
morphology in response to varying the properties material technology, several types of natural and
of the membrane surface. Indeed, cells grown on synthetic biodegradable polymer, including colla-
the smoother membranes such as uorocarbon gen, chitosan, poly(glycolic acid), poly(L-lactic
(FC) and PES membranes displayed a large num- acid), poly(L-lactide-co-glycolide), poly
ber of neuritis with consequent formation of bun- (caprolactone), and polyurethane, have been
dles. The density of axonal network increases the reported as suitable for nerve regeneration,
neurites to become more elaborate and highly although the ideal physicochemical compositions,
branched on the smoother surfaces. surface structure, and functionalization of such
Artificial Kidney 117
materials have not yet been found. It was recently

reported that polycaprolactone-based membranes Artificial Kidney
successfully supported outgrowth and differentia- A
tion of human neuronal cells (Morelli et al. 2015). Sabrina Morelli
References Italy
De Bartolo L, Rende M, Morelli S, Giusi G, Salerno S,

Piscioneri A, Gordano A, Di Vito A, Canonaco M,
Drioli E (2008) Inuence of membrane surface proper-
Articial kidney, a synonym for hemodialyzer, is
ties on the growth of neuronal cells isolated from hip- an extracorporeal device through which blood
pocampus. J Membr Sci 325:139149 may be circulated for removal of elements that
Giusi G, Facciolo RM, Rende M, Alo R, Di Vito A, normally are excreted in the urine.
Salerno S, Morelli S, De Bartolo L, Drioli E, Canonaco
The articial kidney is used to rid the blood of
M (2009) Distinct a subunits of the GABAA receptor
are responsible for early hippocampal silent neuron- metabolic products, to correct electrolyte-water
related activities. Hippocampus 19:11031114 and acid-alkaline balances in acute and chronic
Morelli S, Salerno S, Piscioneri A, Papenburg BJ, Di renal insufciency, to remove dialyzing toxic sub-
Vito A, Giusi G, Canonaco M, Stamatialis D,
stances in cases of poisoning, and to remove
Drioli E, De Bartolo L (2010) Inuence of micro-
patterned PLLA membranes on outgrowth and orienta- excess water in cases of edema. In 1913 the Amer-
tion of hippocampal neurites. Biomaterials ican scientist Abel created an apparatus for dialy-
31:70007011 sis that was the basis for the articial kidney;
Morelli S, Piscioneri A, Salerno S, Rende M, Campana C,
Tasselli F, Di Vito A, Giusi G, Canonaco M, Drioli E,
Dr. Willem Kolff, a Dutch physician, constructed
De Bartolo L (2012) Flat and tubular membrane sys- the rst functioning articial kidney. He treated
tems for the reconstruction of hippocampal neuronal his rst patient in 1943, and in 1945 he was rst
network. J Tissue Eng Regen Med 6:299313 able to save a patients life with hemodialysis
Morelli S, Piscioneri A, Messina A, Salerno S, Al-Fageeh
treatment.
MB, Drioli E, De Bartolo L (2015) Neuronal growth
and differentiation on biodegradable membranes. The articial kidney operates on the princi-
J Tissue Eng Regen Med. 9:106117 ple of the dialysis of substances through a
Morelli S, Piscioneri A, Salerno S, Al-Fageeh M, Drioli E, semipermeable membrane. Dialysis is a result
De Bartolo L (2014) Neuroprotective Effect of
of the differences in the concentrations of sub-
Didymin on H2O2-Induced Injury in Neuronal Mem-
brane System. Cells Tissues Organs 199:184200 stances in the blood and in the dialyzing solu-
Piscioneri A, Morelli S, Mele M, Canonaco M, Bilotta E, tion, which contains glucose and the principal
Pantano P, Drioli E, De Bartolo L (2015) electrolytes of the blood in nearly physiological
Neuroprotective effect of human mesenchymal stem
concentrations without containing any of the
cells in a compartmentalized neuronal membrane sys-
tem. Acta Biomaterialia 24:297308. substances that must be removed from the
Wolfrum B, Mourzina Y, Sommerhage F, Offenhausser body (urea, creatinine, uric acid, sulfates, phos-
A (2006) Suspended nanoporous membranes as inter- phates). Proteins, formed elements of the blood,
faces for neuronal biohybrid systems. Nano Lett
bacteria, and substances with a molecular
6:453457
Zhang N, Yan H, Wen X (2005) Tissue-engineering weight of more than 30,000, do not pass
approaches for axonal guidance. Brain Res Rev through the membrane. During hemodialysis
49:4864 the patients blood is drawn off through a cath-
eter by a pump from the inferior vena cava and
passed inside the chambers of membranes of
the dialyzer; these chambers are washed outside
Artificial Functional Organ by the dialyzing solution, which is supplied by
another pump. Partially puried, the blood is
Membrane Articial Organs returned to one of the surface veins.
118 Artificial Kidney
The membrane used during hemodialysis has reduced induction of inammatory mediators.
evolved signicantly since the earliest reports. While observational studies suggest that high-
Early membranes were natural materials, such as ux hemodialysis is associated with lower and
cellulose, or simple synthetic compositions, such dialysis-related amyloidosis, individual random-
as polysulfone, with symmetric structures and ized controlled trials have not conrmed that
good performance for small solute passage high-ux hemodialysis improves clinical out-
(Fleming 2011). comes. High-ux dialysis membranes have
However, these membranes include signicant larger pore sizes that may increase transportation
interactions with the complement pathway and of dialysate impurities or contaminants from the
immune response. Modifying the hydroxyl dialysate to the bloodstream even when the dial-
groups on cellulose membranes gives similar per- ysate quality complies with acceptable stan-
formance characteristics and less bioactivity and dards, although some evidence suggests that
includes cellulose diacetate and hemophan mem- high-ux membranes may actively absorb endo-
branes. With improvements in chemical engineer- toxins and impede their back transfer from dial-
ing and manufacturing, cellulose-based ysate into the bloodstream (Palmer and Strippoli
membranes have yielded to synthetic membranes. 2013).
The most popular membranes in use today are Although dialysis is currently an available
polyacrylonitrile and polysulfone; however, poly- treatment for chronic renal disease, it replaces
amide, polycarbonate, and polymethyl- only the ltration function of the kidneys, and
methacrylate are available. These materials have thus, it is associated with a number of complica-
improved biocompatibility and reduce comple- tions and is not a permanent solution. Kidney
ment activation; however, they may enhance pro- transplantation performed in a heterotopic loca-
tein adsorption to the membrane in the absence of tion is the only denitive solution for end-stage
further modication. Despite the improved com- renal failure. A potential solution to ease the
patibility of synthetic membranes, the most recent demand for donor kidneys may be the use of tissue
large review does not demonstrate benets con- engineering techniques that could be used to
veyed by biocompatibility. The ultrastructure of develop functional kidneys. A primary challenge
membranes is predominantly the capillary design, in renal tissue engineering, however, is the pro-
or hollow ber structure, in which blood ows duction of tissues with the functional complexity
through a series of small tubes held together in a required for kidney transplantation. The rst step
bundle. This allows for a low-resistance, high in achieving this aim could be to produce acellular
surface area membrane. Membrane properties kidney matrices for use as biological scaffolds that
such as charge, wall thickness, and pore size affect support cell growth and facilitate new tissue for-
function. The radius of the pore is related to the mation by potentiating cellcell interaction. Sev-
ultraltrate ow through the membrane, but over- eral groups have used different perfusion-based
all uid transfer is a function of the mean pore decellularization protocols to produce naturally
size. The number or density of pores and the derived kidney scaffolds (Arenas-Herrera
radius of the pore affect solute transfer in diffusive et al. 2013).
clearances. Membrane characteristics such as ux In the last decade, several studies have been
and permeability to water are determinants of pore performed on the development of the implantable
characteristics and wall thickness (Vigano articial kidneys, but these studies reported only
et al. 2008). developmental concepts, in vitro or in vivo studies
As dialysis membrane technology has using small animals. It is difcult to control cer-
improved, high-ux membranes with larger tain research aspects of an implantable articial
pore sizes have been developed to allow greater kidney to a preclinical level, and these aspects
clearances of larger retention molecules with include long-term maintenance of the
Artificial Liver, Membrane Operations 119
antithrombogenic properties of the systems and

each of its components, and the function of an Artificial Liver, Membrane
implanted, miniaturized system, whose function Operations A
must be maintained for a years duration. Further
progress in nanotechnology and biotechnology is Antonietta Messina and Loredana De Bartolo
required to realize a treatment with an implantable Institute on Membrane Technology, National
kidney (Saito et al. 2011). Research Council of Italy, ITM-CNR, Rende,
New directions in dialysis research include Italy
cheaper treatments, home-based therapies, and
simpler methods of blood purication. These
objectives may be probably obtained with inno- The expression of articial liver denes all the
vations in the eld of articial kidney through the supporting membrane systems able to substitute
utilization of new disciplines such as miniaturiza- or integrate the liver functions when it reaches
tion, microuidics, and nanotechnology. This the 30 % critical threshold of the normal hepatic
research may lead to a new era of dialysis in activity. The liver is the biggest glandular organs
which the new challenges are transportability, in the human body. Placed in the upper section of
wearability, and why not the possibility to develop the abdominal cavity, this dark red organ is well
implantable devices (Ronco et al. 2011). known to be able to self-regenerate from partial
injuries doubling its mass in 1 day. Therefore,
acute or chronic hepatic insufciencies caused
by viral pathology, drug abuse, intoxications, or
References pathological decompensation can result in severe
liver failure (LF). Though implying the inability
Arenas-Herrera JE, Ko IK, Atala A, Yoo JJ of the liver to completely carry out its metabolic,
(2013) Decellularization for whole organ bioengineer-
ing. Biomed Mater 8:014106. doi:10.1088/1748-6041/
excretory, and detoxifying functions, both the
8/1/014106 acute and chronic hepatic insufciency (ALF,
Fleming GM (2011) Renal replacement therapy review. ACLF) can potentially be reversible if one third
Past, present and future. Organogenesis 7:212 of the organ is still functioning; then liver activity
Palmer SC, Strippoli G (2013) High-ux versus low-ux
can return to good levels. However, the LF leads
haemodialysis membranes for end-stage kidney dis-
ease. Nephrology 18:313314 to an accumulation of water-soluble and albumin-
Ronco C, Davenport A, Gura V (2011) The future of the bounded toxins that cause encephalopathy and a
articial kidney: moving towards wearable and minia- multi-organ dysfunction. The solely long-term
turized devices. Nefrologia 31:916
therapy to stem a hepatic insufciency is the
Saito A, Sawada K, Fujimura S (2011) Present status and
future perspectives on the development of bioarticial orthotopic transplantation that ensures a notable
kidneys for the treatment of acute and chronic renal recovery of hepatocytes in number and function-
failure patients. Hemodial Int 15:183192 ality. Actually, due to the limited number of
Vigano SM, Di Filippo S, Manzoni C, Locatelli F (2008)
donors, only a narrow circle of patients in waiting
Membrane characteristics. In: Ronco C, Cruz DN (eds)
Hemodialysis-from basic research to clinical trials, list can be transplanted according to the severity
vol 161, Contribution of nephrology. Karger, Basel, of their disease. Thus, in order to improve
pp 162167 patients prognosis and survival, articial extra-
corporeal liver assist devices have been developed
either bridging the patient to transplantation and
temporarily supporting the failing organ until it is
Artificial Liver able to regenerate. Basically, three different types
of supportive therapies are nowadays available:
Bioarticial Liver bio-articial, articial, and hybrid liver support
120 Artificial Liver, Membrane Operations
systems. Therefore, between them, the articial articial support system for treating the LF. The
systems are the most used in clinic due to the cellular components of the blood are separated
lower costs and their easier handling. Articial from the plasma using a hollow ber lter
liver support systems are cell-free membrane (cellulose diacetate or polyethylene) able to
devices aimed to temporally replace native liver remove toxins and supply lacking components as
detoxication functions through processes of albumin or coagulation factors. However, this
adsorption and ltration in order to improve the method has a disadvantage: it requires a large
clinical state of the patient. The most of articial plasma stock with ensuing higher costs and
liver devices use membrane separation associated bears the risk of infections. Moreover, most of
with columns or suspensions of sorbents (e.g., the toxins that accumulate in the plasma of
charcoal, anion-exchange or cation-exchange patients with liver insufciency are protein
resins) that selectively remove toxins and/or bound; then these methods are not able to ade-
regenerate dialysate or plasma ltrate (Carpentier quately eliminate them. This is the reason why
et al. 2009; Naruse et al. 2007). Conventional new dialysis systems have been developed using
dialysis techniques, such as hemoltration, albumin as scavenger to clear the toxins involved
hemodialysis, or hemodialtration, have been in the physiopathology of the failing liver.
used for the removal of the low molecular weight Between the different albumin dialysis modalities,
water-soluble metabolites. The hemodialysis the single-pass albumin dialysis (SPAD) has
allows to remove water-soluble metabolites, to shown some positive results. The patients blood
re-equilibrate the acid-basic and the electrolytic passes a high-ux dialysis membrane, while an
values in the blood, and to reduce also the ammo- albumin solution streams counter-directionally to
nia concentrations. The hemoltration is carried the ltering blood membrane, accepting toxins
out by ltrating cellulose or polysulfone mem- from the plasma. Therefore, the system works at
branes with a MWCO of 50 KDa. These mem- a very high cost due to the high concentration of
brane devices remove toxic and natural albumin needed in the dialysate to perform the
substances through connective transport allowing plasma detoxication (Naruse et al. 2007;
to partially recover from the hepatic encephalop- Onodera et al. 2006). At present, there are only
athy. The hemoperfusion is performed through two articial extracorporeal liver support systems
extracorporeal devices constituted by active used in clinic: the Molecular Adsorbents
charcoal adsorbing elements covered by a cellu- Recirculating System (MARS1) from Gambro
lose or polysulfone hemocompatible lm. and Fractionated Plasma Separation and Adsorp-
Adsorbing a variety of water-soluble toxins in tion (FPSA), which was designed by Fresenius
the low and middle molecular weight range, the Medical Care (Bad Homburg, Germany) and
hemoperfusion has a signicant disadvantage; in commercialized as Prometheus1 (Carpentier et al.
fact it has been demonstrated that the resulting 2009; Evenepoel et al. 2006).
patients plasma can be highly impoverished of MARS is the most used articial liver support
important biological substances causing impor- system. Introduced in clinics in 1993, it consists of
tant decompensation in patients metabolism. In three components: (A) a conventional dialysis
order to improve the specicity of the detoxica- machine that guides the extracorporeal blood cir-
tion process, membrane reactors containing cuit (MARSFLUX), (B) a supplementary device
immobilized enzymes have been developed. for the albumin circuit made of two different
Hepatic enzymes are denitively linked onto hol- adsorption columns (MARS AC250, MARS
low ber membranes or enclosed in microcap- IE250), and (C) a dialysate circuit for low ux
sules and showed to be able to enhance the dialysis (Rademacher et al. 2011). How it works:
recovery of the patients mental state; hence, the the patients blood is pumped in the blood circuit
percentage of surviving patients does not increase, throughout a double-lumen central venous cathe-
and the clinical application is very limited. The ter. Here a rst column, the MARSFLUX system,
plasma exchange/plasmapheresis is another contains a polysulfone membrane with MWCO of
Artificial Liver, Membrane Operations 121
50 KDa that lters the blood and allows the pas- In the last decades, other dialysis albumin cir-
sage through the membrane of the water-soluble cuits have been developed for LF treatment, such
molecules and the free toxins, retaining the albu- as the continuous veno-venous hemodialysis A
min due to its higher dimension. In the meantime, (CVVHD) and the continuous veno-venous
the second circuit pushes in counter direction with hemodialtration (CVVHDF); therefore, though
respect to blood ux, a 20 % albumin solution that being able to decrease the load of water-soluble
induces the dissociation of the protein-bounded toxins, they cannot provide the complete plasma
toxins freeing the albumin of the plasma. The treatment. The last available articial liver is
dialysate albumin, now saturated, is then restored named selective plasma ltration therapy system
passing through two adsorption columns in the (SEPET). In this system, the patients blood
second circuit: the MARS AC250 lled with passes through a single-use cartridge constituted
active charcoal that retains high molecular weight by hollow bers with MWCO of 100 kDa
molecules (bilirubin) and the MARS IE250 lled allowing molecules with a molecular weight
with the anion-exchange resin cholestyramine that close to 100 kDa to pass the membrane in only
retains the negative-charged molecules. Albumin limited amounts, while albumin, HGF, and several
solution of the second circuit is then restored and clotting factors are retained. Main characteristic of
can be used for a new cycle of treatment. The this process is that a plasma fraction is completely
MARS is recommended for treating both acute discarded during the ltration procedure. This
and chronic liver failure, despite the survival of fraction contains accumulated small molecular
patients is only partially increased. Therefore, weight toxins and free pro-inammatory cyto-
many positive clinical results have been recorded kines; then, to replace the uid loss, a mixture
for the patients in the last years such as the composed of electrolyte solution, human albumin
increase of the arterial pressure values and the solution, and fresh frozen plasma is added to the
systemic vascular resistance index, the attenua- system and collected to the patients blood. This
tion of hyperdynamic circulation, and the signi- SEPET system is designed for use with any com-
cant decrease not only for bilirubin but also mercially available kidney dialysis unit and/or
for creatinine and urea (Evenepoel et al. 2006; plasma apheresis system utilizing hollow ber
Rozga 2006). cartridges (Carpentier et al. 2009; Rozga 2006).
The Prometheus articial system realizes the Many other attempts in developing articial liver
combined removal of the albumin-bounded and systems are conducted nowadays in order to
the water-soluble toxins by means of adsorption improve methods and clinical results. The ulti-
and fractionated plasma separation. It was devel- mate goal is then to develop a liver support device
oped by Falkenhagen in 1999, and the system is that will grant patients a substantial survival ben-
based on FPSA combined with hemodialysis and et compared to standard intensive care now
consists of two circuits (Falkenhagen et al. 1999). available.
In the rst one, the blood is collected through the
double-lumen catheter and allowed to ow through
a polysulfone membrane. Its MWCO of 250 KDa References
allows either the free toxins or the plasma albumin-
bounded toxins to pass through the membrane Carpentier B, Gautier A, Legallais C (2009) Articial and
reaching the second circuit. Here the plasma is bioarticial liver devices: present and future. Gut
58:16901702
puried by a direct adsorption process into the Evenepoel P, Laleman W, Wilmer A, Claes K, Kuypers D,
same two columns of MARS: the anion-exchange Bammens B, Nevens F, Vanrenterghem: Y (2006) Pro-
column and the neutral resin column. Then the metheus versus molecular adsorbents recirculating sys-
blood is recombined with the non-treated fraction tem: comparison of efciency in two different liver
detoxication devices. Artif Organs 30(4):276284,
and dialyzed completely in a high-ux device for Blackwell Publishing, Inc.#
the water-soluble molecules removal (Evenepoel Falkenhagen D, Strobl W, Vogt G, Schre A, Linsberger I,
et al. 2006; Rozga 2006). Gerner FJ, Schoenhofen M (1999) Fractionated plasma
122 Artificial Lung
separation and adsorption system: a novel system for Membrane oxygenators can be used in two
blood purication to remove albumin bound sub- principal modes: to imitate the function of the
stances. Artif Organs 23(1):8186
Naruse K, Tang W, Makuuchi M (2007) Articial and lungs in cardiopulmonary bypass (CPB) and to
bioarticial liver support: a review of perfusion treat- oxygenate blood in longer-term life support,
ment for hepatic failure patients. World J Gastroenterol termed extracorporeal membrane oxygenation
13(10):15161521 (ECMO). A membrane oxygenator consists of a
Onodera K, Sakata H, Yonekawa M, Kawamura A (2006)
Articial liver support at present and in the future. thin gas permeable membrane separating the
J Artif Organs 9:1728 blood and gas ows in the CPB circuit; oxygen
Rademacher S, Oppert M, Jrres A (2011) Articial extra- diffuses from the gas side into the blood, and
corporeal liver support therapy in patients with severe carbon dioxide diffuses from the blood into the
liver failure. Exp Rev Gastroenterol Hepatol
5(5):591599 gas for disposal.
Rozga J (2006) Liver support technology an update. Preliminary efforts to design, fabricate, and test
Xenotransplantation 13:380389 membrane articial lungs have been reported
since the 1950s. The early articial lungs used
relatively impermeable polyethylene or Teon
homogeneous membranes, and when more highly
Artificial Lung permeable silicone rubber membranes were intro-
duced in the 1960s (and as hollow bers in 1971),
Loredana De Bartolo and Sabrina Morelli the membrane oxygenator became commercially
Institute on Membrane Technology, National successful. The introduction of microporous hol-
Research Council of Italy, ITM-CNR, Rende, low bers with very low resistance to mass trans-
Italy fer revolutionized the design of membrane
modules, as the limiting factor to oxygenator per-
formance became the blood resistance (Martin
The articial lung is a medical device used to take and Zwischenberger 2013).
over or supplement the function of the lung which Current designs of oxygenator typically use an
is to oxygenate the blood and remove carbon extraluminal ow regime, where the blood ows
dioxide. outside the gas-lled hollow bers, for short-term
Such a device can replace entirely the pulmo- life support, while only the homogeneous mem-
nary gas exchange or assist the decient gas trans- branes are approved for long-term use.
fer capacity of the natural organ, either The key design considerations in blood oxy-
temporarily, with the hope that the healing process genators include minimizing the resistance to
will eventually repair the diseased organ, or per- blood ow, reducing the priming volume, ensur-
manently, when irreversible lung damage leaves ing easy debubbling at setup, and minimizing
the patient permanently disabled. blood activation and thrombogenicity.
Articial lung can be used as a bridge to lung The fabrication or wrapping of the ber bundle
transplantation, as a support device immediately in a blood oxygenator can be important as the
post-lung transplant, and as a rescue and/or sup- geometry obtained impacts diffusional boundary
plement to mechanical ventilation during the layers, secondary ows, and gas exchange
treatment of severe respiratory failure (Diaz- efciency.
Guzman et al. 2013; Javidfar and Bacchetta Looking ahead, current challenges in
2012). bioarticial lung engineering include creation of
Current articial lungs are also known as blood ideal scaffold materials, differentiation and
oxygenators and are modules composed of bun- expansion of lung-specic cell populations, and
dles of hollow ber membranes designed to bring full maturation of engineered constructs to pro-
blood and gas phases in intimate contact separated vide graft longevity after implantation in vivo
by only the thin walls of the hollow bers. (Sond and Ott 2012).
Asymmetric Ceramic Hollow Fiber Membranes 123
References
Diaz-Guzman E, Hoopes CW, Zwischenberger JB

(2013) The evolution of extracorporeal life support as A
a bridge to lung transplantation. ASAIO J 59:310
Javidfar J, Bacchetta M (2012) Bridge to lung transplanta-
tion with extracorporeal membrane oxygenation sup-
port. Curr Opin Organ Transplant 17:496502
Martin JT, Zwischenberger JB (2013) Articial lung and
novel devices for respiratory support. Semin Thorac
Cardiovasc Surg 25:7075
Song JJ, Ott HC (2012) Bioarticial lung engineering. Am
J Transplant 12:283288
Asymmetric Ceramic Hollow Fiber Membranes,

Fig. 1 An asymmetric aluminum oxide hollow ber
Assisted Transport micro-ltration membrane with an outer diameter of
1.8 mm
Facilitated Transport
extrusion, the most common method of fabrica-

tion involves the incorporation of phase inversion
Asymmetric Ceramic Hollow Fiber to create asymmetric structures during the extru-
Membranes sion process. In this method, a mixture of polymer
binder and ceramic material, suspended in a suit-
Benjamin F. K. Kingsbury able solvent, is extruded through a spinneret into
Imperial College London, London, UK contact with a uid that leads to precipitation of
the polymer phase. As the polymer binder pre-
cipitates, the polymer-ceramic suspension is
Asymmetric ceramic hollow ber membranes are converted into a polymer-ceramic mixed matrix
ceramic tubes with outer diameters typically rang- ceramic hollow ber precursor. The precursor is
ing from 0.5 to 2 mm, utilized for separation then heated to remove the polymer binder and
processes. To be classied as asymmetric, a hol- consolidate the ceramic material by sintering to
low ber must have a nonuniform cross-sectional yield the nal product. Very small hollow ber
structure consisting two or more different mor- diameters can be achieved using this fabrication
phologies. They may be formed from a wide technique with asymmetric cross-sectional struc-
range of ceramic materials depending on their tures consisting of conical micro-channels as well
purpose and may be either dense or porous as a porous ceramic matrix (Fig. 1). Moreover,
according to the type of separation process for multiple layers can be extruded simultaneously to
which they are designed. When combined form asymmetric cross-sectional structures which
together in a module, hollow ber geometry also contain layers of different materials.
gives rise to extremely high membrane surface Asymmetric ceramic hollow ber membranes
area per volume leading to compact membrane can be broadly categorized as either porous or
module designs. Asymmetric membranes offer dense. The former operate on a size exclusion
improved performance by combining a thin sepa- basis usually at low to moderate temperatures
ration layer with a support layer which has high and are commonly fabricated from inert ceramic
permeability. Although asymmetric structures can materials such as aluminum oxide. Asymmetric
be achieved by depositing additional layers on a structures such as that shown in Fig. 1 (Kingsbury
symmetric ceramic hollow ber formed by screw and Li 2009) are prepared in a single step and
124 Asymmetric Hollow Fiber Membranes
consist of a support layer formed from micro- membrane reactors and solid oxide fuel cells. In:
channels and an outer porous separation layer. Basile A, Nunes SP (eds) Advanced membrane science
and technology for sustainable energy and environmen-
The development of asymmetric morphologies tal applications, 1st edn. Woodhead, Cambridge,
such as that shown in Fig. 1 allows for ltration pp 496540
membranes with high ux to be fabricated in a
cost-effective way, with the additional benet of
integrated separation and support layers, leading
to a robust and durable structure. Dense asymmet-
ric ceramic hollow ber membranes operate at Asymmetric Hollow Fiber
higher temperatures with separation occurring by Membranes
lattice diffusion. Membranes fabricated from pro-
ton conducting, ionic conducting, and mixed ionic Lan Ying Jiang
and electronic conducting materials have a num- School of Metallurgy and Environment, Central
ber of potential applications such as oxygen sep- South University, Changsha, China
aration from air, solid oxide fuel cells, and the
partial oxidation of methane to synthesis gas.
For example, an asymmetric hollow ber oxygen The asymmetric membranes can be formed into
separation membrane can be formed from a different geometries such as the at sheet and the
perovskite material, by extruding a hollow ber hollow bers, by choosing the fabricating or cast-
in which a region of micro-channels supports a ing equipment. Hollow bers have annulus cross
thin outer oxygen separation layer (Zydorczak sections with the inner and outer circular skins of
et al. 2009). The micro-channels formed during the membranes being concentric. The outer diam-
the extrusion process provide mechanical strength eter of a hollow ber is less than 1 mm. Their
while at the same time greatly reducing the resis- advantages associated with practice usage include
tance to oxygen permeation compared to a sym- large membrane area per unit volume, self-
metric hollow ber structure. Asymmetric contained mechanical support, good exibility,
multilayer structures formed by co-extrusion and easy handling during the module fabrication
have been developed for solid oxide fuel cells and in the operation, etc. (Matsuura 1994, Peng
and membrane reactor applications in which et at. 2012). These facts have contributed to the
each layer has a different function. Examples lower fabrication cost and higher-performance
include the co-extrusion of an anode and electro- efciency of hollow ber membrane modules.
lyte as an asymmetric hollow ber in a single step By far, the most technically used hollow ber
(Othman et al. 2010), and the co-extrusion of a is produced by phase inversion: non-solvent-
dual-layer hollow ber membrane reactor for induced phase inversion (NIPS), (solvent)
partial oxidation of methane to synthesis gas evaporation-induced phase inversion (EIPS),
comprised of a dense oxygen separation mem- VIPS (vapor sorption-induced phase inversion),
brane on a porous catalytically active support and TIPS (thermally induced phase inversion)
(Wu et al. 2011). (Ulbricht 2006). In single-layer hollow ber
asymmetric membrane formation by NIPS, there
are two coagulations taking place in hollow ber
spinning (shell and lumen surfaces), while there is
References
only one major coagulation surface for an asym-
Kingsbury BFK, Wu Z, Li K (2011) Inorganic hollow ber metric at sheet membrane. The mass exchange
membranes for chemical reaction. In: Basile A, between internal coagulation liquid and lumen
Gallucci F (eds) Membranes for membrane reactors: surface for a hollow ber starts immediately
preparation, optimization and selection, 1st edn. Wiley,
Chichester, pp 117153
after liquid extrusion from the conduit of a spin-
Wu Z, Othman MHD, Kingsbury BFK, Li K (2011) Func- neret. For the contact of shell surface of nascent
tional ceramic hollow ber membranes for catalytic ber and the external coagulation bath, two
Asymmetric Hollow Fiber Membranes 125
different concepts are used. In some cases, there is 3. By adjusting the dope solution concentration,
an interval between nascent membrane extrusion hollow bers with ultrahigh pressure resistance
and immersing in the external coagulant. This can be produced A
contact at the external boundary could also syn- 4. The formation of composite membranes is
chronize with that at the internal boundary by more straightforward and cost effective
directly immersing the spinneret inside the exter- 5. Higher uxes are obtained by circumventing
nal coagulant. The former has a name of dry-jet pore penetration occurring in dip coating.
wet spinning. The mass transfer associated with
the dry step during the nascent bers residence Recently, an idea of immiscibility induced
time in the air gap is short under commercial phase separation (I2PS) was proposed (Ong and
spinning conditions but important in membrane Chung 2013). Phase separation of two immiscible
structure control. Solvent evaporation in the dry dopes renders the formation of a selective layer
step can cause polymer concentration at dope/air between the inner layer and outer layer. The
interface increase. This situation is favorable for degree of swelling of selective layer by feed liquid
forming a relatively compact skin and is often can be moderated as a result of its being shielded
practiced in fabricating membranes for gas sepa- from direct contact with the feed mixture by the
ration or pervaporation. One can easily create this outer layer.
situation by making the temperature of the spin
dope close to the boiling point of the volatile
solvent or non-solvent (Jiang et al. 2004). Rheo-
logical aspect is another choice in ne-tuning
References
hollow ber membrane morphological character-
istics to meet the requirement of separation tasks. He T, Mulder MHV, Strathmann H, Wessling M (2002)
Depending on the polymer intrinsic rheological Preparation of composite hollow ber membranes:
characteristics, air gap and take-up speed are co-extrusion of hydrophilic coatings onto porous
hydrophobic support structures. J Membr Sci
adjusted to deliver polymer chain orientation or
207:143156
more ordered arrangement along the elongation Jiang LY, Chung TS, Li DF, Cao C, Kulprathipanja
direction that improves selectivity of gas separa- S (2004) Fabrication of Matrimid/polyethersulfone
tion membranes (Kneifel and Peinemann 1992). dual-layer hollow ber membranes for gas separation.
J Membr Sci 240(12):91103
Shearing effect that occurs inside the spinneret
Jiang LY, Wang Y, Qiao XY, Chung TS, Lai JY
can also induce orientation, which motivates (2009) Polyimides membranes for pervaporation and
many design strategies of conduit channel near biofuels separation. Prog Ploys Sci 34(11):11351160
the exit (Widjojo et al. 2007). Kneifel K, Peinemann YV (1992) Preparation of hollow
ber membranes from polyetherimide for gas separa-
Composite hollow bers are mostly formed by
tion. J Membr Sci 65:295307
dip coating, interfacial polymerization, etc., as an Matsuura T (1994) Synthetic membranes and membranes
additional step to the porous hollow ber formed separation processes. CRC press, Boca Raton
by phase inversion. Presently, the type of dual- Ong YK, Chung TS (2013) Pushing the limits of high
performance dual-layer hollow ber fabricated via
layer composite asymmetric hollow bers pro-
I2PS process in dehydration of ethanol. AIChE
duced by co-extrusion through a triple-orice J 59:30063018
spinneret represents a new direction Peng N, Widjojo N, Sukitpaneenit P, Teoh MM, Lipscomb
(He et al. 2002). Compared to the single-layer G, Chung TS, Lai JY (2012) Evolution of polymeric
hollow bers as sustainable technologies, Prog Polym
asymmetric hollow bers, the dual-layer hollow
Sci 37:14011424
bers are more advantageous due to the following Ulbricht M (2006) Advanced functional polymer
reasons (Jiang et al. 2009): membranes. Polymer 47:22172262
Widjojo N, Chung TS, Krantz WB (2007)
A morphological and structural study of Ultem/P84
1. Material cost saving by more than 95 %
copolyimide dual-layer hollow ber membranes with
2. Possibility to employ brittle (engineering delamination-free morphology. J Membr Sci
infeasible) but advanced material 294:132146
126 Asymmetric Membrane
step from one material. Phase inversion is a pro-

Asymmetric Membrane cess wherein the thermodynamically stable poly-
mer solution is converted into the instable one by
Lan Ying Jiang changing dope composition. The resultant mem-
School of Metallurgy and Environment, Central branes exhibit graded morphology with a gradual
South University, Changsha, China transition from relative dense structure at one side
of the membrane to fully porous structure on the
other side, in consistence with the characteristics
For almost all the membrane process, ux is one of phase inversion course. In a simplied manner,
of the essential process parameters. The advan- this type of membranes is often described as hav-
tages associated with high ux are small mem- ing a thin skin supported by a porous support. The
brane area and efcient system footprint. These thin skin layer carries out the separation, while the
two factors are often considered by chemical engi- substrate generally presents negligible resistance
neers or managers in their decision-making about to the transport of penetrants. This so-called skin
setting up processes using membrane separation. layer is nonporous for gas separation, reverse
The thinner the required membrane structure, the osmosis, pervaporation, and nanoltration mem-
higher the membrane ux. However, mechanical branes. There also exist several other techniques
strength will be reduced with decreasing mem- pertaining to phase inversion. They are thermally
brane thickness. or pressure induced. The membranes generated
One of the solutions to this problem was asym- have a rather isotropic cross-section morphology.
metric membrane. Loeb and Sourirajan in 1963 To fulll the same target but avoid the disad-
(Loeb and Sourirajan 1964) invented integrally vantages of the integrally skinned asymmetric
skinned asymmetric membrane for reverse osmo- membranes, a series of other types of membrane
sis. They are produced by non-solvent-induced are developed. Figure 1 shows the schematic of
liquid-liquid phase separation processes, in the dense membrane and ve principle types of
which the entire structures are generated in one membranes that are currently used (Paul and
a d
Selective layer
Dense film
(tens of m) Microporous
skin layer
(porosity>1%)
Defect-free
Selective layer
skin layer
b (0.11m) e
Gutter layer
Microporous
substrate Microporous
(100300m) support
c Sealing layer f Selective

(0.11m) asymmetric
layer (material A)
Selective
skin layer
(porosity <104%) Microporous
support
(material B)
Asymmetric Membrane, Fig. 1 Schematic of dense and composite, (d) single-layer composite, (e) multilayer com-
asymmetric membranes. (a) Dense lm, (b) integrally posite, (f) asymmetric composite
skinned asymmetric, (c) multicomponent (caulked)
Atomic Force Microscopy (AFM) 127
Yampolskii 1994): (a) homogeneous dense mem- nonporous skin, has been hindered by the difcul-
brane, (b) integrally skinned asymmetric, ties of obtaining simultaneously the good selec-
(c) caulked, (d) single-layer composite, tivity and the high productivity. In other words, A
(e) multilayer composite, and (f) asymmetric com- the intrinsic property of some advanced material
posite membranes. Usually, a homogeneous cannot be expressed efciently in their asymmet-
membrane is prepared through solution casting. ric membranes. Various physical or chemical
This type of membrane is used for the lab-scale treatments can make the skin layer more perfect
investigation to fundamentally understand the (Jiang et al. 2009). One of the breakthroughs
intrinsic properties of polymers. Various tech- conquering this problem in the membrane devel-
niques are available in the formation of composite opment was made in 1980s by Henis and Tropidi
membranes. They can also be categorized as (Henis and Tripodi 1980). Their invention of coat-
asymmetric membranes morphologically. The ing asymmetric polysulfone membranes with sil-
multilayer composite (Fig. 1d) is an extension of icon rubber effectively sealed the pinholes and led
single-layer composite, and the asymmetric com- to rst generation of commercial gas separation
posite (Fig. 1e) is the combination of the integrally membrane, Prism or Prism Separator.
skinned asymmetric and composite approaches.
In addition to decreasing selective layer thickness,
one of the most attractive aspects of the composite References
membranes is their potential for minimizing mate-
rials cost, because only the selective layer must Guliants VV, Carreon MA, Lin YS (2004) Ordered
mesoporous and macroporous inorganic lms and
use a high-performance while more expensive
membranes. J Membr Sci 235:5372
material. The freedom in material choice and Henis JMS, Tripodi M (1980) Multicomponent mem-
membrane formation technique selection with branes for gas separation. US Patent 4,230,463
composite membranes facilitates the development Jiang LY, Wang Y, Qiao XY, Chung TS, Lai JY
(2009) Polyimides membranes for pervaporation and
of organic/inorganic layered structure that ef-
biofuels separation. Prog Polym Sci 34(11):11351160
ciently combines the characteristics of single Loeb S, Sourirajan S (1964) High ow porous membranes
organic or inorganic material and expands the for separating water from saline solutions. US Patent
application area of membrane separation. The 3,133,132
Paul DR, Yampolskii YP (1994) Polymeric gas separation
aforementioned asymmetric or composite struc-
membranes. CRC Press, Boca Raton
ture can be formed into different geometries such Ulbricht M (2006) Advanced functional polymer mem-
as the at sheet, the hollow bers, etc., by appro- branes. Polymer 47:22172262
priately choosing the fabricating or casting
equipment.
Today, extensive knowledge exists on how to
control the asymmetric membranes structure
including cross-section morphology by the selec- Atomic Force Microscopy (AFM)
tion of material and membrane formation condi-
tions. The polymeric composite membranes are Nidal Hilal and Daniel Johnson
usually made by solution coating, interfacial poly- Centre for Water Advanced Technologies and
merization, and plasma polymerization methods Environmental Research (CWATER), College
on the surface of a prefabricated porous support or of Engineering, Swansea University, Swansea,
by co-extrusion of two or more solutions (Ulbricht Wales, UK
2006). For composite membrane using inorganic
materials, they are prepared by various techniques
from solvent evaporation, in situ growth from The atomic force microscope is a versatile tool
solution, and other alternative methods (Guliants increasingly used for the physical characterization
et al. 2004). The work at commercializing asym- of surfaces and is of great interest for the visual-
metric membranes, especially those having ization and analysis of process surfaces including
128 Atomic Force Microscopy (AFM)
those of membranes used for ltration. It is capa-

ble of resolving features from the micrometer mm
0.20
down to the subnanometer scale and can operate 0.10
in air and liquid environments, allowing mem- 0.00
branes to be studied in environments matching 3
those encountered during their operation (Hilal

et al. 2004; Hilal and Johnson 2010), which 2
mm
allows assessment of effects of, for instance, pH,
ionic strength, and effects of additives on 1
3
membrane structure, a feature not available with 2
0 1 mm
other high resolution imaging applications. In 0
addition, the surface needs no special preparation, 25
providing it remains clear of unwanted contami-
nation, and does not need to be electrically con-
20
ductive, limitations found with some imaging
techniques.
The AFM consists of a sharpened probe 15
% of pores
mounted at the end of a exible cantilever. The
tip of the probe is then used to feel the underly- 10
ing membrane surface, producing a three-
dimensional map of the sample topography. 5
There are three basic imaging modes: contact
mode, where the tip maintains constant contact
0
with the surface, while feedback loops adjust its 0.10 0.15 0.20 0.25 0.30 0.35
height to maintain cantilever deection and hence Pore size, m
force; tapping mode, where the cantilever is
Atomic Force Microscopy (AFM), Fig. 1 3D image of
vibrated close to the resonant frequency and inter- cyclopore microltration membrane and pore-size distri-
mittently contacts the surface with decreased lat- bution for same membrane obtained by AFM
eral forces compared with contact mode; and
noncontact mode in which the probe tip interacts
with attractive forces very close to the surface but
is kept away from hard repulsive interactions. electrical and electrochemical, mechanical, adhe-
Noncontact mode is capable of extremely high sive, frictional and magnetic properties of surfaces
resolution but is the most difcult to attain in of interest.
practice. From the imaging data, several quantita- As well as providing this information, the
tive parameters of interest to membrane technol- AFM probe can be used as a force sensor by
ogists can be obtained including: surface moving the probe into and out of contact with
roughness, pore size, pore-size distribution, as the sample surface and monitoring deection in
well as showing in ne detail the morphology of the microcantilever arm. Through careful calibra-
fouled and unfouled membrane surfaces and the tion of the system this deection can be converted
effects of chemical modication of membrane into force, directly measuring the interaction force
surfaces (Johnson et al. 2012; Kochkodan as a function of probe tip sample separation
et al. 2013). Figure 1 shows a typical image of a distance (Gibson et al. 2004). This allows the
cyclopore microltration membrane and a detection of DLVO, hydrophobic, hydrostatic
resulting pore-size distribution (Bowen and steric interactions, membrane stiffness, and
et al. 1996; Bowen and Hilal 2009). There are a adhesion forces. The sharp tip may be replaced by
large number of expansions for AFM instruments a colloidal particle, which may be functionalized
allowing many other modes allowing mapping of to simulate any of a number of potential
Atomistic Simulations Methods 129
membrane foulants allowing quantitative mea- such as phonon calculations, free-energy optimi-
surements of foulant rejection and attachment zations (molecular mechanics), molecular dynam-
forces. This makes possible the direct quantica- ics (MD), Monte Carlo simulations, and crystal A
tion of membrane fouling properties of different structure prediction, are used to interpret existing
materials under a range of environmental condi- experimental data and predict new phenomena
tions and with only a relatively small sample of and to provide a way forward where experiments
membrane needed. are not yet possible, e.g., under extreme condi-
tions or at atomistic size- and timescales which are
difcult to detect directly (Allen and Tildesley
References
1989; Haile 1992; Frenkel and Smit 2002; Leach
Bowen WR, Hilal N (eds) (2009) Atomic force microscopy 2001; Brenner 2000; Allen 2004; Tocci and
in process engineering. Butterworth-Heineman, Oxford Pullumbi 2011).
Bowen WR, Hilal N et al (1996) Atomic force microscope Molecular modeling is primarily a tool for
studies of membranes: surface pore structures of calculating the energy of a given molecular struc-
cyclopore and anopore membranes. J Membr Sci
110(2):233238 ture, and the goal is to understand and model the
Gibson CT, Johnson DJ et al (2004) Method to determine motion of each atom in the material.
the spring constant of atomic force microscope cantile- Different levels of atomistic simulations exist,
vers. Rev Sci Instrum 75(2):565567 ranging from quantum mechanical models to sta-
Hilal N, Johnson DJ (2010) The use of atomic force
microscopy in membrane characterisation. In: tistical methods. This means solving numerically
Drioli E, Giorno L (eds) Comprehensive membrane the classical or quantum mechanical microscopic
science and engineering. Elsevier Science, Oxford, equations for the motion of interacting atoms or
pp 521538 even deeper electrons and nuclei.
Hilal N, Al-Zoubi H et al (2004) A comprehensive review
of nanoltration membranes: treatment, pretreatment, Quantum mechanical (QM) or ab initio
modelling and atomic force microscopy. Desalination methods describe matter at the electronic level,
170:281308 considering the fundamental particles, electrons
Johnson DJ, Al-Malek SA et al (2012) Atomic force and protons. The equation from which molecular
microscopy of nanoltration membranes: effect of
imaging mode and environment. J Membr Sci properties can be derived is the Schrodinger equa-
389:486498 tion, and various approximations must be intro-
Kochkodan V, Johnson DJ et al (2013) Polymeric mem- duced in order to extend the utility of the method
branes: surface modication for minimizing (bio)col- to polyatomic systems.
loidal fouling. Adv Colloid Interface Sci (in press).
doi:10.1016/j.cis.2013.05.005 Atomistic methods are used to compute molec-
ular properties, which do not depend on electronic
effects; the whole atom is modeled just as a soft
sphere and obeys the laws of statistical mechanics.
Atomistic simulation utilizes analytic potential
Atomistic Simulations Methods
energy expressions (sometimes referred to as
empirical or classical potentials) to describe the
Elena Tocci
systems.
The analytic potential energy functions are
simplied mathematical expressions that
Italy
attempt to model interatomic forces arising
from the quantum mechanical interaction of
Atomistic Simulations electrons and nuclei. Their use is dictated by
the need to model systems with sizes and/or
Atomistic simulations are theoretical and compu- timescales that exceed available computing
tational modeling tools for interpreting what hap- resources, required for quantum calculations,
pens at the atomic scale in solids, liquids, which give no account of the complex elec-
molecules and plasmas. Atomistic simulations, tronic structure of atoms.
130 Atomistic Simulations Methods
Forces between atoms are derived from empir- traditionally split into one-body, two-body,
ical interatomic potentials that are obtained from three-body terms:
tting material properties (e.g., lattice constant,
elastic constants, vacancy formation energy, etc.)
! ! !
from experimental data or QM calculations. They V nonbonded r 1 , r 2 , . . . , r N
may depend on the distance between atoms, X ! XX ! !
angles between bonds, angles between planes, etc. v ri w r i, r j . . .
i i ji
The general form of the total potential of the
N-atom system describes types of interactions,
!
bonded and non-bonded, and can be written as: The v r i term represents an externally applied
potential eld and describes external force elds
X
! ! !
V r 1 , r 2 , . .. , r N V 1 r i
! (e.g., gravitational eld) and external constraining
X X elds (e.g., the wall function for particles

! ! ! ! ! ! ! in a
V2 r i, r j V3 r i, r j, r k . . . chamber). The pair potential w r i , r j w r ij
i, ji i, ji, kj
neglect three-body (and higher order)
interactions.
The part of the potential energy V representing The Lennard-Jones potential is the most com-
bonding interactions will include terms of the monly used form:
following kind:

1X r 2 s 12 s6
V intramolecular K ij1 r ij r eq wLJ r 4e
2 bonds r r
1 X 2
K yijk yijk yeq
2 where s is the diameter and e is the depth of the
bend
potential energy well. If electrostatic charges are
angles
present, the appropriate Coulomb potentials are
1 X X f, m
K ijkl 1 cos mfijkl gm added:
2 m
torsion
angles Q1 Q2
wCoulomb r
4pe0 r
The bonds
typically involve the separation
! !
r ij r i r j between adjacent pairs of atoms where Q1, Q2 are the charges and e0 the permittiv-
in a molecular framework, and a harmonic form ity of the free space.
with specied equilibrium separation has been To be effective, an analytic potential energy
used, although this is not the only possible type. function must possess the following critical
The bend angles yijk are between successive properties:
! ! ! !
bond vectors such as r i r j and r j r k and
involve three atom coordinates. Usually this Flexibility: A potential energy function must be
bending term is quadratic in the angular displace- sufciently exible that it accommodates as
ment from the equilibrium value, although wide a range as possible of tting data.
periodic functions are also used. The torsion Accuracy: A potential should be able to accurately
angles fijkl are dened in terms of three reproduce properties of interest as closely as
connected bonds; hence four atomic coordinates possible.
are used. Transferability: A potential function should be
The part of the potential energy V representing able to study a variety of properties for which
non-bonded interactions between atoms is it was not t.
Autothermal Reforming 131
Computational efciency: Evaluation of the func- steam reforming of the fuel (endothermic reac-
tion should be relatively efcient depending on tion) with the partial oxidation of a small amount
quantities such as system sizes and timescales of the fuel (exothermic reaction) in order to A
of interest, as well as available computing achieve an autothermal reaction that proceeds
resources. without external input of energy (Chang
et al. 2010; Tiemersma et al. 2012).
The major methods are molecular mechanics The most studied autothermal reforming is the
(MM), molecular dynamics (MD), Monte Carlo conversion of methane to hydrogen (Gallucci
(MC), and additionally, there is a whole range of et al. 2009). The overall chemical reactions taking
hybrid techniques which combine features from place in the autothermal reforming of methane
both MD and MC methods. include steam reforming (Eq. 1), water gas shift
(Eq. 2), and total oxidation (Eq. 3). The energy
generated by the oxidation reaction and WGS is
References
used for the SMR:
Allen MP (2004) Introduction to molecular dynamics sim-
ulation computational soft matter: from synthetic poly- CH4 H2 O CO 3H2 (1)
mers to proteins. In: Attig N, Binder K, Grubmuller H,
Kremer K (eds) Lecture notes, John von Neumann CO H2 O CO2 H2 (2)
Institute for computing, Julich, NIC series,
vol 23, ISBN 3-00-012641-4, pp 128
Allen MP, Tildesley DJ (1989) Computer simulation of CH4 2O2 ! CO2 2H2 O (3)
liquids. Clarendon Press, Oxford
Brenner DW (2000) The art and science of an analytic This reaction system can be carried out efciently
potential. Phys Status Solidi B 217:23
Frenkel D, Smit B (2002) Understanding molecular simu-
in a membrane reactor as the extraction of hydro-
lation: from algorithm to applications, 2nd edn. Aca- gen during the reaction shifts the equilibrium
demic, San Diego reactions toward completion at moderate temper-
Haile JM (1992) Molecular dynamics simulation. Wiley, atures, and thus the extent of oxidation reaction to
Chichester
Leach AR (2001) Molecular modelling: principles and
achieve autothermal reforming is moderate.
applications, 2nd edn. Prentice Hall, Harlow One of the problems of autothermal reforming
Tocci E, Pullumbi P (2011) Chapter 1: Multi-scale molec- carried out in membrane reactors is the mismatch
ular modeling approaches for designing/selecting poly- between the oxidation reaction rate and the
mers used for developing novel membranes. In:
Drioli E, Barbieri G (eds) Membrane engineering for
reforming reaction rate. The oxidation is often
the treatment of gases: gas-separation problems with much faster than the reforming, and for this reason
membranes. The Royal Society of Chemistry, Cam- in packed bed membrane reactors, a high-
bridge, UK, pp 128 temperature region is obtained at the beginning
of the bed followed by a low-temperature region
at the end of the bed. This could cause problems to
the membranes that could be damaged by high
Autothermal Reforming temperatures while not working properly (low
ux see Richardson equation) at lower temper-
Fausto Gallucci
atures (Tiemersma et al. 2006; Gallucci et al.
Chemical Process Intensication, Department of
2010).
Chemical Engineering and Chemistry,
To circumvent these problems, uidized bed
Eindhoven University of Technology, Eindhoven,
membrane reactors are often proposed for this
The Netherlands
kind of reaction system, as the solid circulation
inside the reactor allows a virtually isothermal
Autothermal reforming or oxidative steam condition even in case of highly exothermic
reforming is a combination of conventional reactions.
132 Azeotropic Distillation
Examples of autothermal reforming (ATR) 97&partnerID=40&md5=47095694b70255a4e72794

(or oxidative steam reforming) reactions also 5b6ad24016
Tosti S, Borelli R, Santucci A, Scuppa L (2010) Pd-Ag
include the ATR of ethanol and methanol or the membranes for auto-thermal ethanol reforming. Asia
ATR of naphtha (Lin et al. 2010; Tosti et al. 2010; Pac J Chem Eng 5(1):207212, Retrieved from http://
Moreno and Wilhite 2009). All these reactions www.scopus.com/inward/record.url?eid=2-s2.0-77649
have been successfully tested in membrane reac- 116340&partnerID=40&md5=50f41a519d60bbaa07
aa2f96c86792bc
tors for pure hydrogen production.
References
Azeotropic Distillation
Chang H-F, Pai W-J, Chen Y-J, Lin W-H
(2010) Autothermal reforming of methane for produc- Alessandra Criscuoli
ing high-purity hydrogen in a Pd/Ag membrane reactor. Institute on Membrane Technology, National
Int J Hydrog Energy 35(23):1298612992, Retrieved Research Council of Italy, ITM-CNR, Rende,
from http://www.scopus.com/inward/record.url?eid=
2-s2.0-78049471869&partnerID=40&md5=631e9a956f Italy
70441c60cc3fb3987ed76c
Gallucci F, Van Sint Annaland M, Kuipers JAM
(2009) Autothermal reforming of methane with inte-
grated CO2 capture in novel uidized bed membrane Azeotropic distillation is a particular type of dis-
reactors. Asia Pac J Chem Eng 4(3):334344. tillation by which it is possible to separate azeo-
doi:10.1002/apj
tropes (Perry and Green 1984). Azeotropes are
Gallucci F, Van Sintannaland M, Kuipers JAM (2010)
Theoretical comparison of packed bed and uidized mixtures of two or more substances that boil
bed membrane reactors for methane reforming. Int together at a constant temperature. It is, therefore,
J Hydrog Energy 35(13):71427150, Retrieved from impossible to separate them directly by distilla-
http://www.scopus.com/inward/record.url?eid=2-s2.
tion, because a distillate of the same composition
0-77954826477&partnerID=40&md5=569ec26bf5
26471832579c3f965b1d90 of the liquid feed is produced. Azeotropic distilla-
Lin W-H, Liu Y-C, Chang H-F (2010) Autothermal tion is based on the addition of a compound that
reforming of ethanol in a Pd-Ag/Ni composite membrane acts on the volatility of the substances contained
reactor. Int J Hydrog Energy 35(23):1296112969,
into the azeotrope, so that a new azeotrope, made
Retrieved from http://www.scopus.com/inward/record.
url?eid=2-s2.0-78049467904&partnerID=40&md5= of the substances present in the starting azeotrope
0048616ca9b8d29ade703c2eea0ab1bf and the added compound, is formed. The new
Moreno AM, Wilhite BA (2009) Autothermal hydrogen azeotrope can be, then, removed by distillation,
generation from methanol in a ceramic microchannel
and a residual highly rich in one of the substances
network. In: Conference proceedings 2009 AIChE
annual meeting, 09AIChE. Retrieved from http://www. is obtained. A typical example is the distillation of
scopus.com/inward/record.url?eid=2-s2.0-779522829 mixtures of water-ethanol. By classic distillation,
00&partnerID=40&md5=0650ebd8cd9440d3ac32e0 the distillate contains 95 % of alcohol (pure boil-
3b4f859c7d
ing temperature, 78.4 C) and 5 % of water (pure
Tiemersma TP, Patil CS, Sint Annaland MV, Kuipers JAM
(2006) Modelling of packed bed membrane reactors boiling temperature, 100 C), and no further incre-
for autothermal production of ultrapure hydrogen. ments in ethanol purity can be reached, because
Chem Eng Sci 61(5):16021616, Retrieved from this mixture boils, as a unique compound, at
http://www.scopus.com/inward/record.url?eid=2-s2.0-
78.17 C. However, if benzene or cyclohexane is
30344455632&partnerID=40&md5=9fb77a0611e8e6
a89745a4b8625bbebf added to the mixture, a ternary azeotrope is formed
Tiemersma TP, Kolkman T, Kuipers JAM, van Sint that incorporates all the water and has a lower
Annaland M (2012) A novel autothermal reactor con- boiling point than ethanol. Therefore, if the distil-
cept for thermal coupling of the exothermic oxidative
lation is now carried out, the new azeotrope will be
coupling and endothermic steam reforming of methane.
Chem Eng J 203:223230, Retrieved from http://www. recovered as distillate, while practically pure eth-
scopus.com/inward/record.url?eid=2-s2.0-848656377 anol is produced as residue (see Fig. 1). Besides
Azeotropic Distillation 133
Ternary
azeotrope
A
Benzene or
cyclohexane
+ Ternary azeotrope
+ ethanol Distillation
Water-ethanol
azeotrope
Ethanol
Azeotropic Distillation, Fig. 1 Azeotropic distillation of water-ethanol
azeotropic distillation, other methods can be References

also employed for the separation of azeotropes
and, in particular, membrane operations like Perry RH, Green D (1984) Perrys chemical engineers
handbook. McGraw-Hill, Singapore
pervaporation and membrane distillation.
B
b-Lactoglobulin and stabilizer and can be used in the production of

Alpha-Lactalbumin Separation confections (Cayot and Lorient 1997).
There are two types of membrane systems that
Andrew L. Zydney have been explored for the separation of
Department of Chemical Engineering, The b-lactoglobulin and a-lactalbumin: ultraltration
Pennsylvania State University, University Park, and membrane adsorbers (also referred to as mem-
PA, USA brane chromatography). Cheang and Zydney
(2003) demonstrated that b-lactoglobulin and
a-lactalbumin could be separated using commer-
b-Lactoglobulin and a-lactalbumin are the two cial ultraltration membranes by adjusting the
dominant proteins in whey (see Table 1). Although solution pH and ionic strength to enhance the
whey protein concentrates are used as food addi- selectivity. These results were subsequently
tives in the production of a variety of dairy prod- extended to purify b-lactoglobulin and
ucts, meats, and beverages, the unique nutritional, a-lactalbumin from a whey protein isolate using
therapeutic, and functional characteristics of the a two-stage ultraltration process (Cheang and
individual whey proteins are largely unrealized in Zydney 2004).
these whey products. There has thus been consid- Membrane chromatography is of potential inter-
erable commercial interest in the separation of est for the separation of b-lactoglobulin and
b-lactoglobulin and a-lactalbumin to take advan- a-lactalbumin due to the high ow rates that can
tage of their functional and biological properties. be achieved in these systems. For example, Sau
a-Lactalbumin is attractive as a nutraceutical and Fee (2011) examined the use of a mixed mode
because of its high tryptophan content (Cayot and membrane chromatography system containing
Lorient 1997). a-Lactalbumin also provides both cation and anion exchange sites for the frac-
enhanced whippability in meringue-like formula- tionation of whey proteins. b-Lactoglobulin was
tions and may have potential as a contraceptive preferentially bound to the membrane, with the
agent. b-Lactoglobulin is an effective foam a-lactalbumin collected in the ow through.

DOI 10.1007/978-3-662-44324-8
136 Backwashing
b-Lactoglobulin and Alpha-Lactalbumin Separation, Table 1 Physical characteristics of major whey proteins
(From Zydney 1998)
Protein Concentration (g/L) Molecular weight (g/mol) Isoelectric pH
b-Lactoglobulin 2.7 18,362 5.2
a-Lactalbumin 1.2 14,147 4.54.8
Immunoglobulins 0.65 150,0001,000,000 5.58.3
Bovine serum albumin 0.4 69,000 4.74.9
Lactoferrin 0.1 78,000 9.0
Lactoperoxidase 0.02 89,000 9.5
Glycomacropeptide 7
References pressure membrane systems, i.e., microltration

and ultraltration, while high pressure membrane
Cayot P, Lorient D (1997) Structure-function relationships systems, i.e., nanoltration and reverse osmosis,
of whey proteins. In: Damodaran S, Paraf A (eds) Food
do not employ backwashing to control fouling.
proteins and their applications. Marcel Dekker, New
York, pp 225256 Backwashing is most commonly initiated on a
Cheang B, Zydney AL (2003) Separation of alpha- timed cycle; however, it may also be initiated on
lactalbumin and beta-lactoglobulin using membrane a transmembrane pressure set point or a volume
ultraltration. Biotechnol Bioeng 83:201209
ltered set point (AWWA 2013). Backwashing is
Cheang B, Zydney AL (2004) A two-stage ultraltration
process for fractionation of whey protein isolate. performed in stages through a membrane plant in
J Membr Sci 231:159167 order to ensure that water is being produced at all
Sau SM, Fee CJ (2011) Simultaneous anion and cation times; in other words, there is rotation through
exchange chromatography of whey proteins using a
which individual membrane units will be
customizable mixed matrix membrane. J Chromatogr
A 1218:90039009 backwashed, while the majority, at any given
Zydney AL (1998) Protein separations using membrane time, are still maintained in operation. When
ltration: new opportunities for whey fractionation. Int backwashing is no longer considered effective at
Dairy J 8:243250
controlling membrane fouling, as indicated by a
set point loss in ux, permeate production, or high
TMP value, the membrane system will undergo
Backwashing chemical cleaning. Additionally, a chemically
enhanced backwash (CEB) may be put in place
Onita D. Basu before a membrane unit is taken off-line for a full
Department of Civil and Environmental chemical clean. In the event of a CEB, an oxidant,
Engineering, Carleton University, Ottawa, such as chlorine, and acids or bases, such as citric
Canada acid or sodium hydroxide, are added to the back-
wash water to assist in the removal of foulants
from the membrane pores and membrane surface.
Backwashing of a membrane refers to the reversal Backwashing has been found to be helpful in
of ow through a membrane system in compari- removing cake layer fouling that is particulate in
son to the normal ow direction required for per- nature. However, research has shown that
meate production. The purpose of backwashing is backwashing alone does not remove the majority
to assist with removing or loosening foulants from of the cake layer fouling but instead mainly
the membrane surface and within the membrane increases its porosity, effectively loosening the
pores to control membrane fouling. It is a funda- cake layer against the membrane surface
mental membrane operational step to control the (Ye et al. 2011). Therefore, it is important to
rate of transmembrane pressure (TMP) increase include an air scour or cross-ow velocity step
over time. Backwashing is utilized with low with backwashing to effectively decrease the size
Backwashing Frequency 137
and thickness of the cake layer foulants on the Wu J, Le-Clech P, Stuetz RM et al (2008) Effects of
membrane surface (Remize et al. 2010). However, relaxation and backwashing conditions on fouling in
membrane bioreactor. J Membr Sci 324:232
it is important to keep in mind that pore blocking Ye Y, Chen V, Le-Clech P (2011) Evolution of fouling
results in a faster deterioration of the membrane deposition and removal on hollow bre membrane
ux and is more likely to lead to irreversible during ltration with periodical backwash. Desalina- B
fouling than cake layer fouling. The cake layer tion 283:198205
itself can prevent pore blocking by acting like a
prelter for the membrane surface. Thus, if too
much of the cake layer is removed, the membrane
system will have a faster development of pore Backwashing Frequency
blocking over several cycles of backwashing. As
such it is important to congure an optimized Onita D. Basu
backwash procedure with associated air scour or Department of Civil and Environmental
cross-ow velocity step for a particular membrane Engineering, Carleton University, Ottawa,
and water type. When optimizing a system, it is Canada
important to note that research has found that
backwash duration and backwash ux at low
rates, equal to or less than the permeate ux, are Optimization of backwash frequency is an impor-
not likely to have an impact on fouling rate in low tant consideration in terms of membrane operation
fouling systems, although backwash duration was as it uses produced water for the backwash cycle;
found to have an optimum value associated with therefore, the more frequent the backwash, the
net permeate production (Akhondi et al. 2014). lower the efciency of the membrane system
Other research has shown that too high of a back- overall in terms of produced water. In addition to
wash ux may actually exacerbate fouling by the backwash frequency, which is generally every
potentially removing too much of the cake layer 3090 min, consideration should also be applied
and resulting in more irreversible pore fouling for backwash duration and backwash ux. Back-
occurring (Wu et al. 2008). Given that wash duration is how long a membrane system is
backwashing has been shown to clearly assist in backwashed and is typically between 1 and 3 min.
controlling fouling, it important for all membrane In some cases, backpulsing may also be used, and
plants to monitor the effectiveness of their back- this term refers to short burst cycles for reversing
wash procedures on a regular basis with consid- the ow through the membrane that generally last
eration for changes in water quality over different between 0.5 and 5 s. Backwash ux is the uid
seasons. ow rate through the membrane, and it is gener-
ally equal to the initial operating permeate ux of
the membrane system, although it may be higher.
Backwash frequency is linked to the operating
References membrane ux, with a lower ux resulting in a
lower required backwash frequency, while a
Akhondi E, Wicaksana F, Fane AG (2014) Evaluation of
fouling deposition, fouling reversibility and energy higher ux tends to require a higher backwash
consumption of submerged hollow ber membrane frequency (Wang et al. 2008; Rafn et al. 2011).
systems with periodic backwash. J Membr Sci Backwashing is often considered to control
452:319331
American Water Works Association (AWWA) (2013) Com-
reversible fouling; however, it is also tied to the
mittee report: metrics and methods for MF/UF system control of irreversible fouling as well. Research has
optimization. J Am Water Works Assoc 105(2):6576 demonstrated, for instance, that while backwash
Remize PJ, Guigui C, Cabassud C (2010) Evaluation of frequency may not appear to impact on the rate of
backwash efciency, denition of remaining fouling
and characterisation of its contribution in irreversible
reversible fouling, it can inuence the rate of irre-
fouling: case of drinking water production by versible fouling. Specically, longer periods
air-assisted ultra-ltration. J Membr Sci 355:104111 between backwashes (i.e., 60 min vs. 30 min)
138 Bacteria and Spore Removal
resulted in an increase in the irreversible fouling downside of this is that the product properties
rate (Rafn et al. 2012). The importance of this will be inuenced and this is mostly undesirable.
information is that higher irreversible fouling In this respect, membrane microltration
rates result in higher frequencies of chemical could be an interesting alternative when used for
cleaning for a membrane system, which has both cold sterilization. If the bacterial count is not too
an adverse impact on water production and the high and other components are sufciently
economics of running a membrane system. Of smaller, the removal of bacteria and spores can
course, consideration must be taken for the loss rather easily be carried out using dead-end
of system productivity when increasing the fre- microltration.
quency of backwashing as well. As backwash When the size of bacteria and other compo-
frequency increases, the membrane productivity nents overlap, this separation is far from straight-
is concomitantly decreasing; backwashing can forward. An illustrative example can be found in
result in upwards of a 10 % loss in membrane dairy separation, in which cold sterilization of
productivity. It is, therefore, important to weigh milk has been reported. The components that
out the benets and losses associated with increas- are most important for this separation are the milk
ing backwashing frequency and losing productiv- fat globules (cream; typical sizes from 0.1 to
ity versus decreasing the backwash frequency 15 mm), bacteria (0.55 mm), and casein micelles
and potentially increasing the required chemical (20300 nm). Since cream and bacteria overlap in
cleaning of the membrane system. size, the cream is rst removed by centrifugation.
The resulting skim milk receives the cold sterili-
zation treatment. Various researchers have inves-
tigated how far the bacterial count could be
References reduced as listed in Table 1.
From the table, it is clear that a variety of
Rafn M, Germain E, Judd S (2011) Optimising operation
membranes and process conditions have been
of an integrated membrane system (IMS) a
Box-Behnken approach. Desalination 273:136141 used, ranging from very high to low cross-ow
Rafn M, Germain E, Judd S (2012) Inuence of velocities, and application of a uniform trans-
backwashing, ux and temperature on microltration membrane pressure or frequent back pulsing.
for wastewater reuse. Sep Purif Technol 96:147153
These conditions are clearly aiming at processing
Wang L, Wang X, Fukushi KI (2008) Effects of operational
conditions on ultraltration membrane fouling. Desali- under very different conditions, where frequent
nation 229:181191 back pulsing will control the amount of deposited
material, the uniform transmembrane pressure
concept aims at stable ltration conditions along
the length of the membrane. All approaches have
Bacteria and Spore Removal shown interesting results; the log reductions that
can be achieved are around 4 (10,000-fold reduc-
Karin Schroen tion), although it should be noted that these values
Laboratory of Food Process Engineering, are not as high as obtained after regular heat
Wageningen University and Research Centre, treatment.
Wageningen, The Netherlands The highest log reduction (6.6, which is higher
than for regular pasteurization) was claimed for
microsieves, which are silicon plates with very
Bacteria can be of great inuence on many pro- uniform pores prepared by laser interference
cesses, be it in, for example, food, pharma, fer- lithography (Van Rijn and Elwenspoek 1995).
mentation, or water production. Obviously, Although the bacterial reduction was measured
bacteria (and their spores) can be removed by for dead-end ltration using a 0.5 mm microsieve
heat treatment or addition of specic components and SMUF (simulated ultraltrate) spiked with
that inhibit their growth or even kill them, but the Bacillus subtilis, it is expected that the high log
Bacterial Biofilm Formation 139
Bacteria and Spore Removal, Table 1 Comparison of cold sterilization results from various sources
Process conditions cross-ow/pressure, Log
Membrane type and ux UTP, back pulsing reduction Source
Ceramic 1.4 mm; 1.4104 m/s 50 kPa, 7.2 m/s UTP, skim milk Above 3.5 Saboya and Maubois
2000
B
Reversed asymmetric 0.87 mm; 0.51 m/s; back pulsing frequency Between Guerra et al. 1997
1.4104 m/s 0.21 s1, skim milk 4 and 5
Microsieve 0.5 mm Dead-end ltration of spiked SMUF 6.6 Van Rijn and
Kromkamp 2001
Bactocatch: ceramic membranes 68 m/s, skim milk, UTP Holm et al. 1989
count reduction is a result of its narrow pore size Saboya LV, Maubois JL (2000) Current developments of
distribution of the microsieve. microltration technology in the dairy industry. Lait 80:541
Van Dinther AMC, Schron CGPH, Boom RM
In case of the uniform transmembrane pressure (2011) High-ux membrane separation using uid
concept, reduction of bacteria and spores by skimming dominated convective uid ow. J Membr
microltration is carried out near the critical pres- Sci 371(12):20
sure, at which the amount of particles that is Van Rijn CJM, Elwenspoek MC (1995) Micro ltration
membrane sieve with silicon micro machining for
carried toward that membrane by permeate is industrial and biomedical applications. In: Proceedings
counterbalanced by the amount of particles dif- of micro electro mechanical systems (MEMS), Amster-
fusing away from the membrane to the feed solu- dam, p 83
tion. We expect that this can be even taken one Van Rijn CJM, Kromkamp J (2001) Method for ltering
milk. WO Patent 0,209,527. http://www.spgveco.com/
step further, as indicated in the work of van precision+metal/applications/ltration. Accessed July
Dinther et al. (2011), in which particles that size- 2012
wise correspond to milk fat globules and bacteria
were reported to be separated by uid skimming
and lift effects. For this to occur, rst a nonporous
channel is used to induce particle migration after Bacterial Adhesion
which in a porous area the small particles that are
situated close to the wall can specically be Bacterial Biolm Formation
removed. This process has not been demonstrated
at large scale, but it holds great promise since it
would allow for direct separation of bacteria from
full milk without the need of centrifugation.
Bacterial Biofilm
Besides, the separation is no longer determined
Bacterial Biolm Formation
by the pore size of the membrane; metal sieves
from SPG Veco with uniform pores of around
20 mm were used, but by the process conditions
that determine which part of the feed is removed. Bacterial Biofilm Formation
Lidietta Giorno and Napoleone DAgostino

References
Guerra A, Jonsson G, Rasmussen A, Waagner Nielsen E, Italy
Edelsten D (1997) Low cross ow velocity
microltration of skim milk for removal of bacterial
spores. Int Dairy J 7:849
Holm S, Malmberg R, Svensson K (1989) Method and
Synonyms
plant producing milk with low bacterial content. WO
Patent 8:601687 Bacterial biolm; Bacterial adhesion; Biofouling
140 Bacterial Biofilm Formation
Biolms were observed as early as 1674, when the collectors surface. Although useful as a snap-
Antonie van Leeuwenhoek used his primitive but shot of biolm growth, this type of prole is
effective microscope to describe aggregates of limited when considering the intimate processes
animalcules that he scraped from human tooth of cellsubstrate/cellcell interaction. Characklis
surface (Costerton 1999). Since then, more accu- and Marshal later described an eight-step process
rate descriptions of biolms are made. Bacteria which included the formation of an initial condi-
generally exist in one of two types of population: tioning layer, reversible and irreversible adhesion
planktonic, freely existing in bulk solution, and of bacteria, and the eventual detachment of cells
sessile, as a unit attached to a surface or within the from a mature biolm for subsequent
connes of a biolm (Garrett et al. 2008). colonization.
Biolm is a result of many complicated steps. It Anything that may be present within the bulk
includes the formation of a conditioning lm on a uid can through gravitational force or movement
materials surface, the movement of bacteria, an of ow settle onto a substrate and become part of a
attachment process, the growth on material surfaces, conditioning layer. This layer modies substrata
and the breakdown nally. For bacteria, the advan- facilitating accessibility to bacteria. Surface
tages of biolm formation are numerous. These charge, potential, and tensions can be altered
advantages include: protection from antibiotics favorably by the interactions between the condi-
(Godberg 2002), disinfectants (Peng et al. 2002), tioning layer and substrate. The substrate provides
and dynamic environments (Chen et al. 1998). anchorage and nutrients augmenting growth of the
Over the past few decades, biolm growth has bacterial community.
been observed in many industrial and domestic Initially, planktonic microbial cells are
domains. Many industries suffer the ill effects of transported from bulk liquid to the conditioned
biolm growth of one type or another, which can surface either by physical forces or by bacterial
result in heavy costs in cleaning and maintenance. appendages such as agella. The reversible
Biolms occurring in membrane systems may adsorption of a fraction of the cells reaching the
cause severe loss performance and the use of surface normally occurs. Local environmental
costly cleaning procedures to maintain output variables which contribute to bacterial adhesion
and quality. The fouling is often so severe that are factors such as available energy, surface func-
acceptable operation cannot be maintained and tionality, bacterial orientation, temperature, and
membrane replacements are needed. It is neces- pressure conditions. If repulsive forces are greater
sary to understand the biolm formation mecha- than the attractive forces, the bacteria will detach
nism with the aim to propose a solution to contrast from the surface. The probability of this phenom-
this fouling. enon occurs is higher before the formation of the
Bacteria are capable of colonizing almost any conditioning layer. The activation energy for
surface and have been found at extreme condi- desorption of bacteria is low and so it is likely to
tions such as temperatures from 12 C to 110 C occur, underlining the weakness of the bonds.
and pH values between 0.5 and 13. Physical forces associated to bacterial adhesion
Biolm growth occurs by physical, chemical, include the van der Waals forces, steric interac-
and biological processes. Fletcher described the tions, and electrostatic (double layer) interaction,
accumulation of microorganisms on a collecting collectively known as the DVLO (Derjaguin,
surface as a process of three stages: (i) adsorption, Verwey, Landau, and Overbeek) forces, which
or the accumulation of an organism on a collector originally described the interaction of a colloidal
surface, i.e., substrate (deposition); particle with a surface (Rutter and Vincent 1980).
(ii) attachment, or the consolidation of the inter- According to this theory, the total interaction
face between an organism and a collector, often between a surface and a particle is the summation
involving the formation of polymer bridges of their van der Waals and Coulomb interactions.
between the organism and collector; and (iii) col- Since the van der Waals attractive force is domi-
onization, or growth and division of organisms on nant in the vicinity of a surface, particles adhere
irreversibly because they cannot separate from the inhibited in sessile species as motility is restricted
surface by Brownian motion. In contrast, the Cou- and no longer necessary. Simultaneously, expres-
lomb interaction becomes dominant at a distance sion of a number of genes for the production
away from the surface because the van der Waals of cell surface proteins and excretion products
force decreases sharply with distance. Other inter- increases. Surface proteins (porins), such as Opr B
actions that DVLO theory takes into consideration C and Opr E, allow the transport of extracellular
are hydrophobichydrophilic and osmotic (Chang products into the cell and excretion materials
and Chang 2002) and have also been described in out of the cell, e.g., polysaccharides (Hancock
terms of thermodynamic interaction (Gallardo- et al. 1990). The structure of many Gram-
Moreno et al. 2002). negative bacterial polysaccharides is relatively
In real time, a number of the reversibly simple, comprising either homopolysaccharides
adsorbed cells remain immobilized and become or heteropolysaccharides (Sutherland 2001).
irreversibly adsorbed. It has been argued that the These molecules impart mechanical stability and
physical appendages of bacteria (agella, m- are pivotal to biolm adhesion and cohesion and
briae, and pili) overcome the physical repulsive evasion from harsh dynamic environmental
forces of the electrical double layer (De Weger conditions. They consolidate the biolm struc-
1987). Subsequently, the appendages make con- ture. Hall-Stoodley and Stoodley identied the
tact with the bulk lattice of the conditioning differences in gene expression of planktonic
layer stimulating chemical reactions such as and sessile cells, and as many as 57 biolm-
oxidation and hydration and consolidating the associated proteins were not found in the plank-
bacteriasurface bond. Some evidence has tonic prole.
shown that microbial adhesion strongly depends The stationary phase of growth describes a
on the hydrophobichydrophilic properties of phase where the rate of cell division equals the
interacting surfaces (Liu et al. 2004). rate of cell death. At high cell concentration, a
As the stationary cells divide (binary division), series of cell signaling mechanisms are employed
daughter cells spread outward and upward from by the biolm, and this is collectively termed
the attachment point to form clusters. Typically, quorum sensing (Bassler 1999). Quorum sensing
such interactions and growth within the develop- describes a process where a number of
ing biolm form into a mushroom-like structure. autoinducers (chemical and peptide signals in
This structure is believed to allow the passage of high concentrations, e.g., homoserine lactones)
nutrients to bacteria deep within a biolm. After are used to stimulate genetic expression of both
an initial stage, a rapid increase in population is mechanical and enzymatic processors of algi-
observed, otherwise described as the exponential nates, which form a fundamental part of the extra-
growth phase. This depends on the nature of the cellular matrix. The death phase sees the
environment, both physically and chemically. The breakdown of the biolm. Enzymes are produced
rapid growth occurs at the expense of the sur- by the community itself which break down poly-
rounding nutrients from the bulk uid and the saccharides holding the biolm together, actively
substrate. At this stage the physical and chemical releasing surface bacteria for colonization of fresh
contribution to the initial attachment ends and the substrates. Alginate lyase produced by Pseudo-
biological processes begin to dominate. Excretion monas uorescens and Pseudomonas aeruginosa,
of polysaccharide intercellular adhesin (PIA) N-acetyl-heparosan lyase by Escherichia coli, and
polymers and the presence of divalent cations hyaluronidase by Streptococcus equi are exam-
interact to form stronger bonding between cells ples of the enzymes used in the breakdown of
(Dunne 2002). the biolm matrix (Sutherland 1999). Simulta-
Differential gene expression between the two neously, the operons coding for agella proteins
bacterial states (planktonic/sessile) is in part asso- are upregulated so that the organisms have the
ciated to the adhesive needs of the population. For apparatus for motility and the genes coding for a
example, the production of surface appendages is number of porins are downregulated, thus
142 Bacterial Biofilm Formation
completing a genetic cycle for biolm adhesion weak hydrogen bonding forces; (ii) viscous
and cohesion. damping due to polymeric friction and hydrogen
Changes in pH can have a marked effect on bond breakage; and (iii) alignment of the poly-
bacterial growth and as such are frequently mers in the shear direction (Klapper et al. 2002).
exploited in the production of detergents and dis- Such properties change with increased tempera-
infectants used to kill bacteria. Bacteria possess ture. Increasing the temperature of polysaccha-
membrane-bound proton pumps which extrude rides produces a gel-like substance which
protons from the cytoplasm to generate a trans- gradually increases in strength until a critical
membrane electrochemical gradient (Rowland point is reached. At the critical point the gel
2003), i.e., the proton motive force. The passive forms a solution (Villain-Simonnet et al. 2000).
inux of protons in response to the proton motive Such behavior affects the viscosity of the poly-
force can be a problem for cells attempting to saccharides which can affect biolm adherence.
regulate their cytoplasmic pH (Booth The optimum temperature for a microorganism is
1985). Large variations in external pH can over- associated with an increase in nutrient intake
whelm such mechanisms and have a biocidal resulting in a rapid formation of biolm
effect on the microorganisms. Bacteria respond (Stepanovic et al. 2003). Enzymes are responsible
to changes in internal and external pH by of nutrient metabolism; then the formation of a
adjusting the activity and synthesis of proteins biolm is dependent on the presence and reaction
associated with many different cellular processes rates of enzymes. Temperature inuences the
(Olsen 1993). Studies have shown that a gradual reaction rate of enzymes having an impact on the
increase in acidity increases the chances of cell development of the cells. Optimum temperatures
survival in comparison to a sudden increase by result in the healthy growth of bacterial
rapid addition of HCl (Li 2001). This suggests that populations. Temperatures away from the opti-
bacteria contain mechanisms in place which allow mum negatively inuence bacterial enzyme reac-
the bacterial population to adapt to small environ- tion rates, and a reduction of bacterial growth
mental changes in pH. However, there are cellular efciency occurs. Fletcher reported the effect of
processes which do not adapt to pH uctuations so temperature on attachment of stationary phase
easily. One such process is the excretion of cells. Findings showed that a decrease in tem-
exopolymeric substances (polysaccharides). Opti- perature reduced the adhesive properties of a
mum pH for polysaccharide production depends marine Pseudomonad. It is believed that the
on the individual species, but it is around pH 7 for effect was due to a decrease in the bacterial
most bacteria. surface polymer at lower temperatures as well
Both mixed species and pure culture biolms as effects such as reduced surface area. How-
behave like viscoelastic uids. Biolms exhibit ever, Herald and Zottola observed that the pres-
both irreversible viscous deformation and revers- ence of bacterial surface appendages was
ible elastic response and recoil (Ohashi and dependent on temperature. At 35 C cells were
Harada 2004). Extracellular polymeric substances shown to have a single agellum, while at 21 C
like alginate, xanthan, and gellan gum aggregate they had two to three agella, and at 10 C, cells
due to hydrogen bonding to form highly hydrated exhibited several agella. This may suggest that
viscoelastic gels (Stoodley et al. 1999). The pres- the initial interaction between the bacteria and
ence of acetylated uronic acids in the bacterial substrate may increase with a lowering of tem-
alginate of P. aeruginosa biolms increases its perature, increasing the likelihood of adhesion.
hydration capacity. These properties provide the Perhaps the more uniform properties of poly-
biolm with mechanical stability (Stoodley et al. saccharides at lower temperatures increase the
2002). possibility of biolm adhesion, because many
The matrix formed by EPS responds to stress microbial polysaccharides undergo transition
by exhibiting (i) elastic tension due to a combina- from an ordered state at lower temperatures
tion of polymeric entanglement, entropic, and and in the presence of ions to a disordered
state at elevated temperature under low ionic Hall-Stoodley L, Stoodley P (2002) Developmental regu-
environments (Nisbet et al. 1984). lation of microbial biolms. Curr Opin Biotechnol
13:228233
Although there is plenty of information Hancock REW, Siehnel R, Martin N (1990) Outer mem-
describing the effect of temperature on bacterial brane proteins of Pseudomonas. Mol Microbiol
growth in culture, the effect of temperature on the 4:10691075 B
removal of adhered microorganisms is not so well Herald PJ, Zottola EA (1988) Attachment of Listeria
monocytogenes to stainless steel surfaces at
documented. The reports available describe fairly various temperatures and pH values. J Food Sci
radical effects of temperature on adhered bacteria. 53:15491552
Marion-Ferey et al. observed the effect of high Hiemenz PC, Rajagopalan R (1997) Principles of colloid
temperatures (8090 C) on the removal of and surface chemistry. Marcel Dekker, New York
Klapper I, Rupp CJ, Cargo R et al (2002) Viscoelastic uid
biolms. It was discovered that these tempera- description of bacterial biolm material properties.
tures were not effective for biolm removal due Biotechnol Bioeng 80:289296
to baking effects at high temperature, appar- Li Y (2001) Cell density modulates acid adaptation in
ently increasing the adherent nature of the biolm Streptococcus mutans. J Bacteriol 183:68756884
Liu Y, Yang S, Xu H et al (2004) The inuence of cell and
to the surface. substratum surface hydrophobicities on microbial
attachment. J Biotechnol 110:251256
Marion-Ferey K, Pasmore M, Stoodley P et al (2002) Bio-
References lm removal from silicone tubing: an assessment of the
efcacy of dialysis machine decontamination proce-
Bassler BL (1999) How bacteria talk to each other: regu- dures using an in vitro model. J Hosp Infect 53:6471
lation of gene expression by quorum sensing. Curr Marshall KC, Stout R, Mitchell R (1971) Mechanism of the
Opin Microbiol 2:582587 initial events in the sorption of marine bacteria to sur-
Booth IR (1985) Regulation of cytoplasmic pH in bacteria. faces. J Gen Microbiol 68:337348
Microbiol Rev 49:359378 Nisbet BA, Sutherland IW, Bradshaw IJ et al (1984) XM-6:
Chang YI, Chang PK (2002) The role of hydration force on a new gel-forming bacterial polysaccharide. Carbohydr
the stability of the suspension of Saccharomyces Polym 4:377394
cerevisiae application of the extended DLVO theory. Ohashi A, Harada H (2004) Adhesion strength of biolm
Colloids Surf A Physicochem Eng Asp 211:6777 developed in an attached-growth reactor. Water Sci
Characklis WG, Marshal KC (1990) Biolms. Wiley, New Technol 29:281288
York Olsen ER (1993) Inuence of pH on bacterial gene expres-
Chen MJ, Zhang Z, Bott TR (1998) Direct measurement of sion. Mol Microbiol 8:514
the adhesive strength of biolms in pipes by microma- Peng JS, Tsai WC, Chou CC (2002) Inactivation and
nipulation. Biotechnol Tech 12:875880 removal of Bacillus cereus by sanitizer and detergent.
Costerton JW (1999) Introduction to biolm. Int Int J Food Microbiol 77:1118
J Antimicrob Agents 11:217221 Rowland BM (2003) Bacterial contamination of dental unit
De Weger LA, van der Vlugt C, Wijfjes AHM et al (1987) waterlines: what is your dentist spraying into your
Flagella of a plant-growth-stimulating Pseudomonas mouth? Clin Microbiol Newsl 25:7377
uorescens strain are required for colonization of Rutter PR, Vincent B (1980) Microbial adhesion to sur-
potato roots. J Bacteriol 169:27692773 faces. Ellis Horwood, London
Dunne WM (2002) Bacterial adhesion: seen any good Stepanovic S, Cirkovic I, Mijac V et al (2003) Inuence of
biolms lately? Clin Microbiol Rev 15:155166 the incubation temperature, atmosphere and dynamic
Fletcher M (1977) The effects of culture concentration and conditions on biolm formation by Salmonella
age, time, and temperature on bacterial attachment to spp. Food Microbiol 20:339343
polystyrene. Can J Microbiol 23:16 Stoodley P, Lewandowski Z, Boyle J et al (1999) Structural
Fletcher M (1980) Microbial adhesion to surfaces. Ellis deformation of bacterial biolms caused by short-term
Horwood, Chichester uctuations in uid shear: an in situ investigation of
Gallardo-Moreno AM, Gonzalez-Martin ML, Perez- biolm rheology. Biotechnol Bioeng 65:8392
Giraldo C et al (2002) The measurement temperature: Stoodley P, Cargo R, Rupp CJ et al (2002) Biolm material
an important factor relating physicochemical and adhe- properties as related to shear-induced deformation and
sive properties of yeast cells to biomaterials. J Colloid detachment phenomena. J Ind Microbiol Biotechnol
Interface Sci 271:351358 29:361367
Garrett TR, Bhakoo M, Zhang Z (2008) Bacterial adhesion Sutherland IW (1999) Polysaccharases for microbial
and biolms on surfaces. Prog Nat Sci 18:10491056 exopolysaccharides. Carbohydr Polym 38:319328
Godberg J (2002) Biolms and antibiotic resistance: a Sutherland IW (2001) Microbial polysaccharides from
genetic linkage. Trends Microbiol 10:264 Gram-negative bacteria. Int Dairy J 11:663674
144 BAL
Villain-Simonnet A, Milas M, Rinaudo MA (2000) New injected into reservoirs, and the resulting mixture
bacterial exopolysaccharide (YAS34). II. Inuence of generally features stable supersaturated BaSO4,
thermal treatments on the conformation and structure.
Int J Biol Macromol 27:7787 ready to precipitate. While alkaline scales can be
controlled by altering the feed pH prior to ltra-
tion, sulfate scales are insensitive to pH and there-
fore cannot be controlled by acid addition. BaSO4
BAL scales are usually needle shape crystals which can
easily damage the active layer of the membrane,
Bioarticial Liver resulting in loss of integrity (Fig. 1).
Du Pont method is widely applied as predic-
tion method for BaSO4 scale in RO systems
using the salt solubility at 25 C and the solubil-
Barium Sulfate (BaSO4), Fouling ity product, Kc (Boerlage 2001). The Kc
due to expresses the dynamic equilibrium between the
scalant crystal ions in solution and the crystal
Alice Antony and Pierre Le-Clech solid phase, for example, the Kc for BaSO4 is
School of Chemical Engineering, The University expressed as:
of New South Wales, UNESCO Centre for
Membrane Science and Technology, Sydney, Kc Ba2 SO4 2 at equilibrium
NSW, Australia
Kc is determined graphically as a function of ionic
strength for the desired recovery (Boerlage
Barium sulfate or barite, BaSO4 is a nonalkaline et al. 1999). The relationship between Kc and
scale, particularly problematic in membrane sys- ionic strength is derived from data on the effect
tems due to its extremely low solubility, 105 mol/l of individual monovalent and divalent cations on
(equivalent to 2.33 mg/l) in pure water. Also, this barium solubility. When the scalant concentration
scale is extremely difcult to remove and clean product exceeds Kc, the solution is supersaturated
from membrane surfaces if not detected at an and scaling may occur. Du Pont recommends a
early stage and is allowed to age into a hard adher- recovery such that the scalant concentration prod-
ent layer. BaSO4 could originate from various uct is 20 % below the Kc to prevent scaling. This
sources. In the offshore oil industry for example, ensures a safety factor for concentration polariza-
seawater (containing high level of sulfate) is tion which may occur at the membrane surface,
500
400 S Ba
300 S 0
Ca
Ba Ba
200
Si
100 Si CaCa Ba
Ba Ba
Ba
0
0 1 2 3 4 5 6 7
ke V
Barium Sulfate (BaSO4), Fouling due to, Fig. 1 SEM-EDS images showing BaSO4 scaling on RO membrane treating
cooling water from power plant (Picture provided by Dr Alice Antony, The University of New South Wales, Australia)
Basic Electrochemical Relations in Membrane Processes 145
causing a higher localized concentration than in

the bulk solution (Boerlage 2001). The major Basic Electrochemical Relations
limitation of this index is that it lacks background in Membrane Processes: Electron
information on interactions between ionic species Conductivity, Ion Conductivity,
on saturation concentrations; as a result, it also Ohms Law, and Coulombs Law B
neglects the effects of common and uncommon
ions on solubility of scale forming species and the Lidietta Giorno1 and Heiner Strathmann2
1
interactions between ionic species. Institute on Membrane Technology, National
Addition of antiscalant is the common method Research Council of Italy, ITM-CNR, Rende,
to control the BaSO4 scale formation. However, Italy
2
continuous monitoring of the RO performances Universitt Stuttgart, Institute of Chemical
through conventional techniques lacks the sensi- Process Engineering Stuttgart,
tivity to detect the subtle changes occurring across Baden-Wurttemberg, Germany
the membrane, and in particular, early deposition
of BaSO4. This lack of detailed and timely assess-
ment of the fouling state can result in delay in In certain membrane processes, charged compo-
applying the necessary cleaning strategies, whose nents such as salt solutions are separated by neu-
efciency is limited for fully developed BaSO4 tral membranes or membrane carrying xed
scaling layers, known to be particularly resistant positive or negative charges using an electrical
to conventional cleaning methods. When BaSO4 potential as driving force. The behavior of ions
scaling is observed during autopsy analysis, in solution and in a membrane is rather different
crown ethers and concentrated sulphuric acid from that of uncharged components because of the
are generally recommended as cleaning agents, strong interaction of electrical charges with each
causing partial membrane hydrolysis (Van der other and their surrounding. Therefore some fun-
Leeden 1991). Alkaline chelants have also damental electrochemical relations shall be briey
been reported to be efcient cleaning agents reviewed.
(Chesters 2009).
Electron and Ion Conductivity

References and Ohms Law
Boerlage SFE (2001) Scaling and particulate fouling The transport of electric charges, i.e., an electric
in membrane ltration systems. Swets & Zeitlinger, current, can be achieved by two characteristic
Lisse modes (Kortum 1957), i.e., by:
Boerlage SFE, Kennedy MD, Witkamp JG, Van Der Hoek
PJ, Schippers JC (1999) BaSO4 solubility prediction in
reverse osmosis membrane systems. J Membr Sci Transport of electrons through solid material
159:4759 such as metals
Chesters SP (2009) Innovations in the inhibition and ransport of ions through liquids such as elec-
cleaning of reverse osmosis membrane scaling and
trolyte solutions
fouling. Desalination 238:2229
van der Leeden MC (1991) The role of polyelectrolytes in
barium sulfate precipitation. PhD thesis, Technical The conductivity of electrons, i.e., metal con-
University of Delft, Delft ductors, is generally three to ve orders of mag-
nitude higher than the conductivity of electrolyte
solutions. Furthermore, the conductivity of metals
Further Reading is decreasing with increasing temperature, while
Antony A, Low JH, Gray S, Childress AE, Le-Clech P,
the conductivity of electrolyte solutions is increas-
Leslie G (2011) Scale formation and control in high
pressure membrane water treatment systems: a review. ing with the temperature. The most important
J Membr Sci 383:116 difference between electron and ion conductivity,
146 Basic Electrochemical Relations in Membrane Processes
however, is the fact that ion conductivity is always k

Leq (4)
coupled with a transport of mass, while due to the Cza na zc nc
very small mass of an electron virtually no mass is
transported in an electron conductor. In spite of Here Leq is the equivalent conductivity of an
the basic difference in the transport mode between electrolyte; C is the molar concentration of the
electron and ion conductivity, the electrical cur- electrolyte in the solution; za and zc are the charge
rent can be described in both cases by the same numbers of the anion and cation, respectively; and
mathematical relation which states that the elec- na and nc are the stoichiometric coefcients of the
trical current is proportional to the electrical anion and cation, respectively.
potential driving force. This relation is referred The stoichiometric coefcient gives the num-
to as Ohms law which is given by ber of anions and cations in a mole electrolyte, and
the valence gives the number of charges related to
RI (1) an ion. For example, for NaCl nc and na are iden-
tical and 1 and also za and zc are 1. However, for
Here is the electrical potential between two MgCl2 nc is 1 and na is 2, and zc is 2 and za is 1.
electron sources, e.g., between two electrodes in The equivalent conductivity can be expressed
an electrolyte solution expressed in Volt, I is the as the sum of the contributions from its individual
electrical current between the electron sources ions. Thus
expressed in ampere, and R is the electrical resis-
tance expressed in ohm. Leq la lc (5)
The resistance R is a function of the specic
resistance of the material, the distance between Here Leq is the equivalent conductivity and la and
the electron sources, and the cross-section area lc are the equivalent conductivities of anion and
of the material through which the electric current cation, respectively.
passes. It is given by The electrical current passing through an
electrolyte solution under the driving force of
l an electrical potential gradient is proportional
Rr (2)
q to the concentration of the ions in the solution,
the stoichiometric coefcients for cat- and
Here R is the overall resistance, r is the specic anions, and the number of electrical charges
resistance, l is the length, and q is the cross-section carried by one ion, i.e., its valence and the ion
area of the conducting material. mobility in the electrolyte solution. The number
The reversal of the resistance and of the spe- of electrical charges carried by all ions of an
cic resistance, respectively, is the conductivity electrolyte under the driving force of an elec-
and the specic conductivity and thus is trical potential gradient through a certain area
A is given by
1 1
S and k (3) X X
R r Je zi ui ni C e N A D zi FJ i (6)
i i
Here S is the conductivity and k is the specic
conductivity. Here Je is the ux of electrical charges and Ji that
The conductivity of electrolyte solutions of the individual ions; z, u, and n are the charge
depends on the concentration and the valence of number, the ion mobility, and the stoichiometric
the ions in the solution. It is expressed as the coefcient, respectively; C is the concentration of
equivalent conductivity or molar conductivity the electrolyte; e is the charge of an electron, NA
(Harned and Owen 1958). For a solution of a the Avogadro number, an electrical potential,
single electrolyte, the equivalent conductivity is and the subscript i refers to the ions in the solu-
given by tion; and F is the Faraday constant.
The ion mobility u has the dimension be expressed in terms of a relative permittivity er
m2s1V1. The charge of an electron is which is also referred to as dielectric constant of the
e = 1.6019 1019 C, and the Avogadro medium (Atkins 1990). It is given by
number is NA = 6.0232 1023 molecules per
mol or in this case electrical charges per mole. e B
er (9)
The product of e and NA is called the Faraday eo
constant which is F = 96,485 C eq1
(Atkins 1990). Here e, er, and eo are the permittivity of a medium,
The ux of electrical charges represents an the relative permittivity, and the permittivity in the
electrical current, which is for a solution of a vacuum, respectively.
single electrolyte according to Ohms law given Coulombs law describes the interaction
by between two point charges. In some cases it is
desirable to know the force exerted by one
X charge q in a certain distance r to its surrounding.
zi ui ni C F A D
U X i This force is referred to as electric eld. It is
I zi FJ i A (7) given by
R i
l
K
Here I is the current, U is the applied voltage, R is E (10)
the resistance, A is the area through which the q
current passes, D is the voltage difference
Here K is the force, E is the electric eld, and q is
between two points, and l is the distance between
an electrical charge.
the two points.
In determining the electric eld, the direction
of the eld must be taken into account. For a point
electrical charge q, the electric eld is radial sym-
Coulombs Law and the Electric Field
metric (Starzak 1984; Kortum 1957) and given by
Effect on Ions in Solution
q
Electrical charges carried by ions interact with E (11)
each other. Opposite charges attract each other 4pe r 2
and identical charges repel each other. The inter-
action force between two point electrical charges Here E is the electric eld around a point electric
q1 and q2 in a vacuum separated by a distance r is charge q, r is the distance from the electrical
given by charge, and e is the permittivity of the media
surrounding the charge. The electric eld has the
q1 q2 units of newton per Coulomb or volt per distance.
K (8)
4peo r 2 The electric eld around a point charge is illus-
trated in Fig. 1a, which shows the eld strength or
Here K is the interaction force between two point eld density symbolized by density of the lines
electrical charges, q1 and q2 are electrical charges, which is decreasing with the distance from the
r is the distance between the two point electrical point charge. The electric eld between two
charges, and eo is the permittivity of vacuum. point charges of opposite charge is illustrated in
Equation 8 is referred to as Coulombs law and Fig. 1b. In this case the two charges will attract
the electrical charge is expressed in Coulomb. each other. If they carry the same charge, they will
The permittivity depends on the medium repulse each other.
between the electrical charges. In a vacuum The electric eld can be expressed by an elec-
the permittivity is a fundamental constant denoted trical potential gradient. If the electrical potential
as eo. It has the value 8.854 1012 [C2 J1 m1]. varies over a certain distance, the electric eld
If the medium is not a vacuum, the permittivity can associated with this potential is given by the
Basic Electrochemical a b
Relations in Membrane
Processes: Electron
Conductivity, Ion
Conductivity, Ohms
Law, and Coulombs Law,
Fig. 1 Schematic drawing
illustrating the electric eld
lines (a) around a point
charge and (b) between a
positive and a negative
point charge
a b
electrical potential
+
+

+
+ +
+
0
+
+ distance from center ion
+ +

of excess charges
+
concentration
+
+ +
+

+

0
distance from center ion
Basic Electrochemical Relations in Membrane Pro- around a center-ion and (b) the potential and the concen-
cesses: Electron Conductivity, Ion Conductivity, tration of excess charges around a center-ion as function of
Ohms Law, and Coulombs Law, Fig. 2 Schematic distance
drawing illustrating (a) the distribution of counterions
derivative in the direction of the electric eld is decreasing exponentially with the distance from
(Kort
um 1957): the center-ion.
The electrical potential surrounding a center-
4p ion is also decreasing with the distance from
div E div grad r (12)
e the center-ion, and it will reach zero at a certain
distance. This distance is called the Debye length
Equation 12 is a special form of the Poisson dif- and marks the outer boundary of the ion
ferential equation which expresses the change of atmosphere.
the electric eld and the potential, respectively,
around a single charge.
The effect of the electric eld around a center- The Electric Double Layer at the Surface
ion on the distribution of other ions in its sur- Membranes
rounding and the decay of the electrical potential
is illustrated in Fig. 2a, b which shows that the When membranes with xed charges at the sur-
concentration of the excess charges around an ion face are immersed into an electrolyte solution
Basic Electrochemical a b
Relations in Membrane O
H 2O
Processes: Electron H 2
H 2O
Conductivity, Ion
Conductivity, Ohms
+
O
B
2
H2 O
H
Law, and Coulombs Law,
illustrating (a) the closest O
H 2O
H2
approach of hydrated
+ + Dj
H2O
cations to the negatively
O
xed ions of the surface of a
2
H2 O
H
cation-exchange membrane
and (b) the potential
O
H 2O
dissipation in the double H2
layer according to the
+
H2 O
Helmholtz model
+
O 2
HO
H
2
charge, they will generate an electric eld directed layer in parallel at a certain distance from the
into an adjacent solution and attract, due to the membrane surface. This distance is determined
Coulomb forces, ions with the opposite charge. by the radius of the hydrated counterion as indi-
This leads to a layer on the membrane surface in cated in Fig. 3.
which the concentration of the oppositely charged The Helmholtz model resembles a parallel
ions is higher than in the bulk solution. In many plate capacitor where the charge is determined
aspects the xed charges of the membrane behave by the number of charges at the surface and in a
very similar to the charges surrounding a single plane parallel to the membrane surface.
charge in an electrolyte solution. These ions will The relation between the surface charges, the
form a layer at a certain distance from the mem- potential between the negatively and positively
brane surface, which neutralize the xed charges charged layers, and their distance are given by the
at the membrane surface. This double layer at the same equation, which describes the capacitance of
interface between the ion-exchange membrane a parallel plate capacitor (Starzak 1984), i.e.,
and the adjacent solution is kept in place by the
attraction force of the electrical potential gener- A q
Ce (13)
ated by the xed ions of the membrane and the 4 pd U
dispersing force of the thermal motion. This dou-
ble layer at the interface between the membrane Here C is the capacitance, A is the area of the
and the electrolyte solution, which is important for charge surface, d is the thickness of the Helmholtz
certain electrokinetic phenomena such as electro- layer, and q is the charge.
osmosis or streaming potential, can be described Introducing the surface charge density s which
by several models (Wagenen and Andrade 1980). is given by
The simplest one of these models is the Helmholtz
q
model. s (14)
A
and replacing the voltage U by the potential dif-

The Helmholtz Double Layer Model
ference Dj between the two layers
Helmholtz postulates that all counterions neces-
sary to neutralize the surface charge of an
U D (15)
ion-exchange membrane are aligned in a single
excess charges
a b
4 2 1 0
+ +
+
+ + +
+ Dj
+ + + +
+
+
+
+ +

+ +
+
distance distance
Basic Electrochemical Relations in Membrane Pro- double layer as function of the distance from the membrane
cesses: Electron Conductivity, Ion Conductivity, surface and (b) the potential dissipation as function of the
Ohms Law, and Coulombs Law, Fig. 4 Schematic distance from the surface of the membrane
drawing illustrating (a) the excess counterions in the
give a relation between the thickness of the double

O --
layer d, the surface charge, and the potential dif-
ference in the double layer:
104 + +
4p d
D s (16) H H
e
Basic Electrochemical Relations in Membrane Pro-
Here e is the permittivity and s is the surface cesses: Electron Conductivity, Ion Conductivity,
charge density. Ohms Law, and Coulombs Law, Fig. 5 Schematic
drawing illustrating the structure of a water molecule indi-
cating the dipole character due to the uneven distribution of
The Gouy-Chapman Double Layer Model
the electrical charges within the molecule
The Helmholtz model contains a number of rather
drastic simplications. For example, it assumes
that the xed ions form a homogeneous sheet
and that the counterions form a complementary can be calculated by applying the Poisson equa-
parallel sheet not affected by the thermal motion tion which provides a relationship between the
of the solution. In reality this is not the case.
potential and the excess charges at a given point
In the Gouy-Chapman model, the counterions in the solution and under the assumption that the
in the solution are no longer assumed to be aligned system obeys the Boltzmann statistic by
in a sheet but will occupy a nite volume in a
mixture of cat- and anions. The total number of s xD
counterions in this volume is identical to the xed D (17)
e
charges of the membrane surface. Their distribu-
tion, however, is such that the area close to the Here xD is the Debye length which determines the
membrane surface contains the largest excess of thickness of the double layer. The Debye length is
counterions as indicated in Fig. 4 and is very inversely proportional to the ion concentration of
similar to the ion atmosphere around a single the solution. With increasing ion concentration,
charge. the Debye length is decreasing.
In analogy to the potential decay around a The boundary layer at the surface of mem-
single charge, the potential at a certain distance branes carrying xed charges has a signicant
from the membrane surface carrying xed charges effect on the streaming potential and on the
Basic Electrochemical a b --
Relations in Membrane
Processes: Electron
+
+ + +
Conductivity, Ion
--
Conductivity, Ohms
+
--
--
+
B
+
Law, and Coulombs Law, --
+
+
illustrating the distribution --
+
of water molecules +
+
--
+
(a) around a cation and --
+
--
(b) around an anion
+
--
+ +
electric osmotic water transport as well as on the molecules as indicated in Fig. 6 (a) showing water
so-called zeta potential (Lyklema 1995). molecules surrounding a cation and (b) water mol-
ecules surrounding an anion.
Another important result of the molecular
Electrical Dipoles and Intermolecular interaction of water dipoles is the formation of
Forces the hydronium ions from protons and water. In
water the proton recombines readily with the
An electrical dipole consists of two opposite charges water molecule to the hydronium ion in which
attached to a molecule and separated by a certain the three H atoms are no longer distinguishable
distance. An attraction of the charges is prevented and carry the positive charge together. The hydro-
by the rigidity of the molecular structure. The poten- nium ion resembles a tripod with the three
tial produced by the dipole is different from that H atoms on one side and all equally apart from
produced by a single charge. The potential of a the oxygen atom. The rapid formation of the
dipole with a single charge is decreasing with the hydronium ion is the reason for the much faster
distance by the square of the distance. The dipole transport of protons compared to other ions in an
momentum of molecules is measured in Debye aqueous solution.
[D] with D = 3.33 1030 A s m. The dipole
momentum varies substantially depending on the
polarity of the molecules. It is 0 for a completely
nonpolar molecule such as CCl4 or H2 and ca. 1.80 References
D for a polar molecule such as water (Atkins 1990).
Atkins PW (1990) Physical chemistry, 4th edn. Oxford
The high dipole momentum of the water is the result
University Press, Oxford
of the water molecule structure in which the hydro- Harned and Owen (1958) The physical chemistry of elec-
gen atoms are attached to one side of the oxygen trolyte solutions, 3rd edn. Reinhold, New York
atom as illustrated in Fig. 5. Kortum G (1957) Lehrbuch der Elektrochemie. Verlag
Chemie, Weinheim
The strong dipole momentum of water is also Lyklema J (1995) Fundamentals of interface and colloid
the reason for its good solubility properties for science, vol II, Solid-Fluid Interfaces. Academic,
salts. The water dipoles tend to break the generally London
strong Coulomb forces between the negatively Starzak ME (1984) The physical chemistry of membranes.
Academic, New York, pp 4464
and positively charged ions of a salt crystal by
Wagenen and Andrade (1980) Flat plate streaming poten-
surrounding the ions. Thus, in an aqueous solu- tial investigations: hydrodynamics and electrokinetic
tion, each ion is surrounded by a shell of water equivalency. J Colloid Interface Sci 76:305
152 Batch Diafiltration
It should be pointed out that the removal of

Batch Diafiltration microsolutes and the concentration of
macrosolutes (i.e., the reduction of the volume
Zoltn Kovcs of process liquor) are both required for most appli-
Department of Food Engineering, Institue of cations. The term batch dialtration, in its broader
Bioengineering and Process Engineering, Szent sense, may stand for a batch ltration process that
Istvan University, Budapest, Hungary is designed to achieve the twin objectives of con-
centrating and purifying a multi-solute system
according to a specic diluant utilization strategy.
Batch dialtration refers to a pressure-driven In this context, batch dialtration is a complex
membrane ltration process in which a diluant process that may involve a sequence of consecu-
(pure solvent) is added into the feed tank in tive operational steps.
order to enhance the degree of separation of A straightforward way of achieving the dual
macrosolutes from microsolutes. In batch objectives of concentration and fractionation is to
dialtration, in contrast to continuous combine CVD with concentration mode opera-
dialtration, the retentate stream is recirculated tional steps (i.e., in which no diluant is added
to the feed tank, and only the permeate stream is into the feed tank). A so constructed typical
collected separately. During the operation, solute- three-step process, also referred to as traditional
free diluant is introduced into the feed tank to dialtration (TD), involves the following phases:
replace solvent losses as schematically illustrated (i) concentration mode to achieve an intermediate
in Fig. 1. The requirement for an effective separa- macrosolute concentration, (ii) constant-volume
tion is the utilization of a membrane which highly dialtration to wash out the microsolute by a
retains the macrosolute but permeable for the pure solvent introduced into the system, and
microsolute. Thus, depending on the application, (iii) further concentration to the nal desired
microltration, ultraltration, nanoltration, or macrosolute concentration. Beside TD, a number
even reverse osmosis membranes can be applied. of alternative strategies have been proposed.
In its strict, original sense, the term batch These include the sequential dilution dialtra-
dialtration refers to a process that aims at remov- tion (SDD), intermittent feed dialtration
ing the microsolutes from the process liquor. The (IFD), variable-volume dialtration (VVD), pre-
standard way of achieving this purication goal is concentration combined with variable-volume
to employ a constant-volume dialtration (CVD) dialtration (CVVD), and dynamic-volume
process that is probably the most common type of dialtration (DVD).
batch dialtration. In CVD, the feed volume is Dialtration techniques differ in controlling
kept constant by continuously adding a diluant at the addition of the diluant in terms of quantity
a rate equal to the permeation rate. and duration. The differences between the various
Batch Diafiltration,
Diluant u(t)
Fig. 1 Schematic
representation of batch Retentate Permeate q(t)
dialtration conguration
Feed
Batch Diafiltration 153
Batch Diafiltration, Table 1 Diluant utilization strate- V represents the feed tank volume. This initial
gies in batch diafiltration value problem describes the evolution in time
Name a - strategy of the feed volume and the feed concentration
Constant-volume a1 of any solute under the assumption that the
dilution diluant is solute-free and the feed tank is B
Traditional a f0, 1, 0g well-mixed. Note that the time-dependent vari-
dialtration
ables (i.e., permeate ux and the solute rejec-
Variable-volume a const, 0 < a < 1
dialtration tions) are, in a general case, a function of feed
Preconcentration a f0, a1 g, a1 const, 0 < a1 < 1 concentrations and may vary with operation
with variable- conditions (temperature, pressure, cross-ow
volume velocity, etc.).
dialtration
The unique feature of realizing both concen-
Intermittent feed a 0, amax n , 0
dialtration tration and fractionation puts membrane ltra-
tion in an attractive position and compares
Sequential a amax , 0n
dilution favorably with other separation methods or
dialtration even with a sequence of consecutive unit oper-
Dynamic-volume a at 0 at amax ations. In comparison with continuous pro-
dialtration
cesses, batch operations are especially suited
n number of repetition to small-scale operations, require less expensive
automatic controls, and enable a reduced mem-
brane area in order to reach the target (Baker
operations are best described by the proportional-
2004). Most batch plants operate under constant
ity factor a (i.e., the ratio of diluant ow d(t) to
mechanistic membrane pressure adjusted sim-
permeate ow q(t)) as a function of operation time
ply by the retentate valve. There exist, however,
(Foley 2006). For instance, TD process is charac-
other types of process control strategies in engi-
terized with a sequence a(t) = {0, 1, 0} with two
neering practice, such as constant ux or con-
unknown switching times at the end of the rst
stant wall concentration control (van Reis
and of the second time interval. Similarly, CVVD
et al. 1997). These are normally employed
process has two phases with constant a levels
when unfavorable side effects (e.g., enhanced
a(t) = {0, a1} and an unknown switching time.
fouling or product quality deterioration) occur
Table 1 shows the diluant control strategies
that can be associated with the high concentra-
applied in batch processing. Note that the best
tion at the membrane wall.
time-varying prole of diluant addition
needs not necessarily be one of the arbitrarily
predened proles. The dialtration process,
that is, designed by the evaluation of the optimal References
time-varying prole of the diluant ow, is referred
to as dynamic-volume dialtration (Paulen Baker R (2004) Membrane technology and applications.
et al. 2012). Wiley, Chichester
The governing differential equations for a gen- Foley G (2006) Ultraltration with variable volume
dialtration: a novel approach to water saving
eralized batch dialtration process are given as in dialtration processes. Desalination 199(13):
220221
dci ci q Paulen R, Fikar M, Foley G, Kovcs Z, Czermak P (2012)
Ri a, ci 0 ci0 , i 1, 2 Optimal feeding strategy of dialtration buffer in
dt V
dV batch membrane processes. J Membr Sci 411412:
a 1q, V 0 V 0 : 160172
dt van Reis R, Goodrich EM, Yson CL, Frautschy LN,
Whiteley R, Zydney AL (1997) Constant Cwall
where ci is the solute concentration in the feed ultraltration process control. J Membr Sci 130(12):
tank, Ri the rejection of component i, and 123140
154 Bed-to-Wall Mass Transfer Limitations
instance, into the product B, this being repre-

Bed-to-Wall Mass Transfer sentative of all possible reactions. If the reac-
Limitations tion rate is sufciently fast, the transport
through the gas phase becomes limiting and,
Alessio Caravella thus, a signicant concentration/pressure drop
National Institute of Advanced Industrial Science is generated, decreasing in fact the effective
and Technology Research Institute for concentration of species A that can be
Innovation in Sustainable Chemistry, ISC-AIST, involved in the surface reaction with respect
Tsukuba, Ibaraki, Japan to the bulk concentration.
As for the second case (concentration polari-
zation, Fig. 1b), a membrane permeable to both
Bed-to-wall mass transfer limitations. These phe- species A and B is considered, with the species
nomena, also referred to as concentration polar- A being the one to separate and more permeable
ization in the eld of membrane technology (higher transmembrane ux) than the species
(Gallucci et al. 2011), are more generally related B. Because of the effect of the higher ux of the
to the mass transfer resistance that a species species A, the species B is pushed toward the
encounters to reach an active wall. An active membrane wall, generating in this way an obstacle
wall is a surface where physical and/or chemical for the species A to reach the surface. Concentra-
transformations (including adsorption/desorption tion polarization is not desired for two main rea-
phenomenon and solubilization/desolubilization sons. First, it causes a decrease in the permeation
processes) occur. driving force of the species A, as the surface-
The main cases of interest in chemical engi- surface concentration difference is smaller than
neering are reported in Fig. 1. As for the former the bulk-bulk one. Second, it tends to favor the
(Fig. 1a), an impermeable catalytic wall is consid- permeation of the less-permeating species (B),
ered, where the reactant species A reaches the causing the actual membrane separation factor
catalytic surface and is transformed, as an to decrease.
Bed-to-wall mass transfer limitations

Concentration
a b Polarisation
A A
Flux (A) C
Bulk
AB
Bulk
Bulk
Flux (B)
B
B B
Surface
Surface
Surface
Catalytic Wall Membrane Wall

(Impermeable) (Permeable)
Bed-to-Wall Mass Transfer Limitations, reaction takes place and (b) Case of membrane (permeable
Fig. 1 Schematic representations of the two most com- wall) through which all the components can, in general,
mon cases of bed-to-wall mass transfer limitations. (a) pass. Membrane is supposed to be impermeable to the
Case of impermeable catalytic wall on which a chemical species C
Beer Clarification 155
References Beer Clarification, Table 1 Comparison of key

parameters of beer filtered with DE filtration and
Gallucci F, Van Sint Annaland M, Kuipers JAM microfiltration (Lipnizki 2005)
(2011) Modeling of membrane reactors for hydrogen Beer Beer
production and purication. In: Membrane engineering before after DE Beer after
for the treatment of gases. vol 2 Gas separation prob- ltration ltration microltration B
lems combined with membrane reactors, RSC energy Original 11.40 11.37 11.39
and environment series. Roheena Anand Publisher, extract
Books, Cambridge, pp 139, Chapter 10 [%]
Alcohol 3.84 3.83 3.84
[%]
Color 7.20 6.70 7.00
[EBC]
Viscosity 1.62 1.57 1.56
Beer Clarification [MPas]
Turbidity 32.00 0.53 0.41
Frank Lipnizki at 0 C
Alfa Laval Copenhagen, Sborg, Copenhagen, [EBC]
Denmark
3. AlfaBright system by Alfa Laval and

In the traditional brewing process, the beer is Sartorius a process based on the combination
claried after fermentation and before maturation of cassette microltration system with a high-
to remove mainly the remaining yeast but also speed separator as pretreatment
microorganisms and haze. The conventional pro-
cess for beer clarication is the combination of a The membranes established for beer ltration
high-speed separator followed by diatomaceous are all polymeric microltration membranes
earth (DE)/kieselguhr ltration which is based on polyethersulfone with 0.50.65 mm
complemented in some cases by second ltration pores and beer capacities of 0.51.0 hl/(m2 h)
with PVP (polyvinylpyrrolidone) to obtain a very (Buttrick 2007). The resulting beer quality from
clear beer. The major challenge of the conven- the membranes is similar or improved compared
tional process is related to the DE ltration to the DE ltration; in Table 1 a comparison
because the DE can vary in quality and is prob- between beer ltered by DE and microltration
lematic for handling and disposal since it is haz- membranes is given. Other membranes, e.g.,
ardous plus DE ltration generates large amounts ceramic membranes, have been tested for beer
of efuent. Alternatively, from its introduction in ltration but so far have not established them-
2001 (Buttrick 2007) cross-ow microltration selves on the market. The key difference between
has established itself in the brewing industry the three concepts is the use of a high-speed sep-
with over 50 breweries worldwide adopting this arator as pretreatment. In the BMF system a
DE-free beer ltration. Generally, three concepts retentate tank is used to collect the beer solids,
are currently used in the industry: while both the PROFi and the AlfaBright systems
are using high-speed separators as pretreatment
1. Beer Membrane Filtration (BMF) by Pentair before the membrane to remove the beer solids
(previously Norit) a system based on hollow and thus eliminating the need for a retentate tank.
ber microltration cartridges without a high- The systems can be typically run in batch or
speed separator as pretreatment continuous operation depending on the size
2. PROFi Membrane System by Pall and GEA and requirements of the individual brewery. In
Westfalia a combination of hollow ber batch operation, the complete system shifts from
microltration system with a high-speed sepa- ltration into cleaning mode after each batch.
rator as pretreatment Plants with continuous operation are arranged in
156 Beer Dealcoholization
skids/blocks and use sequential cleaning which process based on reverse osmosis can be divided
allows for some skids/blocks to be in standby into four individual operations which are typically
cleaning mode, while other blocks are in ltration carried out in batch mode:
mode. The major challenges for cross-ow mem-
brane beer ltration are the relatively high invest- 1. Pre-concentration: In this step the volume of
ment costs and complexity of the process when the feed beer is reduced. The beer is passed
compared to DE ltration and the increasing through the membrane modules and is then
availability of alternative DE-free lter aids. recycled to the batch tank. The
permeate water and alcohol is removed
from the process, while retentate, concentrated
References beer and avors, is returned to the batch tank.
2. Dialtration: This step is similar to the
Buttrick P (2007) Filtration the facts. Brewer Distiller Int pre-concentration step but dialtration
3(12):1219
water desalted and deoxygenized water is
Lipnizki F (2005) Optimisation and integration of mem-
brane processes in the beverage industry, 10th edn. added to wash out the alcohol. The amount of
Aachener Membran Kolloquium, Aachen, dialtration water added balances the amount
16.-17.03.2005 of permeate removed from the process, and
thus, the level in the batch tank remains con-
stant. This operation is continued until the
desired alcohol concentrations in the beer are
Beer Dealcoholization achieved.
3. Alcohol adjustment: In this step, the taste and
Frank Lipnizki alcohol content is ne-tuned by addition of
Alfa Laval Copenhagen, Sborg, Copenhagen, desalted and deoxygenized water.
Denmark 4. Posttreatment: In order to give the beer its
specic character and to balance taste losses
due to removal of the taste carrier alcohol, the
In the last decades the demand for low alcohol and CO2 levels can be adjusted, and hop extract,
alcohol-free drinks increased, e.g., in Germany syrup, or other avor enhancers are added.
the annual consumption of alcohol-free drinks
nearly doubled from 130.4 l per person in 1980 As an alternative to reverse osmosis, dialysis
to 248.2 l person in 1999, while the annual con- can be used for the dealcoholization of beer. Com-
sumption of alcoholic drinks declined from 179.5 monly, hollow ber modules are used for dialysis
to 156.3 l during the same period (Gebhardt allowing the beer to ow on one side of the mem-
2001). Conventionally beer can be dealcoholized brane and water as dialysate on the other side. The
by distillation, but additionally the membrane process is normally operated in counter-current
process reverse osmosis and dialysis have ow to maximize the concentration gradient over
established themselves for the partial dealcoho- the dialysis membrane and thus the driving force
lization of beer by eight to ten times. The key of the dialysis process. The dialysate is constantly
advantage of membrane processes over distilla- recycled over a steam stripping column to remove
tion is that beer can be dealcoholized at low tem- the alcohols, thus maintaining the driving force of
peratures typically 78 C to minimize the effect of the process while minimizing the dialysate con-
temperature on the beer avor resulting in high- sumption. In addition, in order to minimize CO2
quality low alcohol beer, which can be bottled losses, the feed pressure should be selected close
after nal sterile ltration. to the CO2 saturation pressure, and small amounts
Reverse osmosis is typically carried out in of the CO2 should be added to the dialysate
spiral wound modules, and the dealcoholization (Branyik et al. 2012).
Benzene and Cyclohexane Separation 157
Furthermore, osmotic distillation for beer the fermentation tank volume. In the past this beer
dealcoholization (Russo et al. 2013) and in the tank bottom would be considered lost since
pervaporation for aroma recovery can be consid- it could not be added to the main beer stream.
ered as part of the beer dealcoholization process However, using microltration with tubular
(del Olmo et al. 2012), but so far these processes ceramic membranes or plate-and-frame modules B
are not commercialized. with polymeric membranes, it is possible to
recover a high-quality beer which can be blended
with the main beer stream going toward clarica-
References tion and storage. Using microltration mem-
branes with pores of 0.40.5 mm, it is possible to
Branyik T, Silva DP, Baszczynski M, Lehnert R, Almeida e retain the yeast and allow the beer to pass the
Silva JB (2012) A review of methods of low alcohol
membrane without any major impact on the qual-
and alcohol-free beer production. J Food Eng
2012(108):493506 ity of the beer. In this process, the yeast in the tank
del Olmo A, Blanco CA, Palacio L, Prdanos P, Hernndez bottoms is concentrated from 7 % to 10 % DM to
A (2012) Setting up of a method of pervaporation for approx. 20 % DM and thus 5070 % of the beer in
improving alcohol-free beer. Euromembrane, London,
the tank bottoms can be recovered. The yield of
pp 2327
Gebhardt W (2001) Weltforum der Wein- und recovered extract and alcohols can be further max-
Saftbereitung. F&S Filtrieren Separieren 15(5):229 imized if dialtration water is added. Recovering
Russo P, Liguori L, Albanese D, Crescitelli A, Di Matteo beer from tank bottoms can increase the output of
M (2013) Investigation of osmotic distillation tech-
an average brewery by 1 % of its annual produc-
nique for beer dealcoholization. CEt Chem Eng Trans
32:17351840 tion or 24,000 hl extra for a brewery with an
annual output of 2 million hl (Lipnizki 2005).
The amortization for a microltration beer recov-
ery unit is typically 12 years.
Beer Maturation
References
Frank Lipnizki
Alfa Laval Copenhagen, Sborg, Copenhagen, Lipnizki F (2005) Optimisation and integration of mem-
brane processes in the beverage industry, 10. Aachener
Denmark
Membran Kolloquium, 16.-17.03.2005: preprints. -
Aachen, Mainz, 2005. - 3-86130-409-0.
In the beer production process, the claried and

cooled wort from the brewhouse is transferred
together with yeast to the fermentation tanks for Benzene and Cyclohexane
the primary fermentation which converts the fer- Separation
mentable sugar into alcohol and CO2. The
resulting green beer undergoes a second Tadashi Uragami
fermentation beer maturation under addition Organization for Research and Development of
of sugar or fresh yeasted wort. During the matu- Innovative Science and Technology (ORDIST),
ration, the residual fermentable sugars in the Kansai University, Suita, Osaka, Japan
green beer will be converted to alcohol and the
beer will be saturated with CO2. After the fermen-
tation, the beer is claried and stored in the bright Benzene/Cyclohexane Separation
beer cellar. Remaining in the fermentation tanks
are tank bottoms a mixture of settled yeast cells In the petrochemical industry, the separation of
and beer which are equal approx. 1.52.0 % of benzene (Bz) and cyclohexane (Chx) is the most
158 Benzene and Cyclohexane Separation
feed mixture can preferentially permeate through

the membrane. In a PV process, differences
between the solubility and diffusivity of the mix-
ture components in the polymer membrane and
the relative volatility of the permeants determine
the permeability and selectivity (Binding
et al. 1961; Aptel et al. 1974). In general, PV
exhibits the following characteristics (Uragami
2006, 2010):
1. Selective transport across the nonporous mem-

brane is achieved by a three-step process of
solution, diffusion, and evaporation.
Benzene and Cyclohexane Separation,
Fig. 1 Principle of pervaporation (PV)
2. Because the driving force for permeation is the
vapor pressure for each component rather than
total system pressure, this method is effective
important and difcult processes. Chx is produced for separation of organic liquid mixtures with
in benzene hydrogenation units under Ni or Pd high osmotic pressure.
catalyst. The unreacted Bz is remained in the 3. PV can be applied to the separation and con-
reaction mixture and must be removed to produce centration of mixtures that are difcult to sep-
pure Chx. The separation of benzene and Chx is arate by distillation. For example, it is useful
very difcult by a conventional distillation for the separations of azeotropic mixtures,
because close-boiling point mixtures are formed close-boiling point mixtures, and structural
in the entire range of their compositions. At pre- isomers.
sent, azeotropic distillation and extractive distilla- 4. PV can be used for the removal of certain
tion are applied to this separation. These components in equilibrium reactions.
distillations, however, are complex and need 5. Polymer membrane compaction, a frequent
high energy consumption. In the industry of Chx problem in high-pressure gas separations, is
production, the conventional Bz/Chx separation not encountered in PV because the feed pres-
processes are strongly required. Therefore, many sure is typically low.
studies have investigated the PV properties of
polymer membranes for Bz/Chx separation. A side-chain liquid-crystalline polymer (LCP)
Pervaporation (PV) is a promising membrane was synthesized by the addition of mesogenic
technique for the separation of organic/organic monomer to poly(methylsiloxane) with a Pt cata-
mixtures, as PV can be used to separate organic lyst. When Bz/Chx mixtures were permeated
liquid mixtures such as azeotropic and close- through the LCP membranes by PV at various
boiling point mixtures. The separation mechanism temperatures, the permeation rate increased with
in PV is not based on only relative volatility of increasing benzene concentration in the feed solu-
components in distillation but on the difference in tion and permeation temperature. Although the
sorption and diffusion properties of the feed LCP membranes exhibited Bz/Chx selectivity,
substances. the mechanism responsible for the permeation
Figure 1 illustrates the principles of PV. In this and separation of the Bz/Chx mixtures was differ-
separation process, when a liquid mixture is fed to ent in the glassy, liquid-crystalline state versus the
the upstream side of a polymer membrane and the isotropic state of the LCP membranes. These
downstream side is evacuated, a component in the results suggest that the Bz/Chx selectivity was
Benzene and Cyclohexane Separation 159
moderately inuenced by the change in LCP of 6FDA with various diamines. To obtain high
membrane structure (i.e., a state transformation). permeability as well as high selectivity, a combi-
The balance between the orientation of the nation of the diamines 2,3,5,6-tetramethyl-1,4-
mesogenic groups and the exibility of the silox- phenylene diamine (4MPD), 4,40 hexauoro-
ane chain is very important with respect to perme- isopropylidiene dianiline (6FpDA), and B
ability and Bz/Chx selectivity (Inui et al. 1997, 3,5-diaminobenzoic acid (DABA) as monomers
1998). When benzene/cyclohexane, toluene/ with a crosslinkable group was used. Cross-
cyclohexane, and o-xylene/cyclohexane mixtures linking is necessary to prevent undesirable swell-
were subjected to PV through an LCP membrane ing effects, which generally occur with non-cross-
in the liquid-crystalline state, the permeation rate linked polyimides, especially if high benzene con-
increased with increasing temperature and the centrations are reached during PV. The degree of
LCP membrane exhibited selectivity for the aro- cross-linking was kept constant at 20 %, whereas
matic hydrocarbons. The permeation rate and the ratio of the diamine monomers 6FpDA and
selectivity of the LCP membrane for each mixture 4MPD was varied. The PV experiments were
decreased with increasing molecular size of the performed at 60 C, using Bz/Chx mixtures with
aromatic hydrocarbon in the binary feed mixture benzene concentrations covering the entire con-
(Inui et al. 1998). When Bz/Chx mixtures were centration range. All of the cross-linked polymers
permeated through nematic and smectic side- had excellent chemical and thermal stability in the
chain liquid-crystalline polymer (n- and s-LCP) PV experiments. In all cases, conditioning of the
membranes under various conditions during PV, membrane samples with pure benzene was a suit-
the n- and s-LCP membranes exhibited Bz/Chx able pretreatment to enhance the permeation rate
selectivity. The selectivity of the n-LCP mem- without decreasing the Bz/Chx selectivity signif-
brane changed from solubility-selectivity con- icantly. For the most promising membrane mate-
trolled to diffusion-selectivity controlled upon rial, 6FDA-4MPD/DABA of 4:1 cross-linked
the state transformation of the membrane, induced with ethylene glycol, the PV experiments
by an increase in the permeation temperature. In were performed with a benzene/cyclohexane
contrast, the selectivity of the s-LCP membrane feed mixture of 50/50 (w/w) over a temperature
was governed by diffusion selectivity regardless range between 60 C and 110 C to determine the
of the state of this membrane. At low permeation effect of temperature on the separation
temperatures, the n-LCP membrane in the liquid- characteristics.
crystalline state exhibited lower permeability but
higher selectivity than the s-LCP membrane.
However, at high permeation temperatures, the
relationship between the permeability and References
Bz/Chx selectivity of the n-LCP and s-LCP mem-
Aptel P, Cuny J, Jozenfonvice J, Morel G, Neel J (1974)
branes in the liquid-crystalline state was reversed. Liquid transport through membranes prepared by
These results are a result of differences in the grafting of polar monomer onto poly(tetrauor-
chemical and physical structure of the n-LCP oethylene) lms. II. Some factors determining
pervaporation rate and selectivity. J Appl Polym Sci
and s-LCP membranes (Inui et al. 1998).
18:365
The PV properties of a series of cross-linked Binding RC, Lee RJ, Jennings JF, Mertic EC (1961) Sepa-
4,40 -hexauoro-isopropylidene dianhydride ration of liquid mixtures by pervaporation. Ind Eng
(6FDA)-based copolyimide membranes for the Chem 53:47
Inui K, Miyata T, Uragami T (1997) Permeation and sep-
separation of Bz/Chx mixtures were investigated
aration of benzene/cyclohexane mixtures through
(Ren et al. 2001). The glassy, highly rigid liquid-crystalline polymer membranes. J Polym Sci
copolyimides were obtained by polycondensation Part B Polym Phys 35:699
160 Benzyl-Modified Cellulose
Inui K, Miyata T, Uragami T (1998a) Effect of permeation The benzylation of cellulose can lead to
temperature on permeation and separation of a ben- several useful applications. For example, in
zene/cyclohexane mixture through liquid-crystalline
polymer membranes. J Polym Sci Part B Polym Phys blood hemodialysis, membrane ltration is a
36:281 key process for the replacement of renal func-
Inui K, Miyata T, Uragami T (1998b) Permeation and tion in uremic patients. The hydroxyl groups
separation of binary organic mixtures through a in cellulose can activate complement function,
liquid-crystalline polymer membrane. Macromol
Chem Phys 199:589 which is an undesirable process in hemodialy-
Inui K, Okazaki K, Miyata T, Uragami T (1998c) Effect of sis. A certain amount of the substitution of
mesogenic groups on characteristics of permeation and hydroxyl groups with benzyl groups, being
separation for benzene/cyclohexane mixtures of side- known as synthetically modied cellulose
chain liquid-crystalline polymer membranes. J Membr
Sci 143:93 (SMC), can help minimize the complement
Ren J, Standt-Bickel C, Lichtenthaler R (2001) Separation activation, by balancing the hydrophilic and
of aromatics/aliphatics with crosslinked 6FDA-based hydrophobic domains in the cellulose structure
copolyimides. Sep Purif Technol 223:31 (Hoenich et al. 1997).
Uragami T (2006) Polymer membranes for separation of
organic liquid mixtures. In: Yanpolskii Y, Pinnau I, Another application of benzylation of cellulose
Freeman BD (eds) Materials science of membranes is the fabrication of natural ber composites by
for gas and vapor separation. Wiley, Chichester, the surface modication of cellulose through the
pp 355372 route shown in Fig. 1 (Lu et al. 2004). The ber
Uragami T (2010) Selective membranes for purication
and separation of organic liquid mixtures. In: composites exhibit excellent thermal formability
Drioli E, Georno L (eds) Comprehensive membrane and mechanical properties, enabled by the benzyl
science and engineering, volume 2 membrane opera- cellulose skin layer and natural cellulose core
tions in molecular separations. Elsevier, Amsterdam, (Zhang et al. 2005).
pp 273324
It should be noted that the benzylation of
cellulose, as mentioned above, can produce a
heterogeneous structure with only the skin of
Benzyl-Modified Cellulose the cellulose structure being modied.
A homogeneous benzylation of cellulose has
Hongyang Ma, Benjamin S. Hsiao and been achieved when ionic liquids were used as
Benjamin Chu the solvent for cellulose (Gericke et al. 2012).
Department of Chemistry, Stony Brook N-alkylpyridinium chlorides can be used to
University, Stony Brook, NY, USA dissolve cellulose and then react with benzyl
groups, resulting in benzyl cellulose with a
high degree of substitution. It is possible that
Benzyl-modied cellulose contains covalently the surface modication of brous mats with
bonded benzyl groups in the cellulose structure, benzyl cellulose can be achieved by surface
typically produced by etherication of hydroxyl coating and thereby forming a non-covalently
groups (Bowry and Rintelen 1998) shown in bonded modication approach (Ma et al.
Fig. 1. 2014).
Benzyl-Modified
Cellulose, Cl O
OH
Fig. 1 Synthetically
HO OH HO OH
modied cellulose from O O
O O O
O
benzyl-modied cellulose O NaOH HO OH O
HO OH
by etherication OH n OH n
Bilberry Aroma Recovery by Pervaporation 161
References Inc., California (EEUU), and Agrotechnology &

Food Innovations (A&F) in the Netherlands
Bowry SK, Rintelen TH (1998) Synthetically modied (Willemsen et al. 2004), and theoretical studies
cellulose (SMC) a cellulosic hemodialysis membrane
assessing the recovery and concentration of aroma
with minimized complement activation. ASAIO J 44:
M579M583 compounds by means of PV have demonstrated B
Gericke M, Fardim P, Heinze T (2012) Ionic liquids -prom- the viability of this technology at industrial scale
ising but challenging solvents for homogeneous deriv- (Karlsson et al. 1998; Lipnizki et al. 2002a, b;
atization of cellulose. Molecules 17:74587502
Trifunovic et al. 2006). During the PV, the
Hoenich NA, Wofndin C, Stamp S, Roberts SJ, Turnbull
J (1997) Synthetically modied cellulose: an alterna- aroma compounds contained in a liquid or feed
tive to synthetic membranes for use in haemodialysis? phase permeate selectively through a dense
Biomaterials 18:12991303 hydrophobic membrane due to the chemical
Lu X, Zhang M, Rong M, Yue D, Yang G (2004) The
potential gradient caused by a reduction of the
preparation of self-reinforced sisal ber composites.
Polym Polym Compos 12:297307 component partial pressure in the permeate side.
Ma HY, Hsiao BS, Chu B (2014) Functionalized The aroma compounds are preferentially
electrospun nanobrous microltration membranes transported through the membrane according to
for removal of bacteria and viruses. J Membr Sci
their afnity to the membrane material and their
452:446452
Zhang M, Rong M, Lu X (2005) Fully biodegradable natu- mobility in the membrane matrix, which is the
ral ber composites from renewable resources: all-plant hypothesis of the most accepted mass transport
ber composites. Compos Sci Technol 65:25142525 mechanism in PV known as solution-diffusion.
Bilberry (Vaccinium myrtillus L.) is a species
of shrub and is native from Europe, northwest
Bichromal Particles Synonym: regions of North America, and Northern Asia
Two-Colored Particles (Morton and Macleod 1990). The international
markets demand of bilberrys extract has been
Encapsulation Application steadily growing in the recent years (Foster and
Blumenthal 2012). The major aroma compounds
present in crushed bilberry were identied by gas
chromatography-mass spectroscopy (GC-MS).
Bilberry Aroma Recovery by The results are shown in Fig. 1. Trans-Hex-2-en-
Pervaporation 1-ol was characterized as one of the impact avor
of the bilberry (von Sydow and Anjou 1969).
Immaculada Ortiz and Nazely Diban
Department of Chemical and Biomolecular Membrane Material and Module
Engineering, University of Cantabria, Santander, Configurations
Cantanbria, Spain The most typical organophilic membrane mate-
rials employed in PV of aroma compounds in the
literature are polydimethylsiloxane (PDMS) and
Introduction polyoctylmethylsiloxane (POMS). POMs mem-
branes present higher enrichment factors for
Pervaporation (PV) is a membrane-based technol- aroma compounds than PDMS albeit not signi-
ogy largely studied for aroma compound recovery cantly different. Nonetheless, PDMS membranes
and concentration in the food industry (Pereira are largely more studied due to their stability
et al. 2006). At least two PV pilot plants for during operation and good properties of
aroma compound recovery have been built at processing (Karlsson and Trgardh 1993;
Membrane Technology and Research (MTR) Lipnizki et al. 1999; Urtiaga et al. 1999).
162 Bilberry Aroma Recovery by Pervaporation
kCounts
500
1 4
400
300
200
5
2
100 3
25 50 75 100 minutes
1540 3148 4747 6338 Scans
Bilberry Aroma Recovery by Pervaporation, (4) trans-Hex-2-en-1-ol, and (5) 1-hexadecanol (Adapted
Fig. 1 GC-MS chromatogram of the crushed bilberry: from (Garcia et al. 2008))
(1) trans-hexenal, (2) 2-butyl-1-octanol, (3) 1-hexanol,
With regard to the membrane module congu- (i) The feed composition (ethanol and aroma
rations, plate-and-frame is the most employed compound). The results are presented in
in literature for laboratory scale studies for Fig. 2. As expected, ethanol ux, JEtOH,
their simplicity and versatility for the evaluation increased linearly with higher ethanol con-
of different membrane materials. However, the tent in the feed, while in contrast, aroma
hollow bers membrane (HFM) module congu- compound partial ux, JHex, remained inde-
ration presents important advantages, such as pendent on ethanol concentration but
higher contact area per unit volume of membrane presented a linear dependency with its own
module and thus higher compactness of the equip- concentration in the feed solution. The major
ment and easier operation at industrial scale component ux, water, was independent
(Urtiaga et al. 1999, 2002; Tasselli et al. 2007). on both ethanol and aroma compound com-
position. The enrichment factors found for
Influence of Process Variables the aroma compound, bHex, varied from
Using a model ternary solution (water/ethanol/ 98 to 201.
trans-Hex-2-en-1-ol), the process variables that (ii) The feed temperature, studied in the range
exerted an important inuence on the PV perfor- 35 C < T < 50 C. The components par-
mance of the impact aroma compound (trans- tial uxes (water, ethanol, and trans-Hex-2-
Hex-2-en-1-ol) when using a PDMS HFM mod- en-1-ol) were found to follow an Arrhenius
ule with the characteristics presented at (Garcia dependency with temperature as plotted in
et al. 2008) were: Fig. 3.
Bilberry Aroma Recovery by Pervaporation 163
[Hex] (kg.m3)
a b 0 0.1 0.2 0.3
20 2.00 1.8
18 1.80 1.6
16 1.60 1.4 B
JEtOH (g.m2.h1)
JHex (g.m2.h1)
JHex (g.m2.h1)
14 1.40
1.2
12 1.20
1.0
10 1.00
8 0.80 0.8
6 0.60 0.6
4 0.40 0.4
2 0.20 0.2
0 0.00 0.0
0 2 4 6 0 5 10
[EtOH] (%v/v)
XHex g Hex P0 Hex (Pa)
Bilberry Aroma Recovery by Pervaporation, [Hex] = 0.1 kg.m3, T = 50 C and Re = 772. (b) Effect
Fig. 2 (a) Effect of ethanol concentration in the feed on of trans-hex-2-en-1-ol partial pressure in the feed on its
the permeation ux of the ethanol in () binary water/ permeation ux at 50 C, Re = 772, and [EtOH] = 1%v/v
ethanol and () ternary water/ethanol/Hex mixtures, and (Adapted from (Garcia et al. 2008))
(~) ux of trans-hex-2-en-1-ol in ternary mixtures
Bilberry Aroma Recovery 4.0

by Pervaporation,
3.5
Fig. 3 Effect of the feed
temperature on the partial 3.0
uxes of the components Hex
2.5
when Re = 772,
[Hex] = 0.1 kg.m3, and 2.0 EtOH
[EtOH] = 1%vv (Adapted
In Ji
1.5 Water
from (Garca et al. 2008))
1.0
0.5
0.0
0.5
1.0
3.05 3.10 3.15 3.20 3.25 3.30
1000/T(K1)
In the present application and working under Kanani et al. 2003), similar to the membrane
laminar ow rate conditions, 203<Re<772, thickness herein studied, 196 mm.
the aroma compound transfer ux was indepen-
dent on the feed ow rate. This indicated PV of Multicomponent Systems
that the concentration polarization phenomena PV of multicomponent systems of the bilberry
in the feed boundary layer was negligible in characteristic aroma avor was performed theo-
comparison with the aroma compound mass retically by means of simulation considering the
transfer resistance in the membrane which nor- composition in Table 1 and commercial PDMS
mally occurs with high membrane thickness, HFM modules. The membrane area considered
i.e., 127 and 160 mm (Peng and Liu 2003; was Am = 0.0056 m2 and the membrane
164 Bilberry Aroma Recovery by Pervaporation
Bilberry Aroma Recovery by Pervaporation, Table 1 Feed composition (found in Hirvi and Honkanen 1983), and
simulation results of the permeate composition, enrichment factors, and partial fluxes in the bilberry multicomponent
system after PV with a PDMS HFM module at 30 C and a membrane area Am = 0.0056 m2 (Diban et al. 2008)
Feed composition, CFi Permeate composition, CPi
Component (mg.kg1) (mg.kg1) bi Ji (g.m2h1)
Trans-Hex-2-en-1-ol 0.01 1.21 120.6 3.78 105
n-Hexanol 0.02 4.74 236.9 14.9 105
Trans-Hex-2-en-1-al 0.06 2.78 46.3 8.70 105
Linalool 0.004 0.20 49.3 0.62 105
Phenyl acetaldehyde 0.003 0.02 5.5 0.05 105
Benzyl alcohol 0.08 0.33 4.2 1.05 105
Cis-Hex-3-en-1-ol 0.06 1.64 27.4 5.15 105
Ethanol 7800 106,934 13.7 2.51
Water 992,200 893,057 0.9 20.9
7.00E-05 180
160
6.00E-05
140
JHex(g m2 h1)
5.00E-05
120
4.00E-05 100
bHex
3.00E-05 80
2.00E-05 60
40
1.00E-05
20
0.00E-00 0
0 5 10 15 20 0 5 10 15 20
104 (m) 104 (m)
Bilberry Aroma Recovery by Pervaporation, thickness, d. T = 30 C, [Hex] = 105 kg.m3,

Fig. 4 Change of the partial ux of trans-Hex-2-en-1-ol, [EtOH] = 1%v/v, Re ~6 105, Am = 0.0056 m2
JHex, and enrichment factor, bHex, with the membrane (Adapted from Diban et al. 2008)
thickness, d, was 1.96 104 m. Further details Future Directions

of the simulation conditions can be found in
(Diban et al. 2008). The results of the simulation 1. Experimental assessment of the behavior of
are collected in Table 1. There is a clear change in real solutions to evaluate (i) the interaction
the component prole from the feed solution to between different molecules during PV and
the permeate aroma compound concentrate. (ii) the aroma proles in the permeate.
Moreover, the theoretical evaluation of the 2. Integration of a sensory evaluation of the qual-
inuence of the membrane thickness, d, on the ity of the aroma compound concentrate that,
performance of the PV of trans-Hex-2-en-1-ol independently of the aroma prole, is mainly
was performed. The results are plotted in Fig. 4. affected by the presence of the impact aroma
From these results, it is extracted that a com- compounds in the permeate.
mercially available membrane with a thickness 3. Evaluation of other membrane materials that
of 1.48 104 m led to a compromise situation reduce water mass transfer and enhance the
with important impact aroma compound enrich- aroma compound enrichment factors with
ment factors bHex ~90 and adequate uxes JHex the minimum ux reduction, such
~3 105 g.m2.h1. as POMS.
Bimetallic Electrocatalyst for Fuel Cells 165
References von Sydow E, Anjou K (1969) The aroma of bilberries

(Vaccinium myrtillus L.) I. Identication of volatile
Diban N, Urtiaga A, Ortiz I (2008) Recovery of key com- compounds. Lebensm Wiss Technol 2:7881
ponents of bilberry aroma using a commercial Willemsen JHA, Dijkink BH, Togtema A (2004)
pervaporation membrane. Desalination 224:3439 Organophilic pervaporation for aroma isolation-
Foster S, Blumenthal M (2012) The adulteration of industrial and commercial prospects. Membr Technol B
commercial bilberry extracts. J Am Bot Counc 2:510
96:6473
Garca V, Diban N, Gorri D, Keiski R, Urtiaga A, Ortiz
I (2008) Separation and concentration of bilberry
impact aroma compound from dilute model solution
by pervaporation. J Chem Technol Biotechnol
Bimetallic Electrocatalyst for Fuel
83:973982 Cells
Hirvi T, Honkanen E (1983) The aroma of blueberries.
J Sci Food Agric 34:992998 Dmitri Bessarabov
Kanani DM, Mikhade BP, Balakrishnan P, Singh G,
Faculty of Engineering, DST HySA Infrastructure
Pangarkar VG (2003) Recovery of valuable tea aroma
components by pervaporation. Ind Eng Chem Res Center of Competence, North-West University,
42:69246932 Potchefstroom, South Africa
Karlsson HOE, Trgrdh G (1993) Pervaporation of dilute
organic-waters mixtures. A literature review on model-
ing studies and applications to aroma compound recov-
ery. J Membr Sci 76:121146 General Requirements for Fuel Cells
Karlsson HOE, Trgrdh G, Olsson J (1998) The perfor-
mance of pervaporative units: process simulations. Sep Most of the known today fuel cell systems (e.g.,
Sci Technol 33:16291652
automotive, portable applications, backup power,
Lipnizki F, Hausmanns S, Ten PK, Field RW, Laufenberg
G (1999) Organophilic pervaporation: prospects and etc.) have common requirements that include
performance. Chem Eng J 73:113129 reduction in costs, improvement in performance
Lipnizki F, Olsson J, Trgrdh G (2002a) Scale-up of and durability, increase in tolerance toward impu-
pervaporation for the recovery of natural aroma com-
rities in the feed fuel, etc. (Vielstich et al. 2003a;
pounds in the food industry. Part I: simulation and
performance. J Food Eng 54:183195 Debe 2012).
Lipnizki F, Olsson J, Trgrdh G (2002b) Scale-up of
pervaporation for the recovery of natural aroma com-
pounds in the food industry. Part II: optimization and
Pt-based Electrocatalyst: Key
integration. J Food Eng 54:197205
Morton ID, Macleod AJ (1990) Food avours part C. The Component
avor of fruits. Elsevier, Amsterdam
Peng M, Liu SX (2003) Recovery of aroma compounds The platinum-based electrocatalyst used in PEM
from dilute model blueberry solution by pervaporation.
(proton-exchange membrane, also called polymer
Food Eng Phys Prop 68:27062710
Pereira CC, Ribeiro CP, Nobrega R, Borges CP electrolyte membrane) fuel cells, PAFC
(2006) Pervaporation of volatile aroma compounds (phosphoric acid fuel cell), and DMFC (direct
from fruit juices. J Membr Sci 274:123 methanol fuel cell) is one of the key components
Tasselli F, Cassano A, Drioli E (2007) Ultraltration of
and directly inuences performance, durability,
kiwifruit juice using modied poly(ether ether
ketone) hollow bre membranes. Sep Purif Technol costs, etc. More specically, in the PEM fuel
57:94102 cell, the oxygen reduction reaction (ORR) occur-
Trifunovi O, Lipnizki F, Trgrdh G (2006) The inuence ring at the cathode is known to have slow kinetics,
of process parameters on aroma recovery by hydropho-
leading to large cathodic overpotential losses
bic pervaporation. Desalination 189:112
Urtiaga A, Gorri ED, Beasley JK, Ortiz I (1999) Mass under typical operating conditions. In the case of
transfer analysis of the pervaporation separation of specic application of PEM fuel cells, when
chloroform from aqueous solutions in hollow ber hydrogen is produced from a reformate, an addi-
devices. J Membr Sci 156:275291
tional requirement such as tolerance of the anode
Urtiaga A, Gorri ED, Ortiz I (2002) Mass-transfer model-
ing in the pervaporation of VOCs from diluted solu- electrocatalyst toward the traces of CO in the feed
tions. AIChE J 48:572580 fuel stream must also be met. In order to reduce
166 Bimetallic Electrocatalyst for Fuel Cells
costs of the fuel cells, an increased mass activity electrooxidation of CO. Many metals have been
of the cathode electrocatalyst is required that considered for modifying the activity of the plat-
would allow a decrease in the amount of the inum catalyst, but only a few of them (Ru, Ir, Sn,
electrocatalyst in a fuel cell device. Mo, etc.) lead to improved performances. The
most studied bimetallic electrocatalyst is the fam-
ily of Pt/Ru alloys, which enhance greatly the
Advanced Electrocatalysts rate of oxidation of many alcohols (methanol,
ethanol, etc.).
A search for better ORR electrocatalysts for PEM
fuel cells that meet advance requirements has
resulted in the development of various Pt-based Current Trends
bimetallic compounds (these can include alloys or
intermetallic systems). It is believed that the Research and development in the area of bime-
improved performance (i.e., activity enhance- tallic and ternary electrocatalysts is ongoing.
ment) of bimetallic alloys as electrocatalysts The trend is to make use of unique nanoscale
could be explained by the structural modication structural effects that can be observed in struc-
of Pt 5d orbital, coordination number of Pt, and turally modied alloys (e.g., de-alloyed
modication in adsorption of oxygenated species Pt-based catalysts) and intermetallic systems
from the electrolyte to the Pt or the alloying metal. to enhance greatly ORR. These include novel
For example, alloying results in a lattice contrac- structures observed in thin lm extended sur-
tion, leading to a more favorable Pt-Pt distance for faces (Stamenkovic et al. 2007; Debe 2012) as
the dissociative adsorption of oxygen (Vielstich well as a multilayer Pt-skin surface (Wang
et al. 2003b; Zhang 2008). et al. 2011, 2012).
Pt+Cr, Pt+Ni, and Pt+Co bimetallic
electrocatalysts have been demonstrated to show
two- to threefold increase in their activities for
ORR. The activation energy for oxygen reduction
References
was also shown to be lower than that in
Debe MK (2012) Electrocatalyst approaches and
Pt. However, it is also well recognized that these challenges for automotive fuel cells. Nature 486:4351
bimetallic electrocatalysts are subject to signi- Stamenkovic VR, Mun BS, Arenz M, Mayrhofer KJJ, Lucas
cant degradation rates (sintering/dissolution) as CA, Wang G, Ross PN, Markovic NM (2007) Trends in
electrocatalysis on extended and nanoscale Pt-bimetallic
well as carbon-support corrosion under specic
alloy surfaces. Nat Mater 6:241247
voltage cycling conditions that are common for Vielstich W, Lamm A, Gasteiger HA (eds) (2003a)
automotive drive cycles during fuel cell startup Handbook of fuel cells, vol 1, Fundamentals and survey
and shutdown. of systems. Wiley, New York
Vielstich W, Lamm A, Gasteiger HA (eds) (2003b)
A lot of efforts have also been made during the
Handbook of fuel cells, vol 2, Electrocatalysis. Wiley,
last two decades to improve anode electrocatalyst New York
in fuel cells that are designed to use hydrogen Wang C, Chi M, Li D, Strmcnik D, van der Vliet D,
obtained by means of reforming other fuels. In Wang G, Komanicky V, Chang K-C, Paulikas AP,
Tripkovic D, Pearson J, More KL, Markovic NM,
that case an anode electrocatalyst is required for
Stamenkovic VR (2011) Design and synthesis of
hydrogen oxidation reaction to take place in the bimetallic electrocatalyst with multilayered Pt-skin sur-
presence of CO, as well as to improve perfor- faces. J Am Chem Soc 133:1439614403
mance of the methanol fuel cells. The focus has Wang C, Markovic N, Stamenkovic VR (2012) Advanced
platinum alloy electrocatalysts for the oxygen reduc-
been on the development of bimetallic platinum-
tion reaction. ACS: Catalysis 2:891898
based electrocatalysts to reduce the amount of Zhang J (ed) (2008) PEM fuel cell electrocatalysts and
adsorbed CO and/or to improve performance of catalyst layers. Springer, London
Bioactive Compounds 167
For example, phenolic compounds and avo-

Bioactive Compounds noids from vegetable sources such as cereals,
legumes, olives, and fruits exhibit antioxidant,
Carmela Conidi anti-inammatory properties, and different stud-
Institute on Membrane Technology, National ies have reported protective associations between B
Research Council of Italy, ITM-CNR, Rende, avonoids and cardiovascular disease (CVD) and
Italy cancer.
Hydroxytyrosol, one of the most important
phenolic compounds in olives, olive oil, and
Bioactive compounds are extranutritional con- olive mill wastewaters, presents a strong antioxi-
stituents that typically occur in small quantities in dant and anticancer properties.
foods and can be shown to have an effect on Resveratrol, found in red wine, has antioxi-
human health. dant, antithrombotic, and anti-inammatory prop-
Bioactive compounds are also referred as erties and inhibits carcinogenesis. Lycopene, a
nutraceuticals, a term that reects their existence potent antioxidant carotenoid in tomatoes and
in the human diet and their biological activity. fruits, presents a protection against prostate and
They consist of a wide range of chemical com- other cancers and inhibits tumor cell growth in
pounds with different structures, physiological animals.
activities, and molecular mass between 200 and Organosulfur compounds in onions and mono-
1000 Da (Pennington 2002). terpenes in citrus fruits have shown
Typical bioactive compounds include: anticarcinogenic actions in experimental models,
as well as cardioprotective effects.
Carotenoids (a-carotene, b-carotene, lyco- Bioactive compounds are the main source of
pene, lutein) new drugs, functional food, and food additives. In
Flavonoids (avanols/avans, isoavones/ fact, more than 80 % of food active compounds
isoavonoids, avanones, avones, and more than 30 % of drugs are produced from
anthocyanins) bioactive natural products and the annual growth
Phenolic acids (cinnamic acid, caffeic acid, rate of natural drugs is 20 %.
chlorogenic acid, p-coumaric acid, ferulic As a result, the research of appropriate tech-
acid, citric acid, ellagic acid, vanillic acid, niques for the separation of these compounds
hydroxytyrosol, tyrosol, oleuropein) from complex matrix has attracted much attention
Plant sterols (beta-sitosterol, campesterol, phy- in the last years. Generally, the separation of target
tosterol, saponin, squalene, sigmasterol, stanol) compounds includes extraction and purication
Resveratrol steps.
Probiotics Several techniques have been proposed for the
Omega-3 fatty acids (a-linoleic acid) separation of bioactive compounds including sol-
Proteins vent extraction, irradiation-assisted extraction,
Organosulfur compounds (allyl/diallyl suldes) ultrasound-assisted extraction, heat treatment,
Indoles enzyme-assisted extraction, supercritical uid
Monoterpenes (limonene, perillyl alcohol) extraction, resin-based extraction, and chroma-
tography purication techniques (Azmir
Bioactive compounds have been intensively et al. 2013).
studied in order to evaluate their effects on These separation methods are characterized by
human health, and numerous methodological some drawbacks, such as the degradation of the
approaches have been implemented to clarify compounds of interest due to high temperatures
their biological effect and mechanism of action. and long extraction times (as in solvent
168 Bioactive Compounds
Bioactive Compounds, Table 1 Typical membrane operations for the separation of bioactive compounds from
different sources
Bioactive Membrane
compounds Examples Sources processes
Flavonoids Quercetin, catechin Apple juice, tea, red wine MF, UF, NF,
RO
Narirutin, naringin, hesperidin, Citrus juices MF, UF, NF,
neohesperidin RO
Apigenin, luteolin Artichoke wastewaters UF, NF, RO
Cyanidin, cyanin, myrtillin Red orange and pomegranate juices MF, UF, NF,
RO
Carotenoids a-Carotene, b-carotene Palm oil, carrot juice MF, NF
Lycopene, lutein Carrot, kiwifruit, and tomato juices MF, UF
Phenolic acids Caffeic acid, chlorogenic acid, p-coumaric Apple juice, olive mill, and MF, UF, NF,
acid, ferulic acid artichoke wastewaters RO
Citric acid Citrus juices MF, UF, NF,
RO
Ellagic acid Blackberry juices UF, NF
Tyrosol, hydroxytyrosol oleuropein, gallic Olive mill wastewaters MF, UF, NF,
acid RO
Proteins a-Lactalbumin, b-lactalbumin, bovine Whey, milk MF, UF, NF
serum albumin (BSA)
Immunoglobulines Egg yolk UF
Phycocyanin Marine organisms (microalgae) UF, NF, RO
Omega-3 fatty Linoleic acid Fish-processing wastes MF, UF, NF
acids
extractions) and health-related risks. In addition, concentration of different classes of bioactive

they involve high capital costs. compounds from a complex mixture and in the
Recently, membrane technology has attracted recovery of intermediate compounds (Cassano
attention as an alternative molecular separation and Drioli 2014).
technology due to its advantages such as high Membrane operations in sequential design are
efciency, simple equipment, easy scale-up, and particularly suitable alternative for recovery/con-
low energy consumption. In general, membrane centration of bioactive phenolic compounds from
separations can operate under mild conditions of fruit juices and food-producing wastewaters. MF
temperature, pressure, and shear stress, therefore and UF are typically applied for preliminary treat-
preserving the biological activity of the com- ment while the purication and concentration
pounds to be recovered and the properties of the steps are usually performed by NF and RO
original product; they do not require any extrac- membranes.
tion mass agents such as solvents, avoiding prod- The separation of bioactive proteins from milk
uct contaminations and the need for subsequent and marine sources with UF and NF membranes
purication. results in high selectivity due to the large molec-
Particularly, pressure-driven membrane tech- ular weight differences between the target
nologies such as microltration (MF), ultraltra- compounds.
tion (UF), nanoltration (NF), and reverse Typical examples of membrane operations
osmosis (RO) are extremely efcient for the used for the separation of bioactive compounds
separation, fractionation, purication, and are summarized in Table 1.
Bioadhesion to Membrane Surface 169
References (Woodley 2001). Bioadhesion to membrane sur-

face is often related with biofouling of membrane
Azmir J, Zaidul ISM, Rahman MM, Sharif KM, surface and thus receives much attention in mem-
Mohamed A, Sahena F, Jahurul MHA, Ghafoor K,
brane technology, particularly in biomedical
Norulaini NAN, Omar AKM (2013) Techniques for
extraction of bioactive compounds from plant mate- membrane separation. B
rials: a review. J Food Eng 117:426436 In blood-contact membrane separation, for the
Cassano A, Drioli E (2014) Integrated membrane opera- blood-incompatible membrane materials, the
tions in the food production. De Gruyter, Berlin
platelets in the blood tend to accumulate and
Pennington JAT (2002) Food composition database for
bioactive food components. J Food Comp Anal adhere onto membrane surface due to a biological
15:419434 immune reaction. As a result, thrombus formation
may form and further result in some serious com-
plications. Therefore, the blood-contact mem-
branes such as hemodialysis membranes often
need be modied to prevent the adhesion of plate-
Bioadhesion to Membrane Surface lets on membrane surface during blood ltration.
In membrane science, the experiment of platelet
Li-Ping Zhu adhesion on membrane surface is usually
MOE Key Laboratory of Macromolecular employed to evaluate the hemocompatibility of a
Synthesis and Functionalization, Department blood-contact membrane. In biomedical applica-
of Polymer Science and Engineering, Zhejiang tion, the adhesion of microbes on membrane sur-
University, Hangzhou, Zhejiang, China face often leads to the bacterial infection of
patients and the deterioration of membrane mate-
rials. Therefore, the control of microbial infection
is a very important issue in biomedical membrane
Bioadhesion to membrane surface refers to the separation.
special interaction between biological organisms
to membrane surfaces. The term bioadhesion
generally describes the adherence of materials
(natural or synthetic) to biological surfaces or the Cross-References
rm attachment of biological substances (such as
microorganisms, proteins, cells, biomolecules, Adhesion to Membrane Surfaces
etc.) on material surfaces (Huang et al. 1999). As Adsorption onto Membrane Surfaces
an interface phenomenon, bioadhesion is similar Biofouling
to conventional adhesion, except for the special
characteristics of biological organisms and sur-
References
faces. As such, this term covers the adhesive
properties of both synthetic components and the Huang YC, Zhang ES, Ying K (1999) Progress of bacterial
natural surfaces (such as cells) (Jacek et al. 1999). adhesion on biomaterials. Foreign Med Sci (Biomed
Bioadhesion also includes the use of bioadhesive Eng Fascicle) 22(3):148153
Jacek D, Robert K, Jarosaw K, Hanna MK (1999) Cell
(natural polymeric materials, such as proteins and
adhesion to polymeric surfaces: experimental study and
carbohydrates) to stick two surfaces together simple theoretical approach. J Biomed Mater Res
(Palacio and Bhushan 2012). In medicine and 47:234242
biology, sometimes bioadhesion is also used as a Palacio ML, Bhushan B (2012) Bioadhesion: a review of
concepts and applications. Phil Trans R Soc
synonym of mucoadhesion, which refers to the
A 370:23212347
process whereby synthetic and natural macromol- Woodley J (2001) Bioadhesion: new possibilities for drug
ecules adhere to mucosal surfaces in the body administration? Clin Pharm 40:7784
170 Bioartificial Liver
therapeutic and toxicological effects of various

Bioartificial Liver drugs and chemicals on hepatic metabolism
(Salerno et al. 2013). The feasibility of BALs as
Simona Salerno good and achievable model systems for in vitro
Institute on Membrane Technology, National studies represents an alternative and/or comple-
Research Council of Italy, ITM-CNR, Rende, mentary methods or procedure to in vivo animal
Italy experimentation. BALs as in vitro model systems
allow the reduction of the number of laboratory
animals and provide a faster and more cost-
Synonyms effective way of analysis. The use of human hepa-
tocytes is the only system for a rapid, accurate,
Articial liver; BAL; Biohybrid Articial Liver and highly specic evaluation of effects taking
(BAL) account of the species differences in liver
response, substrate specicity, and enzyme induc-
tion between man and animals.
Critical issues in the design of a BAL are the
Bioartificial Liver conguration, the cell sourcing and density, the
choice of the material for cell adhesion, the oxy-
Bioarticial liver is a system that consists of func- genation and the gas exchanges, the mass trans-
tional liver cells (hepatocytes) supported by an port of nutrients and catabolites, the scale-up, and
articial cell culture environment. In particular it the regulation and safety (Stevens et al. 2014).
incorporates hepatocytes into a bioreactor, or bio- A successful BAL requires favorable culture con-
chemical reactor, in which cells are embedded ditions to maintain cell viability, functionality, and
and/or immobilized, cultured, and induced to per- differentiation in the time. As liver cell sources,
form all the liver-specic functions of the in vivo primary hepatocytes, either human or of xenoge-
native liver: detoxication, biotransformation, neic derivation, human-immortalized hepatoma
excretion, synthesis, and modulation. cell lines, hepatic progenitors, and intrahepatic
BALs have two major uses: in clinic as or extrahepatic stem cells could be used.
bioarticial liver support system (BALSS) to pro- Primary human hepatocytes represent the gold
vide a short-term support for liver failure patients standard for their clinical safety and high liver-
and in in vitro experimentation as a model system specic activity; however, their limited availabil-
for therapeutic and toxicological studies. ity makes them unsuitable for clinical application
BALs were developed as a promising approach and large clinical studies. Most systems are using
for the treatment of patients with acute and fulmi- xenogeneic cells of porcine origin which are more
nant hepatic failure. They dont offer a long-term easily available; nonetheless the metabolic com-
nor a permanent substitute for an individuals liver. patibility is not assured, and the transfer of xeno-
They support liver functions for a relatively short geneic virus and immunological reactions
period of time and are intended to be used either as represent a potential risk. Human-immortalized
a bridge to a transplant or as a temporary support to hepatoma cell lines have the advantages of unlim-
allow the native or post-transplant liver to recover. ited availability and good practicability, but they
BALs with homogeneous and stable in vivo- have low liver-specic activity and cancer-like
like microenvironment are further employed as cell metabolism. A large number of studies have
valid tools to investigate the basic biology of utilized human liver-derived stem cells including
hepatic tissue development, to study liver dis- fetal liver stem cells (hepatoblasts) and liver stem
eases, to develop therapeutic strategies for drug- cells (oval cells) to generate primary hepatocytes;
screening applications, to assess hepatic pathways because of their rarity within the liver tissue, until
of drug biotransformation, and to evaluate both today it is impossible to expand and differentiate
Bioartificial Liver 171
them to a sufcient cell mass. Other studies have xenogeneic implants and protection of cells from
induced extrahepatic stem cells, embryonic or shear stress damages. Furthermore, the selective
adult stem cells, to differentiate into hepatocyte- transport properties of membranes ensure a selec-
like cells; although the pluripotent stem cells may tive transport and an adequate mass transfer of
be a suitable alternative to hepatocytes, their clin- molecules, nutrients, metabolic products, and B
ical use is limited by ethical concerns, for the rst endogenous and exogenous toxins as well as gas
of them, and the risk for tumor development for exchange from/to patients blood or plasma and
both. A promising approach is in the use of to/from the cell compartment.
induced pluripotent stem cells (iPSC) that are The use of at membrane bioreactors allows
adult stem cells reprogrammed for their for control of the internal ow distribution and
pluripotency by genetic manipulations (Lee and perfusion of all cells under a stable oxygen and
Cho 2012). nutrient gradient. The main drawbacks of a at
In an effort to recreate the interactions between conguration are the difculties to build a system
parenchymal and non-parenchymal cells, hepato- with a sufcient cell density and an optimal ratio
cytes have been co-cultured with several liver between the low surface area and the volume of
non-parenchymal cells such as endothelial cells, the system. Most devices tested clinically employ
broblasts, stellate cells, and Kupffer cells, show- hollow ber membrane cartridges. In this cong-
ing remarkable liver-like structure and functions. uration, the cells are compartmentalized, in the
Several types of BALs were developed by intralumen or in the extralumen of hollow bers,
using cell suspensions, alginate beads, and isolated from blood or plasma by the semi-
microcarriers, microcapsules, organoids, porous permeable membrane with a proper molecular
spheroids, collagen sandwich, and matrices. weight cutoff. However, the main drawbacks of
To maintain hepatocyte functions and polarity, a hollow ber conguration concern inherent
its advantageous to mimic the matrix surrounding physical limitations including mass transport
the hepatocytes in the sinusoid of the natural liver. across the ber wall and the hepatocyte mass, a
To this purpose three-dimensional matrices or limited total diffusion surface area, and an
sandwiches conguration are the best environ- increased diffusion distance (Bao et al. 2011).
ments for the long-term maintenance of viability, The design of a successful BAL conguration
functionality, and differentiation state. could be optimized by modelling study in order
Polymeric membranes in at sheet and in hol- to predict and verify the performance of the pro-
low ber congurations are widely used as posed conguration in terms of mass transport and
supporting materials for the design and develop- cellular reaction rates.
ment of BALs.
Membranes offer a substratum for adhesion of
anchorage-dependent cells like the hepatocytes. Cross-References
Morphological characteristics (e.g., pore size,
porosity, roughness) and physicochemical and Articial Liver, Membrane Operations
mechanical properties of the membrane surface Bioarticial Liver Support System (BLSS)
(e.g., wettability, surface free-energy parameters, Bioarticial Organs and Tissue
Youngs modulus) affect cell adhesion and the Bioarticial Organs, Membrane Operations of
maintenance of cell viability, functionality, and Biohybrid Articial Liver (BAL)
differentiation in the time. Membranes with ade- Biohybrid Articial Liver (BAL) Systems
quate molecular weight cutoff (MWCO) ranging Biohybrid Membrane Systems
from 70,000 to 100,000 Da represent a selective Membrane Articial Organs
barrier for the compartmentalization of cells in Membrane Bioarticial Organs
nano- and microstructured environment ensuring Membrane Bioreactors for Cell Growth
immunoisolation and immunoprotection in allo- Tissue Engineering, Membrane Operations of
172 Bioartificial Liver (BAL) Devices
References hemoltration, and plasmapheresis, bioarticial

devices incorporate a functional cell component.
Bao J, Fisher J, Nyberg SL (2011) Liver regeneration and This latter is constituted by living and metaboli-
tissue engineering. In: Bernstein HS (ed) Tissue engi-
cally active hepatic cells, or hepatocytes, which
neering in regenerative medicine, Stem cell biology and
regenerative medicine. Humana Press, New York, are able to fully replace the specic functions of
pp 315332 the in vivo native liver, including biotransforma-
Lee JS, Cho SW (2012) Liver tissue engineering: recent tion, synthesis, and modulation in addition to
advances in the development of a bio-articial liver.
detoxication and excretion. The liver is a vital
Biotechnol Bioproc Eng 17:427438
Salerno S, De Bartolo L, Drioli E (2013) Membrane sys- organ with a wide range of functions in synthesis,
tems in liver regenerative medicine. In: De Bartolo L, detoxication, and regulation; hence, its failure
Bader A (eds) Biomaterials for stem cell therapy: state constitutes a life-threatening condition that in
of art and vision for the future. CRC Press/Taylor &
most cases can be overcome by orthotopic liver
Francis Group, Boca Raton, pp 3764
Stevens KR, Schwartz RE, Ng S, Shan J, Bhatia SN transplantation. Owing the shortage of donor
(2014) Hepatic tissue engineering. In: Lanza R, organs, a variety of alternative treatments was
Langer R, Vacanti JP (eds) Principles of tissue engi- developed to replace some of the livers function
neering, 4th edn. Elsevier Academic Press, London,
in case of liver failure. BLSSs are expected to
Waltham, San Diego, pp 951986
increase spontaneous recovery or at least improve
patient survival until transplantation is possible.
BLSSs employ hepatocytes as implantable
systems or extracorporeal systems (De Bartolo
Bioartificial Liver (BAL) Devices and Bader 2002).
Implantable systems are tissue-engineered
Bioarticial Liver Support System (BLSS) constructs proposed as potential permanent
BLSSs for in vivo liver regeneration. They include
hepatocytes on coated microcarrier beads, within
microencapsulated gel droplets, within biode-
Bioartificial Liver Support System gradable polymeric membranes, or as spheroid
(BLSS) hepatocyte aggregates. They however require a
prolonged period for intraperitoneal engraftment
Simona Salerno and vascularization, and their potential to provide
Institute on Membrane Technology, National long-term hepatic support is still in a very prelim-
Research Council of Italy, ITM-CNR, Rende, inary and experimental phase and untested yet in
Italy clinical trials.
Extracorporeal systems provide metabolic
support for a relatively short time and for the
Synonyms treatment of acute liver failure and patients for
which the orthotopic liver transplantation is the
Bioarticial liver (BAL) devices; Biohybrid arti- only effective therapy. Such systems serve as a
cial liver systems; Cell-based liver support sys- bridge to solid organ transplantation or as a tem-
tems; Extracorporeal BALs; Temporary liver porary support to allow the native and/or post-
support systems transplant liver to regenerate and/or recover prop-
erly. They are not invasive to the body: a shunt
Bioarticial Liver Support System (BLSS) is a veno-venous or arterio-venous is required to con-
bioarticial liver that supports liver functions. In nect the patient to the device. Blood is removed
contrast to purely articial or nonbiological from the vein and circulated through the device
liver support systems in which detoxication pro- where it is processed by living liver cells and then
cess is based on the principle of adsorption and returned to the patient using a dialysis-type pro-
ltration, such as hemodialysis, hemoperfusion, cedure. In some devices the blood plasma
Bioartificial Liver Support System (BLSS) 173
removed from the patient is separated from the compartment. Furthermore, membranes offer a
other constituents by plasma separators before it large surface area for adhesion of anchorage-
circulates trough the device. Extracorporeal dependent cells like the hepatocytes, affecting
BLSSs include hepatocytes in suspension as the maintenance of cell viability, functionality,
aggregates/spheroids or in adhesion to matrices and differentiation in relation to membrane mor- B
and/or membranes into a bioreactor and may be phological characteristics (e.g., pore size, poros-
recharged with fresh hepatocytes for both a con- ity, roughness) and physicochemical and
tinuous or intermittent use. mechanical properties (e.g., wettability, surface
In the design of BAL devices, critical aspects free energy parameters, Youngs modulus).
have to be taken into account such as the cell Several extracorporeal systems have under-
source and density, the culture condition and the gone extensive animal testing and are currently
long-term maintenance of well-differentiated investigated in the early stages of human clinical
hepatocyte functions, the oxygenation and the trials (Zhao et al. 2012). A series of studies
mass transport, and the conguration and the showed that BLSS devices reduced mortality in
scaling-up of the systems. acute liver failure cases.
Several types of extracorporeal BLSSs with
different congurations have been developed: in
packed bed, at membrane sheet, and hollow ber Cross-References
membrane systems (Morelli et al. 2010; Stevens
et al. 2014). The at membrane bioreactors allow Articial Liver, Membrane Operations
a high-density cell culture under sufcient oxy- Bioarticial Liver
genated conditions; however, they present some Bioarticial Organs and Tissue
disadvantages such as the complex up-scaling of Bioarticial Organs, Membrane Operations of
large at plates. Most of the BLSSs are composed Biohybrid Articial Liver (BAL)
of hollow ber cartridges with different polymeric Biohybrid Articial Liver (BAL) Systems
membranes. Hepatocytes are loaded in the Biohybrid Membrane Systems
extralumen space or injected into the lumen of Membrane Articial Organs
the hollow ber membranes. Some of them com- Membrane Bioarticial Organs
prise more independent but interwoven capillary Membrane Bioreactors for Cell Growth
systems in a three-dimensional network for Tissue Engineering, Membrane Operations of
medium inow, medium outow, and oxygen/car-
bon dioxide exchange (Sauer and Gerlach 2002).
References
In both the implantable and extracorporeal sys-
tems, membranes serve as immunoprotective bar- De Bartolo L, Bader A (2002) Development of a hybrid
riers against host defenses. Immunological liver-support device. Minim Invasive Ther Allied
rejection is reduced by means of membranes Technol 11:123134
with molecular weight cutoff (MWCO) of Morelli S, Salerno S, Piscioneri A, Campana C, Drioli E,
De BL (2010) Membrane bioreactors for regenerative
70,000400,000 Da which act as immunologic medicine: an example of the bioarticial liver. Asia-Pac
barrier between the patients blood and hepato- J Chem Eng 5:146159
cytes; thus, xenogeneic or allogeneic implants Sauer IM, Gerlach JC (2002) Modular extracorporeal liver
may be used without an immunosuppression support. Artif Organs 26:703706
Stevens KR, Schwartz RE, Ng S, Shan J, Bhatia SN
therapy. (2014) Hepatic tissue engineering. In: Lanza R,
The selective transport properties of mem- Langer R, Vacanti JP (eds) Principles of tissue engi-
branes ensure a selective transport and an ade- neering, 4th edn. Elsevier Academic Press, London,
quate mass transfer of molecules, nutrients, Waltham, San Diego, pp 951986
Zhao LF, Pan XP, Li LJ (2012) Key challenges to the
metabolic products, endogenous and exogenous development of extracorporeal bioarticial liver sup-
toxins, as well as gas exchange from/to the port systems. Hepatobiliary Pancreat Dis Int
patients blood or plasma and to/from the cell 11:243249
174 Bioartificial Organs and Tissue
autologous primary cells have the advantage to

Bioartificial Organs and Tissue avoid immuno-rejection and/or transmission of
infective agents; however, only some of them
Simona Salerno present a high proliferative capacity. The use of
Institute on Membrane Technology, National embryonic stem cells (ESC) is very attractive
Research Council of Italy, ITM-CNR, Rende, though limited by ethical issues and their tumor-
Italy igenic capacity. The isolation of adult (ASC) or
mesenchymal (MSC) stem cells offers good alter-
native for their ability to activate and differentiate
Synonyms into multiple tissue cell types. Induced pluripotent
stem cells (iPSC), which are ASC reprogrammed
Biofabrication; Engineered organs and tissue; for their pluripotency by genetic manipulations,
Tissue engineering and regenerative medicine have been recently described as a new source of
cells to regenerate tissues.
Biomaterials and scaffolds provide the appro-
Bioartificial Organs and Tissue priate biochemical and biomechanical cues for cell
attachment and proliferation by mimicking the
Articial organs and tissues composed of cells extracellular matrix (ECM) for cells. Due to their
supported by engineered extracellular matrices interaction with living cells, they must be biocom-
such as biomaterials or scaffolds in a suitable patible, nontoxic, and pharmacologically inert.
culture environment. These constructs are pro- These biomaterials can be permanent or biodegrad-
duced by applying principles from different able and made of synthetic or natural polymers or a
established disciplines like engineering, chemis- composite network of both. The natural biological
try, biochemistry, biology, and medicine, with the ECM, in addition to mechanical integrity, contrib-
purpose to repair, replace, preserve, or improve utes in the development, maintenance, and regen-
the functions of damaged organs or tissues and as eration of tissues, and it is itself a dynamic structure
a possible alternative to the growing demand for actively remodeled by the cells with which it inter-
their transplantation. Besides their in vivo appli- acts. Taking into account this consideration, the
cation, human organs and tissue could be used as design characteristics of biomaterials and scaffolds
model systems for drug screening and testing and are major challenges for materials scientists and
as valid tools to investigate the basic biology in should be considered at a molecular chemical
in vitro research. level. Thus, new strategies are focusing on the
The new and functional living organs and tis- development of interactive biomaterials that, by
sue are fabricated using living cells, which are cell-material interactions, are able to activate, stim-
usually associated with a matrix or scaffold to ulate, and trigger cellular specic responses like
guide the tissue and organ development, in a met- cell proliferation and/or differentiation of embry-
abolic and physic microenvironment under tightly onic and/or progenitor cells, driving the self-tissue
controlled conditions, such as bioreactors, for formation and organization into a 3D construct.
their maturation step. ECM-specic molecular structures or specic
The source of cells must be easily accessible, receptors and anchorage sites, as well as growth
with a high proliferative capacity and ability to and differentiation factors, could be incorporated
differentiate into the required cell type. Cells can into cell support components to more closely rep-
be obtained from a fragment of adult tissue licate the native environment of the cells
(primary cells) or from multipotent stem cells. (De Bartolo et al. 2012). Composites of synthetic
They could be of autogeneic, allogeneic, or xeno- or natural polymers with bioactive ceramics or
geneic origin accordingly as they are from the certain glasses can be designed to yield materials
same organism, from a donor of the same species, with a wide range of strengths and porosities for the
or from a different species, respectively. The engineering of hard tissues (Furth and Atala 2014).
Bioartificial Organs and Tissue 175
After the cell seeding onto scaffolds, the mat- potentially functional organ constructs. With this
uration of an organ- or tissue-engineered construct strategy, it has been possible to obtain organs such
takes place in bioreactors. Such systems facilitate as the heart, liver, trachea, and lungs (He and
the reproducible growth and production of Callanan 2013). The so-called organ printing is
bioarticial constructs under tightly controlled based on multicellular self-assembly by a B
conditions, overcoming the limited seeding ef- computer-aided bioprinting technology. 3D
ciency and poor transport of nutrients, oxygen, printers combined with image-processing soft-
and wastes of the static cell culture conditions. ware, operating with bioink particles of spheroidal
The control of mass transport and gas exchange is cell aggregates and biopaper of biocompatible
particularly required for the development of highly ECM hydrogels, are able to form and build com-
perfused and oxygenated bioarticial organs, such plex 3D living structures with complete control of
as the liver and kidney. On the contrary, low oxy- the internal structure, including a vascular tree
gen tension is essential to improve the functions of (Marga et al. 2012).
engineered cartilage or to mimic the normal micro-
environment or niche for stem cells. Further-
more, bioreactors may be used to enhance tissue
formation through mechanical stimulation of the Cross-References
bone, cartilage, blood vessels, and skeletal and
cardiac muscles. In particular, a pulsatile ow is Bioarticial Liver
helpful for the maturation of blood vessels, while a Bioarticial Liver Support System (BLSS)
mechanical stretch strengthens engineered muscles Bioarticial Organs, Membrane Operations of
(Furth and Atala 2014). Biodegradable Polymer for Membranes
Scaffold-based tissue engineering has led to Biohybrid Membrane Systems
signicant results in the reconstruction of various Membrane Articial Organs
tissues. Small parts of bioarticial tissues (i.e., Membrane Bioarticial Organs
skin, cartilage, bone, blood vessels, corneas, uri- Membrane Bioreactors for Cell Growth
nary bladder, left mainstem bronchus, and tra- Tissue Engineering, Membrane Operations of
chea) have been successfully developed in
experimental animals and approved either for References
human use or in clinical trials (Vacanti 2010; de
Mel et al. 2012). De Bartolo L, Leindlein A, Hofmann D, Bader A, de
Currently, tissue engineering only offers the Grey A, Curcio E, Drioli E (2012) Bio-hybrid organs
possibility of recovering lost function, while the and tissues for patient therapy: a future vision for 2030.
Chem Eng Process Process Intensif 51:7987
formation of whole organs and tissues similar to de Mel A, Seifalian AM, Birchall MA (2012) Orchestrating
the natural ones is still not a reality because of the cell/material interactions for tissue engineering of sur-
difculty in providing vascularization, innerva- gical implants. Macromol Biosci 12:10101021
tion, and the same signals of the extracellular Furth ME, Atala A (2014) Tissue engineering: future per-
spectives. In: Lanza R, Langer R, Vacanti JP (eds)
matrix. Principles of tissue engineering, 4th edn. Elsevier
Recently, novel approaches that do not rely on Academic Press, pp 83123
articial scaffolds have been pursued. The most He M, Callanan A (2013) Comparison of methods for
promising ones utilize matrices of decellularized whole-organ decellularization in tissue engineering of
bioarticial organs. Tissue Eng Part B Rev 19:194208
organs or bioprinting-based technologies. The Marga F, Jakab K, Khatiwala C, Shepherd B, Dorfman S,
xenogeneic whole-organ decellularization allows Hubbard B, Colbert S, Gabor F (2012) Toward
the production of complex three-dimensional engineering functional organ modules by additive
extracellular matrix bioscaffolds of the entire manufacturing. Biofabrication 4:022001. doi:10.1088/
1758-5082/4/2/022001, Epub 2012 Mar 12
organ with preservation of the intrinsic vascular Vacanti J (2010) Tissue engineering and regenerative med-
network. These bioscaffolds can then be icine: from rst principles to state of the art. J Pediatr
recellularized by autologous cells to create Surg 45:291294
176 Bioartificial Organs, Membrane Operations of
Moreover membranes, exposing cells to an ade-

Bioartificial Organs, Membrane quate perfusion and oxygenation, are able to guide
Operations of cellular differentiation and functions (Salerno
et al. 2013a).
Simona Salerno The rst concept of membrane-based
Institute on Membrane Technology, National bioarticial organs was based on the development
Research Council of Italy, ITM-CNR, Rende, of implants containing living xenogeneic or allo-
Italy geneic cells enclosed and/or encapsulated in a
selective membrane barrier for cell transplanta-
tion. In such compartmentalized system, mem-
Synonyms branes act as a selective barrier for immune
isolation of cells. They admit the unhindered pas-
Biohybrid membrane systems; Engineered sage of oxygen and small molecules required for
organs; Organotypic membrane systems; Regen- cell metabolism and the release of bioactive com-
erative medicine ponents to the host. In particular, the selective
transport properties of membranes ensure the
transport and an adequate mass transfer of nutri-
Bioartificial Organs, Membrane ents, metabolic products, growth factors, endoge-
Operations nous and exogenous toxins as well as gas
exchange. In the meantime, membranes with ade-
Articial organs realized by taking advantage of quate molecular weight cutoff (MWCO) ranging
membrane technology and operations. They con- from 70,000 to 100,000 Da restrict the passage of
sist of functional cells supported by polymeric larger cytotoxic agents of the immune defense
membranes as scaffold, in a suitable culture envi- system, allowing the immune isolation and
ronment in which cells are maintained and immune protection of cells in allogeneic and
induced to perform all the specic functions of xenogeneic implants. The use of membranes as a
the in vivo native organ. In most cases they incor- selective barrier eliminates the need of chronic
porate cells into membrane bioreactors. immune suppression in the host and allows the
Organ level tissue engineering requires a bio- use of allogeneic and xenogeneic cells, thus
mimetic microenvironment that provides sub- avoiding the cell-sourcing constraints. Further-
strates for cell maintenance and proliferation and more, the compartmentalization of cells in nano-
signaling cues directing morphogenesis and cell and micro-structured environment ensures and
differentiation toward tissue-specic structures protects cells from shear stress damages in mem-
and phenotype. The choice of biomaterials with brane bioreactors (Drioli and De Bartolo 2006).
high biostability and biocompatibility and with Recent advances in mesenchymal stem cell biol-
specic physicochemical, mechanical, and mor- ogy and in their expansion and differentiation
phological properties on the basis of the targeted have made possible the use of autologous cell
tissues or organs is crucial for successful source to avoid the immuno rejection drawback.
bioarticial constructs. In bioarticial organs, polymeric membranes
Progress in membrane technology and in mem- not only act as selective and immune protective
brane operations in understanding and controlling barriers, but they also play a crucial role in cell
micro- and nano-properties and selectivity in trans- adhesion and cell interaction mimicking the extra-
port phenomena currently offer new interesting cellular matrix (ECM) with which cells interact.
opportunities for the design of bioarticial organs. A proper cell-material interaction could trigger
Membranes in at sheet and in hollow ber cell growth, migration, proliferation, differentia-
congurations serve to house cells, supporting tion, and functional activation, thus allowing the
their growth and adhesion and directing their reor- organization of cells in a three-dimensional archi-
ganization in a three-dimensional architecture. tecture and the maintenance of cell viability and
Bioartificial Organs, Membrane Operations of 177
functionality. At the same time, an inappropriate motility, and apoptosis. The signals that are actu-
cell-material interaction could trigger ally responsible for dictating tissue pattern are
programmed cell death or apoptosis. often mechanical in nature, such as compression
New biofabrication approaches to create and on the bone, shear stress on blood vessels, and
develop complex organotypic and organomimetic tension on muscles. The choice of materials with B
systems reckon with molecular, architectural, and specic mechanical properties is strictly related to
biophysical basis of tissue principles which are the target tissue or organs: the bone (hard tissues)
tightly related to developmental biology and mor- necessitates stiff matrices; tendons, ligaments,
phogenesis. Recent advances in the molecular cell skin, brous tissues, muscles, nerves, and blood
biology of morphogenesis, as well as in stem cell vessels (soft tissues) need more elastic and exi-
biology, will aid in the design principles and ble scaffolds. Since both embryogenesis and mes-
architecture for tissue engineering and regenera- enchymal differentiation toward the formation of
tion (Ingber 2014). Organs are made of tissues and new organ rudiments take place under mechanical
tissues are made of cells. To made and tailor tension and compression, differentiation of
bioarticial organs is necessary to start from tissue embryonic and mesenchymal stem cells is
engineering and, therefore, from cell-material strongly inuenced by substrate stiffness. On
interactions. polymeric membranes, an enhanced tensile mod-
The choice of ideal scaffold materials, opti- ulus and strength and a high porosity are able to
mum cell source, and well-dened tissue culture promote capillary development by endothelial
conditions is required for the realization of safe cells, thus favoring the vascularization in
and condent bioarticial substitutes for clinical organotypic membrane systems (Salerno
application. et al. 2011).
Membranes are eligible scaffolds for tissue To tailor the material bioactivity and to more
engineering and bioarticial organs. The majority closely replicate the native and natural microen-
of membranes for the realization of organ mem- vironment of the cells, membranes and biomate-
brane systems are manufactured from homoge- rials could be modied by incorporating
neous polymer solutions by phase inversion. biomolecules and/or ECM-specic molecular
This technique is a very versatile process for the structures. They could also include structural
synthesis of membranes from different polymers, units for the targeted controlled release of bioac-
also biodegradable, and with a wide variety of tive molecules as growth factors or cytokines to be
modulated morphological, physicochemical, released at the right time and at the right place and
mechanical, and transport properties depending to mimic natural signaling pathway. Moreover,
on the target bioarticial organ. the design of biodegradable membranes with a
Porous membranes with specic pore size and programmed and predictable biodegradation rate
porosity inuence cell adhesion and the transport consistent with the tissue formation rate is useful
and/or diffusion of nutrients and metabolites, to for implantable bioarticial constructs. The
and from cells. Morphological and physicochem- research strategies for the next 20 years will
ical surface properties, such as roughness, wetta- focus on the implantation of bioarticial tissues
bility, and free energy parameters, affect cell and the induction of the in situ regeneration and
adhesion and consequently the cell shape either self-repairing of damaged organs and tissues
indirectly, by controlling the adsorption of pro- (De Bartolo et al. 2012).
teins, or directly, by guiding cell spreading and Although the engineering of fully functional
cytoskeleton rearrangements. Topographical bioarticial organs for transplantation is not yet a
cues, including grooves, ridges, wells, and reality for the difculty in providing vasculariza-
nodes, drive the cell orientation and migration tion and innervation, bioarticial tissues with
and the three-dimensional architecture of the increasing complexity reach clinical application
growing bioarticial organ. Mechanical proper- nowadays. Advances in biomaterials and in stem
ties play in control of cell growth, differentiation, cell biology have made possible the realization of
178 Bioartificial Synthetic Membrane
tissues and organotypic bioarticial systems for Salerno S, Campana C, Morelli S, Drioli E, De Bartolo
clinical as well as basic research applications. L (2011) Human hepatocytes and endothelial cells in
organotypic membrane systems. Biomaterials
Membrane systems supporting and stimulating 32:88488859
cell differentiation and proliferation have been Salerno S, De Bartolo L, Drioli E (2013a) Membrane
involved in the biofabrication of active substitutes systems in liver regenerative medicine. In: De
in static and dynamic conditions (Salerno Bartolo L, Bader A (eds) Biomaterials for stem cell
therapy: state of art and vision for the future. CRC
et al. 2013a, b). The dynamic culture models Press/Taylor & Francis Group, Boca Raton, pp 3764
such as the membrane bioreactors may simulate Salerno S, Piscioneri A, Morelli S, Al-Fageeh MB,
the in vivo complex physiological environment Drioli E, De Bartolo L (2013b) Membrane bioreactor
through the uid dynamics modulation and the for expansion and differentiation of embryonic liver
cells. Ind Eng Chem Res 52:1038710395
selective membrane transport, ensuring an ade-
quate mass transfer of nutrients and metabolites.
In such microenvironment a constant turnover of
tissue culture medium augments the gas and nutri-
ent exchanges, which together with the complete Bioartificial Synthetic Membrane
uid dynamics control ensures the long-term
maintenance of cell viability and functions. Mem- Zhiwei Wang
brane bioarticial organs could be used as Department of Environmental Engineering,
implantable or extracorporeal devices for tissue College of Environmental Science and
loss or organ failure and for the designing of Engineering, Tongji University, Shanghai, China
in vitro model systems for drug screening and
diagnostic applications. Examples of membrane
bioarticial organs are the membrane bioarticial Synonyms
liver and the bioarticial liver support systems.
Articial biological membrane; Biological and
synthetic membrane; Synthetic biomembrane
Cross-References
Introduction
Bioarticial Liver
Bioarticial Liver Support System (BLSS)
Bioarticial synthetic membrane is a synthetic
Bioarticial Organs and Tissue
enclosing or separating membrane that functions
Biodegradable Polymer for Membranes
as a selectively semipermeable barrier of articial
Biohybrid Membrane Systems
organs such as articial livers, articial kidneys/
Membrane Articial Organs
dialyzers, and articial lungs/blood oxygenators.
Membrane Bioarticial Organs
Bioarticial synthetic membrane can separate the
Tissue Engineering, Membrane Operations of
contents (e.g., cells) and components of the recip-
ients immune system (Nyberg et al. 2003). The
References importance of a bioarticial synthetic membrane
is suggested by the rapid inux of ultraltrate
De Bartolo L, Leindlein A, Hofmann D, Bader A, de components and the reservation of functional
Grey A, Curcio E, Drioli E (2012) Bio-hybrid organs
and tissues for patient therapy: a future vision for 2030.
cells (e.g., human primary renal proximal tubule
Chem Eng Process Process Intensif 51:7987 cells of bioarticial kidneys), with the inhibition
Drioli E, De Bartolo L (2006) Membrane bioreactor for cell of undesired back diffusion into the recipients
tissues and organoids. Artif Organs 30:793802 system (Ni et al. 2010).
Ingber DE (2014) Mechanobiology, tissue development
and organ engineering. In: Lanza R, Langer R,
Generally, bioarticial synthetic membranes
Vacanti JP (eds) Principles of tissue engineering, that are in contact with biological uids should
4th edn. Elsevier Academic Press, pp 309322 prevent any type of infection and immune
Biobutanol Recovery from Biomass Fermentation Broth 179
response, thrombus formation, and other biologi- References

cal response that could affect the properties of the
uid (Stamatialis et al. 2008). The approaches to Kawakami H (2008) Polymeric membrane materials for
articial organs. J Artif Organs 11(4):177181
the development of biocompatible membranes
Ni M, Teo JCM, Ibrahim MS, Zhang KY, Tasnim F, Chow
include three categories: (1) optimization of syn- PY, Zink D, Ying JY (2010) Characterization of mem- B
thetic polymer, (2) modication of material sur- brane materials and membrane coatings for bioreactor
face, and (3) construction of biomimetic units of bioarticial kidneys. Biomaterials
32(6):14651476
membrane surfaces (Kawakami 2008). So far, a
Nyberg SL, Yagi T, Matsushita T, Hardin J, Grande JP,
variety of materials have been used for Gibson LE, Platt JL (2003) Membrane barrier of a
bioarticial synthetic membranes, including porcine hepatocyte bioarticial liver. Liver Transpl
polymethyl methacrylate, polydimethylsiloxane, 9(3):298305
Stamatialis DF, Papenburg BJ, Girones M, Saiful S,
polysulfone, polyethersulfone, poly(ethylene
Bettahalli SNM, Schmitmeier S, Wessling M (2008)
oxide), and poly(vinyl alcohol). Alternatively, Medical applications of membranes: drug delivery,
membrane surface properties can be modied to articial organs and tissue engineering. J Membr Sci
suppress immune response, and prevalent 308(12):134
Yoshida M, Langer R, Lendlein A, Lahann J (2006) From
methods are comprised of plasma etching, corona
advanced biomedical coatings to multi-functionalized
discharge, UV irradiation, and covalent attach- biomaterials. Polym Rev 46(4):347375
ment (Yoshida et al. 2006).
Nondegradability is an important factor for
bioarticial synthetic membranes. Unexpected
swelling and leaching can result in reduced frac-
ture strength and elastic modulus of the material, Biobutanol Recovery from Biomass
leading to static fatigue or crazing (Stamatialis Fermentation Broth
et al. 2008).
It is very important that bioarticial synthetic Wanqin Jin
membranes can have identical functions as the State Key Laboratory of Materials-Oriented
organs, which is closely related with membrane Chemical Engineering, Nanjing University of
structure. The structure parameters of membranes Technology, Nanjing, China
generally include thickness, pore size, porosity,
and so on. For instance, in articial kidneys, the
pore sizes should be well controlled in order to Biobutanol recovery from biomass acetone-
prevent leakage of useful proteins in blood such as butanol-ethanol (ABE) fermentation broth by
blood corpuscles and albumin (molecular weight pervaporation is a process that couples ABE fer-
69,000 Da) while removing harmful substances mentation with in situ pervaporation recovery
(Kawakami 2008). Microdomain structure is also acetone, butanol, and ethanol from fermentation
important for interaction of membranes with broth.
seeded cells. Differentiated epithelia of human During the past decade, there has been an
primary renal proximal tubule cells have increasing interest for the production of biofuels
been found on the optimal membrane materials from renewable resources because of growing
with/without coatings (Ni et al. 2010), and concerns about global warming and increasing
the concept of developing asymmetric mem- consumption of crude oil. Butanol, as compared
branes is suggested for bioarticial synthetic to ethanol, is less volatile and explosive, contains
membranes. more energy, and can easily mix with gasoline in
any proportion. In addition, butanol can be used
directly or blended with gasoline or diesel without
Cross-References any vehicle retrot, since the air to fuel ratio and
the energy content of butanol are close to gasoline
Biological Membrane (Qureshi and Ezeji 2008). Biobutanol is produced
180 Biobutanol Recovery from Biomass Fermentation Broth
Biobutanol Recovery from Biomass Fermentation Broth, Fig. 1 ABE fermentation-pervaporation-coupled process
from renewable resources (biomass) by acetone the pervaporation process is energy efcient and it
butanol ethanol (ABE) fermentation process. is harmless to the fermentation broth.
However, usually the maximum concentration of PV membranes can be made from either poly-
total solvents (ABE) do not exceed 20 g/L due to meric or inorganic materials, even both of them
the end-product inhibition on microbes, resulting (Qureshi and Blaschek 1999). The polymeric
in high energy cost to recover ABE from the dilute pervaporation membranes for extracting butanol
fermentation broth by distillation. Therefore, sev- from fermentation broth include polydimethyl-
eral in situ product-removal technologies, such as siloxane (PDMS) membranes, poly
adsorption, gas stripping, liquid-liquid extraction, [1-(trimethylsilyl)-1-propyne] (PTMSP) mem-
perstraction, pervaporation, and reverse osmosis, branes, poly(ether block amide) (PEBA) mem-
have been developed. These technologies could branes, liquid membranes, other modied
be coupled with ABE fermentation process to polymer membranes, as well as porous polypro-
reduce the effect of product inhibition and pylene (PP) membranes and polytetrauor-
improve the sugar utilization and solvent produc- oethylene (PTFE) membranes. Among them, the
tivity (Garca et al. 2011). commonly used PDMS membranes with good
Pervaporation (PV) is a membrane process for selectivity and stability are regarded as the most
liquid separation, in which a polymeric or inor- potential PV membranes for butanol recovery
ganic membrane usually serves the separating application. The typical inorganic membrane for
barrier. The application of PV for butanol recov- PV is hydrophobic zeolite membrane, such as
ery from fermentation broth is based on the selec- ZSM-5 and silicalite-1 membrane. The process
tive permeation of organic compounds through of biobutanol recovery from biomass ABE fermen-
the membrane when the nature of the membrane tation broth by pervaporation is also called ABE
is hydrophobic. PV is especially effective for the fermentation-pervaporation-coupled process; its
system in which the concentration of the targeted conventional ow chart is shown in Fig. 1 (Liu
species is low; for instance, the butanol content in et al. 2011). A pervaporation membrane module
the fermentation broth is very low. Furthermore, is connected to the fermentor, and the ABE solvent
Biocatalyst Recycling by Membrane Operations 181
can be selectively and continuously removed from technologies, the end-of-life purication tech-
the fermentation broth and then concentrated in the niques that were applied for the product streams
membrane downstream side. As result, the ABE are being replaced with integrated process solu-
concentration in the broth is kept at low level to tions in closed systems. Exhibiting high activities
facilitate the continuous fermentation process as and selectivity difcult to achieve with chemical B
well as enhance the solvent productivity. In addi- catalysts, biocatalysts have been regarded as ideal
tion, a microltration/ultraltration unit can be catalysts. The widespread use of biocatalysts in
alternatively installed before the pervaporation industrial processes is however largely impeded
module, in order to retain microbes and avoid the by their unfavorably high economic costs. The
biofouling of pervaporation membranes. recycling of the expensive biocatalysts is there-
fore an important goal in applied biocatalysis,
which would in turn promote the development of
References industrial enzyme-mediated processes.
Membrane ltration is an essential tool in the
Garca V, Pkkil J, Ojamo H, Muurinen E, Keiski biotechnological industry and appears to be par-
R (2011) Challenges in biobutanol production: how to
ticularly useful for the purication and concentra-
improve the efciency. Renew Sustain Energy Rev
15:964980 tion of proteins. Membrane technologies could
Liu G, Wei W, Wu H, Dong X, Jiang M, Jin W (2011) therefore be applied to achieve this goal of biocat-
Pervaporation performance of PDMS/ceramic compos- alyst recycling. The application of membrane
ite membrane in acetone butanol ethanol (ABE)
technologies in the eld of biocatalyst (i.e.,
fermentation-PV coupled process. J Membr Sci
373:121129 enzyme and whole cell)-mediated conversions
Qureshi N, Blaschek H (1999) Butanol recovery from could aid in achieving process simplication, as
model solution/fermentation broth by pervaporation: well as facilitating a recycling and reuse of the
evaluation of membrane performance. Biomass
biocatalysts, and in reducing the process costs and
Bioenergy 17:175184
Qureshi N, Ezeji T (2008) Butanol a superior biofuel could be a separation step for the components of
production from agricultural residues (renewable bio- the process efuent stream. The membrane reten-
mass): recent progress in technology. Biofuels Bioprod tion and enzyme recycling methods described
Bioren 2:319330
here are different from systems where the
enzymes are immobilized on or entrapped in a
membrane. The use of membrane technologies
for biocatalyst recycling in biochemical
Biocatalyst Recycling by Membrane processes however depends on a number of
Operations important features of the membranes employed.
Membrane processes can be differentiated on
Yamini Satyawali1, Ehiaze Augustine Ehimen2 the basis of the size and geometry of the particles
and Winnie Dejonghe2 which are aimed to be retained (and recycled), for
1
Separation and Conversion Technology, example, the use of microltration (MF), ultral-
Vlaamse Instelling voor Technologisch tration (UF), nanoltration (NF), and reverse
Onderzoek (VITO), Boeretang 200, Belgium osmosis (RO) techniques. With their ability to
2
Separation and Conversion Technology, Flemish retain the biocatalytic entities and macromole-
Institute of Technological Research (VITO), Mol, cules (i.e., whole cells and enzymes) with dimen-
Belgium sions between 8800 and 0.58 nm, UF and NF
membranes are most widely applied for biocata-
lyst recycling (Dijkstra et al. 2002).
Owing to current industrial demands for environ- Most UF and NF membranes are asymmetric
mentally friendly solutions and cleaner membranes, with the pore sizes on the solute side
182 Biocatalyst Recycling by Membrane Operations
Biocatalyst Recycling by (i) Products

Membrane Operations,
Fig. 1 Biocatalyst Membrane
membrane recycling
Feed
schemes via (i) dead-end Biocatalysts
ltration and (ii) Cross-ow
(loop) techniques
Membrane tube
(ii)
Filtration unit
Feed Products
of the membrane smaller than those on the per- polymers having elongated chains (Dijkstra
meate side, which in turn largely prevents mem- et al. 2002).
brane clogging. The membrane stability under Two major schemes can be applied for the
different operational conditions is also a very recycling of biocatalysts using membrane opera-
important factor which must be considered, since tions. These are the use of (i) dead-end ltration
it indicates the potential biocatalytic processes and (ii) cross-ow (loop) ltration as shown in
which the selected membranes can be suitably Fig. 1.
applied for. This includes a consideration of the With the dead-end ltration technique, the bio-
various possible interactions that the different catalyst is isolated in the reactor and is retained
reacting components and intermediates in the con- using UF or NF membranes. The reactants are
version process could have with the membrane continuously pumped into the reactor with the
surface. Unfortunately, data concerning such product and unreacted substrates permeating
aspects must still largely be obtained by empirical through the membrane for forward processing.
investigations, since extensive data for most However, the accumulation of the catalyst (i.e.,
membranes is still lacking and not widely concentration polarization) near the membrane
available. surface can occur using this method. Using the
The molecular weight cutoff (MWCO), which cross-ow technique, the potential catalyst con-
is dened as the molecular weight at which 90 % centration polarization issue can be prevented,
of the solutes are retained by the membrane, is since the reaction solution is continuously circu-
usually used as the main quantitative criterion for lated through the reactor. This ensures that the
the characterization of the membrane retention. It membrane is continuously swept clean by the
should however be remembered that the pore size uid cross-ow going past the surface which
distribution, charge effects, hydrophilicity, hydro- then minimizes the enzyme build up while
phobicity, and polarity (of the solvents) can also allowing the products and unreacted materials
greatly inuence the permeability and functioning pass laterally across the membrane.
of the membrane (Dijkstra et al. 2002). In addi-
tion, especially in the case of biocatalyst
recycling, the molecular shape and characteristics
are important factors when considering biocata- References
lyst retention using membranes. For example,
Dijkstra H, Van Klink G, Van Koten G (2002) The use of
globular proteins are more efciently retained by ultrasound and nanoltration techniques in homoge-
membranes compared to the retention of exible nous catalyst. Acc Chem Res 35:798810
Biocatalytic and Biochemical Membrane Reactor 183
modied enzymes (10100 kDa) (Prazeres and

Biocatalytic and Biochemical Cabral 1994).
Membrane Reactor The biocatalyst is usually present in the reac-
tor system in two forms: free or immobilized at
Yamini Satyawali1, Ehiaze Augustine Ehimen2 the membrane surface or inside the membrane B
and Winnie Dejonghe2 matrix pores. If the biocatalyst (i.e., enzyme) is
1
Separation and Conversion Technology, present in its free form, its immobilization can
Vlaamse Instelling voor Technologisch be achieved by conning the enzyme to one
Onderzoek (VITO), Boeretang 200, Belgium side of the membrane unit. This can be facili-
2
Separation and Conversion Technology, Flemish tated by size exclusion, electrostatic repulsion,
Institute of Technological Research (VITO), Mol, or enlargement via chemical or physical immo-
Belgium bilization techniques onto an intermediate sup-
port (i.e., inert proteins and gels). As for the
direct immobilization of the biocatalyst onto the
Synonyms membrane, this can be achieved by chemical
binding, physical adsorption, or electrostatic
Biochemical membrane reactors attraction.
Ideally the products resulting from the biocat-
A biocatalytic membrane reactor (also referred to alytic process should permeate through the
as a biochemical membrane reactor) is a system or membrane pores either via a concentration
device that combines and utilizes the separation gradient-induced diffusion or a pressure
capabilities of membranes and the catalytic activ- gradient-induced convection. This would in turn
ities of whole cell or enzyme biocatalysts in one facilitate a continuous removal of products from
unit to facilitate the biochemical conversion of the reaction media, a process which has been
substrates into desired products and the selective considered to be essential for biocatalytic mem-
removal of the products in a single process unit. brane reactor concepts. The selection of the mem-
The removal of the product components from the brane unit to be incorporated into the reactor
reaction site could in turn improve the conversion system must take into account the respective
of the substrates by favoring the product forma- sizes of the enzymes, substrates, and products, as
tion reaction and is thus useful for reactions that well as the chemical characteristics of the reacting
are thermodynamically unfavorable or are species and of the membrane. Here, the solute
product inhibited. The separation of permeable rejection coefcient is an important membrane
solutes can be achieved from the reaction selection parameter for the reactors and should
mixture by the action of a driving force that is be zero and one for the products and the
applied across the membrane. As a result of cou- biocatalysts, respectively (Prazeres and Cabral
pling the conversion reaction with separation, 1994). This would allow for the product/coprod-
lower operating costs could therefore be achiev- uct permeation through the membrane and a com-
able with this type of reactor design compared to plete retention of the biocatalysts inside the
the use of multiple reaction and product separa- reaction system.
tion methods. The congurations of biocatalytic membrane
The complete retention of the biocatalyst reactors can be categorized into two main types
within the reactor system is the most important (Fig. 1) (Giorno and Drioli 2000):
requirement for the successful continuous opera-
tion of biocatalytic membrane reactors. Ultral- 1. The biocatalyst (i.e., whole cells, enzymes, or
tration membranes with a pore size distribution antibodies) are used in solution for the conver-
range (nominal molecular weight cutoff, sion of the substrate to the desired products in a
NMWCO, of 500100 kDa) are the most adequate reactor and further transported to a reaction
for the retention of a majority of native or vessel containing a selected membrane where
184 Biocatalytic and Biochemical Membrane Reactor
1. Retentate recycle
CONVERSION REACTOR
Reactants
Biocatalyst Membrane
Permeate
2.
CONVERSION REACTOR
Reactants
Retentate
Membrane with biocatalyst

encapsulated within
Permeate
Biocatalytic and Biochemical Membrane Reactor, membrane unit, (2) Submerged membrane reactor with
Fig. 1 Main conguration types for biocatalytic mem- the membrane functioning as a catalytic and separation
brane reactors; (1) Conversion reactor combined with a unit (Adapted from Giorno and Drioli 2000)
the biocatalysts (and unreacted substrates) can Advantages Disadvantages

be recycled back to the conversion reactor, Possibility of developing Unfavorable adsorption
with the permeate selectively extracted from continuous conversion and enzyme poisoning
the reaction stream. Here, the biocatalytic processes issues
membrane reactor system might consist of a Improved control of Enzyme deactivation due
reaction systems to shear-related effects
conventional stirred tank reactor combined
Increased productivities Concentration
with a membrane-separation unit. polarization
2. The biocatalyst constituents are Contribute in a favorable Likely product/substrate
immobilized within the membrane of the shift of the reaction toward inhibition at the
matrix. Here, the membrane acts as a sup- product side membrane surface
port system for the biocatalyst as well as a Better conversion rates in Membrane fouling
product-inhibited reactions
separation unit.
Good product enrichment Enzyme leakage
and concentration in
The membrane units of the membrane reactor process stream
system can comprise of membranes which have a Facilitates control of the
at sheet shape (i.e., arranged in a plate and frame molecular weight of
module or a spiral wound module) or tubular hydrozylates
Allows the conducting of
(assembled in a tube and shell module) (Giorno
multiphase reactions,
and Drioli 2000). without emulsication issues
Advantages and disadvantages of Useful rened research
biocatalytic membrane reactors (Prazeres and tool for studying enzyme
Cabral 1994): mechanisms
Biocatalytic Hollow Fiber Membrane Reactor 185
References Where the semipermeable membrane structure

continues to the outer surface of the ber, the
Giorno L, Drioli E (2000) Biocatalytic membrane reactors; matrix is considered to be uniform, and this type
applications and perspectives. Trends Biotechnol
of ber is termed isotropic. In cases where the
18:339349
Prazeres D, Cabral JMS (1994) Enzymatic membrane bio- semipermeable membrane only has a thickness of B
reactors and their applications. Enzyme Microb 0.10.5 mm and is surrounded by a more open
Technol 16:738750 structure, the bers are called anisotropic (Bunch
1988). Although the open spongy structure of
anisotropic bers is often larger than most used
enzymes and microbes that may be employed in
Biocatalytic Hollow Fiber Membrane bioreactors, before such biocatalysts can gain
Reactor entrance to the matrix, the pores on the outer sur-
face (shell side) of the hollow ber need to be more
Yamini Satyawali1, Ehiaze Augustine Ehimen2 than 1 mm in diameter (Bunch 1988). The outer
and Winnie Dejonghe2 pore sizes have however been noticed to vary from
1
Separation and Conversion Technology, different manufacturers, even when comparing
Vlaamse Instelling voor Technologisch anisotropic hollow bers having different ltration
Onderzoek (VITO), Boeretang 200, Belgium characteristics from the same company.
2
Separation and Conversion Technology, Flemish
Institute of Technological Research (VITO), Mol,
Belgium Configuration and Operation
of Biocatalytic Hollow Fiber Membrane
Reactors
Biocatalytic hollow ber membrane reactor refers
to biological or enzyme-mediated reactor systems The hollow ber membrane modules can be either
which employ the use of hollow ber membranes integrated into the biocatalytic reactor system in an
as a support for enzyme or cellular immobiliza- external loop or incorporated directly into the bio-
tion, biocatalyst, and substrate recycling and as a reactor as a submerged membrane unit. In either
means of concentrating selected reaction streams. case surface shear and/or backushing is used to
A hollow ber membrane is a tubular-structured limit the membrane fouling (Fane, et al. 2002).
semipermeable membrane with a center called The immobilization and operation of the
the lumen having diameters in the size ranges biocatalysts in the hollow ber membrane reactors
506000 mm as seen in Fig. 1 (Bunch 1988). Such can be mainly classied into two categories
small diameters facilitate a large membrane area (Hiromi and Norio 1984):
(i.e., active surface) per unit volume for the bio-
reactor system, as well as enable the operation of 1. The biocatalyst (enzyme) solution may be
the processes at higher pressures (Katoh and introduced to the outside of the hollow ber
Fumitake 2009). or to the inner lumen, with the substrate
Biocatalytic Hollow Fiber

Matrix
Membrane Reactor,
Fig. 1 Basic structure of a Semi permeable membrane
hollow ber (Adapted from Lumen
Bunch 1988)
Outer surface
186 Biocatalytic Hollow Fiber Membrane Reactor
Enzyme solution
Hollow fibre membrane
Fraction
collector or
Analyzer
Substrate
Solution
Tubing
Thermostated pump
water
Biocatalytic Hollow Fiber Membrane Reactor, Fig. 2 Schematic drawing of a hollow ber membrane enzyme
reactor (Hiromi and Norio 1984)
solution owing on the opposite side of the achieve maximum levels when compared to
membrane to the enzyme. The substrate mole- the same concentration range using free
cules permeate the membrane and react with enzymes. This is since the actual substrate
the entrapped enzyme. The product then dif- concentration in the enzyme solution cham-
fuses back through the membrane to the sub- ber is always lower than the bulk substrate
strate solution from where it can be further concentration as a result of the enzyme
recovered (Fig. 2). reaction.
2. The enzyme solution is ultraltered into the Furthermore, the dependence of the cata-
porous region of an asymmetric hollow ber lytic rate on process temperature and pH for
membrane, or the hollow ber is dipped into hollow ber biocatalytic reactors are different
the enzyme solution so as to result in the from that of the free enzyme (Hiromi and
absorption of the enzyme molecules into the Norio 1984).
spongy lumen. The substrate is then pumped Regardless of the mode of operation used, bio-
into the reactor as previously highlighted. reactors can be run in parallel or in series. With the
After the enzymatic conversion, the product use of a parallel setup, defective reactors can
permeates through the membrane into the easily be replaced without impacting the opera-
lumen where it is collected. tion of the other reactors while the use of a series
conguration could aid the use of different
Since the biocatalyst entrapment is not biocatalysts when sequential product transforma-
chemically modied, the kinetic behavior of tions are required.
enzymes in hollow ber membrane bioreactors The control of the reaction temperature of the
could be expected to be similar to those biocatalytic hollow ber membrane reactors can
exhibited by free enzymes. However, with hol- be obtained by the use of a jacket with
low ber reactors in the rst category, the thermostated water circulated at the desired reac-
substrate permeation into the membrane has tion temperature (as seen in Fig. 2), by the immer-
been determined to be the rate-limiting step sion of the bioreactor system in a temperature-
(Hiromi and Norio 1984). The rate of the controlled water bath or by containment in a con-
biocatalytic reaction therefore tends not to stant temperature cabinet.
Biocatalytic Membrane 187
Advantages and Limitations

with the Use of Hollow Fiber Biocatalytic Biocatalytic Membrane
Membrane Reactors
Yamini Satyawali1, Ehiaze Augustine Ehimen2
Potential Merits with the Use of Hollow Fiber and Winnie Dejonghe2 B
1
Biocatalytic Membrane Reactors (Bunch 1988; Separation and Conversion Technology,
Luther et al. 1992) Vlaamse Instelling voor Technologisch
Onderzoek (VITO), Boeretang 200, Belgium
1. The possibility of immobilizing and using 2
Separation and Conversion Technology, Flemish
biocatalyst in one step
Institute of Technological Research (VITO), Mol,
2. Potential regeneration of the biocatalyst
Belgium
3. No enzyme losses or washout
4. Ease of process scale-up
5. Ease of reactor cleaning with lower steriliza-
Synonyms
tion requirements
6. Possibility for high volumetric productivity
Immobilization of enzymes
7. Ease of enzyme dosing
8. Achieve homogenous catalysis with no limi-
Biocatalytic membranes refer to the membrane
tation of substance transport
systems which have been equipped to carry out
9. Lower bioreactor capital cost
biochemical reaction function via the integration
10. Lower operational costs due to reduced
of biocatalysts (i.e., whole cells and enzymes)
need for further purication and
on/within the membrane structure. Ultraltration,
immobilization
microltration, and dialysis membranes have
11. Less susceptible to contamination
been adapted to make biocatalytic membranes.
These membranes can be activated so as to pro-
Limitations with the Use of Hollow Fiber Bio-
vide suitable conditions where the intended bio-
catalytic Membrane Reactors (Bunch 1988)
catalyst components can be coupled to it by
1. The bioreactors usually require a particulate physicochemical techniques.
free media. The biocatalytic membranes are used in the
2. The use of organic solvents is usually bioreactor as a carrier or matrix for enzyme immo-
restricted. bilization as well as a selective barrier. Such reac-
3. Fiber rupture. tors are referred to as biocatalytic membrane
4. Decreased stability of the biocatalysts. reactors (BMRs). One of the interesting advan-
5. Difculty in recycling the membrane bers. tages of using biocatalytic membrane is the occur-
ring of transport phenomena that are governed by
convective ows. Convective ows are indeed the
References combination of diffusive and advective ows and
thus are more intense than only diffusive ows. In
Bunch AW (1988) The uses and future potential of micro-
bial hollow-bre bioreactors. J Microbiol Methods case of conventional supports/carriers such as
8:103119 beads, only such diffusive ows govern the pro-
Fane A, Chang S, Chardon E (2002) Submerged hollow cess rendering it less effective. In biocatalytic
bre membrane module - design, options and opera-
membrane there is generally a close contact
tional considerations. Desalination 146:231236
Hiromi K, Norio I (1984) Hollow bre enzyme reactors. between the membrane and the enzymes as a
Trends Biotechnol 2:57 result of the enzyme immobilization. Such a
Katoh S, Fumitake Y (2009) Biochemical engineering. close contact affects the properties of the
Wiley-VCH, Weinheim
immobilized enzymes depending both on the
Luther H, Hirsch S, Schuster E, Weber E (1992) Hollow
bre modules as membrane reactor in biocatalysis. properties of the membrane and the enzyme itself.
Acta Biotechnol 12:133142 One of these enzyme properties, the pH optimum,
188 Biocatalytic Membrane Reactor
often shifts after the immobilization on a certain contain reactive functional groups, i.e., hydroxyl
carrier. Predicting the direction of such a shift in (OH), carboxyl (COOH), and amine (NH2)
pH optimum is sometimes possible by looking at groups. These reactive groups can be introduced
the charge of the used carrier, but such predictions on the membrane via the direct modication of the
are uncertain and merely indicative. Furthermore, polymeric structure of the membrane or by the
enzyme properties are affected by the used immo- preparation of the membrane from suitable mono-
bilization technique, which also will signicantly mers which contain the desired functional groups.
determine the performance of the system. There-
fore, the choice of immobilization technique and
membrane is vital to obtain a high performing References
system, and both need attention when designing
a new biocatalytic membrane system (Jochems Gekas VC (1986) Articial membranes as carriers for the
immobilization of biocatalysts. Enzyme Microb
et al. 2011).
Technol 8:450460
Various immobilization techniques or chemis- Jochems P, Satyawali Y, Diels L, Dejonghe W (2011)
tries are available for the manufacturing of biocat- Enzyme immobilization on/in polymeric membranes:
alytic membranes or for immobilizing enzymes status, challenges and perspectives in biocatalytic
membrane reactors (BMRs). Green Chem
on the membrane which can be realized via
13:16091623
adsorption, entrapment, and chemical coupling
methods such as cross-linking or covalent immo-
bilization. The immobilization of the biocatalysts
can also be achieved via gelation techniques. With
Biocatalytic Membrane Reactor
the gelation method, the biocatalyst is
immobilized in the gel layer formed during ultra-
Biochemical Membrane Reactors
ltration due to the concentration polarization
Immersed Membrane Bioreactor (IMBR)
phenomenon (Gekas 1986).
For the preparation of the entrapped biocata-
lytic membranes, the phase inversion technique is
usually applied. Here, the biocatalysts are intro-
Biocatalytic Membrane Reactor,
duced to the casting solution so as to contain the
Mass Transport in
enzymes or whole cells in the polymeric structure
of the resulting membrane. The use of the encap-
Endre Nagy
sulation method, which is similar to the entrap-
Research Institute of Chemical and Process
ment method, has the biocatalyst contained in the
Engineering, University of Pannonia, Veszprem,
liquid membranes formed via the emulsication
Hungary
of aqueous enzymatic phase with the organic
membrane phase.
The cross-linking with bifunctional agents, i.e.,
Synonyms
glutaraldehyde (which is the most common bifunc-
tional agent), is the simplest demonstrated method
Modeling of biocatalytic membrane reactor
for the chemical immobilization of biocatalyst on
membrane systems. Here, the biocatalysts initially For creating a biocatalytic membrane layer, the
need to be adsorbed onto/within the membrane biological catalyst such as whole cells (bacteria,
structure. The formation of covalent bonds within yeast, mammalian and plant cells) is inoculated
the free amino groups in the enzyme molecule can into the porous membrane layer where they are
then be achieved without any presence of reactive grown, or bioactive molecules such as enzymes
groups on the membrane. are immobilized within the membrane matrix or in
The application of covalent coupling methods a thin layer on the membrane interface of the
is obtainable with membrane systems which membrane layer. A medium is owing through
Biocatalytic Membrane Reactor, Mass Transport in 189
the biocatalytic, mostly asymmetric, plain, or Biocatalytic Membrane Reactor, Mass Transport in,
cylindrical (hollow ber) layer, supplying the Table 1 Expressions of the important biocatalytic
reactions
cells with oxygen and nutrients or the enzyme
with reactant(s) while the wastes and desired Substrate inhibition: Q KMi cc
rmax c
2 =K
c
products are removed (Drioli and Giorno 1999). Substrate inhibition and QK rmax c
B
Mi 1p=Kp cc2 =Kc
competitive product
The performance of a hollow ber or sheet biore-
inhibition:
actor is primarily determined by the momentum
Competitive product QK rmax c
and mass transport rate (Bird et al. 1960) of the Mi 1p=Kp c
inhibition:
key nutrients through the biocatalytic membrane Noncompetitive product Q K rmax c
Mi c 1p=Kp
layer. Thus, the operating conditions inhibition:
(transmembrane pressure, feed velocity) and the
physical properties of membrane (porosity, wall
thickness, lumen radius, matrix structure, etc.) can
considerably inuence the performance of a bio-
Michaelis-Menten (MM) kinetics (c denotes the
reactor, the effectiveness of the reaction. The lim-
concentration):
ited transport of nutrients can cause serious
damage in production. The introduction of con-
d2 c dc vmax c
vective transport is crucial in overcoming diffu- D u 0 (1)
sive mass transport limitation of nutrients dy 2 dy K M c
especially of the sparingly soluble oxygen. The
starting balance equations developed by Navier The source term can be different in biocatalytic
and Stokes (Bird et al. 1960; Seidel-Morgenstern reactions; the most often applied equations are
2010; Nagy 2011) can be essentially simplied, listed in Table 1. The inhibited reaction can take
and the momentum and the mass transport expres- place in both the enzymatic and microbial
sions can often separately be solved in the case of reactions.
biocatalytic processes. The principle of the mass Two important cases can be discussed, namely,
transport of substrates/nutrients into the when the outlet (at y = d) diffusive ow is zero
immobilized enzyme/cells, through a solid, (dc/dy = 0, no sweep phase on the permeate side,
porous layer (membrane, biolm) or through a case A) and dc/dy 6 0 at y = d (there is sweep
gel layer of enzyme/cells, is the same. The struc- phase on the permeate side, case B). The mass
ture and the thickness of this mass transport layer transfer rates are given by Eqs. 2 and 3 for cases
can be very different; thus, the mass transport A and B, respectively, for rst-order reaction as
parameters, namely, diffusion coefcient, D; con- limiting case of MM kinetics (Nagy 2011):
vective velocity, u; the bioreaction rate constant; 2
Pe
their dependency on the concentration; and/or Y2 tanhY PeY
space coordinate are characteristics of the porous D 4
J c0 (2)
layer and the nature of the biocatalysts. Some d Pe
tanhY Y
assumptions made for expression of the differen- 2
tial mass balance equation to the biocatalytic
membrane layer are as follows: reaction occurs with
at every position within the biocatalyst layer; s
mass transport through the biocatalyst layer r
Pe 2 k1 d2
occurs by diffusion and convection; mass trans- Y W2 ; W
4 D
port parameters (diffusion coefcient, convective
velocity, bioreaction rate constant) are constant;
Y
and the process is steady state. Thus, the mass J b c0 cd
balance equation can be given for a plain mem- ePe=2 Pe=2sinhY YcoshY
brane layer as follows, applying, e.g., the (3)
190 Biocatalytic Membrane Reactor, Modeling of
with (c = cd at y = d and c = c at y = 0; pt z pe z r
Pe = nd/D) pM r, z ln pt z (8)
ln1 d=ro ro
D Pe=2tanhY Y k ro pt z pe z
b uM r, z (9)
d tanhY em r ln1 d=ro
The mass transfer rates are given by Eqs. 4 and 5 The axial pressure distribution is given in entry
for cases A and B, respectively, for zero-order Hollow Fiber Enzymatic Reactor, Modeling of.
reaction as limiting case of MM kinetics (k0
denotes the reaction rate constant):

References
D 0 W2 1
J c Pe 1 Pe (4)
d Pe e Bird RB, Stewart E, Lightfoot EN (1960) Transport phe-
nomena. Wiley, New York
Drioli E, Giorno L (1999) Biocatalytic membrane reactor.
and Taylor & Francis, London
Nagy E (2011) Basic equations of the mass transport
D Pe through a membrane layer. Elsevier, Amsterdam
J Seidel-Morgenstern A (2010) Membrane reactors. Wiley-
d 2

VCH, Weinheim
ePe=2 ePe
1 T c 0 cd (5)
sinhPe=2 1T
with
Biocatalytic Membrane Reactor,
s Modeling of
W2 Pe k0 d2
T 2 e 1 Pe 1 ; W
Pe Dc0 Endre Nagy
The ow through this membrane is proportional to Engineering, University of Pannonia, Veszprem,
the transmembrane pressure drop and inversely Hungary
proportional to the uid viscosity (Darcy law,
Eq. 6) (k is the permeability and m is the uid
viscosity): There are two main types of membrane bioreac-
tors: (i) the system that consists of a traditional
k dpM stirred-tank reactor combined with membrane
uM (6)
m dr separation unit and (ii) the membrane that con-
tains the immobilized biocatalysts such as
Then, one can get from Eq. 6, for capillary mem- enzymes, microorganisms, and antibodies and,
brane, as: thus, acts as a support and a separation unit. The
biocatalyst can be immobilized in or on the mem-
brane by entrapment, gelication, physical
1 @ @ ruM k1 @ @p
r r M 0 (7) adsorption, ionic binding, covalent binding, or
r @r @r m r @r @r
cross-linking (Giorno and Drioli 1999; Cabral
et al. 2001). Modeling of stirred-tank reactor, for
When Eq. 7 is integrated twice with respect to r chemical reactions in it, is known (Westerterp
with boundary conditions, at r = ro then pM = pt et al. 1993). This can easily be adapted for bio-
and at r = ro + d then pM = pe, one gets the chemical reactions applying the reaction kinetics
radial pressure and velocity distributions (Nagy given in entry Biocatalytic Membrane Reac-
2011, pp. 178) within the membrane as: tor, Mass Transport in. The momentum and
Biocatalytic Membrane Reactors 191
s

mass, energy transport through membrane as a
16k 1
separation unit, starting with Navier-Stokes equa- l 1
r3o ln1 d=ro B
tion (Bird et al. 1960), strongly depend on the
separation process used (Baker 2004; Drioli
and
et al. 2006; Nagy 2011). Immobilizing biocatalyst B
in the porous membrane matrix of an asymmetric 2 2
membrane or onto membrane interface, the mem- d rk
B 1
brane layer can act as an active catalyst and a ro ro d
permselective layer. The biocatalytic membrane
modules (plane and frame, hollow ber capillary, The mass transfer rates for constant parameters
tube as well as spiral membrane modules) can be are given in entry Biocatalytic Membrane
divided into three sections: feed and permeate Reactor, Mass Transport in for limiting cases.
uid phases and the biocatalytic membrane layer An analytical approach is recommended by
itself. Modeling of a membrane process means the Nagy (2011) for the general Michaeli-Menten
simultaneous treatment of every section starting kinetics or bioreactions with two substrates feed-
with Navier-Stokes equations for momentum, ing on the two sides of the membrane reactor. This
mass, and energy (Bird et al. 1960) adapting method can take into account the effect of the
them to the real operating conditions and complet- in-space and/or time variable or concentration
ing them by source or sink terms (Seidel- dependent parameters on the process.
Morgenstern 2010). The energy balance equation
is mostly negligible in case of biochemical reac-
tion; the process can be regarded as isotherm;
accordingly the continuity (conservation of References
mass), momentum, and component balance equa-
tions (see Eqs. 14 in entry Hollow Fiber Baker RW (2004) Membrane technology and applications,
2nd edn. Wiley, Chichester
Enzymatic Reactor, Modeling of) should be Bird RB, Stewart E, Lightfoot EN (1960) Transport phe-
simultaneously or consecutively (mass transport nomena. Wiley, New York
does not affect the velocities of uids) solved after Cabral JMS, Mota M, Tramper J (2001) Multiphase biore-
suitable simplication of a real system. The cylin- actor design. Taylor & Francis, London
Drioli E, Giorno L (1999) Biocatalytic membrane reactor.
drical effect should be taken into account for Taylor & Francis, London
capillary membrane when the ratio of the mem- Drioli E, Criscuoli A, Curcio E (2006) Membrane
brane thickness and the internal radius is larger contactors: fundamentals, applications and potentiali-
than about 0.05. The second-order pressure gradi- ties. Elsevier, Amsterdam
Nagy E (2011) Basic equations of the mass
ent can be expressed by Eq. (1) (Nagy 2011, p. transport through a membrane layer. Elsevier,
225). From that the radial and the axial velocity Amsterdam
distribution in the feed (lumen) and the shell side Seidel-Morgenstern A (2010) Membrane reactors. Wiley-
can easily be obtained. VCH, Weinheim
Westerterp KR, Swaaij WPM, Beenackers AACM
(1993) Chemical reactor design and operation. Wiley,
d2 pt 8uo mB 1 New York
2 fcoshlL z
dz 2 ro l1 BsinhlL l2
h1 f 1 Bicoshlzg (1)
Biocatalytic Membrane Reactors
with [rk denotes the outer radius of hexagonal
arrangement of the hollow bers; f is the fraction Membrane Bioreactors
of inlet feed directed out through the shell (Nagy Principle of Biocatalytic Membrane Reactors
2011)] (BMR)
192 BMR with Biocatalysts Entrapped Within Pores of Asymmetric Membranes
enzyme solution from porous to selective layer,

Biocatalytic Membrane Reactors where the pores in the selective layer are small
(BMR) with Biocatalysts Entrapped enough to retain enzyme molecules but large
Within the Pores of Asymmetric enough to let products pass through.
Membranes Enzymes are not able to work at high
concentration; hence, the amount of immobilized
Lidietta Giorno catalyst strongly affects reactor performance.
Institute on Membrane Technology, National When the substrate is transported by convection
Research Council of Italy, ITM-CNR, Rende, through the enzyme-loaded membrane, the
Italy residence time is an important parameter to
optimize.
For systems where the substrate is transported
Synonyms by diffusion, quick graphical procedures are avail-
able in literature for evaluating the extent to which
Catalytic membrane bioreactors with biocatalysts external and internal diffusion affects
segregated within the pores of asymmetric mem- immobilized enzyme kinetics (Engasser 1978;
branes; Catalytic membrane reactors with Bailey and Ollis 1986). Suggested models can be
biocatalysts segregated within the pores of asym- used to predict reactor performances in the case of
metric membranes enzymes with relatively simple kinetics (such as
a-galactosidase, invertase, glucose isomerase,
Polymeric and inorganic membranes with asym- urease, fumarase) and when the kinetic and trans-
metric structure have been used to segregate port parameters are known.
(by physical entrapment and adsorption) enzymes The dynamics of substrate conversion depends
and microbial cells within the porous membrane on enzyme kinetics as well as on mass transport
structure (Drioli and Giorno 1999; Drioli conditions.
et al. 1982). Asymmetric hollow bers with selective layer
Polymeric membranes have been cast with either on the lumen side or on the shell side can
microbial cells in the casting solution using the provide an interesting support for immobilizing
phase inversion technique. A number of microor- enzymes. Figure 1 shows the case where the
ganisms can withstand high temperatures and enzymes are effectively immobilized or segre-
organic solvents and have been entrapped in mem- gated within the spongy annular section of an
branes. Microbial enzymes maintain their activity asymmetric hollow ber containing the selective
presumably because of cellular membrane protec- layer on the lumen side.
tion. Cell-loaded membranes appear to be kineti- In this conguration, the substrate solution can
cally active and stable over a long period of time. It be circulated along the shell side of the membrane
is noteworthy that cell entrapment can enhance module and permeated through the enzyme-loaded
microbial activity as compared to cell behavior in membrane under an applied transmembrane pres-
a homogeneous solution, an effect probably due to sure as driving force. As the substrate permeates
cellular membrane permeabilization as a conse- through the membrane containing the enzyme, it is
quence of the entrapment procedure. converted to the reaction product. If the molecular
Biocatalysts can also be entrapped in asymmet- weights of substrate and products are small, both
ric membranes after their preparation, by cross- can permeate through the membrane. This means
ow ultraltration, under operating conditions that the permeate recovered from the lumen side of
that promote fouling, where foulant molecules the membrane contains both reaction product and
are the biocatalysts themselves. The retention of unconverted substrate. Measuring the reduced sub-
enzymes within the porous layer of the asymmet- strate concentration (or the increased product con-
ric membrane is very efcient when permeating centration) in the permeate during time operation, it
BMR with Biocatalysts Entrapped Within Pores of Asymmetric Membranes 193
This gives the conversion at each passage

through the membrane pores that basically work
as high-throughput microreactors.
When the enzyme is also immobilized on the
membrane surface, it is necessary to take into B
account the conversion of the substrate in the
retentate solution, and then the total conversion
is calculated as:

Cr Cp Ci Cr
Total conversion (2)
Ci
where Ci is the initial substrate concentration.

The reactor volume is represented by the void
Biocatalytic Membrane Reactors (BMR)
volume of the porous asymmetric membrane.
with Biocatalysts Entrapped Within the Pores Under convective ow, each macrovoid of the
of Asymmetric Membranes, Fig. 1 Cross-section of asymmetric membrane can work in conditions
BMR with enzyme entrapped within the porous support of continuous stirred tank reactor. If this
assumption is applicable, the reaction rate equa-
tion can be derived by the balance equation at
is possible to calculate the reaction rate and the steady state:
conversion degree.
If the enzyme is immobilized only within the QCf QCp vr V 0 (3)
pores and not on the membrane surface, the con-
centration of the substrate in the retentate stream Resolving for vr, the following is obtained:
does not change, because only the solution that

permeates through the membrane contacts the Q Cf Cp
biocatalyst. The substrate is then continuously vr
V
recycled to the tank and fed to the reactor at
constant concentration (Giorno et al. 2001). where vr is the reaction rate (mmolcm3min1);
If the enzyme is also present on the membrane Q is the ltration ow rate (cm3min1); Cf and Cp
surface, some decrease of substrate concentration are the feed and permeate concentration
in the retentate stream is also observed. A blank (mmolcm3), respectively; and V is the reactor
experiment may conrm that when no enzyme volume (cm3).
is present on the membrane, no conversion occurs. The general behavior of reaction rate and con-
The conversion degree can be calculated as version as a function of residence time assumes
follows: the form qualitatively illustrated in Fig. 2.
Increasing the residence time means decreas-
Cr Cp ing of the ltration ow rate (Q); therefore, reac-
Conversion (1)
Cr tion rate increases as long as higher residence time
means more interaction between the substrate and
where the immobilized enzyme. After the residence time
to saturate enzyme is reached, Cf and Cp in Eq. 3
Cr = concentration of substrate in the retentate are constant, and further increase of residence
solution time (i.e., decrease of ltration ow rate) causes
Cp = concentration of substrate in the permeate a decrease of the reaction rate (Mazzei
solution et al. 2012).
194 Biocatalytic Membrane Reactors with Chemically Bound Enzyme
Biocatalytic Membrane
Reactors (BMR)
with Biocatalysts
Entrapped Within
Reaction rate
Conversion
the Pores of Asymmetric
Membranes,
Fig. 2 Behavior of
reaction rate and conversion
as a function of residence
time in a continuous stirred
tank membrane reactor
Residence time
References
Bailey JE, Ollis DF (1986) Biochemical engineering fun-

Attachments of biocatalysts to membrane can
damentals, 2nd edn. McGraw-Hill, New York, result from ionic binding, cross-linking, and
pp 202227 covalent linking. Since 1954, when protease was
Drioli E, Giorno L (1999) Biocatalytic membrane reactors: covalently bound to diazotized polystyrene,
application in biotechnology and the pharmaceutical
industry. Taylor & Francis, London
enzyme immobilization via covalent bonds has
Drioli E, lorio G, De Rosa M, Gambacorta A, Nicolaus been an established immobilization technique,
B (1982) High Temperature Immobilized-Cell Ultral- usually carried out by means of bi/multifunctional
tration Reactors. J Membr Sci 11:365 reagents, such as glutaraldehyde, (3-aminopropyl)
Engasser JM (1978) A fast evaluation of diffusion effects on
bound enzyme activity. Biochim Biophys Acta 526:301
triethoxysilane (APTES), 1-ethyl-3-(3-dimethyla-
Giorno L, Drioli E, Carvoli G, Donato L, Cassano A (2001) minopropyl) carbodiimide hydrochloride (EDC),
Study of an enzyme membrane reactor with etc.
immobilized fumarase for production of L-malic acid. The crucial aspect is the use of mild reaction
Biotechnol Bioeng 72(1):7784
Mazzei R, Drioli E, Giorno L (2012) Enzyme membrane
conditions that do not denature the enzyme during
reactor with heterogenized beta-glucosidase to obtain the covalent bond formation.
phytotherapic compound: optimization study. J Membr When the extent of initial denaturation is
Sci 390391:121129 acceptable in the economics of the process,
enzymes bound to membranes can be used in
continuous ow reactors.
Table 1 shows the protein functional groups
Biocatalytic Membrane Reactors involved in covalent bond with functional groups
with Chemically Bound Enzyme present or grafted on the membrane, in mild
conditions.
Lidietta Giorno1, Giuseppe Vitola1, Table 2 summarizes most commonly used
Giuseppe Ranieri1 and Francesca Militano1,2 cross-linking agents to bridge enzyme to
1
Institute on Membrane Technology, National functionalized membranes.
Research Council of Italy, ITM-CNR, Rende, Hydrophilic, hydrophobic polymeric mem-
Italy branes as well as inorganic membranes can be
2
University of Calabria, Rende, Calabria, Italy used as support for covalent bound of enzymes.
Polymeric membranes based on natural mate-
rials, such as cellulose and its derivatives
Synonyms (regenerated cellulose, cellulose acetate, etc.)
have been extensively used as support for the
Biocatalytic membrane reactors with covalent covalent immobilization of enzymes. Cellulose
bound enzyme; Immobilized enzyme membranes can be activated by a variety of agents
Biocatalytic Membrane Reactors with Chemically Bound Enzyme 195
Biocatalytic Membrane Reactors with Chemically Bound Enzyme, Table 1 Protein functional groups involved in
covalent bond with membranes in mild conditions
Protein chemical group Structure
Epsilon amino groups of lysine O
H2N B
OH
NH2
Epsilon amino groups of arginine NH O
H2N N OH
H NH2
Beta and gamma carboxyl groups of aspartic and glutamic acids O
HO OH
O NH2
Phenol ring of tyrosine O
OH
NH2
HO
Thiol group of cysteine O
HS OH
NH2
Hydroxyl groups of serine and threonine

O OH O
OH OH
HO
NH2 NH2
Imidazole group of histidine O

H
N OH
N NH2
Indole group of tryptophan O
OH
N NH2
H
Biocatalytic Membrane Reactors with Chemically Bound Enzyme, Table 2 Bi/multifunctional reagents com-
monly used in enzymatic covalent immobilization
Reactive chemical
Cross-linker Structure group
Glutaraldehyde (GA) Aldehydic groups
O O
(3-aminopropyl) triethoxysilane (APTES) O Ethoxyl and amino

Si NH2 groups
O O
Dimethylaminopropane (DAMP) Amino groups

H2N NH2
O, O0 - Bis(2-aminopropyl)-polyethylene glycol PEG Amino groups

500 (bNH2PEG)
H2N NH2
1,3,5-benzene tricarbonyl chloride (or trimesoyl chloride O Cl Carbonyl chloride

TMC) groups
O O
Cl Cl
1-ethyl-3-(3-dimethylaminopropyl) carbodiimide Carbodiimide group

+ Cl
hydrochloride (EDC) N NH
Spacer Arm 0.0 C
N
Dicyclohexylcarbodiimide (DCC) Carbodiimide group

Spacer Arm 0.0 N
C
N
to produce reactive centers for the covalent bind- the corresponding aldehydes by means of sodium
ing of enzymes. Its chemical structure is mainly periodate (NaIO4). The covalent binding of the
composed of hydroxyl groups, and then the most enzyme can be achieved by direct reaction with
common activation methods are based on the con- the aldehyde-active groups on the membrane sur-
version of the hydroxyl groups into more reactive face (Xiao-Jun et al. 2011) or via different spacers.
functional groups. In particular, carbonyl groups The introduction of an appropriate spacer between
are very suitable for enzyme immobilization since the enzyme and the membrane often improves the
they can react with primary amines from enzymes performance of immobilized enzyme, such as
(e-amino group of lysine residues) to form imine. enzyme stability and retained activity. Spacers of
The most used method of forming carbonyl different lengths are generally introduced by using
groups onto the cellulose skeleton is periodate bifunctional agents such as diamines (esp.
oxidation. The hydroxyl groups (secondary alco- ethylenediamine, pentaethylenehexamine, etc.)
hol groups) of the glucose units are oxidized into that are coupled or activated with glutaraldehyde
Biocatalytic Membrane Reactors with Chemically Bound Enzyme 197
H OH H OH H OH
H O periodate H O
O H2N-(CH2)n-NH2 H O
HO oxidation O O
H O O O
OH H O diamine spacer OH
H H H
H H N H
Cellulose H
dialdehyde (CH2)n NH2
B
H OH H OH
OHC (CH2)3 CHO H O H2N
H O
O
O OH O
glutaraldehyde enzyme coupling O OH
activation H H
N H
H (CH2)n (CH2)3 O N H
N H (CH2)n N
C (CH2)3 N
H C C
H C
H H
Biocatalytic Membrane Reactors with Chemically activated with periodate oxidation and a generic diamine
Bound Enzyme, Fig. 1 Schematic representation of an spacer, n represent the number of CH2 groups
enzyme covalently immobilized onto cellulose membrane
before enzyme immobilization (Peng-Cheng interesting for enzyme immobilization is grafting

et al. 2011). The reaction mechanism involves modication which can introduce on the PVDF
the amino groups of the enzyme and the aldehyde backbone functional chains that can be used for
groups of glutaraldehyde on the membrane sur- enzyme immobilization; in addition acting on
face. The schematic representation of an enzyme degree of grafting and type of functional chain
covalently attached onto cellulose-activated introduced, it is possible to affect or not affect
membrane is reported in Fig. 1. membrane bulk properties as, for example,
Poly(vinylidene uoride) (PVDF) membrane hydrophobicity.
has been used either for scientic research or Surface grafting can be performed by high-
industrial application in particular micro- and energy radiation techniques like plasma treatment,
ultraltration process. The great interest to ozonation, and UV or by wet chemistry in harsh
PVDF membrane is due to its extraordinary ther- basic environment. For example, grafting of
mal, mechanical, and chemical resistance in addi- amino groups on PVDF membrane surface can
tion to membrane-forming properties. Its structure be achieved by treatment at high pH and high
consisting of repeated (CH2-CF2)n units con- temperature with diamine reagent. The high pH
fers to PVDF these good properties, but makes promotes uorine elimination from the PVDF
impossible enzyme immobilization on native chains and the grafting of diamino reagent
PVDF in a covalent manner. This disadvantage (Vitola et al. 2015) (Fig. 2).
can be overcome performing PVDF functiona- Subsequently the pendant amino groups can be
lization, in order to introduce on the polymeric activated for enzyme immobilization in mild con-
membrane structure, reactive groups that can ditions using condensing reagent like glutaralde-
covalently couple with enzyme functional groups. hyde (Fig. 3).
Surface modication of existing PVDF mem- Polymeric membrane based on
brane can be performed by coating or grafting. polyethersulfone (PES) is widely used mem-
Coating modication is preferred when only an branes in ultraltration processes due to their
improved hydrophilicity of the membrane surface high performance and low-cost prole. PES mem-
is required, but it is not useful as anchor point for branes have been used also in more sophisticated
enzyme immobilization because of the instability application such as carrier for immobilization of
of the coated layer which interacts with the mem- bioactive molecules in membrane bioreactors. For
brane merely by physical adsorption. More this aim effort has been done to introduce reactive

H2O F
[n
[ CF2 CH CF2 CH2 [n
[ CF2 CH CF CH2 [n
[ CF2 CH= CF CH2
H F

OH
H F

[n
[ CF2 CH= CF CH2 HF
[n
[ CF2 C C CH2 [n
[ CF2 CH= C CH2
+
+
H2N HN
H2N
NH2 NH2
NH2
Biocatalytic Membrane Reactors with Chemically Bound Enzyme, Fig. 2 Possible mechanism of alkaline-induced
DAMP grafting on PVDF
H O O
Na2CO3
+ H2N NH2 NH2 + O H N H N N
pH11 50C Enzyme
DAMP GA
PVDF PVDF-DAMP PVDF-DAMP-GA PVDF-DAMP-GA-Enzyme
Biocatalytic Membrane Reactors with Chemically Bound Enzyme, Fig. 3 Illustration of various steps for PVDF
functionalization with amino groups and subsequent enzyme attachment via cross-linking with glutaraldehyde
groups for biomolecule immobilization without APTES exhibits terminal amino groups avail-
affecting the overall bulk properties. Photochem- able for the attachment of cross-linking reagents
ical surface modication of existing PES mem- such as glutaraldehyde. Terminal aldehydic
brane is the most popular technique for group of glutaraldehyde reacts with the exposed
introduction of reactive groups on PES mem- amino groups of the organosilane by forming a
brane. In fact PES has an intrinsic photoreactivity Schiff base. Final covalent binding occurs
making possible to perform the modication in through formation of a new Schiff base between
mild reaction condition. Monomer like acrylic aldehydic terminal group of glutaraldehyde and
acid has been grafted in this way on PES mem- amino groups exposed from the enzymatic
brane and then activated for enzyme immobiliza- amino acid residues of lysine (S. Tanvir
tion using 1-ethyl-3-(3-dimethylaminopropyl) et al. 2009). Schematic reaction of an example
carbodiimide hydrochloride (EDC). of enzymatic covalent immobilization onto
Inorganic membranes, such alumina, titania, inorganic supports is illustrated in Fig. 4.
silica, etc., can be also used for enzymatic Applications of covalently immobilized mem-
immobilization via covalent binding if brane systems are membrane electrodes for ana-
functionalized in a suitable way. Hydroxyl lytical purposes, reactions of substrates whose
groups present onto inorganic membrane sur- molecular weight is low as compared to mem-
face are not able to directly coupling enzymes. brane molecular weight cut-off, and enzymatic
Activation of hydroxyl groups can be conversion of macromolecules to lower molecular
performed by silanization reagents, such as weight species able to permeate the supporting
(3-aminopropyl)triethoxysilane (APTES). membrane.
Biocatalytic Membrane Reactors with Chemically Bound Enzyme
Biocatalytic Membrane Reactors with Chemically Bound Enzyme, Fig. 4 Schematic representation of stages involved in an example of enzymatic covalent immobilization
onto inorganic membranes
199
B
200 Biocatalytic Membrane Reactors with Covalent Bound Enzyme
References the protein. In this process, enzymatic activity is

signicantly decreased largely due to different ori-
Chen P-C, Huang X-J, Huang F, Ou Y, Chen M-R, Xu Z-K entations of the enzymes with respect to the mem-
(2011) Immobilization of lipase onto cellulose ultrane
brane or to multiple point attachment which also
ber membrane for oil hydrolysis in high performance
bioreactor. Cellulose 18:15631571 renders the active site inaccessible for the sub-
Huang X-J, Chen P-C, Huang F, Ou Y, Chen M-R, Xu Z-K strates or sometimes results in complete denatur-
(2011) Immobilization of Candida rugosa lipase on ation of enzymes. Methods of nonspecic
electrospun cellulose nanober membrane. J Mol
immobilization include adsorption, nonspecic
Catal B Enzym 70:95100
Tanvir S, Pantigny J, Boulnois P, Pulvin S (2009) Covalent covalent binding, entrapment, and encapsulation.
immobilization of recombinant human cytochrome Therefore, it is essential to develop methodologies
CYP2E1 and glucose-6-phosphate dehydrogenase in for site-specic immobilization of the enzymes
alumina membrane for drug screening applications.
with the active (or binding) site directed away
Vitola G, Mazzei R, Fontananova E, Giorno L (2015) from the membrane. This would ensure that bene-
PVDF membrane biofunctionalization by chemical ts of enzyme immobilization could be availed
grafting. J Membr Sci 476:483489 without compromising on their activity.
Specic immobilization or site-specic immo-
bilization of enzymes depends a lot on the type of
enzyme and support material (membranes in the
Biocatalytic Membrane Reactors with
case of biocatalytic membrane reactor). Depending
Covalent Bound Enzyme
upon the enzyme structure and characteristics, var-
ious strategies can be used to conduct site-specic
Biocatalytic Membrane Reactors with Chemi-
immobilization (Buttereld et al. 2001). For exam-
cally Bound Enzyme
ple, if an enzyme does not contain any cysteines
(or any cysteines that are required for activity),
then the site-directed mutagenesis approach may
Biocatalytic Membrane Reactors be used. This approach aims at introduction of
with Site-Specific Immobilized unique cysteines to enzymes. The enzymes are
Enzyme attached on thiol-reactive surfaces through the
sulfhydryl group on the side chain of the intro-
Yamini Satyawali1, Ehiaze Augustine Ehimen2 duced cysteine, which is located on the opposite
and Winnie Dejonghe2 side of the protein from the active site. An example
1
Separation and Conversion Technology, of such an enzyme is subtilisin, a protein normally
Vlaamse Instelling voor Technologisch devoid of cysteine (Vishwanath et al. 1998).
Onderzoek (VITO), Boeretang 200, Belgium Another approach in this eld is gene fusion
2
Separation and Conversion Technology, Flemish which is used to incorporate a peptidic afnity
Institute of Technological Research (VITO), Mol, tag at the N- or C-terminus of the enzyme. The
Belgium enzymes are then attached from this afnity tag to
anti-tag antibodies on membranes. The gene fusion
approach is appropriate when the N- or C-terminus
Membranes are used in bioreactors as selective of the enzyme are away from the active site. Gene
barriers and sometimes also as a support/carrier fusion methods have been employed to investigate
matrix for enzyme immobilization. Such reactors site-specically immobilized alkaline phosphatase
are referred to as biocatalytic membrane reactors (AP) on functionalized membranes (Vishwanath
(BMRs). The immobilization of enzymes on mem- et al. 1997). Another method considered to be of
branes can be of two types: (i) nonspecic immo- biomimetic inspiration is the posttranslational
bilization or (ii) specic immobilization (Fig. 1). modication to incorporate a single biotin moiety
Nonspecic immobilization is often random and on enzymes. The enzymes can be attached
proceeds through the numerous lysine residues on to membranes through a (strept)-avidin bridge.
Biochemical Conversion 201
Active Site
a
b B
Biocatalytic Membrane Reactors with Site-Specific array of similarly oriented proteins, always with the
Immobilized Enzyme, Fig. 1 (a) Random immo- active site away from the polymer surface (Buttereld
bilization of enzymes, showing potential difculties. et al. 2001)
(b) Site-specic immobilization of enzymes to form an
This modication incorporates a biotin moiety on a

specic lysine residue on the enzyme, which Biochemical Conversion
allows the use of an avidin spacer between the
enzyme and the immobilization surface. Biotin, a Yamini Satyawali, Ehiaze Augustine Ehimen and
vitamin, has an afnity for avidin, a protein found Winnie Dejonghe
in egg whites, that is a remarkably strong associa- Separation and Conversion Technology, Flemish
tion. An example of such an enzyme is beta galac- Institute of Technological Research (VITO), Mol,
tosidase immobilization on polyethersulfone Belgium
membranes (Vishwanath et al. 1995).
From the membrane point of view, some of the
approaches for site-specic attachment of the Biochemical conversion refers to the application
enzyme include modifying the membranes by of biologically mediated processes to facilitate the
attaching functional groups such as sugar sand production of a range of chemical products and
polypeptides to which the enzymes eventually power. This conversion route involves the direct
bind (Mazzei et al. 2010). use of living organisms (i.e., whole cells) or
desired active protein constituents (i.e., enzymes
and antibodies) extracted from living cells to aid
the reformation of reactant species to desired
References
products or the synthesis of new chemicals.
Buttereld DA, Bhattacharyya D, Daunert S, Bachas L Biochemical conversion processes thus rely on
(2001) Catalytic biofunctional membranes containing biocatalytic transformation systems for achieving
site-specically immobilized enzyme arrays: a review. synthetic chemistry goals. Historically, such pro-
cesses have been applied for centuries for alcohol
Mazzei R, Drioli E, Giorno L (2010) Biocatalytic mem-
branes and membrane bioreactors. In: Enrico Drioli E, (via fermentation) and cheese (via enzymatic deg-
Giorno L (eds) Comprehensive membrane science and radation of milk proteins) production in different
engineering. Elsevier, Oxford, pp 195212 geographical regions globally using different
Vishwanath S, Bhattacharyya D, Huang W, Bachas LG
techniques but with similar underlying conversion
(1995) Site-directed and random enzyme immobiliza-
tion on functionalized membranes: kinetic studies and principles. Increased knowledge on protein struc-
models. J Membr Sci 108:1 ture (including protein design tools, i.e., rational
Vishwanath SK, Watson CR, Huang W, Bachas LG, design and directed evolution) and function rela-
Bhattacharyya D (1997) Kinetic studies of site-
tionships has further improved scientic under-
specically and randomly immobilized alkaline phos-
phatase on functionalized membranes. J Chem Technol standing of biochemical conversion processes
Biotechnol 68:294 and thus has helped to advance its adaptation,
Vishwanath S, Wang J, Bachas LG, Buttereld DA, implementation, and application (up to industrial
Bhattacharyya D (1998) Site-directed and random
scales) in recent decades.
immobilization of subtilisin on functionalized mem-
branes: activity determination in aqueous and organic The substrates employed for biochemical con-
media. Biotech Bioeng 60:608 version schemes can be derived directly or
202 Biochemical Conversion
indirectly from biomass sources (i.e., plant and fewer environmental issues (Johannes
animal) as in the case of the anaerobic digestion et al. 2006). Compared with chemical and ther-
of biomass for the production of a range of volatile mochemical conversion technologies, the applica-
fatty acids (carboxylic acids with chain lengths tion of biochemical conversion processes has
C1C5) and methane (CH4) which have applica- been highlighted to be generally more efcient
tions in the chemical and energy industry, respec- (with a lower biocatalyst concentration employed
tively. Here, a multistage degradation of the compared to chemical catalysis requirements)
complex polymers and compounds (which make with milder process reaction conditions (usually
up the biomass substrate) to lower molecular with temperature and pH ranges of 2040 C and
weight products is facilitated by a consortium of 58, respectively, for most biochemical pro-
anaerobic bacteria (whole cells). The enzymatic cesses) as opposed to the harsher pressure, tem-
transesterication or esterication of biologically perature, and pH operation conditions usually
obtained triglycerides to fatty acid methyl esters applied for chemical conversions. Biochemical
using the enzymes-lipases is another example of conversion technologies also possess the
such bio-derived substrate for biochemical con- advantage of being modiable to increase
version. The conversion of non-biomass-derived selectivity, stability, and activity of the applied
substrates is also feasible, with the synthesis of biocatalyst to improve the product yields of
new chemical products, obtainable via the trans- the targeted conversion process, the possibility
formation and degradation of chemical reactants of the biocatalyst recycling (which could aid a
with the application of biochemical conversion reduction in the process wastes and costs), and
methods. An example is the asymmetric synthesis potentially having superior environmental ben-
of chiral amines from prochiral ketones (which ets (i.e., nontoxic by-products or side reac-
has a signicant interest from pharmaceutical tions, non-production of process efuents, and
industries), i.e., 4-phenyl-2-butylamine produc- reduced energy process inputs) compared with
tion via the transamination of 2-propylamine and the use of chemical/thermochemical conversion
4-phenyl-2-butanone with o-transaminases used schemes.
as the process biocatalyst. With approximately The main disadvantages with the use of bio-
100 different biocatalyst-mediated processes chemical conversion systems are its inhibition
applied in the pharmaceutical, food, chemical, susceptibility (including substrate and product
agricultural, and cosmetic industries, the product inhibition), the limiting operation ranges where
ranges from biochemical conversion processes biochemical conversions can be applied (since
range from specialist research and commodity biocatalysts usually have an optimal operational
chemicals, to food and fuel products (Wandrey pH, temperature, etc.), and the water requirements
et al. 2000). for most processes (which could potentially be
With the biochemical conversion schemes disadvantageous from a downstream processing
highly dependent on the use of biocatalytic pro- and energetics perspective).
cesses for achieving product synthesis, the major
advantage with the use of this route is the high
selectivity obtainable. This selectivity which can
be chiral (i.e., stereoselectivity), positional (i.e., References
region selectivity), and functional group specic
(i.e., chemo-selectivity) is highly desirable in Johannes T, Simurdiak MR, Zhao H (2006) Biocatalysis.
chemical synthesis since it may offer benets In: Lee S (ed) Encyclopaedia of chemical processing.
Taylor & Francis, Boca Raton
such as the none or reduced use of protecting
Wandrey C, Liese A, Kihumbu D (2000) Industrial bioca-
groups, a minimization of side reactions, ease in talysis: past, present, and future. Org Process Res Dev
downstream separation of products, and possible 4:286290
Biochemical Membrane Reactors 203
modied enzymes (10100 kDa) (Prazeres and

Biochemical Membrane Reactors Cabral 1994).
The biocatalyst is usually present in the reactor
Yamini Satyawali1, Ehiaze Augustine Ehimen2 system in two forms: free or immobilized at the
and Winnie Dejonghe2 membrane surface, or inside the membrane matrix B
1
Separation and Conversion Technology, pores. If the biocatalyst (i.e., enzyme) is present in
Vlaamse Instelling voor Technologisch its free form, its immobilization can be achieved
Onderzoek (VITO), Boeretang 200, Belgium by conning the enzyme to one side of the mem-
2
Separation and Conversion Technology, Flemish brane unit. This can be facilitated by size exclu-
Institute of Technological Research (VITO), Mol, sion, electrostatic repulsion, or enlargement via
Belgium chemical or physical immobilization techniques
onto an intermediate support (i.e., inert proteins
and gels). As for the direct immobilization of the
Synonyms biocatalyst onto the membrane, this can be
achieved by chemical binding, physical adsorp-
Biocatalytic membrane reactor tion, or electrostatic attraction.
Ideally, the products resulting from the biocat-
A biocatalytic membrane reactor (also referred to alytic process should permeate through the mem-
as a biochemical membrane reactor) is a system or brane pores either via a concentration gradient-
device that combines and utilizes the separation induced diffusion or a pressure gradient-induced
capabilities of membranes and the catalytic activ- convection. This would in turn facilitate a contin-
ities of whole cell or enzyme biocatalysts in one uous removal of products from the reaction media,
unit to facilitate the biochemical conversion of a process which has been considered to be essen-
substrates into desired products and the selective tial for biocatalytic membrane reactor concepts.
removal of the products in a single process unit. The selection of the membrane unit to be incor-
The removal of the product components from the porated into the reactor system must take into
reaction site could in turn improve the conversion account the respective sizes of the enzymes, sub-
of the substrates by favoring the product forma- strates, and products, as well as the chemical
tion reaction and is thus useful for reactions that characteristics of the reacting species and of the
are thermodynamically unfavorable or are product membrane. Here, the solute rejection coefcient is
inhibited. The separation of permeable solutes can an important membrane selection parameter for
be achieved from the reaction mixture by the the reactors and should be zero and one for the
action of a driving force that is applied across products and the biocatalysts, respectively
the membrane. As a result of coupling the conver- (Prazeres and Cabral 1994). This would allow
sion reaction with separation, lower operating for the product/coproduct permeation through
costs could therefore be achievable with this the membrane and a complete retention of the
type of reactor design compared to the use of biocatalysts inside the reaction system.
multiple reaction and product separation methods. The congurations of biocatalytic membrane
The complete retention of the biocatalyst reactors can be categorized into two main types
within the reactor system is the most important (Fig. 1) (Giorno and Drioli 2000):
requirement for the successful continuous opera-
tion of biocatalytic membrane reactors. Ultral- 1. The biocatalyst (i.e., whole cells, enzymes, or
tration membranes with a pore size distribution antibodies) are used in solution for the conver-
range (nominal molecular weight cutoff, sion of the substrate to the desired products in a
NMWCO, of 500100 kDa) are the most adequate reactor and further transported to a reaction
for the retention of a majority of native or vessel containing a selected membrane where
204 Biochemical Membrane Reactors
Biochemical Membrane 1. Retentate recycle

Reactors, Fig. 1 Main
conguration types for
biocatalytic membrane
reactors; (1) conversion
reactor combined with a CONVERSION
Reactants REACTOR
membrane unit,
(2) submerged membrane Biocatalyst Membrane
reactor with the membrane
functioning as a catalytic
and separation unit
(Adapted from Giorno and
Permeate
Drioli 2000)
2.
CONVERSION
REACTOR
Reactants
Retentate
Membrane with biocatalyst
encapsulated within
Permeate
Biochemical Membrane Reactors, Table 1 Advantages and disadvantages of biocatalytic membrane reactors
(Prazeres and Cabral 1994)
Advantages Disadvantages
Possibility of developing continuous conversion processes Unfavorable adsorption and enzyme-poisoning
issues
Improved control of reaction systems Enzyme deactivation due to shear-related effects
Increased productivities Concentration polarization
Contribute in a favorable shift of the reaction toward product side Likely product/substrate inhibition at the
membrane surface
Better conversion rates in product-inhibited reactions Membrane fouling
Good product enrichment and concentration in process stream Enzyme leakage
Facilitates control of the molecular weight of hydroxylates
Allows the conducting of multiphase reactions without
emulsication issues
Useful rened research tool for studying enzyme mechanisms
the biocatalysts (and unreacted substrates) can membrane acts as a support system for the
be recycled back to the conversion reactor, biocatalyst as well as a separation unit.
with the permeate selectively extracted from
the reaction stream. Here, the biocatalytic The membrane units of the membrane reactor
membrane reactor system might consist of a system can comprise of membranes which have a
conventional stirred-tank reactor combined at sheet shape (i.e., arranged in a plate and frame
with a membrane separation unit. module or a spiral wound module) or tubular
2. The biocatalyst constituents are immobilized (assembled in a tube and shell module)
within the membrane of the matrix. Here, the (Giorno and Drioli 2000) (Table 1).
Biochemical Processing 205
Cross-References major amount of pharmaceuticals was produced

via chemical synthesis processes. Nowadays, bio-
Biocatalytic and Biochemical Membrane technology productions dominate the pharmaceu-
Reactor tical market, and membrane technology
Membrane Bioreactors operations come to the fore due to the require- B
Principle of Biocatalytic Membrane Reactors ments of higher productivities.
(BMR) In fact, membrane processes have been used in
biochemical processing since well before the start
of the modern membrane industry. Membrane
References
technology for enzyme concentration, analysis of
Giorno L, Drioli E (2000) Biocatalytic membrane reactors; bacteriophages and viruses, preparation of cell-
applications and perspectives. Trends Biotechnol and protein-free ultraltrates from biological solu-
18:339349 tions, and sterile ltration has been widely used.
Prazeres D, Cabral JMS (1994) Enzymatic membrane bio- These systems were however limited to
reactors and their applications. Enzyme Microb
Technol 16:738750 analytical-scale processes due to limitations on
the available membranes and modules. Over the
last two decades, new membranes, modules, and
systems have been developed specically to meet
the requirements of the biotechnology industry
Biochemical Processing (Reis and Zdney 2007).
Particularly, in biochemical processes, with the
Yamini Satyawali1, Ehiaze Augustine Ehimen2 technological advancements such as increased
and Winnie Dejonghe2 product titers of cell cultures from 1 to 3 g/L up
1
Separation and Conversion Technology, to concentrations of 10 g/L, membrane unit oper-
Vlaamse Instelling voor Technologisch ations found a stronger application throughout the
Onderzoek (VITO), Boeretang 200, Belgium production, purication, and formulation of bio-
2
Separation and Conversion Technology, Flemish technology products. The membrane processes
Institute of Technological Research (VITO), Mol, are applied in upstream applications including
Belgium sterile ltration of fermentation media, pH control
solutions, and gases. Furthermore, microltration
may be used for turbid streams, virus ltration to
Biochemical processing refers to the use of bio- protect cell cultures from introduction of viral
logically derived or biochemical techniques to contaminants in media, and ultraltration to
meet the following goals: (i) to facilitate the con- remove, e.g., glycine, hypoxanthine, and thymi-
version of specic substrates to products, (ii) as a dine from serum. On the other hand, in down-
process pretreatment scheme for the production of stream processing, multiple operations are used
intermediates with the aim of achieving increased such as sterile ltration of buffers, products and
production yields or quicker conversion times, gases, and further membranes are integrated as
and (iii) to aid the recovery, recycling, or puri- sieving ltrations or as membrane chromatogra-
cation of a process intermediate or product. phy steps. Ultraltration is used to concentrate
Particularly in biochemical processes such as and buffer exchange (dialtration) products
fermentation, the need for purication and con- throughout the downstream process and is also
centration of the components is highly signicant. used as the method of choice for nal formulation
Cell removal, whole broth clarication, down- of bulk product. Membrane chromatography can
stream recovery, and purication of bioproducts be used for purication of products and raw mate-
from invariably dilute and often complex aqueous rials, e.g., the removal of endotoxins from raw
mixtures form an important aspect of biochemical materials, before using them in downstream
processing. Furthermore, up to a few years, the processing. Membrane technology has high
206 Biochemical Recovery
potential in the eld of biopharmaceutical produc- of the biochemical products. For example, for
tion, and this needs to be fully exploited in the some organic acid fermentations, i.e., citric acid
coming years (Frhlich et al. 2012). and glutamic acid, concentrations of 100 g/L or
In addition, bioconversion processes such as greater can be achieved (Wang 1983). Alterna-
production of biofuels from biomass (biorenery tively, with some more complex, specialized mac-
concept) are also a suitable example of biochem- romolecular products, i.e., vitamin B12 and human
ical processing. Typical technologies that have insulin, concentrations of 1 g/L or less are usually
been incorporated into the process include ion considered to be satisfactory since these products
exchange resins that are being used for the detox- command premium prices (Wang 1983). Apply-
ication of fermentation hydrolyzates, distillation ing schemes for concentrating the product in the
in the recovery of ethanol and ethanol dehydra- medium from which it must be recovered is thus
tion, and membrane separation for the removal of required and usually constitutes a signicant cost.
water from ethanol solution (Huang et al. 2008). Depending on the medium/reaction dilution levels
and the type of product recovery route applied, the
recovery costs and in turn total process costs could
References be quite high. The extent of product concentration
and purication achieved by the applied recovery
Frhlich H, Villian L, Melzner D, Strube J (2012) Mem- process has a direct implication on the product
brane technology in bioprocess science. Chem Ing Tech
costs and the protability of the process, with an
84:905917
Huang HJ, Ramaswamy S, Tschirner UW, Ramarao BV almost linear relationship demonstrated between
(2008) A review of separation technologies in current the selling price of biochemical products and the
and future bioreneries. Sep Purif Technol 62:121 concentrations of the products (Belfort 1989).
van Reis R, Zydney A (2007) Bioprocess membrane tech-
The product purication and its concentration are
nology. J Membr Sci 297:1650
thus the two main considerations and form the basis
guiding the selection of suitable downstream recov-
ery schemes for biochemical processes. Ideal recov-
Biochemical Recovery ery systems should therefore strive for the reduction
of the volumes of the medium, and a purication of
Yamini Satyawali1, Ehiaze Augustine Ehimen2 the desired products while concentrating.
and Winnie Dejonghe2 For the recovery of products from the dilute
1
Separation and Conversion Technology, reaction medium, in addition to the extent of the
Vlaamse Instelling voor Technologisch dilution of the desired product, the medium con-
Onderzoek (VITO), Boeretang 200, Belgium stitution and complexity is also a factor which
2
Separation and Conversion Technology, Flemish must be considered. This is because different clas-
Institute of Technological Research (VITO), Mol, ses of materials are usually present in the biochem-
Belgium ical process streams (i.e., suspended and colloidal
solids, microorganisms and cell debris, macromo-
lecular solutes and microsolutes) (Michaels and
An important step in any biochemical processing Matson 1985). The different size ranges exhibited
or conversion is the recovery of the process inter- by these medium components therefore pose a
mediates or products of interest. This is especially signicant separation issue which must be care-
important since most biochemical processes are fully considered before a recovery scheme is
conducted using very dilute conditions, with the applied. Furthermore, the fact that some biochem-
desired product concentration as required which ical products are normally susceptible to degrada-
in turn varies over a wide range. The nal product tion due to changes in the temperature, ionic
concentration and thus the extent of the product strength, pH, shear rate, and solvent conditions
recovery is highly dependent on the preceding adds to the need for careful planning of the product
process/reaction conditions and the market costs recovery route (Michaels and Matson 1985).
Biochemical Recovery 207
Biochemical Recovery, Table 1 Conventional and membrane bioseparation technologies for product recovery
(Michaels and Matson 1985)
Unit operation Conventional process or equipment Membrane process applicable
Cell harvesting Rotary vacuum drum Crossow microltration (MF)
Filtration Ultraltration (UF) B
Centrifugation
Whole broth clarication Centrifugation Crossow MF, UF
Protein concentration/purication Electrolyte/solvent precipitation UF
Afnity column chromatography Dialtration
Membrane modulated protein
precipitation
UF afnity purication
Membrane-bound afnity ligand
chromatography
Desalting Size-exclusion chromatography Electrodialysis
Microsolute concentration Vacuum evaporation Reverse osmosis
Pervaporation
Membrane distillation
Acid/base recovery Chemical treatment Bipolar electrodialysis
Ion exchange
Solvent extraction Podbielniak extraction Membrane contactor solvent extraction
Different approaches can be applied to achieve The application of membrane techniques and
an optimized recovery of the biochemical prod- its use for biochemical product recovery will be
ucts, and there is no universal optimized recovery the main emphasis of this section.
route applicable to every process; instead, the Table 1 provides a summary of common bio-
recovery options are usually selected to particu- chemical product recovery unit operations and
larly suit the individual process and product case examples of conventional separation techniques
under consideration. The selected recovery pro- usually applied for the desired product recovery.
cess is also quite important especially since the A comparison with the membrane processes
recovery costs for some products have been which can also be applied for such processes are
reported to be considerably greater than the also presented in Table 1.
bioprocessing costs for the production of the The selected biochemical recovery scheme,
product. especially with the use of membrane techniques,
The selection of a suitable, industrial scal- can either be carried out as a downstream process
able recovery process is an important step in the (where the biochemical process is run batchwise)
downstream processing of products. The use of or integrated with the biochemical reaction, i.e.,
recovery methods such as adsorption, centrifu- the use of in situ product recovery (or coproduct)
gation, electrophoresis, chromatographic, sol- methods. With the latter, the biochemical
vent extraction, and membrane processes has processing can be conducted continuously with
been largely applied for biochemical processes. the process reactants, and biocatalysts returned
The different available recovery technologies back or retained in the reactor. The use of such a
can be selected on the basis of the manipulation method, in addition to the concentration and
of specic molecular characteristics of the recovery of the desired product, might also help
desired product such as its molecular size, dif- improve the biochemical process to favor the
fusivity, ionic charge, vapor temperature and product formation (i.e., since its continuous
pressure, solubility, surface activity, and its removal will favor a shift of the reaction equilib-
density. rium toward product production).
208 Biocidal and Self-Polishing Surfaces
References more specically on nontargeted marine livings

due to their high persistence and toxicity. In recent
Belfort G (1989) Membranes and bioreactors: a technical years, some biocidal surfaces containing antimi-
challenge in biotechnology. Biotechnol Bioeng
crobial groups have been designed to kill and/or
33:10471066
Michaels AS, Matson SL (1985) Membranes in biotech- release the contacted microbes without leaching
nology: state of the art. Desalination 53:231258 out biocides. These nontoxic biocidal surfaces
Wang HY (1983) Integrating biochemical separation and have been developed and used in ship antifouling
purication steps in fermentation processes. Ann N Y
to meet the restrictive legislations (Siedenbiedel
Acad Sci 413:313321
and Tiller 2012).
A self-polishing surface means the surface is
able to keep smooth and clean through continuous
polishing from the hydrolysis and dissolution of
Biocidal and Self-Polishing Surfaces paint top layer in moist environment and hydro-
dynamic friction. The term self-polishing has
Li-Ping Zhu been introduced with the development of tributyl-
MOE Key Laboratory of Macromolecular tin self-polishing copolymer (TBT-SPC)-based
Synthesis and Functionalization, Department of antifouling paints (Cavas and Alyuruk 2013).
Polymer Science and Engineering, Zhejiang The rst use of the term in the marine paints
University, Hangzhou, Zhejiang, China industry is directly related with the working
mechanism of TBT-SPC paints. The possible use
of TBT acrylate esters in antifouling paints was
Biocidal and self-polishing surface is a surface rst suggested by Montermoso and co-workers in
that features the capability of preventing the set- 1958. TBT-SPC technology was then patented by
tlement and colonization of biofouling microor- Milne and Hails in 1974. The general working
ganisms through biocide release and surface mechanism of commercially used TBT-SPC
smoothing via a gradual hydrolysis of the paint depends on the hydrolysis of the ester bond
lm in wet environments. Biofouling is dened as between the TBT moiety and the carboxyl group
the undesirable adhesion, adsorption, accumula- of the polymer backbone. TBT acts as a broad-
tion of microorganisms, plants, animals, proteins, spectrum biocide and can be incorporated into
and cells, etc., on a wetted surface, and it is of high antifouling paints to inhibit fouling on a ship
concern in shipping industry. Biofouling on ves- hull up to 5 years. The TBT-SPC paints were so
sels increases hull roughness and hydrodynamic successful that they have been applied in about
drag, which leads to a decrease in vessel perfor- 70 % of commercial ship. However, the release of
mance and an increase in fuel consumption toxic TBT into seawater has brought detrimental
(Kirschner and Brennan 2012). Furthermore, effects on nontargeted marine organisms and
hull maintenance of extensively fouled ships is environments. As a result, many governments as
costly as it is difcult to remove the biolm. It is well as the International Maritime Organization
estimated that the fuel consumption can increase have restricted, even banned, the use of organatins
more than 40 % for a heavily fouled ship. Gov- in antifouling paints. In recent years, many
ernments and industries around the world spend researchers and companies have been devoting
more than US$ 5.7 billion annually to prevent and many efforts to explore non-tin self-polishing
control marine biofouling (Yebra et al. 2004). paints as a replacement of TBT-based paints
A biocidal surface is usually designed by impreg- (Lejars et al. 2012).
nation of coating with biocides which are released
to kill the near microbes and resist biofouling.
However, the release of heavy metal-based bio- Cross-References
cides from antifouling paint lms often leads to an
adverse impact on the marine environment and Biofouling
Biocidal Graphene Oxide Nanosheets 209
References fungi. Cell inactivation by graphene nanomaterials

occurs by a combination of cell entrapment, mem-
Cavas L, Alyuruk H (2013) A letter to editor: on the use of brane disruption, and the generation of reactive
the term self-polishing for antifouling paints. Prog
oxygen species, leading to membrane oxidative
Org Coat 76:16971698
Kirschner CM, Brennan AB (2012) Bio-inspired antifoul- stress. For all these mechanisms, direct contact B
ing strategies. Annu Rev Mater Res 42:211229 between bacterial cells and graphene sheets is
Lejars M, Margaillan A, Bressy C (2012) Fouling release required. The antimicrobial potential of graphene
coatings: a nontoxic alternative to biocidal antifouling
nanomaterials is dependent on the physicochemi-
coatings. Chem Rev 112:43474390
Siedenbiedel F, Tiller JC (2012) Antimicrobial polymers in cal properties of the material such as sheet size,
solution and on surfaces: overview and functional prin- conductivity, and degree of functionalization.
ciples. Polymers 4:4671 The biocidal potential of graphene and graphene
Yebra DM, Kiil S, Dam-Johansen K (2004) Antifouling
oxide sheets has been successfully used to impart
technology-past, present and future steps towards ef-
cient and environmentally friendly antifouling coat- antimicrobial properties to membranes. For water
ings. Prog Org Coat 50:75104 separation membranes, graphene oxide is usually
preferred over reduced graphene due to the more
hydrophilic nature of this material. The incorpora-
tion of graphene oxide sheets into a membrane for
Biocidal Graphene Oxide Nanosheets antimicrobial properties may therefore also
increase the hydrophilicity of the membrane sur-
Francois Perreault
face. Graphene oxide sheets can be incorporated in
School of Sustainable Engineering and the Built
membranes by including the nanomaterials in the
Environment, Arizona State University, Tempe,
polymer solution during fabrication or by post-
AZ, USA
fabrication surface functionalization. When
included in the polymer solution, graphene oxide
Biocidal graphene oxide nanosheets are used to also contributes to the overall mechanical strength
provide antimicrobial properties to surfaces. of the membrane and inuences its porosity and
Graphene is a two-dimensional nanomaterial transport properties (Yu et al. 2013). Since
composed of a single atomic layer of sp2-bonded graphene oxide sheets require direct contact for
carbon atoms. This material came into promi- cell inactivation, only the graphene oxide sheets
nence in 2010 by the awarding of the Nobel on the membrane surface contribute to the mem-
Prize in Physics to Andre Geim and Konstantin brane antimicrobial activity. Graphene oxide sheets
Novoselov for their research on the properties of deeper in the membrane structure do not participate
graphene. Graphene exhibits excellent electrical, in bacterial inactivation. In order to reduce the
mechanical, optical, and thermal properties that amount of material needed to impart antimicrobial
allow its widespread application in photovoltaic properties to a membrane, graphene oxide sheets
devices, transistors, sensors, and electronic visual can be concentrated on the membrane surface via
displays. Graphene possesses the highest theoret- surface functionalization methods. These may
ical specic surface area (2630 m2 g1) due to its involve direct binding of graphene oxide sheets to
two-dimensional, one-atom-thick structure and the membrane using the functional groups of
thus constitutes an excellent candidate for sorbent graphene oxide (Perreault et al. 2014), electrostatic
material used in water decontamination and gas interaction between the membrane surface and
separation (Zhu et al. 2010). Graphene also pos- functionalized graphene oxide (Choi et al. 2013),
sesses good antimicrobial activity and has found or the incorporation of the nanomaterial in a poly-
several applications for the development of antimi- mer composite that is deposited on the membrane
crobial surface coating. Graphene and its oxidized surface (Musico et al. 2014). Due to the hydrophilic
form, graphene oxide, possess a broad-spectrum properties of graphene oxide and the low slip
antimicrobial effect, being effective against gram- length of water on graphene sheets, the addition
positive bacteria, gram-negative bacteria, and of a graphene oxide layer on a membrane surface
210 Biocompatibility of Membranes
for antimicrobial properties do not affect water barriers and support devices. These structures are
transport properties. generally composed of biodegradable polymers,
complemented with additives, such as plasticizers,
emulsiers, and cross-linking agents. To be consid-
References ered biodegradable, all membrane components
must be degraded by the action of microorganisms
Choi W, Choi J, Bang J, Lee J-H (2013) Layer-by-layer and converted into water, carbon dioxide and/or
assembly of graphene oxide nanosheets on polyamide
methane, and new cell biomass.
membranes for durable reverse-osmosis applications.
ACS Appl Mater Interfaces 5:1251012519 The wide range of polymers used in the devel-
Musico YLF, Santos CM, Dalida MLP (2014) Rodrigues opment of biodegradable membranes enables the
DF Surface modication of membrane lters using production of structures with quite diverse proper-
graphene and graphene oxide-based nanomaterials for
ties, nding applications in different areas. It is
bacterial inactivation and removal. ACS Sustain Chem
Eng. 2:15591565 early online. doi:10.1021/sc500044p interesting to note that some of them, beyond
Perreault F, Tousley ME, Elimelech M (2014) Thin-lm being biodegradable, are also biocompatible
composite polyamide membranes functionalized with and/or edible, extending their use to the elds of
biocidal graphene oxide nanosheets. Environ Sci
biomedicine and edible coatings for food products.
Technol Lett 1:7176
Yu L, Zhang Y, Zhang B, Liu J, Zhang H, Song C (2013) Biodegradable polymers obtained by chemical
Preparation and characterization of HPEI-GO/PES synthesis have been used, for example, in food
ultraltration membrane with antifouling and packaging materials (Siracusa et al. 2008) and
antibacterial properties. J Membr Sci 447:452462
medical applications (e.g., fracture xation, dental
Zhu Y, Murali S, Cai W, Li X, Suk J, Potts JR, Ruoff RS
(2010) Graphene and graphene oxide: synthesis, prop- orthopedic implants, articial skin, suture
erties, and applications. Adv Mater 22:39063924 anchors, drug delivery) (Yen et al. 2009;
Armentano et al. 2010; Bettahalli et al. 2011).
They have also been applied for separation of
organic mixtures by pervaporation (Zereshki
Biocompatibility of Membranes
et al. 2010, 2011).
They include the following:
Membrane Biocompatibility
(i) Polyglycolic acid (PGA), an aliphati-
caromatic copolymer, which combines the
Biodegradable Membrane excellent material properties of aromatic
polyethylene terephthalate and the biode-
Isabel Coelhoso1, Filomena Freitas2, gradability of aliphatic polyesters. It is pro-
Vitor D. Alves3 and Maria A. M. Reis2 duced by a polycondensation reaction of
1
LAQV- REQUIMTE, Departamento de glycol and aliphatic dicarboxylic acids,
Qumica, Faculdade de Ciencias e Tecnologia, which may be obtained from renewable
Universidade Nova de Lisboa, Caparica, Portugal resources.
2
UCIBIO-REQUIMTE, Departamento de (ii) Polylactic acid (PLA), a thermoplastic ali-
Qumica, Faculdade de Cincias e Tecnologia, phatic polyester obtained from polymeriza-
Universidade Nova de Lisboa, Caparica, Portugal tion of the lactic acid monomer produced by
3
LEAF Linking Landscape, Environment, microbial fermentation.
Agriculture and Food, Instituto Superior de (iii) Polycaprolactone (PCL), a thermoplastic
Agronomia, Universidade de Lisboa, Lisboa, polymer obtained by chemical synthesis
Portugal using nonrenewable resources (petrochemi-
cal derivatives).
Biodegradable membranes consist of thin-lm-like Among the biodegradable polymers produced

structures usually applied as separating selective by microbial fermentation, polysaccharides (e.g.,
Biodegradable Organic Matter 211
pullulan, hyaluronan, gellan, GalactoPol, curdlan, Eisenbart E (2007) Biomaterials for tissue engineering.
bacterial alginate, bacterial cellulose) (Freitas Adv Eng Mat 9:10511060
Freitas F, Alves VD, Reis MAM (2011) Advances in bacte-
et al. 2011) and polyesters (e.g., polyhydrox- rial exopolysaccharides: from production to biotechno-
yalkanoates) have been applied in lms and edible logical applications. Trends Biotechnol 29:388398
coatings (e.g., pullulan, gellan) (Nieto 2009; Nieto MB (2009) Structure and function of polysaccharide B
Alves et al. 2011). Both microbial polysaccharides gum-based edible lms and coatings. In: Embuscado
ME, Huber KC (eds) Edible lms and coatings for food
and polyesters show a wide range of properties that applications. Springer, New York, pp 57112
may be tuned by manipulating the bioreaction con- Philip S, Keshavarz T, Roy I (2007) Polyhydrox-
ditions. Polyhydroxyalkanoates show a wide range yalkanoates: biodegradable polymers with a range of
of applications, such as in industry (e.g., packag- applications. J Chem Technol Biotechnol 82:233247
Siracusa V, Rocculi P, Romani S, Rosa MD (2008) Biode-
ing, waterproof paperboard), medicine (e.g., bone gradable polymers for food packaging: a review.
plates, osteosynthetic materials, surgical sutures, Trends Food Sci Technol 19:634643
and dressing materials for surgery), and agriculture Yen C, He H, Lee LJ, Winston Ho WS (2009) Synthesis
(e.g., mulch lms) (Philip et al. 2007). and characterization of nanoporous polycaprolactone
membranes via thermally- and nonsolvent-induced
The polymers recovered from natural products phase separations for biomedical device application.
generally used to produce biodegradable mem- J Membr Sci 343:180188
branes include polysaccharides (e.g., starch, cel- Yu L, Dean K, Li L (2006) Polymer blends and composites
lulose, pectin, alginate, carrageenan, chitosan) from renewable resources. Prog Polym Sci 31:576602
Zereshki S, Figoli A, Madaeni SS, Simone S, Jansen JC,
and proteins (e.g., gelatin/collagen, soy protein, Esmailinezhad M, Drioli E (2010) Poly(lactic acid)/poly
gluten). These polymers are widely used to (vinyl pyrrolidone) blend membranes: effect of mem-
develop edible coatings for food products and brane composition on pervaporation separation of etha-
biodegradable lms intended for food packaging nol/cyclohexane mixture. J Membr Sci 362:105112
Zereshki S, Figoli A, Madaeni SS, Galiano F,
(Nieto 2009). Chitosan, alginate, and collagen are Esmailinezhad M, Drioli E (2011) Pervaporation sepa-
also referred to be applied in tissue engineering ration of ethanol/ETBE mixture using poly(lactic acid)/
(Eisenbarth 2007). poly(vinyl pyrrolidone) blend membranes. J Membr
Polysaccharide-based membranes, such as Sci 373:2935
chitosan and sodium alginate, have received much
attention for solvent dehydration by pervaporation,
due to their good lm-forming properties, chemical
resistance, and high permselectivity for water Biodegradable Organic Matter
(Chapman et al. 2008; Yu et al. 2006).
Isabel Coelhoso1, Filomena Freitas2,
Vitor D. Alves3 and Maria A. M. Reis2
1
References LAQV-REQUIMTE, Departamento de Qumica,
Faculdade de Ciencias e Tecnologia,
Alves VD, Ferreira AR, Costa N, Freitas F, Reis MAM, Universidade Nova de Lisboa, Caparica, Portugal
Coelhoso IM (2011) Characterization of biodegradable 2
UCIBIO-REQUIMTE, Departamento de
lms from the extracellular polysaccharide produced
Qumica, Faculdade de Cincias e Tecnologia,
by Pseudomonas oleovorans grown on glycerol
byproduct. Carbohydr Polym 83:15821590 Universidade Nova de Lisboa, Caparica, Portugal
Armentano I, Dottori M, Fortunati E, Mattioli S, Kenny JM
3
LEAF Linking Landscape, Environment,
(2010) Biodegradable polymer matrix nanocomposites Agriculture and Food, Instituto Superior de
for tissue engineering: a review. Polym Degrad Stab
95:21262146
Agronomia, Universidade de Lisboa, Lisboa,
Bettahalli N, Steg H, Wessling M, Stamatialis D (2011) Portugal
Development of poly(L-lactic acid) hollow ber mem-
branes for articial vasculature in tissue engineering
scaffolds. J Membr Sci 371:117126
Chapman PD, Oliveira T, Livingston AG, Li K (2008)
Biodegradable organic matter is organic material,
Membranes for dehydration of solvents by plant, and animal matter with origin in living
pervaporation. J Membr Sci 318:537 organisms, which can be converted by the action
212 Biodegradable Organic Matter
Biodegradable
Polymers
Biopolymers Polymers
(renewable resources) (fossil resources)
PCL
Extracted from Biomonomer Produced by
Biomass Synthesis Microorganisms
PGA
PLA PHA PBSA
Proteins Polysaccharides
Pullulan
Whey protein Pectin

Glucan
Casein Carrageenan
Soy protein Starch
Gluten Cellulose
Biodegradable Organic Matter, Fig. 1 Biodegradable polymers
of microorganisms to water, carbon dioxide, main structural elements of plant and animal exo-
and/or methane and biomass. skeleton (e.g., cellulose, carrageenan, chitin) or
Organic materials can be used to obtain biode- have a key role in the plant energy storage (e.g.,
gradable polymers which are classied according starch).
to the method of production or their source Cellulose and starch are of prime interest as
(Fig. 1): biopolymers because of their availability and
rather low cost. A variety of polysaccharides and
Polymers directly extracted or removed from their derivatives, besides starch and cellulose
biomass such as polysaccharides and proteins derivatives, have been used as biodegradable
Polymers produced by classical chemical syn- membrane-forming matrixes, including alginate,
thesis starting from renewable bio-based pectin, carrageenan, chitin, and various gums.
monomers such as polylactic acid (PLA) Several protein sources have been proposed for
Polymers produced by microorganisms or the preparation of biopolymers, in particular,
genetically modied bacteria such as cereal proteins which are available in large
polyhydroxyalkanoates, bacterial cellulose, amounts as by-products arising from agricultural
xanthan, and pullulan (Mensitier et al. 2011) and biofuel processing activities such as ethanol
production. These protein-rich products include
Polysaccharides are the most abundant macro- spent grain from the brewing and distilling indus-
molecules in the biosphere. These complex car- tries, cereal bran streams from milling, and protein
bohydrates constituted of monosaccharides joined residues from starch extraction activities
together by glycosidic bonds are often one of the (Mensitier et al. 2011).
Biodegradable Polymer for Membranes 213
Microbial biopolymers are naturally synthe-

sized by microorganisms with different functions Biodegradable Polymer
in the microbial cell, including intracellular car- for Membranes
bon or energy storage reserves (e.g., glycogen,
polyesters), structural cell wall components (e.g., Simona Salerno B
chitin, b-glucans), and extracellular biopolymers Institute on Membrane Technology, National
(e.g., exopolysaccharides), often secreted as pro- Research Council of Italy, ITM-CNR, Rende,
tective mechanisms in response to environmental Italy
conditions (Rehm 2010).
A wide range of agro-food and industrial
wastes/by-products have been proposed as alter- Synonyms
native substrates for the production of microbial
biopolymers, including molasses, cheese whey, Biodegradable polymeric membranes; Biopolymers;
palm date syrup, olive mill wastewater (OMW), Polymers for drug delivery; Polymers for tissue
glycerol by-product from the biodiesel industry, engineering
corn-steep liquor, spent malt grains, apple and
grape pomaces, citrus peels, peach pulp, used
oils, and several acid-hydrolysate wastes (e.g., Biodegradable Polymer for Membranes
melon, watermelon, cucumber, tomato, rice),
among others (Freitas et al. 2011; Verlinden Membranes made of biodegradable polymers are
et al. 2011). stable and durable for a specic period of time
The microbial biopolymers produced include related to their intended purpose or use in a par-
polysaccharides, polyamides, polyesters, and ticular application; after that they break down to
polyanhydrides. Depending on their composition result in natural by-products such as gases (CO2,
and molecular weight, they have properties rang- N2), water, biological materials, and inorganic
ing from rheology modiers to bioplastics, which salts.
makes them useful in many industrial applications Biodegradable polymers are naturally and bio-
(e.g., agro-food, cosmetics, pharmaceutical, tex- logically derived or synthetically made and con-
tile) (Rehm 2010). sist of ester, amide, and ether functional groups or
bonds.
Naturally and biologically derived polymers
are found in nature or created by living organisms.
References They include polysaccharides, like starches found
in potatoes or wood, and proteins, such as animal-
Freitas F, Alves VD, Reis MAM (2011) Advances in based whey or plant-derived gluten. Polysaccha-
bacterial exopolysaccharides: from production to bio- rides consist of glycosidic bonds, between a sac-
technological applications. Trends Biotechnol
29(8):388398 charide and an alcohol; proteins of peptide bonds
Mensitieri G, Di Maio E, Buonocore G, Nedi I, Oliviero M, or amide functional groups between amino acids.
Sansone L, Iannace S (2011) Processing and shelf life Synthetic materials are man-made. The most
issues of selected food packaging materials and struc- important group of synthetic biodegradable poly-
tures from renewable resources. Trends Food Sci
Technol 22:7280 mers is polyesters that include polyhydroxy-
Rehm BHA (2010) Bacterial polymers: biosynthesis, mod- butyrate and polylactic acid. They can be
ications and applications. Nat Rev Microbiol synthesized in a number of ways including direct
8:578592 condensation of alcohols and acids, ring opening
Verlinden RAJ, Hill DJ, Kenward MA, Williams CD,
Piotrowska-Seget Z, Radecka IK (2011) Production of polymerizations (ROP), and metal-catalyzed
polyhydroxyalkanoates from waste frying oil by polymerization reactions. A wide variety of
Cupriavidus necator. AMB Express 2011:111 starting materials can be used to synthesize
214 Biodegradable Polymer for Membranes
polyesters, like cyclic dimeric glycolic or lactic eliminated from the body via natural metabolic
acid to form a-hydroxy acids which then poly- pathways. If they are of natural origin, they should
merize into poly(a-esters), poly(b-esters), and be further incorporated in the new forming tissue
poly(g-esters). Other synthetic biodegradable matrix (Williams 2014). Commonly used natural
polymers are polyanhydride and poly(ester biodegradable polymers are different proteic
amide)s, made by condensation, dehydrochlorina- derivatives of the extracellular matrix, such as
tion, dehydrative coupling, and ROP, and poly- collagen, elastin, keratin, and brin, and
urethanes, typically synthesized using a polysaccharidic materials, like chitosan and gly-
diisocyanate, a diol, and a polymer chain cosaminoglycans (GAGs). Although they have all
extender. proved suitable in terms of cell compatibility,
Biopolyesters are produced by both living some issues with potential immunogenicity still
organisms and synthetic processes. These metal- remains. Among the synthetic materials, polyure-
free processes involve the use of bacterial or thanes and poly(ester amide)s were initially used
enzymatic catalysis, with the benet of generally for their biocompatibility, durability, and resil-
regioselective and stereospecic reactions, but ience and are only more recently being investi-
with the drawbacks of the high cost of bacteria gated for their biodegradability. A commonly
and enzymes, long reaction times, and products used synthetic polymer is polylactic acid (PLA),
of low molecular weight. Examples of a polyester which degrades within the human
biopolyesters are poly(glycolic) acid and body to form lactic acid, a naturally occurring
polyhydroxyalkanoates (PHA) (Zhang et al. 2014). chemical which is easily removed from the body.
Membranes made of biodegradable polymers Similar materials are polyglycolic acid (PGA) and
have numerous applications in a variety of elds polycaprolactone (PCL): their degradation mech-
including medicine, agriculture, and packaging. anism is similar to that of PLA, but they exhibit,
In the biomedical eld, they are of signicant respectively, a faster and a slower rate of degrada-
interest in applications such as tissue engineering tion compared to PLA. Polyglycolide, polylactide,
and drug delivery. For these applications biode- polyhydroxybutyrate, chitosan, hyaluronic acid,
gradable polymers must answer to specic and hydrogels are used in orthopedic applications,
requirements: they must be nontoxic, capable of such as bone and joint replacement; poly(2-
maintaining good mechanical integrity until hydroxyethyl-methacrylate), polyethylene glycol,
degraded, and capable of controlled rates of deg- chitosan, and hyaluronic acid are extensively
radation. Their products of degradation require used in the repair of cartilage, ligaments, and
also low or negligible toxicity in terms of both tendons (Chiono et al. 2014; Jagur-Grodzinski
local tissue response and systemic immune 2006).
response. Biodegradable polymeric membranes are of
Biodegradability in tissue engineering is often great interest and benet as drug delivery systems
an essential factor since scaffolds should prefera- and carriers, where it is critical to target and
bly be absorbed by the surrounding tissues with- slowly release the payload drug over time and
out the necessity of a surgical removal. The rate at into a specic site of the body. The polymer
which degradation occurs has to coincide as much slowly degrades into nontoxic and naturally
as possible with the rate of tissue formation: this metabolized smaller fragments releasing in the
means that while cells are building their own meantime the drug in a controlled fashion. Biode-
natural matrix structure around themselves, the gradable polymers for this purpose must be capa-
scaffold should be able to provide structural integ- ble to be formulated for prolonged release of
rity within the body and eventually to break down pharmaceutically active compounds. Encapsulat-
without evoking a foreign body response and ing the therapeutic in a polymer and adding
leaving the newly formed tissue. They should be targeting agents further decreases the toxicity of
Biodegradable Polymeric Membranes 215
the drug to healthy cells. Polyanhydrides are very route is through biological processes which can
attractive in drug delivery because they only be further broken down into aerobic and anaerobic
degrade from the surface and so are able to release processes. Microorganisms and enzymes that
the payload drug at a constant rate. PLA, poly(- have the ability to break down natural polymers
lactic-co-glycolic) acid (PLGA), and PCL have into even simpler units act through oxidation or B
been used to carry anticancer drugs. hydrolysis reactions. Examples of key enzymes
In addition to medicine, biodegradable poly- include proteases, esterases, glycosidases, and
mers are often used to reduce the volume of waste manganese peroxidases.
in packaging materials derived from petrochemi-
cals. One of the most commonly used biodegrad-
able polymers for packaging purposes is PLA, Cross-References
used for a variety of lms, such as wrappings,
shopping bags, or trash bags, as well as con- Bioarticial Organs and Tissue
tainers, including bottles and cups. Since 2002, Bioarticial Organs, Membrane Operations of
PLA has been approved by the Food and Drug Biodegradable Membrane
Administration (FDA) as a safe material to use in Biodegradation
all food packaging. Biopolymers
The breakdown of these polymers depends on Drug Delivery by Membrane Operations
a variety of factors including the polymer proper- Drug Release
ties and the environment in which they are. Com- Drug Delivery, Application of
plete biodegradation occurs when there are no Tissue Engineering, Membrane Operations of
oligomers or monomers left. Polymer properties
that inuence degradation are bond type, molec-
References
ular weight, chain branching, cross-linking,
copolymers, crystallinity, hydrophobicity, and Chiono V, Nardo T, Ciardelli G (2014) Bioarticial bio-
solubility. Factors of the surrounding environ- materials for regenerative medicine applications. In:
ment, either the atmosphere or the body, include Orlando G (ed) Regenerative medicine applications in
pH, temperature, the presence and concentrations organ transplantation. Academic/Elsevier, Waltham,
pp 113136
of microorganisms or enzymes, and amount of Jagur-Grodzinski J (2006) Polymers for tissue engineering,
water. Biodegradable polymers with extremely medical devices and regenerative medicine. Concise
strong carbon backbones are difcult to break, general review of recent studies. Polym Adv Technol
so degradation often starts from the end groups. 17:395418
Williams DF (2014) Strategies for the specication of
Biodegradable polymers with minimal chain tissue engineering biomaterials. In: Orlando G
branching and cross-linking and with low crystal- (ed) Regenerative medicine applications in organ trans-
linity and degree of polymerization often decrease plantation. Academic/Elsevier, Waltham, pp 3137
the number of end groups per unit weight but Zhang Z, Ortiz O, Goyal R, Kohn J (2014) Biodegradable
polymers. In: Lanza R, Langer R, Vacanti JP (eds)
increase their surface area and an easy access for Principles of tissue engineering, 4th edn. Elsevier
reaction with the degradation initiator. Hydropho- Academic Press, London/Waltham/San Diego,
bic polymers cannot easily get in contact with pp 441473
water soluble enzymes decreasing their enzyme
interaction and degradation. There are two pri-
mary mechanisms through which biodegradation
can occur. One is through physical decomposition Biodegradable Polymeric
through reactions such as hydrolysis and Membranes
photodegradation, which can lead to partial or
complete degradation. The second mechanistic Biodegradable Polymer for Membranes
216 Biodegradation
(3) conversion of the intermediates into metabolic

Biodegradation products (Fig. 1). Water soluble polymers, such as
the natural polysaccharides xanthan, alginate, and
Isabel Coelhoso1, Filomena Freitas2, pectin, are usually more easily degraded, while
Vitor D. Alves3 and Maria A. M. Reis2 water insoluble polymers (e.g., polyesters) are
1
LAQV-REQUIMTE, Departamento de Qumica, more difcult to degrade because the process is
Faculdade de Ciencias e Tecnologia, heterogeneous. Abiotic processes, such as chemi-
Universidade Nova de Lisboa, Caparica, Portugal cal degradation, thermal hydrolysis, or
2
UCIBIO-REQUIMTE, Departamento de photodegradation, coexist with the biotic process
Qumica, Faculdade de Cincias e Tecnologia, and often directly induce it by making the materials
Universidade Nova de Lisboa, Caparica, Portugal more easily accessible to the microorganisms.
3
LEAF Linking Landscape, Environment, Hence, environmental conditions (e.g., pH, tem-
Agriculture and Food, Instituto Superior de perature, moisture, salinity, presence/absence of
Agronomia, Universidade de Lisboa, Lisboa, oxygen, nutrient supply) are important factors that
Portugal inuence both the materials themselves and the
microbial activity. The complexity of the materials,
namely, the structure and composition (single poly-
Biodegradation is the process by which microor- mer or polymer blends), the presence of additives
ganisms (bacteria, fungi, algae) change the chem- (e.g., plasticizers), morphology, and crystallinity,
ical structure of a material into naturally occurring are also key issues that directly inuence their
metabolic products, ultimately resulting in its con- accessibility to the microbial enzymes.
version into water, carbon dioxide and/or meth- The standardized methods for testing the bio-
ane, and new cell biomass (Muller 2005). The degradability of polymeric materials are based on
biodegradation of polymeric materials starts with previously developed methods for low-molecular
their depolymerization to generate lower molecu- weight substances. However, they have been
lar weight intermediates that the microbial cells modied to consider their higher complexity. Dif-
can then assimilate. The general mechanism of the ferent biodegradability tests can be performed,
biodegradation process includes three phases: under both controlled laboratory conditions and
(1) secretion of extracellular enzymes by the also natural environmental conditions. In the lab-
microorganisms; (2) generation of short interme- oratory tests, dened environments are created
diates by the action of the secreted enzymes; and using one of two approaches: (1) inoculation of
microorganism microorganism
microorganism
CO2
extracellular low-molecular
enzymes CH4 H2O
weight
intermediates metabolic
products
polymer polymer polymer

(1) (2) (3)
Biodegradation, Fig. 1 General mechanism of the bio- intermediates by the action of the secreted enzymes, and
degradation process: (1) secretion of extracellular enzymes (3) conversion of the intermediates into metabolic products
by the microorganisms, (2) generation of short
Bioelectrochemical MBR 217
Biodegradation, Laboratory tests Field tests

Fig. 2 Biodegradation
tests
Individual Lab reactor Natural

cultures with natural Environment B
matrix (water, soil,
(water, soil, compost,
Mixed compost, landfill)
cultures landfill)
Synthetic Complex Complex

media media media
Defined conditions Variable conditions
synthetic media with individual microorganisms to the microbial environment and the amount
especially selected for a given polymer or with a of 14C generated is estimated. It is a high pre-
mixed microbial culture originating, for example, cision test, but labeled materials are not always
from a wastewater treatment plant or (2) use of available and are expensive.
complex materials collected from the environ- Clear-zone formation: the polymer is placed in
ment (e.g., water, soil, compost) as the matrix for agar plates as very ne particles and inoculated
the tests (Fig. 2). with the microbial culture, being the formation of
Some of the standard testing methods include a clear halo the indication of depolymerization.
(Shah et al. 2008):
Visual observations: macroscopic changes, References

such as roughening of the surface, formation
of hole or cracks, changes in color, etc., can be Muller R-J (2005) Biodegradability of polymers: regula-
tions and methods for testing. In: Steinb uchel A
used as a rst indication of microbial attack,
(ed) Biopolymers, vol 10, General aspects and special
while more sophisticated observations using, applications. Wiley, Weinheim
for example, scanning electron microscopy Shah AA, Hasan F, Hameed A, Ahmed S (2008) Biological
(SEM) or atomic force microscopy (AFM) degradation of plastics: a comprehensive review.
Biotechnol Adv 26:246265
can be made to obtain more information
about the degradation mechanism.
Weight loss measurements: determination of
residual material after being exposed to the Bioelectrochemical MBR
biodegradation conditions.
Changes in mechanical properties and Zhiwei Wang
molar mass: properties, such as tensile strength, Department of Environmental Engineering,
can be used as indicators of degradation since College of Environmental Science and
they are related with the polymers molar mass. Engineering, Tongji University, Shanghai, China
CO2 evolution or O2 consumption: under aer-
obic conditions, the microorganisms use O2 for
carbon oxidation, generating CO2 as one of the Synonyms
major metabolic end product. Hence, CO2 evo-
lution or O2 consumption are good indicators Bioelectrochemical membrane bioreactor (MBR);
of polymer degradation. Bioelectrochemically assisted membrane bioreac-
Radio labeling: materials containing a ran- tor; Electrochemical membrane bioreactor; Mem-
domly distributed 14C marker can be exposed brane bioelectrochemical reactor
218 Bioelectrochemical MBR
Introduction and cathode chambers (Fig. 1a). Proton exchange

membrane (PEM) and anion exchange membrane
Bioelectrochemical MBR is a novel and effective (AEM) are initially utilized in bioelectrochemical
integration of membrane bioreactor process and MBRs because these membranes are designed to
microbial fuel cell technology, which can recover conduct protons and anions while being imperme-
energy from wastewater and simultaneously har- able to other reactants (e.g., O2). Alternatively,
vest clean water for reuse (Wang et al. 2011, substitutes, such as microltration membrane,
2013a, b; Huang et al. 2014). As shown in nylon meshes, and perforated plates (Wang
Fig. 1, a typical bioelectrochemical MBR consists et al. 2012), have been also used in bioelectro-
of an anode chamber (or anaerobic tank) and a chemical MBR to separate anode and cathode
cathode chamber (or membrane tank). In the chambers regarding the construction expenditure
anode chamber, fuel (e.g., organic matter in the (Fig. 1b). As a consequence, coulombic efciency
inuent wastewater) is oxidized by can be substantially decreased due to unwanted
exoelectrogenic bacteria, generating CO2, pro- substrate diffusion.
tons, and electrons. Electrons are then transferred In bioelectrochemical MBRs, anodes are
to the cathode chamber (or membrane tank) always made of carbon cloth, carbon brush, and
through an external electric circuit, while protons carbon paper. The high specic surface area
are transferred to the cathode chamber through the (1002,000 m2/g) of these materials is conducive
separators (e.g., membranes). Electrons and pro- for habitation of exoelectrogenic bacteria. Acti-
tons are nally consumed on the surface of cath- vated carbon can be also used as the anode mate-
odes to nish the reduction reaction (e.g., oxygen rial, especially in the application of wastewater
reduction reaction). In these devices, membrane treatment. Cathodes can be independent or com-
modules are placed in the cathode chamber to bined with membrane modules. For example,
enable a sound solid-liquid separation. stainless steel mesh has been introduced to
Separators with a high conductivity of protons constructed cathodes, and the dynamic membrane
and a low substrate diffusion coefcient are wel- formed on the cathode subsequently enables the
come in bioelectrochemical MBRs in order to ltration process (Wang et al. 2011, 2013b). Com-
achieve a selective mass transfer between anode posite cathodes make the reactor more compact,
a b c External resistance
Membrane Cathode
Anode
H++CO2+e Anode chamber Cathode chamber PEM (or AEM)
Organics
Substitute
PEM (or AEM), independent cathode Substitute separator Composite cathode
Bioelectrochemical MBR, Fig. 1 Schematic representation of bioelectrochemical MBRs

Biofuel 219
which simultaneously alleviates membrane foul-

ing (Liu et al. 2013). Electrochemical MBRs have Biofouling
been also developed by using external power to
replace the generated electricity of MFC. Bacterial Biolm Formation
B
References
Biofuel
Huang J, Wang ZW, Zhu CW, Ma JX, Zhang XR, Wu ZC
(2014) Identication of microbial communities in open
and closed circuit bioelectrochemical MBRs by high-
Diego Luna
throughput 454 pyrosequencing. PLoS ONE 9(4): Department of Organic Chemistry, University of
e93842 Cordoba, Cordoba, Spain
Liu JD, Liu LF, Gao B, Yang FL (2013) Integration of
bio-electrochemical cell in membrane bioreactor for
membrane cathode fouling reduction through electric-
ity generation. J Membr Sci 430:196202 In biofuel production by lipases, membranes are
Wang YK, Sheng GP, Li WW, Huang YX, Yu YY, Zeng innovative devices used for different tasks. In this
RJ, Yu HQ (2011) Development of a novel bioelectro- respect, they are applied as membrane-
chemical membrane reactor for wastewater treatment.
immobilized lipases to get the repeated use of
Environ Sci Technol 45(21):92569261
Wang YP, Liu XW, Li WW, Li F, Wang YK, Sheng GP, expensive biocatalysts in biodiesel fabrication.
Zeng RJ, Yu HQ (2012) A microbial fuel cellmembrane In this way, the enzyme cost which is being cur-
bioreactor integrated system for cost-effective wastewa- rently the main obstacle for its use in biodiesel
ter treatment. Appl Energ 98(0):230235
Wang Z, Huang J, Zhu C, Ma J, Wu Z (2013a)
production can be overcome by lipase immobili-
A bioelectrochemically-assisted membrane bioreactor zation with the aim of reducing the production
for simultaneous wastewater treatment and energy pro- cost by reusing the enzymes. Consequently,
duction. Chem Eng Technol 36:20442050 many research works are currently developed to
Wang YK, Sheng GP, Shi BJ, Li WW, Yu HQ (2013b)
get efcient, robust, and cost effective membranes
A novel electrochemical membrane bioreactor as a
potential net energy producer for sustainable wastewa- able to obtain immobilization of lipases in biodie-
ter treatment. Sci Rep UK 3:1864 sel production.
On the other hand, in the production of biodie-
sel, separation and purication of biodiesel is a
critical technology. Conventional technologies
Bioelectrochemical Membrane used for biodiesel separation, such as gravity set-
Bioreactor (MBR) tling, decantation, ltration and biodiesel puri-
cation by water washing, acid washing, and
Bioelectrochemical MBR absorbents have proven to be inefcient. In this
respect, membrane reactors are being developed
to produce biodiesel from several oils with meth-
Bioelectrochemically Assisted anol via enzymatic catalysis as well as by acid-
Membrane Bioreactor and base-catalysis. Thus, membrane reactors
which exhibit intrinsic characteristics of ef-
Bioelectrochemical MBR ciency, operational simplicity and exibility, rela-
tively high selectivity and permeability, low
energy requirements, good stability under a wide
spectrum of operating conditions, and easy con-
Biofabrication trol have been conrmed in a numerous variety of
applications and operations, such as molecular
Bioarticial Organs and Tissue separations, fractionations, concentrations, and
Tissue Engineering, Membrane Operations of purications within a wide environmental
220 Biofuel Production by Membrane Operations
compatibility. In this respect, novel two-phase the productivity, biomass conversion efcacy,
membrane reactors seems to be particularly useful etc. The signicance of these procedures is more
in removing unreacted oil from the fatty acid than recovery and purication. By in situ removal
methyl ester (FAME) product yielding a high of one of more products in some cases, they can
purity biodiesel. Some research could also be displace the chemical reaction equilibriums or
conducted on technologies for biodiesel produc- relieve the toxic effects constraining the conver-
tion by utilizing membrane reactors to handle sion. With this feature, the system is named as
waste oil in biodiesel production. biofuel production by membrane operations or
Glycerol also causes strong inhibition in the membrane reactor for biofuel production. The
enzymatic methanolysis of triglycerides; therefore membranes incorporated can be passive ones
glycerol removal by dialysis is carried out using that only carry out separation of inhibitory prod-
at sheet membrane modules. Thus, membrane uct or active ones that conduct catalysis of reac-
lter devises are useful tools for obtain cleaning tions and separation simultaneously.
of glycerol in biodiesel production. Membrane In bioalcohol production through biochemical
separation seems to be the most suitable method- conversion, the free cell continuous fermentation
ology for this purpose. Consequently, the involve- eliminates downtime, and cell immobilization fer-
ment of membrane reactor and separative mentation increases cell concentration (Quresh
membrane shows great promise for the separation and Ezeji 2008). However, the cell washout
and purication of biodiesel. Membrane technol- occurs at high dilution rate in the former. As for
ogy are currently being explored and exploited to the latter, the limitation on substrate supply and
overcome the difculties usually encountered in toxic product diffusion inactivates a large amount
the separation and purication of biodiesel. In this of microbes in the latter. It is difcult for the cells
respect, both conventional and most recent mem- located at the bottom of the biolm to get enough
brane technologies used in rening biodiesel are carbon sources for survival, and the products and
being studied (Atadashi et al. 2011). other metabolites that have inhibitory effect on the
viability of microorganisms encounter great resis-
tance in diffusing away from the cell surround-
References ings. As a solution to these problems reducing the
productivity, cell recycle membrane system,
Atadashi IM, Aroua MK, Abdul-Aziz A (2011) Biodiesel which can be called as membrane bioreactor,
separation and purication: a review. Renew Energy
was developed (Quresh and Ezeji 2008).
36:437443
The micro-sized cells suspended in broth are
rejected by the membrane and recycled back to the
reactor as retentate. Figure 1 is the schematic of
the cell recycle membrane system. As a result,
Biofuel Production by Membrane their concentration can be maintained. The sol-
Operations vents containing the products, on the contrary,
pass through the membrane. If the membranes
Lan Ying Jiang used have selectivity toward the target product,
School of Metallurgy and Environment, Central the whole production process will be further
South University, Changsha, China improved in economy, as the system is simplied
and volume of liquid stream to be treated is
reduced.
In integrated membrane separation for biofuel The bioreactor coupling fermentation and
production, the most essential unit is the separa- membrane separation has been applied in
tion of product from the raw material mixtures. biobutanol and bioethanol productions. In the
The membranes are directly combined with the acetone/butanol fermentation using Clostridium
chemical and biochemical conversions to improve acetobutylicum, the function of microbes is
Biofuel Production by Membrane Operations 221
Biofuel Production by Liquid recycle

Membrane Operations, Feed
Fig. 1 Schematic of cell
recycle membrane system
Cell recycle
Product B
Separation 1
Separation 2
Bioreactor
enriched
liquid
Separation 1: microfiltration to retain microbial cells.

Separation 2: membranes that enrich organic component (e.g.
product alcohol) at permeate side.
strongly inhibited by butanol, but acetone has no conditions. Productivity and conversion could be
such toxic effect. Therefore, the purpose to couple adjusted by catalyst-embedded membranes which
the membrane separation is the removal of buta- intensify the contact between reactant and cata-
nol. The productivity of a fermentation integrating lyst. Bernardo et al. explored CO Selox using
pervaporation to remove butanol from the culti- catalytic zeolite (Pt/Na-Y) membranes in a con-
vation medium is up to 6570 % over the one tinuous ow membrane reactor (Bernado
obtained without membrane separation, and the et al. 2008). Depending on the operating condi-
rate of substrate consumption becomes three times tions, the catalytic membranes effectively reduced
faster. That on-site removal of ethanol by PDMS the CO content in H2 from 10,000 to 1050 ppm.
membrane maintains the ethanol concentration Similar to gas reforming mentioned above,
and has also been proved. transesterication encounters the limit of equilib-
Biomass thermochemical conversion technol- rium conversion. FAME or glycerol as product
ogies such as pyrolysis and gasication are not the should be removed during the reaction in order
most important options at present (Demirbas to shift the equilibrium to the product side. One of
2009). As for biogas by anaerobic digestion, gas- the mechanisms used for biodiesel production by
eous metabolites can be easily taken out by strip- membrane reactor is proposed by Dube
ping. Membrane separation is more often et al. (2007). The immiscibility of oil and alcohol
introduced into reforming steps that generate as well as involvement of various surface forces
new products or adjust composition. The steam lead to the formation of an emulsion containing oil
reforming of methane (CH4) is an important way droplets suspended in methanol. The transesteri-
to produce syngas and supply hydrogen (H2) used cation occurs at the surface of the oil droplets.
in fuel cell (Demirbas 2009). It involves two The oil droplets could not pass through the pores
reversible reactions: reforming and water gas of the membrane, whereas the biodiesel product
shift. The rst is endothermic and limited by ther- will pass through the membrane pores along with
modynamic equilibrium. Therefore, high temper- the methanol, glycerol, and catalyst. As a result,
atures are used, making it slow to start up and the reaction equilibrium is shifted to the product
requiring costly refractory materials. The side. Increases in temperature, catalyst concentra-
membrane-based separation proves a cost- tion, and feedstock ow rate can increase the
effective way requiring less severe operation conversion rate.
222 Biofuel Separation by Membranes
References et al. 2001). VMD is liable to wetting by alcohols

and fouling by microorganisms. Yet another type
Bernardo P, Algieri C, Barbieri G, Drioli E (2008) Hydro- of membrane is solid (i.e., propylene)liquid (i.e.,
gen purication from carbon monoxide by means of
oleyl alcohol) combining extraction and
selective oxidation using zeolite catalytic membranes.
Sep Purif Technol 62:629635 pervaporation. It might also be called as mem-
Demirbas A (2009) Bioreneries: current activities and future brane contactor. The process exhibited separation
developments. Energy Convers Manag 50:27822801 of 180 for butanol/water; however, the liquid
Dube MA, Tremblay AY, Liu J (2007) Biodiesel produc-
portion is very unstable. Dehydration by
tion using a membrane reactor. Bioresource Technol
98:639647 pervaporation is for removing water in the recov-
Quresh N, Ezeji TC (2008) Butanol, a superior produc- ered stream so that the fuel specications could
tion from agricultural residues (renewable biomass): be met. It is a much easier task to achieve high
recent progress in technology. Biofuel Bioprod Bioref
separation factor in dehydration as the diffusion of
2:319330
water molecules is in nature faster than organics.
The composite membranes using poly(vinyl
alcohol) (PVA) have been commercialized.
Biofuel Separation by Membranes In biodiesel production, the removal of glyc-
erol as by-product to fatty acid alkyl esters
Lan Ying Jiang (FAAE) is very critical. One of the separation
School of Metallurgy and Environment, Central approaches uses porous membrane (e.g., MF)
South University, Changsha, China (Wang et al. 2009). The immiscibility of glycerol,
biodiesel, and soap added lead to the formation of
reversed micelle binding glycerol at the hydro-
Introduction philic site. The micelle is larger than biodiesel
molecular and thus retained by the membrane.
The purity of biofuels in forms of gas and Membrane separation based on the catalytic
liquid generated in biorenery must meet the pervaporation membrane is investigated in biodie-
in-service requirements (Huang et al. 2008). sel production procedure (Guerreiro et al. 2006).
Several types of membranes have been used Glycerol and methanol are able to form hydrogen
for bioalcohols separation. Pervaporation is bonds with the OH groups in the polymeric mem-
molecular-level liquidliquid separation based brane and are continuously removed from the
on the solutiondiffusion mechanism. The products during the reaction.
polymeric membranes used in pervaporation There exist several approaches for biogas pro-
are essentially dense and inorganic membranes duction; they are gasication, pyrolysis, and aero-
molecularly porous (Jiang and Zhu 2013). bic digestion. Whichever approach, it is necessary
The hydrophobic pervaporation recovers to remove hydrogen sulde (H2S), ammonia
the organics from fermentation solution. Poly- (NH3), and siloxane presenting potential damage
mers such as polydimethylsiloxane (PDMS), (e.g., corrosion, erosion, fouling) to devices used
ethylenepropylenediene monomer (EPDM), in follow-up conversion and pollution to environ-
and polyether block amide (PEBA) exhibit favor- ment. Membrane made of rubbery PDMS has been
able sorption toward organic molecules due to employed to remove the condensable siloxane
their nonpolar functional groups and minimum (Abatzoglou et al. 2006). In H2 or CH4 separation
diffusion discrimination. Vacuum membrane dis- from the other bulk components (mainly CO2) for
tillation (VMD) for recovery of bioalcohols uses various purity requirements, either membranes of
porous membrane, and the separation is based on molecular sieving or reverse selective membranes
vaporliquid equilibrium and Knudsen/Poiseuille relying more on sorption selection, depending on
ow (viscous). In comparison with pervaporation the feed composition, are used (Shao et al. 2009).
using dense membrane, VMD has higher ux but The transmembrane ux and separation factor
slightly lower separation factor (Urtiaga are the key parameters related to the economics of
Biogas 223
the process using membrane separation (Vane References

2005). For pervaporation, for example, only the
permeated components actually undergo Abatzoglou N, Ajhar M, Melin T (2006) Siloxane removal
with gas permeation membranes. Desalination
liquidvapor phase change. Therefore higher sepa-
200:234235
ration factor means higher energy efciency. Inor- Guerreiro L, Castanheiro JE, Fonseca IM, Martin-Aranda B
ganic membranes generally have better separation RM, Ramos AM, Vital J (2006) Transesterication of
performance than polymeric membranes in both soybean oil over sulfonic acid functionalised polymeric
membranes. Cata Today 118:166171
hydrophilic and hydrophobic pervaporation. The
Huang HJ, Ramaswamy S, Tschirner UW, Ramarao BV
rst large-scale pervaporation plant using tubular- (2008) A review of separation technologies in current
type module of zeolite NaA membrane for dehydra- and future bioreneries. Sep Purif Technol 62:121
tion was reported in Japan. In gas separation, to Jiang LY, Zhu JM (2013) Separation technologies for cur-
rent and future bioreneries status and potential of
prevent the plasticization by higher condensable
membrane-based separation. WIREs Energy Environ
molecules (e.g., CO2) and aging of polymeric mem- 2:673690
brane, cross-linking of polymer chains via ultravio- Park JW, Jeong JH, Yoon WL, Rhee YW (2004) Selective
let and ion beam radiations, thermal treatment, and oxidation of carbon monoxide in hydrogen-rich stream
over Cu-Ce/Al2O3 catalysts promoted with cobalt in a
chemical modication is applied.
fuel processor for proton exchange membrane fuel
For an active membrane, such as a catalytic one cells. J Membr Sci 132:1828
used in selective oxidation of CO, the most impor- Shao L, Low BT, Chung TS, Greenberg AR (2009) Poly-
tant requirements for catalyst are (1) highly selec- meric membranes for the hydrogen economy: contem-
porary approaches and prospects for the future.
tive CO oxidation activity at low temperatures and
(2) a wide temperature window for a greater than Urtiaga M, Gorri ED, Ruiz G, Ortiz I (2001) Parallelism
99 % conversion of CO (Guerreiro et al. 2006). To and differences of pervaporation and vacuum mem-
achieve high biodiesel yield via catalytic membrane distillation in the removal of VOCs from aqueous
streams. Sep Purif Technol 2223:327337
brane reactor operation, development of cheap
van Reis R, Zydney A (2001) Membrane separations in
and very active heterogeneous catalysts is also biotechnology. Curr Opin Biotechnol 12:208211
necessary. As environment for separation some- Vane LM (2005) A review of pervaporation for product
times is very harsh, materials chemical and recovery from biomass fermentation processes. J Chem
Technol Biotech 80:603629
mechanical stability become critical to biofuel
Wang Y, Wang X, Liu Y, Ou S, Tan Y, Tang S (2009)
production by membrane operation. One typical Rening of biodiesel by ceramic membrane separation.
case is biodiesel production using membrane Fuel Process Technol 90:422427
reactor, where viscous nature of feed limits the
membrane to tubular ones made of inorganic
materials (e.g., Al2O3/TiO2 ceramic) (Park
et al. 2004). Additionally, strong acid or base Biogas
catalysts during biodiesel production also limit
the selection to inorganic membranes with high Ivo Vankelecom1 and Muhammad Waqas Anjum2
1
resistance to degradation and corrosion. Centre for Surface Chemistry and Catalysis, KU
The challenges faced by porous membranes, Leuven, Leuven, Heverlee, Belgium
2
such as MF in sterile separation product from School of Chemical and Materials Engineering,
yeast and microbial culture, the high concentra- National University of Science and Technology,
tion microbes, and correspondingly high level of Islamabad, Pakistan
cell debris, can cause fouling of membrane sys-
tems. Similar situation happens in biodiesel
transesterication due to oil droplets. Consider- Fossil fuels are still the primary source of energy
able effort is being devoted to developing new by preference. However, as fossil fuels become
membrane modules with improved mass transfer more and more expensive with the possibility of
characteristics for microltration processes (Reis depletion of resources, the quest for alternate
and Zydney 2001). sources of energy is gaining attention. In this
224 Biogas
Biogas, Fig. 1 Schematic

diagram of the anaerobic Organic matter
digestion process for biogas (OM)
generation (Modied from
Basu et al. 2010)
Hydrolysis
(Soluble OM)
Fermentation
(Fatty acids)
Acetate Methanogenesis H2 + C2
CH4 + CO2
whole situation, anaerobic digestion of biological biomethane from organic wastes consists of
resources and biological waste could be a prom- mainly three steps including hydrolysis,
ising alternative energy carrier. acetogenesis, and methanogenesis. Organic sub-
Natural gas normally consists of 9095 % strates can be converted to biogas by a diverse
methane, but in biogas this composition is group of microbes using multienzyme (cellulases,
reduced to 5065 % making it a low-grade natural amylases, lipases, proteases, etc.) systems.
gas which is the product of neutral decomposition Organic material is fed into digesters after
of organic substance of animal or plant origin due grinding to an appropriate size. In the digesters,
to anaerobic bacterial activity. The plants used for these substrates are heated and agitated leading to
biogas production are normally referred to as production of biogas which is collected in a bio-
anaerobic digesters or anaerobic fermenters. The gas container. This gas is fed into an electric
resources used for biogas production include generator which produces electricity and heat.
kitchen waste, dry poultry droppings and animal Biohydrogen generated in fermentation pro-
excrements, remnants of food processing, and cesses (e.g., anaerobic fermentation, photofer-
slaughterhouse leftovers. Four ingredients needed mentation, dark fermentation) has hydrogen and
for biogas production are organic matter, bacteria, carbon dioxide as major ingredients. On the other
anaerobic conditions, and heat. In a controlled hand, biomethane produced by anaerobic diges-
reaction system, the gaseous mixture thus pro- tion of biological wastes has 3840 % carbon
duced can contain up to 70 % of biohydrogen dioxide with smaller amounts of hydrogen sulde
and biomethane, respectively, that can be used along with trace amounts (ppm) of hydrogen,
for commercial applications (Harold 2007). nitrogen, oxygen, and volatiles with 5560 %
The anaerobic digestion process can be classi- methane as a major part (Rasi et al. 2007). Biogas
ed into different sets of complex reactions, as has a caloric value of 3544 kJ g1, which is
shown in Fig. 1. Production of biohydrogen and comparatively higher than other energy resources
Biogas Production 225
Hydrogen production Fuel cell
Gasification
B
Bio Gas
Combined Cycle ELectricity Generation

Thermal power
Anaerobic digesstion/ Hydro

Methane Gas engines/Fuel cell
gasifiction
Biogas, Fig. 2 Possible applications of biogas as energy resource (Modied from Basu et al. 2010)
like petrol, diesel, or LPG and solid fuels like coal,

charcoal wood, etc. Biogas is a potential source of Biogas Production
environmentally benign, clean, and cheap alterna-
tive energy. However, the presence of incombus- Marco Scholz
tible and acid gases, like CO2, not only reduces its Aachener Verfahrenstechnik Chemical Process
caloric value, but their corrosive nature restricts Engineering, RWTH Aachen University, Aachen,
its transportation. One of the many trace compo- Germany
nents includes silicone containing compounds.
Commonly occurring siloxanes in biogas are
known as volatile methyl siloxanes (VMS) that Biogas mainly consists of methane and carbon
include cyclic tri-, tetra-, and pentasiloxane, as dioxide and is saturated with water. Minor com-
well as linear di-, tri-, tetrasiloxane. After components such as hydrogen sulde and ammonia
bustion, undesirable microcrystalline quartz and are also present in the raw biogas. Biogas is pro-
pentasiloxane are produced as they cause the wear duced in on-farm fermenters in the agricultural
& tear of engines and turbines. Puried biogas can production where crops or manure are digested,
be used as a feed to fuel cells or for domestic in sewage plants, and by landlls. It can either be
applications and power generation. Figure 2 rep- used to operate a combined heat and power engine
resents the possible applications of biogas (Basu on site or as natural gas substitute when upgrading
et al. 2010). the raw biogas. Using a combined heat and power
engine (CHP) can be efcient if the produced heat
is utilized (e.g., district heating). However, the
References heat can often not be used. Hence, upgrading the
biogas and injecting the puried gas to the natural
Basu S, Khan AL, Cano-Odena A, Liu C, Vankelecom IFJ gas grid is an interesting alternative compared to
(2010) Membrane-based technologies for biogas sepa- onsite electricity production by CHPs. In order to
rations. Chem Soc Rev 39:750768 upgrade biogas carbon dioxide, hydrogen sulde
Harold House (2007) Proceeding for London Swine
conference-todays challenges. Tomorrows opportuni- and water vapor have to be removed prior to grid
ties 34. http://www.londonswineconference.ca/proceed injection. The term upgrading particularly refers
ings/2007/LSC2007_HHouse.pdf. Accessed 29 Jan to the removal of carbon dioxide as it is the main
2013. gas component to be removed from the methane-
Rasi S, Veijanen A, Rintala J (2007) Trace compounds of
biogas from different biogas production plants. Energy rich raw biogas. The upgraded biogas can be
32:13751380 utilized as a natural gas substitute. Hence, it can
226 Biogas Production
Biogas Production, Table 1 Biogas upgrading technologies

Technology Material Process Ref.
Chemical Diethanolamine DEA scrubbing Hofmann and Plaettner (2006)
absorption Methyldiethanolamine MDEA scrubbing Hofmann and Plaettner (2006)
Monoethanolamine MEA scrubbing Hofmann and Plaettner (2006)
Physical Methanol Rectisol Hofmann and Plaettner (2006)
absorption Polyethylene glycol dimethyl Selexol Hofmann and Plaettner (2006)
ether
Polyglycol dimethyl ether Genosorb Urban et al. (20072008)
Water Pressurized water Hofmann and Plaettner (2006)
scrubbing
Adsorption Activated carbon Pressure swing Hofmann and Plaettner (2006)
adsorption
Molecular sieves Pressure swing Hofmann and Plaettner (2006)
adsorption
Membranes Cellulose acetate Gas permeation Scholz et al. (2013)
Ethyl cellulose Gas permeation Scholz et al. (2013)
Polycarbonate Gas permeation Scholz et al. (2013)
Polyimide Gas permeation Scholz et al. (2013)
Polymethyldisiloxane Gas permeation Scholz et al. (2013)
Polymethylpentene Gas permeation Scholz et al. (2013)
Polyphenylene oxide Gas permeation Scholz et al. (2013)
Polysulfone Gas permeation Scholz et al. (2013)
drive large-scale, central gas turbines or combined established technologies such as amine scrubbing
heat and power cycles in which the produced heat (Pentair 2013) (Table 1).
can be used. In addition, upgraded biogas can be There is no international standard on the gas
utilized as fuel for vehicles or it can be used as specications to be supplied to the natural gas
feedstock for the chemical synthesis (Baerns grid. In general, the methane content should be
et al. 2006). higher than 96 %. Often the gas has to be
Various technologies exist to upgrade biogas supplied at elevate pressures (e.g., >10 bar).
(Hofmann and Plaettner 2006). Physical and The tolerable hydrogen sulde levels vary also
chemical absorption and adsorption are well- depending on the country the biogas plant is
established technologies to upgrade biogas. operated in. The electrical and thermal energy
Recently, gas permeation membranes became demand to drive the upgrading process is typi-
important in biogas upgrading (Scholz cally in the range of 0.150.25 kWh/m3 (STP)
et al. 2013), as gas permeation processes are and 0.30.8 kWh/m3 (STP), respectively. Meth-
energy efcient, easy to handle, and robust, ane recoveries, which are the ratio of the meth-
which is important in the on-farm operation. In ane ow rate of the product gas to the methane
principle, cryogenic separation could also be ow rate of the feed gas, of more than 99 %
applied to separate methane and carbon dioxide, should be achieved to minimize the environ-
but due to the low raw gas ow rates (typically mental impact and to enhance the process eco-
less than 2,000 m3 (STP)/h), it is uneconomical. nomics. However, the product gas requirements
Recently, a membrane cryogenic process was have to be analyzed for each individual biogas
designed which is able to compete with upgrading plant.
Biogas Recovery 227
References digester in which the methane production takes

place, a biogas collection system providing piping
Baerns M, Behr A, Brehm A, Gmehling J, Hofmann H, of biogas to a combustion device, and a gas use
Onken U, Renken A (2006) Technische chemie. Wiley-
device where biogas is combusted to produce heat
VCH, Weinheim
Hofmann PAF, Plaettner A (2006) Einspeisung von biogas or electricity (Agstar 2011). Alternatively, the B
in das erdgasnetz. Technical report. Fachagentur produced biogas can be used on site as fuel for
Nachwachsende Rohstoffe e. V. (FNR), G ulzow, 199 automotives or can be injected in the natural gas
p (mit Frderung des Bundesministeriums f ur
grid. In this case, the gas must be upgraded to
Ernhrung, Landwirtschaft und Verbraucherschutz
(FKZ 22021103)). ISBN 3-00-018346-9 natural gas standards, namely, 98 % methane con-
Pentair: http://www.haffmans.nl/resources/images/267. tent (AEBIOM 2009).
pdfdate: 22 Feb 2013 Although the composition of biogas varies sig-
Scholz M, Melin T, Wessling M (2013) Transforming
nicantly depending on many factors such as the
biogas into biomethane using membrane technology.
Renew Sust Energ Rev 17:199212 type of digester, the average values are reported to
Urban W, Girod K, Lohmann H (20072008) Technologie range between 50 % and 70 % methane and
und Kosten der Biogasaufbereitung und Einspeisung in 3050 % carbon dioxide, as well as hydrogen
das Erdgasnetz. Ergebnisse der Markterhebung
sulde, sulfur compounds, siloxanes, and aromatic
and halogenated compounds (Rasi et al. 2007).
Upgrading (purication) of biogas is highly bene-
cial in terms of increasing the amount of methane
Biogas Recovery per unit volume of biogas, which equals an increase
in its caloric value. In addition to the obvious need
Ivo Vankelecom for removal of carbon dioxide as the major pollut-
Centre for Surface Chemistry and Catalysis, ant, the trace compounds have the potential to
KU Leuven, Leuven, Heverlee, Belgium trigger ozone depletion and the greenhouse effect,
reduce the quality of local air due to formation of
volatile organic compounds (VOCs), and so
The constant increase in energy demand, coupled on. Moreover, sulfur compounds corrode pipelines
with the depletion of fossil fuels, is a concern that and combustion engines, while silicon compounds
cannot be ignored in the industrial and highly oxidize and are deposited on engine parts (Rasi
populated world of today (Basu et al. 2010). Bio- et al. 2007). Currently, biogas upgrading is largely
gas production is a very promising source of achieved through amine scrubbing where carbon
renewable energy that still offers further exploita- dioxide is absorbed into an amine solution, water
tion of its potential (Scholz et al. 2013). Essen- scrubbing where carbon dioxide is absorbed to
tially, microbially controlled biogas production is water at elevated pressures, and pressure swing
an already existing part of the global carbon cycle, adsorption where pressure of the gas mixture is
releasing an estimated 590800 million tons of changed to induce adsorption and desorption of
methane to the atmosphere (Bond and Templeton one gas species. These technologies are in general
2011). Current biogas recovery systems seek to more energy intensive and polluting than mem-
exploit these processes in order to produce energy brane technology (Basu et al. 2010; AEBIOM
from sewage wastewater, animal manure, crop 2009; Scholz et al. 2013). Polymeric membranes
straws, or mixed agricultural wastes (Chen have gained a share in the separations market, as
et al. 2012; Rasi et al. 2011). Usually, a system opposed to the inorganic membranes which are
for biogas recovery consists of four components: a expensive and brittle. Some commercial polymeric
collection system which helps transfer the biogas membranes such as SEPAREX (cellulose acetate)
source to the anaerobic digester, the anaerobic of UOP and polydimethyl siloxane (PDMS) have
228 Biogas Upgrading
already proven or were reported to operate success- industrial waste. The composition of biogas varies
fully in separation of carbon dioxide or siloxanes from the origin of the anaerobic digestion process,
and VOCs, respectively. However, the separation and the main components are methane (CH4) and
of hydrogen sulde still remains a difcult and carbon dioxide (CO2) as shown in Table 1 (Rasi
interesting issue for membrane research. Mixed et al. 2007). Biogas has a very high-energy poten-
matrix membranes composed of inorganic particles tial due to the presence of methane (CH4) and thus
distributed throughout a polymeric matrix are is a great source for renewable energy production,
promising alternatives thoroughly investigated for which can be used for heating, combined heat and
these separations (Basu et al. 2010). power (CHP) generation, vehicle fuel, fuel cell,
and substitute natural gas. Depending on the dif-
ferent end uses, the specic biogas treatment
References should be executed. For the applications as vehi-
cle fuels and natural gas grid injection, CO2
Agstar-EPA (2011). Market opportunities for biogas recov- should be removed from the raw biogas, i.e.,
ery systems at U.S. Livestock Facilities. http://epa.gov/
biogas upgrading, to reduce the corrosion caused
agstar/documents/biogas_recovery_systems_
screenres.pdf. Accessed 31 Jan 2013 by the acid gases and increase the Wobbe index
Basu S, Khan AL, Cano-Odena A, Lui C, Vankelecom IFJ which is directly proportion to the methane con-
(2010) Membrane-based technologies for biogas sepa- centration. However, biogas upgrading process
rations. Chem Soc Rev 39(2):750768
will add extra costs into the biogas production.
Bond T, Templeton MR (2011) History and future of
domestic biogas plants in the developing world. Energy Thus, an optimized upgrading technology in
Sustain Dev 15(4):347354 terms of low energy consumption and high ef-
Chen L, Lixin Z, Changshan R, Fei W (2012) The progress ciency should be developed to reduce the biogas
and prospects of rural biogas production in China.
production cost. Moreover, the minimum emis-
Energy Policy 51:5863
European Biogas Association-AEBIOM (2009). A biogas sion of CH4 should also be addressed since meth-
road map for Europe. http://www.aebiom.org/IMG/ ane has a greenhouse effect of around 23 times
pdf/Brochure_BiogasRoadmap_WEB.pdf. Accessed higher than that of CO2.
31 Jan 2013
Different techniques such as pressure swing
biogas from different biogas production plants. Energy adsorption (PSA), physical absorption (e.g.,
32(8):13751380 water scrubbing), chemical absorption (e.g.,
Rasi S, Lntel J, Rintala J (2011) Trace compounds affect- amines), and membrane system have been inves-
ing biogas energy utilization a review. Energy
tigated for biogas upgrading. The choice of a
Convers Manag 52(12):33693375
Scholz M, Melin T, Wessling M (2013) Transforming suitable technology is mainly dependent on the
biogas into biomethane using membrane technology. specic condition at a plant, such as availability of
Renew Sust Energ Rev 17:199212 low price for heating, electricity, and water, as
well as the amount of gas to be handled. Today,
most of biogas upgrading plants in Sweden are
still using PSA even though the methane content
Biogas Upgrading in the upgraded gas is low (96 %) and methane
loss is quite high (310 %), while some plants
Xuezhong He using water scrubbing technique will additionally
Department of Chemical Engineering, Norwegian produce a lot of wastewater, and electricity con-
University of Science and Technology, sumption is also quite high. Membrane system
Trondheim, Sr-Trndelag, Norway could be favorable for biogas upgrading due to a
series of advantages including the safety and sim-
plicity of the operation and easy maintenance
Biogas is produced in anaerobic digesters from without any hazardous chemicals (Makaruk
biodegradable wastes such as sewage sludge, et al. 2010). Figure 1 shows a schematic diagram
manure, organic fraction of household, and of membrane system for biogas upgrading.
Biogas Upgrading 229
Biogas Upgrading, Table 1 Typical gas composition from different sources (Rasi et al. 2007)
Composition (vol, %)
Process CO2 CH4 N2 O2 H2O H2S/SO2 (ppm)
Farm biogas plant 3738 5558 <2 <1 47 32169
Sewage digester 38.6 57.8 3.7 0 47 62.9 B
Landll 3741 4757 <1 <1 47 36115
CH4 65%
CO2 35%
H2S 200-300ppm
FeCl2 Water saturated
Drier
Filter
Raw biogas
Compressor Cooler
Biomass
Bioreactor Filter
10-30 bar
Propane/ odour
Compressor
Waste solids
Upgraded biogas,
CH4>96% Cooler 2nd membrane 1st membrane
CO2-rich permeate
Biogas Upgrading, Fig. 1 A schematic diagram for biogas upgrading process by membrane systems
As illustrated in the above ow sheet, H2S and specication of natural gas network distribution,
water vapor must be removed before feeding into which makes this environmentally friendly tech-
the membrane system. Compression of gas may nology more competitive compared to the other
vary depending on whether it goes to natural gas conventional technologies currently used.
grid (<80 bar) or will be used for vehicle fuel Makaruk et al. pointed out that membrane system
(~200 bar). The main challenge of a membrane provides enough exibility for heat integration
system is the pretreatment of biogas to protect the within biogas plants (Makaruk et al. 2010), and
membrane materials, especially for the biogas the expected energy requirement for a single-
produced from the sewage treatment plants and produced cubic meter of natural gas substitute is
landll sites where it usually have a lot of mali- estimated to be around 0.3 kWh, which is close to
cious gas components. Polyvinyl amine (PVAm)/ the values reported in an industrial scale demon-
polyvinyl alcohol (PVA) blend xed-site-carrier strator for membrane biogas upgrading plant at
(FSC) membranes for biogas upgrading have been Bruck an der Leitha in Austria (Miltner
investigated by Deng et al. (Deng and Hgg et al. 2008). Moreover, a new carbon membrane
2010). Their results indicated that a membrane company MemfoACT (http://www.memforact.
process with a CH4 recovery of 99 % at a low no) was launched in 2008 in Norway, which
operation cost could be designed to achieve the mainly focuses on the biogas upgrading using
230 Biogas Upgrading by Membrane Processes
their patented hollow ber carbon molecular sieve through the membrane than methane. Thus, the
membranes. Their contributions could promote to methane-rich product gas is supplied on the
bring this technology into the commercial appli- retentate side of the membrane stages.
cation in the near future. Some commercially available membrane mate-
rials have selectivities of as high as 60 (Scholz
et al. 2013). However, by applying these mem-
References branes in a single-stage gas permeation process,
signicant methane losses are observed in order to
Deng L, Hgg M-B (2010) Techno-economic evaluation of achieve the required methane purity of 96 %
biogas upgrading process using CO2 facilitated trans-
(Rautenbach and Welsch 1993). Hence, the
port membrane. IJGGC 4:638646
Makaruk A, Miltner M, Harasek M (2010) Membrane trade-off between methane purity and methane
biogas upgrading processes for the production of natu- recovery cannot be overcome. Therefore, multi-
ral gas substitute. Sep Purif Technol 74:8392 stage gas permeation processes are required to
Miltner M, Makaruk A, Harasek M (2008) Application of
produce a product gas with methane purities of
gas permeation for biogas upgrade operational expe-
riences of feeding biomethane into the Austrian gas more than 96 % while simultaneously providing
grid. 16th European Biomass Conference & Exhibition methane recoveries of more than 99 %. To design
from Research to Industry and Markets, Valencia an upgrading process with high methane
recoveries is particularly important as the
biogas from different biogas production plants. Energy
32:13751380 methane recovery determines the protability of
the biogas upgrading plant and competing
upgrading processes such as amine scrubbing or
pressure swing adsorption show high methane
recoveries.
Biogas Upgrading by Membrane The driving force to operate the gas permeation
Processes process can either be provided by a feed-side
compression or by a permeate-side vacuum
Marco Scholz pump. As the carbon dioxide content in the raw
Aachener Verfahrenstechnik Chemical Process gas can be as high as 50 % and the methane-rich
Engineering, RWTH Aachen University, Aachen, product gas should be provided at elevated pres-
Germany sures, the compression prior to the gas permeation
system is preferred over a system operating with
subambient pressure on the permeate side. Here,
Biogas upgrading, which is the removal of carbon the compressor provides both the driving force for
dioxide from a methane-rich raw biogas, can be the permeation of carbon dioxide and the elevated
accomplished using membranes. Gas permeation pressure of the product gas for natural gas grid
membranes or membrane contactors can be injection.
applied to remove carbon dioxide. Membrane Gas permeation membranes can be combined
contactors are still under investigation, while gas with conventional gas separation technologies in
permeation modules are applied in commercial membrane hybrid processes. Recently, a mem-
biogas upgrading plants (Scholz et al. 2013). In brane cryogenic process was reported, which has
principle, organic as well as inorganic membrane high methane recoveries (Pentair 2013). In this
materials have excellent selectivities to separate particular process, the gas permeation stage is
carbon dioxide and methane. However, Baker applied to provide the methane-rich product gas.
(2001) reports that only a limited number of mem- Since the permeate stream of the membrane stage
brane materials are commercially available as contains signicant quantities of methane, the
membrane modules, and all these membranes are permeate stream is compressed up to 20 bar and
made out of polymers. For commercial polymeric fed to a cryogenic separation unit, which operates
membranes, carbon dioxide permeates faster at 24 C. Here, the carbon dioxide fraction is
Biohybrid Artificial Liver (BAL) 231
liqueed and a methane recovery of 100 % can be blood or plasma of liver failure patients. BAL pro-
obtained. vides temporary support for patients waiting for an
Another option to upgrade biogas is to use a allogeneic liver transplant, and since the liver can
single-stage membrane process and to supply the regenerate, the temporary support provided by
upgraded biogas in one step. The permeate of the BAL may allow time for liver regeneration. B
membrane stage, which contains signicant The bioreactor is an important component of
amounts of methane, is combusted in a combined BAL, because it determines the viability and func-
heat and power engine (Makaruk et al. 2010). The tion of the hepatocytes within it. A successful and
combined heat and power engine requires at least clinically effective bioreactor should mimic the
a methane level of 30 % so that raw biogas can be structure of the liver and provide an in vivo-like
mixed with the permeate stream to adjust the microenvironment for the growth of hepatocytes,
methane level. By applying this particular process thereby maintaining the cells viability and function
design, the electrical energy demand to drive the to the maximum extent. The important issues are the
upgrading process can be provided by the gas choice of cell sources and the design of the bioreac-
engine. However, the produced heat cannot be tor (Ding and Shi 2011). The cell sources provide
used efciently. liver-specic functions, such as detoxication, drug
metabolism, and protein synthesis, while the bio-
reactors maintain the viability and function of cells.
References More efforts are now underway in search for the
best cell resource and best design of bioreactors.
Baker R (2001) Future directions of membrane Considering the several functions that the liver
gas-separation technology. Membr Technol 138:510
performs, the bioreactor for BAL devices has to
Makaruk A, Miltner M, Harasek M (2010) Membrane
biogas upgrading processes for the production of natu- ensure the rapid detoxication of neural and
ral gas substitute. Sep Purif Technol 74:8392 hepatic toxins, the return of liver-specic
Pentair. http://www.haffmans.nl/resources/images/267. hepatotrophic factors, as well as liver-specic
pdf. Date 22 Feb 2013
coagulation factors, back into patients blood,
Rautenbach R, Welsch K (1993) Treatment of landll gas
by gas permeation pilot plant results and comparison and the maintenance of liver cell detoxication
to alternatives. Desalination 90:193207 and synthetic functions until liver tissue regener-
Scholz M, Melin T, Wessling M (2013) Transforming ation or organ transplantation.
biogas into biomethane using membrane technology.
One of the most promising bioreactors is the
Renew Sustain Energy Rev 17:199
membrane bioreactor. Polymeric membranes in
at and hollow ber conguration with different
morphology and chemicalphysical properties
have been used in BAL devices (De Bartolo and
Biohybrid Artificial Liver (BAL) Bader 2001; Kamlot et al., 1996; Kasuya and
Tanishita 2012). Most of the extracorporeal
Sabrina Morelli BALs have not only used cellulose and
Institute on Membrane Technology, National polysulfone derivatives but also native and mod-
Research Council of Italy, ITM-CNR, Rende, ied polypropylene membranes. Morphological
Italy (e.g., pore size, pore size distribution, and rough-
ness) and physicochemical membrane properties
(e.g., surface charge, wettability, and surface free
A biohybrid articial liver (BAL) is a bioarticial energy) affect all the adhesion and metabolic
device which consists of functional liver cells functions of hepatocytes.
supported by an articial cell culture material. It Hepatocytes have been cultured in membrane
incorporates hepatocytes into a bioreactor in which bioreactors in different congurations: between
the cells are immobilized, cultured, and induced to at sheet membranes in a sandwich conguration;
perform the hepatic functions by processing the in the lumen of hollow ber membranes entrapped
232 Biohybrid Artificial Liver (BAL) Systems
in a collagen layer; in the shell of hollow ber represents an important therapeutic strategy for
membranes in monolayer, aggregate, or spheroid patients with acute liver failure.
structure and attached to microcarriers; in a net- Generally, a BAL system consists of functional
work of hollow ber membranes with different liver cells supported by an articial cell culture
functions; in a spirally wound device in which material. In particular, it incorporates hepatocytes
hollow bers are used to provide oxygen to the into a bioreactor in which the cells are
cells; in multibore capillaries; microencapsulated immobilized, cultured, and induced to perform
and in an oxygen-permeable membrane rotating the hepatic functions by processing the blood or
system under microgravity conditions. plasma of liver failure patients. The BAL system
Several designs of BAL devices that are differ- acts as a bridge for the patients until a donor organ
ent in conguration, cell source, and culture tech- is available for transplantation or until liver regen-
nique have currently undergone clinical trials eration. The development of a BAL system
(Morelli et al. 2010). involves many design considerations. It must pro-
vide (1) an adhesion support to the cells; (2) ade-
quate mass transfer of oxygen, nutrients, and toxic
Cross-References substances from the blood or plasma of patients to
the cell compartments and proteins, catabolites,
Bioarticial Liver and other specic compounds produced by cells
from the cell compartment to the blood or plasma;
(3) immunoprotection of cells; and (4) biocompat-
References
ibility. BAL devices are classied by the cell
De Bartolo L, Bader A (2001) Review of a at membrane source, the type of culture system for the hepato-
bioreactor as a bioarticial liver. Ann Transplant 6:4046 cytes, and the conguration of the bioreactor.
Ding YT, Shi XL (2011) Bioarticial liver devices: per- Several BAL systems have been evaluated
spectives on the state of the art. Front Med 5:1519 preclinically in in vitro experiments and in large
Kamlot A, Rozga J, Watanable FD, Demetriou AA (1996)
Review: Articial liver support systems. Biotechnol animal models of liver failure (Morelli
Bioeng 50:382391 et al. 2010). Currently, different types of BAL
Kasuya J, Tanishita K (2012) Microporous membrane- devices are in various stages of clinical evalua-
based liver tissue engineering for the reconstruction of tion, and some of them are listed in Table 1 (van de
three-dimensional functional liver tissues in vitro.
Biomatter 2:290295 Kerkhove et al. 2004).
Morelli S, Salerno S, Piscioneri A, Campana C, Drioli E, Many of these devices use hollow ber mem-
De Bartolo L (2010) Membrane bioreactors for regen- branes (HFMs) as supports for the cultured hepa-
erative medicine: an example of the bioarticial liver. tocytes and as immunoselective barriers between
Asia Pac J Chem Eng 5:146159
the plasma of the patients and the hepatocytes
used in the bioreactor. Membranes also permit
the transport of nutrients and metabolites to cells
and the transport of catabolites and specic meta-
Biohybrid Artificial Liver (BAL) bolic products to the blood. In the membrane bio-
Systems reactors, mass transfer is determined by the
molecular weight cutoff (MWCO) or pore diame-
Sabrina Morelli ter of the membrane and occurs by diffusion
Institute on Membrane Technology, National and/or convection in response to existing trans-
Research Council of Italy, ITM-CNR, Rende, membrane concentration or pressure gradients.
Italy Most of the bioreactors for BAL systems use
membranes with MWCO ranging from 70 to
100 kDa that allow the transport of serum albumin
A biohybrid articial liver (BAL) system is an but exclude proteins with high MW such as
articial extracorporeal supportive device which immunoglobulins and cells.
Biohybrid Artificial Liver (BAL) Systems 233
Biohybrid Artificial Liver (BAL) Systems, Table 1 Membrane BAL systems in clinical evaluation
Bioreactor
BAL system conguration Membrane References
Kiil dialyzer bioarticial liver Plate Cellulose Matsumura
et al. (1987)
B
ELAD Hollow ber Cellulose acetate Sussman
Amphioxus Cell Technology et al. (1992)
LLS Hollow ber Polyamide Gerlach
Charite, Humboldt University, Polyethersulfone et al. (1994)
Germany Polypropylene
HepatAssist Hollow ber Polysulfone Demetriou
Circe Biomedical et al. (1995)
AMC-BAL Spirally wound Nonwoven polyester matrix, Flendrig
University of Amsterdam polypropylene et al. (1997)
BLSS Hollow ber Cellulose acetate Patzer et al. (2002)
Excorp Medical Inc.
BAL TECA Corp. Hollow ber Polysulfone Ding et al. (2003)
One of the rst clinical devices using HFMs testing is a bioreactor from TECA Corp. in which
was developed by Sussman and coworkers, a polysulfone membrane compartmentalizes por-
namely, the extracorporeal liver-assist device cine hepatocytes (Ding et al. 2003).
(ELAD) in which the human hepatocytes were
located outside the hollow ber and blood ows
through the lumen of the hollow bers. This
device was commercialized by Amphioxus Cell
References
Technologies (Sussman et al. 1992). HepatAssist
Circe Biomedical is the most clinically advanced Demetriou AA, Rozga J, Podesta L, Lepage E, Woolf G,
system of its kind. It is an extracorporeal cell- Vierling J, Makowka LE, Moscioni AD, Hoffman A,
based bioarticial liver device, based on the use McGrath M, Kong L, Rosen H (1995) Early clinical
experience with a hybrid bioarticial liver. Scand
of an open membrane hollow ber bioreactor J Gastroenterol 208:111117
(Demetriou et al. 1995). In this system, hepato- Ding YT, Qiu YD, Chen Z, Xu QX, Zhang HY, Tang Q, Yu
cytes are loaded into the extracapillary space, and DC (2003) The development of a new bioarticial liver
the patients plasma ows through the capillary and its application in 12 acute liver failure patients.
World J Gastroenterol 9:829832
lumina of membranes. A more complex system is
Flendrig LM, la Soe JW, Jorning GG, Steenbeek A,
the liver support system (LSS) or the modular Karlsen OT, Bovee WM, Ladiges NC, Te Velde AA,
extracorporeal liver system (MELS) which con- Chamuleau RA (1997) In vitro evaluation of a
sists of a bioreactor with four interwoven inde- novel bioreactor based on an integral oxygenator
and a spirally wound nonwoven polyester matrix for
pendent capillary membrane systems that serve
hepatocyte culture as small aggregates. J Hepatol
different functions (Gerlach et al. 1994). The 26:13791392
BLSS is a hollow ber device that uses porcine Gerlach JC, Encke J, Hole O, Muller C, Ryan CJ, Neuhaus
hepatocytes embedded in a collagen matrix P (1994) Bioreactor for larger scale hepatocyte in vitro
perfusion. Transplantation 58:984988
(Patzer et al. 2002). The Academic Medical Cen- Matsumura KN, Guevara GR, Huston H, Hamilton WL,
ter Bioarticial Liver (AMC-BAL) developed by Rikimaru M, Yamasaki G, Matsumura MS
Flendrig et al. uses a three-dimensional, spirally (1987) Hybrid bioarticial liver in hepatic failure: pre-
wound, nonwoven polyester matrix for hepato- liminary clinical report. Surgery 101:99103
Morelli S, Salerno S, Piscioneri A, Campana C, Drioli E,
cyte attachment with integrated hollow bers for
De Bartolo L (2010) Membrane bioreactors for regen-
oxygen delivery to the cells (Flendrig et al. 1997). erative medicine: an example of the bioarticial liver.
Another BAL system that is currently in clinical Asia Pac J Chem Eng 5:146159
234 Biohybrid Artificial Liver Systems
Patzer JF, Mazariegos GV, Lopez R (2002) Preclinical that mimic the specic features of in vivo
evaluation of the Excorp Medical, Inc, bioarticial environments.
liver support system. J Am Coll Surg 195:299310
Sussman NL, Chong MG, Koussayer T, He DE, Shang TA, Biohybrid membrane systems are successfully
Whisennand HH, Kelly JH (1992) Reversal of fulmi- applied in the eld of tissue engineering and
nant hepatic failure using an extracorporeal liver assist regenerative medicine (Morelli et al. 2009). For
device. Hepatology 16:6065 the development of functional biohybrid mem-
van de Kerkhove MP, Hoekstra R, Chamuleau RAFM, van
Gulik TM (2004) Clinical application of bioarticial brane systems, a number of issues need to be
liver support systems. Ann Surg 240:216230 addressed: morphological, physicochemical,
mechanical, and transport properties of the mem-
brane, the optimal density of immobilized cells,
the interaction of cells with the membrane, the
Biohybrid Artificial Liver Systems differentiation of cells, as well as the maintenance
of viability and metabolic functions in vitro mem-
Bioarticial Liver Support System (BLSS) brane constructs.
Different types of biohybrid membrane sys-
tems have been proposed for the reconstruction
and/or regeneration of many organs and tissues
Biohybrid Membrane Systems (e.g., the pancreas, liver, kidney, skin, and bone
Scharp et al. 1994; Saito et al. 2006; De Bartolo
Sabrina Morelli et al. 2009; Ding and Shi 2011; Gentile
Institute on Membrane Technology, National et al. 2011).
Research Council of Italy, ITM-CNR, Rende, Currently, biohybrid membrane systems are
Italy also developed for the creation of a biomimetic
microenvironment for neural tissue engineering
since they may be used for the in vitro simulation
Biohybrid membrane systems are engineered sys- of human brain functions. Semipermeable hollow
tems based on the combination of biological units, ber membranes are widely used as guidance
cells, or tissues, immobilized on an articial struc- channels in promoting in vitro and in vivo neuro-
ture, the membrane. In these systems, membranes nal regeneration (Zhang et al. 2005; Morelli
act as instructive materials which are capable of et al. 2010, 2012).
supporting tissue/organ formation. Cells have to Generally, biohybrid membrane systems could
make an intimate contact with the surface of the not only have a role in the replacement of injured
membrane but also to develop close cell-cell con- organ or tissue but also accelerate the develop-
nections, which is a precondition for their survival ment of new drugs that may cure patients as an
and high functional activity. alternative to animal experimentation.
Among polymeric materials, membranes in at
and hollow ber conguration are the most attrac-
tive in the use of biohybrid systems for their Cross-References
characteristics of stability, biocompatibility, and
selective permeability. Polymeric membranes Bioarticial Organs, Membrane Operations of
could mimic the extracellular matrix with which Tissue Engineering, Membrane Operations of
cells interact allowing the organization of the cells
into a three-dimensional architecture. The mem-
branes are able to modulate the adhesion, prolif- References
eration, and differentiation of cells which are
De Bartolo L, Salerno S, Curcio E, Piscioneri A, Rende M,
fundamental processes for tissue regeneration by
Morelli S, Tasselli F, Bader A, Drioli E (2009) Human
governing the mass transfer of molecules that hepatocyte functions in a crossed hollow ber mem-
generate a precisely controlled microenvironment brane bioreactor. Biomaterials 30:25312543
Bioinspired Membranes 235
Ding YT, Shi XL (2011) Bioarticial liver devices: per- the archetype. On the one hand, the membrane
spectives on the state of the art. Front Med 5:1519 materials (organic, inorganic, or organic-
Gentile P, Chiono V, Tonda-Turo C, Ferreira AM, Ciardelli
G (2011) Polymeric membranes for guided bone regen- inorganic hybrid) can be synthesized by borrow-
eration. Biotechnol J 6:11871197 ing the fundamental principles of the materials
Morelli S, Salerno S, Piscioneri A, Rende M, Campana C, processing in nature. On the other hand, the B
Drioli E, De Bartolo L (2009) Membranes in regener- advancement in any branch of biological mate-
ative medicine and tissue engineering. In: Drioli E,
Giorno L (eds) Membrane operations: innovative sep- rials science related to structure, function, and
arations and transformations. Wiley VCH, Verlag structure-function relationships can be employed
GmbH & Co. KGaA, Weinheim, pp 433446 and incorporated to rationally design the structure
Morelli S, Salerno S, Piscioneri A, Papenburg BJ, Di or function of bioinspired membrane. As a result,
Vito A, Giusi G, Canonaco M, Stamatialis D,
Drioli E, De Bartolo L (2010) Inuence of micro- the membrane preparation is usually conducted
patterned PLLA membranes on outgrowth and orienta- under very mild conditions (aqueous solution,
tion of hippocampal neurites. Biomaterials ambient temperature and pressure, neutral or
31:70007011 near neutral pH, etc.), the membrane materials
Morelli S, Piscioneri A, Salerno S, Rende M, Campana C,
Tasselli F, di Vito A, Giusi G, Canonaco M, Drioli E, are often easily available and biodegradable, the
De Bartolo L (2012) Flat and tubular membrane sys- membrane structure can be nely tuned, and the
tems for the reconstruction of hippocampal neuronal membrane often displays multiple functions and
network. J Tissue Eng Regen Med 6:299313 high application efciency. To some extent, the
Saito A, Aung T, Sekiguchi K, Sato Y, Vu DM, Inagaki M,
Kanai G, Tanaka R, Suzuki H, Kakuta T (2006) Present bioinspired membrane can be viewed as next-
status and perspectives of bioarticial kidneys. J Artif generation membrane.
Organs 9:130135 At present, several bioinspired platform tech-
Scharp DW, Swanson CJ, Olack BJ, Latta PP, Hegra OD, niques based on fascinating biochemistry and
Doherty EJ, Gentile FT, Flavin KS, Ansara MF, Lacy
PE (1994) Protection of encapsulated human islets biology have been developed. For example,
implanted without immunosuppression in patients inspired by formation process of biosilica, a plat-
with type I or type II diabetes and in nondiabetic control form technique called biomineralization or
subjects. Diabetes 43:11671170 biomimetic mineralization was developed to
Zhang N, Yan H, Wen X (2005) Tissue-engineering
approaches for axonal guidance. Brain Res Rev fabricate polymer-inorganic nanohybrid mem-
49:4864 brane (Pan et al. 2010). By using biomolecules
as catalyst and/or template, silica nanoparticles
were in situ generated and well dispersed in a
polymer matrix, and the as-prepared hybrid mem-
brane synergistically combined the advantages of
Bioinspired Membranes two components. Another platform technique
example is bioadhesion or biomimetic adhe-
Zhongyi Jiang, Yifan Li and Hong Wu sion. This technique originated from mussel-
School of Chemical Engineering and Technology, inspired catecholic chemistry, which was success-
Tianjin University, Nankai, Tianjin, China fully utilized to construct ultrathin and stable
active layer of dense composite membrane
(Li et al. 2009). The other techniques are mostly
Bioinspired membrane can be thought of as a based on self-assembly, in particular layer-by-
comprehensive name for a big family of articial layer (LBL) assembly. LBL assembly technique
membranes of which the design originality arises provides a powerful platform to manipulate mul-
from biological phenomena and principles in tiple interactions, such as covalent bond, hydro-
nature. Compared with biomimetic membrane, gen bonding, charge transfer, hydrophobic, host-
bioinspired membrane is a much broader term guest, and coordination bond interactions
(Rawlings et al. 2012), because a bioinspired (Matsusaki et al. 2012), so as to prepare mem-
membrane does not necessarily employ biological brane with precisely tailored nanostructures (i.e.,
membrane or biologically occurring thin lm as thickness, free volume, hydrophilicity, etc.).
236 Bioinspired Membranes
Bioinspired Membranes,
Fig. 1 Concept of
nanotube-based
hierarchically structured
membrane with durable
high permeability and
selectivity
Future effort on this type of bioinspired mem- transport properties, intelligent properties of
brane may turn to improving the fundamental membranes show great promise in membrane sci-
mechanism and extending the application elds. ence and technology (Tokarev and Minko 2010).
Of course, new bioinspired platform techniques For instance, self-healing membrane, self-
will be continuously explored. cleaning membrane, and many sorts of stimuli-
Selective transport properties and the responsive membranes have been reported (Chen
corresponding membrane structures are another et al. 2011; Wang et al. 2011b; Capadona
major concern in bioinspired membrane realm. et al. 2008; Tokarev and Minko 2010). Such
In biology, the most effective membrane- smart behaviors can endow the membrane with
mediated separation/transport processes often ultra-high stability, switchable permeation ux,
involve complex, nonplanar articulations of struc- alternative driving force, etc. Other properties,
ture that are highly optimized to support the mul- such as mechanical strength, interface adhesion,
tiple functions. Inspired by the hierarchical water retention, etc., have been successfully
structures in biology, rational design and synthesis reinforced by learning from nature (Zeng
hierarchical structures within membranes may et al. 2010; Ma et al. 2010; Wang et al. 2011a).
offer immense opportunities for efcient molecu- Breakthroughs in understanding how a living sys-
lar/ion sieving processes. For example, nanotubes tem works and exploring new inspiration source
have been utilized to mimic biological channels will ensure the sustainable development of
for rapid and selective molecule transport. With bioinspired membrane.
the aid of nanotubes, an ideal concept of vertically
aligned biomimetic channels within a membrane
matrix have been proposed and shown in Fig. 1, References
where membrane permeability, selectivity, and
durability are individually manipulated by the Capadona JR, Shanmuganathan K, Tyler DJ, Rowan SJ,
Weder C (2008) Stimuli-responsive polymer
highly permeable nanopores, the entrance chem-
nanocomposites inspired by the sea cucumber dermis.
istry of the nanotube, and the robust nonporous Science 319:13701374
lling matrix. Here, the aligned nanotubes are Chen WJ, Su YL, Peng JM, Dong YA, Zhao XT, Jiang ZY
extremely like the channel proteins inserted (2011) Engineering a robust, versatile amphiphilic
membrane surface through forced surface segregation
through cell membranes. This concept has par-
for ultralow ux-decline. Adv Funct Mater 21:191198
tially come true recently (Xu et al. 2011), and it Li B, Liu WP, Jiang ZY, Dong X, Wang BY, Zhong YR
requires much more efforts to fulll all the poten- (2009) Ultrathin and stable active layer of dense com-
tials of this concept. Also, many block copoly- posite membrane enabled by poly(dopamine). Lang-
muir 25:73687374
mers have been designed to mimic the Ma J, Zhang MH, Jiang ZY, Nie MC, Liu GX (2010) Facile
hydrophilic-hydrophobic mosaic structure of fabrication of structurally stable hyaluronic acid-based
biomembrane for antifouling (Chen et al. 2011). composite membranes inspired by bioadhesion.
Moreover, many efforts have been dedicated to J Membr Sci 364:290297
Matsusaki M, Ajiro H, Kida T, Serizaw T, Mitsuru
mimicking the proton pump and ion channel so as
A (2012) Layer-by-layer assembly through weak inter-
to accelerate proton transport (Wang et al. 2012; actions and their biomedical applications. Adv Mater
Sharma et al. 2010). In addition to selective 24:454474
Biointerface 237
Pan FS, Cheng QL, Jia HP, Jiang ZY (2010) Facile A biointerface is an interface between biological
approach to polymer-inorganic nanocomposite mem- material (e.g., cells, tissue) and a (commonly)
brane through a biomineralization-inspired process.
J Membr Sci 357:171177 synthetic substrate. Biointerfaces call for
Rawlings AE, Bramble JP, Staniland SS (2012) Innovation advanced design and preparation in order to
through imitation: biomimetic, bioinspired and match the sophisticated (bio) recognition ability B
biokleptic research. Soft Matter 8:26752679 of biological systems. Specically this requires
Sharma M, Yi MG, Dong H, Qin HJ, Peterson E, Busath
DD, Zhou HX, Cross TA (2010) Insight into the mech- combined topographic, chemical, and viscoelastic
anism of the inuenza aproton channel from a structure patterns on surfaces to match proteins at the nm
in a lipid bilayer. Science 330:509512 scale and cells at the micrometer scale (Kasemo
Tokarev I, Minko S (2010) Stimuli-responsive porous 2002).
hydrogels at interfaces for molecular ltration, separa-
tion, controlled release, and gating in capsules and Biointerfaces are typically used in the elds of
membranes. Adv Mater 22:34463462 immunology, surgery, pathology, pharmacology,
Wang JT, Zhang H, Yang XL, Jiang SA, Lv WJ, Jiang ZY, and many others either for the proliferation of
Qiao SZ (2011a) Enhanced water retention by using cells for further application or investigative
polymeric microcapsules to confer high proton conduc-
tivity on membranes at low humidity. Adv Funct Mater study or as a platform for cell manipulation.
21:971978 These studies have become even more topical
Wang M, Janout V, Regen SL (2011b) Hyper-thin organic upon completion of the human genome project,
membranes with exceptional H2/CO2 permeation selec- which has set the formidable challenge of eluci-
tivity: importance of ionic crosslinking and self-
healing. Chem Commun 47:23872389 dating the function of the vast amount of genomic
Wang JT, Yue XJ, Zhang ZZ, Yang Z, Li YF, Zhang H, information within the genotype of an individual.
Yang XL, Wu H, Jiang ZY (2012) Enhancement of A potential medical application of a
proton conduction at low humidity by incorporating biointerface is in the ex situ growth of skin/bone/
imidazole microcapsules into polymer electrolyte
membranes. Adv Funct Mater. doi:10.1002/ organ tissue for subsequent transplantation. For
adfm.201201436 example, burn victims can have their own replace-
Xu T, Zhao N, Ren F, Hourani R, Lee MT, Shu JY, Mao S, ment skin grown in a laboratory via the prolifera-
Helms BA (2011) Subnanometer porous thin lms by tion of their own epidermal cells upon a
the co-assembly of nanotube subunits and block copol-
ymers. ACS Nano 5:13761384 biointerface in a laboratory. Biointerfaces them-
Zeng H, Hwang DS, Israelachvili JN, Waite JH selves can potentially be installed in situ to assist
(2010) Strong reversible Fe3+-mediated bridging in cell regeneration, such as after surgery. For this
between dopa-containing protein lms in water. Proc purpose, biodegradable biointerfaces are of par-
Natl Acad Sci 107:1285012853
ticular interest as they perform their function of
assisting cell regeneration and are then degraded
naturally into nontoxic products which are then
excreted (Hook et al. 2006).
Biointerface The scientic study of cells is important to
improve fundamental understanding of their
Cameron Shearer working process. Biointerfaces are often used as
School of Chemical and Physical Sciences, experimental platforms to study various proper-
Flinders University, Adelaide, South Australia, ties of cells. For instance, neuronal cells are elec-
Australia trically excitable cell that processes and transmits
Institute of Physical Chemistry, University of information through electrical and chemical sig-
Muenster, Muenster, North Rhine-Westphalia, nals. The growth of these cells on specially
Germany designed biointerfaces has allowed scientists to
directly investigate their signal transmission pro-
cess (Robinson et al. 2012).
Synonyms The fabrication of various biointerfaces is of
particular interest for stem cell research. Stem
Biorecognition substrate cells can differentiate into a variety of different
238 Biological and Synthetic Membrane
cell types depending upon their environment Introduction

(Pittenger et al. 1999). Therefore, the careful
design of particular biointerfaces can promote Biological membranes are enclosing or separating
stem cell differentiation into a desired cell type. membranes that act as selectively permeable bar-
For example, stem cells adsorbed on harder sur- riers within living things. Biological membranes
faces tend to differentiate into bone-like cells have dynamic structures composed of phospho-
(Engler et al. 2006). lipid molecules, proteins, and carbohydrates
(Brown and London 1998; Diamond and Wright
1969; Lee 2003). Steroid cholesterol is also found
Cross-References in the biological membranes of eukaryotic
organisms.
Stem Cell The phospholipid component of the biological
membrane plays an important role in the property
of biological membranes. A phospholipid layer
References
consisted of a hydrophilic head that interacts
Engler AJ, Sen S, Sweeney HL, Discher DE (2006) Matrix with polar uids (e.g., water), and its hydrophobic
elasticity directs stem cell lineage specication. Cell tail always meets with the hydrophobic tail of
126(4):677689 another layer (forming a phospholipid bilayer or
Hook AL, Thissen H, Voelcker NH (2006) Surface manip- a liposome) or its own (forming a micelle). This
ulation of biomolecules for cell microarray applica-
tions. Trends Biotechnol 24(10):471477 structure maintains the uidity of cells and organ-
Kasemo B (2002) Biological surface science. Surf Sci elles. The embedded, integral, and peripheral pro-
500(13):656677 teins of biological membranes are used in
Pittenger MF, Mackay AM, Beck SC et al (1999) communication and transportation of chemicals
Multilineage potential of adult human mesenchymal
stem cells. Science 284(5411):143147 and ions. Typically, the phospholipid bilayer acts
Robinson JT, Jorgolli M, Shalek AK et al (2012) Vertical as an impermeable barrier to most molecules and
nanowire electrode arrays as a scalable platform for molecules can only cross the membrane with the
intracellular interfacing to neuronal circuits. Nat aid of proteins.
Nanotechnol 7(3):180184
Biological membranes, in the form of cell
membranes (i.e., plasma membranes), are selec-
tively permeable to ions and organic molecules
and thus could control the movement of sub-
Biological and Synthetic Membrane stances in and out of cells. Cells also contain
intracellular membranes, which act as the bound-
Bioarticial Synthetic Membrane ary of various intracellular organelles, e.g., the
nucleus and the mitochondria.
Biological Membrane
Cross-References
Zhiwei Wang
Department of Environmental Engineering, Bioarticial Synthetic Membrane
College of Environmental Science and Cellular Membranes
Engineering, Tongji University, Shanghai, China
References
Synonyms Brown DA, London E (1998) Functions of lipid rafts in
biological membranes. Annu Rev Cell Dev Biol
Biomembrane 14:111136
Biomaterial 239
Diamond JM, Wright EM (1969) Biological membranes: Natural-based polymers offer the advantage of
the physical basis of ion and nonelectrolyte selectivity. being similar to extracellular matrix components,
Annu Rev Physiol 31:581646
Lee AG (2003) Lipid-protein interactions in biological and they have binding sites for cells and adhesion
membranes: a structural perspective. BBA-Biomembr molecules so that they usually avoid immunolog-
1612(1):140 ical reactions and toxicity. Their principal disad- B
vantage lies in the development of reproducible
production methods, because their structural com-
plexity often makes modication and purication
difcult. Additionally, signicant batch-to-batch
Biological Membrane Reactors variations occur because of their bio preparation
in living organism (plants, crustaceans) while
Membrane Bioreactors their mechanical properties are also limited.
Principle of Biocatalytic Membrane Reactors Chemical modication and cross-linking are com-
(BMR) mon methods for altering mechanical properties
such as elasticity and tensile strength of natural
materials which are often not appropriate for
dynamic physiologic environments. The most fre-
Biomaterial quently used natural biomaterials are the collagen
derived from connective tissues of animals, brin-
Loredana De Bartolo ogen obtained from plasma, and other proteins of
Institute on Membrane Technology, National ECM. Among natural biomaterials, polysaccha-
Research Council of Italy, ITM-CNR, Rende, rides are also frequently used. These natural poly-
Italy mers include agarose, hyaluronan, chitosan, and
alginate. They are derived from plant (seaweed) or
animal origin and are used as hydrogels to support
A biomaterial is any material in contact with a wound healing and also cell growth and differen-
biological system. The rst nonbiological mate- tiation (Mano et al. 2007).
rial entered in contact with the human body seem Synthetic biomaterials different from natural
to date back to the prehistory 9000 years old with materials can be synthesized to give various prop-
the Kennewick Man, who had a spearpoint erties and have high reproducibility, availability,
embedded in his hip. This spearpoint was the lot-to-lot uniformity, and constant quality. They
rst foreign material implant, which was toler- have a well-known structure and properties so that
ated by the body (Ratner et al. 2012). Although mechanical properties, degradation rate, shape,
the use of biomaterials dates back to the ancient and composition can be adjusted to the specic
time, the study of biomaterials is about application. However, the biocompatibility and
60 years old. immune reactions are the big concern because
Biomaterials can be classied on the basis of they often lack sites for cell adhesion therefore
its (i) source in natural or synthetic materials if their interactions with cells could be questionable.
they are from nature or synthesized in laboratory, Synthetic biomaterials include biodegradable
respectively; (ii) chemical nature in organic, inor- polymers as poly(lactic-co-glycolic acid)
ganic, and composites; and (iii) degradation prop- (PLGA), poly-lactic acid (PLA), poly-
erties in biodegradable and nonbiodegradable. caprolactone (PCL), poly(ethylene glycol)
They can be used in biomedical devices for the (PEG), etc. and synthetic polymers (e.g., polyure-
replacement of an organ function or implanted thane, polyetheretherketone, polyethylene,
into the body for the replacement of tissue, polyvinylidene uoride, polyvinyl alcohol,
organ, or body parts. Biomaterials are also used polymethyl methacrylate, polydioxanone, etc.).
every day in dental applications, surgery, and drug For example, PLGA is one of the most frequently
delivery (Williams 2009). used biomaterials applied in neural, bone, and
240 Biomembrane
cartilage tissue engineering for its degradation

rate that can be modulated. PLA is fast reabsorbed Biomembrane
and widely used in tissue engineering applications
alone or blended with other polymers. PCL is a Biological Membrane
very good elastic biodegradable polymer that Cellular Membranes
nds several applications in tissue engineering.
PEG that is used in pharmaceutical technology
to modulate the degradation and absorption of
bioactive proteins is also frequently used in Biomimetic Membrane Approach
bone, cartilage, nerve, liver, and vascular tissue
engineering. Peptide-based biomaterials (e.g., the N. M. Kocherginsky
tripeptide Arg-Gly-Asp, RGD) consisting of short Biomime, Urbana, IL, USA
amino acid sequences exhibit cell-binding
sequences found in the extracellular matrix natu-
ral proteins. Ceramic-based biomaterials that are Biomimetic membrane approach is based on
totally or partly inorganic are applied in bone articial membranes which imitate fundamental
tissue engineering. Mostly they are shaped with properties and functions of biological membranes.
heat-forming porous, brittle materials. Hydroxy- Biomimetics develops simple and efcient syn-
apatite and tricalcium phosphate owing to their thetic systems that mimic the formation, structure,
structural similarity to the mineral phase of bone or function of biological materials, leading to new,
are used alone or in combination with organic more efcient, and cheap applications. Major
polymers like PLA, PLGA, collagen, or chitosan potential applications of biomimetic membranes
in the bone replacement or regeneration are drug delivery, high-throughput screening, sen-
(De Bartolo and Bader 2013; Morelli et al. sors, and separation and purication, possibly
2015). Metals like alumina and titanium alloys coupled with chemical synthesis (Hlix-Nielsen
that are bioinert materials are used as implant 2012). The same membrane can serve an opposite
materials where implants are subjected to extreme purpose and be used as an experimental model of
mechanical load, like articular prostheses, dental biomembranes. The term biofunctional mem-
implants, or heart valves. brane is more appropriate for the membranes,
which include biological components, such as
specic enzymes, aquaporin, antigens, etc.
Biological membranes have a lot of different
References
specic functions but all of them separate either
De Bartolo L, Bader A (2013) Biomaterials for
two aqueous phases or an external gaseous and
stem cell therapy. CRC Press, Boca Raton/New York/ internal aqueous phase. Three major chemical
Oxon components of biological membrane are proteins,
Mano JF, Silva GA, Azevedo HS, Malafaya PB, Sousa RA, lipids, and polysaccharides, and selective perme-
Silva SS, Boesel LF, Oliveira JM, Santos TC, Marques
AP, Neves NM, Reis RL (2007) Natural origin biode-
ability is based on direct diffusion, facilitated
gradable systems in tissue engineering and regenerative transport with carriers, and penetration through
medicine: present status and some moving trends. J R the channels. Coupling with various energy
Soc Interface 4:9991030 sources, including light, adenosine triphosphate,
Morelli S, Salerno S, Holopainen J, Ritala M, De Bartolo L
(2015) Osteogenic and osteoclastogenic differentiation
and different ion concentrations separated by the
in polylactic acid-nanohydroxyapatite ber scaffolds. membrane aqueous solutions, leads to the trans-
J Biotechnol 204: 5362 membrane active transport. Membrane processes
Ratner BD, Hoffman AS, Schoen FJ, Lemons JE can be controlled by temperature, transmembrane
(2012) Biomaterials science: an introduction to mate-
rials in medicine, 3rd edn. Elsevier, Amsterdam
voltage, biochemical mediators, and drugs.
Williams DF (2009) On the nature of biomaterials. Bio- Amphiphilic molecules are usually able to
materials 30:58975909 spontaneously form liquid crystal structures.
Biomimetic Membranes 241
These self-assembled structures, especially lipid controlled by temperature-induced phase transi-

monolayers, multilamellar vesicles, unilamellar tions and even membrane-active organic cations.
liposomes, and bilayer lipid membranes can be The process may be used for water purication.
used as models of biomembranes. To improve Nitrocellulose-based biomimetic membranes
their stability, which is important for practical may also be used in potentiometric drug sensors B
applications, lipid membranes are often (sensitivity often is 105M), and combination
immobilized on a solid support, which may be with electrochemical tests makes this system
metal, glass, semiconductors, and polymers. The attractive in the high-throughput drug screening,
well-known example is Langmuir-Blodgett lms. such as Parallel Articial Membrane Permeabil-
Lipids or their analogs may have an additional ity Assay (PAMPA).
anchor group for tethering to a support. Sulfhy-
dryl or disulde groups are often used for
tethering to the metals such as gold or silver, References
while methyl- and methoxy-substituted silane
groups for tethering to glass, quartz, silica, and Hlix-Nielsen C (ed) (2012) Biomimetic membranes for
mica. Micro- and nano-porous solid supports, sensor and separation applications. Springer, Dordrecht
Kocherginsky NM, Lvovich VF (2010) Biomimetic mem-
including hydrophobic membrane lters, are also branes with nano channels but without proteins. Imped-
used to stabilize thin and nonstable bilayer lipid ance of impregnated cellulose esters lters. Langmuir
membranes. 26:1820918218
Though the thickness of lipid membranes
(~50 ) and some other structural properties are
similar to the characteristics of biological mem-
branes, electrical conductivity and permeability Biomimetic Membranes
for inorganic ions are often four orders of magni-
tude less than those of biomembranes with aque- Zhongyi Jiang, Jing Zhao, Shaofei Wang,
ous channels. Dong Yang and Hong Wu
Nitrocellulose microlters with pores lled School of Chemical Engineering and Technology,
with fatty acids and other lipid analogs are Tianjin University, Nankai, Tianjin, China
another type of stable biomimetic membrane
(Kocherginsky and Lvovich 2010). These impreg-
nated lters have nanolayers of water covering Biomimetic membrane is designed and fabricated
the internal pore surface and separating nitro- by imitating the composition, structure, formation
cellulose and liquid crystal structures inside process, as well as functions of biological mem-
the pores. Aqueous nanolayers imitate aqueous brane. As we know, biological membrane is the
channels in biomembrane and have cation/anion vital structural basis in organism, on which the
selectivity due to the presence of carboxylic essential life activities in cell take place. The
groups immobilized on the nitrocellulose poly- asymmetric and hierarchical bilayer structure
mer chains. Electric resistance and many other formed by the in situ self-assembly of amphiphilic
physically independent properties of these mem- phospholipid and functional proteins (Nielsen
branes, including direct permeability of water, 2009) ensures the biological membrane take
nonelectrolytes and respiratory gases (O2 and important roles in keeping the environmental sta-
CO2), and other properties recalculated for 50 bility of cell, participating in the multiple interac-
thickness, are very similar to parameters for bio- tions of different cells, and efciently achieving
logical membranes. Low molecular mass carriers the conversion of energies, the signal transduction
can be dissolved in the impregnating oils, leading between cells, as well as the conversion and trans-
to imitation of secondary active transport of membrane transport of some substances (Zhu
metal ions and even electrons, driven by trans- et al. 2011). Biological membrane has a very
membrane difference of pH. Kinetics may be sophisticated structure consisting of abundant
242 Biopolymers
species of lipids and proteins, which is difcult to 2010). In addition to act as model of biological
imitate completely due to the constraints of cur- membrane for investigating protein-lipid interac-
rent science and technologies (Eeman and Deleu tion, this type of biomimetic membrane performs
2010). As a simplication and transition, biomi- well as functional membrane in various elds by
metic membranes containing part of (not all) the incorporating different functional proteins:
key components or characteristics of biological (1) energy conversion (energy converter, energy-
membrane are being actively developed in the production devices), (2) signal transduction
last few decades. (biosensor and bioanalysis), and (3) substance
Biomimetic membranes can be divided into transport and conversion (drug delivery,
two categories in terms of structural morphology: immobilized enzyme catalysis, and membrane-
freestanding membrane and supported membrane. based separation). To further exploit the applica-
The freestanding membrane is often utilized for tion efciency and area of biomimetic membrane,
theoretical investigation purpose, in other words, more efforts may be devoted to the following two
as a simplied model for studying the structure aspects: (1) utilize channel proteins directly for
and function of biological membranes. This type the preparation of biomimetic membrane while
of biomimetic membranes mainly includes the acquiring the good control over the channel den-
lipid monolayer membrane and the bilayer lipid sity and the protein activity and (2) deepen the
membrane or named black lipid membrane insight into the structure-function relationship of
(BLM) (Eeman and Deleu 2010; Kim membrane proteins.
et al. 2012). The BLM is suitable for investigating
membrane phase behavior and membrane pro-
cesses such as membrane fusion and molecular References
recognition for its closer structure with biological
membrane, while the lipid monolayer is suitable Duong PHH, Chung TS, Jeyaseelan K, Armugam A,
Chen Z, Yang J, Hong M (2012) Planar biomimetic
for mimicking and evaluating the insertion of
aquaporin-incorporated triblock copolymer mem-
amphipathic compounds (Eeman and Deleu branes on porous alumina supports for nanoltration.
2010). The biomimetic membrane has provided J Membr Sci 409410:3443
an effective platform for exploring the complex Eeman M, Deleu M (2010) From biological membranes to
biomimetic model membranes. Biotechnol Agron Soc
structure and function of biological membrane,
14:19736
even though further endeavors both in biological Kim YR, Jung S, Ryu H, Yoo YE, Kim SM, Jeon TJ
science and fabrication technology are still (2012) Synthetic biomimetic membranes and their sen-
needed. sor applications. Sensors 12:95309550
Li X, Wang R, Tang C, Vararattanavech A, Zhao Y,
The supported membrane often contains a
Torres J, Fane T (2012) Preparation of supported lipid
bilayer residing onto a solid substrate, which del- membranes for aquaporin Z incorporation. Colloids
icately ameliorates the weak stability and short Surf B 94:333340
lifetime of BLM (Li et al. 2012). Furthermore, Nielsen CH (2009) Biomimetic membranes for sensor and
separation applications. Anal Bioanal Chem
the self-assembly of block copolymer (black poly-
395:697718
mer membrane, BPM) as another approach to Zhu B, Li J, Xu D (2011) Porous biomimetic membranes:
form the bilayer can be an alternative of BLM fabrication, properties and future applications. Phys
for its higher stability, controllability, and ability Chem Chem Phys 13:1058410592
to prevent the direct contact of embedded proteins
to solid substrate, which otherwise will inactivate
the proteins (Duong et al. 2012; Nielsen 2009).
The supported BLM or BPM can be prepared by Biopolymers
several methods, of which vesicle collapse tech-
nique is commonly used (Eeman and Deleu Biodegradable Polymer for Membranes
Bioreactor Membrane 243
titanium oxide, have been used in bioreactors, and

Bioreactor Membrane an ideal material should have low afnity with bio-
polymers/foulants, sound mechanical strength, and
Zhiwei Wang good separation efciency and is nonbiodegradable
Department of Environmental Engineering, and cost-effective (Su et al. 2014). B
College of Environmental Science and In recent years, dynamic membranes have been
Engineering, Tongji University, Shanghai, China investigated in MBRs in order to improve mem-
brane performance (Meng et al. 2009). Contrast to
conventional synthetic bioreactor membranes,
Synonyms dynamic membranes can be formed by ltering a
solution containing either inorganic or organic mate-
Membranes used for bioreactors rials (particulate, colloidal, and even soluble sub-
stances) through a coarse support material (Fan and
Huang 2002). The separation efciency of dynamic
Introduction membranes is comparable to microltration and
ultraltration membranes once the dynamic mem-
A bioreactor membrane is a synthetic organic or brane is well formed. There are two basic types of
inorganic membrane used for solid-liquid separa- dynamic membranes: pre-coated and self-forming.
tion in membrane bioreactors (MBRs). The mem- The pre-coated membrane is produced by passing a
branes can be pressure driven (microltration, solution of one or more specic components over the
ultraltration, nanoltration, and so on) and surface of a porous support. The self-forming mem-
osmotically (forward osmosis) or thermally brane is formed by the components in the solution to
driven (membrane distillation); however, be ltered (Meng et al. 2009).
pressure-driven membranes are currently more
popular. In most cases, MBR refers to the combi-
nation of a membrane process with a suspended Cross-References
growth bioreactor that is applied for water/waste-
water treatment and/or for other purposes (e.g., Ultraltration Membrane Bioreactor
fermentation). When used in wastewater treat-
ment, sound retention of suspended solids by bio-
reactor membranes enables the distinctive
References
advantages of MBRs, including a reduced foot- Fan B, Huang X (2002) Characteristics of a self-forming
print, an independent control of sludge retention dynamic membrane coupled with a bioreactor for
time and hydraulic retention time, a higher bio- municipal wastewater treatment. Environ Sci Technol
mass concentration, and a high-quality efuent 36(23):52455251
Judd S (2008) The status of membrane bioreactor technol-
(Judd 2008; Wang et al. 2008; Meng et al. 2009). ogy. Trends Biotechnol 26(2):109116
Bioreactor membrane characteristics such as Meng FG, Chae SR, Drews A, Kraume M, Shin HS, Yang
pore size, porosity, surface charge, roughness, and FL (2009) Recent advances in membrane bioreactors
hydrophilicity/hydrophobicity have been proven to (MBRs): membrane fouling and membrane material.
Water Res 43(6):14891512
impact on MBR performance, especially on mem- Su XY, Tian Y, Zuo W, Zhang J, Li H, Pan XY (2014) Static
brane fouling (Meng et al. 2009). A series of organic adsorptive fouling of extracellular polymeric sub-
and inorganic membrane materials, including stances with different membrane materials. Water Res
polyvinylidene uoride, polyethylene, polyacrylo- 50:267277
Wang ZW, Wu ZC, Mai SH, Yang CF, Wang XH, An Y,
nitrile, polyethersulfone, polytetrauoroethylene, Zhou Z (2008) Research and applications of membrane
polypropylene, polylsulfone, cellulose acetate, ali- bioreactors in China: progress and prospect. Sep Purif
phatic polyamides, aluminum, zirconium, and Technol 62(2):249263
244 Biorecognition Substrate
and anaerobic digestion. A typical bioprocess

Biorecognition Substrate for bioethanol production from lignocellulosic
materials includes the following major steps:
Biointerface pretreatment, hydrolysis, fermentation, and prod-
uct (bioethanol) purication (He et al. 2012).
First, the lignocellulosic material is pretreated to
remove lignin and hemicelluloses. The remaining
cellulose is then hydrolyzed to produce sugars
Biorefinery which are fermented to produce ethanol. Finally,
the ethanol is recovered and puried for commer-
Baoqiang Liao cial application. Separation technologies play
Department of Chemical Engineering, Lakehead an important role in this biorening process,
University, Thunder Bay, ON, Canada including ion-exchange resins that have been
used for the detoxication of fermentation hydro-
lyzates, distillation in the recovery of ethanol, and
A biorenery is a facility that integrates biomass ethanol dehydration. Membrane separation has
conversion processes and equipment to produce been utilized in various stages of this biorening
fuels, power, heat, and value-added chemicals process, including concentration of bioethanol
from biomass (NREL web site 2014). The by using membrane bioreactors and the removal
biorenery concept is analogous to todays petro- of water from ethanol solution (Abels et al. 2013).
leum reneries, which produce multiple fuels and Basically, there are three generations of bio-
products from petroleum (NREL web site 2014). fuels, based on the use of different lignocellulosic
Raw materials for biorenery are renewable lig- biomass. First-generation biofuels include
nocellulosic biomass, which include agricultural bioalcohols, biodiesels, biogas, and biosyngas
residues, wood residues, food wastes, grass, sug- and are typically derived from sugar, starch, and
arcane, algal biomass, etc. The products of vegetable oil (Demirbas 2011). Biofuels from
biorenery include various types of biochemicals agroforestry wastes including the stalks of
(cellulose, hemicellulose, lignin, succinic acid, wheat, corn, wood, and dedicated nonfood-based
lactic acid, lipids, adipic acid, butanol, fatty alco- bioenergy feedstocks (e.g., Miscanthus, willow,
hols and esters, ethyl acetate, glycols, etc.), and poplar) form the basis of the second-
biofuels (bioethanol, biogas, bio-oil), and electric- generation biofuels (Sims et al. 2010). Third
ity. There are three main technical pathways for (3rd)-generation biofuels are derived from algae
biorenery of lignocellulosic biomass: thermo- (Costa and de Morais 2011). The fast-growing
chemical conversion, physicochemical conversion, rate and utilization of CO2 as carbon source
and biochemical conversion. Thermochemical make microalgae a fascinating biomass for bio-
conversion uses heat and catalysis to produce diesel production, due to the high lipid content in
biofuels and bioproducts and includes combus- microalgae. Optimization of processes for effec-
tion, gasication, pyrolysis, and liquefaction of tive and efcient biomass conversion for rst-,
biomass. Physicochemical conversion applies second-, and third-generation biofuels and for
physical and chemical methods to produce true integration in a biorenery will require appro-
biofuels and includes pressing/extraction and priate separation technologies. Compared to other
esterication. Biochemical conversion utilizes separation technologies, the low energy consump-
enzymes and/or microorganisms to convert ligno- tion, greater separation efciency, reduced num-
cellulosic biomass into bioproducts and/or ber of processing steps, and high quality of the
biofuels and includes hydrolysis and fermentation nal product are the main attractions of membrane
Biorefinery with Membrane Operations 245
separation processes in biorening and bioenergy membrane technology in wastewater biorenery

production (de Morais Coutinho et al. 2009). is the development of AnMBR technology.
AnMBR is an integration of membrane separation
with anaerobic digestion technology for bioenergy
References recovery. The use of membrane completely retains B
the slow-growing methanogens and eliminates bio-
Abels C, Carstensen F, Wessling M (2013) Membrane
processes for biorenery. J Membr Sci 444:285317 mass separation problems. To reduce energy usage,
Costa JAV, de Morais MG (2011) The role of biochemical biogas sparging is used for membrane fouling con-
engineering in the production of biofuels from trol. AnMBR has the advantages of small footprint,
microalgae. Bioresour Technol 102:29 high productivity, superior efuent quality for
de Morais Coutinho C, Chiu MC, Basso RC, Ribeiro APB,
Gonalves LAG, Viotto LA (2009) State of art of the water reuse, and system closure. AnMBR has
application of membrane technology to vegetable oils: been successfully used for food processing waste-
a review. Food Res Int 42:536550 water treatment (Christian et al. 2010) and alcohol
Demirbas A (2011) Competitive liquid biofuels from bio- stillage wastewater treatment (Grant et al. 2008) for
mass. Appl Energy 88:1728
He Y, Bagley DM, Leung KT, Liss SN, Liao BQ bioenergy recovery in full-scale plants. This indi-
(2012) Recent advances in membrane technologies cates the successful application of membrane tech-
for biorening and bioenergy production. Biotechnol nology in the production of rst-generation biofuel
Adv 30(4):817858 (biogas). The AnMBR technology achieved a high
National Renewable Energy Laboratory (NREL) website
(2014): www.nrel.gov/biomass/biorenery chemical oxygen demand removal (>98 % and
Sims REH, Mabee W, Saddler JN, Taylor M (2010) An 99.4 %) and bioenergy recovery in food processing
overview of second generation biofuel technologies. wastewater and distillated wastewater treatment.
Bioresour Technol 101:15701580 In the production of bio-ethanol, the rst- and
second-generation of biofuel, membrane separa-
tion technology has been successfully applied to
Biorefinery with Membrane concentrate bio-ethanol from fermentation broth
Operations and replace distillation for dehydration of
bio-ethanol to save energy. In bio-ethanol produc-
Baoqiang Liao tion, the membrane is incorporated with the fer-
Department of Chemical Engineering, Lakehead menters to create a membrane bioreactor for
University, Thunder Bay, ON, Canada biomass separation and bio-ethanol concentration
(Abels et al. 2013). To produce high-purity
bio-ethanol (>99.5 %), pervaporation has been
Membrane separations have been widely used in successfully used for dehydration of bio-ethanol
various stages of biorenery, including upstreams (Oliverio et al. 2010). The principle of membrane
for product manufacturing and downstreams for dehydration of bio-ethanol involves two steps:
product purication. Full-scale biorenery with molecules diffuse through the membrane as per-
membrane operations includes wastewater meate and vaporization of the permeate into vapor
biorenery for bioenergy recovery using anaero- phase through vacuum. The membrane system
bic membrane bioreactors (AnMBR), dehydration can be directly integrated in the rectication of
of bio-ethanol using membrane separations, and existing ethanol plants or new ethanol systems. In
lignin separation in integrated forest biorenery. doing so the energy demand of the complete eth-
In wastewater biorenery, the potential energy anol system is minimized by arranging an opti-
stored in the organic compounds can be recovered mized equipment conguration and strict heat
as bioenergy (methane or hydrogen) through integration. Both polymeric and ceramic mem-
anaerobic digestion. One successful example of branes are used for bio-ethanol dehydration.
246 Biorefractory Pollutants
Full-scale membrane dehydration plants have

been in operations all over the world. Thirty-ve Biorefractory Pollutants
to 70 % reduction in energy consumption was
observed in the downstream of bio-ethanol pro- Marc Cretin1,2 and Geoffroy Lesage3
1
duction, as compared to the distillation processes Institut Europen des Membranes, ENSCM-
(Oliverio et al. 2010). UM-CNRS (UMR 5635), Universit de
In integrated forest biorenery, membrane lter Montpellier, Montpellier Cedex 5, France
2
has been used for lignin recovery from black liquor National School of Chemistry of Montpellier,
in kraft pulp mills. The lignin recovery process University of Montpellier, Montpellier, France
3
includes CO2 precipitation, lignin separation, acid Institut Europen des membranes, IEM,
washing, and product drying. In a pilot testing in UMR-5635, Universit de Montpellier, ENSCM,
lignin precipitation and ltration, a LAROX lter CNRS, Montpellier, France
equipped with a lter cloth from Tamfelt (S-2108-
L1) was used to separate lignin from the aqueous
solution after coagulation (Kouisni and Paleologou A variety of compounds are classied as refractory
2010). A ltration rate of 5060 kg/m2.h was when they are poorly biodegraded and/or exhibit a
achieved with no black liquor pretreatment low value for the ratio of biological oxygen demand
(oxidation). After pretreatment (oxidation) of to chemical oxygen demand (BOD/COD)
black liquor, the ltration rate was increased to (Frimmel et al. 2008; Kopinke et al. 2002). These
150180 kg/m2.h (Kouisni and Paleologou 2010). compounds are resistant or difcult to biodegrade,
A LignoBoost commercial lignin plant with a are toxic or are inhibitory to bacterial growth, and
production capacity of 70 tons lignin/day has been present an increasing problem in water and waste-
in operation in Domtar in Plymouth, NC, USA, water treatment systems (Delay et al. 2015).
since 2013 (Mitchell 2013). Another new full- These compounds could be organic
scale lignin plant using the LignoForce technol- (halogenated organics, aromatic and aliphatic
ogy is being built in Alberta, Canada. hydrocarbons, pesticides, pharmaceutical prod-
ucts) or inorganic (metals, nanoparticles). Their
origin could be anthropogenic or natural.
References
Abels C, Carstensen F, Wessling M (2013) Membrane

processes for biorenery. J Membr Sci 444:285317 Cross-References
Christian S, Grant S, Wilson D, Mccarthy P, Mills D,
Kolakowski M (2010) The rst two years of full-scale
anaerobic membrane bioreactor (AnMBR) operation Persistent Organic Pollutant
treating a high strength industrial wastewater at
Kens Foods Inc. Proc Water Environ Fed 2010(13):
40194033 References
Grant S, Page I, Moro M, Yamamoto T (2008) Full-scale
applications of the anaerobic membrane bioreactor Frimmel FH, Abbt-Braun G, Heumann KG, Hock B,
process for treatment of stillage from alcohol production Ludemann HD, Spiteller M (Eds) (2008) Refractory
in Japan. Proc Water Environ Fed 2008(7):75567570 organic substances in the environment. John Wiley & Sons
Kouisni L, Paleologou M (2010) Method for separating lignin Kopinke FD, Georgi A, Mackenzie K, Kumke M,
from black liquor. US Patent: US 20110297340 A1 Frimmel FH (2002) Refractory Organic Substances in
Mitchell WD (2013) Commercial production of lignin the Environment
from black liquor. Domtar Corporation, Raleigh Delay M., Schwegmann H, Frimmel FH (2015)
Oliverio JL, Filho HG, Boscariol FC, Cote P (2010) Etha- Nanoparticles and refractory organic matter: Interac-
nol dehydration system by SiftekTM polymeric mem- tions and consequences. Journal of Environmental
brane. Proc Int Soc Sugarcane Technol 27:19 Chemical Engineering
Biosensor-Based Membrane 247
molecular membrane. These membranes consist

Biosensor-Based Membrane of one or more lipid bilayer membranes, mimick-
ing cell walls, upon a conducting (often gold)
Cameron Shearer substrate (Cornell 2002). Many schemes have
School of Chemical and Physical Sciences, been reported for modulating the membrane B
Flinders University, Adelaide, South Australia, ion conduction in response to the presence of
Australia a specic target analyte. The most common
Institute of Physical Chemistry, University of membrane-based biosensor was introduced by
Muenster, Muenster, North Rhine-Westphalia, Cornell et al. where a mobile biological ion
Germany channel, initially gramicidin A, is tethered to
an antibody of interest (Cornell et al. 1997).
Under standard conditions, the ion channel is
Synonyms mobile throughout the membrane and predom-
inantly exists in a dimer arrangement with a
Membranes for biosensors substrate-immobilized ion channel. Upon intro-
duction of the target analyte, the analyte binds
A biosensor is an analytical device used for the to both an immobilized antibody and the anti-
quantitative and qualitative detection of a spe- body tethered to the ion channel resulting in a
cic analyte. A biosensor broadly consists of conformational change of the ion channels,
three main components: the biological sensing reducing conductance through the membrane
element which specically interacts with the ana- (Cornell et al. 1997; Cornell 2002). From the
lyte, the transducer which is used to convert the change in conductance, a quantitative value of
signal from the bioelement-analyte interaction the amount of analyte present can be calcu-
into a signal which is readable and reportable lated. Membrane biosensors are promising
by the third component, the reading device for a range of diagnostic studies as the lipid
(Fig. 1). membrane environment is ideal for the long-
Membrane-based biosensors operate by term storage of biological sensing elements
reporting changes in ionic conduction through a (Oh et al. 2008).
Biosensor-Based Membrane, Fig. 1 Schematic of basic mechanism of a membrane-based biosensor based upon the
change in conductance upon interaction between an antibody tethered to the ion channel and target analyte (antigen)
248 Biosurfactant
Cross-References pH values, besides presenting high surface activ-

ity and low critical micelle concentration values
Cell Membrane emerging as promising substitutes of the latter
Lipid Monolayer/Bilayer (Vaz et al. 2012).
Concerning the oil industry, biosurfactants can
be used in the bioremediation of oil pollutants, in
References
the enhanced oil recovery (reducing the capillary
Cornell BA (2002) Membrane-based biosensors, forces that retain the oil in the reservoir rock), in
Chapter 15. In: Frances SL, Chris ART (eds) Optical the treatment of oily sludges, and in the cleanup of
biosensors. Elsevier, Amsterdam, pp 457495 storage tanks. They can also be applied in the
Cornell BA, Braach-Maksvytis VLB, King LG et al (1997) pharmaceutical (due to their antimicrobial activ-
A biosensor that uses ion-channel switches. Nature
387(6633):580583 ity) and agricultural sectors (for the hydrophi-
Oh SY, Cornell B, Smith D et al (2008) Rapid detection of lization of heavy soil), as a pesticide, and in the
inuenza a virus in clinical samples using an ion chan- food industry (Mukherjee and Das 2010; Banat
nel switch biosensor. Biosens Bioelectron 23(7): 1995; Freire et al. 2009). They can also be used in
11611165
the efuent treatment, for metal ion removal from
aqueous solutions (Ramani et al. 2012) and to
enhance the water/oil interaction in the degrada-
tion of high fat content efuents (Damasceno
Biosurfactant et al. 2012).
Regarding membrane separation processes,
Frederico de Araujo Kronemberger biosurfactants can be used in the micellar
COPPE Chemical Engineering Program, enhanced ultraltration, in which small contami-
Federal University of Rio de Janeiro (UFRJ), Rio nants, like heavy metal ions, are bound onto larger
de Janeiro, RJ, Brazil surfactant micelle complexes. These ions associ-
ated with surfactant macromolecules can be easily
retained by an ultraltration membrane module.
A biosurfactant (also called microbial surfactant) El Zeftawy and Mulligan (2011) investigated the
can be dened as a surfactant compound produced use of rhamnolipid biosurfactants (glycolipids) in
by microorganisms. Since surfactants are amphi- the micellar enhanced ultraltration of cadmium,
philic compounds, presenting both hydrophilic copper, nickel, lead, and zinc ions with
and hydrophobic moieties, they tend to move polysulfone hollow-ber ultraltration mem-
toward the interfaces, reducing the surface branes with molecular weight cutoff of 10,000
and/or interfacial tensions. These properties and 30,000. The authors reported nearly complete
make surfactants excellent detergency, emulsier, rejection of the ions, considering a feed concen-
foaming, and dispersing agents (Mukherjee and tration of 5 mg.L1, when using a rhamnolipid
Das 2010). Biosurfactants can be categorized solution above its critical micelle concentration
mainly by their chemical composition and their with both membranes. Even when increasing the
microbial origin, being classied as glycolipids, metal ion content up to 50 mg.L1, the
lipopeptides and lipoproteins, phospholipids and biosurfactant concentration could be adjusted to
fatty acids, polymeric surfactants, and particulate enable their complete rejection, with a permeate
surfactants, and they are produced by a great ux up to 200 L.h1.m2.bar1. Hong
variety of microorganisms, either secreted extra- et al. (1998) also reported the ultraltration of
cellularly or attached to parts of cells (Desai and copper, zinc, cadmium, and nickel ions using a
Banat 1997). The biosurfactants have several polycarboxylic acid-type biosurfactant. The
advantages over chemically synthesized surfac- author used at-sheet cellulose acetate mem-
tants such as lower toxicity, higher biodegradabil- branes with molecular weight cutoff of 1,000
ity, and effectiveness at extreme temperatures or and 3,000, but the biosurfactant was not as
Biosurfactant Production 249
effective as the rhamnolipid reported by El with biosurfactant rhamnolipids: performance for fry-
Zeftawy and Mulligan (2011), leading to lower ing oil degradation and membrane fouling reduction.
Bioresour Technol 126:314320
rejection values. Ramani K, Chandan Jain S, Mandal AB, Sekaran G (2012)
Another interesting application of Microbial induced lipoprotein biosurfactant from
biosurfactant was reported by Qin et al. (2012). slaughterhouse lipid waste and its application to the B
These authors used rhamnolipid biosurfactants to removal of metal ions from aqueous solution. Colloid
Surf B 97:254263
enhance the frying oil degradation and to reduce Vaz DA, Gudia EJ, Alameda EJ, Teixeira JA, Rodrigues
the membrane fouling in a submerged membrane LR (2012) Performance of a biosurfactant produced by
bioreactor, and an increase from 66 % up to 91 % a Bacillus subtilis strain isolated from crude oil samples
in the oil removal efciency was reported. The as compared to commercial chemical surfactants.
Colloid Surf B 89:167174
antifouling property of the biosurfactant was also
conrmed.
Besides, the biosurfactants could be used to
replace their chemically synthesized counterparts
in several other membrane processes, like the Biosurfactant Production
micellar enhanced ultraltration of aromatic alco-
hols and in the emulsion liquid membranes, used Frederico de Araujo Kronemberger
to remove and/or purify metal ions, dyes, ligno- COPPE Chemical Engineering Program,
sulfonate, and lactic acid. Federal University of Rio de Janeiro (UFRJ), Rio
de Janeiro, RJ, Brazil
References
A biosurfactant can be dened as a surfactant
Banat IM (1995) Biosurfactants production and possible compound produced by microorganisms. The
uses in microbial enhanced oil recovery and oil pollu- biosurfactants have several advantages over
tion remediation: a review. Bioresour Technol 51:112
chemically synthesized surfactants such as lower
Damasceno FRC, Cammarota MC, Freire DMG
(2012) The combined use of a biosurfactant and an toxicity, higher biodegradability, and effective-
enzyme preparation to treat an efuent with a high fat ness at extreme temperatures or pH values,
content. Colloid Surf B 95:241246 besides presenting high-surface activity and
Desai JD, Banat IM (1997) Microbial production of sur-
low-critical micelle concentration values emerg-
factants and their commercial potential. Microbiol Mol
Biol R 61:4764 ing as promising substitutes of the latter (Vaz
El Zeftawy MAM, Mulligan CN (2011) Use of et al. 2012). In the present moment, the
rhamnolipid to remove heavy metals from wastewater biosurfactants still present high production costs
by micellar-enhanced ultraltration (MEUF). Sep Purif
Technol 77:120127
in comparison to the chemically synthesized sur-
Freire DMG, Araujo LV, Kronemberger FA, Nitschke factants. That is mainly the result of the low pro-
M (2009) Biosurfactants as emerging additives in ductivity of the microbial strains and the
food processing. In: Passos ML, Ribeiro CP (eds) Inno- inefcient methodology of the bioprocessing.
vation in food engineering: new techniques and prod-
The technological improvement of the production
ucts, Contemporary Food Engineering Series. CRC
Press, Boca Raton, pp 685705 process is essential (Kronemberger et al. 2008). In
Hong J, Yang S, Lee C, Choi Y, Kajiuchi T (1998) Ultra- order to decrease the production costs, the scale-
ltration of divalent metal cations from aqueous solu- up of the whole production process, including
tion using polycarboxylic acid type biosurfactant.
J Colloid Interf Sci 202:6373
upstream and downstream, should be developed.
Mukherjee AK, Das K (2010) Microbial surfactants and One of the bottlenecks of the biosurfactant
their potential applications: an overview. In: Sen R production relies on the fact that most of them
(ed) Biosurfactants, vol 672, Advances in experimental are obtained through aerobic bioreactions. The
medicine and biology series. Springer Science+Busi-
ness Media, LLC, New York, pp 5464
use of the conventional submerged aeration can
Qin L, Zhang G, Meng Q, Zhang H, Xu L, Lv B (2012) lead to the formation of very stable foams, causing
Enhanced submerged membrane bioreactor combined serious operational problems. In order to
250 Biosurfactant Production
overcome that difculty, a nondispersive oxygen- biosurfactants, with 200 L of useful volume
ation process using membrane contactors can be was also described (Kronemberger et al. 2012).
applied (Kronemberger et al. 2008, 2012; Gruber It comprises microltration modules for fresh
and Chmiel 1991). Kronemberger et al. (2008) medium sterilization, nondispersive oxygenation,
investigated the rhamnolipid-type biosurfactant another set of microltration modules with self-
production in bioreactors with a nondispersive backwashing for cell retention, and a reverse
oxygenation device, obtaining productivities osmosis unit used to concentrate the product and
higher to the ones observed in shake asks. This to recover the water as the permeate stream,
system was then successfully used in a fed-batch enabling its reuse and minimizing the efuents.
experiment, in order to assess the potential of a
long-term production (Kronemberger et al. 2010).
Coutte et al. (2010) reported a similar system for
the production of lipopeptide biosurfactants, com-
References
paring internal and external nondispersive
oxygenation. Chen H, Chen Y, Juang R (2008a) Flux decline and mem-
Several authors have been investigating the brane cleaning in cross-ow ultraltration of treated
recovery of biosurfactants, mainly surfactin, fermentation broths for surfactin recovery. Sep Purif
Technol 62:4755
using ultraltration. Chen et al. (2008a) reported
Chen H, Chen Y, Juang R (2008b) Recovery of surfactin
the ux decline in the ultraltration of surfactin from fermentation broths by a hybrid salting-out and
using cellulose ester and polyethersulfone mem- membrane ltration process. Sep Purif Technol
branes with 100,000 Da of molecular weight cut- 59:244252
Coutte F, Lecouturier D, Yahia SA, Leclre V, Bchet M,
off, and the latter was recommended as the best
Jacques P, Dhulster P (2010) Production of surfactin
one for this kind of experiment, even though the and fengycin by Bacillus subtilis in a bubbleless
biosurfactant recovery was a little lower. Isa membrane bioreactor. Appl Microbiol Biotechnol
et al. (2007) investigated a two-step ultraltration 87:499507
Coutte F, Lecouturier D, Leclre V, Bchet M, Jacques P,
recovery system for surfactin. In the rst step
Dhulster P (2013) New integrated bioprocess for the
ultraltration, surfactin was retained by continuous production, extraction and purication of
polyethersulfone or regenerated cellulose mem- lipopeptides produced by Bacillus subtilis in membrane
branes at above its critical micelle concentration. bioreactor. Process Biochem 48:2532
Gruber T, Chmiel H (1991) Aerobic production of
In the second step, with the same kind of mem-
biosurfactants avoiding foam problems. In: Reuss M,
branes, after the disruption of the micelles by the Knackmuss HJ, Chmiel H, Gilles ED (eds) Biochemi-
addition of methanol, the puried surfactin was cal Engineering-Stuttgart. Fischer Verlag, Stuttgart, pp.
recovered in the permeate. Chen et al. (2008b) 212215
Isa MHM, Coraglia DE, Frazier RA, Jauregi P (2007)
also reported the surfactin recovery, but using
Recovery and purication of surfactin from fermenta-
ammonium sulfate salting out, ultraltration, tion broth by a two-step ultraltration process. J Membr
nanoltration, and their hybrid process. The com- Sci 296:5157
bination of salting out and ultraltration was Kronemberger FA, Santa Anna LMM, Fernandes ACLB,
Menezes RR, Borges CP, Freire DMG (2008) Oxygen-
selected due to the reduction of the fouling in the
controlled biosurfactant production in a bench scale
membranes used. bioreactor. Appl Biochem Biotechnol 147:3345
Another point of view should be the whole Kronemberger FA, Borges CP, Freire DMG (2010)
integrated production system supported by mem- Fed-batch biosurfactant production in a bioreactor. Int
Rev Chem Eng 2:513518
brane processes. A system described by Coutte
Kronemberger FA, Freire DMG, Castro AM, Santa Anna
et al. (2013) comprises the nondispersive oxygen- LMM, Borges CP (2012) System for obtaining biolog-
ation of a bioreactor for the production of surfactin ical products. Patent WO2012/079138 A1
and the continuous cell removal and product Vaz DA, Gudia EJ, Alameda EJ, Teixeira JA, Rodrigues
LR (2012) Performance of a biosurfactant produced by
separation using microltration and ultraltra-
a Bacillus subtilis strain isolated from crude oil samples
tion modules, respectively. A pilot scale system, as compared to commercial chemical surfactants. Col-
designed for the production of rhamnolipid-type loids Surf B 89:167174
Biphasic Biocatalytic Membrane Reactors 251
case where the biocatalyst is enantiospecic and

Biotechnology Processes Coupled converts only one of the substrate isomers giving
with Membrane Separation in one step the production and separation of enan-
tiomers in the optically pure form. The most stud-
Membrane Biotechnology ied enzyme in this systems are lipases. They are B
phase transfer catalysts activated by interface.
When lipase is randomly loaded within the mem-
brane and the two phases contact at a certain level
Biphasic Biocatalytic Membrane of the porous thickness, only part of the lipase is
Reactors located at the interface, whilst most of it is either
in the water or in the organic phase (Fig. 2a). This
Lidietta Giorno results in an apparent lower activity of the
Institute on Membrane Technology, National immobilized enzyme.
Research Council of Italy, ITM-CNR, Rende, An advanced development of multiphase reac-
Italy tor systems (emulsion enzyme membrane reactor)
has been reported where the organic/water inter-
face within the pores at the enzyme level is
Synonyms achieved by stable oil-in-water emulsion, pre-
pared by membrane emulsication, so that each
Multiphase enzyme membrane reactors; Two- pore can work as a microreactor containing
separate phase biocatalytic membrane reactors immobilized enzyme. In this way, the enzyme
works in the membrane pores in the same condi-
A schematic representation of a biphasic mem- tions as in the stirred tank reactor, but with no
brane is reported in Fig. 1. The system is suitable shear stress due to stirring. Various congurations
for bioconversions of low water soluble sub- can be obtained. In one case the enzyme is
strates. The enzyme loaded-membrane separates immobilized within the porous matrix and the
two immiscible phases: the substrate is present in emulsion is fed to the membrane and pressed to
the organic phase whilst the product is extracted in pass through the enzyme-loaded membrane
the aqueous phase. Particularly interesting is the (Giorno et al. 2003). As long as the reaction rate
Recemic substrate Membrane Product
COOCH3
COOH
CH3O
CH3O
(S)-naproxen methyl ester (S)-naproxen
Lipase
H2O
COOCH3
COOH
CH3O
CH3O
(R)-naproxen methyl ester (R)-naproxen
Biphasic Biocatalytic Membrane Reactors, Fig. 1 Biphasic membrane reactor for phase transfer catalysis
252 Biphasic Biocatalytic Membrane Reactors
Biphasic Biocatalytic Membrane Reactors, Fig. 2 (b) Schematic representation of the porous membrane
(a) Schematic representation of the organic/aqueous inter- thickness loaded with oil/water emulsion containing lipase
face at a certain level within the porous membrane thick- at the interface. Here the catalytic thickness is much larger
ness, with lipase at the interface in the open conformation
Biphasic Biocatalytic Membrane Reactors, Table 1 Performance of enzyme in multiphase biocatalytic membrane
reactor with different membrane materials
Immob. Enantio-
Membr. Immob. enzyme (S)-naproxen Enantio- selectivity
Type Membr.cut-off (kDa) site (mg/cm2) production (mmol) excess (%) ()
PA 10 Sponge layer 0.3 21 90 20
PA 10 Thin layer 0.08 6 35 2.2
PA 10 Thin layer 0.11 10 64 4.6
PA 50 Sponge layer 0.61 34 91 21
PA 50 Thin layer 0.19 3.5 91 21
PS 30 Sponge layer 1 38 81 10
PS 30 Thin layer 0.17 11 63 4.2
PP 0.2 m Surface 0.14 21 94 25
balances the convective transport rate there is no A comparison between performance of differ-
signicant inuence of fouling. Emulsion enzyme ent membrane materials used in multiphase
membrane reactors can also be prepared by enzyme membrane reactors is reported in Table 1.
immobilizing the emulsion and the enzyme within
the porous structure of the membrane thickness,
so that to have e two separate phase enzyme- References
emulsion loaded membrane, as illustrated in
Fig. 2b (Giorno et al. 2007). This latter congu- Giorno L, Li N, Drioli E (2003) Use of stable emulsion to
improve stability, activity and enantioselectivity of
ration guarantees an high enantioselectivity of lipase immobilized in a membrane reactor. Biotechnol
phase transfer catalysts activated by interface, Bioeng 84(6):677685
thanks to the fact that all immobilized enzyme is Giorno L, DAmore E, Mazzei R, Piacentini E, Zhang J,
located at the oil/water interface. Drioli E, Cassano R, Picci N (2007) An innovative
approach to improve the performance of a two separate
These conguration improved the selectivity
phase enzyme membrane reactor by immobilizing
and productivity of the biocatalytic system as lipase in presence of emulsion. J Membr Sci
well as its catalytic stability. 295:95101
Blending Modification of Membranes 253
discontinuity are frequent phenomena, which take

Blending Modification of Membranes place in solution or during fabrication of the com-
posite (Gugliuzza et al. 2007; De Lorenzo
Annarosa Gugliuzza et al. 2012). These events occur mainly when the
Institute on Membrane Technology, National mixed compounds do not exhibit compatibility B
Research Council of Italy, ITM-CNR, Rende, and/or miscibility at all. In this respect, the glass
Italy transition temperature (Tg) could be a powerful
tool for examining the degree of miscibility of
binary blends, e.g., mixed polymers. When one
When using the word blend, we mean a mixture Tg is obtained, the blend is considered fully mis-
of two or more components with low or high cible. If two Tg values are measured and a depen-
molecular weight, organic or inorganic nature, dence on the composition is estimated, the blend
and solid or liquid state. Physical or chemical is partially miscible. Finally, immiscible blends
blends with various degree of complexity can be show distinct and independent Tg values, as illus-
produced depending on the nature of the mixed trated by Fig. 2.
materials. Generally, in a blend we distinguish A set of equations can be useful to predict the
between a continuous lm (hosting matrix) and a x-dependence of Tg, where x is the composition.
dispersed phase (ller) consisting of polymers, The Fox equation (1956) is the simplest one and is
oligomers, nanoparticles, nanotubes, nanowires, traditionally used for binary systems, thereby giv-
zeolites, and so many others (Fig. 1). ing indication based on the properties of pure
The blending is used for making new formula- components:
tions with superior performance, when compo-
nents with complementary functions are well 1 x1 x2
(1)
interlaced and uniformly distributed. Major con- T g T g1 T g2
cerns are compatibility, uniformity, adhesion, and
minimum loading for maximum surface area where Tg is related to the blend, whereas Tgi and xi
(Gugliuzza 2013; Gugliuzza et al. 2006). Segre- are glass transition temperature and composition
gation, nano-connement, phase separation, and of the i-esimo pure component, respectively.
Blending Modification of Membranes, Fig. 1 Examples of miscible and immiscible blends

254 Blending Modification of Membranes
Blending Modification
of Membranes, Fig. 2 A
schematic representation of
the dependence of Tg on
composition in binary
polymer blends
The Gordon and Taylor equation (1952) regarded as the result of the buoyancy of van der
includes an additional experimental parameter Waals forces and/or hydrogen bonds established
(kGT) that considers unequal contributions of the at their interface (De Luca et al. 2009). Chemical
components: functionalization and/or compatibilizing agents
are two viable routes for strengthening these
x1 T g1 kGT x2 T g2 reciprocal interactions and dispersing the phases
Tg (2)
x1 kGT x2 in continuous volumetric spaces. This needs to
promote desired events in every region of the
Two other more complex relations are the Kwei blend.
(1984) and Brostow (2008), Eqs. 3 and 4, which In membranes, the blend is a common practice
take in account additional parameters such as KKW , to reinforce or produce specic properties, includ-
q, and ai upon complexity of the polymer blend ing transport, wetting, pores and mechanics, and
and/or copolymers: so many others. Transport properties can be mod-
ied depending on the phase behavior and related
x1 T g1 kKW x2 T g2 morphology. Permeability, selectivity, or conduc-
Tg qx1 x2 (3)
x1 kKW x2 tivity can be varied with respect to the upper
bound relationship as a function of the orientation
of the dispersed phase and/or the nature of the
T g x1 T g1 x2 T g2 x1 x2 : a0 a1 2x1 1

interface generated between the various compo-
nents (Robeson 2010). Indeed, the molecular sep-
a2 2x1 12 a3 2x1 13
aration can be addressed via diffusion or sorption
(4) events dependently on free volume distribution
and/or binding sites density.
In all equations the index 2 is referred to the Similarly, the blending can give new solutions
component with the highest Tg, whereas x2 is 1-x1. to dry-wet membranes; changes in supramolecu-
A good blend depends on the degree of inter- lar chemistry can modulate the wettability, mak-
action between the hosting matrix and the external ing the membrane usable for numerous times
surface area of the dispersed phase and can be without damages and/or special treatments.
Block Copolymer Membrane 255
This approach is also preferred when pores factor for water sorption into polymer membranes.
must be generated through polymer networks. J Phys Chem B 111(30):88688878
Kwei TK (1984) The effect of hydrogen bonding on the
Pore-forming agents are often blended with poly- glass transition temperatures of polymer mixtures.
mers and, then, removed by washing treatments. J Polym Sci Lett 22(6):307313
This takes the advantage of generating empty Robeson LM (2010) Polymer blends in membrane trans- B
spaces inside dense polymer networks, but the port processes. Ind Eng Chem Res 49:1185911865
major difculty is often the complete removal of
the ller from the matrix. In this respect, a good
compatibility between pore-forming agent and
polymer matrix is necessary to produce a large
number of uniform and regular pores through the Block Copolymer Membrane
network.
Finally, the blend is often preferred to provide Mario Malinconico
composites with mechanical, chemical, and ther- Research Director, Institute for Polymers,
Composites and Biomaterials (IPCB-CNR),
mal properties beyond those observable for the
single elements. Enhanced tensile properties Pozzuoli, Naples, Italy
along with reinforced resistance to high tempera-
tures and harsh environments can be obtained by
blending organic and inorganic materials in dif- Block Copolymers
ferent ratios.
Block copolymers are polymeric materials in
which macromolecules of two or more different
References homopolymers (blocks) are chemically bonded
together to form complex macromolecules with
Brostow W, Chiu R, Kalogeras IM, Vassilikou-Dova linear (di-, tri-, multiblock copolymers) or
A (2008) Prediction of glass transition temperatures: nonlinear (mixed arm star, block, or graft copoly-
binary blends and copolymers. Mater Lett mers) architecture (Fig. 1).
62:31523155
De Lorenzo L, Tocci E, Gugliuzza A, Drioli E (2012) Pure
The routes of synthesis of block copolymers
and modied co-poly (amide-12-b-ethylene oxide) are two:
membranes for gas separation studied by molecular
investigations. Membranes 2(3):346366 1. Creating centers or active sites (radical,
De Luca G, Gugliuzza A, Drioli E (2009) Competitive
hydrogen-bonding interactions in modied polymer
anionic, cationic) of a polymer chain that can
membranes: a density functional theory investigation. subsequently trigger the polymerization of a
J Phys Chem B 113(16):54735477 second monomer. If the location of the terminal
Fox TG (1956) Inuence of diluent and of copolymer active center is not specied, this denition
composition on the glass temperature of a polymer
also includes graft copolymers.
system. Bull Am Phys Soc 1:123125
Gordon M, Taylor JS (1952) Ideal copolymers and the 2. Condensation between functional groups
second-order transitions of synthetic rubbers. located at the end of the polymer or
1. non-crystalline copolymers. J Appl Chem prepolymer:
2(9):493500
Gugliuzza A (2013) Nanocomposite membranes for mem- Am X Y Bn ! Bn XY Am
brane reactors. In: Basile A (ed) Handbook of mem-
brane reactors volume I: fundamental materials science,
design and optimisation. Woodhead Publishing,
If the species are bifunctional, multiblock
Cambridge copolymers can be obtained. Despite the general
Gugliuzza A, Fabiano R, Garavaglia MG, Spisso A, Drioli indications just described, the interest for the pro-
E (2006) Study of the surface character as responsible duction of block copolymers is focused on anionic
for controlling interfacial forces at membranefeed
interface. J Coll Interf Sci 303(2):388403
initiators and polymerization processes having
Gugliuzza A, De Luca G, Tocci E, De Lorenzo L, Drioli living character. This way, derived from the orig-
E (2007) Intermolecular interactions as controlling inal studies of M. Szwarc (1956), provides
256 Block Copolymer Membrane
ABC triblock Mixed arm

AB diblock
star block
AB
ABA triblock
(AB)n star (AB)n multiblock
Block Copolymer Membrane, Fig. 1 Linear and nonlinear architecture of block copolymers
copolymer sequences well dened, with desired new monomer will inevitably make some living
molecular weights, structures, and compositions. chains dead because of impurities, leading to the
All this is achieved with diene monomers, vinyl resultant block copolymer with lower block num-
(nonpolar), esters, ethers, or cyclic suldes. bers and broad molecular weight distribution.
The key technologies for producing block Therefore, the sequential addition method can
copolymers with anionic initiators can be indi- only be used to make copolymers with a few
cated schematically in the following way: blocks, such as triblock copolymers. Another lim-
itation of the sequential addition of different
Sequential polymerization of the monomers comonomers is their reaction compatibility: each
(especially in the case of di-blocks) added monomer must be sufciently reactive so
Polymerization of the second monomer with that the chain can propagate. Often, a living
the functionalized prepolymers of the rst A block can initiate comonomer B, but a living
monomer B block cannot initiate comonomer A.
Use of reactions of coupling between reactive In the second approach, multiblock copoly-
terminals and a compound reagent that mers could be in principle prepared by coupling
becomes the junction site: (Yagci and Tasdelen 2006) different polymer
chain di-ends terminated with suitable reactive
2Am Bn ! Am Bn X Bn Am groups or by linking (Yagci and Tasdelen 2006;
Rahman et al. 2007) di-end-functionalized poly-
Use of bifunctional initiators mer chains using so-called difunctional linking
agents in solution. However, both coupling and
The synthesis of multiblock copolymers can be linking reactions are extremely ineffective when
obtained by an alternately sequential addition of long polymer chains (Mw >5 103) are used as
different comonomers into a living copolymeriza- the precursors because most of the reactive end
tion system or, alternatively, by the coupling of groups are segregated inside the polymer chains
difunctional homopolymeric precursors. coiled in good solvent. Moreover, for long poly-
In the rst case, anionic polymerization or, mer chains, the concentration of the reactive end
more recently, living radical polymerization groups is too low to have an effective coupling or
methods can be in principle used. In real cases linking reaction because the overall polymer con-
(Ekizoglou and Hadjichristidis 2002; Buzdugan centration cannot be too high. Therefore, the
et al. 2000; Reuter et al. 1991), each addition of a essential problem is how to expose and
Block Copolymer Membrane 257
concentrate the reactive end groups of long poly- structures are formed in which the contact area
mer chains, while the overall polymer concentra- between the incompatible microdomains is min-
tion will not be increased. imized. The minimization of the interfacial area
between the domains involves a reduction in
interfacial energy for which even if from an B
Block Copolymers in Membranes entropic point of view the individual polymer
chains prefer a conformation random coil, in
Copolymers, and in particular block copolymers the case of block copolymers, macromolecules
(BCPs), in recent years have aroused great interest tend to assume more extended conformations at
in the eld of specialty polymer applications, the interface between the two blocks so as to
including nanotechnologies (Park et al. 2003; allow similar blocks to organize into microstruc-
Krausch and Magerle 2002; Hamley 2003; tures that minimize the ratio (surface exposed/
Lazzari and Lopez-Quintela 2003; Lu and Sastry volume).
2007; Hawker and Russell 2005; Li et al. 2005; The frequency and size of these domains
Segalman 2005; Cheng et al. 2006; Stoykovich depend on the lengths of the regular blocks and,
and Nealey 2006; Krishnamoorthy et al. 2006; Li therefore, from the molecular masses.
and Ober 2006; Fasolka and Mayes 2001; Darling Block copolymers nd many interesting appli-
2007; Kim and Hinsberg 2008; Bates and cations in membrane technology.
Fredrickson 1999; Abetz and Simon 2005; van Block copolymers have found application in a
Zoelen and ten Brinke 2009), given their ability class of rubbery membranes, having the unique
to self-organize into structures with periodicity ability to separate CO2 from gas streams, with a
down to the nanometer scale. consequent increase in performance compared to
If the different polymer blocks are chemically that of the analogous homopolymers (Guan and
incompatible, phase separation occurs with spon- De Simone 1994).
taneous segregation of different macromolecules The polyethylene oxide [PEO] is, for example,
in different microdomains (Fig. 2). Microdomains one of the most interesting materials for the sep-
formed are not arranged in a random but show a aration of CO2, especially when cross-linked with
regular arrangement which involves the formation other polymers or incorporated into block copol-
of periodic structures. ymers (Lin et al. 2006; Pethe et al. 2008). Because
This phenomenon, called self-assembly, of its amorphous character and of its exibility, it
determines the spontaneous formation of suitably directs the permselectivity of CO2
nanostructures. It is associated with the com- through the polymer matrices, especially when
petition between the tendency to separation of the length of the segment, the weight, and the
phase, due to the incompatibility of the blocks, molecular ratio with the other blocks of the copol-
and chemical connectivity that prevents it; periodic ymer complexes are changed.
Block Copolymer
Membrane,
Fig. 2 Schematic of self- A
assembly in BCPs due to
phase separation
block copolymer
phase-separated
block copolymer
258 Block Copolymer Membrane
Bondar et al. (2000) observed that the perme- Guan Z, De Simone JM (1994) Fluorocarbon-based
ability of CO2 increases with the amount of Heterophase Polymeric Materials: Block Copolymer
Surfactants for Carbon Dioxide Applications. Macro-
polyether [PEO or PTMEO] when segmented molecules 27:55275532
block copolymers are incorporated into the Hamley IW (2003) Nanotechnology with soft materials.
polyether-b-polyamide, where they form the seg- Angew Chem Int Ed 42:1692
regated phase of nylon-6 [PA6] or nylon-12 Hawker CJ, Russell TP (2005) BCP lithography: merging
bottom-up with top-down processes. MRS Bull
[PA12]. In particular, they have observed that the 30:952
copolymers 57PEO/PA6 and 55PEO/PA12 Kim HC, Hinsberg WDJ (2008) Surface patterns from block
exhibit a better selectivity to pure gas CO2/N2 copolymer self-assembly. J Vac Sci Technol A 26:1369
56 (PCO2 66 barrer) and CO2/H2 9.8 (PCO2 Krausch G, Magerle R (2002) Nanostructured thin lms by
self-assembly of block copolymers. Adv Mater 14:1579
20 barrer), respectively, compared to more con- Krishnamoorthy S, Hinderling C, Hcinzelmann H (2006)
ventional polymers rubbery and glassy. Nanoscale patterning with block copolymers. Mater
Higher performance was observed when more Today 9:40
than one unit polar (PEO and PA6) were included Lazzari M, Lopez-Quintela MA (2003) Block copolymers as
a tool for nanomaterial fabrication. Adv Mater 15:1583
in the polymer chain, suggesting the critical role Li MQ, Ober CK (2006) Block copolymer patterns and
of interesting polar intermolecular interactions templates. Mater Today 9:30
between the polymer chain and the penetrating Li MQ, Coenjarts CA, Ober CK (2005) Patternable Block
molecules. Membranes PEO/PBT have been ad Copolymers. Adv Polym Sci 190:183
Lin H, Van Wagner E, Swinnea JS, Freeman BD, Pas SJ,
hoc designed in order to obtain block copolymers Hill AJ, Kalakkunnath S, Kalika DS (2006) Transport
with high performance for the separation of CO2 and Structural Characteristics of Crosslinked Poly(eth-
(Zhao et al. 2008). Separation factors for the pairs ylene oxide) Rubbers. J Membr Sci 276:145161
of gases, such as CO2/H2, CO2/N2, and CO2/CH4, Lu W, Sastry AM (2007) Self-assembly for semiconductor
industry. IEEE Trans Semicond Manuf 20:421
were obtained in different ratios of the polymer Park C, Yoon J, Thomas EL (2003) Enabling Nanotech-
chain. When mixed with the polyether glycol nology with Self Assembled Block Copolymer Pat-
[PEG], a signicant improvement of the perme- terns. Polymer 44:6725
ability and selectivity for various block copoly- Pethe VV, Wang HP, Hiltner A, Baer E, Freeman BD
(2008) Oxygen and carbon dioxide permeability of
mers was observed. EAA/PEO blends and microlayers. J Appl Polym Sci
110:14111419
Rahman MS, Sama LS, Lee JS (2007) Quantitative In
References Situ Coupling of Living Diblock Copolymers for the
Preparation of Amphiphilic Coil-Rod-Coil Triblock
Abetz V, Simon PF (2005) Phase Behaviour and Morphol- Copolymer Poly(2-vinylpyridine)-b-poly(n-hexyl iso-
ogies of Block Copolymers. Adv Polym Sci 189:125 cyanate)-b-poly(2-vinylpyridine). Macromolecules
Bates FS, Fredrickson GH (1999) Block Copolymers- 40(26):92799283
Designer Soft Materials. Phys Today 52:32 Reuter H, Berlinova IV, Horing S, Ulbricht J (1991) The
Bondar VI, Freeman BD, Pinnau I (2000) Gas transport anionic block copolymerization of ethylene oxide with
properties of poly(ether-b-amide) segmented block copol- tert-butyl methacrylate. Diblock and multiblock copol-
ymers. J Polym Sci Part B Polym Phys 38:20512062 ymers. Eur Polym J 27(7):673680
Buzdugan E, Ghioca P, Stribeck N (2000) Synthesis and Segalman RA (2005) Patterning with block copolymer thin
properties of styrene-isoprene multiblock copolymers. lms. Mater Sci Eng R Rep 48:191
Mater Plast 37(3):138144 Stoykovich MP, Nealey PF (2006) Block copolymers and
Cheng JY, Ross CA, Smith HI, Thomas EL (2006) conventional lithography. Mater Today 9:20
Templated self-assembly of block copolymers: top- Szwarc M (1956) Living Polymers. Nature 178:1168
down helps bottom-up. Adv Mater 18:2505 van Zoelen W, ten Brinke G (2009) Thin lms of
Darling SB (2007) Directing the self-assembly of block complexed block copolymers. Soft Matter 5:1568
copolymers. Prog Polym Sci 32:1152 Yagci Y, Tasdelen MA (2006) Mechanistic transformations
Ekizoglou N, Hadjichristidis NJ (2002) Synthesis of model involving living and controlled/living polymerization
linear tetrablock quaterpolymers and pentablock methods. Prog Polym Sci 31(12):11131170
quintopolymers of ethylene oxide. Polym Sci Part Zhao Y, Cao Y M, Ding X L, Zhou MQ, Liu J H and Yuan
A Polym Chem 40(13):21662170 Q (2008) Poly(ethylene oxide) induced cross-linking
Fasolka MJ, Mayes AM (2001) Block copolymer thin lms: modication of Matrimid membranes for selective sep-
physics and applications. Annu Rev Mater Res 31:323 aration of CO2. J Membrane Sci 320, 179184
Blood Cell Origins 259
engraftment time and reduce the graft failure rate,

Blood Cell Origins particularly in developing this therapy for adult
patients (Higuchi et al. 2010). HSCs cultured in
A. Higuchi 2D materials and in porous membranes can be
Department of Chemical and Materials possible to expand ex vivo. B
Engineering, National Central University, Ex vivo expansion of HSCs is a hot topic due
Chung-Li, Taoyuan, Taiwan to the shortage of HSCs in clinical application.
HSCs can be dened by surface marker of cells
such as CD (cluster of differentiation). The cells
Blood cell origins are from the hematopoietic of CD34+CD45 or CD34+CD38 are generally
stem cells (HSCs). HSCs can generate all types of believed as HSCs, which are measured from ow
the blood cells such as the erythrocytes (red blood cytometry where + indicates the expressing
cells), lymphocytes (T cells, B cells, and NK cells and indicates non-expressing cells.
(natural killer) cells), monocytes, macrophages, CD133+CD34 cells are believed as more pro-
neutrophils, basophils, eosinophils, megakaryo- genitor HSCs than CD34+CD38 cells.
cytes/platelets, and dendritic cells. HSCs are Efcient cell separation is an important issue for
found in the bone marrow of adults, which includes the successful isolation and purication of HSCs.
the femurs, hip, ribs, sternum, and other bones. Centrifugation, afnity column chromatography,
Cells can be obtained from the iliac crest part of uorescence-activated cell sorting (FACS), mag-
the pelvic bone, using a special needle and a netic cell selection (MACS), and membrane ltra-
syringe; the cells removed are of two forms, i.e., tion are techniques typically employed for cell
as a smear and as a core biopsy. HSCs can be also separation (Higuchi et al. 2009). Highly puried
found in the umbilical cord blood (UCB, 1575/ cellular preparations are obtained using
ml) and peripheral blood (1/ml). The low number FicollPaque method followed by FACS or
of HSCs and small volume (50150 ml) obtainable MACS in conjunction with a uorescently labeled
from a single donor of UCB limit direct transplan- antibody as the cell surface marker, while the cell
tation of UCB to the treatment of pediatric patients. separation by membrane ltration method is less
A density of 1.7 107 CD34+ cells (HSCs)/kg is specic, but is simple and easy operation, and ste-
necessary for the transplantation of UCB into rility can maintain during the process. Only half an
patients (Higuchi et al. 2010). Therefore, UCB hour is necessary to purify HSCs from the blood by
transplantation has been limited to children with the membrane ltration method, while the conven-
an average weight of 20 kg. Numerous efforts have tional FicollPaque followed by MACS method
been made to expand HSCs ex vivo to improve needs around 8 h (Higuchi et al. 2009) (Fig. 1).
Blood Cell Origins,

Fig. 1 HSCs are blood cell
origins, which are
multipotent stem cells
giving rise to all of the
blood cell types
260 Blood Filtration
References done by applying a negative pressure to the dial-

ysate compartment of the dialyzer. This pressure
Higuchi A, Yang ST, Li PT, Ruaan RC, Chen WY, gradient causes water and dissolved solutes to
Chang Y, Chang Y, Tsai EM, Chen YH, Wang HC,
move from blood to dialysate and allows the
Hsu ST, Ling QD (2009) Permeation of blood cells
from umbilical cord blood through surface-modied removal of several liters of excess uid during a
polyurethane foaming membranes. J Membr Sci typical 3- to 5-h treatment. Hemodialysis treat-
339:184188 ments are typically given in a dialysis center
Higuchi A, Yang ST, Li PT, Tamai M, Tagawa T, Chang Y,
three times per week.
Chang Y, Ling QD, Hsu ST (2010) Direct Ex Vivo
expansion of hematopoietic stem cells from umbilical Hemodialysis membranes are typically pre-
cord blood on membranes. J Membr Sci 351:104111 pared from cellulose materials or polysulfone-
polyvinylpyrrolidone (PVP)-blended materials.
Cellulose is hydrophilic and can be used as a
hemocompatible material, whereas polysulfone
Blood Filtration is one of engineering plastics and needs to add a
hydrophilic and hemocompatible material as a
A. Higuchi blending material. PVP shows relatively good
Department of Chemical and Materials hemocompatibility, and a more important fact is
Engineering, National Central University, that PVP can be blended well with polysulfone.
Chung-Li, Taoyuan, Taiwan PVP is also used as a porogen in dialysis mem-
branes of polysulfone. There is a recent demand
for hemodialysis membranes that remove the low
Blood ltration is performed by hemodialysis molecular weight proteins such as b2-myoglobin
membranes, leukocyte removal lter, and plasma (MW 11500) and endotoxin (subunit of
separation membranes. MW = 5,00020,000), and useful albumin in
Blood is composed of plasma solution and the plasma should be recovered by the mem-
blood cells. Several valuable proteins such as branes. Polysulfone hollow bers blended with
immunoglobulin, factor VIII, factor IX, albumin, PVP have been widely used as suitable hemodial-
vitronectin, etc. are contained in plasma solution. ysis membranes which satisfy this requirement
Blood cells are erythrocytes (RBC, red blood (Higuchi et al. 2002).
cells), lymphocytes (T cells, B cells, and NK Leukocyte removal lter: White blood cells
(natural killer) cells), monocytes, macrophages, (leukocytes) generate many adverse reactions dur-
neutrophils, basophils, eosinophils, megakaryo- ing blood-transfusion therapy, which are graft-
cytes/platelets and dendritic cells, and hematopoi- versus-host disease (GVHD), platelet refractori-
etic stem cells (HSCs). HSCs are found in the ness, nonhemolytic febrile transfusion reaction,
bone marrow of adults, which includes femurs, and infection of viruses, such as human
hip, ribs, sternum, and other bones. T-lymphotropic virus (HTLV), cytomegalovirus
In hemodialysis, the patients blood is pumped (CMV), and human immunodeciency virus
through the blood compartment of a dialyzer, (HIV) (Higuchi 2010). It was found that most of
exposing it to a dialysis membrane. The dialyzer the viruses infect specic type of leukocytes, such
is composed of thousands of tiny synthetic hollow as granulocytes, monocytes, lymphocytes,
bers. The ber wall acts as the semipermeable B lymphocytes, T-helper cell (CD4+ cell), and
membrane. Blood ows through the bers, dialy- suppressor/cytotoxic T cells (CD8+ cell)
sis solution ows around the outside of the bers, (Higuchi 2010). Therefore, removal of leukocytes
and water and wastes move between these two in RBC and platelet concentrates as well as whole
solutions. The cleansed blood is then returned blood components is essential to prevent the
via the circuit back to the body. Ultraltration adverse effect of contaminated leukocytes. Leu-
occurs by increasing the hydrostatic pressure kocytes can be removed using a lter comprised
across the dialyzer membrane. This usually is of nonwoven fabric or sponge materials as a lter
Blood Separation 261
medium. The mechanism of leukocyte removal on tissue composed of several blood cells in plasma.
the lters comprised of nonwoven fabric is based The cellular elements of red blood cells (RBCs),
on the adsorption of leukocytes, while that com- platelets, and white blood cells make up 45 % of
prised of sponge materials is based on the sieving the volume of whole blood. Another 55 % is
effect and adsorption. plasma, which contains 78 % of the plasma pro- B
Plasma separation membrane: Plasma- teins and 9293 % of water (Higuchi 2010).
exchange therapy has been increasingly applied Figure 1 shows typical blood after centrifugation
clinically over the past few years. Membrane with and without addition of Ficoll-Paque (Ficoll-
plasma separation has been used since 1979, Hypaque) solution. After centrifugation of blood
which is similar to hemodialysis and without addition of Ficoll-Paque (Ficoll-
hemoltration. Plasma separation from whole Hypaque) solution (native blood centrifugation),
blood is now performed routinely. blood can be separated into plasma layer, platelet
The materials of plasma separation membranes and leukocyte layer, and RBC layer (Fig. 1a). In
are typically made of nitrocellulose, polysulfone, this case, each layer contains other contaminant
and polypropylene. cells, e.g., RBC layer contains 96 % of RBCs, 3 %
of leukocytes, and 1 % of platelets in blood cells.
Platelets and leukocytes are also included in
Cross-References plasma layers. Platelet-rich plasma is necessary
to use for the evaluation of biocompatibility of
Blood Cell Origins biomaterials (Higuchi et al. 2003). Platelet-rich
Blood Separation plasma is obtained by centrifugation of peripheral
Blood Treatment Membrane blood or umbilical cord blood at 3,000 rpm.
Platelet-poor plasma is used for plasma protein
adsorption on biomaterials for the evaluation of
References
biocompatibility of the biomaterials (Higuchi
Higuchi A (2010) Separation and purication of stem and et al. 2003). Platelet-poor plasma is obtained by
blood cells by porous polymeric membranes. In: centrifugation of peripheral blood or umbilical
Driolli E, Giorno L (eds) Comprehensive membrane cord blood at a relatively high speed of
science and engineering. Elsevier, Cambridge 3,000 rpm. Mononuclear cells including hemato-
Higuchi A, Shirano K, Harashima M, Yoon BO, Hara M,
Hattori M, Imamura K (2002) Chemically modied poietic stem cells (HSCs) cannot easily be
polysulfone hollow bers with vinylpyrrolidone hav- obtained by centrifugation of native blood. There-
ing improved blood compatibility. Biomaterials fore, Ficoll-Paque (or Ficoll-Hypaque) solution
23:26592666 was injected into blood sample, and the mixed
solution was centrifuged at 400 g for
3040 min at 20 C (Fig. 1b). The upper layer
contains plasma and platelets. Mononuclear cells
Blood Separation including lymphocytes (T cells, B cells, and NK
(natural killer) cells) and HSCs can be isolated
A. Higuchi from the upper second layer (Fig. 1b). When
Department of Chemical and Materials specic cells such as HSCs, T cells, N cells, or
Engineering, National Central University, NK cells should be isolated, MACS or FACS are
Chung-Li, Taoyuan, Taiwan applied. Direct application of MACS and FACS to
isolate the specic blood cells is difcult due to
large quantity of RBCs in blood samples. After
Blood separation can be performed by centrifuga- the mononuclear cells were isolated by Ficoll-
tion, magnetic cell selection system (MACS), Paque method, residual RBCs were removed by
uorescence-activated cell sorting (FACS), and the treatment of lysing solution and then HSCs
membrane ltration method. Blood is a living (CD34+ cells) can be isolated by MACS or FACS
262 Blood Separation
Blood Separation,
a Blood contains
Fig. 1 Blood components
after centrifugation of 55% of plasma.
native blood (a) and before Plasma Concentration of plasma
and after centrifugation of protein is 7 - 8%
blood with Ficoll-Paque
solution (b) Platelets
Blood contains
Lymphocytes
45% of blood cells
(96% of red blood cell,
Red Blood cell 3% of leukocytes, and
1% of platelet)
b
Blood sample
Plasma, Platelets
Centrifugation
Lymphocytes
Ficoll-Paque
Ficoll-Paque
Red blood cells,
Granulocytes
Blood Separation, a b
Fig. 2 Schematic
mechanism of the
separation method of cells
by a magnetic cell selection
system
Magnet Magnet Magnet Magnet
treatment using antibody of CD34+ (Chen MACS. The magnetic beads attaching the mono-
et al. 2012). clonal antibody are separated by magnetic force to
MACS is a sophisticated cell separation collect the specic marked cells. The MACS
method, in which magnetic beads attaching a needs to use an expensive antibody conjugated
monoclonal antibody as the cell-surface marker with magnetic beads to bind to the target cells
are mixed with cells. Figure 2 shows the sche- for the detection of the cells. Both cell separation
matic mechanism of the separation method by an methods using FACS and MACS are not
Blood Treatment Membrane 263
applicable if the antibodies to the specic markers Driolli E, Giorno L (eds) Comprehensive membrane
on the surface of the target cells have not been science and engineering. Elsevier, Cambridge
Higuchi A, Sugiyama K, Yoon BO, Sakurai M, Hara M,
established. Sumita M, Sugahara S, Shirai T (2003) Serum protein
Blood cell separation through membrane ltra- and platelet adsorption on pluronic-coated polysulfone
tion was recently reported by several researchers membranes. Biomaterials 24:32353245 B
(Komai et al. 1998; Yasutake et al. 2001; Higuchi Higuchi A, Yamamiya S, Yoon BO, Sakurai M, Hara
M (2004) Peripheral blood cell separation through
et al. 2004, 2008). Typical blood cell separation surface-modied polyurethane membranes. J Biomed
membranes are leukocyte removal lter Mater Res Part A 68A:3442
(membrane) and HSC purication membranes. Higuchi A, Sekiya M, Gomei Y, Sakurai M, Chen WY,
HSC separation from peripheral blood and Egashira S, Matsuoka Y (2008) Separation of hemato-
poietic stem cells from human peripheral blood through
umbilical cord blood through surface-modied modied polyurethane foaming membranes. J Biomed
polyurethane membranes by membrane ltration Mater Res Part A 85A:853861
method was reported (Higuchi et al. 2004, 2008, Higuchi A, Yang ST, Li PT, Tamai M, Tagawa T, Chang Y,
2010). Peripheral blood or umbilical cord blood Chang Y, Ling QD, Hsu ST (2010) Direct ex vivo
expansion of hematopoietic stem cells from umbilical
was permeated though the surface-modied mem- cord blood on membranes. J Membr Sci 351:104111
branes by ltration. HSCs are more adhesive cells Komai H, Naito Y, Fujiwara K, Takagaki Y, Noguchi Y,
than RBCs, platelets, and lymphocytes. There- Nishimura Y (1998) The protective effect of a leuko-
fore, HSCs remained to adhere on the membranes cyte removal lter on the lung in open-heart surgery for
ventricular septal defect. Perfusion 13:2734
during permeation of blood. The membrane- Yasutake M, Sumita M, Terashima S, Tokushima Y,
adhering HSCs were rinsed with phosphate buffer Nitadori Y, Takahashi TA (2001) Stem cell collection
saline and subsequently human serum albumin or lter system for human placental/umbilical cord blood
dextran solution as surfactant solution was perme- processing. Vox Sang 80:101105
ated through the membranes. The HSCs can be
harvested in the recovery solution of human
serum albumin or dextran solution. The mem-
brane ltration method of blood separation should
be useful, because centrifugation instrument is not Blood Treatment Membrane
necessary to use and antibodies targeting specic
cells are not used in the method, which cause A. Higuchi
contamination of antibodies in the blood cell Department of Chemical and Materials
samples. Engineering, National Central University,
Chung-Li, Taoyuan, Taiwan
Cross-References
Blood treatment membrane is categorized as
Blood Cell Origins dialysis membrane, leucocyte removal lter, and
Blood Filtration plasma separation membrane.
Blood Treatment Membrane
Dialysis Membrane
References
Chen LY, Chang Y, Shiao JS, Ling QD, Chang Y, Chen

Dialysis is a process for removing waste and
YH, Chen DC, Hsu ST, Lee H, Higuchi A (2012) Effect excess water from the blood and is used primarily
of the surface density of nanosegments immobilized on to provide an articial replacement for lost kidney
culture dishes on ex vivo expansion of hematopoietic function in people with renal failure. Dialysis was
stem and progenitor cells from umbilical cord blood.
Acta Biomater 8:17491758
used for those with an acute disturbance in kidney
Higuchi A (2010) Separation and purication of stem and function (acute kidney injury), or progressive but
blood cells by porous polymeric membranes. In: chronically worsening kidney function.
264 Blood Treatment Membrane
Dialysis works on the principles of the diffu- biocompatibility. However, these biocompatible
sion of solutes and ultraltration of uid across a materials cannot be used as blending materials in
semipermeable membrane. Blood ows by one polysulfone dialysis membranes due to low
side of a semipermeable membrane, and a special mixing with polysulfone. PVP is also used as a
dialysis uid ows by the opposite side. porogen in dialysis membranes of polysulfone.
A semipermeable membrane (dialysis membrane, There is a recent demand for hemodialysis mem-
dialyzer) is a thin layer of material that contains branes that remove the low-molecular-weight pro-
the appropriate size of pores. Smaller solutes teins such as b2-myoglobin (MW 11,500) and
(urea, NaCl) and uid pass through the mem- endotoxin (subunit of MW = 5,00020,000) and
brane, but the membrane blocks the passage of useful albumin in the plasma should be recovered
larger substances such as red blood cells and large by the membranes. Polysulfone hollow bers
proteins (albumin, globulin). This is the ltering blended with PVP have been widely used as suit-
process taking place in the kidneys, when the able hemodialysis membranes which satisfy this
blood enters the kidneys and the larger substances requirement (Higuchi et al. 2002).
are separated from the smaller ones in the glomer-
ulus (Daugirdas et al. 2006).
In hemodialysis, the patients blood is pumped Leukocyte Removal Filter
through the blood compartment of a dialyzer,
exposing it to a dialysis membrane. The dialyzer White blood cells (leukocytes) generate many
is composed of thousands of tiny synthetic hollow adverse reactions during blood-transfusion ther-
bers. The ber wall acts as the semipermeable apy, which are graft-versus-host disease (GVHD),
membrane. Blood ows through the bers, dialy- platelet refractoriness, nonhemolytic febrile trans-
sis solution ows around the outside of the bers, fusion reaction, and infection of viruses, such as
and water and wastes move between these two human T-lymphotropic virus (HTLV), cytomega-
solutions (Daugirdas et al. 2006). The cleansed lovirus (CMV), and human immunodecient
blood is then returned via the circuit back to the virus (HIV) (Higuchi 2010). It was found that
body. Ultraltration occurs by increasing the most of the viruses infect specic type of leuko-
hydrostatic pressure across the dialyzer mem- cytes, such as granulocytes, monocytes, lympho-
brane. This usually is done by applying a negative cytes, lymphocytes-B, T helper cell (CD4+ cell),
pressure to the dialysate compartment of the dia- and T-cell suppressor/cytotoxic cells (CD8+ cell).
lyzer. This pressure gradient causes water and HTLV-1 and HIV mainly infect T helper cell,
dissolved solutes to move from blood to dialysate while CMV mainly infects granulocytes, mono-
and allows the removal of several liters of excess cytes, and lymphocytes. GVHD was mainly gen-
uid during a typical 3- to 5-h treatment. Hemo- erated by T helper cell and T-cell suppressor/
dialysis treatments are typically given in a dialysis cytotoxic cells (CD8+ cell) (Higuchi 2010).
center three times per week (Daugirdas Therefore, removal of leukocytes in RBC and
et al. 2006). platelet concentrates as well as whole blood com-
Hemodialysis membranes are typically pre- ponent are essential to prevent the adverse effect
pared from cellulose materials or polysulfone- of contaminated leukocytes. Leukocytes can be
polyvinylpyrrolidone (PVP)-blended materials. removed using a lter comprised of nonwoven
Cellulose is hydrophilic and can be used as a fabric or sponge materials as a lter medium.
hemocompatible material, whereas polysulfone The mechanism of leukocyte removal on the l-
is one of the engineering plastics and needs to ters comprised of nonwoven fabric is based on the
add a hydrophilic and hemocompatible material adsorption of leukocytes, while that comprised of
as a blending material. PVP shows relatively good sponge materials is based on the sieving effect and
hemocompatibility and the more important fact is adsorption.
that PVP can be blended well with polysulfone. Filtration methods have several advantages
There are a lot of materials reported excellent compared to other methods of removing
Blowdown Water 265
leukocytes such as centrifugation. Virus contami- Driolli E, Giorno L (eds) Comprehensive membrane
nation is lower in blood components during the science and engineering. Elsevier, Cambridge
Higuchi A, Shirano K, Harashima M, Yoon BO, Hara M,
process in the ltration method than in the centri- Hattori M, Imamura K (2002) Chemically modied
fugation method due to mild operation and the polysulfone hollow bers with vinylpyrrolidone hav-
ease of operation under sterilized conditions. ing improved blood compatibility. Biomaterials B
Leukocyte removal lters were typically made 23:26592666
of polyurethane (PU) foaming membranes where
the pore was made by salt leaching method and
nonwoven fabric. The pore structure of both lters
is found to be completely different, although the Blowdown Water
pore size of those lters was almost the same from
capillary ow porometer measurements. The Aamer Ali
mechanism of leukocyte removal (i.e., separation Department of Environmental and Chemical
of leukocyte from plasma and other blood cells) in Engineering, University of Calabria, Rende,
leukocyte removal lters is based on leukocyte Italy
adsorption on the lters. The adsorption of leuko- Institute on Membrane Technology, National
cytes was affected signicantly by lter materials, Research Council of Italy, ITM-CNR, Rende,
pore structure, and pore size. Italy
Plasma Separation Membrane Operational principle of steam and water boilers,

cooling towers, evaporative coolers, and evapora-
Plasma-exchange therapy has been increasingly tive condensers is based upon the evaporation of
applied clinically over the past few years. Mem- water. Dehydration from these equipment leaves
brane plasma separation has been used since behind a solution enriched with nonvolatile sub-
1979, which is similar to hemodialysis and stances mainly including salts of calcium, magne-
hemoltration. Plasma separation from whole sium, sodium, and silica, requiring the discharge
blood is now performed routinely. The materials of contaminant-rich water. The drained volume of
of plasma separation membranes are typically water in order to maintain an acceptable level of
made of nitrocellulose, polysulfone, and impurities in the evaporation equipment is known
polypropylene. as blowdown water or bleed water or simply
blowdown. In addition to the salts, blowdown
water can also contain bacteria, algae, and patho-
gens. Increased concentration of minerals and
Cross-References other impurities is undesirable for several appli-
cations. For example, in steam boiler, the evapo-
Blood Cell Origins ration of pure water can lead toward the scale
Blood Filtration formation, water carryover with steam, and block-
Blood Separation age of the piping system, while in case of cooling
Dialtration tower, concentrated blowdowns cause scaling and
corrosion problems even in downstream equip-
References ment. The concentration of salts and other solid
species in blowdown water is much higher than
Daugirdas JT, Van Stone JC, Boag JT (2006) Hemodialysis the original feedwater introduced.
apparatus. In: Daugirdas JT, Peter G, Blake PG, Todd S, In boiler, the blowdown water is discharged
Ing TS (eds) Handbook of dialysis. Lippincott Williams
& Wilkins, Riverwoods from the surface and bottom to get rid of oating
Higuchi A (2010) Separation and purication of stem and and settled impurities, respectively. On one side,
blood cells by porous polymeric membranes. In: less amount of blowdown water may cause the
266 Blue Energy
scale formation, water carryover, local heating quality equal or higher than that of makeup water,
leading toward the failure of boiler tubes, and various membrane operations have been tested.
less heat exchange rate, while the excessive dis- Nanoltration and reverse osmosis have shown
charge of blowdown water leads toward increased the capability to treat the water to required purity
water usage, loss of energy associated with the level but usually require a pretreatment step. On
discarded steam, large amount of makeup water, the other hand, forward osmosis and membrane
and excessive use of chemicals. Therefore, the distillation have shown more interesting results in
quantity of blowdown water must be optimized. terms of less severe pretreatment requirements,
Scarceness of freshwater, stringent environ- operational stability, and less fouling problems.
ment regulations, and cost reduction are the driv-
ing forces for treatment of blowdown water.
State-of-the-art treatment techniques for cooling References
tower blowdown water include lime-soda soften-
ing and thermal evaporation. The former removes Kochi Membrane Systems, Inc. (2014) Case study: ultra-
ltration reduces boiler blowdown at Indias largestpa-
hardness but involves the use of chemicals and
perboard mill, available online at: http://www.
produces sludge. Thermal evaporation, on the kochmembrane.com/PDFs/Case-Studies/KMS_ITC_
other hand, is highly energy intensive. Due to Case_Study.aspx
environmental concern, the concept of zero liquid Yu X, Yang H, Lei H, Shapiro A (2013) Experimental
evaluation on concentrating cooling tower blowdown
discharge for treatment of blowdown water has
water by direct contact membrane distillation. Desali-
gained popularity recently. The main tools to nation 323:134141. doi:10.1016/j.desal.2013.01.029
achieve zero liquid discharge under consideration
include membrane ltration, electrodialysis, and
evaporation.
In boilers, chemical pretreatment is applied to Blue Energy
reduce the scaling problem and to enhance the heat
transfer rate. These chemicals along with the other Rong Wang1, S. T. V. Sim2 and
contaminates present into the boiler feedwater Anthony Gordon Fane2
1
form sludge that settles down at the bottom of the School of Civil and Environmental Engineering,
boiler. Consequently, the blowdown becomes Singapore Membrane Technology Centre,
richer in impurities. Alternative membrane-based Nanyang Environment and Water Research
pretreatment can be applied to tackle these prob- Institute, Nanyang Technological University,
lems. For example, ultraltration has been used to Singapore, Singapore
2
affectively remove colloidal silica from the Singapore Membrane Technology Centre,
feedwater that imparts a vigorous effect in control- Nanyang Environment and Water Research
ling the boiler blowdown (Kochi membrane sys- Institute, Nanyang Technological University,
tems, Inc. 2004). A more rened membrane Singapore, Singapore
process such as nanoltration or reverse osmosis
can further improve the consequence.
As far as the posttreatment of boiler blowdown Introduction
is concerned, there have not been signicant
efforts devoted, probably due to high concentra- The energy crisis is becoming one of the most
tion and very complex nature of the waste stream. pervasive problems facing the world today plagued
However, advanced membrane operations, espe- with rapid economic growth and exponential
cially membrane distillation, can be appropriate increase in the global population. The key solution
candidate due to high temperature and concentra- is sustainable energy that combines renewable
tion of the stream and less fouling tendency of the energy with high energy efciency to balance the
MD process (Yu et al. 2013). In case of cooling growing energy needs without jeopardizing the
tower, where the objective is to recover water with earths climate. Of great interest is the potential of
Blue Energy 267
blue energy associated with osmotic power that hydraulic pressure (DP = 0). Water ows
can be harvested from the salinity gradient between from a less concentrated feed (F) to a more
two water streams such as seawater and freshwater, concentrated draw solution (DS).
as seawater is globally distributed (Lewis
et al. 2011), available all day and all year round. PRO is an intermediate osmotic process between B
Two emerging membrane techniques, pressure RO and FO. As long as the osmotic pressure differ-
retarded osmosis (PRO) and reverse electrodialysis ence across the semipermeable membrane is greater
(RED), are the primary methods proposed for than the applied pressure on a high-salinity stream,
recovering the osmotic energy. water will continuously ow from a low-salinity
stream at a low pressure into the pressurized high-
PRO Working Principles salinity stream. The increased volume of pressur-
Three osmotic processes are illustrated in Fig. 1: ized water is able to drive a turbine for power
generation (Thorsen et al. 2009), resulting in the
(a) Reverse osmosis (RO) that utilizes hydraulic recovery of the chemical potential energy from the
pressure (DP) to overcome the osmotic pres- difference in osmotic pressures of the two water
sure gradient (Dp) across a semipermeable streams, as shown schematically in Fig. 2.
membrane (DP > Dp), leading to the net
water movement from the concentrated feed RED Working Principles
(F) to the permeate (P) side. Unlike PRO, RED primarily depends on ion trans-
(b) Pressure retarded osmosis (PRO) that applies a port through the membranes when waters of dif-
hydraulic pressure lower than the osmotic pres- ferent salinities are placed on opposite sides of
sure gradient to the high-salinity stream (draw either cation or anion exchange membranes.
solution (DS)) (DP < Dp), giving a net water These membranes are stacked in an alternating
ux in the direction toward the draw solution. pattern between a cathode (reduction electrode)
(c) Forward osmosis (FO) that uses the osmotic and an anode (oxidation electrode) (Post
pressure gradient as the driving force without et al. 2007). Another distinctive difference is that
Blue Energy, a b c
Fig. 1 Osmotic processes P P
using semipermeable
membranes: (a) RO; (b)
PRO; (c) FO (arrow
indicates the direction of
water ow)
P F F DS F DS
P > 0 < P < > P = 0
Blue Energy,
Fresh water Fresh water @ low pressure
Fig. 2 Schematic of PRO
process for power
generation membrane
Water flow
Brine
P Diluted
Brine @ high pressure brine
Pump Turbine
generator
268 Blue Energy
Cathode Electrode
Less Saline Solution
(River water)
C
External Na+ Cl Na+ Cl Saline Solution
Load (Sea water)
A
Less Saline Solution
(River water)
Anode Electrode
Blue Energy, Fig. 3 A simple RED process showing the transport of ions where A is an anion exchange membrane and
C is a cation exchange membrane
the RED system is not pressurized like PRO to et al. 2012), while RED requires not only a mem-
obtain the osmotic energy mechanically. Instead, brane of minimal electrical resistance but also the
direct electrical energy can be harvested from the whole stack design with a low electrical resis-
movement of the electrons required to maintain tance. Normally, seawater and freshwater are
electroneutrality of the overall solution as the ions used as a pair for both PRO (Lewis et al. 2011;
from the saline solution are transported toward Thorsen et al. 2009; Ramon et al. 2011) and RED
either the anode or cathode electrodes, as shown (Ramon et al. 2011; Dugoecki et al. 2010) pro-
schematically in Fig. 3. cesses to harvest salinity gradient energy. Other
Specically, electroneutrality is maintained via salinity gradient pairs such as concentrated brine
reduction at the cathode and oxidation at the from seawater desalination (Enomoto et al. 2010)
anode via an external electric circuit that results or hypersaline lakes such as the Dead Sea (Loeb
in electrons being transferred from the anode to 1976, 2001; Lacey 1980) are also possible
the cathode. Electrical energy can then be resources for utilization. It is expected that large-
harnessed directly when an external load is scale applications for PRO and RED would be
connected to the circuit through the electrical cur- near river estuaries, hypersaline lakes, or seawater
rent (movement of electrons) and the potential desalination plants. In particular, the co-location
difference over the electrodes. Although the of PRO/RED with seawater desalination plants
applied electrical potential difference will reduce would allow the recovered energy from the RO
the driving force of the ions when the external brines to be used to desalinate the water.
load is connected to the circuit, ions will continue In the oil and gas and mining industries, large
to be transported toward the cathode and anode as amounts of water streams with high total
long as the electrochemical potential difference is dissolved salts (TDS) are produced from the
greater than the electrical potential difference. hydraulic methods used in the harvesting of gas
from coal beds (McBeth et al. 2003), oil from
Challenges and Potential Large-Scale subsurface formation (Ahmadun et al. 2009), or
Applications from mining activities (Balaba and Smart 2012).
The PRO process requires a robust membrane that Thus, the recovery of osmotic energy from such
can withstand high-operating pressures and produced water streams is also possible for large-
achieve a reasonable high water ux (Chou scale applications.
Bore Liquid: Solvent and Nonsolvent 269
References
Bore Liquid: Solvent and Nonsolvent
Ahmadun F-R, Pendashteh A, Abdullah LC, Biak DRA,
Madaeni SS, Abidin ZZ (2009) Review of technologies
Khayet Mohamed
for oil and gas produced water treatment, J Hazard
Mater 170:530551 Department of Applied Physics, Universidad B
Balaba RS, Smart RB (2012) Total arsenic and selenium Complutense de Madrid, Madrid, Spain
analysis in Marcellus shale, high-salinity water, and
hydrofracture owback wastewater, Chemosphere
89:14371442
Chou S, Wang R, Shi L, She Q, Tang CY, Fane AG Bore injection liquid is an internal coagulant
(2012) Thin-lm composite hollow ber pumped through the inner tube of the spinneret
membranes for pressure retarded osmosis (PRO) used for fabrication of hollow bers by the
process with high power density, J Membr Sci 389:
dry/wet spinning technique or the wet spinning
2533
Dugoecki P, Dbrowska J, Nijmeijer K, Wessling technique. Figure 1 shows schematic illustrations
M (2010) Ion conductive spacers for increased power of the cross-sections of the spinneret used in these
generation in reverse electrodialysis, J Membr Sci techniques. The spinneret has a tube-in-orice
347:101107
structure with typical dimensions 0.51 mm
Enomoto H, Fujitsuka M, Hasegawa T, Kuwada M,
Tanioka A, Minagawa M (2010) A feasibility study inner diameter and 0.92 mm outer diameter.
of pressure-retarded osmosis power generation The bore liquid is one of the controlling param-
system based on measuring permeation volume using eters of the wet spinning and dry/wet spinning
reverse osmosis membrane, Electr Eng Jpn 173:
techniques used to fabricate hollow bers. It
11291138
Lacey RE (1980) Energy by reverse electrodialysis, Ocean must be a non-solvent or a non-solvent mixture
Eng 7:147 exhibiting a low afnity with the polymer solution
Lewis A, Estefen S, Huckerby J, Musial W, Pontes T,
Torres-Martinez J (2011) In: Edenhofer O, Pichs-
Madruga R, Sokona Y, Seyboth K, Matschoss P,
Kadner S, Zwickel T, Eickemeier P, Hansen G,
Schlomer S, Stechow CV (eds) IPCC special report a
on renewable energy sources and climate change Polymer
mitigation. Cambridge University Press, Cambridge/ solution
New York
Loeb S (1976) Production of energy from concentrated
brines by pressure-retarded osmosis: I. Preliminary Bore
technical and economic correlations, J Membr Sci liquid
1:4963
Loeb S (2001) One hundred and thirty benign and
Bore
renewable megawatts from Great Salt Lake? The b liquid
possibilities of hydroelectric power by pressure-
retarded osmosis, Desalination 141:8591
McBeth IH, Reddy KJ, Skinner QD (2003) Chemistry of
trace elements in coalbed methane product water, Water Polymer Polymer
Res 37:884890 solution solution
Post JW, Veerman J, Hamelers HVM, Euverink GJW,
Metz SJ, Nymeijer K, Buisman CJN (2007) Salinity-
gradient power: Evaluation of pressure-retarded
osmosis and reverse electrodialysis, J Membr Sci
288:218230
Ramon GZ, Feinberg BJ, Hoek EMV (2011) Membrane-
based production of salinity-gradient power, Energy
Environ Sci 4:44234434 Bore Liquid: Solvent and Nonsolvent,
Thorsen T, Holt T (2009) The potential for power Fig. 1 Transversal (a) and lateral (b) cross-sections of a
production from salinity gradients by pressure retarded common spinneret used in dry/wet spinning and wet spin-
osmosis, J Membr Sci 335:103110 ning techniques
270 Bore Liquid: Solvent and Nonsolvent
(dope solution). The solubility parameter of the When a strong bore liquid coagulant such as
bore liquid must be different from that of the water for PVDF polymer is used for the fabrication
polymer to be coagulated. The interactions of a hollow ber membrane, a dense and smooth
between solvent and coagulant and between sol- surface with no obvious pores is formed. However,
vent and polymer play important roles affecting when a weaker coagulant is employed such as
the rate of polymer precipitation (coagulation, methanol, which may be mixed with water, the
solidication). For poly(vinylidene uoride) roughness of the internal surface of the formed
(PVDF) polymer solution as an example, water hollow ber may increase considerably due par-
is a strong coagulant commonly used for fabrica- tially to the formation of pores. Figure 2 shows the
tion of hollow bers. In the dry/wet spinning or effects of the bore liquid coagulant on the cross-
wet spinning, the polymer solution under gas sectional structure of porous hydrophobic hollow
pressure is extruded from the annular space of bers prepared by the dry/wet spinning technique
the spinneret, whereas the internal coagulant (Khayet et al. 2002). Water and 50 % (by volume)
(bore liquid) is driven through the central tube of ethanol in water were used as internal and external
the spinneret either by gravity force or by means coagulants. Different cross-sectional structures can
of a circulation pump. Coagulation of the internal be observed. In Fig. 2a prepared with water as
surface of the nascent ber starts immediately internal and external coagulants, long nger-like
after its extrusion from the spinneret, whereas voids are formed near the inner surface, while
the external surface experiences coalescence, smaller cavities are formed near the outer surface.
and orientation of polymer aggregates through Between the inner and outer layers, sponge-like
the air gap before gelation in the external coagu- structure appears. The inner layer is reduced
lation (i.e., non-solvent or non-solvent mixture) when ethanol is added to the internal coagulant as
bath. The air gap is the vertical distance between can be seen in Fig. 2b. A uniform sponge-like
the spinneret and the liquid coagulation bath and it structure extends over the entire hollow ber
may vary from 0 (wet spinning) to more than 1 m cross-section, in Fig. 2c, when ethanol is added to
(dry/wet spinning). The bore liquid and the sol- both the internal and external coagulants. The addi-
vent exchange by diffusion and then the polymer tion of ethanol delays the coagulation process, and
solution become thermodynamically unstable. a long nger-like structure changes to a short
Demixing may take place from the bore side or nger-like structure and further to a sponge-like
lumen side and from the shell or outside of the structure. The slow coagulation of the hollow
as-spun ber. Two types of demixing are possi- bers can be explained on the basis of the mutual
ble: liquidliquid demixing and gelation or crys- diffusivity of solvent/non-solvent exchange and
tallization. Gelation is not a phase separation solubility parameters of the materials involved in
process whereas crystallization involves a hollow ber spinning. The addition of ethanol in
solidliquid demixing. In order to determine the water reduces the diffusion of non-solvent into
composition or temperature at which the solution the nascent hollow ber and consequently
is no longer thermodynamically stable, turbidity decreases the rate of precipitation (coagulation,
or cloud points must be determined. Cloud points precipitation). The surface roughness increased
are dened as the moment when the solution when ethanol was added to the bore liquid coag-
changes from clear to turbid. To localize the ulant, which is due to the increase of the pore
cloud point of a ternary system, a possible size. The pores at the inner surfaces of the hollow
method is titration of polymer solution bers are larger than those at the outer surfaces,
containing polymer and solvent with and the pore size increases both at the inner and
non-solvent or with a mixture of solvent and at the outer surfaces as ethanol was added to the
non-solvent at a given temperature until perma- coagulants (Khayet et al. 2002). A decrease of
nent turbidity is detected. The turbidity point can the porosity of PVDF hollow bers was also
be detected by turbidity measurements or light observed with the addition of ethanol in the
scattering methods. bore liquid.
Boric Acid Separation by Membrane Contactor 271
Bore Liquid: Solvent and Nonsolvent, Fig. 2 Cross- coagulant, (b) internal coagulant 50 % ethanol in water
sectional scanning electron microscopy (SEM) structure of and external coagulant water, (c) 50 % ethanol in water
PVDF hollow ber membranes prepared by wet/dry spin- coagulant (Reprinted from Khayet et al. (2002), Copyright
ning technique with different coagulants: (a) water 2002, with kind permission from Elsevier)
References water up to around 7 ppm in seawater and can be

even higher in wastewaters discharged from some
Khayet M, Feng CY, Khulbe KC, Matsuura T (2002) Prep- industrial plants. Due to its toxic effects on both
aration and characterization of polyvinylidene uoride
humans and plants, the World Health Organiza-
hollow ber membranes for ultraltration. Polymer
43:38793890 tion (WHO) has xed at 0.3 ppm the maximum
allowed concentration of boron in water. Different
are the techniques that can be used for controlling
the boric acid content of water, like adsorption,
Boric Acid Separation by Membrane ion-exchange resins, solvent extraction, electrodi-
Contactor alysis, multiple stages of reverse osmosis units
(working at different pH), and membrane
Alessandra Criscuoli contactors. In particular, the effectiveness of
Institute on Membrane Technology, National membrane contactors has been conrmed using
Research Council of Italy, ITM-CNR, Rende, various congurations to treat waters containing
Italy boric acid.
Boric acid was selectively separated through
supported liquid membranes containing inside
Boron is usually found in water as boric acid, and micropores 1,3-diols (carrier) dissolved into
its concentration ranges from few ppb in river o-dichlorobenzene. The membranes were
272 Boric Acid Separation by Membrane Contactor
Boric Acid Separation by Organic-filled

Membrane Contactor, micropore
Fig. 1 Transport of boric
acid through a supported
liquid membrane
Boric acid
Feed water Alkaline aqueous phase
Boric Acid Separation by

Membrane Contactor,
Fig. 2 Transport of boric
acid in a hydrophilic
membrane
Boric acid
Distilled water
Feed water
hydrophobic and separated the aqueous stream extractant were also observed by Park and Lee
containing boric acid from an alkaline aqueous (1995). Moreover, they found good efciencies
phase, representing the so-called receiving of anion exchange membrane contactors for the
phase; see Fig. 1 (Bachelier et al. 1996). removal of boric acid from liquid radioactive
The removal of boric acid was also obtained by wastes.
employing hydrophilic microporous membranes,
sending distilled water as extractant phase. In this
case, the membrane pores were lled with the References
aqueous feed, and, by properly acting on the pres-
Bachelier N, Chappey C, Langevin D, Mtayer M,
sures of the two phases, the interface was located
Verchre J-F (1996) Facilitated transport of boric acid
at the membrane surface-extractant side, so that through supported liquid membranes. J Membr Sci
the boric acid removal occurred by simply diffu- 119:285294
sion from the feed to the extractant (Fig. 2). For Criscuoli A, Rossi E, Cofone F, Drioli E (2010a) Boron
removal by membrane contactors: the water that
practical implementations, an integrated mem-
puries water. Clean Technol Environ Pol 12:5361
brane system where the extractant was continu- Criscuoli A, Rossi E, Cofone F, Drioli E (2010b) Boron
ously regenerated and recycled to the membrane removal by membrane contactors: the water that
contactor was also proposed and designed puries water. Additional information. Clean Technol
Environ Pol 12:63
(Criscuoli et al. 2010a, b).
Park JP, Lee KJ (1995) Separation of boric acid in liquid
Successful performances of hydrophilic micro- waste with anion exchange membrane contactor. Waste
porous membranes using distilled water as Manag 15:283291
Boron Reduction 273
and animals when found at high concentration in

Boron Reduction irrigation or drinking water. The effects of boron
on plants and mammals were summarized by
Piotr Dydo the WHO (2009), Kabay et al. (2010), and
Silesian University of Technology, Gliwice, Hilal et al. (2011). Among the reported effects, B
Poland the adverse impact of boron on the male repro-
ductive system in rats, mice, and dogs was
presented. Based on analysis of toxicokintetic of
Synonyms boron, the WHO proposed a guideline value for
boron in drinking water of 2.4 mg/L in 2011. In
Boron removal areas with high natural boron levels, local regula-
tory and health authorities are, however, advised
to consider values in excess of 2.4 mg/L by
Characteristics assessing exposure from other sources
(e.g., food).
The boron element has an average concentration The low drinking water boron content limit
in the Earths crust equal to about 10 mg/kg. Due implies the necessity for a reduction of boron
to its strong afnity toward oxygen, it exists in content in drinking water, especially when seawa-
nature mainly in the form of boric acids or ter or boron-rich underground waters (e.g., geo-
borates. Boric acid and borates are mainly used thermal waters) are to be treated. There could also
in the glass and ceramic industry to obtain boro- be a need for boron removal from industrial and
silicate glass, insulation berglass, ame retardant municipal efuents that contain more than the
berglass, ceramic glazes, and porcelain WHO guideline value. Conventional methods
enamels. Boron compounds are also used as a for water treatment do not signicantly remove
ame retardant in plastics and cellulosic boron. The methods proven to be efcient in a
insulation, neutron absorbers, herbicides (when reduction of boron content in waters are either
at high concentrations) or fertilizers (when at adsorption or membrane based. Adsorption-
low doses), and components of washing powders based methods for boron removal include coagu-
and soaps. lation with metal hydroxides, adsorption on clays,
According to the WHO (2009) and (2011), y ashes, and activated carbon (WHO 2009).
average boron concentration in surface waters Also, successful boron reduction with conven-
does not exceed the value of 0.5 mg/L. However, tional anion exchange resins or resins
the amount of boron in these waters depends upon functionalized with N-methyl D-glucamine
the presence of boron-bearing minerals in the (Chillon Arias et al. 2011) was reported. Among
proximity of water reservoirs. It may also be the membrane-based methods, boron removal by
affected by the discharge of municipal and reverse osmosis is reported most commonly. Also
industrial efuents into the environment. On some achievements in boron removal with
the other hand, naturally occurring boron is ion-exchange membranes as well as boron
present in ground waters at a wide concentration removal by electrodialysis are presented. In addi-
range from <0.3 to >100 mg/L. This includes tion, some hybrid systems, which combine mem-
strongly mineralized, naturally carbonized brane ltration with sorption, are proposed, e.g.,
geothermal waters. Considerable amounts of the adsorption-membrane ltration (AMF) sys-
boron are also present in oceans, with an average tem. In this system, microparticulate boron selec-
concentration of 4.5 mg/L as reported by the tive resin suspension is recirculated in the
WHO in 2009. retentate. Boron species remains adsorbed on the
Despite the fact that boron is an important resin, while the boron-depleted water is continu-
nutrient, it manifests toxic action against plants ously separated by microltration. According to
274 Boron Removal
Kabay et al. 2008, the main advantage of the AMF

over other methods of boron separation lies in the Boron Removal by Electrodialysis
fast removal of boron, ease of sorbent separation,
and low operating pressure. Piotr Dydo
According to the WHO (2009), the Silesian University of Technology, Gliwice,
abovementioned methods for boron removal are Poland
likely to be prohibitively expensive, and blending
with waters of low boron content may be the only
economically feasible option to reduce high con- As in the case of boron removal by reverse osmo-
centrations of boron in water. sis, the effectiveness of boron removal by electro-
dialysis (ED) is strongly affected by the aqueous
chemistry of the boron species. In aqueous solu-
tions, boric acid may exist in the form of boric
Cross-References
acid, metaborate, and polyborates. Boric acid as a
weak electrolyte (pKa = 9.2) dominates in
Boron Removal by Electrodialysis
diluted aqueous solutions of a pH equal to or
Boron Removal by Reverse Osmosis
less than 9, while at higher pH borate ions domi-
Boron Removal with Ion-Exchange
nate. Therefore, the effectiveness of boron trans-
Membranes
port across ion-exchange membranes (IEM) in the
electrodialysis process needs to be discussed with
References regard to the type of the species that dominates in
the dilute.
Chillon Arias MF, Valero i Bru L, Prats Rico D,
Varo Galvan P (2011) Comparison of ion
exchange resins used in reduction of boron in desali-
nated water for human consumption. Desalination The Effectiveness of Boric Acid
278:244249 Transport in ED
Hilal N, Kim GJ, Somereld C (2011) Boron removal from
saline water: a comprehensive review. Desalination
273:2335
Boric acid, H3BO3, is a small and electrically
Kabay N, Bryjak M, Schlosser S, Kits M, Avlonitis S, neutral species. The reported boric acid removal
Matejka Z, Al-Mutaz I, Yuksel M (2008) Adsorption- efciencies and electric current efciencies in ED
membrane ltration (AMF) hybrid process for boron are poor when compared to ionic species (Melnik
removal from sweater: an overview. Desalination
223:3848
et al. 1999; Turek et al. 2005, 2007, 2008a; Kabay
Kabay N, Guler E, Bryjak M (2010) Boron in seawater and et al. 2008; Banasiak and Schafer 2009; WHO
methods for its separation a review. Desalination 2009). In fact, boric acid uxes across IEMs are
261:212217 so low that the possibility of boric acid separation
WHO (2009) Boron in drinking-water. Background docu-
ment for development of WHO guidelines for drinking-
from ionic species, strong acids (Melnik
water quality. World Health Organization. http:// et al. 2005, 2007) and salts (Turek et al. 2005,
whqlibdoc.who.int/hq/2009/WHO_HSE_WSH_09.01_ 2007; Bandura-Zalska et al. 2009; Dydo 2012a),
2_eng.pdf. Accessed 17 June 2012 by the use of electrodialysis is considered.
WHO (2011) Guidelines for drinking-water quality,
4th edn. World Health Organization. http://whqlibdoc.
Employing this method, ionic species are effec-
who.int/publications/2011/9789241548151_eng.pdf. tively transported across IEMs while boric acid
Accessed 17 June 2012 remains in the ion-depleted dilute solution.
The type of the membrane, the dilute boron
concentration, the presence of ions in the dilute,
and the electric current density were shown to
Boron Removal affect the rate of boric acid transport across IEM
to a great extent (Melnik et al. 1999; Yazicigil and
Boron Reduction Oztekin 2006; Kabay et al. 2008; Dydo 2012a).
Boron Removal by Electrodialysis 275
The reported mechanism for boric acid transport upon the pH of the dilute, the type of the mem-
across IEMs is diffusion (Dydo 2012b); however, brane, dilute boron concentration, and the kind of
the ux of boric acid was found to increase with the ion cotransported across the membrane.
an increase in the ux of ionic species. This was Yazicigil et al. in 2006 reported that at a dilute
suggested to be the result of a kind of ion-coupled pH of approx. 9, there is a maximum boron trans- B
transport of boric acid across IEMs. The effective- port rate, and at higher pH the rate of boron
ness of such a transport was found to decrease in transport decreases. These results are contradic-
the following ion series: sulfate nitrate tory to those presented by Melnik et al. (1999)
chloride. It was also reported that most of the (heterogeneous membranes), Turek et al. (2007),
boric acid was transported across anion-exchange and Kabay et al. (2008), according to which
membrane (AEM). there is a continuous increase in the rate of boron
Turek et al. 2008b reported high boron uxes transport with dilute pH even when above 9.0.
and exceptionally high electric current efcien- However, the Yazicigil et al. (2006) results seem
cies of up to 220 % when boric acid was to be justied by high boron concentration
transported from ion-depleted, neutral or acidic, (0.1 mol/L) in their experiments. Similar behavior
dilute into the alkaline concentrate of pH >11. was observed by Ayyildiz and Kara in 2005.
Such a behavior was identied to be the result of It is agreed that an increase in dilute boron con-
Donnan dialysis, in which hydroxyl acts as a centration causes an increase in boron (borate)
carrier for boric acid. However, in the light of ux (Yazicigil et al. 2006; Kabay et. al 2008;
recent reports, simple boric acid diffusion should Turek et al. 2008a). It is also agreed that boron
explain the results as well. (borate) is transported faster in the presence of
Tentative results on boron removal by chloride ions than in the presence of sulfates
electrodialysis units equipped with ion-exchange (Yazicigil et al. 2006; Kabay et al. 2008) or
spacers were also presented by Oren et al. in 2006. nitrates.
The possibility of up to an 80 % reduction of In 2005 and 2008a, Turek et al. analyzed the
boron from water containing 4.5 mg/L was effect of salinity on the electric current efciency
reported. of boron (borate) transport. They found that the
observed electric current efciencies are low as
long as dilute salinity remains high. During the
The Effectiveness of Borate initial part of the experiments, no, or almost no,
Transport in ED boron was transported from the dilute. Then, i.e.,
when more than 90 % of the Cl was removed
Melnik et al. (1999), Yazicigil et al. (2006), and from the dilute, a dramatic increase in boron
Kabay et al. (2008) reported that as the pH of (borate) electric current efciency was observed.
boron-containing water was brought up to just This increase was accompanied by an increase in
above the value of 9, a signicant increase in the boron ux. Such an effect can be identied as the
rate of boron transport across IEMs is observed. result of low borate mobility (diffusivity) when
Moreover, Melnik et al. 1999 reported that in the compared to other ions present in the waters, i.e.,
case of heterogeneous membranes, an increase in Cl. As long as there are ions of higher mobility
the boron transport rate across anion-exchange than borates in the dilute, boron (borate) ux
membrane at pH >9 is accompanied by a decrease remains low. However, even in an ion-depleted
in the rate of boron transport across cation- dilute, the electric current efciency of boron
exchange membrane (CEM). It is clear that (borate) transport has not exceeded the value of
under these conditions (pH >9), borates are the 30 %. The remaining percentage of the electric
boron species that dominates and their transport current was probably utilized for hydroxyl ion
across AEMs rather than CEMs should be transport. A considerable drop in the pH of the
discussed. The effectiveness of borate transport dilute was reported afterwards. This resulted in a
across AEMs in ED systems was found to depend drop in dilute pH and in a consequent reduction in
276 Boron Removal by Reverse Osmosis
the ux of boron since all the borate present was Turek M, Dydo P, Trojanowska J, Bandura B (2007) Elec-
converted into boric acid. So the low electric trodialytic treatment of boron-containing wastewater.
current efciency of borate transport in an ED Turek M, Bandura B, Dydo P (2008a) Electrodialytic
system and the necessity for deep dilute deminer- boron removal from SWRO permeate. Desalination
alization make the economic feasibility of borate 223:1722
removal by ED questionable. Turek M, Bandura B, Dydo P (2008b) The inuence of
concentrate alkalinity on electrodialytic boron trans-
port. Desalination 223:119125
WHO (2009) Boron in drinking-water. Background
document for development of WHO guidelines for
Cross-References
drinking-water quality. World Health Organization.
http://whqlibdoc.who.int/hq/2009/WHO_HSE_WSH
Boron Reduction _09.01_2_eng.pdf. Accessed 17 Jun 2012
Boron Removal by Electrodialysis Yazicigil Z, Oztekin Y (2006) Boron removal by electro-
dialysis with anion-exchange membranes. Desalination
190:7178
References
Ayyildiz HF, Kara H (2005) Boron removal by ion

exchange membranes. Desalination 180:99108
Banasiak LJ, Schafer AI (2009) Removal of boron, uoride
and nitrate by electrodialysis in the presence of organic Piotr Dydo
matter. J Membr Sci 334:101109 Silesian University of Technology, Gliwice,
Bandura-Zalska B, Dydo P, Turek M (2009) Desalination
of boron-containing wastewater at no boron transport.
Poland
Dydo P (2012a) The effect of process parameters
on boric acid transport during the electrodialytic Reverse osmosis (RO) is a commonly applied
desalination of aqueous solutions containing selected
salts. Desalination. In press. doi:10.1016/j.bbr.2011.
membrane technique of desalination capable of
03.031 producing permeates of drinking-water quality.
Dydo P (2012b) The mechanism of boric acid transport In most cases, the boron content in the product
during an electrodialytic desalination process. J Membr should not exceed the WHO 2009 and 2011
Sci 407408:202210
Kabay N, Arar O, Acara F, Ghazal A, Yuksel U, Yuksel
guideline value of 2.4 mg/L. The effectiveness
M (2008) Removal of boron from water by electrodi- of boron rejection by reverse osmosis membranes
alysis: effect of feed characteristics and interfering ions. is, however, strongly affected by its concentration
Desalination 223:6372 in the feedwater and the aqueous chemistry of
Melnik L, Vysotskaja O, Kornilovich B (1999) Boron
behavior during desalination of sea and underground
boron specie (Kabay et al. 2010; Hilal
water by electrodialysis. Desalination 124:125130 et al. 2011). In aqueous solutions, boric acid may
Melnik L, Goncharuk V, Butnyk I, Tsapiuk E (2005) exist in the form of boric acid, metaborates, and
Boron removal from natural and wastewaters using polyborates. Boric acid as a weak electrolyte
combined sorption/membrane proces. Desalination
185:147157
(pKa = 9.2) dominates in diluted aqueous solu-
Melnik L, Goncharuk V, Butnyk I, Tsapiuk E (2007) tions of a pH less than 9. Since the pH of most
Development of the sorption-membrane green tech- naturally occurring waters is lower than 9, the rate
nology for boron removal from natural and wastewa- of boric acid transport determines the effective-
ters. Desalination 205:206213
Oren Y, Linder C, Daltrophe N, Mirsky Y, Skorka J,
ness of boron removal by RO.
Kedem O (2006) Boron removal from desalinated sea- According to Kabay et al. 2010, the rejection of
water and brackish water by improved electrodialysis. boric acid by RO membranes is, however, low due
Desalination 199:5254 to its small size and lack of electric charge. The
Turek M, Dydo P, Ciba J, Trojanowska J, Kluczka J,
Palka-Kupczak B (2005) Electrodialytic treatment
reported boron rejection coefcients range from
of boron-containing wastewater with univalent around 20 % in the case of old RO membranes to
permselective membranes. Desalination 185:139145 approx. 90 % in the case of modern RO
Boron Removal by Reverse Osmosis 277
membranes (WHO 2009; Kabay et al. 2010; Hilal stages (if possible at elevated pHs) and a blending
et al. 2011). of all the permeates produces water of the desired
The effectiveness of boron removal from sea- quality (Hilal et al. 2011; Faigon and Hefer 2008).
water by RO can be enhanced by adjusting the This design was successfully applied, e.g., in the
feedwater pH to above the value of 9.25. The full-scale seawater treatment plant in Eliat. B
monoborate anion, which dominates in diluted
solutions at pH > 9.5, was found to be rejected
more effectively than boric acid due to its larger Cross-References
size and discrete charge (Kabay et al. 2010; Hilal
et al. 2011). The reported boron rejection coef- Boron Reduction
cients at pH = 11 exceeds 99 % in the case of Boron Removal by Electrodialysis
dense seawater reverse osmosis membranes. Also, Boron Removal with Ion-Exchange
in the case of boron-rich geothermal waters and Membranes
industrial wastewaters, high boron rejection is
observed only at pH > 10.5 (Koseoglu
et al. 2010; Dydo et al. 2005). Furthermore, it References
was shown that the effectiveness of boron
Dydo P, Turek M, Ciba J, Trojanowska J, Kluczka J (2005)
removal can be further enhanced by creating Boron removal from landll leachate by means of
borate complexes (esters) with polyhydroxyl alco- nanoltration and reverse osmosis. Desalination
hols (Geffen et al. 2006; Dydo et al. 2012), 185:131137
although again, under alkaline conditions only. Dydo P, Nem I, Turek M (2012) Boron removal and its
concentration by reverse osmosis in the presence of
The results of mechanistic studies on boron polyol compounds. Sep Purif Technol 89:171180
rejection by reverse osmosis membranes Faigon M, Hefer D (2008) Boron rejection in SWRO at
conducted by Sagiv and Semiat (2004), Hyung high pH conditions versus cascade design. Desalination
and Kim (2006), Hun et al. (2009), and Tu 223:1016
Geffen N, Semiat R, Eisen MS, Balazs Y, Katz I, Dosoretz
et al. (2010) showed that permeabilities of boric CG (2006) Boron removal from water by complexation
acid are at least one order of magnitude larger than with polyol compounds. J Membr Sci 286:4551
those of monoborate ion. It proves more intensive Hilal N, Kim GJ, Somereld C (2011) Boron removal from
diffusion of boric acid than of borate across RO saline water: a comprehensive review. Desalination
273:2335
membranes. Also, the reection coefcients for Hun PVX, Cho S-H, Moon S-H (2009) Prediction of boron
boric acid (<0.985) are considerably lower than transport through seawater reverse osmosis membranes
those specic for monoborate (>0.995) as using solution-diffusion model. Desalination
reported by Hyung and Kim (2006). This, in 247:3344
Hyung H, Kim J-H (2006) A mechanistic study on boron
turn, indicates strong boric acidwater interaction. rejection by sea water reverse osmosis membranes.
Feedwater pH is said to have a dominant J Membr Sci 286:269278
impact on boron rejection by RO membranes Kabay N, Guler E, Bryjak M (2010) Boron in seawater and
(Hilal et al. 2011). Among other RO parameters methods for its separation a review. Desalination
261:212217
that affect boron rejection are: operating pressure Koseoglu H, Harman BI, Yigit NO, Guler E, Kabay N,
(increase), feed temperature (decrease), feedwater Kitis M (2010) The effects of operating conditions on
salinity (decrease), and recovery (decrease). boron removal from geotermal waters by membranes
However, RO systems cannot be directly oper- processes. Desalination 258:7278
Sagiv A, Semiat R (2004) Analysis of parameters affecting
ated at high feedwater pH due to the possibility of boron permeation through reverse osmosis membranes.
membrane scaling with insoluble calcium and J Membr Sci 243:7987
magnesium basic compounds. Therefore, several Tu KL, Nghiem LD, Chivas AR (2010) Boron removal by
modications to the operation of RO and its reverse osmosis membranes in seawater desalination
applications. Sep Purif Technol 75:87101
design have been proposed. In general, a cascade WHO (2009) Boron in drinking-water. Background docu-
design for RO systems in which a fraction of the ment for development of WHO guidelines for drinking-
rst stage permeate is treated in the following RO water quality. World Health Organization. http://whqlibdoc.
278 Boron Removal with Ion-Exchange Membranes
who.int/hq/2009/WHO_HSE_WSH_09.01_2_eng.pdf. accompanying ions in the feed solution, and the

Accessed 17 June 2012 type of the carrier anion in the receiving solutions.
WHO (2011) Guidelines for drinking-water quality, 4th edn.
World Health Organization. http://whqlibdoc.who.int/ The effect of pH of the feed solution was found to
publications/2011/9789241548151_eng.pdf. be complex. At high boron concentration
Accessed 17 June 2012 (0.1 mol/L), maximum boron ux was observed
at around the pH of 9.5, while in the case of a
diluted solution (0.001 mol/L), maximum boron
ux was observed at the maximum examined
Boron Removal with Ion-Exchange feedwater pH of 11.5. Such behavior was
Membranes explained by the formation of polyborate ions at
high boron concentrations and their absence in
Piotr Dydo diluted solutions. The maximum rate of boron
Silesian University of Technology, Gliwice, transport at the pH of around 9.5 in the case of
Poland concentrated boron solution was later conrmed
by Yazicigil et al. in 2006. Ayyildiz and Kara in
2005 reported also that the pH of the receiving
Apart from boron removal by electrodialysis, phase affects boron ux with its maximum at
ion-exchange membranes are able to transport around a pH of 9.5. Moreover, the boron-
boron in the so-called Donnan dialysis (DD) or accompanying anions, chlorides, bicarbonates,
diffusion dialysis processes (Fig. 1). and sulfates, were found to affect the rate of
Ayyildiz and Kara (2005) examined the effec- boron transport in DD with a maximum observed
tiveness of boron transport across anion-exchange in the presence of bicarbonate. On the other hand,
membranes in a DD process. They found that the highest boron transport rates were observed
boron ux depends upon the membrane, concen- with sodium chloride in the receiving solution.
tration of boron in the feed solution, pH of the feed Neosepta AHA and AMH membranes produced
and receiving solution, presence of the similar uxes of boron, while the superiority of
Boron Removal with Ion-Exchange Membranes, Fig. 1 Boron Removal with Ion-Exchange Membranes
Breakthrough Pressure 279
AFN was clearly seen. In 2011, Kir et al. showed needs to be done to enhance the kinetics of the
that plasma modication of the existing anion- diffusive transport of borate during DD.
exchange membranes may result in a signicant
enhancement of the rate of boron transport in a
DD process. Cross-References B
The mechanistic study on boric acid trans-
port across cation-exchange membranes Boron Reduction
(CEMs) in a DD process was presented by Boron Removal by Electrodialysis
Goli et al. in 2010. It was reported that the Boron Removal by Reverse Osmosis
membrane diffusion coefcients for boric acid
depend not only upon the manufacturer of the
References
CEM but also upon the type of the cation pre-
sent in the membrane. The same behavior is Ayyildiz HF, Kara H (2005) Boron removal by ion
reported for arsenite. In general, in the presence exchange membranes. Desalination 180:99108
of monovalent cations in the membrane, higher Bryjak M, Pozniak G, Kabay N (2007) Donnan dialysis of
uxes of boron were observed than in the pres- borate anions through anion exchange membranes: a
new method for regeneration of boron selective resins.
ence of divalent cations. Such a behavior was React Funct Polym 67:16351642
later conrmed by Dydo in 2012. The existing Dydo P (2012) The mechanism of boric acid transport
differences in boric acid membrane diffusion during an electrodialytic desalination process.
coefcients were rationalized by Goli J Membr Sci 407408:202210
Goli E, Hiemstra T, Van Riemsdijk WH, Rahnemaie R,
et al. (2010) as being the result of the change Malakouti MJ (2010) Diffusion of neutral and ionic
in the mean viscosity of the solution conned in species in charged membranes: boric acid, arsenite,
membrane pores. and water. Anal Chem 82:84388445
In 2007, Bryjak et al. proposed a Donnan Kir E, Gurler B, Gulec A (n.d.) Boron removal from
aqueous solution by using plasma-modied and
dialysis-based method for the regeneration of unmodied anion-exchange membranes. Desalination
nely divided boron selective resin (BSR) 267: 114177
DOWEX XUS 43594.00. In this process, BSR Yazicigil Z, Oztekin Y (2006) Boron removal by electro-
slurry with boron adsorbed on it is fed into the dialysis with anion-exchange membranes. Desalination
190:7178
DD feed compartment. It was assumed that there
are always some minute amounts of borate in the
feedwater at equilibrium with the BSR. These
amounts were subject to transport into the receiv-
Breakthrough Pressure
ing solution as a result of Donnan dialysis, which
should ultimately result in complete boron
Renzo Di Felice
removal from the BSR. The net effect of the
University of Genova, Genova, Italy
process would be a regenerated resin in the chlo-
ride form and boron (borate)-rich receiving solu-
tion. As reported by Bryjak et al. (2007),
The pressure required for a liquid to enter the
considerably high boron uxes were observed
membrane pore is called breakthrough pressure.
during the course of a DD regeneration of boron
This parameter is of practical relevance as the
containing BSR. Boron desorption was found to
membrane will not work properly if liquid enters
be the phenomena that governs kinetics of the
the pores reducing in this way the contact ef-
boron transport in the process.
ciency. Breakthrough pressure can be estimated
Unfortunately the tentative, mostly mechanis-
by using the Young-Laplace equation
tic report on DD boron removal presented in this
chapter does not provide much information about
2s cos y
the effectiveness of boron removal from water nor P
r
its nal concentration. It seems that a lot of work
280 Breath Figure Membranes
where s is the liquid surface tension, y the contact passes through the membrane, while the solute
angle, and r the radius of membrane pore. It can be (product) is retained. Suppose the solution can
seen that high breakthrough pressure can be be viewed as a mixture containing a macromolec-
obtained by using liquids with high surface ten- ular product, A, and a micromolecular impurity,
sion, by reducing the pore size, and by increasing B. Now suppose that the buffer contains a desired
the contact angle. However, there are practical component, C a salt perhaps. Then, buffer dis-
limitations. For example, most of the physical sol- placement involves the reduction in the concen-
vents used for bulk CO2 removal have low surface tration of B while keeping the concentration of
tensions, and the reduction in the membrane pore A constant and raising the concentration of C to
size beyond certain limit results into reduced trans- the desired level. This can be done using constant
port of the gas through the membrane due to diffu- volume (continuous) dialtration, discontinuous
sion limitations. The higher contact angle for a dialtration by dilution, or discontinuous
given solvent is favored by the membrane material dialtration by volume reduction.
having low surface energies. If constant volume dialtration is used and
In order to give an idea, a hydrophobic poly- B has an apparent rejection coefcient of zero,
propylene membrane with a pore size of 0.05 mm the concentration of B declines according to the
has a breakthrough pressure larger than 1 MPa expression (Cheryan 1998)
(Wiesler 1996).
cB cB0 expV b =V s
References
where Vb is the volume of buffer added while Vs is
Wiesler F (1996) Ultrapure water. UP130427 the (constant) volume of solution. Assuming that
component, C, also has a rejection coefcient of
zero, its concentration in the retentate solution
Breath Figure Membranes increases according to the expression
Honeycomb Membrane Structure cc cc0 1 expV b =V s
where cc0 is the concentration of C in the buffer

tank. If the product, A, has a rejection coefcient
BTEX Removal by PV on Polymeric of 1.0, its concentration will remain unchanged.
Membranes
Toluene Removal by PVon Polymeric Membranes

References
Cheryan M (1998) Ultraltration and microltration,

2nd edn. CRC Press, Boca Raton
Buffer Displacement in Diafiltration
Greg Foley Buffer Flush in Diafiltration

School of Biotechnology, Dublin City University,
Dublin, Ireland Greg Foley
School of Biotechnology, Dublin City University,
Dublin, Ireland
Buffer displacement involves the use of
dialtration to replace the solvent (non-product)
fraction of a macromolecular solution, with a A buffer ush is simply the replacement of an
buffer. The basis of the operation is that the sol- impure, product-containing solution with buffer
vent, including low molecular weight impurities, by subjecting it to dialtration.
Bulk Biotech Industry 281
The most common way of performing a in 1928 paving the way for the industrial scale
buffer ush, at least in a membrane context, production of antibiotics and amino acids. The
is to use constant volume (continuous) discovery and increased understanding of the
dialtration in which a certain volume of DNA created the foundation to use molecular
buffer is added to a given volume of impure engineering to recombine DNA leading to the B
solution while that solution undergoes batch second wave of biotechnology processes in the
ultraltration, all the while keeping the solu- 1980s. The current third and so far nal wave of
tion volume constant. biotechnology the white biotechnology aims
The greater the ratio of buffer volume (Vb) to to replace the C2/C3 chemistry based on fossil
solution volume (Vs) ( Dialtration Factor), the fuels such as oil and gas by biotechnological
greater will be the reduction in undesired impuri- processes. It is foreseen that in the near future,
ties. For example, if an impurity has a rejection up to 20 % of all chemical products with a market
coefcient of 0.2 and the dialtration factor is 5, value of approx 250 billion will be produced by
the reduction in impurity concentration is given biotechnology. Approximately 60 % of the prod-
by (Foley 2013) ucts produced by white biotechnology will be
intermediated chemicals used in the pharmaceuti-
c=c0 e1sV b =V s e510:2 0:018 cal industry, and the remaining 40 % will be bio-
polymers and special chemicals for various
References industries (Festel et al. 2004).
Since the 1970s, cross-ow membrane pro-
Foley G (2013) Membrane ltration a problem solving cesses have established themselves in the
approach with MATLAB. Cambridge University Press, downstream processing of the biotechnology
Cambridge, UK
industry for the recovery and purication of
the products. It is foreseen that membrane
processes will also play an important role in
white biotechnology and the related concept of
Bulk Biotech Industry bioreneries. Similar to petroleum reneries,
bioreneries are aiming for the full utilization
Frank Lipnizki of biomass for the simultaneous production of
Alfa Laval Copenhagen, Sborg, Copenhagen, biofuels, biochemicals, heat, and power
Denmark (Axegrd 2005). The integrated production of
biomaterials can be based on, e.g., sugar,
starch, and cellulose-based feedstock and as
The term bulk biotech industry is referring to such extend current sugar, starch, and pulp
the use of biological processes on industrial scale factories. Membrane processes as highly selec-
to produce bulk products. Examples of products tive and energy-saving processes are well
from the bulk biotech industry are antibiotics, suited to play an important part in bioreneries
enzymes, organic and amino acids, vitamins, and thus the white biotechnology.
bioalcohols, and biopolymers. One of the earliest The key membrane technologies for the bulk
biotechnological processes adopted by humans is biotech industry are microltration (MF), ultral-
the production of alcohol by fermenting fruits tration (UF), nanoltration (NF), and reverse
around 50006000 years ago. The production of osmosis (RO). Other membrane technologies
lactic acid by Pasteur in 1857 is often considered such as membrane contactors (MC), electrodialy-
to be the beginning of modern biotechnology sis (ED), pervaporation (PV), and vapor perme-
followed by the industrial scale production of ation (VP) plus membrane bioreactors for
citric acid by Pzer in 1923 (Chotani continuous fermentation are less established but
et al. 2007). The rst wave of biotechnology have, nevertheless, the potential to become
started with the discovery of penicillin by Fleming increasingly important.
282 Bulk Liquid Membrane
References water-immiscible organic liquid membrane solu-

tions may be unied under the term bulk organic
Axegrd P (2005) The future pulp mill a biorenery? hybrid liquid membrane (BOHLM) systems (for
Presentation at the 1st international biorenery work-
more details, see Kislik 2010a).
shop, Washington, DC
Chotani GK, Dodge TC, Gaertner AL, Arbige MV Bulk LM processes with water-soluble carriers
(2007) Industrial biotechnology: discovery to delivery. are dened as bulk aqueous hybrid liquid mem-
In: Kent JA (ed) Kent and Riegels handbook of indus- brane (BAHLM) systems (for more details, see
trial chemistry and biotechnology, 11th edn. Springer,
Kislik 2010b). Regenerable water-soluble polyionic
Berlin
Festel G, Knll J, Gtz H, Zinke H (2004) Der einuss der complexants are used as suitable aqueous liquid
biotechnologie auf produktionsverfahren in der membrane carriers. These polyelectrolytes typically
chemieindustrie. Chem Ing Technol 76:307312 have a very high effective concentration of charged
groups and could constitute highly selective
complexants. The BAHLM technology is based
on a combination of liquid membrane (LM) process
Bulk Liquid Membrane and dialysis (D) in the case of neutral hydrophilic
membranes or Donnan dialysis (DD) in the case of
Vladimir S. Kislik ion-exchange membranes used.
Campus Givat Ram, Casali Institute of Applied BOHLM and BAHLM technologies achieve
Chemistry, The Hebrew University of Jerusalem, the necessary transport and selectivity character-
Jerusalem, Israel istics to have potential for commercial applica-
tions. Applications of the BOHLM processes are
mainly in metal separation, wastewater treatment,
Bulk liquid membrane (BLM) consists of a bulk biotechnologies, drugs recovery-separation,
aqueous feed and receiving phases separated by a organic compounds, and gas separation. Selective
bulk organic, water-immiscible liquid phase. The separations of alkali, alkali earth, rare earth, heavy
feed and receiving phases may be separating from metal ions, precious metals, etc., are studied by
the LM by microporous supports or may be with- many authors using all above-described tech-
out supports (layered BLM). niques. Recovery and separation of carboxylic
Many technologies that were developed and and amino acids from fermentation broth have
tested in the last two decades have to be included been tested using layered BLM, rotating, creep-
in the BLM group. These are hybrid liquid mem- ing, spiral-type FLM, HFLM, HLM, and
brane (HLM), hollow-ber liquid membrane MHS-PV techniques of the BOHLM processes.
(HFLM), hollow-ber-contained liquid membrane Separation of ethylene, benzene, propanol, olen,
(HFCLM), pertraction, owing liquid membranes and aromatic amines from organic liquid mixtures
(FLM), membrane-based extraction and stripping, and of volatile organic compounds (VOC) and
multimembrane hybrid system (MHS), and mem- phenol from wastewater was investigated using a
brane contactor systems. All these systems are based rotating lm module, spiral-type FLM, and
on membrane-based nondispersive (as the means hollow-ber and layered LM techniques. High
for blocking the organic reagent from mixing with separation factors (>1,000) in pilot- and
the aqueous feed and strip solutions) selective industrial-scale experiments were found. In the
extraction coupled to permselective diffusion of last two decades, BOHLM techniques were inten-
solute-extractant complexes and selective stripping sively used in analytical chemistry for separation
of the solute in one continuous dynamic process and preconcentration of metals, organic acids,
(see Fig. 1). A great number of terms for similar organic, and pharmaceutical compounds.
bulk LM processes confuse the readers. The terms BAHLM is a new technology and few studies
vary by membrane type used (hollow ber, at are known up to date. In metal separations copper-
neutral, ion-exchange sheets) or by module design. cadmium recovery from chloride aqueous solu-
All abovementioned bulk LM processes with tions and cadmium, copper, and zinc separation
Butanol-Water Mixtures: Separation by Pervaporation 283
Bulk Liquid Membrane, Fig. 1 Schematic transport models of a bulk organic hybrid liquid membrane (BOHLM)
system with a, d hydrophobic membranes and b, e hydrophilic or ion-exchange membranes
from wet-process phosphoric acid are studied. chemical separations & wastewater treatment, 1st edn.
BAHLM systems were tested for the separation Elsevier, Amsterdam, pp 201276, Ch. 5
Kislik V (2010b) Bulk Aqueous Hybrid Liquid Membrane
of carboxylic acids such as lactic, citric, and acetic (BAHLM) processes with water-soluble carriers: appli-
or their anions. Continuous separation of different cation in chemical and biochemical separations. In:
isomeric mixtures of organic compounds has been Kislik V (ed) Liquid membranes principles and appli-
studied by means of a hollow-ber-contained liq- cations in chemical separations, 1st edn, Elsevier,
Ch. 6, pp 277326
uid membrane, HFCLM.
In recent years, integrated hybrid systems
incorporating two or more functions in one mod-
ule, for example, biotransformation and separa-
tion, become of great interest to researchers. The Butanol-Water Mixtures: Separation
BOHLM systems, integrating reaction, separa- by Pervaporation
tion, and concentration functions in one apparatus
(bioreactor) attracted great interest in the last few Wanqin Jin
years. Bioreactors combine the use of specic State Key Laboratory of Materials-Oriented
biocatalyst for the desired chemical reactions, Chemical Engineering, Nanjing University of
with repeated or continuous application of it Technology, Nanjing, China
under very specic conditions. Such techniques
were termed hybrid membrane reactors.
Because of the shortage of crude oil and the envi-
ronmental requirement, the utilization of renew-
References able resources for energy production, such as
using the potential acetone-butanol-ethanol
Kislik V (2010a) Bulk hybrid liquid membrane with organic
water-immiscible carriers: application to chemical, bio-
(ABE) fermentation process to produce butanol
chemical, pharmaceutical, and gas separations. In: Kislik as a biofuel, has been receiving increased atten-
V (ed) Liquid membranes principles and applications in tion in recent years. However, the fermentation
284 Butanol-Water Mixtures: Separation by Pervaporation
process suffers from severe inhibition due to the concentration of butanol is low. As an alternative,
high toxicity of butanol to microorganisms even at when the butanol concentration is high and the
low concentrations (46 gL1). In order to elim- water content is low, the hydrophilic membranes
inate the inhibition, in situ product recovery sys- should be used for the water removal, so that the
tems have been developed (Vane 2005). pervaporation process is economic for butanol
Pervaporation (PV) is considered to be a particu- production.
larly promising recovery technique, because of its The hydrophilic pervaporation polymeric
general advantages in energy saving and environ- membranes used for dehydration of butanol-
mental issue, particularly no medium ingredients water mixtures include: poly (vinyl alcohol)
removed from the fermentation broth and no (PVA) membrane, polyimide membrane, chitosan
harmful effects on microorganisms. membrane, sodium alginate membrane, polyelec-
The hydrophobic pervaporation membranes trolyte, etc. (Chapman et al. 2008). PVA is the
used for butanol recovery and concentrate most commonly used hydrophilic polymeric
include: polydimethylsiloxane (PDMS) and its membrane, but it often shows a lack of chemical,
modied copolymer and mixed matrix mem- mechanical, and thermal stability. In recent years,
branes, liquid membrane, poly[-1- the PI membrane began to be used for dehydration
(trimethylsilyl)-1-propyne] (PTMSP) membrane, because of its excellent mechanical and thermal
poly(ether block amide) (PEBA) membrane, stability. The main inorganic membranes used for
polypropylene (PP) membrane, polytetrauor- butanol dehydration are silica membrane (Verkerk
oethylene (PTFE) membrane, etc. (Vane 2005; et al. 2001) and A-type zeolite (NaA) membrane
Liu et al. 2011). PDMS, as the most typical hydro- (Morigami et al. 2001).
phobic polymeric membrane, is widely investi-
gated and proved to be stable and prospective for
industrial applications. The operating parameters References
of PV, including feed temperature (T), feed con-
centration (C), feed ow rate and permeate pres- Chapman PD, Oliveira T, Livingston AG, Li K (2008)
sure (P), etc., have signicant impacts on Membranes for the dehydration of solvents by
membrane performance. Take the PDMS/ceramic pervaporation. J Membr Sci 318:537
Liu GP, Hou D, Wei W (2011) Pervaporation separation of
membrane, for example, ux increased as temper- butanol-water mixtures using PDMS/ceramic compos-
ature increased, while the separation factor ite membrane. Chin J Chem Eng 19:15
decreased slightly. As the butanol concentration Morigami Y, Kondo M, Abe J, Kita H, Okamoto K (2001)
increased, the total ux increased while the sepa- The rst large-scale pervaporation plant using tubular-
type module with zeolite NaA membrane. Sep Purif
ration factor changed slightly. The total ux Technol 25:251260
increased with the ow rate, but the separation Vane LM (2005) A review of pervaporation for product
factor changed little (Liu et al. 2011). recovery from biomass fermentation processes. J Chem
Pervaporation using hydrophobic membranes Technol Biotechnol 80:603629
Verkerk AW, van Male P, Vorstman MAG, Keurentjes JTF
might not be enough for concentrating butanol (2001) Description of dehydration performance of
for direct reuse. Hydrophobic membranes were amorphous silica pervaporation membranes. J Membr
preferred for concentrating butanol when the Sci 193:227238
C
Cadmium Rejection by NF et al. 2009). Bulk liquid membranes (BLM) usu-

ally consist of tow aqueous feed and stripping
Cadmium Separation by NF phases, separated by a water-immiscible liquid
membrane phase with a relatively small surface
area. Emulsion liquid membranes have a very
high surface area per unit of volume and low
Cadmium Separation by Liquid thickness. The problem is that the vehicles have
Membranes to be produced before the process; they have to be
stable enough to have minimum leakage, but still
Naser Dalali not very stable so that they could be destroyed
Phase Separation and FIA Lab, Department of after the separation. In SLM, usually organic liq-
Chemistry, Faculty of Science, University of uid is imbedded in small pores of a polymer sup-
Zanjan, Zanjan, Iran port and is kept there by capillary forces. If the
organic liquid is immiscible with the aqueous feed
and strip streams, SLM can be used to separate the
Cadmium separation by liquid membranes: two aqueous phases.
Cadmium occurs as a minor component in the Hollow ber modules are usually more expen-
zinc ores. It has many common industrial applica- sive, but they offer much higher surface area per
tions. Despite wide applications, cadmium is unit of module volume. In the simplest case, the
known to be a toxic element, and the occupational organic solvent with a carrier is lling the pores of
exposure is by inhalation of fumes and ingestion the microporous walls of hollow tubes. Transport
of soluble cadmium compounds. This metal is an with liquid membranes is a kinetic process
environmental hazard. Therefore, due to its indus- and controlled by diffusion rate of analyte through
trial, chemical, or environmental importance, it the membrane. The membrane separation pro-
must be separated prior to determination or recov- cesses operate at ambient temperature and thus
ery from original matrices. Liquid membranes use less energy than thermal separations. The
broadly classied into bulk, emulsion, and basic concept for a metal cation to transport
supported liquid membranes (SLM) have found through a liquid membrane (regardless of mem-
wide applications for separation of metal species brane type) is its conversion to (i) an uncharged
(including cadmium) during the recent years hydrophobic species in the feed-membrane phase
(Dalali et al. 2012; Kumbasar 2008; Carletto interface and (ii) a hydrophilic species in the

DOI 10.1007/978-3-662-44324-8
286 Cadmium Separation by NF
membrane-stripping phase interface. Therefore,

the metal ion should either form an ion pair com- Cadmium Separation by NF
plex or a chelate in the presence of a hydrophobic
organic salt or chelating agent as carrier in the Z. V. P. Murthy
membrane phase to fulll such requirement. On Department of Chemical Engineering,
the other hand, the selectivity of transport or sep- S.V. National Institute of Technology, Surat,
aration efciency depends upon the stability of Gujarat, India
metal complexes (ionic or neutral) or salts formed
in feed, membrane, and stripping phases, which in
turn is affected by the type and concentration of Synonyms
salt in the tow aqueous phases and of carrier
(extractant) in membrane phase. The separation Cadmium rejection by NF; Cadmium wastewater
of cadmium by a BLM as a chloride complex via treatment by NF
its ion pair formation with trioctyl methyl ammo-
nium chloride (Aliquat 336) in benzene as organic Heavy metals present in wastewaters cause seri-
carrier into EDTA solution as stripping phase ous problems to human beings and other liv-
(Dalali 2012 was inuenced by concentration of ing organisms. Different methods are available
NaCl in the feed phase, concentration of carrier to separate heavy metals such as precipitation,
(Aliquat 336), and type and concentration of strip- coagulation, occulation, ion exchange, electro-
ping solution (EDTA). The rate of transport dialysis, reverse osmosis (RO), nanoltration,
depends upon the area and thickness of membrane etc. (Wang et al. 2007). Nanoltration (NF) is
or the interfaces. The larger the surface area and relatively new and it falls between RO and ultra-
the less the thickness of interface, the faster will be ltration (UF). NF can be operated at much lower
the mass transfer. The effective separation of cad- pressures than RO. NF can achieve high removal
mium is achieved in the above, e.g., within 4 h, of divalent and multivalent ions, but only moder-
while the similar separation by an ELM has been ate removal of monovalent ions is possible (Wang
achieved within 10 min time (Kumbasar 2008), et al. 2006). Nanoltration can be characterized
due to the larger surface area and the thinner by molecular weight cutoff (MWCO) between
interface in ELM. 200 and 1000 due to the pore diameter range of
0.52 nm, and also NF membranes are either
positively charged or negatively charged. NF per-
mits the retention of small molecules in the molec-
References ular mass range of 2002000 Da (Wang
et al. 2011). Cadmium is mainly toxic to the
Carletto JS, Luciano RM, Bedendo GC, Carasek E (2009)
Simple hollow ber renewal liquid membrane extrac- kidney and can also cause bone demineralization.
tion method for pre-concentration of Cd(II) in environ- Cadmium is one of the endocrine disruptor com-
mental samples and detection by ame atomic pounds (WHO 2013). The International Agency
absorption spectrometry. Anal Chim Acta 638:4550
for Research on Cancer (IARC) has classied
Dalali N, Yavarizadeh H, Agrawal YK (2012) Separation
of zinc and cadmium from nickel and cobalt by facili- cadmium as a human carcinogen (Group 1) on
tated transport through bulk liquid membrane using the basis of occupational studies (European Food
trioctyl methyl ammonium chloride as carrier, J Ind Safety Authority 2009). Cadmium is an odorless,
Eng Chem 18:10011005.
silver-white, blue-tinged malleable metal or
Dalali N (2012) Analytical applied chemistry, phase sepa-
ration and FIA Lab, department of chemistry, faculty of grayish-white powder with an oxidative state of
science, University of Zanjan , Zanjan 45195-313, +2, and it forms salts such as sulfate and nitrate. In
Iran; J Ind Eng Chem 18:10011005 the environment, cadmium is not usually present
Kumbasar RA (2008) Transport of cadmium ions from zinc
as a pure metal but as a mineral, such as cadmium
plant leach solutions through emulsion liquid mem-
brane containing Aliquat 336 as Carrier, Sep Purif oxide, cadmium chloride, cadmium sulfate, cad-
Tech 63:592599 mium sulde, etc. The chlorides and sulfates are
Cadmium Separation by NF 287
Cadmium Separation by NF, Table 1 Summary of the reflection coefficients for cadmium separation by NF-300
membrane
Feed concentration, mg/L Reection coefcient
CdSO4 NiSO4 CdCl2 NiCl2 CdCl2 NiSO4 CdCl2
5 0.9863 0.9951 0.8861 0.9051 0.8749 0.9961 ----
50 0.9842 0.9938 0.8567 0.8753 0.8455 0.9905 0.7556
100 0.9822 0.9924 0.8233 0.8419 0.8121 0.9864 0.7227 C
150 0.9801 0.9914 0.7921 0.8106 0.7809 0.9849 0.7197
200 0.9781 0.9897 0.7620 0.7805 0.7508 0.9834 0.6749
250 0.9760 0.9883 0.7177 0.7363 0.7065 0.9819 0.6644
most easily dissolved in water. Cadmium metal feed rate 15 L/min, and applied pressure
itself does not disappear from the environment, 420 atm. It can be seen from Table 1 that the
but it may change forms; hence, knowing the reection coefcient decreases with increase in
particular form of cadmium is very important feed solute concentration. It is observed that
when determining the risk of potential adverse the reection coefcient for each salt increases
health effects (US Department of Health and with co-ion valency. Recently, modeling of
Human Services, July 1999). When RO and NF cadmium removal by nanoltration using
are compared in separating cadmium from waste- genetic programming was reported (Okhovat
water, the results showed that RO was capable of and Mousavi 2012). A mechanism related to
treating wastewater with an initial concentration cadmium salt transport through a nanoltration
of 500 ppm and reducing the ions concentration to membrane was proposed by Garba et al. (2000),
about 3 ppm (99.4 % removal), whereas the aver- and it was found that in most cases convection
age removal efciency by NF was 97 % (Qdais plays the most important role.
and Moussa 2004). Cadmium separation from
wastewaters by NF assumes importance due to
its low-energy consumption and ease of operation. Cross-References
By using an NF-300 membrane system (Murthy
and Chaudhari 2008, 2009; Chaudhari and Nanoltration (Transport Phenomena)
Murthy 2010), the separation of cadmium from NF Membrane Characterization Methods
dilute aqueous solutions having different cad- NF Membranes
mium salts was reported. In wastewaters generally Nickel, Recovery of
cadmium is not alone but is present with other
metals such as nickel. Separation data were
reported for CdCl2-water system, CdCl2-NiCl2- References
water system, CdSO4-NiSO4-water system, and
Chaudhari LB, Murthy ZVP (2010) Separation of Cd and
CdCl2-NiSO4-water system (Murthy and
Ni from multicomponent aqueous solutions by
Chaudhari 2008, 2009; Chaudhari and Murthy nanoltration and characterization of membrane using
2010). The separation capability of an NF mem- IT model. J Hazard Mater 180:309315
brane can be related with its reection coefcient. European Food Safety Authority (2009) Scientic opinion
of the panel on contaminants in the food chain on a
If reection coefcient is zero, it means the sepa-
request from the European Commission on cadmium in
ration of a solute is zero, and similarly if it is one, food. EFSA J 980:1139
the separation is 100 %. Table 1 (Murthy and Garba Y, Taha S, Gondrexon N, Cabon J, Dorange
Chaudhari 2008, 2009; Chaudhari and Murthy G (2000) Mechanisms involved in cadmium salts trans-
port through a nanoltration membrane: characteriza-
2010) shows the separation capability of NF-300
tion and distribution. J Membr Sci 168:135141
membrane at different feed concentrations. All the Murthy ZVP, Chaudhari LB (2008) Separation of cadmium
experiments are performed at a pH of 5.0 0.1, ions and estimation of membrane transport parameters
288 Cadmium Wastewater Treatment by NF
of a nanoltration membrane. Indian J Chem Technol found in the earths crust at concentrations rang-
15:107112 ing from 0.1 to 5 ppm, primarily as sulde min-
Murthy ZVP, Chaudhari LB (2009) Separation of binary
heavy metals from aqueous solutions by nanoltration erals in association with zinc ores, zinc-bearing
and characterization of the membrane using Spiegler- lead ores, and/or complex copper-lead-zinc ores.
Kedem model. Chem Eng J 150:181187 The salts of cadmium chloride and cadmium sul-
Okhovat A, Mousavi SM (2012) Modeling of arsenic, fate are soluble in water. Most of the cadmium is
chromium and cadmium removal by nanoltration pro-
cess using genetic programming. Appl Soft Comput extracted through mining. Cadmium is mainly
12:793799 used in batteries (83 %), pigments (8 %), coatings
Qdais HA, Moussa H (2004) Removal of heavy metals and platings (7 %), stabilizers for plastics (1.2 %),
from wastewater by membrane processes: a compara- nonferrous alloys, photovoltaic devices, and other
tive study. Desalination 164:105110
U.S. Department of Health and Human Services uses (0.8 %). Hence, naturally cadmium is mainly
(1999) Toxicological prole for cadmium, U.S. Depart- recovered from used batteries (Toxicological pro-
ment of Health and Human Services, Public Health le for cadmium 2012). Cadmium is recovered
Service, Agency for Toxic Substances and Disease from spent NiCd batteries, copper-cadmium
Registry
Wang LK, Hung YT, Shammas NK (2006) Handbook of alloy scrap, some complex nonferrous alloy
environmental engineering, volume 4: advanced phys- scrap, and cadmium-containing dust from electric
icochemical treatment processes. Humana Press, arc furnaces (EAF). In 2006, the US steel industry
Totowa generated about 0.7 million tons of EAF dust,
Wang LK, Hung YT, Shammas NK (2007) Handbook of
environmental engineering, volume 5: advanced phys- typically containing 0.0030.07 % cadmium
icochemical treatment technologies. Humana Press, (Mineral Commodity Summaries 2007).
Totowa Recently, membrane separations are being used
Wang LK, Chen JP, Hung YT, Shammas NK (2011) Hand- to recover/extract the cadmium from different
book of environmental engineering, volume 13: mem-
brane and desalination technologies. Humana Press, waste streams. Cadmium present in the dust
Totowa form was most efciently collected by membrane
WHO (World Health Organization) (2013) State of the lters (United Nations Environment Programme:
science of endocrine disrupting chemicals 2012. Chemicals Branch 2010). Cadmium, along with
http://www.who.int/ceh/publications/endocrine/en/
other metal ions, was recovered from acidic leach
solutions by emulsion liquid membrane (ELM)
using trioctylamine (TOA) as extractant
(Kumbasar 2009). The results showed
Cadmium Wastewater Treatment (Kumbasar 2009) that up to 98 % cadmium recov-
by NF ery was possible after 10 min contact time by
using ELM (Table 1). Jha et al. (2012) and
Cadmium Separation by NF Safarzadeh et al. (2007) show the possibility of
extraction and separation of cadmium from differ-
ent solutions containing other metallic ions. The
cadmium extraction efciencies of three phospho-
Cadmium, Recovery of ric acid derivatives such as bis-(2-ethylhexyl)
phosphoric acid (D2EHPA), 2-ethylhexyl
Z. V. P. Murthy phosphonic acid mono-2-ethylhexyl ester
Department of Chemical Engineering, (PC-88A), and bis-(2,4,4-trimethyl pentyl)
S.V. National Institute of Technology, Surat, phosphinic acid (Cyanex 272) have been reported
Gujarat, India using supported liquid membrane (SLM) (Fig. 1)
(Parhi et al. 2009). Also, several researchers
reported the recovery of cadmium by SLM with
Recovery of cadmium from aqueous efuents different extractants (He et al. 2007). Polymer-
by membrane operations. Cadmium is a metal enhanced ultraltration (UF) and nanoltration
Cadmium, Recovery of 289
Cadmium, Recovery of, Table 1 Effect of extractant concentration on the extraction rate of Cd (Span 80: 6 %;
kerosene: 8692 %; stripping solution: 18 mL NH3 solution of 6 M; mixing speed: 250 rpm; Cd concentration of feed
solution: 100 mg/L; HCl concentration of feed solution: 1.0 M; phase ratio: 3/5; feed solution: 250 mL) (Reproduced with
permission from Elsevier)
C/C0
Extractant concentration, % 0 min 2 min 4 min 6 min 10 min
2 1 0.121 0.03 0.01 0.01
C
4 1 0.101 0.01 0.01 0.01
6 1 0.09 0.01 0.01 0.01
8 1 0.083 0.01 0.01 0.01
Cadmium, Recovery of, 100

Fig. 1 Percentage
extraction of Cd at different
times with D2EHPA, 80
PC-88A, and Cyanex 272.
% Cd extraction
Experimental conditions:
60
pH 7.5, 600 mol/m3
extractant in membrane
phase, 900 mol/m3 H2SO4 40
in strip solution, and CYANEX-272
120 mL/min ow rate PC88A
(Reproduced with 20 D2EHPA
permission from Elsevier)
0
0 1 2 3 4
Time X 103, s
(NF) are also being used to recover cadmium hybrid liquid membrane (SDHLM) containing double
with encouraging results (Caizares et al. 2008; carrier. J Membr Sci 305:3647
Jha MK, Kumar V, Jeong J, Lee J (2012) Review on
Ennigrou et al. 2009; Gonzlez-Muoza et al. 2006). solvent extraction of cadmium from various solutions.
Hydrometallurgy 111112:19
Kumbasar RA (2009) Extraction and concentration study
of cadmium from zinc plant leach solutions by emul-
References sion liquid membrane using trioctylamine as extractant.
Hydrometallurgy 95:290296
Caizares P, Prez A, Camarillo R, Mazarro R (2008) Mineral Commodity Summaries (2007) U.S. Department
Simultaneous recovery of cadmium and lead from of the Interior, U.S. Geological Survey
aqueous efuents by a semi-continuous laboratory- Parhi PK, Das NN, Sarangi K (2009) Extraction of cad-
scale polymer enhanced ultraltration process. mium from dilute solution using supported liquid mem-
J Membr Sci 320:520527 brane. J Hazard Mater 172:773779
Ennigrou DJ, Gzara L, Romdhane MRB, Dhahbi M (2009) Safarzadeh MS, Bafghi MS, Moradkhani D, Ilkhchi MO
Retention of cadmium ions from aqueous solutions by (2007) A review on hydrometallurgical extraction and
poly(ammonium acrylate) enhanced ultraltration. recovery of cadmium from various resources. Miner
Chem Eng J 155:138143 Eng 20:211220
Gonzlez-Muoza MJ, Rodrguez MA, Luquea S, Toxicological prole for cadmium, U.S. Department of
lvareza JR (2006) Recovery of heavy metals from Health and Human Services Public Health Service,
metal industry waste waters by chemical precipitation Agency for Toxic Substances and Disease Registry
and nanoltration. Desalination 200:742744 Sept 2012, Atlanta
He D, Gu S, Ma M (2007) Simultaneous removal and United Nations Environment Programme: Chemicals
recovery of cadmium (II) and CN from simulated Branch, DTIE, Final Review of Scientic Information
electroplating rinse wastewater by a strip dispersion on Cadmium, Version of Dec 2010
290 Capillary
Capillary Capillary Flow
Andras Koris Andras Koris

Corvinus University of Budapest, Budapest, Corvinus University of Budapest, Budapest,
Hungary Hungary
The word capillary originated from the Latin Capillary ow means the movement of the uid
adjective capillaris, which means pertaining to in the internal gaps and on the surface of the solid
the hair. Possibly, the scientic phenomenon was material, where the driving force is the molecular
rst observed between contiguous hairs, for interaction between the solid and uid material
example, within a paintbrush. (see Capillary Force). If the uid is incom-
In medicine and biology, capillary is the pressible and Newtonian, the ow is laminar
smallest of a bodys blood vessel and is part of through a pipe of constant circular cross section
the microcirculation. The word capillary, in the that is substantially longer than its diameter, and
nonmedical sense, means narrow tube. since there is no acceleration of uid in the pipe,
In membrane science the noun capillary is the total ow rate (Q) could be estimated
both used to name the inner spaces of a porous according to Poiseuilles law (Briant et al. 1989):
media and to specify a typical membrane (see
Capillary Membrane). The classication of d4 p DP
porous membrane lter types is mainly based Q
128 m L
on the nominal diameter of the capillaries in the
membrane. Articial capillaries in porous media where d represents capillary diameter, DP is the
are produced by interfacial polymerization, or total pressure drop along the channel, m is the uid
the natural porosity in solid materials is adjusted viscosity, and L represents the capillary length.
to the desirable nominal diameter. The shape and For velocity calculation, let us consider a lam-
size of the pores in the membrane are very inar horizontal ow through a straight circular
diverse. channel of length L and a radius r. The mean
The pore theory of membrane transport ow velocity u through the channel is given by
through the capillaries considers that both con- (Landau and Lifshitz 1987):
vection and diffusion contribute to solute trans-
port across the porous membrane, the two r 2 DP
processes being impeded by steric hindrance u
8m L
at the entrance of the pores and by frictional
forces within the pores; the same steric hin- where DP is the total pressure drop along the
drance and frictional resistance terms were channel and m is the uid viscosity.
used in the convection and diffusion terms of In the general case of two immiscible uids:
the solute transport equation (Pappenheimer
et al. 1951). r 2 DP Pc
u
8meff L
References where
Pappenheimer JR, Renkin EM, Borrero LM (1951) Filtra-
x Lx
tion, diffusion and molecular sieving through periph- meff mw ma
eral capillary membranes. Am J Physiol 167:13 L L
Capillary Force 291
is the effective viscosity; mw and ma are viscosities where DAB represents the diffusivity of salt in
for phase w and phase a uids, respectively; x is water, e is the porosity, and k is a correction factor
the location of the interface measured from the for the distance (when it is nonlinear; general
inlet; Pc is the capillary pressure across the inter- value = 1.22.5).
face; and a0 is a constant associated with the The diffusion of a liquid in capillaries has
channel cross-sectional shape. three types: Knudsen diffusion, molecular
Assuming that the uid system is in static (Ficks) diffusion, and transition region diffusion C
equilibrium, the capillary pressure at an interface (Geankoplis 1972).
in a circular channel is (Young-Laplace eq.): The power-law is an improvement to the
Poiseuille's equation by generalizing the ow to
4g cos y include non-Newtonian effect. The beauty of this
Pc
dc model is its simplycity. However there are other
aspects of the ow properties of uids which this
where g represents the interfacial tension, y is the model fails to examine, one of them being the
contact angle, and dm represents the capillary presence of yield stress (Mazumdar 1992)
diameter.
Since the determination of the number of active
capillaries during membrane ltration is very dif-
cult, statistical approach from Darcy can be References
applied basically. Darcys law is a simple propor-
Briant J, Guy Parc JD (1989) Rheological properties of
tional relationship between the instantaneous dis- Lubricants, ch. 3., p. 69-71
charge rate through a porous medium, the Geankoplis CJ (1972) Mass transport phenomena.
viscosity of the uid, and the pressure drop over Rinehart and Winston, Holt
a given distance: Landau LD, Lifshitz EM (1987) Fluid mechanics, 2nd edn.
Pergamon Press, Oxford
Mazumdar J (1992) Biouid Mechanics, p. 86-87, World
k A DP
Q Scientic
mL
where Q is overall ow rate, k represents perme-

ability of the medium, A is the cross-sectional area
Capillary Force
to ow, DP is the pressure drop, m represents
viscosity, and L is the length over which the
Andras Koris
pressure drop is taking place.
Corvinus University of Budapest, Budapest,
Due to the concentration polarization phenom-
Hungary
enon, gel layer forming, membrane fouling, and
pore blocking, the global model was further
developed for different membrane techniques,
Capillary force is the force which generates move-
see, e.g., resistances-in-series model or solution-
ment of a liquid along the surface of a solid caused
diffusion model.
by the attraction of molecules of the liquid to the
The diffusion of gasses, liquids, and solids in
molecules of the solid. For example, molecules of
solids is quite important in mass transfer opera-
water are naturally attracted to each other and
tions. The diffusion is greatly affected by the size
form temporary hydrogen bonds with each other;
and shape of pores and capillaries. For example, in
their attraction for each other on the surface of a
porous solids, the effective diffusion of salts
liquid, for example, gives rise to surface tension.
(A) in water (B) is described as:
But they are also attracted in a similar way to other
molecules, called hydrophilic molecules, such as
D_f_A, eff g DAB those in the sides of a narrow glass tube inserted
k2
292 Capillary Membrane
Capillary Force, Fig. 1 Contact of three uids
into water. These attractive forces can draw water

up against the force of gravity to a certain degree.
Where three different uid interfaces are
contacting each other along one line, equilibrium
exists only when the vector polygon of the
stresses is closed. This requirement determines
the angle between the edges of disjunctive sur-
faces (Pattantys 1961) (Fig. 1).
Sometimes |C13| > |C12| + |C23| occurs.
Because vector C13 is always larger, equilibrium
is not possible; therefore, uid no. 2 is perma-
nently moving on the no. 1 and 3 uid interface. Capillary Force, Fig. 2 Contact when one solid phase is
present next to two uids
For example, mineral oil unfolds the water surface
even to molecule thick layer when the available
area is sufcient.
One angle of the vector triangle is given in
pressure difference to the shape of the surface or
cases where instead one of the liquids a solid
wall, and it is fundamentally important in the
material presents. In Fig. 2 this angle is 180 .
study of static capillary surfaces. It is a statement
The requirement of equilibrium: C13 = C23 +
of normal stress balance for static uids meeting at
C12*cosa.
an interface, where the interface is treated as a
In case of solid particle or material, |C13| >
surface (scpacp).
|C12| + |C23| can also exist sometimes, and
no. 2 liquid could totally cover the solid surface
(1). This movement is usually limited by impuri-
References
ties on the solid surface. For mercuryairglass
system, |C13| < |C23| and a > 90 . Slyi I (ed) (1961) Pattantys gpsz s villamosmrnkk
The YoungLaplace equation is a nonlinear kziknyve, 2. Alaptudomnyok Anyagismeret.
partial differential equation that describes the cap- Mszaki Knyvkiad
illary pressure difference or capillary force
sustained across the interface between two static
uids, due to the phenomenon of surface tension
or wall tension, although usage on the latter is Capillary Membrane
only applicable if assuming that the wall is very
thin. The YoungLaplace equation relates the Capillary Membrane Module
Capillary Membranes 293
concept applied, asymmetric capillaries are used

Capillary Membrane Module with their skin on the inside or on the outside
(Mulder 1997).
Joerg Balster
Application Technology, Evonik Fibres GmbH,
Schoering, Austria References
C
Mulder M (1997) Basic principles of membrane technol-
Synonyms ogy, 2nd edn. Kluwer, Dordrecht
Scholz M, Wessling M, Balster J (2011) Design of mem-
brane modules for gas separations. In: Drioli E
Capillary membrane; Capillary module (ed) Membrane engineering for the treatment of
gases: gas-separation problems with membranes. RSC
A number of different membrane modules based Publishing, p 5
on capillaries are possible, and all are based on the
same membrane geometry with different dimen-
sions as shown in Fig. 1.
The capillary membrane module consists of a Capillary Membranes
large number of capillaries assembled together in
a pressure vessel similar as a hollow ber module Andras Koris
(Mulder 1997). The free ends of the capillaries are Corvinus University of Budapest, Budapest,
potted with agents such as epoxy resins, polyure- Hungary
thanes, or silicone rubber. The capillaries are self-
supporting. Two types of module arrangements
can be distinguished (Mulder 1997): Articial capillary membranes are one type of the
tubular-shaped membrane lters. With capillary
Where the feed passes through the bore of the membranes, the membrane serves as a selective
capillary barrier, which is sufciently strong to resist ltra-
Where the feed enters the module on the shell tion pressure. Because of this, the ow through
side capillary membranes can be both inside out and
outside in.
The choice between the two concepts is mainly The diameter of capillary membranes is much
based on the application where parameters such as smaller than that of tubular membranes, namely,
pressure, pressure drop, type of membrane 0.55 mm. Because of the smaller diameter, the
available, etc. are important. Depending on the chances of plugging are much higher with a
Capillary Membrane
Module, Fig. 1 Available outer D ~ 0.08-0.8 mm
types of capillary Hollow fiber membranes
membranes adapted from inner D ~ 0.04-0.5 mm
Scholz et al. (2011)
inner D ~ 0.5-5 mm
Capillary membranes
outer D ~ 0.8-7 mm
Tubular membranes outer D ~ 5 -25 mm

294 Capillary Membranes
benets. For example, supported zeolite mem-

branes have been synthesized under microwave
heating in order to reduce synthesis time in the
work of Sebastian et al. (2010) to prevent support
dissolution and to reproducibly obtain a thin
defect-free zeolite layer. The MFI-type zeolite
membranes were synthesized on ceramic capil-
laries, with a high membrane surface area-to-
volume ratio (>1,000 m2 m3), which is by far
higher than that of classical tubular supports
Capillary Membranes, Fig. 1 Ceramic capillary mem-
brane (Source: www.kerafol.com)
(<500 m2 m3).
In a study of Zhu et al. (2010), alumina
microltration membranes were prepared on the
capillary membrane. A benet is that the packing inner surface of alumina capillary support (outer
density is much greater. The rst capillary mem- diameter, 4 mm; inner diameter, 2.5 mm) by a
brane was fabricated from polymers, but nowadays dip-coating method. Scanning electron micros-
the technology development enables the produc- copy (SEM) observation, gas bubble pressure
tion of inorganic capillary modules (Fig. 1). (GBP) method, and membrane permeation test
were carried out to evaluate membrane perfor-
mance. Two major effects in preparation of
Description of a Capillary Membrane crack-free membrane, capillary ltration, and
Module lm coating upon the thin support were studied.
Bonyadi and Mackley (2012) reported the
A complete description of a membrane module experimental results for the processing of micro-
requires the simultaneous solution of local trans- capillary lm (MCF) membranes. MCFs are lms
port equations that describe the ow and transport with embedded multiple hollow capillaries and
conditions. In a capillary membrane module with can be considered as a hybrid geometry between
a permeable membrane wall, three regions of ow at sheet and hollow bers. Compared to the
should be considered: ow in the lumen, ow conventional membrane geometries, MCFs
within the membrane matrix, and ow in the potentially provide a higher surface area per unit
extracapillary space. The pressure distribution in volume, better mechanical strength, ease of han-
the membrane is determined by application of the dling, and more efcient module fabrication.
overall balance of linear momentum. The utiliza- MCF membranes were fabricated out of ethylene
tion of this conservation principle is complicated vinyl alcohol (EVOH) copolymer through a solu-
by the fact that an external force must be applied tion extrusion followed by a nonsolvent induced
to keep the polymer membrane stationary. In order phase separation (NIPS) process.
to describe the behavior of a membrane module, Solvent resistant nanoltration (SRNF) is a
three submodels are required: two that describe membrane separation process allowing for an ef-
the transport on either side of the membrane and a cient separation of small molecules of
third model that characterizes the separation prop- 2001,000 g mol1 from organic solvents. The
erties of the membrane and any porous support application of SRNF in industry applications is
material (Nagy 2012). currently hindered by a limited choice of SRNF
membranes and congurations. In the work of
Dutczak et al. (2011), SRNF composite capillary
Membrane Development membranes made of an a-alumina support and a
selective poly(dimethylsiloxane) (PDMS) top
Besides polymeric and ceramic capillary mem- layer were prepared and studied. We combine
branes, other types are also developing due to its the advantages of a ceramic support such as high
Capillary Membranes 295
mechanical, thermal, and chemical stability with Suitability of capillary modules to the decolor-
very good separation properties of the PDMS ization of both synthetic and actual dye solutions
coating. by ultraltration was investigated as well. The
process involved capillary membranes made of
polysulfone and modied polysulfone. Mem-
Possible Applications brane modules (UFTA PS10, UFTA PS30, and
UFTA PSA50) of various molecular cutoffs were C
The number of practical applications of capillary applied (Majewska-Nowak et al. 1996).
membranes is also expanding. A new combined Investigations concerning the application
method is reported to localize the sites of enzyme of ceramic and capillary microltration and
immobilization and to determine its catalytic ultraltration membranes for the purpose of
activity on a polymeric capillary membrane treatment of natural water were carried out by
reactor activity (Mazzuca et al. (2006)). Bodzek and Konieczny 1998. The effectiveness
b-Glucosidase from olive fruit was selected as of ultra- and microltration was assessed by vol-
enzyme model because of its suitable relevance umetric measurements of the permeate ux as
in the industrial processing of foods, in biotech- well as by microbiological and physicochemical
nology, and in pharmaceuticals and for its activity analyses.
against the synthetic substrate 5-bromo-4-chloro- Analysis of oxygen permeation uxes through
3-indolyl-b-d-glucopyranoside which develops Ba0.5Sr0.5Co0.8Fe0.2O3d (BSCF) capillary
an insoluble dyed product. The enzyme was phys- membranes, fabricated via a phase inversion spin-
ically immobilized within 30 kDa cutoff capillary ning technique using polysulfone as binder,
polysulfone membranes, and results obtained by showed a signicant limiting role of the surface
means of a polyclonal antibody against oxygen exchange kinetics. Within the studied
b-glucosidase and the synthetic substrate clearly temperature and oxygen partial pressure ranges,
showed a coherent localization of the immobili- the activation of core and shell sides of the BSCF
zation enzyme sites and its activity. capillary with praseodymium oxide (PrOx)
Commercial at sheet membranes were com- resulted in an increase in permeation rate of
pared to capillary ones (Van der Bruggen about 300 % (Kovalevsky et al. 2011).
et al. 2003). For the capillary membrane, the Capillary membrane diffusion scrubber
water ux could, however, easily be increased (CMDS) was scaled down to match with capillary
and maintained at a stable level by a combination electrophoresis (CE), a quick separation tech-
of forward ushing and airushing, which is not nique that requires nothing more than some
possible with the at sheet membranes. Further- nanoliters of sample and, when combined with
more, the water permeability for the capillary capacitively coupled contactless conductometric
membrane was 315 times higher than for the detection (C4D), is particularly favorable for
commercial at sheet membranes, which leads to ionic species that do not absorb in the UVvis
lower operating pressure and a correspondingly region, like the target analytes formaldehyde,
lower energy consumption. For the capillary formic acid, acetic acid, and ammonium (Coelho
membrane, rejections of organic and inorganic et al. 2010).
compounds were satisfactory to reach for COD Ceramic tube and ceramic capillary mem-
and conductivity standards in one step starting branes and their effect on permeate ux, oil reten-
from the surface water. Rejections for most at tion, and specic energy consumption were
sheet membranes were comparable to the rejec- compared in the work of Gspr et al. (2011).
tions obtained with the capillary membrane, but The results, obtained with a stable oil-in-water
the rejection of ions was usually higher, except for microemulsion as feed, demonstrated that the
the N30F and NF PES 10 membranes. Low ion use of novel ceramic capillary membranes at opti-
rejections are advantageous for drinking water mal operating cross-ow velocity and transmem-
production because demineralization is avoided. brane pressure is reasonable.
296 Capillary Module
References
Capillary Tubes
Bodzek M, Konieczny K (1998) Comparison of ceramic
and capillary membranes in the treatment of natural
Andras Koris
water by means of ultraltration and microltration.
Desalination 119(13):191198 Corvinus University of Budapest, Budapest,
Bonyadi S, Mackley M (2012) The development of novel Hungary
micro-capillary lm membranes. J Membr Sci
389:137147
Coelho LHG, Melchert WR, Rocha FR, Rocha FRP, Gutz
IGR (2010) Versatile microanalytical system with porous Def. (Engineering) A tube sufciently ne so that
polypropylene capillary membrane for calibration gas capillary attraction of a liquid into the tube is
generation and trace gaseous pollutants sampling applied signicant. (General Physics) A glass tube with
to the analysis of formaldehyde, formic acid, acetic acid
a ne bore and thick walls, used in thermometers,
and ammonia in outdoor air. Talanta 83(1):8492
Dutczak SM, Luiten-Olieman MWJ, Zwijnenberg HJ, etc. (Medical) Glass capillary tubes are used for
Bolhuis-Versteeg LAM, Winnubst L, Hempenius MA, the collection of blood in a variety of healthcare
Benes NE, (. . .), Stamatialis D (2011) Composite cap- settings, including hospitals, ambulatory care
illary membrane for solvent resistant nanoltration.
facilities, physicians ofces, blood donation
J Membr Sci 372(12):182190
Endre N (2012) 7 Transport of uid phase in a capillary facilities, and blood testing centers.
membrane. Basic equations of the mass transport
through a membrane layer. p 177192
Gaspar I, Koris A, Bertalan Z, Vatai G (2011) Comparison
of ceramic capillary membrane and ceramic tube mem-
Basic Description
brane with static mixer inside. Chem Pap 65(5):596602
Kovalevsky A, Buysse C, Snijkers F, Buekenhoudt A, A capillary tube works by its being tension pulling
Luyten J, Kretzschmar J, Lenaerts S (2011) Oxygen between the circumference and the inner wall of
exchange-limited transport and surface activation of
the tube. The restriction and metering of the liquid
Ba0.5Sr0.5Co0.8Fe0.2O3-d capillary membranes.
J Membr Sci 368(12):223232 will maintain pressure differential. See more at
Majewska-Nowak K, Kabsch-Korbutowicz M, Winnicki Capillary Force.
T (1996) Capillary membranes for separation of dye Capillary tubes can be made of different mate-
particles. Desalination 105(12):91103
rials like plastic, glass, stainless steel, and
Mazzuca S, Giorno L, Spadafora A, Mazzei R, Drioli
E (2006) Immunolocalization of b-glucosidase ceramics.
immobilized within polysulphone capillary membrane Wall thickness of microhematocrit capillary
and evaluation of its activity in situ. J Membr Sci tubes is 0.01 mm.
285(12):152158
Sebastian V, Mallada R, Coronas J, Julbe A, Terpstra RA,
Dirrix RWJ (2010) Microwave-assisted hydrothermal
rapid synthesis of capillary MFI-type zeolite-ceramic Applications
membranes for pervaporation application. J Membr Sci
355(12):2835
It can be a device used to measure kinematic
Van der Bruggen B, Hawrijk I, Cornelissen E,
Vandecasteele C (2003) Direct nanoltration of surface viscosity. Kinematic viscosity is the time it takes
water using capillary membranes: comparison with at for a xed amount of uid to move through a
sheet membranes. Sep Purif Technol 31(2):193201 measured length of the capillary tube with gravity
Zhu J, Fan Y, Xu N (2010) Preparation and characterization
alone.
of alumina membranes on capillary supports: effect of
lm-coating on crack-free membrane preparation. Chin In Membrane Science once capillary tubes
J Chem Eng 18(3):377383 present in the lter media, they are called
pores. Symmetric pore structure is usually
formed by pressing and sintering, extruding and
stretching, irradiation and etching, template
Capillary Module leaching, or phase inversion. On the other hand,
capillaries can be found in capillary membranes
Capillary Membrane Module as well.
Carbon Capture and Storage (CCS) 297
Capillary tubes are used in analytical applica- This is primarily a separation of CO2 from N2, as
tions, such as chromatography, gas and liquid lines, well as water and other minor components, such
and measurement components in remote thermom- as CO, SOx, and NOx. The separation process
eter systems (with or without compensation). occurs at lower pressures, low CO2 partial pres-
An indispensible capillary tube is used for cal- sures, and often around ambient temperature.
culating blood clotting time or determining melt- Oxy-fuel combustion is pure oxygen ring of the
ing points. combustion process. This excludes N2 from the C
Also a capillary tube is used in refrigeration combustion zone and produces a ue gas
and air-conditioning systems. It is a copper tube consisting of CO2 and H2O. Carbon capture is
that has a small internal diameter. A capillary tube achieved through water removal. However, the
functions by dropping pressure due to its small very high ame temperature often dictates ue
internal diameter. The refrigerant leaves the con- gas recycling, and critically this strategy requires
denser and enters the capillary tube and when it the separation of high purity O2 from air, which is
reaches the capillary tube falls in pressure and commonly achieved through cryogenic methods.
keeps the pressure constant. It limits the maxi- Precombustion capture is the gasication of the
mum amount of the refrigerant that can be charged carbon-based fuel into synthetic gas, also called
in the refrigeration system. syngas, consisting of CO and H2. This syngas is
an important industrial feed gas that is used in a
wide range of processes, such as Fischer-Tropsch
synthesis. For power generation, H2 production is
Carbon Capture and Storage (CCS) often maximized through the water gas shift reac-
tion, resulting in high-pressure gas mixtures of
Colin A. Scholes CO2, CO, and H2. Separation of CO2 is achieved
Department of Chemical and Biomolecular before H2 undergoes combustion. This separation
Engineering, University of Melbourne, Parkville, strategy takes advantage of the high pressure and
VIC, Australia CO2 partial pressure before combustion.
Carbon capture has also been applied to other
industrial applications, such as natural gas sweet-
Carbon capture and storage (CCS) is a proposed ening, for the removal of acidic gases to meet
strategy to separate carbon dioxide from industrial pipeline specications, as well as in natural gas
gases and store that carbon on a geological time- reforming for ammonia production.
scale. This is a response to the perceived threat A range of separation technologies have been
industrial carbon dioxide emissions have in pro- applied to carbon capture, including solvent sepa-
ducing anthropogenic climate change (IPCC ration, membrane gas separation, adsorption tech-
2007). CCS is divided into three distinct stages, nology, cryogenic separation, and mineral
the capture of the carbon dioxide from the emis- carbonation. The current commercial technology
sion source, the compression and transportation of of choice is physical and chemical solvent technol-
CO2 to the storage site, and the nal storage of ogy, with membrane gas separation having been
CO2 for long periods of time (IPCC 2005). commercial proven in natural gas sweetening.
The capturing of CO2 is predominately associ- Transportation of captured CO2 to the storage
ated with large stationary point sources and in site is undertaken through shipping (low temper-
particular coal-red power stations. The strategies ature and low pressure) or pipelines (high pressure
to capture carbon emissions from these sources and ambient temperature). The method of trans-
can be classied into three distinct approaches: portation is dependent on distance and geography
post-combustion capture, oxy-fuel combustion, between the source and storage sites.
and precombustion capture (Table 1). Storage of carbon for extended periods of time
Post-combustion capture is the separation of can be achieved through geosequestration, deep
CO2 from ue gases after the combustion process. ocean storage, or mineralization. Geosequestration
298 Carbon Clothes in Fuel Cells
Carbon Capture and Storage (CCS), Table 1 Carbon capture strategy conditions and gas compositions
Post-combustion Oxy-fuel combustion Precombustion
Pressure (bar) 1 1 30
Composition (mol. frac.)
CO2 0.11 0.87 0.40
N2 0.62
H2 0.55
O2 0.04 0.05
H2O 0.23 0.08 0.02
CO 0.03
is the injection of supercritical CO2 into geological component in the fuel cell which performs
formations, generally sedimentary basin, over a several functions such as to distribute reactant
kilometer underground. Importantly, the geological gases, water management, and electrical con-
formation must contain a suitable reservoir rock, duction. Carbon cloth is one of the two pre-
with high permeability and porosity for CO2, as ferred materials (other one is carbon paper) for
well as a seal rock above the reservoir that forms GDLs in polymer electrolyte membrane fuel
the necessary geological trap. Possible geological cells (PEMFC) and direct methanol fuel cells
sites for CO2 geosequestration include disused oil (DMFC).
and gas elds, saline aquifers, as well as deep coal Carbon cloth is commercially available in
seams. Currently, CO2 is geosequestrated commer- thickness of about 300450 mm. Carbon cloth
cially as part of enhanced oil recovery procedures. is made by weaving carbon bers and therefore
is more exible and resilient than carbon paper-
based GDL layers. Because carbon cloth is elec-
References trically conductive, it also provides pathway for
current collection. Therefore carbon cloth GDL
IPCC (2005) IPCC special report on carbon dioxide
capture and storage. Prepared by working group III of acts as electrical connection between the bipolar
the Intergovernmental Panel on Climate Change. Cam- plate and the catalyst layer (reaction layer)
bridge University Press, Cambridge, UK where the electrons are liberated/consumed in
IPCC (2007) IPCC the physical science bases. Contribu-
the electrochemical reactions involved inside
tion of working group 1 to the fourth assessment report
of the Intergovernmental Panel on Climate Change. the fuel cell. Carbon cloth offers good porosity
Cambridge University Press, Cambridge, UK to allow easy passage of reactants (i.e., gas) to
diffuse to the catalyst layer for more uniform
reaction. The good porosity of carbon cloth also
helps in the removal of products of the fuel cell
Carbon Clothes in Fuel Cells reaction, i.e., water and oxygen depleted air,
while the heat is transferred away from the
Rajnish Kaur Calay reaction site to other components inside the
Institute for Building, Energy and Material cell. It is important to prevent the ooding of
Technology, University of Troms, Navrik, the GDL; therefore carbon cloth is made wet
Norway proof typically with a hydrophobic coating
Energy Technology Research Group, Narvik (e.g., PTFE or Teon). Carbon cloths are often
University College, Lodve Langes gt 2, Navrik, highly compressible, reaching up to 4050 %
Norway compression when in a stack. This compression
can signicantly change their electrical and gas
permeation properties and therefore careful con-
Carbon cloth is used as gas diffusion layers sideration is required when using carbon cloth
(GDL) in fuel cells. The GDL is an essential as a GDL.
Carbon Membrane 299
The second generation of CO2-selective mem-

Carbon Dioxide (CO2) Separation by branes are based on polyimides. These types of
Membranes membranes generally display higher permeabil-
ities and selectivities for CO2 than cellulose
Colin A. Scholes acetate-based membranes. These have a CO2
Department of Chemical and Biomolecular permeance range of 10200 GPU and a CO2/
Engineering, University of Melbourne, Parkville, CH4 selectivity range of 530. Polyimide mod- C
VIC, Australia ules are predominately hollow ber design and
have been commercialized by UBE Industries
and Air Liquide. Other CO2-selective membranes
Carbon dioxide (CO2) separation by membranes has that have been commercialized are based on
been proposed for a range of carbon capture and peruoropolymers and are currently only applied
storage strategies. To date, CO2-selective mem- in niche processes (Merkel et al. 2006).
branes have only been commercialized for natural A wide range of other nonporous polymeric
gas sweetening, though a range of membrane pilot membrane materials have been reported in the liter-
plants are currently assessing the feasibility for CO2 ature for CO2 separation, such as polybenzoxazoles,
separation from syngas and ue gas. In all of these polysulfones, polycarbonates, and polypyrroles
applications, the CO2-selective membrane is based (Powell and Qiao 2006). Facilitated transport mech-
on nonporous polymeric membranes. These operate anism selectivity for CO2 has also been developed,
on the solution-diffusion mechanism and take especially for syngas applications. Many of these
advantage of the strong CO2 solubility within the materials display high CO2 permeances and CO2
polymeric matrix relative to other gases, such as N2 selectivities relative to other gases present. How-
and H2. In particular, glassy polymeric membranes ever, none of these membrane materials have been
are applied to CO2 separation from ue gas and successfully commercialized to date.
natural gas, because of the high permeabilities and
selectivities possible for these types of polymers.
References
While, for CO2 separation from syngas, rubbery
polymeric membranes are the focus, because these Merkel TC, Pinnau I, Prabhakar R, Freeman BD
can provide the CO2/H2 selectivity needed for CO2 (2006) Gas and vapour transport properties of peruor-
separation. Molecular sieving and surface diffusion opolymers in materials science of membranes for gas
membranes based on porous inorganic and and vapor separation. Wiley, Chinchester
Powell CE, Qiao GG (2006) Polymeric CO2/N2 gas sepa-
nanoporous carbon materials have been developed ration membranes for the capture of carbon dioxide
that are CO2 selective. However, there is currently from power plant ue gas. J Membr Sci 279:149
no industrial plant or pilot trials of these membranes. Scholes CA, Stevens GW, Kentish SE (2012) Membrane
The rst CO2-selective membrane to be commer- gas separation applications in natural gas processing.
Fuel 96:1528
cialized was cellulose acetate, for the natural gas
sweetening industry. Generally, these membranes
are a combination of cellulose acetate, cellulose
diacetate, and cellulose triacetate and have CO2 Carbon Membrane
permeabilities on the order of three to six barrer
and CO2/CH4 selectivities of 2540. Currently, cel- May-Britt Hgg
lulose acetate membranes make up to 80 % market Department of Chemical Engineering, Norwegian
share of membranes in natural gas processing, and University of Science and Technology,
modules of both spiral wound and hollow ber Trondheim, Norway
designs can be purchased (Scholes et al. 2012). Cel-
lulose acetate-based membranes are currently com-
mercialized under the names Cynara (part of Carbon membranes are usually characterized as
Cameron) and UOP Separex (part of Honeywell). carbon molecular sieves (CMS membranes) as
300 Carbon Membrane
they basically separate based on molecular size. The rst carbon membranes were prepared
This means that the pores can be tailored for a from carbonization of cellulose hollow bers by
specic gas mixture, permeating the smallest mol- Koresh and Soffer (1983). Since then various
ecule and retaining the larger ones. polymers have been used such as polyimides,
There are two main ways of preparing the CMS polyacrylonitrile (PAN), poly(phenylene oxide)
membranes: pyrolysis of polymeric hollow bers (PPO), cellulose derivates, and others. The sepa-
or performing pyrolysis of a polymer sprayed onto ration properties of CMS membranes are very
a ceramic support tube. In both cases a pyrolysis attractive as their sieving properties easily exceed
protocol needs to be developed for the specic the upper Robeson boundary of permeability ver-
polymer and pores to be obtained. Important vari- sus selectivity trade-off relationship (Robeson
ables to be controlled during the pyrolysis is 2008) this makes it also attractive for industrial
heating rate, soak time, nal carbonization tem- applications. The making of the membranes
perature, and which gas to use as purge (typically may thus involve six steps: material selection,
an inert gas or CO2) vacuum may also be used. material functionalization, precursor pre-
Each of these variables will inuence the nal paration, pretreatment, carbonization, and possi-
pore size and separation properties of the mem- bly posttreatment (Fig. 1). Functionalization
brane. The nal carbonization temperature is usu- usually means adding a porogen or a metal salt
ally in the range of 500900 C. Carbon to the precursor. The porogen will degrade during
membranes are fairly easy to produce on lab the carbonization and thereby increase the pore
scale as much is known on how carbonization volume. When the additive is a metal salt (doping
conditions affect separation properties (Geiszler the precursor), the nal CMS membrane will con-
and Koros 1996; He and Hgg 2012). However, tain the metal which may functionalize the mem-
the scaling up to commercial-size modules may be brane by showing high afnity to one of the
more of a challenge as commented on permeating gases (Lie and Hgg 2005). Then if
further down. using posttreatment, surface modiers may adjust
Carbon Membrane,
Fig. 1 Comparing the
CO2/CH4 Robeson upper
bound for dense and
thermally rearranged
(TR) polymer membranes
to the carbon membranes
and the region for industrial
applicability data for
CMS membranes and
industrial applicability
region added to the original
Robeson plot (see Hgg and
He 2011)
Carbon Membrane 301
a b c
Porogens Doping/spacers Surface modifiers

Carbon Membrane, Fig. 2 Schematic model of different ways of tailoring microporous carbon. The dashed line shows
the limit of an original slit-shaped pore (Lie and Hgg 2005)
Carbon Membrane,
Fig. 3 Left: SEM picture of
a CMS ber; wall thickness
is 16 mm. Right: A bundle of
CMS hollow bers ready
for mounting in a pilot
module as illustrated by the
company MemfoACT in
2013 on their homepage
the pore size. A big advantage of carbon mem- especially studied for high-temperature applica-
branes is that they may be used for high- tions (e.g., H2/CO2 separation and as carbon
temperature gas separations. This advantage membrane reactors). The special challenge of
may, however, be compromised if the surface making these membranes is to obtain a defect-
modier is a degradable polymer at process oper- free carbon layer on the top of the
ation temperature. The different ways of tailoring support compatibility between support and car-
microporous carbon membranes are illustrated in bon layer may also be a problem. Supported CMS
Fig. 2 (Lie and Hgg 2005). membranes have been studied by several.
Although showing excellent separation perfor- A method by using spin-coat technique was
mance, the self-supported hollow ber mem- developed already in 1996 by Rao and Sirkar
branes may be quite brittle, thus the scaling up (1996). Stainless steel disks (Acharya and Foley
and module making are challenging. The Norwe- 1999) as well as porous a-alumina tubes (Zhou
gian company MemfoACT did successfully dem- et al. 2003) have been reported as suitable support
onstrate on pilot scale the use of carbonized self- for pyrolized carbon membranes. Acharaya and
supported cellulose hollow bers for upgrading of Foley studied the separation of O2/N2 and
biogas (CO2/CH4 separation) (Fig. 3). The scaling obtained a separation factor of 12. A challenge
up to full-size commercial module has however of concern with respect to any CMS membrane is
not yet been reported (in 2015). the possibility of aging, whereby separation per-
The supported carbon membranes will provide formance may decline over time; hence, the mem-
better mechanical strength than self-supported branes need to be regenerated. Several ways are
bers but will naturally have lower packing den- proposed, such as using heat, purging with air or
sity. These types of carbon membranes have been propane, or using low-voltage direct current at
302 Carbon Nanotube Membranes (CNT Membranes)
intervals (Lie and Hgg 2006). The research on Rao MB, Sirkar S (1996) Performance and pore character-
CMS membranes is however continuously mov- ization of nanoporous carbon membranes for gas sepa-
ration. J Membr Sci 110:109
ing forward, and is well documented, and example Robeson LM (2008) The upper bound revisited. J Membr
is Fu et al. (2015). The main challenge still seems Sci 320:390400
to be scaling up to commercial-size modules. Yoshino M, Nakamura S, Kita H, Okamoto K, Tanihara N,
Kusuki Y (2003) Olen/paran separation perfor-
mance of carbonized membranes derived from asym-
metric hollow ber membranes of 6FDA/BPDA-
Potential of Applications DDBT copolyimide. J Membr Sci 215:169183
Zhou W, Yoshino M, Kita H, Okamoto K (2003) Prepara-
Due to the possibility of tailoring the pores, the tion of gas permeation properties of carbn molecular
sieve membranes based on sulfonated phenolic resin.
CMS membranes have a very nice potential J Membr Sci 217:5567
of separating gas pairs which are difcult to
separate using other type of membranes and
gases which are much alike in structure and
physical properties, typically O2/N2 and C3H8/
C3H6 (olen/parafns in general) (Yoshino Carbon Nanotube Membranes
et al. 2003). Other type of gas separations can (CNT Membranes)
typically be CO2/CH4 (both as biogas and as
natural gas) and recovery of hydrogen from gas Enrica Fontananova1 and Enrico Drioli1,2,3,4
1
streams such as H2/CO2 (precombustion) and H2 Institute on Membrane Technology, National
from renery gasses. Research Council of Italy, ITM-CNR, Rende,
Italy
2
Department of Environmental and Chemical
Engineering, University of Calabria, Rende,
References
Italy
3
Acharya H, Foley H (1999) Spray-coating of nanoporous WCU Energy Engineering Department,
carbn membranes for air separation. J Membr Sci Hanyang University, Seoul, S. Korea
4
161:15 Center of Excellence in Desalination
Fu S, Sanders E, Kulkarni SS, Koros WJ (2015) Carbon Technology, King Abdulaziz University, Jeddah,
molecular sieve membrane structureproperty relation-
ships for four novel 6FDA based polyimide precursors. Saudi Arabia
Geiszler VC, Koros WJ (1996) Effects of polyimide
pyrolysis conditions on carbon molecular sieve Carbon nanotubes (CNTs) are allotropic forms of
membrane properties. Ind Eng Chem Res
35:29993003 carbon that can be conceptually visualized as a
Hgg MB, He X (2011) Carbon molecular sieve mem- cylindrical nanostructure formed by rolling a
branes for gas separation, chapter 15. In: Drioli E, graphene sheet along a lattice vectors (m,n) in
Barbieri G (eds) Membrane engineering for the treat- the graphene plane (Fig. 1).
ment of gases, membrane engineering handbook. RCS
Publishing, Cambridge CB4 0WF, UK. ISBN:978-1- Depending from the number of concentric cyl-
84973-239 inders, it is possible to have: single-, double-, or
He X, Hgg MB (2012) Structural, kinetic and perfor- multiwalled carbon nanotubes (SWCNTs,
mance characterization of hollow ber carbon mem- DWCNTs, MWCNTs). The indices (m,n) deter-
branes. J Membr Sci 390391:2331
Koresh J, Soffer A (1983) Molecular sieve carbon mem- mine the diameter and chirality (i.e., the chiral
brane. Part I. Presenting a new device for gas mixture angle between hexagons and the tube axis) of the
separation. Sep Sci Technol 18:723734 CNTs (Dai 2002). The CNTs can be in the zigzag,
Lie JA, Hgg MB (2005) Carbon membranes from cellu- armchair, or chiral form with metallic or semicon-
lose and metal loaded cellulose. Carbon 43:26002607
Lie JA, Hgg MB (2006) Carbon membranes from cellu- ducting properties (Table 1).
lose: synthesis, performance and regeneration. CNTs, thanks to their diameter in the nanome-
J Membr Sci 284:7986 ter range and their atomically smooth surfaces,
Carbon Nanotube Membranes (CNT Membranes) 303
Carbon Nanotube a
Membranes
(CNT Membranes),
Fig. 1 Schematic
honeycomb structure of a n
graphene sheet (sp2 m
hybridized carbon network)
(a) and image of single-,
double-, and multiwalled
C
carbon nanotubes
(SWCNTs, DWCNTs,
MWCNTs), (b)
SWCNT DWCNT MWCNT
offer unique systems for fast molecular transport It is interesting to note that graphitic carbon
(Hummer et al. 2001; Kalra et al. 2003). Molecu- nanomaterials like CNTs are usually introduced at
lar dynamics (MD) simulations showed spontane- lower content (2 wt%) in mixed matrix mem-
ous and continuous lling of a CNT with a branes in comparison with three-dimensional
one-dimensionally ordered chain of water inorganic nanollers like TiO2 and ZrO2 (usually
molecules. blended at loading 5 wt%, up to 60 wt%), thank
Membranes made of ordered arrays of carbon to their high specic surface, elevated aspect ratio,
nanotubes well aligned along the direction per- and the intrinsic properties of graphitized struc-
pendicular to the substrate surface have been ture. In addition, the relatively easy functiona-
already synthesized in which the transport occur lization of the surface of carbon nanomaterials
through the inner core of the CNTs (Majumder render them ideal candidate to tailor the poly-
et al. 2011). High permeability enhancement than mer/nanoller interface.
predictions (>10 for gases and >1000 for liquids) CNTs were succesfully entrapped in mixed
have been obtained with these membranes. How- matrix membranes made of various polymeric
ever, the small area (typically in the order of few materials by several techniques including: disper-
mm2), the long and complex fabrication process, sion in the casting solution and successive phase
and the poor mechanical stability of these type of separation, entrapping in the membrane pores
CNT membranes, limit the practical applicability using a polymer binder, in situ crosslinking of a
of these interesting nanostructured systems. On polymer matrix around oriented CNTs (Ismail
the contrary, more easy is the scale up of CNT et al. 2009). The resulting mixed matrix CNT
membranes in which the CNTs are mixed with a membranes offer, in many cases, relevant advan-
polymer in the form of mixed matrix membranes. tages in comparison with the polymeric samples.
The main limitations of these membranes are Poly(vinyl alcohol) (PVA)/MWCNTs membranes
related to the poor dispersion of the CNTs in the were realized for pervaporation application,
polymeric matrix and the moderate increase of the obtaining signicant improvement in Youngs
membrane performance. modulus and thermal stability, as compared to
304 Carbon Nanotube Membranes (CNT Membranes)
Carbon Nanotube Membranes (CNT Membranes), of this composite showed an increased membrane
Table 1 Different forms of a SWCNT and its electrical mechanical strength relative to the PA polymeric
properties
membranes.
(n, m) Form Electrical conductivity Mixed matrix polyimide membranes were also
(n, 0) Zigzag Metallic when n is multiple prepared using functionalized MWCNTs (aminated
of 3, otherwise,
semiconducting or oxidized). The MWCNTs functionalization
improved their dispersion in the casting solution
in comparison to pristine MWCNTs and, as a con-
sequence, in the formed membranes. The mem-
branes containing functionalized MWCNTs
showed better performance in the rejection of
dyes (enhanced ux and reduced fouling, with sim-
ilar or higher rejection), with respect to reference
(n, m) Armchair Metallic
with
polymeric membranes (without MWCNTs). These
n=m results were attributed to the formation of
low-resistance pathways for solvent transport at
the interface between the MWCNTs and the poly-
meric chains. Moreover, the MWCNTs reduced the
severe membrane fouling caused by the absorption
of the positively charged dye Safranine (used as a
model of organic pollutant) with respect the
polymeric membrane, by a screening effect of the
(n, m) Chiral Metallic when (n-m)/3 is an
with m 6 integer, otherwise,
attractive electrical interactions between the
0 and n semiconducting Safranine and the membrane surface characterized
by a negative value of zeta potential (Grosso
et al. 2014).
The presence of oxygen-containing polar
groups on oxidized MWCNTs, resulted also in a
good dispersion in polyvinylidene diuoride
(PVDF) membranes, allowing the formation of
mixed matrix membranes with a lower fouling
tendency in comparison with the bare polymeric
samples (Fontananova et al. 2014).
pure PVA membranes (Peng et al. 2007). The MWCNTs were also immobilized in the pores
entrapment of MWCNTs in polyethersulfone of a hydrophobic membrane improving the
(PES) membranes reduced fouling problems in water-membrane interactions to promote vapor
water treatment (Celik et al. 2011). Mixed matrix permeability in membrane distillation process
membranes consisting of sulfonated carbon (Gethard et al. 2011). In this case, the CNTs dis-
nanotubes (sCNTs) and sulfonated poly persion was forced under vacuum into the pores of
(ethersulfone ether ketone ketone) (SPESEKK) a polypropylene (PP) membrane, using PVDF as
were also fabricated via the solution casting binder.
method (Zhou et al. 2011). The proton conductiv-
ity of the SPESEKK membrane increased while
the methanol permeability decreased as the sCNTs
content increased. MWCNTs were covalently References
linked to aromatic polyamide (PA) membranes
Celik E, Park H, Choi H, Choi H (2011) Carbon nanotube
by a polymer grafting process (Shawky et al. blended polyethersulfone membranes for fouling con-
2011). Measurements of mechanical properties trol in water treatment. Water Res 45:274282
Carbon Porous Membranes Modified with Enzymes 305
Dai H (2002) Carbon nanotubes: synthesis, integration,

and properties. Acc Chem Res 35:10351044 Carbon Porous Membranes Modified
Fontananova E, Bahattab MA, Aljlil SA, Alowairdy M,
Rinaldi G, Vuono D, Nagy BJ, Drioli E, Di Proo with Enzymes
G (2014) From hydrophobic to hydrophilic polyvinyli-
deneuoride (PVDF) membranes by gaining new Christophe Innocent
insight into materials properties. RSC Adv Institut Europen des Membranes, ENSCM,
5:5621956231
Gethard K, Sea-Know O, Mitra S (2011) Water desalina- CNRS, Universit de Montpellier, Montpellier, C
tion using carbon-nanotube-enhanced membrane distil- France
lation. Appl Mater Interfaces 3:110114
Grosso V, Vuono D, Bahattab MA, Di Proo G, Curcio E,
Al-Jilil SA, Alsubaie F, Alfe M, Nagy BJ, Drioli E,
Fontananova E (2014) Polymeric and mixed The carbon porous tubes (from Novasep-Orelis-
matrix polyimide membranes. Sep Purif Technol France) have been characterized by average pore
132:684696 diameter of 3 mm and an inner diameter of 0.6 cm
Hummer G, Rasaiah JC, Noworyta JP (2001) Water con- (Fig. 1).
duction through the hydrophobic channel of a carbon
nanotube. Nature 414:188190 The immobilization of enzyme on the carbon
Ismail AF, Goh PS, Sanip SM, Aziz M (2009) Transport tube surface has been carried out by dipping in
and separation properties of carbon nanotube-mixed enzyme solution and then addition of pyrrole
matrix membrane. Sep Purif Technol 70:1226 monomer just before starting the anodic electropo-
Kalra A, Garde S, Hummer G (2003) Osmotic water trans-
port through carbon nanotube membranes. Proc Natl lymerization at +0.9 V vs. Ag/AgCl. The oxidation
Acad Sci U S A 100:1017510180 of pyrrole allowed the entrapment of enzyme in
Majumder M, Chopra N, Bruce JH (2011) Mass transport polypyrrole coated on the electrode surface.
through carbon nanotube membranes in three different A glucose/O2 biofuel cell is based on tubular
regimes: ionic diffusion and gas and liquid ow. ACS
Nano 5:38673877 cathode (coated with laccase/ABTS on the exter-
Peng F, Hu C, Jiang Z (2007) Novel ploy(vinyl alcohol)/ nal surface for dioxygen reduction) and the tubu-
carbon nanotube hybrid membranes for pervaporation lar anode (coated with glucose oxidase enzyme
separation of benzene/cyclohexane mixtures. J Membr (GOD) and hydroxyquinoline sulfonic acid as
Sci 297:236242
Shawky HA, Chae S-R, Lin S, Wiesner MR (2011) Syn- mediator) on the internal surface for glucose
thesis and characterization of a carbon nanotube/poly- oxidation.
mer nanocomposite membrane for water treatment. The two electrodes were soaked in an unstirred
Desalination 272:4650 10 mM glucose solution. Glucose diffused
Zhou W, Xiao J, Chen Y, Zeng R, Xiao S, Nie H, Li F, Song
C (2011) Sulfonated carbon nanotubes/sulfonated poly through the polypyrrole lm to be oxidized by
(ether sulfone ether ketone ketone) composites for the GOD at the anode. In order to prevent the
polymer electrolyte membranes. Polym Adv Technol presence of dioxygen in the vicinity of GOD,
22:17471752 dissolved oxygen was supplied to the system by
Carbon Porous
Membranes Modified
with Enzymes,
Fig. 1 Photography of
carbon tube (photo IEM)
306 CarmanKozeny Equation
the pressure drop for laminar ow through a

packed bed of solids. It was originally developed
by Kozeny in 1927, using the simplied model of a
number of parallel capillary tubes of equal length
and diameter to describe the packed bed (McCabe
et al. 2005). The Kozeny equation is given by
Dp KV o m 1 e2
2 2 (1)
L Fs Dp e3
where Dp is the pressure drop, L is the total height

of the bed, Vo is the supercial or empty-tower
velocity, m is the viscosity of the uid, e is the
porosity of the bed, Fs is the sphericity of the
particles in the packed bed, Dp is the diameter of
the equivalent spherical particle, and K is an
empirical constant, which depends on bed tortu-
Carbon Porous Membranes Modified with Enzymes, osity. The sphericity of a particle is dened by
Fig. 2 Scheme of concentric biofuel cell (thesis of G.
MERLE, University Montpellier 2, 2008)
6=Dp
Fs (2)
av
circulating through the inner cavity of the
biocathode, then before being reduced at the where av is the specic surface area of the particle.
external surface by the BOD. If the bed is made of uniform particles, the av of
Cylindrical and porous carbon tubes were used the particle represents the specic surface of the
as original conducting support for the compart- bed. The av of the particle is simply the surface-
ment of the bioelectrodes, for enzyme immobili- volume ratio, i.e., av = sp/vp, where sp is the sur-
zation and transport of dissolved oxygen via face area of the particle and vp is the volume of the
diffusive ow through the porosity (Fig. 2). The particle. It follows that for a sphere with the diam-
resulting glucose/oxygen biofuel cell is operated eter Dp, av = 6/Dp and Fs 1:0.
at a maximum power density of 42 mWcm2at In 1937 Carman applied Eq. 1 to experimental
0.3 V and 37 C in phosphate buffer pH 7.4. results of ow through packed beds and found that
K = 180 (Foust et al. 1980). Other references also
suggested the value of K = 150 (Seader et al. 2011).
The CarmanKozeny equation indicates that the ow
CarmanKozeny Equation is proportional to the pressure drop and inversely
proportional to the uid viscosity. This statement is
Boguslaw Kruczek also known as Darcys law, which is often used to
Department of Chemical & Biological describe ow of liquids through porous media.
Engineering, University of Ottawa, Ottawa, ON,
Canada
Extension of the CarmanKozeny
Equation
Introduction
The CarmanKozeny equation is applicable for
The CarmanKozeny equation (or the ow through beds at particle Reynolds number
KozenyCarman equation) is a relation to calculate (Rep) up to about 1.0, and
CarmanKozeny Equation 307
V o Dp r kg s1 m2) through the membrane is then

Rep (3)
m obtained by multiplying the ow velocity by uid
density and membrane porosity:
where r is the density of uid. As the ow rate
through a packed bed increases, the pressure drop erD2
becomes a stronger function of the supercial N Dp (7)
velocity. At Rep > 1,000 the pressure drop in
32mLM C
packed beds is described by an empirical relation
often referred to as the BurkePlummer equation In real porous membranes, the pores may not be
(McCabe et al. 2005): cylindrical and straight, and Eq. 7 is modied by
using the procedure developed by Carman and
Kozeny, in which the pore diameter is replaced
Dp 1:75rV 2o 1 e
(4) by the hydraulic diameter (DH):
L Fs Dp e3
An equation covering the entire range of ow 4Ap

DH (8)
rates is obtained by assuming that the pressure P
drop due to viscous losses (Eq. 1) and the pressure
drop due to the kinetic energy losses (Eq. 4) are where Ap is the cross-sectional area of the pore
additive. The result is called the Ergun equation and P is the wetted perimeter of the pore.
(McCabe et al. 2005): Expressing Ap and P in terms of the specic sur-
face area and the porosity of the membrane leads
Dp KV o m 1 e2 1:75rV 2o 1 e to an alternative expression for DH (Seader
2 2 (5) et al. 2011):
L Fs Dp e3 Fs Dp e3
4e
The sphericity often does not appear in Eq. 5, DH (9)
av 1 e
which implies the particles in the bed are nearly
spherical (Fs 1). The actual pore length is longer than the mem-
brane thickness and can be represented by t LM,
where t is a tortuosity factor > 1. Substituting
CarmanKozeny Equation in Membrane
Eq. 9 and the tortuosity factor into Eq. 7 gives
Processes
e3 r
The transport of uid in ultraltration (UF) and N Dp (10)
microltration (MF) processes is driven by a 21 e2 a2v mtLM
pressure difference across the membrane (Dp). The
UF and MF membranes are porous with the pore Equation 10 is an alternative representation of the
size ranges of 2100 nm and 0.110 mm, respec- CarmanKozeny equation. Recognizing that
tively (Mulder 1996). Because of very small pore Fs Dp 6=av , the actual uid velocity in the bed
sizes, the ow velocity in an idealized straight cylin- V = Vo/e, replacing L by LM, and multiplying by
drical pore of a UF or MF membrane is governed by the uid density r, Eq. 1 can be rearranged to
the HagenPoiseuille law (Seader et al. 2011):
e3 r
N Dp (11)
D 2
K=361 e2 a2v mLM
V (6)
32mLM
If K = 150, the comparison of Eq. 10 with Eq. 11
where D is the pore diameter and LM is the mem- leads to t = 2.08; on the other hand, with
brane thickness. The mass velocity of uid (N in K = 180 t = 2.5. In other words, the numerical
308 Carrier-Mediated Transport
value of the empirical constant in the resistance to the ow through the membrane
CarmanKozeny equation depends on the tortu- (RM). Consequently, the expression for the mass
osity of the packed bed, which in general ranges velocity of uid can be written as follows:
from 2.0 to 2.5. In the case of the porous mem-
branes, the tortuosity can be greater than that of Dp
N (14)
the packed bed. m R C R M
The porosity and the specic surface area of
porous membrane can be determined experimen- and
tally. If the membrane tortuosity along the mem-

brane thickness is known, Eq. 10 allows to predict K 1 e2 a2v mLC
the mass velocity of uid through the membrane RC (15)
36 e3 r
under a given pressure gradient across the mem-
brane. Alternatively, for a given mass uid veloc-
21 e2 a2v mtLM
ity, Eq. 10 can be used to predict the required RM (16)
e3 r
pressure gradient across the membrane.
The mass velocity of uid through the mem-
If there is no adsorption of solute particles inside
brane can also be expressed in terms of bulk-ow
the membrane pores, RM is constant, but since the
permeability (PM), which is a measure of the
cake thickness (LC) increases with the amount of
membrane productivity (Seader et al. 2011):
feed processed through the membrane, RC
increases accordingly. Equations 14, 15, and 16
PM are analogous to those used to describe the con-
N Dp (12)
LM ventional cake ltration.
It follows from the comparison of Eq. 10 with

Eq. 12 that the bulk-ow permeability in porous
UF and MF membranes is given by References
Foust AS, Wenzel LA, Clump CW, Maus L, Andersen BL

e3 r (1980) Principles of unit operations. Wiley, New York
PM (13) McCabe W, Smith JC, Harriot P (2005) Unit operations of
21 e2 a2v mt chemical engineering. McGraw-Hill, New York
The rearranged form of the original ogy. Kluwer, Dordrecht
Seader JD, Henley EJ, Roper DK (2011) Separation pro-
CarmanKozeny equation given by Eq. 11 can cess principles. Wiley, Hoboken
be used to predict the bulk-ow permeability of
microltration membranes prepared by sintering.
This method of membrane formation involves
compressing a powder consisting of particles of
a given size and sintering at elevated temperatures Carrier-Mediated Transport
(Mulder 1996). Consequently, the nal membrane
has a structure similar to that of the packed bed. Gerardo Len
The CarmanKozeny equation is also applica- Departamento de Ingeniera Qumica y
ble for the analysis of membrane processes in Ambiental, Universidad Politcnica de Cartagena,
which the transport of uid through the membrane Cartagena, Spain
is associated with cake formation on the mem-
brane surface (Mulder 1996). In this case, the
uid which permeates through the membrane Transport mechanism used to improve ux, per-
must overcome two resistances in series, the resis- meability, and selectivity in membrane separa-
tance to ow through the cake (RC) and the tions. It is also known as type 2 facilitated
Carrier-Mediated Transport 309
transport or carrier-facilitated transport (Ho and Li the bulk feed phase across the aqueous boundary
1992). A carrier reagent is incorporated in the layer to the feed phase/membrane phase interface;
membrane phase to transfer the diffusing specie (2) formation of carrier-target specie compound;
from the feed phase to the product phase across (3) diffusion of the carrier-target specie compound
the membrane phase. Although xed carriers across the membrane; (4) breakage of the carrier-
(bonded within the membrane solid matrix either target specie compound at the membrane phase/
by electrostatic forces or by covalent linkages) product phase interface, releasing both the target C
have been described, liquid membranes employ specie and the carrier; (5) diffusion of the released
mobile carriers, i.e., compounds (soluble in the target specie from the membrane phase-product
membrane phase but insoluble in the feed and the phase interface to the bulk strip phase; (6) return
product phases) which are added to the membrane of the released carrier across the membrane to the
phase to form with the specie to be separated into feed phase/membrane phase interface (Marr and
a reversible compound or complex, soluble in the Knopp 1982). In this way, each carrier molecule is
membrane. This compound is formed in the inter- able to transport solute molecules as many times
face feed phase/membrane phase, it is transported as necessary, so that only a small amount of the
through the membrane phase due to its concentra- carrier is required in the membrane phase even for
tion gradient, and it is broken in the membrane achieving a high degree of separation. As the
phase/product phase interface. Then, the specie to process ends when the concentration of the specie
be separated is liberated in the product phase, and to be separated equalizes on both sides of the
the carrier diffuses in the contrary way through the membrane, carrier-mediated transport can be
membrane phase, as a consequence of its concen- improved by incorporating a chemical reaction
tration gradient, to the opposite interface, to initi- on the receiving side in two ways. The rst is by
ate a new separation cycle. So, the transport of the reducing the concentration of the specie to be
target species can be broken down to the follow- separated to practically zero on the product side
ing steps: (1) diffusion of the target specie from through an auxiliary reaction in the product phase,
a b
Feed Membrane Product Feed Membrane Product
Phase phase phase Phase phase phase
(Oct)3N NaOH CLXD
AOH M+ (aq) M+ (aq)
(aq) Pic- (aq) Pic- (aq)
(Oct)3NH+ -OA NaOA + [M CLXD]+ Pic-
H2O
M+
AOH
Pic-
c d
Feed Membrane Product Feed Membrane Product
Phase phase phase Phase phase phase
M+2 2(HR)2 2H+ DC18C6 4SCN-

MCl4= (aq)
2 K+ (aq)
2H+ MR2(HR)2 M+2 [M(SCN)4]=
(KDC18C6)2[MCl4]
2 K+ 4 Cl-
M+2
M+ K+
Cl-
H+
Carrier-Mediated Transport, Fig. 1 (a) Carrier- (CLXD), (c) carrier-mediated counter-transport of a metal-
mediated transport of an acid with tri-n-octylamine, (b) lic ion by an organophosphorus carrier (HR)2, and (d)
carrier-mediated co-transport of a metallic ion and its carrier-mediated co-transport of a metallic ion and accom-
accompanying anion with a calixarene derivative panying anion and cation with a crown ether (DC18C6)
310 Cascade Diafiltration
to form a compound insoluble in the membrane Consider a two-stage cascade. Constant vol-
and release the carrier to begin a new separation ume dialtration is performed in the rst stage to
cycle. The second is by maintaining the driving maximize the recovery of the product which
force for the transport of the target specie through passes through the membrane. The product-
maintaining a concentration gradient of a second containing permeate enters the second stage
specie transported in the opposite direction. where it undergoes batch ultraltration using a
Carrier-mediated transport combines the process membrane which retains the product.
of extraction and stripping in a single unit opera- Cascades can be operated in open-loop or
tion, and it offers the possibility of transporting a closed-loop congurations (Charcosset 2012). In
component against its own concentration gradient the open-loop conguration, the permeate from
(Ho and Li 1992; Kislik 2010). In carrier- the second stage goes to drain. In the closed-loop
mediated transport, two different species can be conguration, the permeate from the second stage is
transported at the same time, giving a coupled used as the diluant for the rst stage, thus reducing
transport, in which the membrane contains a carrier substantially the volume of fresh diluant required.
which can only lead to transport when two different
species present themselves at the same time. If the
transport of the two species occurs in the same References
direction (from the feed to the product phase), we
can speak of co-transport. On the other hand, if it Charcosset C (2012) Membrane processes in biotechnol-
occurs in opposite directions (one is transported ogy and pharmaceutics, 1st edn. Elsevier, Boston
from the feed to the product phase and the other
from the product to the feed phase), we can speak
of counter-transport. Different examples of carrier-
mediated transport are shown in Fig. 1. CassieBaxter Model
Wolfgang M. Sigmund and Shu-Hau Hsu

References Department of Materials Science and
Engineering, University of Florida, Gainesville,
Ho WSW, Li NN (1992) Emulsion liquid membranes. Deni- FL, USA
tions. In: Ho WSW, Sirkar KK (eds) Membrane handbook, Department of Energy Engineering, Hanyang
1st edn. Van Nostrand Reinhold, New York, pp 597610
Kislik VS (2010) Introduction, general description, denitions University, Seoul, Republic of Korea
and classication. Overview. In: Kislik VS (ed) Liquid
membranes, 1st edn. Elsevier, Amsterdam, pp 115
Mar R, Knopp A (1982) Liquid membrane technology. The Wenzel regime is usually recognized as
A survey of phenomena, mechanisms and models. Int
Chem Eng 22:4460 homogeneous wetting, since the liquid
completely penetrates into the grooves. While
under some circumstances, especially the increase
of the surface roughness, vapor pockets may be
Cascade Diafiltration trapped underneath the liquid yielding a compos-
ite interface. This heterogeneous wetting is usu-
Greg Foley ally described by the Cassie-Baxter (CB) model,
School of Biotechnology, Dublin City University, from which the apparent contact angle (yCB) is
Dublin, Ireland given by equation: (Cassie and Baxter 1944)
cos yCB f s cos yc f v cos yv (1)

Cascade systems may be used when the product is
in the permeate of a microltration or ultraltra- where yc is the intrinsic contact angle on the
tion system. original smooth surface and fs and fv are the area
Casting 311
Liquid
q cs
Substate
C
CassieBaxter Model, Fig. 1 Vapor pockets are trapped Cassie-Baxter model (Hsu et al. 2011, reproduced with
between the grooves and the liquid droplet. This state is permission from Elsevier Ltd.)
essential for superhydrophobicity and is described by
fractions of the solid and vapor on the surface,

respectively. Since f s f v 1 and yv = 180 Casting
(implying that a suspended liquid droplet in air
is a perfect sphere), Eq. 1 can be rewritten as Francesco Galiano
follows: Institute on Membrane Technology, National
cos yCB f s cos yc 1 1 (2) Italy
From Eq. 2, it can be found that droplets will have Casting is the main technique applied for the
a higher apparent contact angle if less area is in preparation of at polymeric membranes using
contact with the solid substrate. The CB equation phase inversion process (such as NIPS, VIPS,
simply indicates the contact angle can be etc.). The procedure for casting at sheet mem-
increased even when the intrinsic contact angle branes can be summarized as follows:
of a liquid on the original smooth surface is less
than 90 . 1. Preparing the casting solution by dissolving
New developments in the eld show that there the polymer in the appropriate solvent
may need to be taken other considerations 2. Placing a suitable casting plate (a belt, a glass,
for contact angles and contact angle hysteresis. or other supports)
It is suggested that the contact line model is a 3. Adjusting the desired casting knife thickness
better approach to predict the superhydro- 4. Pouring the polymeric solution on the casting
phobicity of surfaces (Gao and McCarthy 2007, plate
2009) (Fig. 1). 5. Uniformly spreading out the polymeric solu-
tion across the casting plate (Fig. 1)
Then, based on the way the phase separation

References occurs, the following techniques can be
distinguished:
Cassie ABD, Baxter S (1944) Wettability of porous sur-
faces. Trans Faraday Soc 40:546550
Wet-casting process: the cast lm is immersed
Gao L, McCarthy TJ (2007) How Wenzel and Cassie were
wrong. Langmuir 23:37623765 into a coagulation bath, and due to solvent/
Gao L, McCarthy TJ (2009) An attempt to correct the nonsolvent exchange, the precipitation of the
faulty intuition perpetuated by the Wenzel and Cassie polymer occurs.
Laws. Langmuir 25:72497255
Dry-casting process: the volatile solvent is left
Hsu SH, Woan K, Sigmund W (2011) Biologically inspired
hairy structures for superhydrophobicity. Mater Sci to evaporate from the cast lm leading to the
Eng 72:189201 precipitation of the polymer.
312 Casting Solution
Casting, Fig. 1 Procedure for casting at sheet membranes at lab scale and methods for inducing phase separation
Dry-wet-casting process: the cast lm is

exposed to a nonsolvent vapor (generally Casting Solution
water) for a xed time and then immersed
into a coagulation bath. Francesco Galiano
Thermal-induced phase separation: the cast poly- Institute on Membrane Technology, National
mer solution prepared at elevated temperature is Research Council of Italy, ITM-CNR, Rende,
cooled down causing phase separation and poly- Italy
mer precipitation (Drioli and Giorno 2009).
Manual and automatic casting machines are Casting solution is used to prepare membranes by
used both at laboratory and at industrial scale. casting method. Polymer and solvent are the main
In the latter case, the casting machine is on purpose components of the solution system, but various
created for specic applications. Commercial additives can be also added. The choice of the
membranes are, usually, cast on porous supports polymer is of primary importance in preparing
made up, for instance, of a nonwoven polyester the casting solution. The polymer, in fact, must
material in order to improve their mechanical sta- be soluble in the selected solvent in an appropriate
bility, obtaining composite membranes. concentration which is strictly related to nal
membrane application. Low concentrations of
polymer, in fact, generally lead to a membrane
References with a porous structure, while high concentrations
of polymer produce membranes with a more
Drioli E, Giorno L (2009) Membrane operations: innova-
tive separations and transformations. Wiley-VCH, dense structure. The typical solution casting,
Weinheim, Chapt 2, p.27 with a polymer concentration ranging from 15 to
Casting Solution Additives 313
20 wt.%, is used for the preparation of porous ber, in polymeric membrane preparation via
ultraltration membranes. On the contrary, a solu- phase inversion. The two main components of a
tion casting containing higher concentrations casting solution are the selected polymer (P) and a
of polymer, up to 25 wt.%, is used for the prepa- suitable solvent (S) (or a diluent, for membrane
ration of reverse osmosis, gas separation, and preparation via TIPS). Different types of additives
pervaporation membranes. can be introduced in the casting solution compo-
Even the choice of the solvent is not arbitrary sition to improve/enhance some selected mem- C
but based on specic requirements. The solvent brane properties, depending on the application.
must dissolve the polymer, and it must be miscible Additives can be divided in two categories: solu-
with a nonsolvent, when it is used as coagulation ble and insoluble additives. Soluble additives can
medium, taking into account both kinetic and be, in general, salts, as LiCl and LiClO4; small
thermodynamic aspects. Due to the fast demixing molecules, as glycerol or ethylene glycol (EG); or
in water, aprotic solvents, such as dimethyl form- polymers, as polyethylene glycol (PEG) or poly-
amide, dimethyl acetamide, or dimethyl sulfox- vinyl pyrrolidone (PVP), with different molecular
ide, are preferred if membranes with a porous and weights (Mw). Most soluble additives are intro-
asymmetric structures have to be obtained. On the duced to enhance the formation of membrane
contrary, solvents causing a slow precipitation of pores and are called pore former or pore
the polymer, such as acetone or tetrahydrofuran, forming. They are normally leached out from
are used for the production of dense membranes. the membrane matrix in the coagulation bath.
Addition of various modiers in the solution However, high Mw additives, as PVP, can be
casting allows to adjust membrane properties on retained in the membrane structure. This effect
the basis of nal membrane application. can be exploited to tailor membrane wettability,
Cosolvents, nonsolvents, and llers of different since PVP is hydrophilic. Similarly, a class of poly
types, such as pore forming and cross-linking (ethylene oxide)-b-poly(propylene oxide)-bpoly
agents, can be added to the casting solution (ethylene oxide) triblock copolymers, referred to
(Drioli and Giorno 2009). Although these addi- as Pluronic , can be used as additives to improve
tives are used in low concentrations, they are membrane hydrophilicity, thanks to their stable
determinant for the nal membrane performances incorporation in the polymeric matrix. The effect
and applications. of soluble additives on the membrane morphology
depends on a trade-off between thermodynamic
and kinetic factors. Soluble additives can increase
References the thermodynamic instability of the casting
Drioli E, Giorno L (2009) Membrane operations: innova-
solution. This can be detected from a shift of
tive separations and transformations. Wiley-VCH, the binodal curve toward the P/S axis, in ternary
Weinheim, p 30, Chapter 2 phase diagrams, which indicates less non-solvent
(NS) tolerance of the P/S/Additive system, with
respect to the P/S system having the same poly-
mer concentration. This normally promotes
Casting Solution Additives phase inversion, resulting in more porous struc-
tures, and may enhance macrovoids formation as
Silvia Simone well. However, additives can also increase the
Institute on Membrane Technology, National dope viscosity. This will hinder the S/NS
Research Council of Italy, ITM-CNR, Rende, exchange and, hence, delay phase inversion. As
Italy a result, membrane structure can shift to more
spongy type, while macrovoids formation may
be reduced. For instance, PVP affects both the
A casting solution is normally prepared to be thermodynamic and kinetic of the phase-
lmed as a at sheet, or extruded as a hollow inversion process due to its hydrophilicity and
314 Casting Solution Additives
its effect on the dope viscosity (Simone wettability and endow membranes with self-
et al. 2010). Hence, the effect of PVP on the cleaning, antifouling, and antibacterial properties
nal membrane morphology and properties (see, for instance, the work by Song et al. 2012).
depends both on its concentration and molecular Zeolites and carbon-based llers can be used to
weight. On the one hand, PVP concentration is overcome the typical trade-off between mem-
directly connected with the thermodynamic of brane productivity and selectivity in different
the P/S/Additive/NS system. On the other hand, applications, as gas separation and pervaporation
the casting solution viscosity increases with (see Clarizia et al. 2004; Vu et al. 2003; Dobrak
polymer concentration and molecular weight. et al. 2010). Other insoluble additives, as silica
Finally, the additive leaching from the polymeric particles, can be etched out, using proper acid/
matrix is more difcult as its Mw increases. Sev- alkaline treatment, and work as pore formers
eral examples of casting solution soluble addi- (Hashim et al. 2011).
tives effect on membrane morphology and
properties can be found in literature. For
instance, the analysis of PVP effect on the kinetic
References
and thermodynamic of phase inversion during
preparation of polysulfone (PS) membranes was Clarizia G, Algieri C, Drioli E (2004) Filler-polymer com-
carried out by Lee et al. (2003). Mansourizadeh bination: a route to modify gas transport properties of a
et al. (2010) found that lithium chloride acted as a polymeric membrane. Polymer 45:56715681
phase-inversion promoter additive and was Dobrak A, Figoli A, Chovau S, Galiano F, Simone S,
Vankelecom IFJ, Drioli E, Van der Bruggen B (2010)
leached out in the coagulation bath in the prepa- Performance of PDMS membranes in pervaporation:
ration of poly(vinylidene uoride) (PVDF) hol- effect of silicalite llers and comparison with SBS
low ber membranes. This resulted in bers with membranes. J Colloid Interface Sci 346:254264
reduced pore size, high surface porosity, and low Guillen GR, Pan Y, Li M, Hoek EMV (2011) Preparation
and characterization of membranes formed by
mass transfer resistance. nonsolvent induced phase separation: a review. Ind
Susanto et al. (2009) reported that Pluronic Eng Chem Res 50:37983817
improved the wettability of polyethersulfone Hashim NA, Liu Y, Li K (2011) Preparation of PVDF
(PES) membranes, thanks to its amphiphilic, sur- hollow ber membranes using SiO2 particles: the effect
of acid and alkali treatment on the membrane perfor-
factant properties, and was stably integrated in the mances. Ind Eng Chem Res 50:30353040
membrane matrix. Other relevant examples for Lee KW, Se BK, Nam ST, Han MJ (2003) Trade-off
each additive type can be found in the review by between thermodynamic enhancement and kinetic hin-
Guillen et al. (2011). drance during phase inversion in the preparation of
polysulfone membranes. Desalination 159:289296
Insoluble additives can be also used to tailor Mansourizadeh A, Ismail AF, Matsuura T (2010) Effect of
certain membrane properties. Inorganic llers, as operating conditions on the physical and chemical CO2
zeolites, metal oxides (as TiO2, ZnO, Al2O3), absorption through the PVDF hollow ber membrane
hydrophobic/hydrophilic cloisite, and carbon- contactor. J Membr Sci 353:192200
Simone S, Figoli A, Criscuoli A, Carnevale MC,
based llers (graphene oxide, nanotubes, fuller- Rosselli A, Drioli E (2010) Preparation of hollow
enes), can be incorporated, usually as powders of bre membranes from PVDF/PVP blends and their
nanometric size, in the casting solution, for pre- application in VMD. J Membr Sci 364:219232
paring mixed-matrix membranes (MMM). In this Song H, Shao J, He Y, Liu B, Zhong X (2012) Natural
organic matter removal and ux decline with
case, ller stability in the membrane matrix, as PEGTiO2-doped PVDF membranes by integration of
well as its uniform dispersion, is among the key ultraltration with photocatalysis. J Membr Sci
issues. For example, TiO2 nanoparticles (NPs) 405406:4856
show photocatalytic, antibacterial, antifouling, Susanto H, Stahra N, Ulbricht M (2009) High performance
polyethersulfone microltration membranes having
and UV-cleaning properties. They can be intro- high ux and stable hydrophilic property. J Membr
duced in the membrane matrix to enhance Sci 342:153164
Catalyst Immobilization in Membranes 315
Vu DQ, Koros WJ, Miller SJ (2003) Effect of condensable

impurity in CO2/CH4 gas feeds on performance of Catalyst Immobilization
mixed matrix membranes using carbon molecular
sieves. J Membr Sci 221:233239 in Membranes
Ivo Vankelecom
Centre for Surface Chemistry and Catalysis,
Catalyst Entrapment KU Leuven, Leuven, Heverlee, Belgium C
Ivo Vankelecom
Centre for Surface Chemistry and Catalysis, KU By incorporating catalysts in membranes, a
Leuven, Leuven, Heverlee, Belgium new environment is created around the catalyst
which can in several ways inuence the cata-
lytic activity and selectivity of the catalyst. The
Catalyst entrapment refers in a membrane context membrane can block certain inhibiting com-
to the connement of a catalyst in a membrane. pounds from the catalyst surroundings by steric
The catalyst can be either a homogeneous or het- exclusion or by decreasing the sorption ten-
erogeneous catalyst, or an enzyme. Two basic dency of that compound and/or its diffusion to
ways exist for the entrapment. In the rst the catalyst. Alternatively, it can facilitate the
approach, the (bio)catalyst is dispersed or transport of desired compounds to the reactive
dissolved in the polymer solution, which after sites. When incorporated in a membrane, other
forming into a at-sheet or hollow-ber mem- reactor setups can become available, e.g., with
brane can be turned to into a dense or porous the two reagents contacting the membrane (and
catalytic membrane. If of a porous nature, the thus the immobilized catalyst) from the different
pore sizes should obviously be smaller than the sides.
(bio)catalyst to prevent leaching. If of a dense Different methodologies exist to immobilize
nature, steric occlusion can also help to retain homogeneous catalysts in membranes: via a
the catalyst inside the polymer, but this will also covalent or coordinative bond between the cata-
happen when the afnity for the polymer matrix is lyst and the membrane, via electrostatic interac-
much higher than for the reaction mixture, hence tions between them, through steric occlusion of
prohibiting catalyst dissolution in the contacting the catalyst in a matrix or through physical
reaction uid (Vankelecom 2002). In the second adsorption of the catalyst on the membrane sur-
approach, encapsulation can be applied (McMorn face or in its pores. For homogeneous catalysts in
and Hutchings 2004). Connement of a liquid specic, extra advantages appear upon immobi-
which contains the (bio)catalyst is then taking lization in a membrane. This can be in fact a
place within small capsules enclosed by a polymer strategy to heterogenize the catalyst; it allows
or a surfactant. good dispersion of the homogeneous catalyst
into the reaction environment and nally also
offers an easy way for reuse. Additives that
References boost the catalyst performance can also be
co-immobilized. On the other hand, leaching of
McMorn P, Hutchings GJ (2004) Polymeric Membranes in the catalyst or these additives from the membrane
Catalytic Reactors for Vankelecom. Chem Soc Rev can be a problem under certain liquid-phase reac-
33:108 tion conditions.
Vankelecom IFJ (2002) Heterogeneous enantioselective
catalysts: strategies for the immobilisation of homoge- When heterogeneous catalysts are
neous catalysts Paul McMorn and Graham J Hutchings immobilized in polymeric membranes, this is
Chem Rev 102(10):3795 most commonly done by dispersing them in
316 Catalyst Loading
the membrane casting solution. Key is then to

nd a good solvent for the polymer that initially Catalyst Retention by Membranes
allows creation of a good dispersion of the
heterogeneous catalyst and that is subsequently Ivo Vankelecom
evaporated fast enough to avoid settling of the Centre for Surface Chemistry and Catalysis,
catalyst particles, leaving an even membrane KU Leuven, Leuven, Heverlee, Belgium
behind.
When coupling catalysis to membrane separa-

tions, the catalyst retention expresses in general
References the amount of catalyst that does not pass the
membrane while ltering the reaction mixture
Vankelecom IFJ (2002) Catalyst Immobilization in Mem- over a membrane. Such mixture can either con-
branes. Chem Rev 102(10):3795
tain a homogeneous or heterogeneous catalyst.
The membrane used to retain it will then need to
have a pore size able to retain the dissolved/
dispersed catalyst by size. If dense membrane is
applied, good retention can also be realized by
Catalyst Loading
lowering the afnity of the membrane material
for the catalyst or the diffusion of the catalyst
Ivo Vankelecom
through it. The ltration can be driven by a
Centre for Surface Chemistry and Catalysis, KU
variety of driving forces (pressure, concentra-
Leuven, Leuven, Heverlee, Belgium
tion, electrical, thermal), depending on the reac-
tion type that is run and the conditions under
which the reaction is taking place. The separa-
The term catalyst loading refers to the amount
tion can run at the same time as the reaction in a
of catalyst present in a catalytic membrane, most
continuous hybrid process, or the removal of the
commonly expressed as mol/g or g/g.
catalyst can take place batchwise. In the latter
If the catalyst is incorporated in the membrane
case, the conditions for the separation can be
already during the membrane preparation
different from the conditions during reactions;
(in contrast to a post-synthesis catalyst deposi-
hence, both can be optimized separately
tion), dispersion problems can appear, especially
(Vankelecom 2002).
at higher loadings. This can be reected in aggre-
Retention and rejection by a membrane refer to
gate formation for heterogeneous catalysts and in
the same effect of withholding a compound by a
cluster formation or even precipitation for homo-
membrane, but by denition, retention refers to a
geneous catalysts. When the catalyst is present
concentration of the removed compound in the
during membrane preparation, side effects might
feed, while rejection refers to the concentration
complicate the preparation of the catalytic mem-
in the retentate (Mulder 1991). In continuous pro-
brane, such as interference of the catalyst with the
cesses or when using large feed volumes from
solvent evaporation or the polymer curing or set-
which only small quantities are ltered, both def-
tling of the catalyst.
initions become equal.
Many of such membrane-assisted catalytic
reactions have been described, where the mem-
References brane separation process is either pervaporation
(very popular to shift the equilibrium of the reac-
Vankelecom IFJ (2002) Chem Rev 102(10):3795 tion to the right by removing the products of the
Catalytic (PtY) Membranes 317
reaction), vapor permeation, gas separation, are a mixture of CO and H2. Additional hydrogen
ultraltration (solvent resistant), nanoltration, can be recovered by a lower-temperature gas-
reverse osmosis, or dialysis. They all aim at an shift reaction. At the end of this process, the
easy recovery of the catalyst or at operating in amount of carbon monoxide is equal to 1 %.
continuous or semi-batch mode. A very high This CO amount is not tolerated by the PEM
catalyst retention is required, since the continu- fuel cells. In order to avoid cell performance
ous loss of small amounts of catalyst adds up degradation, carbon monoxide concentration C
seriously after treating several batches or in a must be kept below 10 ppm in the hydrogen
long-term continuous process (Brinkmann feed streams (Rohland and Plzak 1999); CO
et al. 1999). Tricks to increase the catalyst reten- selective oxidation (selox) is an interesting and
tion can be to enlarge the catalyst by creating economic approach for a deep purication of
polymeric, dendrimeric, or hyperbranched ver- H2-rich gas streams from carbon monoxide
sions of the homogeneous catalysts (Van Koten (Avgouropoulos and Ioannides 2003). During
et al. 2002; Muller et al. 2005). this process two competitive reactions take
place, carbon monoxide and hydrogen oxidation,
and so a selective catalyst is required to avoid
References hydrogen consumption. Among the various
catalysts used, the better is the platinum
Brinkmann N, Giebel D, Lohmer G, Reetz MT, Kragl
U (1999) J Catal 183:163 because it represents a good compromise
Koten G, Dijkstra HP, van Klink GPM (2002) Acc Chem between cost and performance. It demonstrated
Res 35(9):798 the possibility to use Pt-Y membranes as innova-
tive devices to perform the CO selective oxida-
ogy. Kluwer, Dordrecht
Muller C, Nijkamp MG, Vogt D (2005) Eur J Inorg Chem tion (Hasegawa et al. 2002). These systems
20:4011 permit to disperse catalytic nano-sized particles
Vankelecom IFJ (2002) Chem Rev 102(10):3795 in a thin zeolite layer assuring a good contact
between reactants and catalytic sites and at the
same time a low pressure drop and reduced
Catalytic (PtY) Membranes bypass (Fig. 1).
Catalytic zeolite membrane Pt-loaded demon-
Catia Algieri strated the possibility to reduce the CO concen-
Institute on Membrane Technology, National tration from 10,000 ppm down to 1050 ppm
Research Council of Italy, ITM-CNR, Rende, (Bernardo et al. 2006, 2008). The membrane
Italy with the best permeation properties exhibited the
best catalytic performance (amount of CO at the
outlet stream equal to 10 ppm). A negligible effect
Today, about the 6070 % of hydrogen is proof the feed pressure was obtained with the
duced by steam reforming. The reaction products low-permeance membranes. On the contrary,
Catalytic (PtY)
Membranes,
Fig. 1 Scheme of a
catalytic zeolite membrane
reactor
318 Catalytic Crystals
positive effect was found by using more perme- aluminum (Szostak 1998). The resulting structure
able membranes. The effect of the pressure on CO is very regular with pore size at molecular scale.
conversion can be explained considering two dif- Their adsorption properties change varying the
ferent paths for the reactants: through the zeolite Si/Al ratio during the synthesis because it deter-
channels (rst path) and defects (second path). CO mines the number of cations in the structure and
molecules permeating into the zeolite channels so their hydrophilicity (Szostak 1998). When the
interact with the catalyst (Pt) with very high prob- cation is a proton, a Brnsted acid site is formed in
ability owing to the narrow space of the zeolite the zeolite structure (Abott and Guerzoni 1992).
pores. In this case, the effect of pressure is negli- In addition, different transition metals can be
gible on the conversion. On the contrary, the inter- loaded (by cationic exchange) into the zeolite
action of the CO molecules with the catalyst channels (Howe 2004). These inorganic crystals
depends on the pressure owing to the large vol- are used at industrial level as adsorbents for the
ume of the support pores. purication and separation of different substances
and as exchangers in laundry detergents
(Smulders et al. 2003). Furthermore, they are
largely used as catalysts (Weitkamp 2000). In
References catalysis, the best performances of the zeolite
with respect to the other systems are the high
Avgouropoulos G, Ioannides T (2003) Selective CO oxi- concentration of acid sites and their high thermal
dation over CuO-CeO2 catalysts prepared via the urea-
nitrate combustion method. Appl Catal A 224:155167
and hydrothermal stability. Besides, their shape
Bernardo P, Algieri C, Barbieri G, Drioli E (2006) Catalytic selectivity plays a very important role to orient
(Pt-Y) membranes for the purication of H2-rich the reaction toward the production of the desired
streams. Catal Today 118:9097 product (Chen et al. 1989). They are used in
Bernardo P, Algieri C, Barbieri G, Drioli E (2008) Hydro-
gen purication from carbon monoxide by means of
different acid-catalyzed hydrocarbon transforma-
selective oxidation using zeolite catalytic membranes. tions such as cracking, isomerization, and alkyl-
Sep Purif Technol 62:629635 ation which are very important reactions in the
Hasegawa Y, Sotowa K-I, Kusakabe K, Morooka S (2002) petrochemical industry, and increasing attention is
The inuence of feed composition on CO oxidation
using zeolite membranes loaded with metal catalysts.
also focused in environmental catalysis and syn-
Microporous Mesoporous Mater 53:3743 thesis of ne chemicals (Xu et al. 2006). The
Rohland B, Plzak V (1999) The PEMFC-integrated CO oldest and largest application of the catalytic zeo-
oxidation-a novel method of simplifying the fuel cell lites in oil rening is the uid catalytic cracking
plant. J Power Sources 84:183186
(Martnez and Corma 2011). Hydrocracking is the
second largest application of zeolite in this eld.
In this last process, zeolites are used for their high
stability toward the deactivation by coking or by
Catalytic Crystals nitrogen compounds (Ward 1993). In the petro-
chemical process, very relevant is the production
Catia Algieri of the most important isomer of the xylene (para-
Institute on Membrane Technology, National xylene) because it is used for the production of
Research Council of Italy, ITM-CNR, Rende, terephthalic acid that in turn is used for the syn-
Italy thesis of polyester resin and bers. This process is
carried out by using ZSM-5 zeolite taking advan-
tage of its shape selectivity property and low
Zeolites are crystalline microporous materials. coking tendency (Corma et al. 2011). A further
They are composed of AlO45 and SiO4 tetrahe- development of zeolite chemistry (e.g., synthesis
dra that build a network of channels and cavities. of structures with more large pores) will open new
Extra-framework ions are located in the pores and possibilities for processes in chemical, ne chem-
cavity to balance the negative charge of the ical, and biomass transformations.
Catalytic Membranes 319
References
Catalytic Membranes
Abott J, Guerzoni FN (1992) Roles of Brnsted and Lewis
sites during cracking of n-octane on H-mordenite.
Jose M. Sousa
J Catal 85:173188
Chen NY, Garwood WE, Dwyer FG (1989) Shape selective Chemistry Department, School of Life and
catalysis in industrial applications. Marcel Dekker, Environment Sciences, University of
New York Tras-os-Montes e Alto Douro, Vila Real, C
Corma A, Portilla MT, Martnez C, Valencia S, LIopis FJ
Portugal
(2011) Appl Catal A Gen 393:257268
Howe RF (2004) Zeolite catalysts for dehalogenation pro- LEPABELaboratory for Process Engineering,
cesses. Appl Catal A Gen 27:311 Environment, Biotechnology and Energy,
Martnez C, Corma A (2011) Inorganic molecular Chemical Engineering Department, Faculty of
sieves: preparation, modication and industrial application
Engineering, University of Porto, Rua Roberto
in catalytic processes. Coord Chem Rev 255:15581580
Smulders E, Rahse W, von Rybinski W, Steber J, Sung E, Frias, Porto, Portugal
Wiebel F (2003) Toxicology. In: Smulders E (ed) Laun-
dry detergents. Wiley, New York
Szostak R (1998) Molecular sieves: principles of synthesis
A catalytic membrane is a membrane with cata-
and identication, 2nd edn. Blackie Academic and
Professional, London lytic properties. The catalytic activity can be
Ward J (1993) Hydrocracking processes and catalysts. Fuel intrinsic to the material itself, as in the case of
Process Technol 35:5585 membranes made of Pd, TiO2, and H-ZSM-5 zeo-
Weitkamp J (2000) Zeolite and catalysis. Solid State Ion
lite, which are catalytic for specic reactions, or
132:17518
Xu B, Sievers C, Hong SB, Prins R, van Bokhoven JA can be obtained by coating the external or the
(2006) Catalytic activity of Brnsted acid sites in zeo- internal (porous) surfaces of the membrane with
lites: intrinsic activity. J Catal 244:163168 the catalyst (metal or oxides) or even by occluding
the catalyst (metal nanoclusters, zeolites, acti-
vated carbon, metal complexes) inside a dense
polymer matrix (Gryaznov 1986, 1992; Irusta
Catalytic Diffuser et al. 1998; Gobina and Hughes 1994; Saracco
and Specchia 2000; Itoh 2000; Dittmeyer
Contactor-Type Catalytic Membrane Reactor
et al. 2001; Piera et al. 2001; Julbe 2005; van
Dyk et al. 2003; Marcano and Tsotsis 2002; Basile
2012; Fritsch and Peinemann 1995).
Catalytic Membrane Bioreactors with Catalytic membranes can be inorganic
Biocatalysts Segregated Within the (metallic, ceramic, or carbon made) or polymeric
Pores of Asymmetric Membranes in their nature. Examples of dense inorganic cat-
alytic membranes are the ones made of palladium,
Biocatalytic Membrane Reactors (BMR) with which are catalytically active for hydrogenation
Biocatalysts Entrapped Within the Pores of reactions (though presenting low activity), while
Asymmetric Membranes porous catalytic inorganic membranes can be
made of a variety of materials, namely, alumina,
silica, titania, and zeolites among others. Porous
polymeric membranes are usually made, for
Catalytic Membrane Reactors with example, of polysulfone, polyacrylonitrile, or
Biocatalysts Segregated Within the polypropylene, while dense polymeric mem-
Pores of Asymmetric Membranes branes are prepared from PDMS
(polydimethylsiloxane) and other silicones, PVA
Biocatalytic Membrane Reactors (BMR) with (polyvinyl alcohol), peruoropolymers,
Biocatalysts Entrapped Within the Pores of polyimides, or polyamides among other polymers
Asymmetric Membranes (Basile and Gallucci 2011; Basile 2012).
320 Catalytic Membranes
Membranes in general can be permselective or the case of zeolites, surface diffusion has a rele-
nonpermselective. Catalytic permselective mem- vant contribution. Also in these cases the mem-
branes are characterized by two important param- branes are permselective (Basile and Nunes
eters concerning their separation ability: 2011).
Permselectivity, which describes the ability of A catalytic membrane is used to carry out
the membrane to transport the different compo- different types of reactions, namely, chemical,
nents of a mixture at different rates, and the per- biochemical, electrochemical, or photocatalytic,
meability, which quanties the total amount in a so-called membrane reactor. The type of
permeated by the membrane when subjected to membrane to be used, porous or dense, polymeric
specic operation conditions. Nonpermselective or inorganic, etc., depends on several factors,
membranes can be described by their permeabil- namely, the type of reaction. For example, dense
ity, but they are not able to discriminate between polymeric catalytic membranes are used for
the different components of a mixture (Marcano low-temperature reactions where the
and Tsotsis 2002). permselectivity is important for shifting the con-
The mass transport mechanism suitable to version beyond the thermodynamic equilibrium
describe the permeation rate through a (catalytic) value based on feed conditions (e.g., selective
membrane is primarily a function of the type of removal of water in esterication reactions) or to
membrane. For dense membranes, the accepted enhance the reaction selectivity (e.g., hydration,
model is the sorption (solution) diffusion (the epoxidation, isomerization, and hydrogenation
permeant species in gas or liquid phase sorbs reactions). Dense polymeric catalytic membranes
into the membrane at the higher chemical poten- can also be used as contactors for promoting reac-
tial side, diffuses through the membrane subjected tions between two immiscible phases with segre-
to the driving force gradient, and desorbs on the gated feed (e.g., oxyfunctionalization of
opposite side). These membranes are usually hydrocarbons with hydrogen peroxide reactions)
permselective. Concerning porous membranes, (Basile and Gallucci 2009).
the total mass transport results from different con- Another type of reaction that has been studied
tributions, namely, a diffusive component is the photocatalytic removal of contaminants
(activated, surface, Knudsen and bulk) and a vis- from water, either using porous or dense poly-
cous component (Poiseuille ow). The contribu- meric catalytic membranes. The catalyst typically
tion of each one depends on the operating used is TiO2, occluded or coated, but also
conditions and on the pores size. For pore radius occluded Fe in an H2O2 oxidant medium (photo-
between 1 and 50 nm, Knudsen diffusion is nor- Fenton process) has been considered (Basile and
mally the predominant mass transfer mechanism Nunes 2011).
(Knudsen diffusion is characterized by the mean- Catalytic nonpermselective porous mem-
free path of the traveling species being much branes, polymeric or inorganic, can be used to
larger than the pore radius and is independent of improve the contact between the reactants and
any pressure gradient along the pore), and the the catalyst, in order to obtain higher reaction
membranes present some permselectivity, related conversions. Particularly in the case of triphasic
with the ratio of the species molar mass. For (gas/liquid/solid) reactions, usually limited by the
membranes with larger pores, namely, diffusion of the gaseous reactant, porous mem-
macroporous, the mass transfer mechanism is a brane reactors show conversion advantages. By
viscous ow if a total pressure gradient across the controlling the pressure of the gas and liquid
porous membrane is present or bulk diffusion if ows, it is possible to shift the reactants to meet
not. Neither of these situations allows any the catalyst zone (Fig. 1).
permselectivity. When the membrane has micro- The same segregated feeding strategy can also
pores and the species adsorb signicantly, as in be used in cases of reactions where a strict control
Catalytic Membranes 321
Catalytic Membranes,
Fig. 1 Porous catalytic
membrane for wastewater
treatment (Reprinted with
permission from Raeder
2010)
of the reactants is important, namely, in the case of Gryaznov VM (1992) Platinum metals as components of
very fast or highly exothermic reactions. catalyst-membrane systems. Platinum Met Rev 36:70
Irusta S, Pina MP, Menendez M, Santamaria J (1998)
In the case of reactions where a complete Development and application of perovskite-based cat-
conversion of some components, like VOCs, is alytic membrane reactors. Catal Lett 54:69
of high importance, porous catalytic membranes, Itoh N, Haraya K (2000) A carbon membrane reactor. Catal
permselective or not, are usually used in a ow- Today 56:103
Julbe A (2005) Zeolite membranes a short overview. In:
through mode, that is, all the reactants are fed Cejka J, van Bekkum H (eds) Zeolites and ordered
conjointly. In these cases, it is critical controlling mesoporous materials: progress and prospects. Studies
the residence time of the reactants to guarantee a in surface science and catalysis, Elsevier science &
complete conversion (Westermann and Melin technology, vol 157. p 135
Marcano JS, Tsotsis TT (2002) Catalytic membranes and
2009). membrane reactors. Wiley, Weinheim
Piera E, Tellez C, Coronas J, Menendez M, Santamaria
J (2001) Use of zeolite membrane reactors for selectiv-
References ity enhancement: application to the liquid-phase oligo-
merization of i-butene. Catal Today 67:127
Raeder H (2010) Wastewater oxidation using catalytic
Basile A (ed) (2012) Handbook of membrane reactors. contactor a revolutionary catalytic membrane reactor
Woodhead, Cambridge, UK (in press) for wastewater treatment. SINTEF Materials and
Basile A, Gallucci F (eds) (2009) Modeling and simulation Chemistry. https://www.sintef.no/globalassets/upload/
of membrane reactors. Nova, New York materialer_kjemi/energikonvertering-og-materialer/
Basile A, Gallucci F (eds) (2011) Membranes for mem- dokumenter/watercatox-screen.pdf
brane reactors: preparation, optimization and selection. Saracco G, Specchia V (2000) Catalytic combustion of
Wiley, New York propane in a membrane reactor with separate feed of
Basile A, Nunes S (eds) (2011) Advanced membrane sci- reactants. IV. Transition from the kinetics- to the trans-
ence and technology for sustainable energy and envi- port-controlled regime. Chem Eng Sci 55:3979
ronmental applications. Woodhead, Cambridge, UK van Dyk L, Miachon S, Lorenzen L, Torres M, Fiaty K,
Dittmeyer R, Hollein V, Daub K (2001) Membrane reac- Dalmon JA (2003) Comparison of microporous MFI
tors for hydrogenation and dehydrogenation processes and dense Pd membrane performances in an extractor-
based on supported palladium. J Mol Catal A Chem type CMR. Catal Today 82:167177
173:135184 Westermann T, Melin T (2009) Flow-through catalytic
Fritsch D, Peinemann K-V (1995) Catalysis with homoge- membrane reactors principles and applications.
neous membranes loaded with nanoscale metallic clus- Chem Eng Process 48:1728
ters and their preparation. Catal Today 25:277283
Gobina E, Hughes R (1994) Ethane dehydrogenation using
a high-temperature catalytic membrane reactor. Further Reading
J Membr Sci 90:11 Seidel-Morgenstern A (ed) (2010) Membrane reactors:
Gryaznov VM (1986) Hydrogen permeable palladium distributing reactants to improve selectivity and yield.
membrane catalysts. Platinum Met Rev 30:68 Wiley, New York
322 Catalytic Methanol Steam Reforming
surplus of steam is required. Therefore, practical

Catalytic Methanol Steam Reforming systems operate at S/C ratios between 1.3 and
2.0. Because an excess of steam is required to
G. Kolb maintain catalyst durability, apart from hydrogen
Product Sector Energy and Chemistry, Fraunhofer and carbon dioxide, the reformate contains sig-
ICT-IMM, Mainz, Germany nicant amounts of unconverted steam and to a
lower extent some unconverted methanol and
Methanol receives increasing interest as future carbon monoxide. The latter is (at least for con-
energy carrier for distributed and mobile power ventional Cu/ZnO catalysts) formed by the
generation, because it can be produced via sus- reverse water-gas shift reaction as shown in
tainable routes and has a high storage stability and Eq. 2 (Fig. 1):
energy density when compared with liqueed or
compressed hydrogen. When treated by reforming
CO2 H2 ! CO H2 Og DH0298
technology, it serves as a hydrogen source for fuel
cells (Kolb 2008). kJ
40:4 (2)
Methanol is converted to hydrogen-containing mol
reformate by endothermic steam reforming (STR)
(Kolb 2008) as shown in Eq. 1: When not only steam but also air is fed to the
reformer, the reaction is named oxidative metha-
nol steam reforming:
CH3 OHl H2 Ol ! CO2
3 H2 DH0298 (1) CH3 OH S=CH2 O 0:5 O=CO2
kJ 0:579=21N2 ! CO2 3 O=CH2
131
mol S=C O=C 1H2 O 0:579=21O=CN2
The highest hydrogen concentration in the (3)

reformer product is achieved at S/C 1. In order
to reduce the carbon monoxide formation and S/C is the molar steam to carbon (here metha-
to avoid coke deposition on the catalyst, a nol) ratio; O/C is the ratio of atomic oxygen in the
Catalytic Methanol
Steam Reforming, 70
Fig. 1 Equilibrium gas H2
30
composition of methanol
60
steam reforming vs. S/C
ratio of the feed; pressure CO
5 bar; temperature 200 C 50
(Schmidt et al. 1994)
Content / %
Content / %
20
40 CO2
30
H2O
20 10
CH3OH p = 5 bar
10
T = 200C
0 0
0 1 2 3 4
M = H2O/CH3OHratio [mol/mol]
Catalytic Methanol Steam Reforming 323
oxidant to carbon in the fuel. The hydrogen con- methanol dehydration and is also observed as a
tent in the reformate is then derived from the by-product:
following equation:
2CH3 OH ! CH3 OCH3 H2 O (10)
xH2 3 O=C=3 S=C
(4)
0:579=21O=C The reaction is thermodynamically favored by C
lower reaction temperatures (Lwin et al. 2000).
Catalyst development for methanol steam
Oxidative steam reforming involves also the
reforming as hydrogen source for fuel cells had
exothermic partial oxidation of methanol:
been mainly focused on ZnO-containing catalysts
such as Cu/ZnO and Pd/ZnO (either supported by
CH3 OHl 0:5 O2 ! CO2
the ZnO itself or b alumina carrier) owing to their
2 H2 DH0298 (5) high activity and low selectivity toward carbon
kJ monoxide. However, these catalyst formulations
155
mol show a number of drawbacks. Once converted
The enthalpy of the reaction at a temperature into their active (reduced) state, Cu/ZnO catalysts
T is: show pyrophoricity when exposed to an oxidizing
environment. At reaction temperatures exceeding
DH0298 1 O=CDHTR STR 300 C, the catalyst deactivates, which originates
from copper sintering. Cu/ZnO is an excellent
O=CDHTR POx (6)
water-gas shift catalyst, which is its current main
For oxidative steam reforming with S/C = 1.5 industrial application eld. Consequently, carbon
and O/C = 0.2 at 350 C reaction temperature, monoxide is formed by the reverse water-gas shift
the wet reformate has a hydrogen content of reaction as soon as all the methanol is converted,
57.4 Vol.%, and the reaction enthalpy amounts which happens under conditions of partial load of
to 15 kJ/mol, i.e., the reaction is still endothermic. a practical reformer reactor. Pd/ZnO catalysts are
Heat losses are not included into the calculations higher in activity when compared to Cu/ZnO,
above which means that additional energy is mainly because this catalyst type allows higher
required to compensate for. In a practical, self- operating temperature. Even at reaction tempera-
sustaining reactor, this energy could be gained tures exceeding 300 C, the selectivity toward
from additional oxygen supply to the feed or by carbon monoxide remains low, because a Pd/Zn
separate or integrated anode off-gas combustion. alloy is formed (Chin et al. 2002, 2003). However,
Frequently observed by-products of methanol Pd/ZnO catalysts are known to bear always the
steam reforming are formic acid and methyl for- risk of metallic palladium formation, which results
mate (HCOOCH3), which are not tolerated by the in carbon monoxide formation much higher than
anode electrocatalyst of low-temperature PEM the values formed over Cu/ZnO or Pd/Zn alloy.
fuel cells. They are intermediate products of meth- Pd/In2O3/Al2O3 and Pt/In2O3/Al2O3 catalysts are
anol steam reforming (Takahashi et al. 1985): an alternative to the catalysts discussed above
(Men et al. 2010; Kolb et al. 2011). They are
2CH3 OH ! HCOOCH3 2H2 (7) even more active than Pd/ZnO.
HCOOCH3 H2 O ! CH3 OH HCOOH (8)

References
HCOOH ! CO2 H2 (9)
Chin Y-H, Dagle R, Hu J, Dohnalkova AC, Wang Y (2002)
Formic acid decomposition is a fast reaction Steam reforming of methanol over highly active
(Jiang et al. 1993). Dimethyl ether is formed by Pd/ZnO catalyst. Catal Today 77:7988
324 Catalytic Reactions, Membrane Operations of
Chin Y-H, Wang Y, Dagle RA, Li XS (2003) Methanol processes can be synergistically combined
steam reforming over Pd/ZnO: catalyst preparation and with catalytic reactors targeting conversion,
pretreatment studies. Fuel Process Technol 83:193201
Jiang CJ, Trimm DL, Wainwright MS (1993) Kinetic selectivity, or safety enhancements. These effects
mechanism for the reaction between methanol and may be obtained by selective product extraction
water over a Cu-ZnO-Al2O3 catalyst. Appl Catal and purication, normally originating conver-
A 97:145158 sion and selectivity enhancements, or by segre-
Kolb G (2008) Fuel processing for fuel cells. Wiley-VCH,
Weinheim gated or distributed feed of reactants, aiming
Kolb G, Keller S, Pecov S, Pennemann H, Zapf R (2011) selectivity and/or safety enhancements. When
Development of micro-structured catalytic wall reac- a reactor is synergistically combined with a
tors for hydrogen production by methanol steam membrane process, the unit is called membrane
reforming over novel Pt/In2O3/Al2O3 catalysts chemi-
cal. Eng Trans 24:133138 reactor. Usually, membrane and reactor are
Lwin Y, Daud WRW, Mohamad AB, Yaakob Z (2000) integrated in the same housing (Marcano and
Hydrogen production from methanol steam reforming: Tsotsis 2002).
thermodynamic analysis. Int J Hydrogen Energy Membranes can be inorganic (metallic,
25:4753
Men Y, Kolb G, Zapf R, OConnell M, Ziogas A (2010) ceramic, carbon) or polymeric in their nature.
Methanol steam reforming over bimetallic Pd-In/Al2O3 Membranes can be permselective or
catalysts in a microstructured reactor. Appl Catal non-permselective. Permselective membranes
A 380:1520 are characterized by two important parameters:
Schmidt VM, Brckerhoff P, Hhlein B, Menzer R,
Stimming U (1994) Utilization of methanol for poly- permselectivity, which describes the ability of
mer electrolyte fuel cells in mobile systems. J Power the membrane to transport the different compo-
Sources 49:299312 nents of a mixture at different rates, and perme-
Takahashi K, Kobayashi H, Takezawa N (1985) Chem Lett ability, which quanties the total amount
6:759
permeated by the membrane when subjected to
specic operation conditions, normalized by the
membrane thickness. The mass transport mecha-
nism is described by the sorption diffusion
Catalytic Reactions, Membrane models for dense and microporous membranes.
Operations of For porous, different contributions should be con-
sidered: diffusive (activated, surface, Knudsen,
Adelio Mendes and bulk) and viscous (Poiseuille ow). The con-
Faculty of Engineering, University of Porto, tribution of each one depends on the operating
Porto, Portugal conditions, pore size, penetrant mass and size
and the surface, and penetrant nature (Basile and
Gallucci 2011).
Catalytic reactions (chemical, biochemical, pho- There are various potential industrial applica-
tochemical, electrochemical, and photoelectro- tions that take advantage of combining a catalytic
chemical) are promoted or enhanced in their rate reactor with a membrane process, integrated or
by a catalyst. Most of the chemical catalytic reac- not in the same device, operating either in liquid
tions are heterogeneous; the catalyst is normally a and/or gas phases. One of the main membrane
solid, while reactants are in uid phase. In the studied functions has been the selective removal
petrochemical industry, for example, the catalyst of components from the reaction medium.
is usually composed of micro- or nanoclusters of a Biorening and biofuel production, for example,
noble metal in a support, usually a metal oxide. In take advantage of process integration of a biore-
the ne chemical and organic synthesis, on the actor and a permselective membrane, as is the
other hand, it is common that the reactions be case of bioethanol and acetic acid production: the
homogeneous. continuous removal of the main product from the
Membrane is a permselective medium or reaction medium decreases or even eliminates
interface between two uid phases. Membrane the potential reaction inhibition (Ma et al. 2009;
Catalytic Reformer Off-Gas 325
He et al. 2012). Also the removal of fermentation Basile A, Nunes S (eds) (2011) Advanced membrane sci-
inhibitors generated during the pretreatment pro- ence and technology for sustainable energy and envi-
ronmental applications. Woodhead Publishing Limited,
cess of lignocellulosic material for the second- Cambridge
generation bioethanol production is important, He Y, Bagley DM, Leung KT, Liss SN, Liao BQ
because of the negative impact in the (2012) Recent advances in membrane technologies
ethanol yield and productivity and in the cell for biorening and bioenergy production. Biotechnol
growth inhibition (He et al. 2012). Still in the
Adv 30:817858
Ma Y, Wang J, Tsuru T (2009) Pervaporation of water/
C
selective removal of a reaction product from the ethanol mixtures through microporous silica mem-
reaction medium, it can be referred to as the branes. Sep Purif Technol 66:479485
selective removal of water in esterication Marcano JS, Tsotsis TT (2002) Catalytic membranes and
membrane reactors. Wiley-VCH Verlag GmbH,
reactions and the selective removal of hydrogen Weinheim
in dehydrogenation and water-gas shift reactions Seidel-Morgenstern A (ed) (2010) Membrane reactors:
for a conversion shifting beyond the thermody- distributing reactants to improve selectivity and yield.
namic equilibrium based on feed conditions. Wiley, New York
In case of using palladium-silver membranes in
the dehydrogenation and water-gas shift reac- Further Reading
Basile A (ed) (in press) Handbook of membrane reactors.
tions, high-purity hydrogen can be obtained Woodhead Publishing Limited, Cambridge
(Marcano and Tsotsis 2002; Basile and
Nunes 2011).
The membrane can also play a role of a selec-
tive distributor, dosing a reactant along the reac-
tion medium. This approach has been extensively Catalytic Reformer Off-Gas
used in consecutive-parallel reaction schemes,
especially in partial oxy-dehydrogenation or oxi- G. Kolb
dation of alkanes and oxidative coupling of Product Sector Energy and Chemistry, Fraunhofer
methane, using dense ceramic or metallic mem- ICT-IMM, Mainz, Germany
branes permselective to oxygen (silver, yttrium-
stabilized zirconia, perovskites and related oxides,
or composite membranes involving a mixture of The product mixture formed during reforming of
ionic and electronic conducting materials, usually all kind of fuels such as hydrocarbons and alco-
oxides and metals). Separating the hydrocarbon hols is named reformate (Kolb 2008). The prod-
and oxygen feed, the reaction is carried on in ucts of the reforming reaction are hydrogen and
much safer conditions, and the possibility of ther- carbon monoxide (see entry Catalytic
mal optimization and oxygen concentration con- Reforming of Methane and Other Hydrocarbons)
trol along the reaction medium can be used to unless methanol is applied as feed, which pro-
improve the selectivity for the intermediate duces mainly carbon dioxide instead. The thermo-
desired oxygenated product. Though in less dynamic equilibrium of the reforming reaction
extension, also proton-conducting membranes itself is equivalent to full conversion of all kinds
have been considered as distributors for hydroge- of fuels except for methane. The conversion of
nation reactions (Marcano and Tsotsis 2002; methane is limited by the equilibrium of the
Seidel-Morgenstern 2010). methanation reaction, which is the reverse reac-
tion of methane steam reforming:
References 3H 2 CO ! H 2 O CH 4
Basile A, Gallucci F (eds) (2011) Membranes for mem- kJ

brane reactors: preparation, optimization and selection. DH R 253:7 (1)
mol
Wiley, New York
326 Catalytic Reformer Off-Gas
Catalytic Reformer 80
Off-Gas, Fig. 1 Dry
reformate composition of
isooctane partial oxidation 70
in the presence of steam
(nitrogen excluded); 60
S/C = 3.0; O/C = 1.0;
Product Distribution(%)
GHSV = 8,776 h1 (Moon
H2
et al. 2001) 50
CO
CO2
40
CH2
30
20
10
0
450 500 550 600 650 700 750 800
Temperature (C)
The formation of methane is favored by lower In case the reformate is dedicated as hydrogen
temperature and elevated pressure, and therefore supply for PEM fuel cells, the carbon monoxide
usually reformate from higher hydrocarbon content has to be minimized. The situation is
reforming contains some methane (see Fig. 1). different, when high temperature fuel cells such
Methane is tolerated by most fuel cell types up as SOFC are applied, because they are capable of
to a concentration of 5 vol.%. Over sufciently converting carbon monoxide.
active catalyst and at reaction temperatures Further undesired by-products, which are fre-
exceeding 600 C, a second reaction occurs up quently found in reformate originating from alco-
to its thermodynamic equilibrium, the water-gas hol fuels, are acetaldehyde, ethers such as dimethyl
shift reaction: ether and acetone. Light hydrocarbons, namely,
ethylene and propylene, are frequent by-products
CO H2 O ! CO2 H 2 of higher hydrocarbon reforming and an indicator
for catalyst activity losses which then result in
kJ
DH 0298 40:4 (2) release of unconverted fuel later on.
mol Unconverted fuel and by-products are
undesired components in reformate because they
The thermodynamic equilibrium of the
cannot be converted by fuel cells and reduce the
reaction is favored by lower temperature (see
efciency of the energy conversion process. On
Fig. 1).
top of that, they also poison the fuel cell catalyst in
Reformate contains besides hydrogen and car-
most cases.
bon monoxide, depending on type of reforming
Because hydrocarbon fuels frequently contain
applied, in case of steam reforming unconverted
sulfur components, which originate from intended
steam and for partial oxidation signicant amounts
odorant addition in the case of gaseous fuels such
of nitrogen originating from the air feed. Reformate
as natural gas (methane) and liqueed petroleum
originating from autothermal reformers contains
gas (LPG, propane/butane) and from the sulfur
both unconverted steam and nitrogen because
species present in the crude oil for higher
both steam and air are fed to the process.
Catalytic Reforming of Methane and Other Hydrocarbons 327
hydrocarbons, a sulfur removal is required when References

fuel cells are connected to the reformer. While
gaseous fuels can be relatively easily puried Kolb G (2008) Fuel processing for fuel cells. Wiley-VCH,
Weinheim
through adsorption processes, sulfur removal
Moon DJ, Sreekumar K, Lee SD, Lee BG, Kim HS (2001)
from liquid fuels is more complicated, mostly Studies on gasoline fuel processor system for fuel-cell
owing to the lower adsorption capacity in the powered vehicles application. Appl Catal A 215:19
liquid phase. Because the reforming catalyst itself C
is normally rather sulfur tolerant and converts
sulfur species to hydrogen sulde, the sulfur
removal can be performed through conversion
over zinc oxide in the gas phase downstream the Catalytic Reforming of Methane
reformer: and Other Hydrocarbons
H 2 S ZnO ! ZnS H 2 O (3) G. Kolb

Product Sector Energy and Chemistry, Fraunhofer
To convert the carbon monoxide which is poi- ICT-IMM, Mainz, Germany
soning for low-temperature PEM fuel cells to
carbon dioxide, the water-gas shift reaction (see
Eq. 2) is applied in dedicated reactors downstream Reforming of hydrocarbons is possible by
the reformer, which operate at lower temperature conversion with either oxygen from air (partial
to shift the equilibrium of the reaction. oxidation), steam, or a combination of both
To reach the low concentration of carbon mon- named oxidative steam reforming. Only in the
oxide required for low-temperature PEM fuel cells, case of a self-sustaining operation of an oxidative
which is in the range between 10 and 100 ppm, steam reformer, when the partial oxidation reac-
preferential oxidation (see Eq. 4) or selective tion produces sufcient heat for preheating the
methanation reactions (see Eq. 1) are applied: feed, supply of the endothermic steam reforming
with energy, and compensation of the heat losses
CO 0:5 O2 ! CO2 of the reactor to the environment, this process
should be named autothermal reforming (Kolb
kJ
DH 0298 283 (4) 2008).
mol The general formula of hydrocarbon steam
reforming is provided below:
The preferential oxidation reaction is
accompanied by the undesired side reaction of y
part of the hydrogen present in the reaction Cx Hy x H 2 O ! x CO x H 2 (1)
mixture: 2
H 2 0:5 O2 ! H 2 O The product mixture of the endothermic reac-

tion is named reformate, which contains besides
kJ
DH 0298 242 (5) hydrogen and carbon monoxide some
mol unconverted steam, possibly unconverted fuel
Usually preferential oxidation requires a mini- and carbon dioxide, which is formed through con-
mum of excess air corresponding to an atomic secutive water-gas shift:
oxygen to carbon monoxide O/CO ratio between
1.5 and 2.0. Under these conditions and full con- CO H 2 O ! CO2 H 2
version of carbon monoxide, between 0.5 and
1.0 mol of hydrogen is lost for each mole carbon kJ
DH 0298 40:4 (2)
monoxide converted. mol
328 Catalytic Reforming of Methane and Other Hydrocarbons
Catalytic Reforming
of Methane and Other
Hydrocarbons,
Fig. 1 Equilibrium
conversion of methane
steam reforming versus
reaction temperature for
different S/C ratios and
system pressure (Joensen
and Rostrup-Nielsen 2002)
This reaction obviously increases the hydrogen requirements of processes switched downstream
concentration of the reformate. Under operating the reformer.
conditions of hydrocarbon steam reforming over While methane steam reforming is a reversible
sufciently active catalyst, thermodynamic equi- reaction, this is not the case for other hydrocarbon
librium of water-gas shift is achieved. fuels.
Steam reforming of higher hydrocarbons Partial oxidation is the conversion of fuels to
shows also formation of methane through reverse synthesis gas under oxygen decient atmosphere:
methane steam reforming (methanation):
x
Cx H y O2 3:76N 2 ! x CO
2 (4)
3H 2 CO $ H 2 O CH 4 y x
H2 3:76 N 2
2 2
kJ
DH 0298 253:7 (3)
mol The reaction is signicantly faster than steam
reforming. It requires only air feed and no water
Increasing reaction temperatures and higher evaporation to raise steam, which simplies the
atomic steam to carbon (S/C) ratio increase the process. However, the carbon monoxide concen-
methane conversion according to the equilibrium tration is signicantly higher compared to steam
of methane steam reforming reaction, while reforming, and the dilution through the nitrogen
increasing system pressure has the opposite effect. from the air is considerable. When hydrocarbons
The equilibrium conversion of methane steam are converted by partial oxidation, some combus-
reforming at various S/C ratio and system pres- tion occurs as an undesired side reaction (Joensen
sure is shown in Fig. 1. However, elevated pres- and Rostrup-Nielsen 2002), especially because an
sure is required if the reforming process is excess of air is fed into practical systems. The
combined with membrane separation and indus- water produced by the combustion process allows
trial steam reforming plants work at operating water-gas shift to take place to a certain degree.
pressures exceeding 20 bar owing to similar Another typical by-product of partial oxidation
Catalytic Wall Reactor 329
Catalytic Reforming of Methane and Other Hydrocarbons, Fig. 2 Ethylene formation versus time on stream for
autothermal reforming of synthetic diesel (Kang et al. 2007)
is methane, which is formed according to reac- applying ultrasonic injector for effective fuel delivery.
tion (3). Coke formation is a critical issue in partial J Power Sources 172:845852
Kolb G (2008) Fuel processing for fuel cells. Wiley-VCH,
oxidation. It is formed through the reaction of Weinheim
carbon monoxide with hydrogen, the Boudouard Yoon S, Kang I, Bae J (2009) Suppression of ethylene-
reaction, and cracking of higher hydrocarbons. induced carbon deposition in diesel autothermal
Figure 2 shows the breakthrough of reforming. Int J Hydrog Energy 34:18441851
unconverted fuel for oxidative steam reforming
of synthetic diesel fuel. Catalyst deactivation dur-
ing reforming of heavier hydrocarbons is indi-
cated by the increasing formation of light Catalytic Wall Reactor
hydrocarbons, predominantly ethylene. Yoon
et al. (2009) explained the preferred ethylene for- G. Kolb
mation in mixtures of parafnic and aromatic Product Sector Energy and Chemistry, Fraunhofer
feedstock by homogeneous decomposition of par- ICT-IMM, Mainz, Germany
afnic components. The initial linear increase of
the breakthrough is followed by an exponential
increase later on. The micro-reactors discussed in the current sec-
tion cover plate-heat-exchanger technology (see
Fig. 1 (top) with channels in the meso- and micro-
References scale, which are coated with catalyst in one or
both ow paths, which makes them suitable for
Joensen F, Rostrup-Nielsen JR (2002) Conversion of heterogeneously catalyzed reactions.
hydrocarbons and alcohols for fuel cells. J Power
If only one ow path of the heat-exchanger is
Sources 105:195201
Kang I, Bae J, Yoon S, Yoo Y (2007) Performance coated with catalyst, it transforms into a catalytic
improvement of diesel autothermal reformer by wall heater as shown in Fig. 1 (center). Catalytic
330 Catalytic Wall Reactor
a Exothermic Heat millimeter. In such channels the ow regime is

Reaction Removal usually laminar, diffusion paths for heat and mass
transfer are short, and the surface to volume ratio is
T small. Consequently surface effects dominate vol-
umetric effects. Compared to macroscopic devices,
Conventional a higher share of wall material exists. Thus the heat
Heater
transfer through the wall material contributes sig-
nicantly to the overall heat transfer and requires
b Exothermic Heat
consideration when designing such reactors.
Reaction on Removal
Wall In practical micro-reactors suited for gas phase
T reactions, the channel dimensions are ranging
between 250 mm and 1,000 mm height and from
Catalytic 250 mm to several millimeters width. Such micro-
Wall Heater reactors carry normally thousands of channels
operated in parallel. The high number and rela-
c Exothermic Endothermic
tively large dimensions of the channels create a
Reaction on Reaction
Wall pressure drop in the order of few to few tens of
T mbar. The high degree of parallelization makes
the scale up of such micro-reactors relatively sim-
Catalytic ple provided that the equipartition of the uid ow
Wall Reactor through all channels is ensured.
Obviously straight (micro-)channels are the
Catalytic Wall Reactor, Fig. 1 Threedifferent reactor
concepts: conventional heater (top), catalytic wall heater simplest possible design. Other geometries such
(center), and catalytic wall reactor (bottom) (Redenius as sinusoidal channels or n geometries have the
et al. 2001) potential for increasing the heat transfer. Increas-
ing the heat transfer at constant pressure drop is
wall heaters improve the thermal management of not feasible for most channel geometries. Low
exothermic reactions such as combustion and par- pressure drop is critical for practical systems
tial oxidation of fuels and of reactions related to owing to the energy consumption required for
CO removal from synthesis gas dedicated as the compression of gases.
hydrogen source for fuel cells, namely, water-gas A high heat conductivity of the heat-exchanger
shift and preferential oxidation of CO. Alterna- wall material reduces the efciency of plate heat
tively, endothermic reactions can be supplied exchangers (Stief et al. 1999) especially if their
with energy in such a reactor (catalytic wall length is similar to their width because isothermal
cooler). temperature proles will be created especially for
When the catalyst is coated onto the walls of construction material of high heat conductivity
the second ow path, the plate heat exchanger such as aluminum. It is therefore recommended to
turns into a catalytic wall reactor as shown in increase the heat-exchanger length and to reduce
Fig. 1 (bottom). This reactor type allows the direct the wall material thickness and, if possible, to use
coupling of exothermic reactions (such as steam construction material of low heat conductivity if a
reforming) and exothermic reactions (such as cat- temperature prole is required in the catalytic wall
alytic combustion). The heat source and the heat reactor for reactions such as water-gas shift (Zalc
sink are then only separated by the few hundred to and Lfer 2002; Baier and Kolb 2007).
thousand micrometer metal foils between both However, heat-exchanger reactors for coupling
catalyst coatings. exothermic and endothermic reactions such as
Micro-reactors for practical applications are catalytic combustion and steam reforming,
chemical reactors carrying a multitude of channels respectively, should be operated under isothermal
with at least one dimension not exceeding one conditions. A cocurrent ow arrangement should
Catalytically Modified (Activated) MIEC Membranes 331
be chosen (Frauhammer et al. 1999) which has as Ba0.5Sr0.5Co0.8Fe0.2O3d (BSCF) and

been proven for practical applications (Wichert La0.6Sr0.4Co0.2Fe0.8O3d (LSCF), as these mate-
et al. 2011; OConnell et al. 2009). rials show the highest oxygen ux.
The amount of catalyst which can be introduced The mechanism of oxygen transport through
into the plate-heat-exchanger reactor is a drawback the MIEC membranes involves three different
more than counterbalanced by the better catalyst steps (Fig. 1):
utilization through the improved heat and mass C
transfer. However, catalysts of higher activity are
1. The oxygen dissociation and reduction on the
required to compensate the smaller quantities.
membrane surface (interface I)
2. The incorporation into the crystal lattice
References as oxygen anions and the migration
through the lattice via oxygen vacancy defect
Baier T, Kolb G (2007) Temperature control of the water- sites
gas shift reaction in microstructured reactors. Chem 3. The combination of the oxygen anions on the
Eng Sci 62:46024611
membrane sweep side (interface II)
Frauhammer J, Eigenberger G, von Hippel L, Arntz
D (1999) A new reactor concept for endothermic high-
temperature reactions. Chem Eng Sci 54:36613670
OConnell M, Kolb G, Schelhaas KP, Sch urer J, Tiemann D, Thus the oxygen permeation ux through a
Ziogas A, Hessel V (2009) Development and evaluation MIEC membrane is determined by the bulk diffu-
of a microreactor for the reforming of diesel fuel in the sion across the membrane and the surface
kW range. Int J Hydrogen Energy 34:62906303 exchange reactions.
Redenius JM, Schmidt LD, Deutschmann O (2001) Milli-
second catalytic wall reactors: I Radiant burner. AIChE The bulk diffusion resistance of the material
J 47:11771184 can be improved by reducing the membrane
Stief T, Langer O-U, Schubert K (1999) Numerical inves- thickness. The resistance from bulk diffusion
tigations of optimal heat conductivity in micro heat plays an important role when the membrane sep-
exchangers. Chem Eng Technol 21:297302
Wichert M, Men Y, OConnell M, Tiemann D, Zapf R, aration layer is relatively thick. The oxygen ux
Kolb G, Butschek S, Frank R, Schiegl A (2011) Self- is proportional to the reciprocal of the membrane
sustained operation and durability test of a 300 W-class thickness. So the ux can be improved by
micro-structured LPG fuel processor. Int J Hydrogen decreasing the separation layer thickness. The
Energy 36:34963504
Zalc JM, Lfer DG (2002) Fuel processing for PEM fuel membrane surface exchange reaction becomes
cells: transport and kinetic issues of system design. the rate-limiting step when the thickness of
J Power Sources 111:5864 the separation layer is decreased beyond a certain
value (LC). Then the oxygen ux can still be
further increased by improving the surface
Catalytically Modified (Activated) exchange kinetics. This can be done by applying
MIEC Membranes a porous layer of the membrane material itself
to increase the surface area or by adding a
Bart Michielsen and Marijke Jacobs catalytic coating of a material with enhanced
Sustainable Materials Management, Flemish oxygen exchange properties to the membrane
Institute for Technological Research VITO, surface.
Mol, Belgium An example of the rst case is shown in Fig. 2
where a 10 mm thick porous layer of BSCF is
shown on top of the dense separation layer.
Dense ceramic membranes which are made from The extra surface area improved the oxygen ux
mixed ionic and electronic conducting (MIEC) of the membrane by 210 %.
oxides can selectively separate oxygen from air Pd has better oxygen exchange properties than
at temperatures above 700 C. The most used LSCF itself. So another strategy, performed by
MIEC membrane materials are perovskites, such Yacou et al., is coating the membranes with Pd
332 Catalytically Modified (Activated) MIEC Membranes
Catalytically Modified
(Activated) MIEC
Membranes,
Fig. 1 Different sections
involved in oxygen
transport during oxygen
permeation (Sunarso
et al. 2008)
(Activated) MIEC
Membranes,
Fig. 2 Details of the
morphology of the oxygen
activation layer after the
permeation test (Baumann
et al. 2011)
Catalytically Modified (Activated) MIEC Membranes, Fig. 3 SEM micrographs of the surface of Pd modied LSCF
membranes (Yacou et al. 2011)
Cell Adhesion 333
(Activated) MIEC
Membranes,
Fig. 4 Percentage
improvement of the O2 ux
through some Pd modied
LSCF membranes
compared to the original
LSCF membrane, as a
C
function of the operating
temperature (Yacou
et al. 2011)
as shown in Fig. 3. This results in an even higher cell adhesion molecules, which are integral mem-
improvement ratio as shown in Fig. 4. This way brane proteins that have cytoplasmic, transmem-
the oxygen ux is improved by 350 %. So a small brane, and extracellular domains. The
coating layer can have a big effect on the mem- extracellular domains of adhesion molecules
brane oxygen ux without compromising the extend from the cell and bind to other cells or
mechanical stability. matrix by binding to other adhesion molecules
of the same type (homophilic binding), binding
References to other adhesion molecules of a different type
(heterophilic binding), or binding to an interme-
Baumann S et al (2011) Ultrahigh oxygen permeation ux diary linker which itself binds to other adhesion
through supported Ba00.5Sr0.5Co0.8Fe0.2O3d mem-
molecules. Adhesion molecules belong to four
Sunarso J et al (2008) Mixed ionic-electronic conducting major families: cadherins, immunoglobulin-like
(MIEC) ceramic-based membranes for oxygen separa- adhesion molecules, integrins, and selectins.
tion (review). J Membr Sci 320:1341 Cadherins cause adhesion via homophilic binding
Yacou C et al (2011) Palladium surface modied
to other cadherins in a calcium-dependent manner.
La0.6Sr0.4Co0.2Fe0.8O3d hollow bres for oxygen sep-
aration. J Membr Sci 380:223231 As is the case for their role in desmosomes and
adherens junctions, cadherins ultimately anchor
cells through cytoplasmic actin and intermediate
Cell Adhesion laments. Immunoglobulin-like adhesion mole-
cules are involved in both homophilic and
Loredana De Bartolo heterophilic binding.
Institute on Membrane Technology, National The well-studied members of this group are the
Research Council, ITM-CNR, Rende, Italy neural cell adhesion molecules (N-CAMs), which
are expressed predominantly in nervous tissue,
and the intercellular cell adhesion molecules
Cell adhesion is the binding of a cell to a surface, (ICAMs). Integrins are a diverse and large group
extracellular matrix (ECM), or another cell using of heterodimeric glycoproteins. The two subunits,
334 Cell Adhesion in Bioartificial Organs
designated as alpha and beta, both participate in medium (or secreted by the cells), or directly, e.g.,
binding. Integrins participate in cellcell adhesion by guiding cell spreading with suitable surface
and are of great importance in binding and inter- topography (De Bartolo et al. 2002). Such prop-
actions of cells with components of the extracel- erties resulted in being critical to cellsubstratum
lular matrix such as bronectin. Importantly, interaction and have to be considered in the choice
integrins facilitate communication between the of material suitable for biomedical device.
cytoskeleton and extracellular matrix, allow each
to inuence the orientation and structure of the
other. Selectins are expressed primarily on leuko-
References
cytes and endothelial cells and, like integrins, are
important in many host defense mechanisms De Bartolo L, Morelli S, Bader A, Drioli E (2002) Evalu-
involving those cells. In contrast to other cell ation of cell behaviour related to physico-
adhesion molecules, selectins bind to carbohy- chemical properties of polymeric membranes to be
drate ligands on cells, and the resulting binding used in bioarticial organs. Biomaterials
23(12):24852497
forces are relatively weak. Cell adhesion is De Bartolo L, Gugliuzza A, Morelli S, Cirillo B, Gordano
believed to be the rst and dominant step for cell A, Drioli E (2004) Novel PEEK-WC membranes with
growth. low plasma protein afnity related to surface
Cells adhere strongly to some materials, but free energy parameters. J Mater Sci Mater Med
15:877883
not to others (Hynes 2002). This is determined De Bartolo L, Rende M, Morelli S, Giusi G, Salerno S,
by the special structure of individual cell mem- Piscioneri A, Gordano A, Di Vito A, Canonaco M,
branes and material surface properties. The initial Drioli E (2008) Inuence of membrane surface proper-
cellular events that take place at the biomaterials ties on the growth of neuronal cells isolated form hip-
pocampus. J Membr Sci 325:139149
interface mimic to a certain extent the natural Grifn L, Naughton G (2002) Tissue engineering current
adhesive interaction of cells with the extracellular challenges and expanding opportunities. Science
matrix (Grifn and Naughton 2002). 259:10091014
Cell adhesion on a substrate such as a mem- Hynes RO (2002) Integrins: bidirectional, allosteric signal-
ing machines. Cell 10:673687
brane is a multistep process that involves, in
sequence, adsorption of ECM proteins onto the
surface, recognition of ECM components by cell
receptors, cytoskeletal rearrangements, and cell
spreading. In particular, immediately after the Cell Adhesion in Bioartificial Organs
biomaterial comes into contact with cell environ-
ment, protein adsorption to its surface occurs. Loredana De Bartolo
This happens within seconds, long before the Institute on Membrane Technology, National
rst cells reach the surface. Consequently, cells Research Council, ITM-CNR, Rende, Italy
almost never come into direct contact with the
material surface; they rather interact with the
layer of adsorbed proteins. This layer mediates Binding of cells to a surface, extracellular matrix,
the cell adhesion and also provides signals to the or other cells in an articial device is used to
cell through the cell adhesion receptors. The replace a nonfunctioning organ. Correct cellular
membrane properties (roughness, physicochemi- adhesion is essential in maintaining multicellular
cal, mechanical, and transport) especially the sur- structure and allows cells to interact dynamically
face characteristics play an important role in the with adjacent cells and the extracellular matrix. In
adhesion process (De Bartolo et al. 2004, 2008). bio articial organs the cell adhesion process
Surface free energy, electric charge, and morphol- involves cellcell and cellnatural or articial
ogy might all affect cell attachment and its behav- matrix/material that plays essential roles in overall
ior either indirectly, e.g., by controlling tissue architecture and proper physiological func-
adsorption of the proteins present in the culture tions of the device.
Cell Adhesion in Bioartificial Organs 335
The functional units of cell adhesion are typi- isolated cells are compartmentalized in a poly-
cally multiprotein complexes made up of three meric membrane which provides a number of
general classes of proteins: the cell adhesion mol- important functions for the success of these
ecules/adhesion receptors, the extracellular matrix devices. Membranes should act as barriers to
(ECM) proteins, and the cytoplasmic plaque/ immunocompetent species present in the patients
peripheral membrane proteins. The cell adhesion blood and should permit the rapid passage of key
receptors are usually transmembrane glycopro- metabolites such as nutrients and oxygen from the C
teins that mediate binding interactions at the extra- surrounding to the cell compartment (Curcio
cellular surface and determine the specicity of et al. 2005). In such devices cells are contacting
cellcell and cellECM recognition. They include with membranes, the surface properties of mem-
members of the integrin, cadherin, immunoglob- brane could affect the various bioresponses. Thus,
ulin, selectin, and proteoglycan (for example, one major approach of the materials scientists has
syndecans) superfamilies. At the extracellular sur- been to try to inuence the extent and the charac-
face, the cell adhesion receptors recognize and ter of the cell response by modifying the surface
interact with either other cell adhesion receptors composition and properties of the polymer. The
on neighboring cells or with proteins of the response of cells to different material properties is
ECM. ECM proteins are typically large glycopro- a complex process and even minute changes in
teins, including the collagens, bronectins, composition of the substrate produce amplied
laminins, and proteoglycans that assemble into differences in cell responses. Although surface
brils or other complex macromolecular arrays. properties are often derived from the bulk proper-
Owing to their binding to adhesion receptors, they ties of the materials, the bulk materials do not
can also be tightly associated with the cell surface. entirely dene them, because the used substrates
At the intracellular surface of the plasma mem- are coated with proteins almost immediately after
brane, cell adhesion receptors associate with cyto- implantation in the body or immersion in culture
plasmic plaque or peripheral membrane proteins. media. Surface chemistry and topography deter-
Cytoplasmic plaque proteins serve to link the mine the identity, quantity, and conformational
adhesion systems to the cytoskeleton, to regulate change of these adsorbed proteins. In particular,
the functions of the adhesion molecules, and to the roughness and pore size of polymeric mem-
transmit signals initiated at the cell surface by the branes seem to play an important role since they
adhesion receptors. Several substrates different have been shown to inuence the viability and
for the shape and physico-chemical properties metabolic rates of cells (De Bartolo et al. 2008).
are used for cell adhesion in articial devices. Modication of surface chemistry including
Natural substrates such as ECM proteins grafting of functional groups, peptides, and pro-
(collagen, bronectin, vitronectin, laminin) or teins represents a strategy to control cell responses
polymers (chitosan, polylactic acid, polylysine, in in vitro and in vivo systems. The most common
etc.) and synthetic substrates such as polystyrene, peptide immobilized onto surfaces is RGD amino
polycarbonate in the shape of scaffolds, gels, acid (arginine-glycine-aspartic acid) sequence
sponge, and membranes can be used in devices. that stimulates cell adhesion and growth since
Semipermeable membranes made from this peptide represents the minimal adhesion
polyethersulfone, polysulfone, polycarbonate, domains of the most ECM proteins (De Bartolo
polytetrauoroethylene, modied polyether- et al. 2005). The immobilization of specic moi-
etherketone, chitosan, polycaprolactone, eties that interact with specic receptors of cell
polylactic acid, polyglycolic acid, and copoly- membrane is a challenge to enhance the selectiv-
mers in at and hollow ber congurations are ity of the membrane with respect to a cell type in
used for adhesion of different type of anchorage- order to employ it in tissue engineering for the
dependent cells (hepatocytes, endothelial cells, reconstruction of a specic tissue. For example
neuronal cells, keratinocytes, progenitor cells, the immobilization of galactose motifs on the
osteoblasts). In the case of bioarticial organs, surface could enhance the specic interaction
336 Cell Culture
with hepatocytes owing to the specic binding S. Ringer demonstrated that a beating animal heart
between the galactose moiety and the asialogly- can be maintained in salt solutions. First attempts
coprotein receptor present on hepatocyte of cell cultures are dated 1885, when W. Roux
cytoplasmatic membrane (De Bartolo et al. 2006). observed for a week embryonic chick cells using
only a saline solution as medium. Therefore, it is
in 1907 that the father of the cell culture technique
References R. Granville Harrison published experimental
results on the growth of frog nerve bers in a
Curcio E, De Bartolo L, Barbieri G, Rende M, Giorno L,
lymph clot held by the hanging drop method.
Morelli S, Drioli E (2005) Diffusive and convective
transport through hollow ber membranes for liver cell Beneting from the strict aseptic culture condi-
culture. J Biotechnol 117:309321 tions introduced by A. Carrel, the nobel prize for
De Bartolo L, Morelli S, Lopez L, Giorno L, Campana C, medicine and physiology in 1992, this experimen-
Salerno S, Rende M, Favia P, Detomaso L, Gristina R,
tation carried out for several weeks, leaded the
dAgostino R, Drioli E (2005) Biotransformation and
liver specic functions of human hepatocytes in culture way to the development of an enormous variety
on RGD-immobilised plasma-processed membranes. of cell culture techniques and to the progresses of
Biomaterials 26(21):44324441 the sterile conditions and the employ of culture
De Bartolo L, Morelli S, Rende M, Salerno S, Giorno L,
media more and more completed and balanced in
Lopez LC, Favia P, dAgostino R, Drioli E (2006)
Galactose derivative immobilized glow discharge composition (Karp 2010). Cell culture setup. The
processed PES membranes maintain the metabolic specic requirements for a cell culture depend
activity of human and pig liver cells. J Nanosci mainly on the nal application the work is fated;
Nanotechnol 6:23442353
therefore, there are some specic instruments and
De Bartolo L, Rende M, Morelli S, Giusi G, Salerno S,
Piscioneri A, Gordano A, Di Vito A, Canonaco M, parameters that must be used and observed. To the
Drioli E (2008) Inuence of membrane surface proper- basic equipment made of cell culture hood, incu-
ties on the growth of neuronal cells isolated form hip- bator, water bath, centrifuge, refrigerator and
pocampus. J Membr Sci 325:139149
freezer, cell counter, microscopes, liquid nitrogen,
cryostorage container, and sterilizer, other addi-
tional supplies are requested, as vessels, reagents,
media, pipettes, and so on. Moreover, either the
Cell Culture working areas or the operators must ensure the
aseptic conditions essential for avoiding biologi-
Antonietta Messina and Loredana De Bartolo cal and chemical contaminations. Contaminants
Institute on Membrane Technology, National like bacteria, yeasts, molds, viruses, and myco-
Research Council of Italy, ITM-CNR, Rende, plasma are not easy to eradicate and may waste the
Italy experimentations (Alberts et al. 2002).
In a living organism, the cells survive due to
the vascular system that brings the nutrients and
Cell culture refers to a homogeneous group of removes waste product derived from the metabo-
cells able to grow and proliferate in controlled lism. In vitro, this function is entrusted to the
articial environments for many generations. culture medium. Classied in balanced salt solu-
Generally derived from animal and plant, these tions (PBS, Hanks BSS, DPBS), basic media
eukaryotic cells are maintained in sterile vessels (MEM, DMEM), completed media (RPMI 1640,
where a culture medium reproduces the nutritional Iscoves DMEM, Leibovitzs L-15), and serum-
and chemical-physical conditions of the native free media (HAM F10, HAM F12), this specic
environment. isotonic solution is constituted by elementary sub-
Some historiographical notes. The cell culture stances and nutrients, as glucose, salts, amino
technique is only relatively a new science. acids, growth factors, carbohydrates, vitamins,
Though the 1950s are considered the cell culture lipids, and proteins, and a supplement of antibi-
years, actually, already in the nineteenth century, otics is also requested for the inhibition of the
Cell Culture 337
bacterial growth (penicillin 10,000 ug/ml; strep- from a tissue fragment. When cells are immedi-
tomycin 10,000 U/ml). A vitally important sup- ately placed in a suitable culture environment after
plement of the medium is also the animal serum mechanical removal from a tissue, the so-called
(FBS, NCS, HF). The serum, used in concentra- primary cultures are dened. To degrade the extra-
tion of 520 %, is the main source of growth and cellular matrix responsible of the cell layer unity
adhesion factors (PDGF, EGF, IGF), hormones, and to obtain a cell suspension, the tissue is enzy-
lipids, and energetic sources. Moreover, it regu- matically treated (collagenase, trypsin, protease), C
lates cell membrane permeability and protects the and the highly heterogeneous cell population is
cells from mechanical and chemical stresses, and cultured in specic media able to separate single
its buffer power balances the pH value at 77.7. cell types interacting with the cytoplasmatic mol-
All the cell types need proper culture medium and ecules expressed onto the plasmatic membranes
supplements that must be refreshed regularly in (antigens). Despite having a nite life span due to
aseptic condition to guarantee their survival the phenomenon of senescence, these cells prolif-
(Alberts et al. 2002; Karp 2010). erate until they occupy all the substrate available,
Culture conditions. In articial systems, the reaching the conuence. At this stage, in order to
cell is traditionally cultured in plastic or glass avoid the degeneration and their death, the cells
vessels purposely made with different shapes are enzymatically removed (e.g., trypsin, papain,
and properties (asks, Petri dishes, bottles). collagenase in combination with EDTA) and
They are put in incubators where a temperature transferred into a fresh growth medium in a new
of 37 C and a controlled atmosphere (95 % vessel (Hayick 1998; Karp 2010). This proce-
air 5 % CO2) are set in order to mimic the native dure is called subculturing (passaging or splitting)
environment. In the eld of tissue engineering and and transforms primary cultures in cell lines or
bioarticial organs, isolated cells are cultured on subclones, where the cells with the highest growth
several types of biomaterials including polymeric capacity predominate resulting in a degree of
and composite membranes, extracellular matrix genotypic and phenotypic uniformity in the pop-
components, and batch and dynamic systems. ulation. Moreover, when a subpopulation is fur-
Depending on the cell types, there are two basic ther selected, this cell line becomes cell strains
systems for growing cells in culture: the adherent able to acquire additional genetic changes. The
culture system and the suspension system. The survival ability of normal cell cultures lasts only
adherent culture system has been specically for a limited number of passages; however, some
developed for the growth of anchorage-dependent cells can become immortal due to a spontaneous
cells that derive from solid tissues and require to or a viral-/chemical-induced modication.
adhere to a substrate in order to survive and pro- Acquiring the ability to proliferate indenitely
liferate. On the basis of their morphology (shape like the tumor cells, the characteristics of these
and appearance) or on their functional character- transformed continuous cell lines are (i) the fast
istics, these cells are divided into two different growth, (ii) the ability to grow up to higher cell
typologies: the epithelial-like cells that appear density, (iii) reduced serum and anchorage depen-
attened and polygonal in shape and the dence that allow the culture in suspension more
broblast-like cells, elongated and bipolar. The than in adherent vessels, (iv) different in pheno-
suspension culture, on the contrary, refers to the types from native tissue, and (v) stop expressing
culture of anchorage-independent cells, as the tissue-specic genes (Hayick 1998; Alberts et al.
blood or lymphatic cells that survive oating in 2002; Karp 2010). In order to preserve cell char-
suspension and assuming a spherical shape acteristic from further transformation or contami-
dened lymphoblast-like (Alberts et al. 2002; nations and to ensure the reproducibility of
Karp 2010). experimentation in time, a surplus of cells,
Isolation of cells. In order to set up a culture, derived from subculturing, can be treated with a
the early stage is to acquire an already established cryoprotective agent (DMSO or glycerol) and
cell line or strain or to isolate the desired cells stored at a temperature below 130 C in liquid
338 Cell Encapsulation
nitrogen directly in the laboratories. Working When sieving is ensured by an ultrathin mem-
stocks of cell lines are reposted in many cryopres- brane formed around a liquid core, the term of
ervation banks including the American Type Cul- capsule is appropriate. When sieving is ensured
ture Collection (ATCC) and the European by the whole core of the sphere, the term of bead
Collection of Cell Cultures (ECACC) (Alberts is preferred and the porosity is homogenous
et al. 2002). across the structure.
The cell culture technology opens up new per- In case of cell encapsulation for 3D tissue engi-
spectives in many elds, as the manufacturing of neering, the goal is to host cells of a specic organ
vaccines and other products of biotechnology, and type into a porous scaffold in order to maintain their
above all they became the turning point for the original functions and recreate a tridimensional
development of the tissue engineering, science organization mimicking the tissue to be replaced.
entrusted of the study and production of articial Although many applications (such as pancreas,
tissues and the regeneration of tissue and organs, liver) have been developed in the last 30 years,
either in 2D or 3D systems. Nowadays, the cell there is still no clear consensus on the best choice
cultures are the model system and the perfect for biomaterials and cells, whatever the organ type.
instrument for many applications: (i) production The materials hosting the cells should offer
of pharmacological substances (antibodies, vac- biocompatibility, adapted surface properties, and
cines, growth factors, etc.), (ii) toxicity testing of adequate porosity to allow the transfer of nutrients,
drug effects, (iii) study of basic cell physiology, oxygen, metabolites, and low molecular weight
(iv) cancer research, (v) genetic engineering, proteins such as albumin, but also immune protec-
(vi) tissue engineering and regenerative medicine, tion and mechanical stability. They can be native or
and (vii) gene therapy. benet from chemical and biological modications
to better mimic the cells niche and create micro-
environments to control their response. Alginate is
References up to now the materials of reference. Alginate
beads or capsules hosting cells are easy to produce
Alberts B, Johnson A, Lewis J, et al (2002) Table 8-3: some
by extrusion techniques, using coaxial air ow,
landmarks in the development of tissue and cell culture.
In: Molecular biology of the cell, 4th edn. Garland electrostatic force, or vibrating noozle to adjust
Science, New York the bead diameter. Encapsulation can be performed
Hayick L (1998) A brief history of the mortality and at room or body temperature, at physiological pH,
immortality of cultured cells. Keio J Med 47(3):174182
and in isotonic solutions. The alginate bead or
Karp GC (2010) Cell and molecular biology: concepts and
experiments, 6th edn. Wiley, New York capsule might be coated in some approaches with
a poly-L-lysine layer (PLL) to improve mechanical
stability and reduce external porosity Orive
et al. (2006). Other materials such as collagen,
Cell Encapsulation chitosan, gelatin, and agarose can also be
employed, alone or in combination with alginate.
Cecile Legallais The spherical shape of beads or capsules pro-
UMR CNRS 7338 Biomechanics and vides them with interesting characteristics that can
Bioengineering, Universit de Technologie de be considered advantageous to other porous scaf-
Compiegne (UTC), Compiegne, France folds with plate or cylindrical (hollow ber, for
instance) geometries. This spherical shape is con-
sidered as the most efcient one for the promotion
Encapsulation can be generally dened as a proof mass transfer. Indeed, the sphere brings the best
cess that allows the entrapment of substances in a surface-to-volume ratio, leading to proper solutes
spherical-shaped porous material Chang (1964). bidirectional diffusion, provided that the sieving
Cell Recycle Membrane Fermentor 339
a b
cells plasma/blood
nutrients
IgG
oxygen C
metabolites
barrier cells
Cell Encapsulation, Fig. 1 Principle of biohybrid constructs. (a) Description of the mass transfer expected between cell
compartment and host. (b) Application to microencapsulation
properties of the material are tuned. The size of the

beads or capsules related can be adjusted by the Cell Recycle Membrane Fermentor
manufacturing process. Usually, their diameter
does not exceed 5001000 mm to avoid cell necro- Lidietta Giorno
sis in the center. In addition, beads or capsules of Institute on Membrane Technology, National
this size can be easily frozen. Microencapsulation Research Council, ITM-CNR, Rende, Italy
decreases cell damage during cryopreservation and
allows thus an easy storage before use, leading to
readily available encapsulated cells when needed. Synonyms
The choice of cells depends on the targeted
application as organ supply. For clinical applica- Cell-Recycle Membrane (Bio)Reactor (CRMR);
tion of tissue engineering, only primary cells Continuous Membrane Fermentor (CMF)
(from animal or human origin) and cell lines
have been employed. To reconstitute the functions A cell-recycle membrane fermentor is a hybrid
of a failing organ or tissue, multiple setup of traditional fermentor combined with a
pre-differentiated stem cell populations could be membrane ltration process (Fig. 1). In literature
used and/or combined in biomaterials to form they are also referred to as Cell-Recycle Mem-
biohybrid constructs (Fig. 1). brane (Bio)Reactor (CRMR); continuous mem-
brane fermentor (CMF); etc. All of them have the
same basic concept of catalytic membrane reac-
References
tors, which combine the chemical conversion
Chang TM (1964) Semipermeable microcapsules. Science (in the specic case a fermentation) with the sep-
146(3643):524525 aration of the product by membrane operations.
Orive G et al (2006) Biocompatibility of alginate-poly- The system consists of a fermentor connected to
L-lysine microcapsules for cell therapy. Biomaterials
a membrane separation unit. The system as a whole
27(20):36913700
can be seen as a catalytic membrane reactor in
which the biocatalyst is compartimentalized in the
bulk reaction by the membrane module. Cells,
Cell Membrane macromolecules, and solids are retained by the
membrane and recirculated to the fermentor. Prod-
Cellular Membranes uct and other small molecules permeate through
340 Cell Recycle Membrane Fermentor
Cell recycle
Filter UF Membrane
Feed Permeate
Fermentor
Cell Recycle Membrane Fermentor, Fig. 1 Schematic of a cell-recicle membrane fermentor (or continuous membrane
fermentor) (Drioli and Giorno, 1999)
the membrane and are removed from the fermentor. replaced with fresh medium at regular intervals of
The volume of fermentor is kept constant by time to maintain good process performance.
adding fresh feed at the same rate as the permeate The process variables in a continuous mem-
ow rate. The product is diluted in the fermentor brane fermentor are
solution; concentration is reduced and product
inhibition is avoided or at least limited. Since the Flow rate of fermentation broth Q (l/h)
substrate (glucose, sucrose) and low molecular Dilution rate, D Qp=V h1 ; where Qp =
weight nutrients permeate through the membrane permeate ow rate; V = fermentation volume
together with the product, it is convenient to start Bleed rate, B (h1)
the ultraltration process when the concentration Starting time of ultraltration after onseat of
of nutrients is decreased and that of product is fermentation
increased. It must be taken into account that the
concentration of substrate should not be lower than These parameters have to be optimized case by
the limiting value for the survival of cells. case. In general, production rate is related to the
Since in a continuous fermentation mode, the dilution rate, in particular the production rate
microorganism is continuously fed to the reactor, increases with increasing dilution rate. The pro-
its density is increased considerably. This fact ductivity of continuous fermentation is higher
allows the reactor to operate at high biocatalyst compared to batch fermentation. However, nal
concentration. On the other hand, an optimal cell product concentration in the permeate is lower
density must be maintained in the fermentor by than the case of batch fermentation.
using an appropriate bleed rate. In fact, high dense The economic advantage of high productivity
and viscous solutions are difcult to be treated by process decreases rapidly when product concen-
membrane processes. Furthermore, the fouling tration decreases. Lactic acid production by con-
created by the macromolecules decreases mem- tinuous membrane fermentor is interesting on an
brane separation efciency. This means that part economic basis when production is on a relatively
of the fermentation broth has to be removed and small scale and the lactic acid concentration is as
Cell Recycle Membrane Fermentor 341
high as possible. The main reason for this is the The specic growth rate (SGR) can be
high cost for water removal. expressed as a linear function of the specic sub-
It is expected that the performance of the con- strate utilization rate, KS, to give
tinuous membrane fermentor can be improved by SGR Y:KS D (4)
integrating it in a continuous downstream process
which selectively separates the lactic acid and where
recycles the substrate and nutrients, permeated C
through the UF membrane, back to the fermentor KS = the specic substrate utilization rate, mass/
solution. For the downstream processing, electro- mass time
dialysis, membrane-based solvent extraction, and D = the organism decay coefcient, 1/time.
nanoltration ae among the most suitable
methods. The kinetics of substrate removal in an anaero-
When kinetic models for the growth and fer- bic reactor actually determines the solid retention
mentation of a specic kind of cell or microbe on time (SRT) required for a given fermentation ef-
the corresponding substrate medium are available, ciency. In turn, the presence of the UF unit allows a
mathematical modeling of membrane fermentors ne control of SRT, making process control easier.
can be accomplished according to mass balances When high biomass concentrations are achieved,
and the kinetics. longer SRTs can be maintained at lower reactor
A steady-state mass balance on a continuous volumes.
cell recycle fermentor over the vessel alone is as The volumetric loading to a reactor, VL, is
follows: dened as
a0 F X l a 0 F X V m
X 0 (1) VL Q Sf=V Sf=t (5)
on cells, and where
X0 Y D Sf S=m
(2) V = reaction volume, L3
Q = ow rate, L3 T1
on limiting growth substrate to give a cell concen-
tration value in the vessel equal to High biomass concentrations permit operation
at higher volumetric loading rates. Reactor design
X0 Y Sf S=l a0 C a0 (3) can be carried out by estimating the value of
volumetric loading necessary to achieve a given
where efuent quality. This design criterion is com-
monly used for biological systems where the eval-
a0 = recycle ratio uation of biomass concentration within the reactor
F = feed ow rate, L3T1 is difcult. This is not the case with cell recycle
Sf = feed substrate concentration, ML3 fermentors, so the VL or the SRT approach can
S = substrate concentration, ML3 also be used as design criterion.
X0 = cell concentration, ML3 In continuous membrane fermentors productivity
X10 = cell concentration entering the membrane is a function of viable cell concentration within the
module, ML3 fermentor. Cell recycle allows to operate at higher
0
X2 = cell concentration in the recycle stream, microbial cell concentrations than in conventional
ML3 batch fermentors, reducing the volume required for a
m* = specic cell growth rate, T1 given productivity and hence the capital costs. On
Y = organism yield coefcient, i.e., mass of cells the other hand, the viscosity of the cell slurry
formed/mass of nutrients increases with cell concentration, and concentration
C = concentration factor polarization phenomena are usually signicant.
342 Cell Separation
Systems such as rotating membrane modules, Traditionally cells are separated through centrifuga-
backushing operations, etc. can be used to over- tion technique for the different density and size. This
come fouling and expected concentration polari- technique only requires the resuspension of the cells
zation problems. Provided cell resistance to shear in an appropriate buffer and the knowledge of their
stress, concentration polarization is usually con- approximate composition and density or size. Cells
trolled by operating the reaction system at high of different masses and densities are pelleted
recirculation rates. accordingly, with the densest cells pelleting rst
When concentration polarization occurs, and at comparatively low centrifugation speeds,
permeate uxes become invariant with the trans- while the smallest, lightest cells require much faster
membrane pressure, and increases in the perme- speeds. The pellets can then be collected and
ation rate can be achieved only by appropriate resuspended in the desired buffer. Another method
uid dynamics conditions. The design of the UF consists to use antibody which are able to recognize
membrane unit in the polarized regime must relate cell membrane proteins. All cells have an array of
the ux to the optimal level of reactor volatile proteins on their surface membrane, some of which
suspended solids (VSS) or total suspended solids are found only on particular cells. The antibodies are
(TSS) according to Michaels gel theory, to give often tagged with a uorescent compound
(uorophore) so that the whole cell suspension can

J KS ln Cg =Cb (6) then undergo ow cytometry, which will identify
and then isolate each cell individually. The anti-
where bodies can be also combined with a tiny magnet.
The cells are then applied to a magnetic column that
KS = the overall mass transfer coefcient, LT1 is capable of retaining these tagged cells when a
Cg = apparent solid concentration in the gel, magnetic eld is generated within it. Cells can be
ML3 separated through a biochemical method, which
Gb = concentration in the bulk, in the specic involves perturbing a biochemical process that is
case, concentration of VSS or TSS, ML3 required by the cell for its growth and/or survival,
and it should only be performed if a drastic manip-
High recirculation and dilution rates help in ulation will not detrimentally affect other steps in
maintaining low levels of inhibitory products. the experiment. Biochemical separations include
blocking DNA synthesis, e.g., with hydroxyurea,
References and serum deprivation, i.e., growing cells in
serum-free media for a specic amount of time.
Drioli E, Giorno L (1999) Biocatalytic membrane reactors: Membrane processes are used for separation of
application in biotechnology and the pharmaceutical blood cells from other blood components through
ltration. Cells are isolated by using membranes
with suitable molecular weight cutoff that permit
the passage of all components excluding only cells.
Cell Separation
Membrane ltration has also been used for the
industrial separation of blood cells. Blood for trans-
Loredana De Bartolo
plantation is typically passed through membrane
lters to eliminate leukocytes, which can help pre-
Research Council, ITM-CNR, Rende, Italy
vent infection by viruses such as human immuno-
deciency virus and hepatitis C virus. Compared
Cell separation is the process aimed at separating with other cell separation methods, membrane l-
cells. Cells can be separated on the basis of different tration is simple and inexpensive, and it is easy to
size, shape, physicochemical characteristics, spe- maintain sterility during the process. Various porous
cic molecules, or receptors present over their plas- polymeric membranes have been used for separa-
matic membrane (Orfao and Ruiz-Arguelles 1996). tion by ltration of different types of marrow
Cellophane 343
stromal cells (KUSA-A1 osteoblasts and H-1/A Cellophane was derived from the French words
preadipocytes), due to the different cell size cellulose and diaphane (transparent). Cellophane
(Higuchi et al. 2005). Separation of hepatocytes was invented by Swiss chemist Jacques
and broblasts has been realized through surface- E. Brandenberger while employed by
modied polyurethane membranes combining the Blanchisserie et Teinturerie de Thaon (Wikipedia
ltration process with the use of negatively charged free encyclopedia). In 1900, he started his rst try.
membranes (Higuchi and Tsukamoto 2004). After almost 10 years, 1910, Brandenberger C
improves his lm by adding glycerin to soften the
material. By 1912 he had constructed a machine to
References manufacture the lm and cellophane was patented
that year (Carlisle 2004).
Higuchi A, Tsukamoto Y (2004) Cell separation of hepa-
tocytes and broblasts through surface-modied
Cellophane manufacture process includes
polyurethane membranes. J Biomed Mater Res A many steps. The rst step is to dissolve cellulose
71A(3):470479 in alkali and carbon disulde to have a viscose
Higuchi A, Shindo Y, Gomei Y, Mori T, Uyama T, solution. The second step is the extrusion of the
Umezawa A (2005) Cell separation between mesen-
chymal progenitor cells through porous polymeric
viscose solution through a slit into a bath of dilute
membranes. J Biomed Mater Res B Appl Biomater sulfuric acid and sodium sulfate to reconvert the
74(1):511519 viscose into cellulose lm. The third step is the
Orfao A, Ruiz-Arguelles A (1996) General concepts about passing of the lm through several more baths,
cell sorting techniques. Clin Biochem 29(1):59
one to remove sulfur, one to bleach the lm, and
one to add glycerin to prevent the lm from
becoming brittle.
The properties of cellulose membrane are inti-
Cell-Based Liver Support Systems mately related to their structure and morphology
which are mainly controlled by conditions, nature,
Bioarticial Liver Support System (BLSS) and mechanism of coagulation. Therefore, to
design desired morphology and properties, the
cellulose membrane coagulated in different coag-
ulation baths such as water and sulfuric acid aque-
Cell-Recycle Membrane ous solution (Fink et al. 2001; Hongo et al. 1999;
(Bio)Reactor (CRMR) Inamoto 1996; Manabe 1987; Abe and Mochizuki
2002; Bang 1999; Rosenau 2001). Owing to its
Cell Recycle Membrane Fermentor high production costs and environmental pollu-
tion problems, it has great signicance to develop
simple producing process for cellulose membrane
using new and powerful organic solvent
Cellophane N-methylmorpholine N-oxide (NMMO) to dis-
solve cellulose straightaway without any complex
M. S. Mohy Eldin formation (Fink et al. 2001). The obtained cellu-
Polymer Materials Research Department, lose membranes exhibited excellent mechanical
Advanced Technologies and New Materials properties and permeation characteristics (Fink
Research Institute (ATNMRI), City for Scientic et al. 2001; Abe and Mochizuki 2002; Bang
Research and Technology Applications (CSRTA), 1999; Rosenau 2001).
New Borg El-Arab City, Alexandria, Egypt Applications such as packaging for a variety of
food items, base for such self-adhesive tapes, bat-
tery, and dialysis tubing (Visking tubing) and as a
Cellophane is made from regenerated cellulose in release agent in the manufacture of ber glass and
the form of thin and transparent lms. The name of rubber products have been recognized.
344 Cellophane Membranes
References types of solvents for cellulosic materials especially

ionic liquid solution. According to these characters,
Abe Y, Mochizuki A (2002) Hemodialysis membrane pre- different applications of cellophane membranes
pared from cellulose/N-methylmorpholine-N-oxide
have been recognized such as food packaging
solution. I: effect of membrane preparation conditions
on its permeation characteristics. J Appl Polym Sci which represents the oldest one (Stocking and
84:23022307 Mueller 1955). Hemodialysis cellophane mem-
Bang YH, Lee S, Park JB, Cho HH (1999) Effect of coag- branes are one of the rst applications in dialysis
ulation conditions on ne structure of regenerated cellu-
biological uid (blood) to eliminate unwanted sub-
losic lms made from cellulose/N-methylmorpholine-
N-oxide/H2O systems. J Appl Polym Sci 73:26812690 stances such as urea and creatinine. The cellophane
Carlisle R (2004) Scientic American inventions and dis- membranes were used in planer form to dialyzed
coveries. Wiley, Hoboken, p 338. ISBN 0-471-24410-4 blood against electrolyte solution in articial kidney
Fink HP, Weigel P, Purz HJ, Ganster J (2001) Structure
(Mollison and Graydon 1977). Separation between
formation of regenerated cellulose materials from
NMMO-solutions. J Prog Polym Sci 26:14731524 two electrolytes as ionic barriers (battery separator)
Hongo T, Inamoto M, Iweata M, Matsui T, Okajima K is one of the most known industrial applications of
(1999) Morphological and structural formation of the cellophane membranes. As a separator, it has
regenerated cellulose membranes recovered from its
achieved the following objectives: (1) immobiliza-
cuprammonium solution using aqueous sulfuric acid.
J Appl Polym Sci 72:16691678 tion of the electrolyte by absorption in the cellulose,
http://en.wikipedia.org/wiki/Cellophane (2) elimination of the gases evolved in the operation
Inamoto M, Miyamoto I, Hongo T, Iwata M, Okajima K of the cell, (3) prevention of mutual penetration of
(1996) Morphological formation of the regenerated cel-
the active materials, and (4) homogeneous and
lulose membranes recovered from its cuprammonium
solution using various coagulants. Polym J 28:507512 reversible deposition of the negative active material
Manabe SI, Kamata Y, Iijima H, Kamide K (1987) Some (Zn) from the electrolyte (Claudia and Evans
morphological characteristics of porous polymeric 1964). As ionic conductors, ion exchange cello-
membranes prepared by micro-phase separation
phane membranes have been developed as fuel cell
method. Polym J 19:391404
Rosenau T, Potthast A, Sixta H, Kosma P (2001) The ion exchange membranes (Abu-Saied et al. 2012,
chemistry of side reactions and byproduct formation 2013). The cellophane membranes act to conduct
in the system NMMO/cellulose (Lyocell process). Prog hydrogen ions and separate fuel from crossover to
Polym Sci 26:17631837
the anode compartment in the polyelectrolyte fuel
cells. Due to cellophane membrane hydrophilicity,
low nonspecic adsorption (Zou et al. 2001;
Cellophane Membranes Crawford et al. 1999), and good chemical stability,
they were adapted as base for afnity separation
M. S. Mohy Eldin chromatography membranes for proteins. Mohy
Polymer Materials Research Department, Eldin et al. 2009 developed immobilized metal ion
Advanced Technologies and New Materials afnity membranes from poly(glycidyl methacry-
Research Institute (ATNMRI), City for Scientic late) grafted cellophane membranes for separation
Research and Technology Applications (CSRTA), of b-galactosidase enzyme. Later on Mohy Eldin
New Borg El-Arab City, Alexandria, Egypt et al. 2011 developed immobilized metal ion afnity
membranes from poly(methacrylic acid) grafted
cellophane membranes for separation of His-Tag
Cellophane membranes are regenerated transpar- chitinase.
ent cellulose membranes of high hydrophilicity,
good mechanical properties, and biodegradability,
biocompatibility, and gas barrier characters. The
References
crystallinity and porosity of the membranes have
been controlled through the regeneration conditions Abu-Saied MA, Elzatahry AA, El-Khatib KM,
over the last decades. Recently, these characters are Hassan EA, El-Sabbah MM, Drioli E, Mohy
under control after investigating and using new Eldin MS (2012) Preparation and characterization of
Cellular Membranes 345
novel grafted cellophane-phosphoric acid doped Living cells are divided into several compart-
membranes for proton exchange membrane fuel ments and sub-compartments by a membrane.
cell (PEMFC) applications. J Appl Polym Sci
123:37103724 Each compartment has a specic function and
Abu-Saied MA, Fontananova E, Drioli E, Mohy Eldin MS structure and the interactions between compart-
(2013) Sulphonated poly (glycidyl methacrylate) ments are highly regulated by specic signals.
grafted cellophane membranes: novel application in The corresponding delimited spaces or compart-
polyelectrolyte membrane fuel cell (PEMFC).
J Polym Res 20:187192 ments are called organelles; in eukaryotic cells the C
Claudia C, Evans (1964) Characterization of cellophane as main are endoplasmic reticulum, Golgi apparatus,
an ionic barrier, NOLTR 63-225 nucleus, mitochondria, lysosomes, endosomes,
Crawford J, Ramakrishnan S, Periera P, Gardner S, and peroxisomes. Each membrane-enclosed
Coleman M, Beitle R (1999) Immobilized metal afn-
ity membrane separation: characteristics of two mate- organelle contains a specic set of proteins free
rials of differing preparation chemistries. Sep Sci or on the membrane surface that regulate many
Technol 34:27932802 vital biochemical processes.
Mohy Eldin MS, Soliman EA, Hassan, Abu-Saied MA Cellular membrane is a complex system with
(2009) Immobilized metal ions cellophane PGMA
grafted E a membranes for afnity separation of multifunctional properties, acting as a separating/
b-galactosidase enzyme: preparation and characteriza- contacting barrier within or around cell and cell
tion. J Appl Polym Sci 111:26472656 components. This system controls the selective
Mohy Eldin MS, Abdel Rahman S, El-Fawal GF transport of components and regulates the passage
(2011) Preparation and characterization of grafted cel-
lophane membranes for afnity separation of his-tag through the different compartments. In addition
chitinase. Adv Polym Technol 30:191202 biological membranes have mechanical and elas-
Mollison AN, Graydon WF (1977) Cellulosic tic properties given by the interaction with intra-
ion-exchange membranes for hemodialysis. J Biomed cellular skeleton and, if present, the extracellular
Mater Res 11:563575
Stocking GW, Mueller WF (1955) The cellophane case and cell wall (such as fungi, etc.).
the new competition. Am Econ Rev 45:2963 In general, the biological membrane is formed
Zou H, Luo Q, Zhou D (2001) Afnity membrane chro- by a continuous, water-impermeable, double layer
matography for the analysis and purication of proof phospholipids containing a myriad of macro-
teins. J Biochem Biophys Methods 49:199240
molecules and their complexes through which
transport is promoted. According to the uid
mosaic model, the cell membrane is a
two-dimensional uid of proteins and lipids
Cellular Membranes (Singer and Nicolson 1972).
The biological membrane is a dynamic and
Fabio Bazzarelli, Rosalinda Mazzei and uid structure, which implies that the molecules
Lidietta Giorno have a substantial degree of mobility in the
Institute on Membrane Technology, National two-dimensional membrane plane. This allows
Research Council, ITM-CNR, Rende, Italy rapid lateral diffusion of lipids and proteins
within the planar membrane surface. In contrast
to lateral movement, the transbilayer or ip-
Synonyms op diffusion of lipids from one leaet in
the bilayer to the other occurs very slowly.
Biological membrane; Biomembrane; Cell mem- Transbilayer movement requires that a polar or
brane; Plasma membrane charged head group leave its aqueous environ-
ment and move into the hydrophobic interior of
Cellular membrane (or biomembrane) is the bio- the bilayer.
logical membrane that surrounds and protects the The uidity of a membrane depends on its
cell and governs the transport of matter, energy, composition, the temperature, and how the differ-
and information between the external and internal ent components are packed and organized in the
(and vice versa) of the cell. membrane. In general, membrane uidity is
346 Cellular Membranes
decreased by lipids with long or saturated fatty component of cellular homeostasis. The selective
acid chains, whereas lipids with short or unsatu- passage of solutes across the cell membrane, a
rated fatty acid chains tend to increase the uidity physiological property known as membrane per-
of a membrane (Bittar and Bittar 1997). There are meability, is mediated by the presence of mem-
three major types of lipids found in the biological brane transport proteins that span the
membrane: phospholipids, cholesterol, and glyco- phospholipid bilayer. Transport mechanisms may
lipids. Phospholipids form a bilayer in which the be distinguished thermodynamically according to
nonpolar regions of the lipid molecules in each their ability to mediate active or passive transport.
layer face the core of the bilayer and their polar The active transport is dened as movement of a
head groups face outward, interacting with the solute across membranes from areas of low con-
aqueous phase on either side. centration to areas of high concentration. This
Cholesterol is another important component process is done by carrier proteins and uses energy
found in the eukaryotic cell membranes. It has a to transfer solutes against their concentration gra-
rigid structure that stabilizes the membrane and dient. The most common source is the hydrolysis
reduces the natural mobility of the complex lipids of ATP (adenosine triphosphate). Others include
in the plane of the membrane. Increasing amounts light energy and the energy stored in ion gradient.
of cholesterol make it more difcult for lipids and In addition, exocytosis and endocytosis are other
proteins to move in the membrane. transport mechanisms used for the movement of
Proteins are the second major component of large molecules across the membrane. During
the cell membrane, present in about equal propor- exocytosis the macromolecules are contained
tion by weight with the lipids. Proteins are embed- within vesicles which fuse with the cell membrane
ded in this bilayer sheet, held by hydrophobic to release their contents to the outside. In endocy-
interactions between the membrane lipids and tosis the reverse takes place: the substances to be
hydrophobic domains in the proteins. taken in by the cell are gradually enclosed by a
Membrane proteins can be classied according region of cell membrane until a vesicle containing
to their relationship with the lipid bilayer into two the substances is formed. Regarding the passive
types: peripheral proteins (or extrinsic proteins) transport, it is dened as movement of a solute
that are outside of the lipid bilayer without from a region of high concentration on one side of
completely crossing the membrane and integral the cell membrane to a region of lower concentra-
proteins (or intrinsic proteins) that cross the lipid tion on the opposite side (no energy is required to
bilayer only once or many times. move a substance). The speed of solute diffusion
The proteins embedded in the cell membrane depends on the difference of concentration, the
play a variety of roles. Many peripheral mem- size of the molecules, and the possible interactions
brane proteins are enzymes, and many of the diffusible substance with water. The diffu-
membrane-spanning integral proteins are carriers sion across the membrane can be described by
or channels (aquaporins) for the movement of Ficks law:
water-soluble molecules and ions into and out of
the cell. Another important role of membrane pro- Dc
teins is structural; for example, certain membrane J DA
Dx
proteins in the erythrocyte help maintain the
biconcave shape of the cell. Finally, some mem- where J is the rate of diffusion, D is the diffusion
brane proteins serve as highly specic receptors coefcient, A is the membrane area, Dc is the
on the outside of the cell membrane to extracellu- concentration difference across the membrane,
lar molecules, such as hormones. The selective and Dx is the thickness of membrane.
and regulated passage of molecules and ions The principal driving force for the passive dif-
across the cell membrane is an essential fusion of an uncharged solute across the plasma
Cellulase in Membrane Bioreactors 347
membrane is the difference of concentration

between the inside and the outside of the cell. In Cellulase in Membrane Bioreactors
the case of an electrically charged solute, such as
an ion, diffusion is also driven by the membrane Katalin Bela-Bako
potential, which is the electrical gradient across Research Institute of Bioengineering, Membrane
the membrane (Considine 2013). The cell Technology and Energetics, University of
employs three main mechanisms to carry out the Pannonia, Veszprem, Hungary C
passive transport:
Simple diffusion is the passage of lipid-soluble Membrane reactors have been developed mainly
solutes across the plasma membrane. for catalytic reactions, using various membrane
Osmosis: in this process, the spontaneous modules (e.g., at sheet, hollow bers). If
movement of water across a membrane driven biocatalysts (enzymes or cells) are applied, the
by a gradient of water concentration and the equipments can be operated as membrane biore-
movement of water from a solution of high actors (Drioli and Giorno 1999). One of the main
water concentration (low concentration of sol- advantages of the membrane bioreactors is that
ute) toward a solution with a lower concentra- they make recovery and reuse of biocatalysts
tion of water (high solute concentration). possible.
Osmosis is a passive transport mechanism Usage of membrane bioreactors is especially
that tends to equalize the total solute benecial for polymer degradation processes,
concentrations. where the small size product may have inhibitory
Facilitated diffusion is the movement of water- effect, but can be separated easily by a proper
soluble solutes and ions through a channel or porous membrane. This is the case of enzymatic
across the transmembrane transporter. hydrolysis of polysaccharides, like cellulose or
starch. The long polysaccharide chain as well as
In addition to lipid and proteins, in the cell the biocatalyst (enzyme or cell) are rejected by the
membrane, there are carbohydrates, which are membrane, while the product (glucose) passes
attached to the proteins forming glycoproteins, through the membrane into the other phase. In
or to some of the lipid classes, forming the such a system, continuous uptake of substrate
glycosphingolipids. Protein-bound carbohydrate and release of product without loss of biocatalysts
residues are on the extracellular surface of the can be achieved.
cell membrane and take part in cell-cell interac- The huge amount of biomass produced annu-
tions, including those of the immune system ally worldwide is mainly composed of cellulose;
(Stein and Litman 2015). its application for various purposes (industrial raw
material, bioethanol, etc.) is extremely important
nowadays. In these cases cellulose should be
References degraded by hydrolysis, resulting in glucose
(monomer of cellulose). Enzymatic cellulose
Bittar E, Bittar N (1997) Principles of medical biology. Jai hydrolysis is carried out by various cellulases.
Press Ltd, Greenwich
In general, hydrolysis of cellulose is more dif-
Considine RV (2013) Plasma membrane, membrane trans-
port, and resting membrane potential, 4th edn. cult to achieve than that of the other polysaccha-
Lippincott Williams & Wilkins, Philadelphia, pp 2439 rides. The process is inuenced by three main
Singer SJ, Nicolson GL (1972) The uid mosaic model of factors: the nature of the enzymes; the structure
the structure of cell membranes. Science 175:720731
of the substrate; and the enzyme-substrate inter-
Stein WD, Litman T (2015) Channels, carriers, and pumps,
an introduction to membrane transport. Elsevier, actions (Manseld et al. 1999). The action of the
Amsterdam cellulase enzyme system includes three main
348 Cellulose Derivatives
groups: endoglucanases, cellobiohydrolases, and

cellobiases (-glucosidase). These enzymes work Cellulose Derivatives
in a synergistic mode to degrade cellulose, there-
fore it would be more effective to apply the whole Hongyang Ma, Benjamin S. Hsiao and
complex enzyme system. One of the most effec- Benjamin Chu
tive cellulase systems can be obtained from the Department of Chemistry, Stony Brook
fungus Trichoderma reesei. University, Stony Brook, NY, USA
The main difculty in cellulose hydrolysis is
attributed to the dual nature of the substrate. The
accessibility of the crystalline and amorphous Cellulose derivatives used in membrane fabrica-
regions is different; moreover, the solubility of tion are mainly referred to cellulose esters, in
the substrate is limited. which cellulose acetate, cellulose diacetate, cellu-
Finally in the enzyme-substrate interactions, it lose triacetate, and cellulose nitrite are typical
should be considered that hydrolysis of cellulose components. There are other cellulose esters,
is a surface phenomenon. Before the reaction including cellulose acetate propionate, butyrate,
itself, the enzyme has to be bound to the substrate, octanoate, benzoate, methyl carbonate, and
i.e., the process requires the adsorption of the methoxyacetate, which have also been used to
biocatalyst onto the insoluble substrate prior to fabricate membranes, with/without combination
the hydrolysis. of cellulose acetate (Fig. 1). Membranes fabri-
In a recent work, a membrane bioreactor was cated from single or multiple cellulose derivatives
used with a hairy membrane layer (Bla-Bak can be used in all elds of water purication, gas
et al. 2006) that offered a special surface area for separation, and biomedical applications, where
simultaneous adsorption of the substrate and the morphology (e.g., pore size) and structure of
enzyme. In this way the cellulose-cellulase the membrane can be exibly adjusted through
layer formed provided a further advantage: dif- different preparation pathways (Ulbricht 2006).
fusion resistance between the substrate and Another widely used cellulose derivative is cel-
enzyme was reduced. Since solid particles pre- lulose ether (e.g., ethyl cellulose), which has been
sent might cause troubles (fouling), it required demonstrated in fabrication of membranes for gas
special conditions for the construction of the separations, such as hydrogen, oxygen, nitrogen,
membrane reactor. To avoid blockage, tubular carbon dioxide, and methane (Mulder 1996).
type membrane module was used in the experi- Membranes fabricated from cellulose deriva-
ments. The simultaneous, reversible immobiliza- tives for separation applications offer several
tion of the biocatalyst and the substrate onto the advantages: (1) a variety of cellulose derivatives
membrane surface was found to improve the with different physical and chemical properties are
efciency of the process. available; (2) good chemical resistance and
mechanical properties are attainable; (3) cellulose
derivatives are cost-effective; and (4) readily avail-
able surface functionalization pathway can offer
References new opportunities to generate new functional
membranes. For example, the nanobrous mem-
Bla-Bak K, Koutinas A, Nemestthy N, Gubicza L,
Webb C (2006) Continuous enzymatic cellulose hydro- brane fabrication approach (e.g., electrospinning to
lysis in a tubular membrane bioreactor. Enzyme Microb produce nanobers) using cellulose derivatives, in
Technol 38:155161 combination with surface grafting of functional
Drioli E, Giorno L (1999) Biocatalytic membrane reactors. molecules and/or inclusion of nanoscale additives
Taylor and Francis Group, London
Manseld SD, Mooney C, Saddler JN (1999) Substrate and to create a nanocomposite structure, has expanded
enzyme characteristics that limit cellulose hydrolysis. the family of cellulose membranes and opened
Biotechnol Prog 15:804816 many new separation applications (Ma et al. 2012).
Central Nervous System in Relation to Membranes 349
Cellulose Derivatives, R2O OR3

Fig. 1 Cellulose
derivatives for membrane O O
fabrication (Turbak 1970) OR1 n
R1 = R2 = R3 = H Cellulose
R1 =
O
, R2 = R3 = H Cellulose acetate C
O
R1 = R2 (R3) = , R2 (R3) = H Cellulose diacetate
O
R1 = R2 = R3 = Cellulose triacetate
R1 = R2 = R3 = NO2 Cellulose nitrite

O
R1 (R2, R3) = Cellulose acetate propionate
O
R1 (R2, R3) = Cellulose acetate butyrate
O
R1 (R2, R3) = Cellulose acetate benzoate
O
R1 (R2, R3) = Cellulose acetate methyl carbonate
O
O
R1 (R2, R3) = Cellulose acetate methoxy acetate
O
References The spinal cord attends the communication path-

way between the brain and the periphery, and like
Ma HY, Hsiao BS, Chu B (2012) Functional nanobers and the brain, it is made of a gray matter, constituted
applications. Wood Publishing Limited, UK,
by the bodies of the neurons, and a white matter,
pp 331370, Chapter 15
Mulder M (1996) Basic principle of membrane technology. constituted by their myelinated axons, serving the
Springer, India impulse transmission. Moreover, a special blood-
Turbak AF (1970) Membranes from cellulose and cellulose spinal cord barrier made of endothelial cells and
derivatives, Applied Polymer Symposia No. 13.
astrocytes protects the spinal cord from the bio-
Interscience Publishes, Wiley, New York
Ulbricht M (2006) Advanced functional polymer mem- logical and chemical changes coming from the
branes. Polymer 47:22172262 periphery. It was long believed that regenerative
processes in the CNS were completely absent,
then after an injury, the recovering of the axonal
function was impossible (Kandel et al. 2000). In
Central Nervous System in Relation the early 1980s, therefore, it was discovered that
to Membranes CNS retinal neurons were able to regenerate in
bioengineered devices of peripheral nerves,
Loredana De Bartolo and Antonietta Messina highlighting that the nearly absent ability to
Institute on Membrane Technology, National regenerate the central neurons does not lies in
Research Council, ITM-CNR, Rende, Italy the cells themselves but in their surrounding envi-
ronment. After physical trauma, hypoxic events,
and degenerative pathologies, a rapid severing of
The central nervous system (CNS) is the most the axons occurs, and in the spinal cord, cell death
complex organ in the human body. It consists of is recorded in high extent due to the lost electro-
two components, the brain and the spinal cord. lytic homeostasis and the accumulation of free
350 Central Nervous System in Relation to Membranes
radicals and toxic factors that stimulate apoptotic new approaches appeared to be necessary for the
events. While the peripheral nervous system treatment of the CNS injury. Due to the develop-
(PNS) consists of a particular glial cell population ment of biomaterials and new processing methods
able to proliferate and produce specic factors for the scaffold design, tissue engineering and
(BDNF, CNTF, GDNF) involved in the axonal chemotherapeutic strategies are the main
regeneration across the area of injury, the oligo- approaches for the development of regenerative
dendrocytes constituting the main part of the cen- strategies able to guide axonal outgrowth. A wide
tral glia are not able to provide similar support range of natural and synthetic biomaterials is
after damage. In the injury site of the central available to prepare neural scaffolds and releasing
nervous system, they secrete numerous inhibitor devices (Tysseling and Kessler 2011). Different
factors for the axonal regeneration, including classes of biodegradable polymers are available,
NOGO proteins, NI-35, chondroitin sulfate pro- including polycaprolactone, polyurethanes,
teoglycans (NG2), and keratan sulfate proteogly- polylactic derived, polyglicolic acid, chitosan,
cans. The produced signaling factors that play agarose, polysaccharides, collagen and various
neuroprotective and regenerative roles appear, ECM proteins derived. Due to their controlled
then, to be not enough to oppose this inhibition bioerosion, these materials allow the design of
and the inammatory events occurring after the different bioengineered devices, with different
damage. Furthermore, astrocytes, another compo- shapes and conguration purposely made for
nent of the central glia, became proliferative. In bridging the nerve growth or the factors release.
order to protect the severed site from external Any biomaterial used to prepare neural scaffolds
molecules and the spreading of apoptotic events, possess, in addition to biocompatibility, an appro-
they form a tight network called a glial scar, which priate range of permeability, porosity, and biome-
obstruct the axonal regeneration, creating a wall chanical properties serving to avoid the
that axons and neurogenerative factors cannot compression of the regenerating nerves or the
overtake (Horner and Gage 2002). Thus, manipu- sudden collapse of the device in the treated site.
lation of the chemical and biological environment Actually, three different strategies are available
around the injured site represents the key of the for the treatment of the CNS injury: the
treatment of the CNS damages. Specically, in entubulation models, the drug delivery systems,
order to promote the axonal regeneration of the and the cell therapy. All of them are categorized as
central neurons, the neurogenerative and membrane system devices due to their mass trans-
neuroprotective factor levels should increase port properties they allow the diffusion, release,
(Recknor and Mallapragada 2006). Nowadays, it and removal of the different factors involved after
is well known how neural cells respond to their an injury.
microenvironment, enhancing adhesion, prolifer- The entubulation strategy consists in the
ation, and migration along the regenerating site. transplantation of tubular scaffolds purposely
Smooth surfaces, patterned supports, and struc- made to mimic the peripheral nerve tissue. Con-
tures with high mass transport properties are the sidering that the PNS is characterized by a regen-
main characteristics allowing the neural regener- erative ability due to proliferation of glial cells
ation both in vivo and in vitro systems (Morelli and the secretion of neuroprotective and
et al. 2012). Since decades ago, the gold standard neuroregenerative factors, many devices have
of the nervous system injury treatment is the been developed with the aim of entrapping the
transplantation of nerve grafts able to substitute central axons in a protect environment, where
the malfunctioning or lost areas of the system. inhibitor factors are not allowed to pass through
Despite the implantation of nerve grafts allowed and neuroregenerative signals are enhanced.
to reach important results in the recovery of move- These bioarticial tubular nerve grafts are meant
ment and sensibility of the PNS, the central ner- to ensure a correct diffusion of nutrients and oxy-
vous system is hard to be treated, particularly gen and purposely designed with specic thick-
when inner regions of the brain are severed, then ness, mechanical strength, dimension, and
Central Nervous System in Relation to Membranes 351
porosity degree. Tubular nerve grafts made of delivery strategy. When a nerve guide or other
PLLA, PCL and PGA, and collagen and chitosan bioarticial implants are not appropriate due to
already extensively used for peripheral nerve the impossibility to reach the severed central ner-
regeneration are nowadays applied for the treat- vous area, the delivery system of neurogenerative
ment of CNS injury models, alone or in conjunc- and neuroprotective factors represent the alterna-
tion with brain-derived neurotrophic factor tive strategy. A few molecules of natural and
(BDNF) to promote the axonal regeneration synthetic origin showed to promote the neural C
(Recknor and Mallapragada 2006; Morelli regeneration. Steroids, minocycline, interleukin
et al. 2012). Cell-adhesive proteins such as colla- 10, erythropoietin, and neurotrophins, like NGF,
gen, laminin, and bronectin, which bind strongly BDNF, and CNTF, are just some of the
the integrin receptors on the cell surfaces, assist neuroprotective agents implied in the inhibition
the nerve regeneration. They are often used to coat of the inammatory response, microglial activa-
the lumen of the bioarticial channels or to mod- tion, and apoptotic event after a central injury
ify the biomaterial used for the device preparation (Kandel et al. 2000). Induction of the
in order to enhance the cell response and the regeneration-associated genes expression
regenerative process. The patterning of polymeric (RAGs) or the Rho-Rock pathway inhibition is
tubular membranes with short oligopeptides some of the neurogenerative properties of the
derived from ECM proteins is another way to neurotrophins. Moreover, they promote their
increase the neurite outgrowth and the cytoskele- remyelination, the sprouting and growth of new
ton organization in the regenerating neural tissue. neurons with functional recovery. Whether the
Along similar lines, the incorporation of factors possess neuroprotective or
hydrogels into the nerve tubes have been evalu- neurogenerative ability, the delivery strategies
ated with encouraging results for CNS applica- are the same. Growth factors, proteins, and drugs
tions. Collagen, hyaluronic acid, agarose, are encapsulated into the devices made of poly-
alginate, chitosan, methyl cellulose, xyloglucan, mer, hydrogels, or proteins and injected in the
matrigel, brin, polysia-based hydrogels, peptide spinal cord through epidural and intrathecal routes
hydrogels, as well as synthetic hydrogels, such as for the spinal cord regeneration. Hydrogels, lipo-
polyacrylamide and polyethylene glycol somes, brin, collagen-based, and other poly-
hydrogels, revealed to be the best alternative to meric delivery devices obtained from natural and
the classical polymeric nerve guidance grafts. synthetic biomaterials have been implanted sub-
Bioengineered tubular scaffolds have been com- cutaneously near the spinal cords injured site,
bined also with cells providing a promising new allowing the continuous release of the
approach for the CNS injury treatment. Since neuroprotective and neurogenerative agents and
1998, it is well known that remyelination of cen- the partial recovery of the axonal functionality
tral axons can be promoted by using Schwann and (Zhong and Bellamkonda 2008). Cell therapy.
olfactory glial cells entubulated in bioarticial The drug/biological delivery strategies are easy
scaffolds, and moreover, Schwann cells, when with respect to the entubulation methods, but in
seeded in combination with neurotrophic factors truth, they have one disadvantage: the constant
or hormones, facilitated the axon regeneration. drug dose is achievable only with reinjection of
Therefore, the strategy is to entubulate neural new delivery devices. The cell therapy, instead,
stem cells (NSC), which possess the ability to relies on a longer production of the therapeutic
differentiate into neurons, astrocytes, or oligoden- factors through the cell stimulation in the injured
drocytes in presence of specic growth factors site or the transplantation of autologous, allo-
and appropriate chemical environment. Despite genic, or xenogenic cells. Generally, the cell ther-
the risk of uncontrolled proliferation and some apy is used for the treatment of the
problems in the maintenance of cell viability, the neurodegenerative pathologies and of the chronic
use of the neural stem cells in membrane systems pain. In order to implant the cells in the interested
is still the most promising approach. Drug site, specic polymeric carriers are available.
352 Ceramic Membrane for Pervaporation
A few trial models suggested that this kind of DR (eds) The biomedical engineering handbook: tissue
approach could enhance the chances of the patient engineering and articial organs. Taylor & Francis,
New York
to recover partially the neuronal functionality, Tysseling VM, Kessler JA (2011) Biomaterials for central
opposing the symptoms and the cognitional prob- nervous system regeneration. Tissue Eng Neurolog
lems. Polymeric membranes ensure a consistent Neurosurg 533:455466
advantage for that kind of application, which con- Zhong Y, Bellamkonda RV (2008) Biomaterials for the
central nervous system. J R Soc Interface 5: 957975
sists in the chance to control cell viability and the
diffusion of nutrients and at the same time, the
protection of the cellular population from the
immunological reaction of the body. Genetically
engineered cells encapsulated into hollow ber Ceramic Membrane
membranes are able to produce and release ciliary for Pervaporation
neurotrophic factor (CNTF) in patients affected
by amyotrophic lateral sclerosis. Parkinsons dis- Wanqin Jin
ease is another degenerative pathology treatable State Key Laboratory of Materials-Oriented
with the cell transplantation. In this case, Chemical Engineering, Nanjing University of
dopamine-producing cells, dispersed also in dif- Technology, Nanjing, China
ferent hollow ber membranes of alginate,
showed an increase of the recovery and an anal-
gesic effect in the patient (Cho and Borgens 2012; Ceramic membrane for pervaporation, which is
Morelli et al. 2012). Therefore, the low percentage also called inorganic pervaporation membrane,
of cells surviving after implantation and the host belongs to inorganic microporous membranes
rejection and the inammatory response represent made from silica, alumina, or zeolite, which has
the main disadvantages of the cell therapy strate- a narrow pore size distribution typically with
gies. For this reason, despite the encouraging pore diameter smaller than 1 nm. Inorganic
results available, the complexity of the central pervaporation membranes have better structural
nervous system made the regenerative strategies stability without the problems of swelling or com-
always promising and not conclusive at the same paction relative to organic membranes. They usu-
time, and further investigation must be done in ally can withstand harsh chemical environments
order to ensure a safe and functional approach for and high temperatures (Wee et al. 2008). Most of
the recovery after an injury. the inorganic pervaporation membranes are asym-
metrical, with several macroporous support layers
providing the mechanical strength and a micropo-
References rous selective top layer providing the selectivity.
According to the membrane materials, inorganic
Cho Y, Borgens RB (2012) Polymer and nano-technology pervaporation membranes generally have two
applications for repair and reconstruction of the central kinds: zeolite membranes and silica membranes.
nervous system. Exp Neurol 233:126144
Zeolites are microporous aluminosilicate min-
Horner PJ, Gage FH (2002) Regeneration in the adult and
aging brain. Arch Neurol 59:17171720 erals with crystalline structures that have uniform
Kandel E, Schwartz J, Jessell T (2000) Formation and and molecular-sized pores, which have been used
regeneration of synapsis. In: Principles of neuronal extensively as catalysts and adsorbents. Zeolite
science, 4th edn. McGraw-Hill Medical, New York,
pp 10871114
membranes are synthetized by the deposition of
Morelli S, Piscioneri A, Salerno S, Rende M, Campana C, continuous polycrystalline zeolite layers on
Tasselli F, di Vito A, Giusi G, Canonaco M, Drioli E, porous supports, rst reported by Suzuki in 1987
De Bartolo L (2012) Flat and tubular membrane sys- (Suzuki 1987). Zeolite structures are made up of
tems for the reconstruction of hippocampal neuronal
network. J Tissue Eng Regen Med 6:299313
TO2, where T represents tetrahedral framework
Recknor JB, Mallapragada SK (2006) Nerve regeneration: atoms, such as Si, Al, B, Ge, Fe, and P. Most
tissue engineering strategies. In: Bronzino JD, Peterson zeolites contain Si with other metals substituted
Ceramic Membrane for Pervaporation 353
Ceramic Membrane
for Pervaporation,
Fig. 1 Atomic stick
representations for the
frameworks of CHA (a),
MFI (b), and MOR (c)
zeolite structures. The
nodes represent tetrahedral
framework atoms and the
C
sticks represent oxygen
bridges (Reproduced from
Bowen et al. 2004)
into the framework (Bowen et al. 2004). The dehydration, organic/organic separations (e.g.,
zeolite pores are made from rings in the frame- xylene isomers, methanol/methyl-tert-butyl ether
work and are designated by the number of oxygen mixtures, methanol/benzene mixtures), and acid
atoms making up the ring. Small-pore zeolites separations (e.g., acetic acid/water mixtures).
include those structures made up of eight-member In 2000, the rst industrial-scale separation
oxygen rings, medium-pore zeolites have plant, based on NaA zeolite membrane for
10-member rings, and large-pore zeolites have pervaporation dehydration of organic-water mix-
12-member rings. Examples of each of these are tures, was developed by Mitsui Engineering and
shown in the ball-and-stick representations of zeo- Shipbuilding Co. Ltd. in Japan (Morigami
lite frameworks in Fig. 1 (Bowen et al. 2004). et al. 2001). Until Now, NaA zeolite membranes
Several zeolite structures such as MFI, LTA, have been commercialized by the Busan Nano-
MOR, and FAU are widely reported as zeolite Research Institute Inc. (BNRI, a subsidiary of
membranes used for pervaporation, due to their Mitsui), SMART Chemical company (UK),
unique characteristic such as pore structure, Inocermic GmbH (Germany), and Nanjing Jiusi
adsorption properties, and their mechanical and Hi-Tech Co. (China) (Gascon et al. 2012). Unfor-
chemical stability, as described in Table 1 tunately, NaA zeolite membranes are very sensitive
(Baerlocher et al. 2002). Zeolite membranes are to even the slightest acidity.
most often prepared by hydrothermal synthesis, Another kind of ceramic membrane for
including in situ synthesis, secondary growth pervaporation is silica membranes, which now
method, continuous ow synthesis method, are commercially available and developed by
microwave synthesis, etc. (Wee et al. 2008). By The Netherlands Energy Research Corporation
utilizing the properties of molecular sieving and (ECN) (Verkerk et al. 2001). The microporous
preferential adsorption, the current pervaporation silica membranes consist of four support layers
applications of zeolite membranes mainly include of a- and g-alumina, and the selective top layer at
alcohol (e.g., ethanol, iosopropanol, butanol) the outer wall of the tube is made of amorphous
354 Ceramic Membranes
Ceramic Membrane for Pervaporation, Table 1 Characteristics of the zeolite applied in membrane and an example
of its application in pervaporation (Reproduced from Baerlocher et al. 2002)
Zeolite
A Y ZSM-5 Mordenite
Structure type LTA FAU MFI MOR
Si/Al ratio 1 2.3 81 4
Cations Na Na, Ca Na Na
Pore size 0.41 nm 0.74 nm 0.74 nm 0.51 nm 0.55 nm 0.65 nm 0.70 nm
0.41 nm 0.53 nm 0.56 nm 0.34 nm 0.48 nm
0.26 nm 0.57 nm
Channel network Three dimensional Three dimensional Three dimensional One dimensional
Application Dehydration of Separation of Separation of Separation of benzene/
organic mixture methanol/MTBE xylene isomers p-xylene mixture
mixture
silica. Silica membranes are mechanically and

chemically robust. Not only are the silica mem- Ceramic Membranes
branes stable in a normal industrial environment,
they can even be used to remove water from Bo Wang
condensation reactions like esterications, where Department of Chemical Engineering, Imperial
strongly acidic conditions predominate. However, College London, London, UK
silica membranes are not stable against water, and
thus the addition of oxides such as ZrO2 and TiO2
is often employed to improve the membrane sta- Ceramic membranes are made of ceramic mate-
bility toward water and alkali (Wee et al. 2008). rials, and to be specic, they must undergo a high-
temperature sintering process to fall into the
ceramic membranes category. Ceramic mem-
References branes belong to inorganic membranes category
and need to be distinguished with other types such
Baerlocher C, Meier WM, Olson DH (2002) Atlas of
as metallic membranes or other hybrid inorganic
zeolite framework types, 5th edn. Elsevier, Amsterdam
Bowen TC, Noble RD, Falcone JL (2004) Fundamentals membranes.
and applications of pervaporation through zeolite mem-
Gascon J, Kapteijn F, Zornoza B, Sebastin V, Casado C,
Coronas J (2012) Practical approach to zeolitic mem-
Membrane Types
branes and coatings: state of the art, opportunities,
barriers, and future perspectives. Chem Mater Ceramic membranes can be divided to two types:
24:28292844 porous and dense. Porous ceramic membranes
Morigami Y, Kondo M, Abe J, Kita H, Okamoto K (2001)
have found their industrial usages in various l-
The rst large-scale pervaporation plant using tubular-
type module with zeolite NaA membrane. Sep Purif tration processes, while dense membranes are still
Technol 25:251260 at stages not more than pilot plants.
Suzuki H (1987) Composite membrane having a surface Porous ceramic membranes separate sub-
layer of an ultrathin lm of cage-shaped zeolite and
stances based on size exclusion; therefore, they
processes for production thereof. US Patent 4,699,892
Verkerk AW, van Male P, Vorstman MAG, Keurentjes JTF can be further categorized as microltration, ultra-
(2001) Description of dehydration performance of ltration, and nanoltration membranes
amorphous silica pervaporation membranes. J Membr according to the pore size. Dense ceramic mem-
Sci 193:227238
branes separate one or two particular species
Wee S-L, Tye C-T, Bhatia S (2008) Membrane separation
process-pervaporation through zeolite membrane. Sep (usually oxygen or hydrogen) based on the ionic
Purif Technol 63:500516 conductivity of membrane materials, and they
Ceramic Membranes 355
often need to be used at high temperatures to materials is oxygenion conductor, for example,
increase their ionic conductivity. uorite type oxides including fully stabilized zir-
conia and doped ceria, which are commonly used
in SOFC, and perovskite type oxides that can be
Membrane Materials used alone to separate pure oxygen from air at
high temperatures. Another important category is
Porous ceramic membranes are usually used in proton conductor, which also have attracted wide C
liquid ltration processes and mostly in aqueous interests to use them in SOFC and high-
environments; therefore, the membrane material temperature hydrogen separation.
must be compatible with water aside from proper
mechanical strength. Most commonly used mem-
brane materials are alumina, titanium oxide, zir- Membrane Structure and Fabrication
conia, and silicon carbide. Alumina is
predominant among all ceramic membrane mate- Most ceramic membranes use asymmetric struc-
rials because of its availability and relatively low tures to obtain good strength and high permeation
cost, and it also provides strong mechanical prop- uxes, especially for porous membranes,
erties, inert chemical nature, good thermal stabil- although symmetric structures are also found in
ity, and good surface hydrophilicity to the dense ceramic membranes. In an asymmetric
membranes. Zirconia or stabilized zirconia gives structure, a relatively thick and porous ceramic
similar membrane characteristics as alumina supporting layer is used, which is the source of
membranes but much better mechanical proper- mechanical strength of the membrane, and a thin
ties; however, their price is also higher. Titanium separating layer is coated/deposited onto the sup-
oxide has better hydrophilicity and hydrothermal port to provide selectivity for the membrane.
stability than alumina and zirconia, but its The supporting layer can be formed by extru-
mechanical strength is moderate; therefore, tita- sion, or by pressing, followed by partial sintering.
nium oxide is normally used only for the separa- The support is then coated/deposited with a
tion layer supported by alumina membrane desired separating layer. Different coating/depo-
supports. Silicon carbide membranes are more sition techniques have been developed for differ-
permeable than other ceramic materials and they ent purposes, such as dip coating and wash
often give impressive uxes, but the major issue is coating, solgel process, and chemical or physical
the fabrication cost, which is much higher than vapor deposition techniques.
others, due to the expensive raw material and
extremely high sintering temperatures, as well as
the protecting atmosphere required for sintering. Applications
Dense ceramic membranes are used at high
temperatures, and these membranes normally are Porous ceramic membranes have been commonly
made of ion-conductive materials or mixed ion used in water and wastewater treatment, food and
and electron-conductive materials. For the former beverage processing, pharmaceuticals, and chem-
ones, an external circuit is needed to pass elec- ical processing. The most widely used membrane
trons to compensate the charge of ionic species types are microltration and ultraltration mem-
when ions move from one side to the other side, branes, although the demand of nanoltration
and work can be done within the external circuit. ceramic membranes is also increasing in recent
This is the principle of solid oxide fuel cells years.
(SOFCs), and the dense membrane acts as the Dense ceramic membranes have not really
electrolyte in a SOFC. For the later ones, electrons found their industrial usages, but some pilot plants
can pass the membrane directly, and separation of of oxygen permeable ceramic membranes with a
gas species can be completed by its own. The considerable scale are running in the USA and
most important category of ion-conductive Europe, and some commercial SOFC products
356 Ceramic Multilayer Membranes
are also available in the market. Some key techni- sintering (Burggraaf and Cot 1996). In general,
cal obstacles need to be addressed before these the process of multilayering involves preparation
dense ceramic membranes and the membrane- of multilayer through colloidal or polymeric
based technologies can enter into industrial sol-gel route. Precursors for colloidal sol-gel pro-
usages, such as high-temperature sealing and sys- cess may be organometallic compounds (e.g., alu-
tem stability. minum or titanium based). This is hydrolyzed
using excess of water. The sol is further peptized
at certain temperature. It is also possible to pre-
References pare multicomponent sols. Supported gel layers
are formed on membrane support by dip coating
Armstrong PA (2015) Development of ITM oxygen tech- the supports with freshly prepared sols. Obtained
nology for integration in IGCC and other advanced
gel layers are dried at room temperature followed
power generation. United States. doi:10.2172/
1224800. http://www.osti.gov/scitech/servlets/purl/ by calcinations (Van Gestel et al. 2002).
1224800 Surface of the multilayer membrane can also
Kneer R, Toporov D, Forster M, Christ D, Broeckmann C, be modied for different applications. Surface
Pfaff E, Zwick M, Engels S, Modigell M (2010)
modication of hydrophilic g-Al2O3/anatase-
OXYCOAL-AC: towards an integrated coal-red
power plant process with ion transport membrane- TiO2 multilayer membranes by a silane coupling
based oxygen supply. Energy Environ Sci 3(2): treatment in order to obtain NF membranes
198207 applicable in nonpolar solvents has been studied
Pfaff EM, Kaletsch A, Broeckmann C (2012) Design of a
(Van Gestel et al. 2003).
mixed ionic/electronic conducting oxygen transport
membrane pilot module. Chem Eng Technol 35(3): Multilayer ceramic membranes have got huge
455463 potential for gas permeation applications which
include separation of CO2 or H2 from methane
and other hydrocarbons, adjustment of H2/CO
ratio in syngas, separation of air into nitrogen
Ceramic Multilayer Membranes and oxygen, and recovery of helium and methane
from biogas. Inorganic membranes also offer pro-
Swachchha Majumdar found potential in applications for gas separation
Ceramic Membrane Division, CSIR-Central in a high-temperature catalytic reaction and are
Glass and Ceramic Research Institute, Kolkata, potentially useful for the separation of
West Bengal, India trace contaminants such as ammonia, hydrogen
sulde, and carbonyl sulde from coal gasier
fuel gases (Othman et al. 2004).
A ceramic macroporous substrate is required to
build multilayers over the substrates. Separation
properties of a multilayer ceramic membrane nor- References
mally depend upon the microstructure of the top
Burggraaf AJ, Cot L (1996) Fundamentals of inorganic
layer (Li 2007). However, there may be necessity
membrane science and technology. Elsevier,
of intermediate layers between the macroporous Amsterdam
substrate and top layer which give certain advan- Li K (2007) Ceramic membranes for separation and reac-
tages like preventing sudden differences in pore tion. Wiley
Othman MR, Mukhtar H, Ahmad AL (2004) Gas perme-
size of support and top layer. The multilayering is ation characteristics across nano-porous inorganic
carried out by a coating technique. When a tubular membranes. IIUM Eng J 5(2):17
(or at) substrate is withdrawn from a suspension Van Gestel T, Vandecasteele C, Buekenhoudt A,
or colloidal solution in a well-dened manner, a Dotremont C, Luyten J, Leysen R, Van der
Bruggen B, Maesc G (2002) Alumina and titania mul-
wet and more or less dense dispersion layer of
tilayer membranes for nanoltration: preparation, char-
well-dened thickness covering the substrate can acterization and chemical stability. J Membr Sci
be further processed through drying, calcining, or 207:7389
Ceramic Supported Polymer Composite Membranes in Pervaporation 357
Van Gestel T, Van der Bruggen B, Anita B, Dotremont C, et al. 2007) and poly(vinyl alcohol) (PVA) (Zhu
Luyten J, Vandecasteele C, Maes G (2003) Surface et al. 2010) are the typical hydrophobic and
modication of g -Al2O3/TiO2 multilayer, membranes
for applications in non-polar organic solvents. J Membr hydrophilic membrane materials, respectively.
Sci 224:310 To keep a low transport resistance meanwhile
Venkataraman V, Rath LK, Stem SA (1992) Potential ensures to obtain a defect-free selective layer;
applications of microporous inorganic membranes the mean pore size of the porous ceramic support
to the separation of industrial gas mixtures, Key
engineering materials, vols 6162, pp. 347352 ranges from a few nanometers to submicron size. C
Like most composite membranes, polymer/
ceramic composite membranes possess high
selectivity from the polymeric layer and high
Ceramic Supported Polymer ux from the porous ceramic support. Compared
Composite Membranes with traditional polymer composite membranes
in Pervaporation using organic support, ceramic-supported poly-
mer composite membranes show advantages of
Wanqin Jin chemical, mechanical, and thermal stability. Con-
State Key Laboratory of Materials-Oriented sidering the practical application, the congura-
Chemical Engineering, Nanjing University of tions of polymer/ceramic composite membrane
Technology, Nanjing, China usually include tubular and hollow ber.
Polymer/ceramic composite membrane is
always fabricated by dip-coating method in
Ceramic-supported polymer composite mem- which a certain polymer solution is deposited on
branes in pervaporation are also called polymer/ the outer or inner surface of the pretreated ceramic
ceramic composite pervaporation membranes and support. In general, some key preparing parame-
are a type of composite membranes consisting of a ters such as polymer molecular weight, polymer
thin dense polymeric separation layer and a concentration, cross-linking degree, dip-coating
porous inorganic ceramic support layer time, and surface properties of ceramic support
(as shown in Fig. 1). Theoretically, the polymeric play an important role in membrane preparation.
materials used for preparing composite mem- The separation performance of polymer/ceramic
branes are also suitable to produce polymer/ composite membrane is determined by the integ-
ceramic composite membranes. For instance, rity of the top polymeric layer and the thickness of
polydimethylsiloxane (PDMS) (Xiangli the separation layer and interfacial layer created
Ceramic Supported Polymer Composite Membranes in Pervaporation, Fig. 1 Structure and morphology of
polymer/ceramic composite pervaporation membrane
358 CH4/N2 Separation
by the penetration of polymer solution into the

porous ceramic support (Liu et al. 2012). CH4/N2 Separation
According to the membrane materials of the
polymeric separation layer, the applications of A. Baudot
ceramic-supported polymer composite mem- Physics and Analysis Division, IFP Energies
branes in pervaporation include organic com- nouvelles, Solaize, France
pounds recovery from aqueous solution (Xiangli
et al. 2008), biofuel production (Liu et al. 2012),
gasoline desulfurization (Xu et al. 2010), solvent The US pipeline specication requires an inert
dehydration (Zhu et al. 2010), and hybrid process content lower than 4 % in natural gas. It is esti-
(Liu et al. 2011; Lv et al. 2012). For example, the mated that 14 % of US present proven gas
PDMS/ceramic composite membrane can be reserves contain more than 4 % nitrogen (Baker
applied for pervaporation recovery of ethanol and Lokhandwala 2008; Lokhandwala
from dilute water solution. With 5 wt.% ethanol/ et al. 2010). The most trivial solution consists in
water in the feed at 40 C, the membrane exhibits diluting small ow rates of nitrogen-concentrated
a high ux of 1.64.5 kg/m2h and separation natural gas with main streams of natural gas
factor of 6.58.5 with ethanol concentration in containing a low concentration in inert gases.
the permeate of 2631 wt.%. However, if this is not feasible (for instance,
when all the surrounding gas elds contain high
concentrations in nitrogen), it is necessary to
References install a nitrogen removal plant to comply with
the regulatory natural gas composition specica-
Liu GP, Wei W, Wu H, Dong XL, Jiang M, Jin WQ tions. Cryogenic distillation is the dedicated tech-
(2011) Pervaporation performance of PDMS/ceramic nology for large-scale nitrogen removal
composite membrane in acetone butanol ethanol (ABE) operations (60,000600,000 Nm3/h capacity for
fermentation-PV coupled process. J Membr Sci
373:121129
operation life higher than 10 years). At the very
Liu GP, Wei W, Jin WQ, Xu NP (2012) Polymer/ceramic end of the 1990s, 26 such plants were in operation
composite membranes and their application in on US ground (Lokhandwala et al. 2010). Two
pervaporation process. Chin J Chem Eng 20:6270 technologies have been recently introduced in
Lv BD, Liu GP, Dong XL, Wei W, Jin WQ (2012) Novel
reactive distillation-pervaporation coupled process for
order to solve this challenge at small and medium
ethyl acetate production with water removal from scale: pressure-swing adsorption (PSA) with
reboiler and acetic acid recycle. Ind Eng Chem Res molecular sieves (Molecular Gates by the
51:80798086 Engelhard company (Mitariten 2001, 2004) and
Xiangli FJ, Chen YW, Jin WQ, Xu NP (2007)
Polydimethylsiloxane (PDMS)/ ceramic composite
gas permeation, which has been promoted by
membrane with high ux for pervaporation of MTR Inc. since the end of the 1990s
ethanol-water mixtures. Ind Eng Chem Res (Lokhandwala et al. 2010).
46:22242230 In the mid-1990s, MTR Inc. evaluated at lab
Xiangli FJ, Wei W, Chen YW, Jin WQ, Xu NP (2008) Opti-
mization of preparation conditions for polydimethyl-
scale (on 12 m2 membrane modules) a wide
siloxane (PDMS)/ceramic composite pervaporation array of composite membranes which selective
membranes using response surface methodology. membranes were made of various glassy or
J Membr Sci 311:2333 rubbery polymers (Lokhandwala et al. 2010).
Xu R, Liu GP, Dong XL, Jin WQ (2010) Pervaporation
separation of n-octane/thiophene mixtures using
Intuitively, rubbery polymers proved to be
polydimethylsiloxane/ceramic composite membranes. more selective toward methane (due to its
Desalination 258:106111 higher sorption, induced by its more condens-
Zhu YX, Xia SS, Liu GP, Jin WQ (2010) Preparation of able behavior than nitrogen), while glassy poly-
ceramic-supported poly(vinyl alcohol)-chitosan com-
posite membranes and their applications in
mers, especially uorinated ones, proved to be
pervaporation dehydration of organic/water mixtures. more selective toward nitrogen (due to diffu-
J Membr Sci 349:341348 sive selectivity).
CH4/N2 Separation 359
MTR Inc. evaluated the economic feasibility of whether the membrane material was nitrogen-
a membrane-based nitrogen removal operation on selective or methane-selective. Therefore,
a 12,000 Nm3/h natural gas stream entering the two-stage recycle schemes were studied involving
process at 32 bar and containing 10 % nitrogen. only nitrogen-selective membranes or only
The target of the membrane operation was to methane-selective membranes or a combination
reach the 4 % nitrogen specication in the product of both. Only the use of methane-selective mem-
gas while the waste gas was containing at least branes proved to offer interesting performances C
50 % nitrogen and the overall methane recovery (Lokhandwala et al. 2010), though the resulting
yield was to be higher than 93 %. Based on the process schemes were leading to signicant extra
performances of the various tested membrane cost due to the presence of two compressors: the
materials, the study led to the conclusion that a rst compression unit was used to pressurize the
single-stage membrane operation was not viable, methane-enriched permeate produced by the rst
Nitrogen removal (4 to 8% N2) 1 2 3 4 5

2
(NitrosepTM, MTR Inc.)
Flow rate (103 Nm3/h) 12 9.7 2.3 1.5 0,8
Pressure (bar) 70 14 ? ? 5.6 1 3 5
N2 8 4 26 50 12.6
Composition (% mol.)
CH4 92 96 92.5 50 87.4
CH4 recovery = 86 %
550 m membrane area 4
690 kW (theoretical compression power)
stream 5 used as compressor fuel
stream 4 to flare
2
3
Nitrogen removal (8 to 15% N2 ) 1 2 3 4 5 6
Flow rate (103 Nm3/h) 12 9 ? ? 1,9 1,1
Pressure (bar) 70 14 ? ? ? 3.5 1 6
N2 15 4 8.7 50 65 26
CH4 85 96 81.3 50 35 74
CH4 recovery = 86 %
1,800 m membrane area 4 5
1,860 kW (theoretical compression power)
stream 6 used as compressor fuel
stream 5 to flare
4
3
Nitrogen removal (15 to 30% N2 ) 1 2 3 4

Flow rate (103 Nm3/h) 12 9 ? ? 1
Pressure (bar) 70 14 ? ?
N2 15 80 11 4
CH4 85 20 89 96
CH4 recovery = 96 %
3,500 m membrane area
2,680 kW (theoretical compression power)
part of stream 1 used as compressor fuel
stream 2 to flare
2
CH4/N2 Separation, Fig. 1 Permeation-based nitrogen removal from natural gas (After Lokhandwala et al. 2010)
360 CHA Zeolite Membrane
membrane stage to the pipeline pressure, while the References

second one was aiming at boosting the pressure of
the permeate stream issued from the second-stage Baker RW, Lokhandwala K (2008) Natural gas processing
with membranes: an overview. Ind Eng Chem Res
in order to recycle it at the input of the rst
47:21092121
membrane stage. With this scenario, the total Lokhandwala KA, Pinnau I, He ZJ, Amo KD, DaCosta
equipment investment cost was in a 48 million AR, Wijmans JG, Baker RW (2010) Membrane sepa-
US dollar range (in 2009), while the required ration of nitrogen from natural gas: a case study from
membrane synthesis to commercial deployment.
membrane area was approximately equal to
4,000 m2. This appears to be acceptable by gas Mitariten M (2001) New technology improves nitrogen-
processors as the payback time of this type of removal economics. Oil Gas J 99:4244
installation was estimated to be 1 year. Mitariten M (2004) Economic nitrogen removal.
Hydrocarb Eng 9:5357
The rst eld test of these membranes was
therefore conducted with a 40 in-long module
equipped with few m2 of methane-selective mem-
branes for more than 1 year on a shut-in gas well
containing 15 % nitrogen operated by Butcher CHA Zeolite Membrane
Energy in Southern Ohio. The performances of
the spiral-wound module were very similar to Tsuneji Sano
those observed at lab scale, and its permeation Department of Applied Chemistry, Hiroshima
properties remained unchanged for the rst 6- University, Graduate School of Engineering,
month period. After this proof of principle, MTR Higashi-Hiroshima, Japan
Inc. designed different membrane-based nitrogen
removal process schemes, trade-named
Nitrosep, with various module/compressor con- Zeolites are a class of crystalline aluminosilicates
gurations depending on the nitrogen concentra- with highly regular and open microporous struc-
tion to be addressed (Fig. 1). tures. More than 200 types of zeolite frameworks
Up till now, 12 Nitrosep units have been have been identied by the Structure Commission
installed in the industry. Few examples depicting of the International Zeolite Association. Zeolite
the applications addressed by this membranes membranes combine the great advantages of inor-
process are listed hereafter: ganic membranes, such as temperature stability
and resistance against solvents, with the molecu-
A very small unit was installed in Southern lar sieving effect.
Kentucky in order to upgrade 200 Nm3/h of Zeolite chabazite (CHA, where the three char-
natural gas containing 7 % nitrogen. This acters indicate the framework type) with Si/Al
two-membrane module unit was able to ratios of 2 ~ 3 is known to possess a three-
recover 80 % of the natural gas with a nitrogen dimensional pore system with large ellipsoidal
content of 3.8 %. Part of the residue gas was cages (6.7 10 ) that are accessible via eight-
used as compressor engine fuel, while the rest membered ring windows (3.8 3.8 ). Figure 1
was vented. shows the framework structure of CHA zeolite
A much larger unit was installed in Rio Vista, (IZA web. 2013). High-silica CHA (Si/Al ratio
California (14,000 Nm3/h treatment capacity). > 5) has attracted great interest owing to its ther-
The membrane operation was aiming at mal and acid stabilities, and hence, application of
upgrading a 16 % nitrogen containing natural this material to membrane has been widely inves-
gas to a gas stream offering 10 % more heating tigated. The high-silica CHA zeolite membrane
value (lowering thus the nitrogen content to can separate light-gas mixtures of CO/N2, CO2/
9 %). The membrane system involved three CH4, H2/CH4, and H2/n-C4H10 with notably
stages in series and allowed a methane recov- higher selectivity than that allowed by the Knud-
ery yield of 95 % for pipeline delivery. sen mechanism (Kalipcilar et al. 2002). The high-
CHA Zeolite Membrane 361
hydrothermal reaction is carried out at 130 C

for 7 days using a Teon -lined autoclave. After
cooling the autoclave, the support tube is recov-
ered, washed with distilled water, and dried over-
night in air at room temperature. Finally, the
support tube is calcined at 550 C for 20 h.
Figure 2 shows scanning electron micrographs C
(SEM) of the surface and cross section of the
CHA membrane on the a-alumina support
(Yamanaka et al. 2012). The high-silica
CHA-type zeolite membrane prepared on the
a-alumina tube is connected with the stainless
steel tube using heat shrink tubing, and the tube
is subsequently set in the conventional batch-type
pervaporation apparatus. An acetic acid aqueous
solution (50/50 wt%) is used as the feed at
75 C. The compositions of the feed and
the permeate are determined by FID-gas chroma-
CHA Zeolite Membrane, Fig. 1 Framework structure of tography (5MS capillary column). The perme-
CHA zeolite (IZA web. 2013)
ation ux and separation factor, a(H2O/
CH3COOH), are calculated from the following
silica CHA zeolite membrane also exhibits excel- equations:
lent dehydration performance for water/alcohol
mixtures (Hasegawa et al. 2010). Flux kg=m2 h
As the high-silica CHA zeolite has only been
weight of permeate, kg
synthesized using expensive N,N,N-trimethyl-1-
membrane area, m2 permeation time, h
adamantammonium cation (TMAda+) as a
structure-directing agent (SDA). Very recently, (1)
success was achieved in synthesizing high-silica
CHA zeolites with Si/Al ratios of 521 from FAU Separation factor aH2 O=CH3 COOH
zeolite using the benzyltrimethylammonium cat- CH2 O =CCH3 COOH Permeate
(2)
ion (BTMA+) instead of the expensive TMAda+. CH2 O =CCH3 COOH Feed
The high-silica CHA zeolite synthesized by the
interzeolite conversion of FAU zeolite has supe- where the CCH3 COOH and CH2 O are the weight
rior acid stability (structural and compositional fractions of acetic acid and water, respectively.
stabilities) as compared to the CHA zeolite syn- The permeate ux and separation factor,
thesized using TMAda+ (Yamanaka et al. 2012). a(H2O/CH3COOH), are 7.9 kg/m2 h and ca.
Polycrystalline high-silica CHA zeolite mem- 2500, respectively, and the membrane perfor-
branes can be formed by the secondary growth mance is identical to that of commercially
of seed crystals on the outer surface of porous available NaLTA zeolite membranes used for the
supports such as a porous a-alumina tube. Seed dehydration of alcohol solution (Sato et al. 2008).
crystals are applied to the outer surface of the The long-range time courses of both the separa-
support tube by rubbing in order to implant tion factor and the ux are listed in Table 1
the seed crystals for nucleation. The secondary (Yamanaka et al. 2012). The CHA membrane
growth solution with a chemical composition prepared by the interzeolite conversion of FAU
of SiO2:0.03Al2O3:0.2BTMA:0.1NaCl:10H2O is zeolites has high potential for use in the separation
prepared from the dealuminated FAU, BTMAOH, of water from acidic organic solvents and is not
NaCl, and distilled water. Thereafter, a limited to acetic acid.
362 Characterization of Porous and Dense Membranes
CHA Zeolite Membrane, a b

Fig. 2 SEM images of (a)
outer surface and (b) cross
section of CHA zeolite
membrane (Reproduced
from Yamanaka
et al. (2012) with the CHA
permission of Elsevier)
Al2O3 support
5 m 5 m
CHA Zeolite Membrane, Table 1 Time course of

CHA-type zeolite membrane performance for dehydration Characterization of Porous
of 50 wt% acetic acid aqueous solution at 75 C and Dense Membranes
(Reproduced from Yamanaka et al. (2012) with the permis-
sion of Elsevier)
Lidietta Giorno1, Enrico Drioli1,2,3,4 and
Permeation
time/h a(H2O/CH3COOH) Flux/kg/m2 h
Heiner Strathmann5
1
1 2500 7.96
22 2480 7.88 Research Council, ITM-CNR, Rende, Italy
2
77 2500 7.91 Department of Environmental and Chemical
125 2380 7.80 Engineering, University of Calabria, Rende,
170 2480 7.80 Italy
3
WCU Energy Engineering Department,
Hanyang University, Seoul, S. Korea
4
Center of Excellence in Desalination
Technology, King Abdulaziz University, Jeddah,
References
Saudi Arabia
5
Hasegawa Y, Abe C, Nishioka M, Sato K, Nagase T, Universitt Stuttgart, Institute of Chemical
Hanaoka T (2010) Formation of high ux CHA-type Process Engineering Stuttgart, Baden-
zeolite membranes and their application to the dehy- Wurttemberg, Germany
dration of alcohol solutions. J Membr Sci 363:318324
International Zeolite Association Web site (2013) http://
www.iza-online.org/
Kalipcilar H, Bowen TC, Noble RD, Falconer JL Synonyms
(2002) Synthesis and separation performance of
SSZ-13 zeolite membranes on tubular supports. Chem
Rejection
Mater 14:34583464
Sato K, Sugimoto K, Nakane T (2008) Preparation of
higher ux NaA zeolite membrane on asymmetric Membrane characterization is a key part in mem-
porous support and permeation behavior at higher tem- brane research and development. In fact, the design
peratures up to 145 C in vapor permeation. J Membr
of membrane processes and systems depends on
Sci 307:181195
Yamanaka N, Itakura M, Kiyozumi Y, Ide Y, reliable data relating to membrane properties. To
Sadakane M, Sano T (2012) Acid stability evalua- identify a membrane type to be used for a particular
tion of CHA-type zeolites synthesized by interzeolite process, it is crucial to know its properties.
conversion of FAU-type zeolite and their membrane
Since membranes are very different in their
application for dehydration of acetic acid aqueous
solution. Microporous Mesoporous Mater 158: properties and applications, a large number of
141147 different techniques are required for their
Characterization of Porous and Dense Membranes 363
characterization. Some of the most important microltration to remove suspended materials, is

membrane characterization procedures described used as a test solution. The ux is determined at
in the literature or recommended by membrane different applied pressures and plotted as a func-
manufacturers are illustrated. tion of the pressure. From the slope of the plot, the
Porous micro- or ultraltration membranes hydrodynamic permeability is obtained. Prior to
are generally characterized in terms of their the ux measurements, the membranes should be
ux, their pore size, their pore size distribution, pressurized by ltration of pure water at a hydro- C
and their molecular mass cutoff. Other important static pressure that is at the upper limit of the
characteristics are the nature of the membrane recommended operating pressure for one to two
material (whether it is hydrophilic or hydropho- hours to compress the membrane to a stable value
bic or carrying positive or negative charges), the that will not further deform when used at lower
structure of the membrane (whether it is symmet- pressures. Possible preservatives, often present in
ric or asymmetric), and their mechanical, chem- membranes, must be eliminated by a washing
ical, and thermal stability. The methods used to procedure.
characterize porous membranes include simple Some inorganic membranes, like Al2O3-based
ltration and retention measurements, micro- types, show an apparent internal pore clogging
scopic techniques, surface tension measure- when deionized water is used in the pure water
ments, and the determination of pore radii by ux measurements. In these cases an electrolyte
liquid/gas or liquid/liquid replacement. The solution is used for determining the membrane
chemical and mechanical stability at various tem- ux, e.g., a NaCl or KCl solution. Since there
peratures is usually measured in long-term tests. exists no standard method for the determination
Which of the membrane characterization of the pure water ux, companies use their own
methods are applied depends on the use of the standards, which makes a reliable comparison of
membrane. pure water uxes rather difcult.
Typical pure water uxes of microltration
membranes vary from 500 to 50,000 L m2 h1
The Pure Water Flux of Micro- bar1, while pure water uxes of ultraltration
and Ultrafiltration Membranes membranes vary from 50 to 800 L m2 h1 bar1.
In practical applications, uxes are normally
A characteristic property of micro- and ultraltra- much lower, and a steep ux decline is observed
tion membranes is their pure water ux or pure during the rst period of operation. This ux
water permeability. The ux J through a porous decline is mainly determined by membrane foul-
medium is proportional to the applied hydrostatic ing which is generally more pronounced for
pressure and inversely proportional to the solvent microltration than for ultraltration.
viscosity. It is expressed in Darcys law, which is
given by:
Microscopic Techniques
Lp Dp
Jv (1)
A variety of microscopic techniques are used to
study membrane structures, such as scanning elec-
Here Jv is the membrane ux, Lp is the hydrody- tron microscopy, eld emission electron micros-
namic solvent permeability, is the viscosity of copy, transmission electron microscopy, and
the solution passing through the pores of the atomic force microscopy. The advantage of these
membrane, and Dp is the applied pressure. techniques is that direct visual information of the
For the determination of the pure water ux, it membrane morphology is obtained. The different
is very important to use a certain water quality. microscopic techniques require different prepara-
Generally deionized or distilled water with a con- tion methods and have different resolution. Scan-
ductivity of <5 mS cm1, which is treated by ning electron microscopy, e.g., has a resolution of
up to 5 nm and provides good information on the a resolution of ca. 1 nm (Fritsche et al. 1992;
structures of micro- and ultraltration membranes Hilal and Johnson 2010). Sample preparation is
including pore sizes and pore shapes. To reach a fairly simple, since no conductive coating is
high resolution, high electron beam energy has to required. Therefore, samples need not be dried
be applied, which can damage the membrane and exposed to vacuum. AFM can operate in air
samples. Another problem is that in electron and liquid phases; therefore, membranes can be
microscopy, a conductive layer must coat the studied in the environments used during their
samples. This is generally done in a high vacuum, operation. This is very valuable for determining
which can lead to a change of the structure of inuence of physical-chemical parameters (such
membranes made from hydrophilic polymers. as pH, ionic strength, charge) and functional
The pore sizes and pore shapes of ultraltration molecules on membrane surface properties.
membranes can also be determined by transmis-
sion electron microscopy which has a resolution
of 0.40.5 nm or by eld emission electron The Mechanical Properties
microscopy which has a resolution of of Membranes
0.60.7 nm. The transmission electron micros-
copy technique requires a signicantly more com- The mechanical stability and swelling behavior of
plex preparation procedure and is mainly applied the membranes is determined by the degree of
to determine the surface porosity of ultraltration cross-linking or crystallinity of the polymer
membranes (Fane et al. 1981). Because of the high matrix and the concentration of the xed charges.
resolution, pores in ultraltration membranes can The structure of membranes based on a highly
be visualized by eld emission electron micros- crystalline polymer such as the uorocarbon poly-
copy (Kim et al. 1990). mers are studied using transmission electron
Sample preparation for scanning electron microscopy and small angle X-ray diffraction
microscopy and eld emission electron micros- (Gierke et al. 1982). The degree of cross-linking
copy analysis is relatively easy compared to the is closely related to the swelling behavior and the
sample preparation for transmission electron mechanical properties of the membranes. But it
microscopy. To visualize the cross sections of a also affects the permselectivity. Infrared spectros-
membrane, e.g., it is immersed in ethanol or a copy measurements can provide some informa-
water/ethanol mixture prior to cooling in liquid tion on the type and degree of cross-linking of
nitrogen. The sample is fractured providing a ion-exchange membranes.
rather clean fracture and glued with conductive A mechanical characterization of ion-exchange
glue to a sample holder and sputtered with a thin membranes involves the determination of its
layer of gold. The sputtered layer should be thin- swelling and dimensional stability in different
ner than the structure features of the membrane to test solutions and tensile strength and hydraulic
be studied. Recently, low-vacuum scanning elec- permeability measurements. All mechanical char-
tron microscopy, i.e., so-called environmental acterization tests should be carried out with well-
scanning electron microscopy, has become avail- equilibrated membranes in a controlled environ-
able which makes it possible to scan a sample ment. The tensile strength and information
without a conductive coating. However, the reso- concerning the plastic or elastic deformation of a
lution of these microscopes is limited to magni- membrane is obtained from a stress versus strain
cations of ca. 7500. Figures 1 and 2 show typical diagram as shown schematically in Fig. 2.
pictures of membrane structures that can be The strain versus stress curves of ion-exchange
obtained by scanning electron microscopy membranes generally show three distinct areas. At
techniques. relatively low strain, the membranes show elastic
Besides electron microscopy, atomic force deformation; with increasing strain, the mem-
microscopy is also applied to study membrane branes show plastic deformation, and at a certain
surface structures. Atomic force microscopy has point they break. Since the mechanical properties
Characterization of Porous and Dense Membranes, track etching, (d) a porous cellulose nitrate membrane
Fig. 1 Scanning electron micrographs of (a) a sintered prepared from a homogeneous polymer solution by water
membrane prepared from a polymer powder, (b) a mem- vapor precipitation, and (e) a porous polypropylene mem-
brane prepared by stretching an extruded polytetrauor- brane precipitated by thermal gelation from a hot polymer
oethylene lm perpendicular to the direction of extrusion, solution
(c) of a capillary pore polycarbonate membrane made by
of membranes generally change drastically with membranes equilibrated in water or different solu-
the water content of the membranes, they must be tions similar to those used in practical
determined with dry membranes and with applications.
Characterization
of Porous and Dense
stress
Membranes,
Fig. 2 Stress versus strain
diagram of a membrane
sample indicating the elastic
and plastic deformation and
the breakpoint
break point
elastic plastic
strain
deformation deformation
The swelling of membranes depends on a num- water content of a membrane, a sample is equili-
ber of different parameters such as the nature of brated in a test solution. After removing the sur-
the basic membrane polymer, the nature of the face water from the sample, the wet weight of the
ion-exchange groups and their concentration in swollen membrane is determined. The sample is
the membrane, the counterions, the cross-linking then dried at elevated temperature over phospho-
density, the homogeneity of the membrane, and rus pentoxide under reduced pressure until a con-
the composition of the solution with which the stant weight is obtained. The water content of a
membrane is in contact. Especially the concentra- membrane is obtained in weight percent by:
tion of the solution has a signicant effect on the
W wet W dry
water sorption of the membrane because of wt% swelling 100 (2)
osmotic effects that are directly related to the W wet
chemical potential difference of the water in the
Here Wwet and Wdry are the weight of a membrane
membrane and in the solution. The state of the
sample in the wet and the dry state.
water in the membrane can also be rather differ-
ent. Part of the water is so-called free water.
Another part of the water is strongly bound within Membrane Separation Properties
the hydration shell of the counterions and the Determined by Filtration Test
xed charges of the membrane, while again
another part is more loosely bound to the basic Both micro- and ultraltration are pressure-driven
polymer matrix. The different water structures in membrane processes using porous membranes.
ion-exchange membranes are studied by differen- The separation of various components is based
tial scanning calorimetry, infrared spectroscopy, on a sieving mechanism, i.e., this membrane
and nuclear magnetic resonance (Komoroski and retains particles or molecules that are larger in
Mauritz 1982). Studies of the water structure size than the pores of a membrane. A relation as
within the membrane and its interaction with the a function of the particle size and the pore size for
membrane polymer, xed charges, and counter- the membrane retention can be dened by a rela-
ions require more sophisticated spectroscopic tion between the pore and the particle radius
measurements. The total water uptake of the (Ferry 1936). If the particles are spherical and
membrane in equilibrium with an electrolyte solu- have a radius rs and the membrane pores are
tion, however, can be determined quite easily by cylindrical with a radius rp, the following relation
measuring the weight difference between a mem- can be applied to describe the membrane solute
brane in the wet and dry state. To determine the rejection:
the experiment is performed for a very short time

rs 2 2
R 1 1 (3) period. Accumulation of solutes at the membrane
rp
surface due to concentration polarization effects
can also obscure the retention measurements and
For the radius of the molecule, either the radius of must be considered when the true, i.e., intrinsic,
gyration or the hydrodynamic radius can be used. retention shall be determined. Using low pressures,
The Ferry model is rather simplied, but gives a low feed concentrations, and a high degree of agi- C
good approximation of the expected retention tation can more or less eliminate these effects.
when particle size and the membrane pore radius Manufactures of ultraltration membranes
are known. generally make use of marker molecules, e.g.,
dextrans and proteins, for the determination of
the molecular weight cutoff value of their mem-
Retention and Molecular Weight Cutoff branes. In addition to the molecular weight cutoff
of a membrane, the sharpness of the cutoff is
To describe the retention of ultraltration mem- important for the practical application. The sharp-
branes, usually the molecular weight cutoff is ness of the cutoff of a membrane is determined by
used. The molecular weight cutoff determination measuring the retention of the membrane for com-
is very sensitive to the experimental conditions, ponents of different molecular weights and plot-
i.e., the solutes and their concentration used in the ting the retention versus the molecular weight of
test, the applied pressure, and the temperature. the retained components. The sharpness of the
Therefore, a comparison of retention data given cutoff of a membrane depends mainly on the
by different manufacturers is generally pore size distribution of the membrane. For prac-
problematic. As test conditions for determining tical applications, the cutoff of a membrane
the ultraltration membrane cutoff, a transmem- should be as sharp as possible.
brane pressure of 100 kPa, a feed solution con-
centration of 0.1 %, and a test temperature of
25 C are recommended. Furthermore, a maxi- The Bacterial Challenge Test
mum of agitation of the feed solution at the mem-
brane surface is required to minimize Micro- and ultraltration membranes are often
concentration polarization effects, and small used in the food and pharmaceutical industry for
amounts of solution should be ltered to assure a the sterilization of solutions. In this application, it
constant feed concentration. is important that membranes retain undesired
The retention is generally expressed in % and microorganisms as completely as possible, i.e.,
dened by: the retention for microorganisms should be close
to 100 %. Therefore, these membranes are
CP exposed to a bacterial challenge test. In this test,
R 1 100 (4)
Cf the membrane is challenged with a solution
containing a minimum concentration of organ-
Here R is the retention or rejection of a membrane isms such as Pseudomonas diminuta per cm3 of
and Cf and Cp are the feed and the permeate solution (Cheryan 1998). The data obtained in
concentration. bacterial challenge tests are normally expressed
The retention of molecules having identical in terms of the log reduction value, which is given
molecular weights may be rather different when by the logarithms of the ratio of feed to ltrate
their shape is different. Globular proteins, e.g., are concentration. The bacterial challenge test is a
rejected easier than branched polysaccharides or destructive test and is therefore not suited as an
exible polymers (Porter 1979). Adsorption of on-line quality control test. For this application,
the solute to the membrane surface, e.g., might the nondestructive diffusion test and the bubble
result in high apparent retention values when point integrity test are generally applied.
Characterization force
of Porous and Dense
Membranes,
Fig. 3 Schematic drawing rp
that illustrates the capillary
effect of a membrane pore
wetting liquid
Membrane Properties Determined by The term h r g in Eq. 5 can be replaced by the

Membrane Pore Size Measurement pressure p. This results in the Laplace equation,
which is given by:
In addition to ltration tests, micro- and ultral-
tration membranes are characterized by determin- 2s cos f
rp (6)
ing their pore size and pore size distribution. p
A number of techniques such as liquid/gas and
liquid/liquid displacement, mercury intrusion, The bubble point method is a fairly simple tech-
permporosimetry, thermoporosimetry, etc. are nique where the pores of a membrane are lled
used for the determination of the membrane with a liquid. From the bottom side of the mem-
pore size and pore size distribution. One of the brane, nitrogen gas or air is introduced with
more simple methods is the so-called bubble increasing pressure. At a certain pressure, the
point test. gas will replace liquid in the largest pores perme-
ating the membrane, and a bubble rising from the
The Bubble Point Test membrane surface can be observed.
The bubble point test is a structure integrity test in Equation 6 referred to as Laplace equation
which the largest pore or hole in a membrane is gives a relation for the surface tension between
determined. The bubble point method is based on the liquid and the gas and the required pressure to
the capillary effect of small pores due to surface open up pores of a certain size. In the bubble point
tension forces. In equilibrium these forces are test, mostly water or isopropanol is used, since the
balanced with gravity of the liquid in the pores. surface tension of water/air is approximately 3.5
The principle is shown in Fig. 3. times higher than for isopropanol/air.
This equilibrium between surface tension Other techniques to determine the pore size
forces and gravity of the liquid in the pore can and pore size distribution of a porous membrane
be expressed by the following equation: that are also based on the Laplace equation, but
cover different pore size ranges when the same
2pr p s cos f r 2p phrg (5) pressure range is applied, are mercury
porosimetry, gas-liquid displacement, and liquid-
Here rp is the pore radius, s is the surface tension liquid displacement.
of the liquid in contact with air, cosf is the contact
angle between the liquid and the wall of the pore, The Mercury Porosimetry
h is the height of the liquid column, r is liquid In mercury porosimetry, a dry membrane is
density, and g the acceleration of gravity. exposed to a certain volume of liquid mercury.
Characterization
of Porous and Dense minimum pore size
Membranes,
illustrating the gas/liquid dry flow
gas flow
displacement measurement
to determine the pore size
and pore size distribution of
microltration membranes mean flow pore size C
50% of dry flow
wet flow
pressure
bubble point
(maximum pore size)
The pressure applied to the mercury is slowly pressure and gas ow is observed when the pres-
increased, while simultaneously the amount of sure is further increased. Figure 4 shows schemat-
mercury forced into the porous structure is mea- ically typical gas uxes obtained as a function of
sured. The largest pores will be lled rst and the pressure with a dry and a liquid-lled membrane.
required pressure for the mercury to penetrate the The mean ow pore size is given by the point
porous structure increases with decreasing pore where the 50 % dry ow curve crosses the
size. The required pressure corresponds to a cer- wet ow curve. The maximum pore size is
tain pore radius and the amount of mercury that given by the point where a rst ow through the
disappears in the membrane corresponds to the liquid-lled membrane is observed. The mini-
total volume of these pores. The interfacial ten- mum pore size is given by the point where the
sion of mercury and air is very high. Therefore, uxes of the dry and the liquid-lled membrane
relatively high pressures are required to ll small become identical. The Laplace equation describes
pores. In polymer membranes, only pores having the relationship between the pore radius and the
a radius in excess of 1 mm can be detected reliably applied pressure.
by mercury porosimetry. Furthermore, this tech-
nique does not give distinction between dead-end Liquid-Liquid Displacement
pores and interconnected pores. Liquid-liquid displacement differs from
gas-liquid displacement in the displacing
Gas-Liquid Displacement medium; a second liquid instead of a gas displaces
The gas-liquid displacement technique is identical the liquid inside the pores. The two liquids applied
to the bubble point method. The pores of the should be immiscible. A typical liquid pair used is
membrane are generally lled with an organic water/isobutanol. Both liquids are rst saturated
solvent which has a lower surface tension than with each other before one of the liquids is used to
water. The solvent is forced out of the membrane ll the pores and the other liquid is applied as
pores by nitrogen gas, which is introduced with replacement liquid. The experimental setup for
increasing pressure. As the pressure increases, the the determination of pore sizes by liquid-liquid
liquid will be replaced by nitrogen in the largest displacement is shown in Fig. 5. A slow increase
pores rst, and a convective gas ow through the replacement liquid pressure pushes the liquid
these pores will occur. The gas ow is measured out of the largest pores rst. With increasing in
as a function of the applied pressure. When all the pressure, the liquid in the smaller pores will also
pores are opened, a linear relationship between be replaced. The pore size distribution can be
damper
p recorder
HPLC
pump
liquid A
computer
membrane digital
containing liquid B balance
Characterization of Porous and Dense Membranes, Fig. 5 Schematic representation of a typical liquid-liquid
displacement setup (Capannelli et al. 1983)
calculated from a plot of the ux versus the using a sensitive differential scanning calorimetry.
applied pressure (Capannelli et al. 1983; Wienk Permporometry is a technique based on gas
et al. 1994). adsorption/desorption measurements where
Whether isobutanol or water is used as pores are either lled by controlled condensation
displacing liquid depends very much on the nature of a vapor phase or where lled pores are emptied
of the membrane material. Due to swelling phe- by controlled evaporation of the liquid inside the
nomena, different results might be obtained pores. The relation between the relative vapor
depending on the liquid placed inside the pores. pressure and the pore radius is given by the Kelvin
The relation between the pore radius and the equation. Thermoporometry is based on the
pressure required to open pores of a certain size microcalorimetric analysis of a solid-liquid tran-
is again described by the Laplace equation, and sition using a sensitive differential scanning
the pore size distribution can now be calculated calorimetry.
when the ux as a function of the pressure is Table 1 shows the pressures required to open
measured. pores of 10 nm for different displacement sys-
Like in gas-liquid displacement, only active tems. A second advantage is that membranes are
pores contributing to transport are taken into characterized in a wetted state, and therefore the
account. Pore diameters of approximately test conditions are closer to the real operating
5100 nm can be analyzed using liquid-liquid conditions.
displacement. The advantage of liquid-liquid dis-
placement over gas-liquid displacement is the
lower interfacial tension between liquid pairs Dense Membrane Characterization
compared to a gas/liquid pair. Therefore, pores
of the same size can be opened at much lower Dense membranes are generally used for the sep-
pressures in liquid-liquid displacement. aration of low molecular mass materials such as
There are more methods that can be used to gases, salts, or solvents. The transport mechanism
determine pore sizes in porous membranes in these membranes is based on the solution and
(Cuperus et al. 1992a, b) such as permporometry diffusion of components within the membrane
or thermoporometry, which is based on the micro- matrix. The intrinsic properties of the basic mem-
calorimetric analysis of a solid-liquid transition brane material are often more relevant for
Characterization of Porous and Dense Membranes, The methods used to determine the structural
Table 1 Different porosimetry techniques and the influ- properties of ion-exchange membranes are similar
ence of the interfacial tension on the pressure required
opening pores of 10 nm to those used for characterizing other polymer
membranes. Because ion-exchange membranes
Pressure
are used mainly in separation processes with elec-
Technique System (mN/m) (Bar)
trical potential driving forces, their permeability
Mercury
porosimetry (G-L)
Air/mercury 480 749
and selectivity are determined under experimental C
Bubble point (G-L) Air/water 72.3 145 conditions closely related to their practical use,
Coulter porometry Air/Porol 16 32 and their properties are expressed in terms com-
(G-L) monly used in electrochemistry (e.g., electrical
Gas-liquid Air/isobutanol 20.7 41 resistance, ion-transfer numbers, charge densities)
displacement (G-L) (Cussler 1971, 1984).
Liquid-liquid Water/ 1.85 3.7
displacement (L-L) isobutanol
Liquid-liquid Water/ 0.35 0.7
displacement (L-L) isobutanol/ References
methanol
Berens AR, Hopfenberg HB (1982) Diffusion of organic
vapors at Low concentrations in glassy PVC, polysty-
rene and PMMA. J Membr Sci 10:283
Capannelli G, Vigo F, Munary S (1983) UF characteriza-
characterizing their transport behavior than is the tion methods. J Membr Sci 15:289
macroscopic membrane structure. Characteriza- Cheryan M (1998) Ultraltration and microltration hand-
book. Lancaster Publishing Company, Lancaster
tion methods used with microporous structures
Cuperus FP, Bargeman D, Smolders CA (1992a)
have therefore been supplemented with other pro- Permporometry. The determination of active pores in
cedures, such as sorption and diffusion measure- ultraltration membranes. J Membr Sci 71:75
ments or determination of the glass transition Cuperus FP, Bargeman D, Smolders CA (1992b) Critical
points in the analysis of membrane pore structure by
temperature and crystallinity of the basic mem-
thermoporometry. J Membr Sci 66:45
brane material. For ion-exchange membranes, the Cussler EL (1971) Membranes which pump. AICHE
measurement of electrokinetic properties is J 17:13001303
required. Cussler EL (1984) Diffusion. Academic, Cambridge, MA
Fane AG, Fell CJD, Waters AG (1981) The relationship
In addition to the determination of diffusion and
between membrane pore characteristics and ux for
partition (solubility) coefcients, measurements ultraltration membranes. J Membr Sci 9:245
providing information on the physical structure of Ferry JD (1936) Ultralter membranes and ultraltration.
the polymer, such as differential scanning calorim- Chem Rev 18:373
Fritsche A.K., Arevalo AR, Moore MD, OHara C (1992)
etry, X-ray studies, or determination of the free
Surface Structure and Morphology of Polyacrylonitrile
volume in a polymer, are important. These Membranes by Atomic Force Microscopy. J. Membr
methods are described in Koros and Chem Sci. 46:167
(1987), Berens and Hopfenberg (1982), and Gierke TD, Munn GE, Wilson FC (1982) Morphology of
peruorosulfonated membrane products wide- angle
Morel and Paul (1982). Permeability and selectiv-
and small-angle x-ray studies. ACS Symp Ser
ity are the most interesting parameters; they are 180:195216
determined in permeation tests and expressed in Hilal N, DJ Johnson (2010). The use of atomic
terms of rejection coefcients (in reverse osmosis) force microscopy in membrane characterisation. Com-
prehensive Membrane Science and Engineering. E.
or separation coefcients (in gas separation and
Drioli and L. Giorno. Oxford, Elsevier Science.
pervaporation). These coefcients are usually not 521538
constant but a function of composition, tempera- Kim KJ, Fane AG, Fell CJD, Suzuki T, Dickson MR
ture, and pressure. Furthermore, boundary layer (1990) Quantitative microscopic study of surface char-
acteristics of ultraltration membranes. J Membr Sci
effects at the membrane surface often obscure the
54:89
determination of membrane uxes and selectivities Komoroski RA, Mauritz KA (1982) Peruorosulfonated
in practical measurements. ionomer membranes. In: Eisenberg A, Yeager HL (eds)
372 Charged Mosaic Membrane
ASC symposium series 180. American Chemical Soci- groups, the chemistry of the ligand and the covalent
ety, Washington, DC linkage, and the properties of the spacer arm used to
Koros WJ, Chern RT (1987) Separation of gaseous mix-
tures using polymer membranes. In: Rousseau RW attach the ligand to the membrane (Zydney 2011).
(ed) Handbook of separation process technology. One of the rst applications of charged ultra-
Wiley, New York ltration membranes was in the recovery of
Morel G, Paul DR (1982) CO2 sorption and transport in electropaint. Electrodeposition of a charged paint
miscible polyphenylene oxide/polystyrene blends.
J Membr Sci 10:273 resin is used extensively in the industrial painting
Porter MC (1979) Membrane ltration. In: Schweizer PA of metallic surfaces in automobiles and large
(ed) Handbook of separation techniques for chemical household appliances. The paint resins are organic
engineers. Mac Graw-Hill, New York polymers with attached anionic or cationic
Wienk IM, Folkers B, van den Boomgaard T,
Schmolders CA (1994) Critical factors in the determina- groups. Ultraltration has been used to recover
tion of the pore size distribution of UF membranes using the charged electropaint from the dilute solution
a liquid displacement method. Sep Sci Technol 29:1433 produced by washing the excess paint off of the
metal part. Ultraltration can also be applied
directly to the paint in the electrodeposition bath,
with the permeate (essentially water) used for
Charged Mosaic Membrane subsequent rinsing steps (Zeman and Zydney
1996). The use of charged membranes (having
Mosaic Membranes the same polarity as the electropaint) provides
signicantly greater ltrate ux with less fouling
than neutral or oppositely charged membranes.
Charged ultraltration membranes can be used
Charged Ultrafiltration Membrane to signicantly improve the inherent tradeoff
between the permeability and selectivity of an
Andrew L. Zydney ultraltration membrane (Mehta and Zydney
Department of Chemical Engineering, 2005). The rate of solute transport through a
The Pennsylvania State University, charged ultraltration membrane is determined by
University Park, PA, USA a combination of steric (size based) and electro-
static interactions (Mehta and Zydney 2006). Elec-
trostatic interactions strongly effect the partitioning
Electrically charged ultraltration membranes can of charged solutes into the membrane pores. For
be used to reduce the fouling and increase the example, the presence of positively charged groups
retention of like-charged species. Most polymeric on the pore surface causes a strong electrostatic
membranes have a net negative charge in solution exclusion of positively charged species from the
due to the presence of trace anionic groups (e.g., membrane pores signicantly increasing the selec-
carboxylic acids) and/or the preferential adsorption tivity of the membrane. This makes it possible to
of negatively charged ions from the aqueous solu- employ charged ultraltration membranes with rel-
tion. It is also possible to cast membranes from atively large pore size and thus with very high
polymers containing xed charge groups, e.g., a permeability, with the required selectivity achieved
positively charged membrane can be developed by electrostatic exclusion of the protein.
using polymers containing xed amine groups. There is also considerable interest in performing
However, the most common method for generating protein separations by charged UF membranes. In
charged ultraltration membranes is by surface this case, the charged ultraltration membrane pro-
modication of a base polymer through the attach- vides very high retention of like-charged proteins,
ment of appropriate anionic (e.g., carboxylic or enabling uncharged proteins and smaller impurities
sulfonic acid) or cationic (e.g., quaternary amine) to be washed into the permeate by a dialtration
groups. The overall performance of these mem- process (see Fig. 1). The solution pH and ionic
branes is determined by the density of the charge strength can be adjusted to obtain high-resolution
Chemical and Electrochemical Equilibrium in Membrane Systems 373
Charged Ultrafiltration Membrane, Fig. 1 Positively charged membrane provides high retention of positively
charged protein while allowing transmission of neutral proteins
protein separations (van Reis et al. 1999). Van Reis

and Zydney (2007) have discussed a number of Chemical and Electrochemical
separation processes using charged ultraltration Equilibrium in Membrane Systems
membranes, including a nonafnity process for
purifying a monoclonal antibody from harvested Lidietta Giorno1, Heiner Strathmann5 and
cell culture uid. Enrico Drioli1,2,3,4
1
2
References Engineering, University of Calabria, Rende,
Italy
3
Mehta A, Zydney AL (2005) Permeability and selectivity WCU Energy Engineering Department,
analysis for ultraltration membranes. J Membr Sci Hanyang University, Seoul, S. Korea
249:245249 4
Mehta A, Zydney AL (2006) Effect of membrane charge Center of Excellence in Desalination
on ow and protein transport during ultraltration. Technology, King Abdulaziz University, Jeddah,
Biotechnol Prog 22:484492 Saudi Arabia
van Reis R, Zydney AL (2007) Bioprocess membrane 5
Universitt Stuttgart, Institute of
technology. J Membr Sci 297:1650
van Reis R, Brake JM, Charkoudian J, Burns DB, Zydney Chemical Process Engineering Stuttgart,
AL (1999) High performance tangential ow Baden-Wurttemberg, Italy
ltration using charged membranes. J Membr Sci
159:133142
Zeman LJ, Zydney AL (1996) Microltration and ultral-
tration: principles and applications. Marcel Dekker, Equilibrium in a system is achieved when all com-
New York ponents are in the same thermodynamic state, i.e.,
Zydney AL (2011) High performance ultraltration mem- the Gibbs free energy of the system is constant.
branes: pore geometry and charge effects. In: Oyama
ST, Stagg-Williams SM (eds) Inorganic, polymeric,
Equilibrium between two systems such as a mem-
and composite membranes: structure, function, and brane and an adjacent solution or two solutions
other correlations. Elsevier, Amsterdam, pp 333352 separated by a membrane is obtained when both
374 Chemical and Electrochemical Equilibrium in Membrane Systems
systems are in the same thermodynamic state, i.e., product Kw which is 1.008 1014 [mol2 L2] at
when the Gibbs free energies in both systems are 25 C (Atkins 1990). It is
identical. The Gibbs free energy of a system is a
function of various state variables such as pressure, H3 O OH K H2 O2 K w
electrical potential, and the activity of individual 1014 at 25 o C (3)
components. A system composed of water and an
electrolyte which is partially dissociated will be in The Kw value is a function of temperature. How-
equilibrium when dissociated and non-dissociated ever, at constant temperature and in electrolyte
components are in equilibrium. Considering two solutions with low to moderate ion concentra-
systems separated by a membrane which is imper- tions, the water concentration does not change
meable to some components and permeable to very much and Kw can be considered constant,
others, the two systems will be in equilibrium and in pure water, the concentrations of H3O+-
when all components that are able to permeate the and OH ions are identical and are 107 [mol L1]
membrane have identical electrochemical poten- at 25 C.
tials in both systems. This may also be the case if In electrolyte solutions containing an acid or a
the components do have differences in pressures, base, the concentrations of H3O+- and OH ions
electrical potentials, or the activities in the two are no longer identical, and their concentrations
systems as long as these differences compensate are expressed by the so-called pH value (potentia
each other and the sum of the state variables, i.e., hydrogenii). The pH value is dened as the nega-
the Gibbs free energy in both systems is identical. tive logarithm of the H3O+-ion concentration to
Examples for two systems being in equilibrium are the basis of 10. Thus, is
the osmotic equilibrium, in which pressure and
activity differences compensate each other, and H3 O 10pH (4)
the Donnan equilibrium which is obtained when
an electrical potential difference is compensated by and
activity difference. For membrane processes, the OH 1014pH (5)
dissociation equilibrium in a system as well as the
osmotic and the Donnan equilibrium between two Pure water or neutral electrolyte solutions at 25 C
systems are of importance and will be discussed in have a pH value of ca 7. Acid exhibits pH values
more detail. <7 and bases >7.
Water is dissociated to a certain extent into An important parameter for acids and bases is
hydronium and hydroxide ions, i.e., H3O+ and the dissociation constant expressed in the
OH ions according to the following equilibrium so-called pK value. The dissociation of acids and
relation: bases follows the general scheme:
2H2 O , H3 O OH
K1
(1) acids : HA H2 O , H3 O A (6)
The water dissociation equilibrium constant K is K2

bases : BOH , B OH (7)

H3 O OH
K (2) Here, K1 and K2 are the dissociation equilibrium
H2 O2
constants, and HA and BOH represent the
The brackets [] denote concentrations of the var- non-dissociated acid and base.
ious components in equilibrium. The acid and base equilibrium can be
In pure water, the concentration of water is expressed by the following relations:
55.6 mol L1. The product of K [H2O]2 is H3 O A
constant and referred to as the water dissociation K 1 H2 O K a (8)
HA
Chemical and Electrochemical Equilibrium phases consist of a solvent and a solute indicated by the
in Membrane Systems, Fig. 1 Schematic drawing illus- subscripts l and s. C and m refer to concentration and
trating the osmotic phenomenon. It shows two systems (0 ) chemical potential, p to hydrostatic, and p to osmotic
and (00 ) separated by a semipermeable membrane. The pressure, respectively, and J to the membrane ux
and number of molecules present in a solution and

B OH given by:
K2 Kb (9)
BOH
A B
a (12)
Here, Ka and Kb are the dissociation constants for nA1 AB A vBJ AB B
the acid and the base.
The dissociation constant between strong and Here, a is the dissociation degree and AB is the
weak acids or bases can vary by several orders of total concentration of the non-dissociated electro-
magnitude. Therefore, the dissociation constant is lyte. A and B+ are anion and cation
generally expressed by the so-called pKa and pKb concentrations.
values which are dened similarly to the pH value The degree of dissociation a has values between
as the negative logarithm of the dissociation con- 1 for a totally dissociated electrolyte and 0 for a
stant, i.e., completely non-dissociated electrolyte.
From practical experience, it is well known
pka log Ka 10pK a (10) that the degree of dissociation of weak acids or
and bases depends on the pH value of the solution. In
pkb log Kb 10pK b (11) solutions with high pH values, weak acids have a
higher degree of dissociation than in solutions
Multivalent acids or bases such as H2SO4 or Ca with low pH values.
(OH)2 which have more than one dissociable pro- If two aqueous salt solutions of different concen-
ton or hydroxide ion dissociate in consecutive trations are separated by a membrane which is per-
steps and, therefore, have more than one pKa or meable for the solvent, e.g., water, but impermeable
pKb value, respectively. for the solute, e.g., salt, a transport of water from the
In practical applications the degree of dissoci- more dilute solution in the more concentrated solu-
ation of an electrolyte is of interest. It is dened as tion is observed. This phenomenon, which is illus-
the ratio of dissociated molecules to the total trated in Fig. 1, is referred to as osmosis.
376 Chemical and Electrochemical Equilibrium in Membrane Systems
The gure shows a schematic diagram of a reverse osmosis

membrane separating two solutions consisting of
a solvent, e.g., water, and a solute, e.g., a salt. The dilute solution
two solutions are indicated by (0 ) and (00 ).
solvent flux
The membrane is assumed to be permeable to the
solvent but impermeable to the solute. Depending
on the concentrations and hydrostatic pressures in Dp
the two phases separated by the membrane, four
Dp=Dp
different situations can be distinguished:
concentrated solution
(a) The hydrostatic pressures in the two phases
separated by the membrane are equal, but the osmosis
solute concentration in solution (0 ) is higher
Chemical and Electrochemical Equilibrium
than the one in solution (00 ). In this case, the in Membrane Systems, Fig. 2 Solvent ux between
osmotic pressure in solution (0 ) is higher than two solutions of different concentrations through a strictly
that in solution (00 ), and there will be a ow of semipermeable membrane as function of the hydrostatic
pressure applied to the more concentrated solution
solvent from the more diluted solution (00 ) into
the more concentrated solution (0 ). This situa-
tion is referred to as osmosis. is shown as a function of the hydrostatic pressure
(b) The two phases separated by the membrane applied to the more concentrated solution.
have different hydrostatic pressures, but the As long as the applied hydrostatic pressure is
difference in hydrostatic pressure is equal to lower than the osmotic pressure, the difference
the difference in the osmotic pressures between the two solutions solvent will ow
between the two solutions acting in opposite from the more dilute solution into the more con-
direction. This situation is referred to as centrated solution by osmosis. When the hydro-
osmotic equilibrium, and there will be no static pressure exceeds the osmotic pressure
ow of solvent through the membrane, difference, the ow is reversed and solvent will
although the concentrations in the two solu- ow from the more concentrated solution to the
tions are different. dilute solution.
(c) The two phases separated by the membrane The osmotic pressure of a solution p is propor-
have different hydrostatic pressures, but the tional to the solute concentration. It can be calcu-
difference in hydrostatic pressure across the lated, however, from the activity of the solvent in
membrane is larger than that in the osmotic a solution and is given by:
pressure and is acting in opposite direction.
Thus, solvent will ow through the membrane RT
PS P l p ln asl (13)
from the solution (0 ) with the higher solute Vl
concentration into the solution (00 ) with the
lower solute concentration. This phenomenon For dilute solutions, the osmotic pressure can be
is referred to as reverse osmosis. expressed to a rst approximation by the concen-
tration of the individual components in the solu-
The ux of solvent between two homogeneous tion (Strathmann 2004):
solutions of different concentrations separated by
X
a semipermeable membrane, which is only per- p RT gi Ci (14)
meable for the solvent, to the more concentrated i
solution as a function of a hydrostatic pressure
applied is illustrated in Fig. 2. Here, the solvent If the solute is a salt, as is the case in many
ux between two solutions of different concentra- practical applications, the concentration of the
tions through a strictly semipermeable membrane individual ions must be considered when
determining the osmotic pressure, i.e., the degree constant; the subscript i refers to individual com-
of dissociation, and the stoichiometric coefcients ponents and the superscripts m and s refer to the
of the salt have to be considered for determining membrane and the electrolyte solution, respec-
the osmotic pressure. tively. The potential difference between the mem-
The osmotic pressure of an aqueous solution brane and the solution is referred to as Donnan
with salts is given by (Robinson and Stokes 1959): potential Don.
X The Donnan potential between an electrolyte C
p RT gi nic nia Cis (15) solution and an ion exchange membrane cannot
i be measured directly. It can, however, be calcu-
Here, g is the osmotic coefcient; the subscripts n, lated from the ion activities in the solution and the
c, and a refer to a stoichiometric coefcient, cat- membrane and by the pressure difference between
ion, and anion. the membrane phase and the adjacent solution, pm
In discussing the osmotic pressure, it was ps, which is referred to as swelling pressure and
assumed that two solutions separated by a mem- which is identical to the osmotic pressure differ-
brane are in chemical equilibrium. If the solution ence between the solution and the membrane.
contains charged components, i.e., ions, and the Introducing the osmotic pressure into Eq. 17
membrane is permeable for at least one ionic com- gives the Donnan potential as a function of the ion
ponent, the membrane will be in equilibrium with and the water activities in the membrane and the
the adjacent solution if the electrochemical poten- solution:
tial of all ions in the membrane and the solution are
1 as
equal. Thus, for each ion in equilibrium, it is Don RT ln mi V i Dp (18)
zi F ai
e
mmi e
m si mm
i zi F mi zi F
m s s
(16) The numerical value of the Donnan potential don
can be calculated either from the cation or anion
Here, mi and em i are the chemical and the electro- activities. For a single salt, thus is
chemical potential of the component i, is the
electrical potential, and F is the Faraday constant. 1 as
Don RT ln mc V c Dp
The superscripts m and s refer to the membrane zc F ac
and to the solution, respectively.
1 asa
Thus, the electrochemical potential of an ion is RT ln m V a Dp (19)
za F aa
composed of two additive terms. The rst is the
chemical potential and the second is the electrical Equation 19 gives a general relation for the cation
potential multiplied by the Faraday constant and and anion distribution at the interface between a
the valence of the ion. Introducingthe chemical solution and an ion exchange membrane.
potential mi dmi Vdpi RTdlnai into Eq. 16 It should be noted that the value of the Donnan
and rearranging gives the electrical potential dif- potential is negative for the cation exchange mem-
ference between the membrane and the adjacent brane and positive for the anion exchange mem-
solution (Donnan and Guggenheim 1932) to brane in equilibrium with a dilute electrolyte
solution.

1 asi The Donnan equilibrium describes the electro-

m s
RT ln m V p p Don
s m
zi F ai chemical equilibrium of an ion in an ion exchange
(17) membrane, and an adjacent solution and can be
calculated for a single electrolyte by rearranging
Here, is the electrical potential, a is the ion Eq. 19:
activity, V is the partial molar volume, z is the z1 z1
asa a am c
eRT
Dp V s
valence, F is the Faraday constant, p is the pres- c zc vc (20)
sure, T is the absolute temperature, and R is the gas am
a asc
378 Chemical Binding of Biomolecules to Membranes
Here, z is the valence, n is the stoichiometric

coefcient of the electrolyte, F is the Faraday Chemical Binding of Biomolecules
constant, R is the gas constant, T is the absolute to Membranes
temperature, V is the partial molar volume, and a is
the activity; the subscripts a and c refer to anion Christophe Innocent
and cation, and the superscripts s and m refer to Institut Europen des Membranes, ENSCM, CNRS,
the solution and the membrane, respectively. Universit de Montpellier, Montpellier, France
The Donnan equilibrium is an important rela-
tion in electro-membrane processes since it deter-
The immobilization in electropolymerized poly-
mines the distribution of co- and counterions
mers such as polypyrrole has been developed
between the membrane and the adjacent solution.
for a variety of biomolecules and provides
References Et2O, AlLiH4
Atkins PW (1990) Physical Chemistry, fourth edition,

N N
Oxford University Press, Oxford
Donnan FG, Guggenheim EA (1932) Exact thermodynam-
ics of membrane equilibrium. Z Phys Chem
A 162:346360 NH2
Robinson RA, Stokes RH (1959) Electrolyte solutions. N
Butterworths, London 1 2
Strathmann H (2004) Ion-exchange membrane separation
processes. Elsevier, Amsterdam Chemical Binding of Biomolecules to Membranes,
Fig. 1 Scheme of synthesis of amino propyl pyrrole
Chemical Binding
of Biomolecules
a
to Membranes, O O
Fig. 2 Laccase
immobilization on modied NH2
electrode with H2N laccase
polyaminopyrrole; (a)
nucleophilic attack on the
glutaraldehyde, (b)
reduction of imine function
(IEM picture)
b
N N laccase
N N laccase
H H
Chemical Cleaning of Membranes 379
very stable environment for the biocatalyst. The Fouling in membrane ltration is problem-
advantages of the polypyrrole are the control of atic but inevitable as it occurs with the retention
lm thickness and its contribution in electron of contaminants that accumulate on the mem-
transfer between the biocatalysts and the conduc- brane surface. The causes of fouling are often
tive support. specic, depending upon feed water constitu-
Amino pyrrole is synthetized according the ents, the membrane, and the operating regime.
following scheme (Fig. 1). Cleaning protocols are typically recommended C
Electropolymerization of 1-(2-cyanoethyl)pyr- by the membrane manufacturers with some
role monomer (1) in acetonitrile 0.1 M NBu4PF6 cleaners being proprietary. We can minimize
was performed by controlled-potential oxidation the rate of fouling, but we cannot eliminate
of the pyrrole at +1.01 V vs. SCE. Reduction of fouling, unless very extensive pretreatment is
the nitrile to amine function was carried out using applied. Types of fouling are presented along
a large excess of LiAlH4 in dried ether at room with cleaning, physical and chemical conditions
temperature. that have signicant impacts on cleaning ef-
Enzyme is grafted on the modied tube (with ciency, and compatibility of membrane mate-
amino polypyrrole) by using glutaraldehyde as rials with cleaning agents.
coupling agent (Fig. 2). Membrane fouling is a complicated phenome-
Fixation of laccase (enzyme which catalyzed non and results from numerous causes. In spite of
the oxygen reduction) on the modied electrode its complexity, electrostatic and hydrophobic/
has been investigated and applied to the fabrica- hydrophilic interactions that involve both mem-
tion of biocathode for enzymatic biofuel cell brane and fouling materials are recognized to have
(Servat et al. 2007). signicant bearing. This is especially for more-
difcult-to-clean membrane fouling dominated by
natural organic matter (NOM) and microbes.
References Membrane fouling is referred to the ux decline
of a membrane lter caused by the accumulation
Servat K, Tingry S, Brunel L, Querelle S, Cretin M, of certain constituents in the feed on the surface of
Innocent C, Jolivalt C, Rolland M (2007) Modication
the membrane or in membrane matrix. Certain
of porous carbon tubes with enzymes: application for
biofuel cells. J Appl Electrochem 37:121127 fouling materials can be removed hydraulically
(backwash or scrubbing); however, most can be
removed by chemical means such as clean-in-
place (CIP) or chemical cleaning.
Chemical Cleaning of Membranes The technique of investigating a fouled
membrane membrane autopsy identies the
C. V. Vedavyasan cause of poor membrane performance and hence
Department of Technology, Annamalai gives the opportunity to rectify or mitigate the
University, Chidambaram, Tamil Nadu, India problem and improve future plant design.
Introduction Membrane Fouling and Chemical

Cleaning
Cleaning of membranes is carried out to enhance
its permeability followed by performance in a Chemical cleaning is an integral part of membrane
membrane-based treatment plant. The Pareto prin- process operation that has a marked effect on the
ciple 80/20 rule that 80 % of effects result from performance and economics of the process.
20 % of the causes best describes the discussion, A decline in RO membrane performance can be
where membrane fouling is the cause (the 20 %) caused by fouling (organic and inorganic fouling
and cleaning is the effect (the 80 %). and biofouling), operational mishaps, and
380 Chemical Cleaning of Membranes
membrane element quality. The behavior and fre- For difcult-to-remove foulants silicates and
quency of RO system fouling are affected by feed organics cleaning should be performed when the
water source and quality, pretreatment require- normalized permeate ow decline is 1015 %. In
ments, pretreatment chemicals, and system design many cases, commercial membrane cleaners are
and operation. During normal operation over a required but they increase OPEX.
period of time, RO membrane elements are sub-
ject to fouling by suspended or sparingly soluble Inorganic Fouling/Scaling
materials in the feed water. According to the type Inorganic fouling or scaling is caused by the accu-
of foulants, ve categories of membrane fouling mulation of inorganic precipitates such as
are presented. metal hydroxides and scales on membrane sur-
face or within pore structure. Precipitates are
Inorganic fouling/scaling formed when the concentration of chemical spe-
Particle/colloid fouling cies exceeds their saturation limits. Scaling is a
Biological fouling (bacterial bioslime, algae, major concern for reverse osmosis (RO) and
mold, or fungi) nanoltration (NF) membrane processes. RO and
Organic fouling NF membranes reject inorganic species. These
Sulfate scale of calcium, barium, or strontium species, if allowed to build up, form a concen-
trated layer of concentration polarization. For
Generally RO membrane suppliers recom- microltration (MF) and ultraltration (UF), inor-
mend membrane cleaning when: ganic fouling due to concentration polarization is
much less; however, they can exist due to interac-
Normalized permeate ow declines by 10 % tions between ions and foulants (organic poly-
Pressure drop declines by 10 % mers) via chemical bonding. Coagulation and
Normalized salt passage increases by 10 % oxidation, if not designed or operated properly,
might introduce metal hydroxides on membrane
Well-designed, well-operated RO systems are surface or within pore structure. Inorganic fouling
usually cleaned one to four times per year. Irre- and scaling can be a signicant problem for
versible membrane element performance decline makeup water of caustic solution used for chem-
such as membrane compaction and intrusion conical cleaning.
tributes to the 0.7 ow factor. Based on this per-
meate ow prole, the systems stabilized Particulate/Colloidal Fouling
performance is a ow factor of 0.7. Cleaning Accumulation of particulate material (dirt, sand)
criteria should be based on actual stabilized mem- on the membrane surface plugging leads to
brane element performance. The typical cleaning particulate fouling. Algae, bacteria, and certain
criteria (10 % normalized permeate ow decline natural organic matters fall into this category.
or 0.9 ow factor) are not applicable, because true However, they are different from inert particles
membrane element performance is a permeate and colloids such as silts and clays. In most cases,
ow factor of 0.7. Cleaning performed at a 0.7 particles and colloids do not really foul the mem-
ow factor has no effect but increases the OPEX brane because the ux decline is largely reversible
(operating expenses). by hydraulic cleaning measures such as backwash
Therefore, RO system cleaning criteria must be and air scrubbing.
customized to actual stabilized element perfor-
mance. The criteria can be customized to site con- Microbial/Biological Fouling
ditions. For instance, calcium carbonate is a foulant Microbial fouling is a result of formation of
that can be easily and effectively removed with a biolms on the membrane surface. Once bacteria
hydrochloric acid solution at pH 2. For sites with get attached to the membrane, they start to multi-
low energy costs, cleaning may be carried out when ply and produce extracellular polymeric sub-
the normalized permeate ow decline is 20 %. stances (EPS) to form a viscous, slimy, hydrated
gel. The severity of microbial fouling is intrinsi- Figure 1 is a photograph of fouled reverse
cally related to the quality of feed water like osmosis membranes. Deposits of silt are com-
environmental conditions for microbial growth, posed of suspended particulates of all types that
availability of nutrients, and abundance of accumulate on the membrane surface. Sources of
microbes. Organic-based deposits resulting from silt are organic colloids, iron corrosion products,
bacterial slimes, fungi, molds, etc. can be difcult precipitated iron hydroxide, algae, and ne partic-
to remove, if the feed is plugged. Plugging of the ular matter. Silt density index(SDI) testing is a C
feed makes it difcult to introduce and distribute widely accepted method for estimating the rate
the cleaning solutions. To inhibit additional at which colloidal and particle fouling will
growth, it is important to clean and sanitize not occur in water purication systems, especially
only the RO system but also the pretreatment, using reverse osmosis (RO) or nanoltration
piping, dead legs. (NF) membranes.
Bacteria present a signicant fouling problem Shown in Fig. 2 are scanning electron micro-
in most of the RO plants. As they grow and graphs of the surface of the RO membrane after
multiply, they normally produce biolm, organic 4 and 16 days. (A) Rod-shaped gels embedded in
lms of lipopolysaccharides, to encase and protect an extracellular brillar material structure (square 1).
the bacteria. These biolms cover the membrane Compact aggregates are visible on top of this biolm
surface resulting in fouling. (square 2). The RO membrane surface is visible at
the bottom (under the biolm layer) as a rough-
appearing texture. (B) Typical microcolony formed
on the surface of the membrane after 16 days (Fig. 3).
Shown above are series of photographs of
fouled reverse osmosis membranes (A) and their
feed-side spacers (B). The membranes and
spacers were removed from the ow cells after
4 (column 4 d), 8 (column 8 d), 16 (column 16 d),
and 32 (column 32 d) days of operation. The
direction of the feed water ow along the length
of each ow cell was from left to right.
Organic Fouling
Organic fouling is signicant in membrane
Chemical Cleaning of Membranes, Fig. 1 Fouled RO ltration with source water containing relatively
membrane high natural organic matters (NOM). Surface
Chemical Cleaning of Membranes, Fig. 2 SEM imaging of biolms

4d 8d 16d 32d
a
Feed flow
Chemical Cleaning of Membranes, Fig. 3 Fouled membranes
Chemical Cleaning of Membranes, Table 1 Effects of operating strategies on fouling

Type of fouling Hydraulic cleaning Feed chlorination Feed acidication Chemical cleaning
Inorganic ++ ++
Particulate ++ ++
Microbial + ++ +a ++
Organic + ++
No effect or negative effects
+Some positive effects
++Always positive effect
a
In conjunction with feed chlorination
water (lake, river) typically contains higher Chemicals for silica inhibition
NOM than groundwaters, with exceptions. For Chemicals for silica dispersion (Neofotistou
source water high in NOM, organic fouling is the and Demadis 2004)
most signicant factor in ux decline. NOM typ- Chemicals for silica scale removal from mem-
ically contains about 50 % humic substances. brane surfaces
Some organic molecules (oils, greases, humic/
fulvic acids, surfactants), occurring either natu- Silica inhibitors can retard the polymerization
rally or synthetic, coat and plug the membrane of silica, which exceeded its solubility, i.e.,
pores. phosphonate-based chemicals. Dispersants can
The effects of various operating strategies work by placing a surface charge on the polymer-
against different types of fouling are summarized ized silica and cause repulsion and dispersion of
in Table 1. As indicated, chemical cleaning is the these polymerized structures into water phase
most effective strategy for all types of membrane (Dudley and Baker 1999).
fouling. Polymerization of dissolved silica (i.e., super-
saturation and catalyzed by hardness) can be
difcult to remove. This type of silica fouling is
Chemical Cleaning of Silica Fouling different from colloidal and particulate silica,
which may be associated with either metal
Chemicals for silica fouling control can be hydroxides or organic matter. It might be dif-
divided into three categories: cult to remove polymerized silica by typical
chemical cleaning process. However, polymer- Chemical Cleaning of Membranes, Table 2 Catego-
ized silica scales removed with a high pH ries of membrane cleaning chemicals
cleaning solution (pH of 1011) were reported. Category Functions Typical chemicals
Caustic (sodium hydroxide) at maximum pH Caustic Hydrolysis, NaOH
allowed by the membrane manufacturer will solubilization
remove polymerized silica scales but it will Oxidants/ Oxidation, NaOCl, H2O2,
take many hours to remove a silica scale. Colloi-
disinfectants disinfection peroxyacetic acid
C
Acids Solubilization Citric acid, nitric
dal silica and particulate silica coating, not asso- acid
ciated with either metal hydroxides or organic Chelating Chelation Citric acid, EDTA
matter, can be removed from membrane surface agents
by physical ushing. Surfactants Emulsifying, Surfactants,
MF and UF membranes can be of organic poly- dispersion, detergents (SDS,
surface sodium dodecyl
mers or inorganic materials, whereas NF and RO conditioning sulfate)
membranes are all organic polymers only.
Low pH CIP (clean-in-place)

Membrane Cleaning Chemicals Acid ushing with permeate and nonoxidizing
biocide
Selection and Use of Cleaning Chemicals
The rst time cleaning is performed, it is Once the cause of membrane fouling is identi-
recommended to contact the manufacturer of the ed, various cleaning chemicals can be used to
equipment, the RO element manufacturer, and remove foulants from the membrane and to
service supplier. The cleaning chemicals can be restore the membrane ux. Chemicals commonly
generic or proprietary. used for cleaning MF and UF membranes in water
It is not unusual to use a number of different industry fall into ve categories, as shown in
cleaning chemicals in a specic sequence to Table 2.
achieve optimum cleaning. A high pH cleaning
is used to remove foulants like mineral scale or Caustic
metal hydroxides/oxides. The order of high and Caustic is used to clean membranes fouled by
low pH cleaning solution could be reversed or organic and microbial foulants. The function of
only one solution is required under the situation. caustic is twofold: (1) hydrolysis and (2) solubili-
Detergents are added to aid in the removal of zation. There are a number of organic materials
heavy biological and organic matter, while other including polysaccharides, and proteins that can
foulants need a chelating agent like EDTA for the be hydrolyzed by caustic. Cellulose-based mem-
removal of colloidal material, organic and biolog- branes based on polysaccharides hydrolyze;
ical material, and sulfate scale. An improper selec- hence, such membranes are to be used in a limited
tion of cleaning chemical or the protocol thereof pH range. This is due to the hydrolysis of
can make the entire cleaning exercise worse. cellulose a simple polysaccharide consisting
Iron deposits and scaling require a simple low of thousands of glucose linked by 1,4-b-glucoside
pH cleaning. For most complex fouling, the bonds . Fats and oils also react with caustic
sequence below may be followed: through saponication, generating water-soluble
soap micelles.
Flushing with permeate and with nonoxidizing An important function of caustic is to increase
biocide at the end of ushing negative charges of humic substances. Hence,
High pH CIP (clean-in-place) with temperature they are easier to be removed from membranes.
versus pH Humic substances contain functional groups that
Flushing with permeate till pH on the brine are organic acids. Their acidity ranges from pKa
side <8.5 (negative logarithm of the dissociation constant)
of 1.2 (oxalic) to 13 (phenol), with an average pKa compatibility of membrane and lter components
of 4.2 (Thurman 1985). During caustic cleaning, limits the type and the maximum allowable con-
pH of cleaning solution can be as high as 13. centration of a chemical to be used during cleaning.
Concentration of cleaning chemicals can affect
Acids and Chelating both the equilibrium and the rate of reaction. The
Acids are used primarily for removing scales and concentration prole of cleaning chemicals within
metal dioxides from fouling layers. When mem- the fouling layer is a function of the concentration
brane is fouled by iron oxides, citric acid is quite of cleaning chemicals in the bulk liquid phase.
effective because it not only dissolves iron oxide Therefore, the concentration of cleaning chemicals
precipitates but also forms complex with iron. The not only need to maintain the reasonable reaction
removal of divalent cations by either acids or rate (kinetics need) but also need to overcome mass
chelating reagent such as EDTA improves the transfer barrier imposed by the fouling layer. In
cleaning of membranes fouled by organic practice, the concentrations of cleaning chemicals
foulants. are usually high enough to satisfy the kinetic need.
It is the mass transfer that sets the lower boundary
Surfactants for the concentration of cleaning chemicals.
Surfactants are compounds that have both hydro- Temperature can affect membrane cleaning by:
philic and hydrophobic structures. They can form
micelles with fat, oil, and proteins in water and Changing the equilibrium of a chemical
help to clean the membranes fouled by these reaction
materials. One interesting but less clear aspect is Changing the reaction kinetics
how surfactants affect membrane fouling domi- Changing the solubility of fouling materials
nated by NOM (natural organic matter). As the and/or reaction products during the cleaning
surfactant is nonionic, the interaction between the
membrane and the surfactant is dominated by From mass transfer point of view, dynamic
hydrophilic and hydrophobic reaction. Since the cleaning involving circulating cleaning solutions
membrane is hydrophobic, hydrophilic tail of the through the system can be more effective than
surfactant is preferably adhered to the membrane simply static cleaning such as soaking.
surface and hydrophilic head is orientated toward
aquatic phase. Soaking in surfactant has been used
as a method for surface modication to improve RO Membrane Element Cleaning
hydrophilicity or wettability of certain membranes. and Flushing Procedures
The RO membrane elements can be cleaned in

Operating Parameters Affecting place in the pressure tubes by recirculating the
Cleaning Efficiency During Chemical cleaning solution across the high-pressure side of
Cleaning the membrane at low pressure and high ow.
A cleaning unit is needed to do this. RO cleaning
Because membrane cleaning is essentially procedures may vary depending on the situation.
conducted through chemical reactions between The time required to clean a stage can range from
cleaning chemicals and fouling materials, factors 4 to 8 h.
that affect the mass transfer and chemical reac- A general procedure for cleaning the RO mem-
tions such as concentration, temperature, length of brane elements is:
cleaning period, and hydrodynamic conditions
affect the efciency of cleaning. Perform a low-pressure ush at 60 psi (4 bar)
When selecting cleaning conditions, ensure the by pumping clean water from the cleaning tank
compatibility of cleaning chemicals with mem- through the pressure tubes to drain for several
brane media and lters and systems. Chemical minutes. Flush water should be clean water of
RO permeate or DI (deionized) quality water for 1560 min. A conductivity meter can be
and be free of hardness, transition metals, and used to test for removal of cleaning chemicals,
chlorine. such that the ush water to drain is within
Mix a fresh batch of the selected cleaning 1020 % of the feed conductivity. A pH
solution in the cleaning tank. The dilution meter can also be used to compare the ush
water should be clean water of RO permeate water to drain to the feed pH.
or DI quality and free of hardness, transition Once all the stages of a train are cleaned, and C
metals, and chlorine. the chemicals ushed out, the RO plant can be
Circulate the cleaning solution through the pres- restarted and placed into a service rinse. The
sure tubes for about 1 h or as required. At the RO permeate should be diverted to drain until
start, send the displaced water to drain so you do it meets the required quality of the process. It
not dilute the cleaning chemical and then divert normally takes few hours to few days for the
up to 20 % of the most highly fouled cleaning RO permeate quality to stabilize, especially
solution to drain before returning the cleaning after high pH cleanings.
solution back to the RO cleaning tank. For the
rst 5 min, slowly throttle the ow rate to one
third of the maximum design ow rate. This is Conclusion
to minimize the potential plugging of the feed
path with a large amount of dislodged foulant. As the water treatment industry grows at a remark-
For the second 5 min, increase the ow rate to able pace, the ability to treat aggressive surface
two third of the maximum design ow rate and and waste waters that are notorious for having
then increase the ow to two third of the max- high organics, colloids, and biological fouling
imum design ow. potential continues to be challenging.
An optional soak and recirculation sequence Using a generic or commercial alkaline
can be used, if necessary. The soak time can cleaning chemical signicantly affects OPEX,
be from 1 to 8 h depending on the manufac- while the payback primarily depends on the
turers recommendations but not to exceed chemicals used. Cleaning to reduce a facility
maximum pH and temperature limits for spe- OPEX depends on water temperature, energy
cic elements. costs, and chemical costs. Establish a cleaning
Upon completion of the chemical cleaning strategy that is most effective and well tested;
steps, a low-pressure cleaning rinse with clean analyze the foulants followed by an autopsy
water (RO permeate or DI quality and free of protocol. Selection of an appropriate pretreatment
hardness, transition metals, and chlorine) is with good operation and maintenance
required to remove all traces of chemical from practices will lead to cost-effective cleaning
the cleaning skid and the RO skid. Drain and regime. The cleaning efciency depends on the
ush the cleaning tank; then completely rell type of cleaning agent and its concentration. The
the cleaning tank with clean water for the net effect is that the efciency increases with
cleaning rinse. Rinse the pressure tubes by increasing chemical concentration. Operating
pumping all of the rinse water from the cleaning parameters such as cross-ow velocity,
tank through the pressure tubes to drain. turbulence in the vicinity of membrane surface,
Once the RO system is fully rinsed of cleaning temperature, pH, and duration of cleaning affect
chemical with clean water from the cleaning the cleaning process. Analyze the right cleaning
tank, a nal low-pressure cleanup ush can be strategy on a case-by-case basis. It is
performed with pretreated feed. The permeate recommended that the MSDS (Material Safety
line should be open to drain. Feed pressure Data Sheet) of the cleaning chemicals be procured
should be less than 60 psi (4 bar). This nal from the chemical supplier and knowledge be
ush continues till the ush water is clean and applied in the handling and storage of the
is free of any foam or residues. This ush goes chemicals.
386 Chemical Industry and Membrane Operations
Cross-References materials, and energy utilization, as well as

waste generation, in the process intensication
Cleaning Cycle of Fouled Membranes logic (Stankiewicz and Moulijn 2000). Membrane
Cleaning Effectiveness processes have several advantages than conven-
Cleaning Efciency tional separation techniques (e.g., distillation,
Fouling Index extraction, absorption, and adsorption), including
Fouling in Membranes modularity and compactness, operational exibil-
Fouling Release ity, no need for energy-intensive phase changes,
Fouling Release Membranes or potentially expensive adsorbents and/or
Natural Organic Matter, Removal and Fouling of difcult-to-handle solvents.
The separation of hydrogen from nitrogen in
ammonia plants was the rst large commercial
Further Readings
application of membrane-based gas separation
Baker, J.S; Dudley, L.Y (1999) Biofouling in membrane (Bernardo et al. 2009). Compact and modular
systems: a review. Water supply 17(1):17381 membrane modules, mainly in the form of hollow
Hijos G, Artal J, Gogeascoa IA, Bidaurrazaga JA bers, are widely used for nitrogen production to
(2012) Adelaide desalination plant process design: inert fuel tanks, also aboard aircrafts, and for
energy use optimization. IDA J 4(2):5256
Johnson J, Mercer G (2012) Seawater desalination plant blanketing chemical, methanol, and liqueed gas
optimization based on membrane replacement and shipments to enrich oxygen for medical uses, for
cleaning. Desalination Water Reuse 4(2):3438 hydrogen recovery and purication, for air and
Liu C, Caothein S, Hayes J, Caothuy T (2012) Membrane gas dehydration, and for ratio adjustment of gas
chemical cleaning: from art to science. http://www.
membranecleaning.com/pdf mixtures.
Nghiem LD, Schaefer AI (2006) Fouling autopsy of Membranes represent a valid alternative to
hollow-bre MF membranes in wastewater reclama- chemical treatment in the wastewater treatment,
tion. Desalination 188:113121 producing a cleaner and more reliable efuent for
Technical Services Bulletin 107.20. Selection and use of
cleaning chemicals. Resource document. Hydranautics. discharge or recycle, reducing, at the same time,
http://www.membranes.com.pdf. Accessed 6 July 2010. chemical, analytical, and labor costs. Membrane
technology has a signicant role in the production
of different grades of process water and in a recy-
cle/reuse logic. Depending on the nature and var-
iability of the feed supply, ultraltration or
Chemical Industry and Membrane microltration membranes can effectively per-
Operations form the pretreatment for reverse osmosis, remov-
ing bacteria and solid particles (Lefebvre and
Paola Bernardo Moletta 2006). Membrane bioreactors, by com-
Institute on Membrane Technology, National bining the traditional biological degradation
Research Council, ITM-CNR, Rende, Italy with a solidliquid separation by means of
low-pressure membrane ltration (ultraltration
or microltration), are appropriate solutions in
Chemical industry is a complex of processes, the industrial wastewater treatment to reduce
involved in the manufacture of chemicals and plant footprint and operating costs (Marrot
their derivatives. Generally, a production cycle et al. 2004).
comprises the conversion of raw materials and Filtration membranes are variously used for
the product recovery and purication. Separation concentration and/or recovery purposes in differ-
processes are among the most energy-intensive ent industrial cycles: in the production of pig-
stages. Membrane technology is widely accepted ments, surfactants, electronics, textiles, pulp, and
and well established in the chemical industry, paper, in the industrial laundering, and in the
allowing to decrease equipment size, raw chemical and mineral processing. Acid- and
Chemical Looping 387
base-stable membranes, based on ceramic mate- Koltuniewicz A (2010) 4.05 integrated membrane oper-
rials, are suited for the concentration and demin- ations in various industrial sectors. In: Drioli E, Giorno
L (eds) Comprehensive membrane science and engi-
eralization of dyes in acids; the decolorization, neering. Elsevier, Oxford, pp 109164. ISBN
clarication, or demineralization of spent acid 9780080932507
and caustic solutions for recycle; and the concen- Lefebvre O, Moletta R (2006) Treatment of organic pollu-
tration, desalination, and softening of in-process tion in industrial saline wastewater: a literature review.
and waste streams and for COD and BOD reduc-
Water Res 40:36713682
Marrot B, Barrios-Martinez A, Moulin P, Roche N (2004)
C
tion in waste streams such as pulp and paper waste Industrial wastewater treatment in a membrane biore-
efuents. actor: a review. Environ Prog 23(1):5968
Pervaporation allows organic dewatering, Shirasaki Y, Tsuneki T, Ota Y, Yasuda I, Tachibana S,
Nakajima H, Kobayashi K (2009) Development of
especially from ethanol and isopropanol (Van membrane reformer system for highly efcient hydro-
der Bruggen 2009). The development of solvent- gen production from natural gas. Int J Hydrog Energy
stable membranes has permitted the progress of 34(10):44824487
organic solvent nanoltration which is applied for Stankiewicz A, Moulijn JA (2000) Process intensication:
transforming chemical engineering. Chem Eng Prog
nonthermal solvent recovery; in situ recycling of 96(1):2234
organic solvents; purication, and concentration Van der Bruggen B (2009) Fundamentals of membrane
of intermediates; product upgrading; and contin- solvent separation and pervaporation. In: Drioli E,
uous recovery of homogeneous catalysts (White Giorno L (eds) Membrane operations: innovative sep-
arations and transformations. Wiley-VCH Verlag
2006). Moreover, signicant progresses have GmbH & Co. KGaA, Weinheim. doi:10.1002/
been done in the recent years in developing other 9783527626779.ch3
unit operations such as membrane reactors White LS (2006) Development of large-scale applications
(Shirasaki et al. 2009) and membrane contactors in organic solvent nanoltration and pervaporation for
chemical and rening processes. J Membr Sci
(Klaassen et al. 2005), covering all the chemical 286:2635
engineering unit operations. Membrane
contactors, accepted into the beverage industry
for carbonation, deaeration, and nitrogenation,
are effectively used in degassing applications
and corrosion control. Chemical Looping
The range of application of the membrane
technology could be extended with the availabil- Fausto Gallucci
ity of new membrane materials able to withstand Chemical Process Intensication, Department of
more demanding process conditions of tempera- Chemical Engineering and Chemistry, Eindhoven
ture, pressure, and aggressive conditions. Other University of Technology, Eindhoven,
interesting options are related to hybrid systems in The Netherlands
which membrane technology is combined with
another separation process, improving overall
process efciency and offering the opportunity Among different strategies for CO2 capture, a new
of developing, reengineering, or retrotting chem- emergent technology is focusing the attention:
ical processes (Koltuniewicz 2010). chemical-looping combustion (CLC). This novel
concept could reduce up to 50 % of the costs of
CO2 capture in regard to current technologies. For
References that reason, the IPCC remarked CLC as one of the
cheapest technologies for CO2 capture. It was in
Bernardo P, Drioli E, Golemme G (2009) Membrane gas 1987 when Ishida et al. introduced by the rst time
separation: a review/state of the art. Ind Eng Chem Res the term chemical-looping combustion for power
48(10):46384663
Klaassen R, Feron PHM, Jansen AE (2005) Membrane
generation and later as a powerful technology for
contactors in industrial applications. Chem Eng Res CO2 capture (Ishida and Jin 1994; Ishida
Des 83(A3):234246 et al. 1987). The technology is referred to as a
388 Chemical Membrane Reactors
cycling system where a solid with high oxygen much higher hydrogen recovery can be obtained
capacity called oxygen carrier transfers the oxygen at lower temperatures.
needed to convert the fuel. By this way, CO2 pro-
duced by the combustion of the fuel is not diluted
in the N2 (from the air) because direct contact
between fuel and air is avoided. Hence a CO2-
References
rich stream is produced at the exit point of the
Adanez J, Abad A, Garcia-Labiano F, Gayan P, de Diego
reactor. It implies lower energy requirements for LF (2012) Progress in chemical-looping combustion
CO2 capture than other strategies where CO2 and reforming technologies. Prog Energy Combust
and N2 are mixed. The main drawback of this Sci 38:215282
Ishida M, Jin H (1994) A new advanced power-generation
technology is the lack of maturity, and for that
system using chemical-looping combustion. Energy
reason the interest in the system has increased 19:415422
during the last years as can be discussed in the Ishida M, Zheng D, Akehata T (1987) Evaluation of a
recent review by Adnez et al. (2012). More than chemical-looping-combustion power-generation sys-
tem by graphic exergy analysis. Energy 12:147154
700 different oxygen carriers have been already
Rydn M, Lyngfelt A (2006) Using steam reforming to
tested, and the performance of the concept has produce hydrogen with carbon dioxide capture by
been tested in different CLC units in the size chemical-looping combustion. Int J Hydrog Energy
range 0.3120 kWth. 31:12711283
Rydn M, Lyngfelt A, Mattisson T (2008a) Chemical-
Chemical looping systems have not been
looping combustion and chemical-looping reforming
exclusively focused on fuel combustion for in a circulating uidized-bed reactor using Ni-based
power generation but also for fuel reforming oxygen carriers. Energy Fuel 22:25852597
(mainly methane) for syngas production. This Rydn M, Lyngfelt A, Mattisson T, Chen D, Holmen A,
Bjrgum E (2008b) Novel oxygen-carrier materials for
concept is called chemical-looping reforming
chemical-looping combustion and chemical-looping
(CLR). The CLR occurs if at the fuel reactor the reforming; LaxSr1xFeyCo1yO3d perovskites and
CH4 is fed with steam and with a lower amount of mixed-metal oxides of NiO, Fe2O3 and Mn3O4. Int
oxygen (compared to the CLC) contained in the J Greenhouse Gas Control 2:2136
oxygen carriers. In this way, the fuel conversion
gives a reformed syngas with high H2 concentra-
tion. The system behaves like an autothermal
reformer, and the heat for the endothermic reac- Chemical Membrane Reactors
tion is directly supplied by burning part of the
methane. It has been operated for short times in Lidietta Giorno
lab-scale systems at moderate pressures (Rydn Institute on Membrane Technology, National
and Lyngfelt 2006; Rydn et al. 2008a, b). In total Research Council, ITM-CNR, Rende, Italy
around 100 h operation of CLR have been dem-
onstrated worldwide (Adanez et al. 2012). The
main research in literature has been focused on Chemical membrane reactors are membrane reac-
oxygen carrier studies for reforming of natural tors using chemical catalysts, mainly of inorganic
gas. The typical reactor scheme of CLR is similar origin.
to the well-known CLC but differs in the desired Both dense and porous membranes are used in
product, which in CLC is heat and in CLR membrane reactors. Dense membranes are made
is syngas. Recently, Gallucci et al. proposed of polymers, metals and their alloys, and solid
the application of a chemical-looping reforming oxides. Dense metal membranes mostly consist
where the fuel reactor is replaced by a membrane- of Pd, Pt, Ru, Rh, Ag, and their alloys. Pd mem-
assisted uidized bed reactor. This new concept branes and its alloys are used for dehydrogenation
named membrane-assisted chemical-looping reaction. They have high selectivity toward
reforming (MA-CLR) gives many advantages hydrogen, while Ag membranes are selective
over the previous concepts especially because a toward oxygen.
Chemical Membrane Reactors 389
re
porous inert film
(region c) rs
mesoporous film rm
catalitically active
ri
(region b) C
tube side
(region a)
shell side filled with catalyst

(region d )
Chemical Membrane Reactors, Fig. 1 Packed bed membrane reactor with the membrane formed of a mesoporous lm
catalytically active and a porous inert support layer
Solid oxides (ZrO2, BiO3, Y2O3) and solutions In the tube side (region a):
of mixed solid oxides (such as perovskite-type
oxides) have been used as dense membranes
@Cai 1 @ @Ca
selective for oxygen or hydrogen. They are used Vf Dai r i rb ai va (1)
@z r @r @r
in catalytic membrane reactors for partial and total
oxidation reactions.
Mesoporous membranes made of alumina, tita- with 0 r ra and 0 z L:
nia, and zirconia are also used to prepare mem- Vf is the supercial uid velocity (m s1); Cia
brane reactors. Microporous membranes made of is the concentration of species i in region 1; Dia is
zeolites have been also prepared. the effective radial diffusivity through the packed
As a general example of modeling of chemical bed in the tube; rb is the reactor bed density; ai is
membrane reactors, here is reported the case of a the stoichiometric coefcient; and va is the appar-
packed bed membrane reactor with the membrane ent reaction rate (mol kg1 s1).
formed of a mesoporous lm catalytically active In the reactive membrane (region b):
and a porous support layer inert (Fig. 1). The tube
and shell sides are occupied by packed bed
1 @ @Cbi
catalysts. Dbi r rm a i v b (2)
r @r @r
The reaction occurs in gas phase. The gas
transport through the mesoporous membrane
layer takes place by Knudsen diffusion. Surface ri r rm
ow and diffusion through the porous region are
considered to be negligible. The reaction could rm is the membrane density; Dbi is the membrane
take place inside the catalytically active mem- diffusivity; and vb is the observed reaction rate.
brane and/or in the packed beds placed in the In the macroporous support (region c):
tube- and/or shell-side compartments. Based on
these assumptions, the model equations in cylin- 1 @ @Cc
Dci r i 0 (3)
drical coordinates and the corresponding bound- r @r @r
ary conditions for each region showed in the
Fig. 1 are described as follows: rm r rs
390 Chemical Pretreatment
In the shell region (region d): porosity and the tortuosity of the membrane), Dik
is the corresponding effective Knudsen diffusion
coefcient, and x is the molar fraction. B0 corre-
@Cdi 1 @ @Cd
Ud Ddi r i r b ai v d (4) sponds to the permeability coefcient for convec-
@z r @r @r
tive ow in m2, and m is the viscosity in Ns/m2.
Dij, Dik, and B0 are empirical parameters, which
rs r re depend on the structure of the porous membrane.
0zL
Catalytic Membrane Reactors

with Nonselective Membrane (CNMR) Chemical Pretreatment
In these types of membrane reactors, the mem- Graeme K. Pearce

brane may act as a contactor device (catalytic) or Membrane Consultancy Associates Limited,
may serve to distribute reactants in an uniform Reading, RG8 7EJ, UK
way (inert). In modeling such reactors, Knudsen
diffusion as well as molecular diffusion and con-
vective transport may occur. The dusty gas model Definitions
of transport is typically utilized. Membrane reac-
tors using mesoporous membranes and operating The term chemical pretreatment refers to the use
under a transmembrane pressure gradient may of coagulant chemicals to form occulant (oc)
also use this model. For such reactors, an isother- particles in a water-based feed stream to enhance
mal model that consists of mass balance equations contaminant removal.
in every membrane compartment can be used. For
example, for compartments where reaction
occurs, the mass balance is Objectives
The purpose of chemical pretreatment is to entrap

J i V ri (5)
particulates and to some extent dissolved contam-
inants, for subsequent removal by gravity settle-
while in nonreactive regions ment or otation, or by removal in a physical
ltration step (see granular media pretreatment
J i 0 (6) and physical pretreatment).
where Ji is the ux of species i and Vri is the Process Principles and Mechanisms
corresponding reaction rate.According to the
dusty gas model, Ji is described as: Water treatment relies on settlement to remove
large particles, followed by ltration to remove
Xn
Xi J i Xj J i Ji 1 ner ones. A wide variety of particulate and
xi dissolved contaminants can be found in water
j1, j6i
PD ij PD ik RT
(7) and wastewater sources (Parsons and Jefferson,
J i B0 P
1 Pi 1 . . . n 2005). Larger particles are held in suspension in
PRT mDik
a owing stream and settle rapidly if the feed is
stored. As particle size becomes ner, settling
where Dij is the effective binary diffusion coef- velocities reduce sharply and below a size of
cient of components i and j (already includes the 10 mm, rates become so low that it is not practical
Chemical Pretreatment 391
to remove these contaminants by settlement pro- treatment process. If derived from organic matter
cesses alone, as illustrated in Table 1. such as algae, they may cause fouling in down-
In a typical water source, ne particles are held stream processes such as RO, or could be respon-
in suspension in a stable dispersed state. There are sible for toxic residuals in treated water.
forces of gravity attraction between particles, but Coagulation is the process used to destabilize
for small particles, the forces are not sufcient to the dispersed suspended matter as illustrated in
lead to natural agglomeration, which would aid Fig. 1. The colloids tend to have a negative charge C
their removal by settlement. Indeed, forces of at the surface which generates a signicant repul-
electrostatic repulsion tend to prevent particles sion between particles. The coagulant neutralizes
coming together, particularly for submicron the surface charge and overcomes the repulsive
sizes, and the random forces of Brownian motion forces. In the gure, trivalent Al3+ ions are used to
contribute to maintaining colloids and macromol- full this role.
ecules in suspension. Hydrophilic molecules and ionic species have
Colloids can give rise to turbidity and affect strong bonds of hydration to water molecules
other aesthetic water quality parameters such as which assist their stable dissolution in water. In
color and so need to be removed in the water contrast, hydrophobic molecules do not bond to
water and are easy to separate from a bulk water
medium. However, naturally occurring dissolved
Chemical Pretreatment, Table 1 Settling velocities for organic molecules and colloids are often amphi-
common particles and colloids philic in nature having a portion of the molecule
Particle diameter which is hydrophilic (water loving) and a portion
mm Type Settling velocity which is oleophilic (oil loving). The oleophilic
10 Pebble 0.73 m/s portion is likely to also be hydrophobic (water
1 Coarse sand 0.23 m/s
hating), giving rise to the synonymous term
0.1 Fine sand 0.6 m/min
amphipathic.
0.01 (10 mm) Silt 8.6 m/day
Common contaminants in surface water
0.0001 (0.1 mm) Large colloids 0.3 m/year
include humic and fulvic acids which are respon-
0.000001 (1 nm) Small colloids 3 m/million year
sible for color and can give rise to disinfection
Chemical Pretreatment,
a
Fig. 1 Colloid Large
destabilization Repulsive
Force
b Al3+
Reduced Al3+
Al3+ Repulsive
Force Al3+ Al3+
Al3+ Al3+
Al3+
Al3+
Al3+ 3+
Al
Al3+ Al3+
Inorganic coagulants
Inorganic polymeric coagulants
Organic polymeric coagulants
The most common inorganic coagulants used in

water treatment are Al3+ salts added as aluminum
chloride or aluminum sulfate (alum) and Fe3+salts
added as ferric chloride or ferric sulphate (Bache
and Gregory, 2007). The coagulation reaction of
alum is shown below:
Al2 SO43 :14 H2 O 6HCO3 , 2AlOH3 #
Chemical Pretreatment, Fig. 2 Floc particle 6CO2 14H2 O 3SO4 2
The optimum pH for the hydrolysis reaction

shown is 5.56.5. However, note that the reaction
by-product residuals after chlorination. These
conditions are only satised for feed water with a
macromolecules are examples of colloids with
signicant level of bicarbonate alkalinity.
strong amphiphilic character. Coagulants assist
According to the equation above, 1 mol of alum
in the removal of these colloids by displacing
consumes 6 mol of bicarbonate, so in a soft, low
hydration water from the hydrophilic portion of
alkalinity water, there may not be sufcient natu-
the molecule and associating with the hydropho-
rally occurring calcium bicarbonate to support the
bic portion through charge attraction. The Al3+
required reaction. In this case, the reaction gener-
sites then attract hydration water and the partly
ates sulfuric acid, causing the pH to drop sharply,
neutralized particle agglomerates with other coag-
as shown in the following equation. Note that lime
ulated particles to create the rapid buildup of a oc
addition would be required to neutralize:
particle as shown in Fig. 2.
The initial dispersion phase takes between
10 and 120 s. Once the coagulation process has Al2 SO43 :14 H2 O , 2AlOH3 # 8H2 O
been initiated, the process of occulation can then 3H2 SO4 2
proceed, i.e., the agglomeration of the
destabilized particles. The growth phase can take Ferric chloride consumes half the molar concen-
up to 40 min, though in commercial processes, it tration of bicarbonate in its hydrolysis reaction
is normal to use considerably less time than this. since there is only one ion of Fe3+ for each ferric
The initial seed for the oc to form is normally a chloride molecule, whereas the sulfate in alum is
small particle or colloid, but as it grows, dissolved divalent. The hydrolysis equation is:
organics become associated with the oc due to
their charge or surface character and can therefore FeCl3 3HCO3 , FeOH3 # 3CO2 3Cl
be removed from the bulk liquid phase. Coagula-
tion can therefore be used to reduce total The optimum pH for Fe3+ is 4.55.5, but there is a
suspended solids and turbidity, and to some extent wider range of pH that can be used in different
dissolved organics, thereby providing a signi- circumstances of 49. Thus, Fe3+ is less pH sen-
cant impact for the improvement of water quality. sitive than Al3+ and normally provides better tur-
bidity, color, and natural organic matter (NOM)
removal at a given molar concentration. As with
Coagulant Chemical Categories Al3+, acidity will be created if the bicarbonate
concentration is insufcient, requiring lime
Dosing chemicals fall into three categories: addition.
Chemical Pretreatment 393
Due to the higher density of oc particles with advantage in reducing sludge disposal costs,
Fe3+, the sludge has better settleability and is often and this is likely to become an even more impor-
preferred for clarication applications. Fe3+ pro- tant factor in the future, outweighing chemical
duces a higher sludge mass. Al3+ can struggle in purchase cost considerations in some
cold water conditions, but is often preferred for circumstances.
dissolved air otation, since this process relies on
the low density of the oc particle. In some coun- C
tries Fe3+is preferred to Al3+ for health reasons. Coagulation Conditions
The basic coagulation chemistry of both inor-
ganic and organic polymeric coagulants is similar Establishing the correct conditions for coagula-
to the straightforward inorganic coagulants, but tion rst requires a selection of the correct coagu-
the pH tolerance may be wider and the effective- lant from the options described in the previous
ness of the coagulation and subsequent occula- section. The appropriate coagulant concentration
tion reactions may be improved. The most and pH will then need to be determined, as well as
common coagulant in the second category is poly- the correct physical conditions. These require-
meric aluminum chloride (PACl), which is used to ments may vary depending on the process selected
overcome problems associated with Al. PACl has for the formation and removal of the
better binding with clays, etc. due to higher charge oc. Choosing the right coagulant often requires
density. In general water treatment, poly- a review of comparable applications in terms of
aluminum coagulant hybrids can be used, but feed water type and temperature, with similar
these are not often encountered in UF process requirements. If there is no direct compar-
pretreatment, though aluminum chlorohydrate ison, several alternatives might need to be
(ACH) is found to be effective. Coagulants with explored.
a higher charge density can be more effective at The next step is to determine the best condi-
neutralizing the negative charge of colloids and tions for the coagulant chosen. This is normally
will tend to initiate agglomeration more rapidly, carried out by conducting jar tests. In these tests, a
growing stronger oc particles. number of beakers are set out with the feed
The third category of coagulant is the organic and coagulant at different concentrations and pH
polymeric category. The most common example conditions. A typical jar test set up is shown in
in this group is the cationic polymeric coagulant Fig. 3.
PDADMAC which has a high charge density giv- As well as varying coagulant concentration,
ing good removal of clays and organics. It can be another series of tests may be carried out at a
typically used at 10 % of the inorganic coagulant promising concentration with a variation of the
concentration and has low sludge yield. Some pH from say 5.07.5 in increments of 0.5 of a pH
cationics can bind to membrane surfaces, so UF unit between jars. The jars are mixed in a consis-
users are very cautious and the often chemically tent way with the addition of the coagulant and the
related category of oc aid is also used with great time noted for oc formation. A typical procedure
care due to the risk of attachment. Fortunately, would use 1 min of rapid mixing at 100150 rpm
oc aids are not normally needed in membrane to ensure quick dispersion of the coagulant
pretreatment. throughout the mixture, followed by a period of
In summary, both Fe and Al compounds are 14 min at a slow mixing speed of say 2530 rpm.
widely used in chemical pretreatment, but Fe is The emergence of oc is noted, and the turbidity
more often selected for use in membrane of the supernatant is measured once stirring has
pretreatment, utilizing low dosing concentrations. stopped. It is also instructive to observe oc
PACl is sometimes successfully employed, but settleability. A plot of residual turbidity against
high-performance organic coagulants are less pH as shown in Fig. 4 will indicate the optimum
often used due to cost and concerns about fouling. coagulation condition, which in the case of the
The low coagulant concentrations provide a key graph above is 6.3
Chemical Pretreatment,
Fig. 3 Jar test showing
different Al3+
concentrations with a
15 min contact time
Chemical Pretreatment, 20
Fig. 4 Residual turbidity
versus pH
Turbidity Remaining (TU)
15
10
5.0 5.5 6.0 6.5 7.0 7.5 8.0

pH
Process and Implementation The key issue for coagulant addition is to

ensure a rapid dispersion of the dosed chemical
In water and wastewater treatment, coagulation is after the injection point so that it is well dispersed
often required to improve the removal perfor- through the bulk liquid, followed by an opportu-
mance of conventional processes such as sand nity for the oc to grow in relatively low shear
lters. For a conventional process ow scheme, conditions. Rapid dispersion requires a good
the coagulation step is normally conducted prior mixing stage which can be achieved by a dosing
to ltration using a clarier or dissolved air ota- injection point immediately prior to a pump or
tion (DAF) unit. Membrane processes are less strainer, thereby using the pump impeller or
reliant on coagulation, but they are still used strainer media to achieve the mixing. Alterna-
quite frequently. Wastewater reuse applications tively an in-line mixer can be used in the pipe
always use coagulant and it is used in over half just after the injection point, or if in a tank, a
of surface water capacity and also quite frequently bafed owpath is created. It is important that
in seawater treatment. However, the use of coag- process shear is applied immediately after injec-
ulants in groundwater treatment is unusual. tion, i.e., within the rst few seconds. Once the
Chemical Vapor Deposition 395
oc has formed and starts to grow, high shear danger of coagulant fouling, potentially both of
operations may break up the oc, and it could be the UF or MF membrane and/or of any down-
difcult to get the oc to regrow. This can cause stream RO. This may occur due to overdosing or
high turbidity and fouling of downstream unit oc disruption from shear, either as a result of
operations. pumping or excessive transmembrane pressure
In general, coagulation in membrane (TMP). In seawater applications, the current
pretreatment ow schemes may be considered trend is to eliminate in-line dosing during good C
for the following reasons: feed quality periods and to use a low and variable
dose during periods of poor or variable quality.
Improve general feed quality to increase Different membranes respond differently to
UF/MF ux coagulants, and so selection of the coagulant
Ameliorate feed quality excursions such as procedure is often carried out through pilot trails
turbidity spikes and, in the case of seawater or comparison with other operations using simi-
pretreatment, algal bloom, and red tide events lar feeds.
Reduce dissolved organic concentration to
lessen biofouling
References
Operational Issues
Bache DH, Gregory R (2007) Flocs in water treatment.
IWA Publishing, London. ISBN 1843390639
It is important for a coagulant to produce a oc Parsons SA, Jefferson B (2005) Introduction to potable
which will be effectively removed in downstream water treatment processes. Blackwell publishing,
process operation. If the removal is to be achieved Oxford. ISBN-13: 978-1-4051-2796-7
by settlement in a clarier, the oc needs to be
large, dense, and fairly robust. For ltration
through a granular media lter, the oc needs to
be large and robust. Sometimes, a polymeric coag- Chemical Vapor Deposition
ulant or oc aid could be used as a oc strengthener
or enlarger to assist growth of the oc to improve Hiroki Nagasawa and Toshinori Tsuru
removal by settlement or ltration, and make it less Department of Chemical Engineering, Hiroshima
likely that the oc will be broken up through pro- University, Higashi-Hiroshima, Japan
cess shear. If it is to be removed in a dissolved air
otation unit, the oc needs to be low in density. If
the oc is formed by direct in-line injection just Chemical vapor deposition (CVD) is a method for
prior to a membrane ltration stage, it is likely that depositing thin lms of various materials (Smith
a microoc would be sufcient. 1995). CVD process involves a series of gas phase
Although clariers or DAF are sometimes and surface reactions of volatile precursors, which
used upstream of the membrane in some are induced by, for example, heat, light, and/or
pretreatment applications, a signicant propor- plasma-discharge. CVD has been used extensively
tion of membrane systems use in-line dosing at in a number of industries such as semiconductor,
low coagulant dose with a minimal contact time electronics, and optics industries.
at 2030 s. This has provided the opportunity for CVD techniques have also been utilized in the
membrane systems to cut coagulant doses radi- fabrication of inorganic molecular sieve mem-
cally from levels of several ppm used for clari- branes. Among the types of CVD technique, ther-
cation or DAF to 1 ppm or less (as Fe3+) for mal CVD has been most extensively studied since
in-line dosing. Gavalas and coworkers rst reported successful
The main concern in the use of coagulants synthesis of the hydrogen-selective silica mem-
with membrane pretreatment processes is the branes in 1989 (Gavalas et al. 1989). There are
396 Chemical Vapor Deposition
Chemical Vapor
Deposition,
of (a) one-sided and (b)
counter-diffusion CVD
processes
two types of reaction geometries available for example, Gu and Oyama prepared a silica mem-
thermal CVD, as shown in Fig. 1: one-sided brane on the substrate with a graded g-alumina
CVD and counter-diffusion CVD (Nomura intermediate layer by the thermal decomposition
et al. 1997; Nakao et al. 2000). For the case of of tetraethoxysilane (TEOS) at 600 C and
the one-sided CVD, all of the reactants are fed to reported H2 permeance of 5.0 107 mol/(m2 s
the same side of the porous substrate, and then, the Pa) at 600 C, with H2/CH4 selectivity of 5,900
lm-forming reactions occur at the surface of the (Gu and Oyama 2007). Ohta et al. prepared silica
substrate. On the other hand, for the case of the membranes on the g-alumina intermediate layer
counter-diffusion CVD, the reactants are fed to by using oxidation of tetramethoxysilane
the opposite side of the substrate. The reactants (TMOS), phenyltrimethoxysilane (PTMS), or
diffuse into the porous substrate and the reactions dimethoxydiphenylsilane (DMDPS) at 600 C
take place at the location where the reactants and reported that larger pores were formed when
contact each other. Therefore, the deposition the number of phenyl groups on the precursor was
occurs inside the pores and the pore size will be increased (Ohta et al. 2008). The DMDPS-derived
decreased. The deposition continues until the pore membrane showed a high H2 permeance of the
size becomes small enough to reduce diffusion of order of 106 mol/(m2 s Pa) at 300 C, with H2/
the reactants. As a result, uniform pore size can be SF6 selectivity of 6,800.
obtained by using the counter-diffusion CVD. Plasma-enhanced CVD (PECVD) is another
The hydrogen-selective silica membranes type of CVD technique utilized for the mem-
reported by Gavalas et al. were prepared by brane fabrication. PECVD offers a low
depositing a dense silica layer on the top of processing temperature compared with thermal
Vycor glass membranes by the oxidation of SH4 CVD due to the high reactivity of plasma. Sev-
at 450 C (Gavalas et al. 1989). The obtained eral research groups reported the PECVD-
membranes were reported to have high H2/N2 derived membranes prepared at room tempera-
selectivities as high as 3,000 but low H2 ture by using several kinds of precursors such as
permeances of 2.1 108 mol/(m2 s Pa) at hexamethyldisiloxane (HMDSO) (Roualdes
450 C. Until then, a great deal of progress has et al. 2002; Tsuru et al. 2011; Nagasawa
been made by several research groups. For et al. 2013), hexamethyldisilazane (HMDS)
Chemometrics 397
(Kafrouni et al. 2009), or bis(dimethylamino) CVD in the opposing reactants geometry. Microporous
dimethylsilane (BDMADMS) (Kafrouni Mesoporous Mater 37:145152
Nomura M, Yamaguchi T, Nakao S (1997) Silicalite mem-
et al. 2010). Typical PECVD-derived membranes modied by counterdiffusion CVD technique.
branes showed moderate selectivity (e.g., Ind Eng Chem Res 36:42174223
He/N2 selectivity of <10 at 25 C and ~30 at Ohta Y, Akamatsu K, Sugawara T, Nakao A, Miyoshi A,
150 C) with He permeance of the order of Nakao S (2008) Development of pore size-controlled
108107 mol/(m2 s Pa), probably because the
silica membranes for gas separation by chemical vapor
deposition. J Membr Sci 315:9399
C
PECVD-derived layer remains to have high con- Roualdes S, Sanchez J, Durand J (2002) Gas diffusion and
tent of organofunctional groups which make the sorption properties of polysiloxane membranes pre-
structure more loose (Roualdes et al. 2002; Tsuru pared by PECVD. J Membr Sci 198:299310
Smith DL (1995) Thin-Film Deposition: Principles and
et al. 2011; Nagasawa et al. 2013; Kafrouni Practice, McGraw-Hill, Inc.,
et al. 2009, 2010). Selectivity of PECVD- Tsuru T, Shigemoto H, Kanezashi M, Yoshioka T (2011)
derived membranes can be improved by 2-step 2-step plasma-enhanced CVD for low-temperature fab-
PECVD sequence, which involves HMDSO/Ar- rication of silica membranes with gas-separation per-
formance. Chem Commun 47:80708072
PECVD followed by HMDSO/O2-PECVD
(Tsuru et al. 2011; Nagasawa et al. 2013). More-
over, the membrane prepared via 2-step PECVD
showed a good thermal stability, although the
membrane was fabricated at room temperature. Chemometrics
The 2-step PECVD-derived membrane showed
high He and H2 permeances of 5.2 107 and Rocco Caliandro
2.4 107 mol/(m2 s Pa) at 400 C, respec- Institute of Crystallography, CNR, Bari, Italy
tively, with He/N2 and H2/N2 permeance ratios
of 4,200 and 1,900, respectively (Tsuru
et al. 2011; Nagasawa et al. 2013). Chemometrics can be dened as information
aspects of chemistry, where statistical and math-
ematical methods are used to produce good data
and to extract relevant information from measured
References data (Wold and Sjostrom 1998). The rst aim can
be achieved by using design of experiments (DoE)
Gavalas GR, Megiris CE, Nam CW (1989) Chem Eng Sci to provide a small number of information-rich
44:18291835 experiments. Multivariate data analysis and
Gu Y, Oyama ST (2007) High molecular permeance advanced data visualization can be employed for
in a poreless ceramic membrane. Adv Mater
19:16361640 the second purpose. In fact, multivariate projec-
Kafrouni W, Rouessac V, Julbe A, Durand J (2009) Syn- tion methods can be used to simplify complex
thesis of PECVD a-SiCxNy:H membranes as molecu- data and thus make the visualization easier. Fur-
lar sieves for small gas separation. J Membr Sci thermore they can be used for classication of
329:130137
Kafrouni W, Rouessac V, Julbe A, Durand J (2010) Syn- samples and to predict outcomes.
thesis and characterization of silicon carbonitride lms Theory of DoE and literature with references
by plasma enhanced chemical vapor deposition covering the eld of experimental design and
(PECVD) using bis(dimethylamino)dimethylsilane optimization can be found in Box et al. 1978;
(BDMADMS), as membrane for a small molecule gas
separation. Appl Surf Sci 257:11961203 Gabrielsson et al. 2002; Lundstedt et al. 1998;
Nagasawa H, Shigemoto H, Kanezashi M, Yoshioka T, Mandenius and Brundin 2008. Instrumentation
Tsuru T (2013) Characterization and gas permeation developed in the eld of process analytical chem-
properties of amorphous silica membranes prepared via istry (PAC) supplies data about the state of a
plasma enhanced chemical vapor deposition. J Membr
Sci 441:4553 process (Callis et al. 1987). Combination of
Nakao S, Suzuki T, Sugawara T, Tsuru T, Kimura S (2000) PAC instrumentation and multivariate analysis
Preparation of microporous membranes by TEOS/O3 provides tools for effective process monitoring
398 Chemometrics
60 1
Calculated concentration
Sample 0
CBZ I
50 0.8
CBZ-SAC
40
0.6
30 SAC
0.4
20
CBZ III 0.2
10
CBZ I
0 0
5 10 15 20 25 30 35 40 0 0.2 0.4 0.6 0.8 1
Measured concentration
Chemometrics, Fig. 1 X-ray diffraction proles and calibration plot of a PLS model for quaternary carbamazepine
saccharin mixtures
and control enabling detection of multivariate system, provide similar information content. Col-
relationships between different variables such linear variables can be combined and described by
as raw materials, process conditions, and end fewer, so-called factors or latent variables, which
products. describe the underlying structure in the data. In
DoE represents a special case of predictive modeling, the prime aim is to separate information
modeling. The objective of predictive modeling is from noise and nd the crucial patterns in the data.
to determine the relationship between several Following the introduction of computers and
x-variables (often called independent or explana- computerized measurement techniques, latent
tory variables) and one or more y-variables variable methodology has penetrated nearly all
(dependent or response variables). This objective areas where complex systems are measured and
can be achieved by means of a regression model, modeled, and it is especially powerful when huge
where the observed result, i.e., response (y), is amounts of data are produced and systematic
described as a function of the x-variables. Models approaches are needed to reveal the information
can be seen as tools to describe reality. Empirical in the data.
models based on the experimental data can be In the framework of membrane technology,
estimated and used for interpretation and predic- chemometrics can be used for process optimiza-
tion. All models are more or less erroneous, since tion, which includes quality control of compo-
there are always noise and other irrelevant features nents, and for characterizing the products of the
in the data. Experimental error is produced by both process. For example, in the process of membrane
known and unknown disturbing factors that may crystallization, chemometrics allows to nd the
confound important effects wholly or partially. best setting of parameters ruling the process
This can be reduced and sometimes almost elimi- (solvent ux, precipitant and solute concentra-
nated by using DoE and statistical analysis. tions, temperature, geometry, amount of solution),
Multivariate data analysis has proven to be a to control the quality of the membranes and plates
powerful tool when combined with advanced used, and to characterize the crystals produced.
characterization techniques. Apart from regres- Regression models allow to predict crystallization
sion models, it also uses latent variable methods, products given a set of (unexplored) parameter
such as principal component analysis (PCA) and setting. As an example, the X-ray diffraction pro-
partial least squares (PLS), for classication and les measured for mixtures formed by polymorph
prediction purposes. Measured variables, which III of carbamazepine, saccharin, and cocrystal
describe partially or fully the same property of a carbamazepine saccharin produced by
Chiral Drug Separation by Membranes 399
membrane crystallization have been processed by sales of single-enantiomer drugs have been grow-
PLS, which are able to supply precise quantitative ing (Maier et al. 2001; Li and Haynie 2007).
predictions on their concentrations through the There are some techniques to prepare single-
calibration plot shown in Fig. 1 (Caliandro enantiomer drugs in pharmaceutical industry. One
et al. 2013). is chemical synthesis using chiral building blocks
as starting materials. Another is asymmetric syn-
thesis using enantioselective catalysts like chiral C
organocatalysts or biocatalysts. Optical resolution
References of enantiomers in a racemate is another important
technique in chiral drug industry. The method of
Box GEP, Hunter WG, Hunter JS (1978) Statistics for
optical resolution can be applied to not only nal
experimenters. Wiley, New York
Caliandro R, Di Proo G, Nicolotti O (2013) Multivariate racemic drugs but also racemic intermediates
analysis of quaternary carbamazepinesaccharin mix- appeared in preparation steps.
tures by X-ray diffraction and infrared spectroscopy. We already have some techniques to separate
J Pharm Biomed Anal 7879:269279
one enantiomer from the other isomer: preferen-
Callis JB, Illman DL, Kowalski BR (1987) Process
analytical-chemistry. Anal Chem 59:A624 tial crystallization, chiral chromatography,
Gabrielsson J, Lindberg NO, Lundstedt T (2002) Multivariate enantioselective extraction, enzymatic kinetic res-
methods in pharmaceutical applications. J Chemometr olution, membrane separation, and so on. Among
16:141160
these techniques, membrane-based method has
Lundstedt T, Seifert E, Abramo L, Thelin B, Nystrom A,
Pettersen J, Bergman R (1998) Experimental design some advantages such as low energy demand,
and optimization. Chemom Intell Lab Syst 42:340 continuous operation, and easy scale-up. The
Mandenius CF, Brundin A (2008) Bioprocess optimization important factor of membrane-based
using design of experiments methodology. Biotechnol
enantioselective separation is the way how to
Prog 24:11911203
Wold S, Sjostrom M (1998) Chemometrics, present and give the function of chirality recognition to mem-
future success. Chemom Intell Lab Syst 44:314 branes. In a supported liquid membrane, a liquid
containing enantioselective carriers is impreg-
nated into a porous support membrane (see the
topic Supported liquid membrane for
Chiral Drug Separation by enantioselective separation). In enantioselective
Membranes solid membranes, chirality recognition exists in
main polymer chain, chiral polymer branch,
Keiji Sakaki doped chiral selector, or cavity formed by molec-
National Institute of Advanced Industrial Science ular imprinting (see the topic Enantioselective
and Technology (AIST), Research Institute for transport of amino acids by membrane
Sustainable Chemistry, Tsukuba, Ibaraki, Japan operations).
Although the industrial application of
membrane-based chiral separation has been very
An enantiomer is one of two stereoisomers that few until now, one successful example exists in
are mirror images of each other and chiral drug production. Diltiazem is a calcium
non-superposable. This feature of an enantiomer channel blocker used for hypertension and angina
is called chirality. treatment, and a lipase-immobilized membrane
Chirality is a very important key word in phar- reactor was applied to the enantiomeric separation
maceutical industry, because more than half of of chiral intermediate of this drug (Lopez and
pharmaceutically active ingredients are chiral Matson 1997). In the separation process,
(Sekhon 2010). In chiral drugs, one enantiomer a biphasic membrane system is used. Figure 1
shows desirable bioactivity, but the other enantio- shows the concept of enantiomeric separation in
mer frequently does the less bioactivity or some- a biphasic membrane reactor using enzyme
times toxic side effects. Therefore, worldwide immobilized on the membrane surface. The
400 Chlorination Process
Chiral Drug Separation

Organic phase (shell side)
by Membranes,
Fig. 1 Concept of (R), (S)-isomers (ester) (R)-isomer (ester)
enantiomeric separation in a
biphasic enzyme membrane
Lipase-immobilized UF membrane
reactor
(S)-isomer (acid)
Aqueous Phase (lumen side)
immobilized enzyme on a membrane hydrolyzes

preferentially one isomer ester in the organic feed Chlorination Process
phase, and the produced acid is extracted into the
aqueous phase. In this system, reaction and sepa- Zhi Wang
ration occur simultaneously in/on membranes. Chemical Engineering Research Center, School
Membrane technology has a potential to of Chemical Engineering and Technology, Tianjin
improve the application of enzymatic reactions University, Tianjin, China
to the industrial scale production of chiral drugs.
The recent progress of enzymatic membrane reac-
tors was reviewed elsewhere (Rios et al. 2004). Chlorination process, in chemistry, is a halogena-
The most important factor for successful opti- tion reaction using chlorine. Because chlorine is a
cal resolution of enantiomers in a racemate is the highly efcient disinfectant, chlorine is widely
design of a chiral selector tting for the separa- used in the treatment of water. In this case, the
tion, because many drugs have complicated steric chlorination process is the addition of chlorine or
structures. One of the measures devised to deal chlorine compound. Aqueous free chlorine solu-
with this problem is the implementation of com- tions contain three species in equilibrium, Cl2,
binatorial strategies (Maier et al. 2001). Large HClO, and ClO-. At present, it is generally
libraries of potential selectors are generated, and accepted that the chlorination is closely related
the systems with desirable properties are extracted to the major chlorinating species in free chlorine
from the libraries. solution and their concentration, both of which
depend on the pH of the solution.
For membranes, the chlorination is the reaction
of chlorine with membrane materials. The chlo-
References rine is widely used as the pretreatment agent dur-
ing many kinds of membrane process because of
Li B, Haynie DT (2007) Chiral drug separation. In: Ency-
clopedia of chemical processing. Taylor & Francis
their oxidability. However, this oxidizing chlorine
Online, pp 449458 can react with membrane, resulting in membrane
Lopez JL, Matson JL (1997) A multiphase/extractive failure. Usually, the hydrogen atom in the mem-
enzyme membrane reactor for production of diltiazem brane is substituted by the chlorine atom. Many
chiral intermediate. J Membr Sci 125:189211
Maier NM, Franco P, Lindner W (2001) Separation of enan-
materials used for membrane fabrication contain
tiomers: needs, challenges, perspectives. J Chromatogr active hydrogen. One of the most famous ones is
A 906:333 polyamide, which is widely used for the fabrica-
Rios GM, Belleville MP, Paolucci D, Sanchez J (2004) tion of many kinds of membranes such as reverse
Progress in enzymatic membrane reactors a review.
osmosis membrane (RO), nanoltration mem-
Sekhon BS (2010) Enantioseparation of chiral drugs an brane (NF), gas separation membrane,
overview. Int J PharmTech Res 2:15841594 pervaporation membrane, and forward osmosis
Chlorine-Resistant Polymeric Membranes 401
membrane (FO). The polyamide is easily to be coagulant residuals on polyamide membrane perfor-
chlorinated because the nitrogen atom on amide mance. Desalination 150:1530
Gabelich CJ, Frankin JC, Gerringer FW, Ishida KP, Suffet
bond is susceptible to chlorine due to the electro- IH (2005) Enhanced oxidation of polyamide mem-
philic character of the carbonyl groups. The chlo- branes using monochloramine and ferrous iron.
rination process of polyamide consists of the J Membr Sci 258:6470
substitution of chlorine atom for hydrogen atom Glater J, Mc CJW, Mc CSB, Zachariah MR (1981) The
of amide bond, accompanied by weakening or
effect of halogens on the performance and durability of
reverse-osmosis membranes. In: Synthetic membranes.
C
destruction of the hydrogen bonds between poly- American Chemical Society, Washington, DC,
mer chains (Xu et al. 2013). This process consists pp 171190
of two steps: positive chlorine attachment to the Glater J, McCutchan JW, McCray SB, Zachariah MR
(1982) Halogen interactions with typical reverse osmo-
amide unit at the O atom with lone electron pair sis membranes. In: Proceedings of water reuse sympo-
followed by rapid rearrangement to the N-Cl sium, Washington, DC, pp 13991409
(Avlonitis et al. 1992; Jensen et al. 1999; Challis Jensen JS, Lam YF, Helz GR (1999) Role of amide nitro-
and Challis 2010). For aromatic polyamide, the gen in water chlorination: proton NMR evidence. Envi-
ron Sci Technol 33:35683573
chlorination process consists of not only the sub- Kwon YN, Leckie JO (2006a) Hypochlorite degradation of
stitution process but also the chlorination of aro- crosslinked polyamide membranes: I. Changes in
matic ring through an intermolecular chemical/morphological properties. J Membr Sci
rearrangement after amide chlorination, known 283:2126
Kwon YN, Leckie JO (2006b) Hypochlorite degradation of
as the Orton rearrangement which involves the crosslinked polyamide membranes: II. Changes in
transfer of the N-chlorine atom from a side hydrogen bonding behavior and performance.
amino group to a phenyl ring forming various J Membr Sci 282:456464
aromatic substitution products under acidic con- Murphy AP (1991) Accelerated deacetylation of cellulose
acetate by metal salts with aqueous chlorine. Res
dition (Kwon et al. 2006a, b). Although the chlo- J Water Pollut Control Fed 63:177180
rination process could also be affected by various Tessaro IC, da Silva JBA, Wada K (2005) Investigation of
other factors (Gabelich et al. 2002, 2005; Tessaro some aspects related to the degradation of polyamide
et al. 2005; Murphy 1991), it is generally accepted membranes: aqueous chlorine oxidation catalyzed by
aluminum and sodium laurel sulfate oxidation during
that this chlorination is closely related to the major cleaning. Desalination 181:275282
chlorinating species in chlorine solution and their Xu J, Wang Z, Wei X, Yang S, Wang J, Wang S (2013) The
concentration (Glater et al. 1981, 1982). chlorination process of crosslinked aromatic polyamide
The chlorination usually causes serious degra- reverse osmosis membrane: new insights from the
study of self-made membrane. Desalination
dation of membrane chemical structure, ulti- 313:145155
mately resulting in membrane failure. Therefore,
developing novel membranes with high chlorine
resistant property is very important for the wide
application of membrane separation technology.
Chlorine-Resistant Polymeric
Membranes
References Pejman Ahmadiannamini
Avlonitis S, Hanbury WT, Hodgkiess T (1992) Chlorine
Biomedical Engineering, University of Arkansas,
degradation of aromatic polyamides. Desalination Fayetteville, AR, USA
85:321334
Challis BG, Challis JA (2010) Reactions of the
carboxamide group. In: Zabicky J (ed) Amides
A chlorine-resistant polymeric membrane is
(1970). Wiley, Chichester, pp 731857
Gabelich CJ, Yun TI, Coffey BM, Suffet IHM mainly referred to a polymeric reverse osmosis
(2002) Effects of aluminum sulfate and ferric chloride (RO) desalination membrane that can withstand
402 Chlorine-Resistant Polymeric Membranes
exposure to chlorine and preserve its separation Cl2 + H2O HCIO + H+ + Cl-
characteristics under such a harsh condition.
Currently, commercially available RO mem- HCIO H+ + CIO-
branes are derived from two classes of polymers:
cellulose acetate (CA) and aromatic polyamide Cl2, HOCl, and OCl are in equilibrium, and
(PA) (Li and Wang 2010). CA membranes are depending on the pH, different distributions of
relatively low cost and tolerant to limited free aqueous chlorine species are observed (Deborde
chlorine. However, CA membranes suffer from and von Gunten 2008). The established
some disadvantages such as a narrow operating chlorinating strength of these species is
pH range (4.57.5), susceptibility to biological Cl2 ~ HOCl > OCl (Soice et al. 2003).
attack, structural compaction under high pressure, The amide nitrogen of the membrane is highly
and low upper temperature limit. On the other vulnerable to chlorine attack because of electron-
hand, PA thin lm composite (TFC) membranes withdrawing effect of carbonyl group. Upon
feature thin highly selective interfacially polymer- exposure to free chlorine, N-H group is chlori-
ized layers, which exhibit superior ux and salt nated to N-Cl group which can reversibly
rejections, wider operating temperature and pH form the initial amid by treatment with reducing
range, and higher stability to biological attack, as agents. The aromatic rings are also susceptible to
compared with CA membranes (Park et al. 2008). attack by chlorine because it is an electron-
Thus, PA TFC membranes used in RO are most rich region. Two possible chlorination mecha-
preferred by the desalination industry. However, nisms are proposed for aromatic rings, i.e., direct
PA membranes encounter signicant drawbacks chlorination of the aromatic ring and
in desalination processes, namely, membrane deg- Orton rearrangement, which involves initial chlo-
radation under continuous exposure to trace rination of amid nitrogen followed by an
amount of chlorine and deterioration of their intermolecular rearrangement, forming various
performances. aromatic substitution products (Fig. 1) (Raval
Chlorine is the most common industrial oxidiz- et al. 2010).
ing biocide in water treatment that is used for dis- According to the chlorination mechanisms,
infection of domestic water and for the removal of chlorine resistance of PA membranes largely
tastes and odors from water. It is also typically used depends on the chemical structures of the diamine
in RO treatment processes in order to control micro- components used. Aliphatic PA reversibly reacts
organisms that biofoul and clog the membrane. with chlorine to yield N-chlorinated amide. Ter-
Consequently, water to be puried is often chlori- tiary PAs are inactive towards oxidative chlorine.
nated, to disinfect it and ultimately inhibit biofoul- The Orton rearrangement takes place only when
ing of the membranes, and then dechlorinated amide linkage is directly connected with benzene
before being fed to membrane desalination units. ring. Generally, the chlorine resistance increases
After passing through the membranes, the water is in the order of PA synthesized from aromatic,
then rechlorinated before being sent to the distribu- cycloaliphatic, and aliphatic diamines,
tion network. However, this requires additional respectively.
equipments and chemicals and increased operating Thus, the following modications can be con-
cost for the plant, accordingly. sidered as potential strategies to enhance the chlo-
When chlorine gas is added into water, it is rine resistance of PA membranes (Geise et al. 2010):
hydrolyzed to form hydrogen ion, chloride, and
hydrochlorous acids. The hydrochlorous can be 1. Replacing chlorine-sensitive amidic hydrogen
further ionized to produce hypochlorite ions on the amide linkages with other moieties, e.g.,
according to the following reactions (Geise methyl (CH3) or phenyl (C6H5)
et al. 2010):
Chlorine-Resistant Polymeric Membranes 403
O O H H O O H H
Cl2
C C N N C C N N
Irreversible direct aromatic chlorination
n Cl n
Cl C
Re 2,
HC e t
ve IO ibl en
rs
ibl an ers em
e d ev ng
ch Cl Irr ra
lor O- ar
ina re
tio ton
n Or
O O Cl H
C C N N
Chlorine-Resistant Polymeric Membranes, composite RO membrane by controlled chlorine treatment,

Fig. 1 Proposed mechanism of chlorination of aromatic 945949, Copyright (2010) with permission from
PA (Reprinted from Desalination, 250, Raval HD, Trivedi Elsevier)
JJ, Joshi SV, Devmurari CV, Flux enhancement of thin lm
2. Replacing the aromatic ring bonded to the References

amide nitrogen with aliphatic chain or cyclics
3. Prevention of Orton rearrangement by adding Deborde M, von Gunten U (2008) Reactions of chlorine
with inorganic and organic compounds during water
protective groups at the possible chlorination
treatmentkinetics and mechanisms: a critical review.
sites on the aromatic rings Water Res 42(12):1351
Geise GM, Lee H, Miller DJ, Freeman BD, McGrath JE,
Other than PA, polysulfone has much better Paul DR (2010) Water purication by membranes: the
role of polymer science. J Polym Sci B: Polym Phys
chlorine resistance as it has stronger chemical
48(15):16851718
bonds. However, due to hydrophobic nature of Li D, Wang H (2010) Recent developments in reverse
polysulfone, introduction of controlled levels of osmosis desalination membranes. J Mater Chem
hydrophilic groups, e.g., SO3H, while retaining 20:45514566
Park HB, Freeman BD, Zhang Z, Sankir M, McGrath
its physical properties is necessary for polysulfone JE (2008) Highly chlorine-tolerant polymers
RO membranes. The polymer chain cleavage and for desalination. Angew Chem Int Ed
side reactions, which can make it difcult to con- 47(32):60196024
trol the degree of sulfonation and molecular Raval HD, Trivedi JJ, Joshi SV, Devmurari CV (2010)
Flux enhancement of thin lm composite RO mem-
weight and consequently drop the mechanical
brane by controlled chlorine treatment. Desalination
and the thermal properties of polymer, can occur 250(3):945949
in the post-sulfonation method. The direct synthe- Soice NP, Maladono AC, Takigawa DY, Norman AD,
sis of the functionalized polysulfone from a Krantz WB, Greenberg AR (2003) Oxidative
degradation of polyamide reverse osmosis
sulfonated monomer is found to be more advan-
membranes: studies of molecular model compounds
tageous since the associated challenges can be and selected membranes. J Appl Polym Sci
avoided (Park et al. 2008). 90(5):11731184
404 Chloromethylation and Amination of Polymers (for AEM)
Chloromethyl groups can be introduced via

Chloromethylation and Amination Friedel-Crafts mechanism typically by reacting
of Polymers (for AEM) polymer with a chloromethylating agent, such as
chloromethyl methyl ether (CH3OCH2Cl) or
Pejman Ahmadiannamini chlorotrimethylsilane (C3H9ClSi), using a metal-
Biomedical Engineering, University of Arkansas, lic chloride (e.g., FeCl3, ZnCl2, or SnCl4) as cat-
Fayetteville, AR, USA alyst. Cross-linking of benzene rings with
methylene bridges is encountered as a side reac-
tion which would result in lower ion exchange
Chloromethylation (generally haloalkylation) of capacity of nal product according to the reaction
polymeric membranes followed by an amination presented in Fig. 1 (Jones 1952):
reaction has been widely used for preparation of Alternatively, chloromethylation of polymers
anion exchange membranes (AEMs). Signicant can be carried out using reaction mixtures which
efforts have been made to prepare AEMs from produce chloromethyl methyl ether in situ. These
insoluble polymeric lms such as polyethylene reaction mixtures are also known as
(PE), polypropylene (PP), polytetrauoroethylene chloromethylating complexes and may be, for
(PTFE), and polyvinylidene uoride (PVDF). example:
Such AEMs are typically prepared by grafting
vinyl monomers such as styrene onto polymeric Mixtures of paraformaldehyde and
lms and subsequent chloromethylation and hydrochloric acid
amination modications. Also, soluble polymers Mixtures of formaldehyde polymers and alu-
such as polystyrene (PS), polyether sulfone minum chloride
(PES), polyether ketone (PEK), or their blends Mixtures of methylal, thionyl chloride, and
or block copolymers have been widely used for catalyst
preparation of AEMs. In such case, charged moi- Mixtures of methylal, sulfuryl chloride plus
eties can be introduced to the dissolved polymer chlorosulfonic acid, or chlorosulfonic acid
via chloromethylation and amination reactions, only
either prior to the casting of polymer or after the Mixtures of methylal, chlorosulfonic acid, and
casting into a lm. Posttreatment, i.e., cross- an oxygen-containing polar liquid
linking of membranes, is often required to
improve the chemical stability of AEMs prepared Among these complexes, the most commonly
from soluble polymers (Xu 2005). used are the ones that contain chlorosulfonic acid.
CH3OCH2Cl
R R CH2Cl
SnCl4
R + R CH2Cl R CH2 R
Chloromethylation and Amination of Polymers (for AEM), Fig. 1 Chloromethylation and methylene bridge
crosslinking of polymers
Chromium Transport Through Ion Exchange Membranes 405
Chloromethylation R R
and Amination
of Polymers (for AEM),
Fig. 2 Amination reaction R1
of chloromethylated
N H
polymers
NH3 R2 R1
Cl N+H3 R
N+H Cl
C
R2
1
R R1
R N H N R3 R
H R2
CH2Cl
R1 R1
Cl N+H2 N+ Cl
R2
R3
The reaction can be represented as (Boutier

et al. 1980): Chromium Transport Through Ion
Exchange Membranes
CH3 OH CHOH SO3 HCl
! CH3 OCH2 Cl H2 SO4 Eugenio Bringas and Immaculada Ortiz
Department of Chemical and Biomolecular
Positively charged groups are introduced to the Engineering, University of Cantabria, Santander,
chloromethylated polymer via amination reaction. Cantabria, Spain
The functional group of the AEMs can be weakly
basic, such as primary, secondary, or tertiary
amino groups which are obtained via reaction of Occurrence and Removal of Hexavalent
chloromethylated polymer with ammonia or with Chromium from Polluted Waters
primary or secondary amines, respectively. The
amination reaction with tertiary amines, known The widespread use of chromium in industrial
as quaternarization, leads to AEMs with strongly applications such as leather tanning, metallurgy,
basic functional groups (quaternary ammonium) electroplating processes, and so forth has caused
(Helfferich 1995) (Fig. 2). chromium contamination of surface and ground-
waters (Pagilla and Canter 1999; Lin 2002;
Rengeraj et al. 2001). Public concerns arise as a
References result of the sufcient evidence of chromium
(VI) carcinogenicity in humans and the
Boutier JL, Grammont PDA, Herbin JEE (1980) Process bioaccumulation into ora and fauna (IARC
for the chloromethylation of polymers for the produc- 1990; Kimbrough 1999).
tion of anionic exchange resins, United States Patent, Conventional processes reduce Cr6+ to Cr3+,
4,225,644
Helfferich FG (1995) Ion exchange. Dover Publications,
which is less toxic, less soluble, and less mobile
New York than Cr6+. Trivalent chromium is then converted
Jones GD (1952) Chloromethylation of polystyrene. Ind to the hydroxide form and removed after precipi-
Eng Chem 44(11):26862693 tation (Hashim et al. 2011). Other technologies
Xu T (2005) Ion exchange membranes: state of their devel-
opment and perspective. J Membr Sci 263(12):129
have been reported in literature as feasible
406 Chromium Transport Through Ion Exchange Membranes
Chromium Transport 2 1 0 1 2 3 4 5 6 7 8 9 10
Through Ion Exchange 2
Membranes,
Fig. 1 Speciation diagram 1 Cr2O72-
of Cr6+ in aqueous solution
Log [Cr] total

0
(Bringas et al. 2006a)
H2CrO4 HCrO4- CrO42-
1
4
2 1 0 1 2 3 4 5 6 7 8 9 10
pH
alternatives to achieve the removal and recovery while in the acid range (1< pH <7), the coexis-
of hexavalent chromium from wastewaters: tence of hydrogen chromate (HCrO4) and
adsorption and ion exchange (Mohan and Pittman dichromate (Cr2O72) is expected. Therefore, the
2006; Rivero et al. 2004), membrane processes extraction of Cr6+ takes place through an anion-
(Bohdziewicz 2000) and separation processes exchange mechanism, and thus, solvating and
based on the use of selective liquid membranes, basic extractants are required.
especially emulsion liquid membranes (ELMs) The extraction of chromium by solvating
(Salazar et al. 1992) supported liquid membranes extractants is based on the formation of an organic
(SLMs) (Kentish and Stevens 2001), intermediate by the interactions between an elec-
non-dispersive solvent extraction (NDSX) tron donor (usually a phosphinic or a phosphoric
(Bringas et al. 2012), and the emulsion pertraction moiety) and an electron acceptor (usually a pro-
technology (EPT) (Bringas et al. 2006a). ton) as shown by the following reaction which
occurs in acid media:
Formulation of Selective Liquid S H X $ SHX (1)

Membranes for Hexavalent Chromium
Removal where S is the solvating carrier and X- represents
the target species.
The main characteristics of the compounds to The commercial solvating extractants that
be employed as selective carriers in the formula- employed the most on the removal of hexavalent
tion of selective liquid membranes in industrial chromium are TBP (tributyl phosphate marketed
applications are as follows: (i) high selectivity by Lanxess) (Venkateswaran and Palanivelu
toward the target species, (ii) minimum extrac- 2005), Cyanex 921 (trioctylphosphine oxide,
tion capacity of competing compounds, (iii) ease TOPO, marketed by Cytec Industries) (Alguacil
of regeneration, (iv) high solubility in low-cost et al. 2004), and Cyanex 923 (a mixture of four
solvents, (v) long-term stability, (vi) low cost, trialkylphosphine oxides marketed by Cytec
and (vii) not exhibiting toxic and carcinogenic Industries) (Alguacil et al. 2004).
properties. Tertiary amines and salts of quaternary amines
Figure 1 shows the distribution of hexavalent represent the most important groups of basic
chromium species in aqueous solution as a func- extractants employed in the extraction of
tion of pH and the total chromium concentration hexavalent chromium being the extraction mech-
(Bringas et al. 2006a). It is noticed that at basic pH anism described by the following ion-exchange
values, the major species is chromate (CrO42), reactions:
Salts of quaternary amines (chloride form) Removal of Hexavalent Chromium from

Polluted Groundwaters by the EPT
QCl X $ QX Cl (2) Technology
Tertiary amines
Bringas et al. (2006a, b, 2007) reported the meth-
T HY $ TH Y Protonation (3) odology for the optimal design of selective
liquid membrane processes as efcient alterna- C
tives for the separation and selective recovery of
TH Y X $ TH X
hexavalent chromium from polluted groundwa-
Y Ion exchange (4) ter generated by efuent leaking from surface
deposition of industrial wastes. Furthermore,
where QCl is the chloride salt of a quaternary amine, other competitive anionic species (mainly sulfate
T is a tertiary amine, and HY is a mineral acid. and chloride anions) were also present in the
groundwater due to the specic location being
Aliquat 336 (N-methyl-N,N-dioctyloctan-1- close to the shore. The process should be
ammonium chloride marketed by BASF) designed to (i) reduce the chromium concentra-
(Bhowal et al. 2012) and Alamine 336 (mixture tion in the polluted water below the disposal limit
of tertiary amines with C8 and C10 groups (0.5 mg L1) and (ii) obtain a concentrated chro-
marketed by BASF) (Bringas et al. 2006b) are mium solution (>20 g L1) with negligible con-
the most reported commercial quaternary and ter- centrations of sulfate and chloride anions.
tiary amines employed in the removal of The design methodology is briey illustrated in
hexavalent chromium species. Fig. 2.
1.FUNDAMENTALS
Problem Statement Selection of Extraction (EX) and Back-Extraction (BEX) MBSX Technology Selection
Agents
- Effluent characterisation - Proposal of EX and BEX reactions - Process configuration
- Objectives - Kinetic or equilibrium Patten - Operational conditions
- Obtention of the equilibrium/kinetic parameters
2. TECHNICAL EVALUATION
Kinetic Analysis of the Separation - Concentration Process Mathematical Modeling of the Separation - Concentration Process
- Process Viability - Mass transfer hypothesis
- Analysis of the influence of the operation variables on the EX - Solute mass balances in the different fluid phases
and BEX kinetics - Mass transfer coefficients
- Model validation
3. PROCESS DESIGN 4. ECONOMICAL EVALUATION
- Removal and recovery requirements (Constrains) Detailed cost evaluation of the selected alternative
- Preliminary capital cost definition (Objective Function) - Capital cost evaluation
- Proposal of design alternatives (Network Superestructures) - Operation cost evaluation
- Mathematical modeling of the network superestructures - Analysis of the investment recovery
(Opimization Model)
- Application of the mathematical optimization techniques to select
the alternative with minimum total cost
Chromium Transport Through Ion Exchange Membranes, Fig. 2 Design methodology of selective liquid mem-
brane processes
Synthetic solutions that simulate the extraction mechanism and the equilibrium param-
anionic composition of the real groundwater eters for the multicomponent system, Alamine
[Cr6+] = 5001,300 mg L1, [SO42] = 336/chromium/sulfate/chloride.
9003,000 mg L1 and [Cl] = 3001,000 mg The emulsion pertraction technology (EPT)
L1, were employed as feed solutions. The pH of which combines the advantages of ELM, i.e.,
the feed solution was adjusted to 1.5 using H2SO4. high efciency due to large surface area for
The organic solution was formulated as follows: mass transfer, and SLM, i.e., extraction and strip-
10 % (v/v) of Alamine 336 as the anionic ping are performed in one step without disper-
extractant, 3 % (v/v) of Pluronic PE 3100 as sion of the organic phase into the aqueous feed
surfactant, and Isopar L uid as solvent. The phase, was selected as the most suitable congu-
emulsion was prepared by dispersing the stripping ration of the selective liquid membrane technol-
solution (NaOH 36 mol L1) in the organic ogy. More details about the technology can be
phase at a volume ratio organic/stripping = 4/1. found elsewhere (Bringas et al. 2006a; Bringas
After dening the separation system, the exper- and Ortiz 2014). Kinetic experiments were
imental and theoretical description of the extrac- performed in a laboratory experimental setup
tion equilibria was performed following the containing a membrane contactor Liqui-Cel
guidelines reported in Fig. 2. In general, extraction ExtraFlow 2.5 8 with ber X-30 that pro-
systems involving ionic species are generally vides an effective membrane area of 1.4 m2. Fig-
thought to reach chemical equilibrium at the inter- ure 3 shows the evolution with time of the
face, and thus equilibrium models of the chemical chromium concentration in the different uid
reactions are proposed. Table 1 reported the phases (feed, organic, and stripping) and a
Chromium Transport Extraction reaction Equilibrium parameter

Through Ion Exchange
HCrO
4 H R3 N , R3 NH HCrO4
K1 = 4.83 mol2 m6
Membranes,
Table 1 Extraction
7 2H 2R3 N , R3 NH 2 Cr 2 O7
Cr 2 O2 K2 = 0.741 mol4 m12
mechanism and equilibrium
HSO
4 H R3 N , R3 NH HSO4
K3 = 1.37 103 mol2 m6
parameters
R3 NH HSO4 Cl , R3 NH Cl HSO
4
K4 = 1.52
2 100
Aqueous
[Cr+6](mol/m3) stripping
80 Phase
1,5 Aqueous
[Cr+6](mol/m3)
Inlet Phase Stripping

60
1 Outlet
40
0,5
20
0 0
0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6
t (h)
Feed Organic Stripping
Chromium Transport Through Ion Exchange Membranes, Fig. 3 Qualitative and quantitative evolution of the
chromium concentration in the uid phases (Feed solution: ows in continuous mode; Emulsion: recirculated)
photograph of the uid phases at the end of the and (ii) the equilibrium parameters of the extrac-
experiment. Similar extraction and back- tion reactions reported in Table 1. The model
extraction results were reported for sulfate and described satisfactorily the kinetics and selectiv-
chloride anions (Bringas et al. 2006a). ity of the separation-concentration EPT process
The kinetic results conrmed the high selectiv- under a wide range of operational conditions
ity of the carrier Alamine336 toward chromium and served as a tool to perform the optimal
species. The total chromium concentration design of the separation process (Bringas C
decreased from 500 mg L1 at the module inlet et al. 2006a).
to 0.8 mg L1 at the module outlet in one pass Finally, a novel design methodology based on
when the residence time in the module is 0.014 h. the application of mathematical optimization was
The chromium concentration in the stripping developed in order to determine the optimal EPT
phase reached an average value of 14.6 g L1 process conguration able to achieve, at minimum
after 1.3 h, and thus, a concentration factor of cost, the separation and concentration objectives,
18,733 was reached before depletion of the strip- (i) the contaminant removal up to the required
ping reagent. From the analysis of the results concentration level ([Cr6+] <0.5 mg L1) and,
obtained for the competitive species, it was con- (ii) the selective contaminant recovery to obtain
cluded that the back-extraction process was not a concentrated solution that can be reused else-
selective because all the species were totally back- where ([Cr6+] 20 g L1). The different design
extracted from the organic phase. For this reason alternatives were represented by means of net-
the selectivity had to be controlled in the work superstructures that were modeled through
extraction step. the developed EPT mathematical model, generat-
The kinetic analysis was completed with ing optimization problems that were solved fol-
the proposal of the multicomponent separation lowing global optimization techniques (Bringas
mathematical model of the EPT process with et al. 2007).
the corresponding design parameters: (i) mass The procedure was illustrated for the treatment
transport coefcients in the aqueous feed phase of an incoming groundwater stream with a known
boundary layer (kL), in the membrane (km), and ow rate of 2.5 m3 h1 and a chromium concen-
in the organic phase stagnant layer (koAv) tration of 400 mg L1 (7.7 mol m3). Comparison
for the different species involved in the of the results obtained from the analysis of the
separation-concentration process (see Fig. 4) different design alternatives concluded that a
Chromium Transport Parameter Value Obtaining

Through Ion Exchange
Membranes, KL,Cr 0.217 . (Q1/3
2
)(m/h)
Fig. 4 Values of the mass
transport coefcients in the KL,Sulphate 0.234 . (Q1/3)(m/h)
Mass Transport Coefficients
EPT mathematical model Aqueous Phase Stagnant Layer

2
(kLi)
KL,Chloride 0.317 . (Q1/3
2
)(m/h) Calculated
KL,H+ 0.256 . (Q1/3

2
)(m/h)
Km,Cr 1.2 103 m/h

Membrane (kmi)
Km,Sulphate/ Chloride 1.5 103 m/h
Parametric
Organic Phase Stagnant Layer Estimation
.A
(ko)
Ko v 2.94 104h1
1,41 m3/h 1,41 m3/h

7,7 mol/m3 2 0,041 mol/m3
A1=248 m
2,5 m3/h 2,5 m3/h

2,5 m3/h 0,041 mol/m3 0,01 mol/m3
2 2
A total = 591 m m3a A3=150 m
7,7mol/m3
S4a
3 3
1,09 m /h 1,09 m /h
3
7,7 mol/m3 2 0,041 mol/m
A2=193 m
a 3 a 3
Q = 2,5 m /h Q = 2,5 m /h
3 3
[Chromium(VI)] in = 7,7 mol/m [Chromium(VI)] out 0,01 mol/m
Chromium Transport Through Ion Exchange Membranes, Fig. 5 Optimal process conguration for the aqueous
feed solution
3
0,13 m /h
3
Co =0,002 mol/m
3
Cg = 0,1 mol/m
2
A1 = 248 m
3
0,24 m /h 0,13 m3/h
Co = 0,002 mol/m3 Co = 0,42 mol/m3
3
Cg = 0,1 mol/m Cg = 400 mol/m
3
S4e
3
0,11m /h 0,24 m3/h
3
Co = 0,002 mol/m Co = 0,002 mol/m3
3
Cg = 0,1 mol/m Cg = 400 mol/m3
2 2
m2e A2=193 m A3 = 150 m
3 0,24 m3/h
0,24m /h 3
3 Co = 0,42 mol/m
Co = 0,24 mol/m
Cg = 396,9 mol/m3
Cg = 223,2 mol/m3
Atotal = 591m2 0,048 m3/h
Cg = 0,1 mol/m3
0,048 m3/h
Cg = 400 mol/m3
[Chromium(VI)]back-extration 380 mol/m3
Chromium Transport Through Ion Exchange Membranes, Fig. 6 Optimal process conguration for the emulsion
phase
network conguration with three treatment mod- Future Directions

ules combined in series and parallel and with a
total membrane area of 591 m2 was the most Future development will require research and
suitable process conguration to comply, at min- development activities in the following areas:
imum cost, with the concentration specications
in the outlet streams. Figures 5 and 6 show the 1. Evaluation of the operation cost (energy
optimal networks for the aqueous feed solution and material costs) to be considered in the
and the emulsion phase, respectively. formulation of the objective function
Citric Acid Recovery by Electrodialysis 411
employed to determine the optimal process Lin C (2002) The chemical transformations of chromium
network in natural waters-A model study. Water Air Soil Pollut
139:137
2. Analysis of the long-term performance of the Mohan D, Pittman CU (2006) Activated carbons and low
separation process to evaluate the stability of cost adsorbents for remediation of tri- and hexavalent
the selective carrier which guarantees the pro- chromium from water. J Hazard Mater 137:762
cess economic viability Pagilla KR, Canter LW (1999) Laboratory studies on reme-
3. Rigorous evaluation cost of the optimal pro-
diation of chromium contaminated soils. J Environ Eng
125:243
C
cess network Rengeraj S, Yeon KH, Moon SH (2001) Removal of chro-
mium from water and wastewater by ion-exchange
resins. J Hazard Mater B87:273
Rivero MJ, Primo O, Ortiz MI (2004) Modellins of Cr
(VI) removal from polluted groundwaters by ion
References exchange. J Chem Technol Biotechnol 79:822
Salazar E, Ortiz MI, Urtiaga AM (1992) Kinetics of the
Alguacil FJ, Lpez-Delgado A, Alonso M, Sastre AM separation of chromium(VI) with emulsion liquid
(2004) The phosphine oxides Cyanex 921 and Cyanex membranes. Ind Eng Chem Res 31:1523
923 as carriers for facilitated transport of chromium Venkateswaran P, Palanivelu K (2005) Studies on recovery
(VI)-chloride aqueous solutions. Chemosphere 57:813 of hexavalent chromium from plating wastewater by
Bhowal A, Bhattacharyya G, Inturu B, Datta S (2012) supported liquid membrane using tri-n butyl phosphate
Continuous removal of hexavalent chromium by emul- as carrier. Hydrometallurgy 78:107
sion liquid membrane in a modied spray column. Sep
Purif Technol 99:69
Bohdziewicz J (2000) Removal of chromium ions (VI) from
underground water in the hybrid complexation-
ultraltration process. Desalination 129:227
Bringas E, Ortiz I (2014) Emulsion pertraction technology
for zinc recovery. In: Drioli E, Giorno L (eds) Encyclo- Citric Acid Recovery by
pedia of membranes. Springer-Verlag Berlin Heidel- Electrodialysis
berg 2014. doi:10.1007/978-3-642-40872-4_646-1
Bringas E, San Romn MF, Ortiz I (2006a) Separation and
recovery of Anionic pollutants by the emulsion Gerald Pourcelly
pertraction technology. Remediation of polluted Institut Europeen des Membranes, CC 047,
groundwaters with Cr(VI). Ind Eng Chem Res 45:4295 Universite Montpellier II, Place Eugene Bataillon,
Bringas E, San Romn MF, Ortiz I (2006b) Removal of Montpellier, France
anionic pollutants from groundwaters using Alamine
336: chemical equilibrium modeling. J Chem Technol
Biotechnol 81:1829
Bringas E, Karuppiah R, San Romn MF, Ortiz I, Carboxylic acids such as lactic, succinic,
Grossmann IE (2007) Optimal groundwater remedia- gluconic, citric, and tartaric are widely used in
tion network design using selective membranes. Ind
Eng Chem Res 46:5555 food processing, detergent manufacture, and bio-
Bringas E, San Romn MF, Urtiaga AM, Ortiz I (2012) degradable plastic production (Bailly
Membrane contactors (NDSX and EPT): an innovative et al. 2001). Their production at the industrial
alternative for the treatment of efuents containing scale is mainly achieved by mean of fermentation
metallic pollutants. Int J Environ Waste Manag 9:201
Hashim MA, Mukhopadhyaya S, Sahu JN, Sengupta from molasses, starch hydrolysates, or sugars.
B (2011) Remediation technologies for heavy metal Traditional processes to obtain these carboxylic
contaminated groundwater. J Environ Manage 92:2355 acids consist on the precipitation of both the
IARC (1990) IARC monographs on the evaluation of the acids and their salts, so they can be isolated
carcinogenic risk of chemical in humans, Chromium,
Nickel and Welding, Environmental and Experimental from the rest of the components of the raw mate-
Data. IARC, Lyon rial. Then, they are placed in an acid medium
Kentish SE, Stevens GW (2001) Innovations in separations (sulfuric acid) to generate the acid form.
technology for the recycling and re-use of liquid waste A latter concentration step is carried out by evap-
streams. Chem Eng J 84:149
Kimbrough DE, Cohen Y, Winer AM, Creelman L, oration followed by a crystallization. This gener-
Mabuni C (1999) A critical assessment of chromium ates large volumes of efuents with high salt
in the environment. Crit Rev Environ Sci Technol 29:1 contents. For example, typical yields of 1 kg of
412 Cleaning Cycle of Fouled Membranes
Citric Acid Recovery by

Electrodialysis,
Fig. 1 Scheme of a process
of carboxylic acid
production from
fermentation. (EDC)
concentration ED step,
(EDBM) ED with bipolar
membranes
citric acid are obtained per 2 kg of gypsum which acids targeted around membrane operations: design of
is very difcult to dispose (Bialey and Ollis the concentration step by conventional electrodialysis.
1986). In order to reduce this environmental Bialey JE, Ollis DF (1986) Biochemical engineering fun-
impact, the design of alternative production damentals. MacGraw-Hill, Singapore
scheme was investigated. Extraction, adsorption, Novalic S, Jagschits F, Okwor J, Kulbe KD (1995) Behav-
and membrane technologies, like electrodialysis iour of citric acid during electrodialysis. J Membr Sci
108:201205
(ED), were proposed to replace precipitation Xu T, Yang W (2002) Citric acid production by electrodi-
(Novalic et al. 1995). A complete scheme for alysis with bipolar membranes. Chem Eng Process
carboxylic acid recovery is depicted in Fig. 1. 41:519524
For EDC step of citric acid salts to the three
dissociation constants of citric acid, the conduc-
tivity is mainly inuenced by the pH value and the
concentration. More ion dissociation leads to for- Cleaning Cycle of Fouled Membranes
mation of further citric-acid-charged complexes in
the solution which in turn causes electrical resis- Hongyu Li
tance reduction of the solution. The maximum ion UNESCO Centre for Membrane Science and
density and consequently the minimum electrical Technology, The University of New South Wales,
resistance of the solution is obtained at pH Sydney, Australia
8 (Novalic et al. 1995). The EDBM step,
which is water splitting by membrane electrodial-
ysis, provides an attractive complement to the Membrane Fouling
fermentation technology by removing the
product acid while simultaneously providing an Membrane fouling is an unavoidable process in
equivalent amount of base for use in adjusting the almost all industrial membrane applications. It
pH in the fermentor (Fig. 1). Moreover, the refers to the deposition of foulants (solutes or
produced citric acid is usually at a relatively particulates) on the membrane surface and/or
high concentration (0.5 M) so that the subsequent inside the membrane pores during membrane sep-
purication via crystallization or other aration process. Figure 1a demonstrates a typical
techniques is relatively inexpensive (Xu and membrane fouling structure in ultraltration
Yang 2002). (UF) and microltration (MF) systems.
The consequence of membrane fouling is the
deterioration of membrane separation perfor-
References mance, which can manifest in two possible ways:
Bailly M, Roux-de-Balman H, Aimar P, Lutin F, Cheyran 1. Decreased productivity, which can be observed
M (2001) Production processes of fermented organic in the reduction of permeate ux at a constant
Cleaning Cycle of Fouled Membranes 413
a the membrane as well as understanding the foul-

ing mechanisms of operation conditions such as
applied pressure, shear rate on the membrane sur-
Foulants face, operational temperature, the length of mem-
brane operation, and the degree of membrane
fouling could provide great insight for the selec-
Membrane tion of membrane cleaning protocols and the C
cleaning frequencies.
For example, microltration (MF) and ultral-
b c
Flux Transmembrane tration (UF) membranes used in the dairy industry
Pressure for concentration of milk or removal of bacterial
and spores would encounter components in the
feed and potential foulants such as proteins as
whey and casein, carbohydrates, fats, and min-
erals. Membranes used in reverse osmosis
(RO) of seawater may experience organic fouling
Time Time caused by humics, fulvic acid, carboxylic acid,
Cleaning Cycle of Fouled Membranes, and extracellular polymeric substances (EPS) as
Fig. 1 Structure of membrane fouling layer in UF or MF well as biofouling, which may involve adsorption
ltration system (a) and observation of ux (b) or trans- of organic matter to membrane surface, continu-
membrane pressure changes during fouling process ous adhesion of microorganisms, growth of
adhered cells, or subsequent formation of biopoly-
pressure operation mode (Fig. 1b) or increase mer matrix. Inorganic fouling or scaling initiated
in transmembrane pressure in a constant ux by supersaturation of minerals on the membrane
operation mode (Fig. 1c). surface is also common in RO, UF, and membrane
2. Changed separation efciency in terms of distillation (MD) processes.
selectivity of particular species, due to the Operation conditions of a membrane separation
altered surface characteristics by the buildup process also affect properties of foulant accumu-
of fouling layer such as pore size, surface lated on the membrane. For example, higher trans-
charge, or hydrophobicity. membrane pressure in RO, UF, and MF processes
leads to higher convection of foulants to
While membrane fouling is unavoidable in the membrane surface, resulting in a higher level
most cases, the degree of fouling can be managed of concentration polarization and eventually foul-
by using: ing. Higher pressure can also cause compaction
of already formed foulant layer and make it
(a) Appropriate membranes for a particular more difcult to clean. Higher shear near the mem-
application brane surface causes higher back transport of
(b) Appropriate pretreatment processes for the foulant away from the concentration polarization
feed before the membrane process layer and could reduce the formation of fouling
(c) Well-managed hydrodynamics in the mem- layer on the membrane. On the other hand, the
brane system and optimized operation system back transport of bigger foulants occurs preferen-
incorporating membrane cleaning process tially over smaller ones, which could lead to the
formation of foulant layer with smaller species,
which could contribute in higher resistance and is
Considerations for Membrane Cleaning more difcult to remove. Operation at higher tem-
perature should lead to higher ux; however, dena-
Understanding the properties of the feed including turation of certain species such as protein could also
the main components as the potential foulants on lead to more difculty in cleaning foulant layers.
414 Cleaning Cycle of Fouled Membranes
Cleaning Cycle of Fouled

Membranes, Draining
Fig. 2 Stages of a cleaning
cycle
Rinsing
Chemical cleaning 1
Draining and Rinsing
Chemical cleaning 2
Draining and Rinsing Optional
Sanitizing
Membrane Cleaning Process Rinsing should normally be conducted at the same

temperature as cleaning to avoid compaction of
The aim of membrane cleaning is to disrupt and foulants and normally takes 520 min.
remove foulants from the membrane surface and During the chemical cleaning process, suf-
restore the membrane separation capability. The cient contact between the cleaning agent and the
process is normally comprised of steps (as shown foulant is probably the rate limiting factor for
in Fig. 2) including draining of the feed solution cleaning reactions. This requires efcient transport
from the membrane system, rinsing of loosely of chemical to the foulant sublayer and rapid
bound foulants with clean water, and chemical removal of the dislodged foulant away from the
cleaning through breaking the cohesive and adhe- membrane. Appropriate chemicals should be
sive bonds between the foulants and between the selected that can attack the structure of foulant by
foulant and the membrane. Sequential cleaning cleavage of bonds within the foulant molecules
with a second chemical is also common for rou- and between the foulant aggregates, or solubilize
tine operations over repeated ltration and the foulant deposits.
cleaning cycles over a long period.
Draining and rinsing before application of
chemicals in membrane cleaning is very impor- Chemicals for Membrane Cleaning
tant. Without these steps, the majority of the
chemical cleaner would be consumed by the resid- Selection of right chemical is most important for
ual feed instead of the foulants on the membrane achieving an efcient cleaning reaction and pro-
surface and thus achieve low cleaning efciency. cess. Common chemicals used for membrane
Cleaning Cycle of Fouled Membranes 415
Cleaning Cycle of Fouled Membranes, type of cleaning agent followed by cleaning with
Table 1 Common chemicals used for membrane cleaning the second chemical, can prove more effective in
and their cleaning mechanisms
achieving higher overall efciency. For example,
Chemical type (and acidic cleaners are often used in conjunction with
simple chemical Cleaning mechanisms and
examples) applications alkaline cleaners to chelate calcium and magne-
Caustic or alkaline Hydrolysing, cleavage of sium ions that would otherwise form insoluble
(NaOH, KOH) disulphide bonds for protein soap residuals. The acidication wash can also C
aggregates for cleaning of been applied to neutralize residual alkalinity and
protein foulants remove mineral deposits that may have formed
Saponicate fats and lipids
Dissolve inorganic and during the alkaline cleaning or alkaline based
organic foulants enzyme cleaning.
Acidic (HCl, H3PO4, Solubilization and chelation
nitric acid, citric acid) for scale compounds and
metal oxides Cleaning Conditions
Dissolving precipitates
Enzymatic cleaning Cleave specic peptide
(Protease, Lipase bounds in the proteins Operating chemical cleaning at appropriate con-
based commercial Operate at conditions when ditions such as circulation of the cleaning solution
enzymes) enzyme activity was highest over the membrane surface provides mixing and
Surfactants or wetting Lower the surface tension of shear force to the foulants from the membrane
agents (Anionic, cleaning solution and increase
cationic surfactants) solubility
surface to the bulk.
Displace foulants and prevent With the selected chemicals and cleaning pro-
redeposition tocols, parameters such as concentration of cleaning
Improve contact between agent, cleaning time, temperature, and hydrody-
cleaning chemical and
deposits
namic conditions during the cleaning process can
Oxidants (Free Oxidize foulant and increase all affect the overall cleaning efciency. Chemical
chlorine, NaOCl) hydrophilicity compatibility between the membrane and other l-
Sanitizers and Used at the end of cleaning ter components to the cleaning chemicals is also an
disinfectants cycle or as part of the storage important factor in selection of suitable cleaning
regime between uses
chemicals. Membrane systems made of high chem-
ical tolerances would allow greater freedom in
cleaning include alkaline or caustic chemicals selecting the composition, strength of cleaning solu-
such as sodium hydroxide used for hydrolysis tions as well as the conditions for cleaning. On the
and solubilize foulants, acids such as citric acid other hand, milder cleaning conditions would have
and nitric acid with solubilization function. Sur- much less detrimental effect on the membrane
factants and detergents are also frequently used integrity and life span as well as being more envi-
for emulsifying, dispersion, and surface condition ronmentally benign. Cleaning at higher temperature
purposes as shown in Table 1. Many membrane could lead to higher cleaning reaction and lower
companies provide cleaning advice and proce- viscosity with high cleaning efciency. However,
dures to their membrane products, some also optimized cleaning condition should be considered
have their own proprietary products. Many with the following factors in mind.
cleaning products are formulated with a mixture
of chemicals that could include alkaline Running costs in cleaning chemicals,
(or acidic), surfactants, and disinfectants. Down-time,
Given their specic functions in cleaning Membrane integrity over long period of time and
mechanisms and the complexities of the feed and Environmental impact for discharge of
foulants, sequential cleaning, cleaning with one cleaning chemicals.
416 Cleaning Effectiveness
Summary membrane life while also minimizing expense

and time.
Membrane cleaning is an essential part of mem- Factors that inuence the selection of mem-
brane operation system. Design and selection of brane cleaning regime are listed in Table 1. Nor-
cleaning protocol should start with the identication mally the chemical and temperature tolerance of
of potential foulants and the understanding of the membranes will be recommended or can be
membrane fouling process and mechanisms, obtained from the membrane manufacturer. Inor-
followed by selection of appropriate chemical ganic membranes have a much higher tolerance of
agents to the specic foulants. An effective cleaning harsh chemicals and a higher temperature range
process should start with proper draining of the feed than polymeric membranes.
and rinsing of the system prior to application of The application with specic feed solution and
chemicals. Consideration of clean process design operation conditions will dictate the foulants on the
should include the membrane separation perfor- membrane and the strength of bonding between the
mance over many repeated cycles of fouling and foulant and foulant and the membrane. One exam-
cleaning, the effect on membrane integrity as well ple is membranes used in dairy application where
as overall operation and disposal costs. the milk concentration would experience mem-
The following reading materials are brane fouling with protein as well as lactose and
recommended for more information. minerals. The cleaning of membranes used in dairy
applications normally involves caustic cleaner in
daily basis. On the other hand, membranes used in
reverse osmosis (RO) would experience fouling by
References scale of calcium, ions, colloidal deposits, organic
matter, and biological materials. The frequency of
Li H, Chen V (2010) Chapter 10 Membrane fouling and
cleaning in food and bioprocessing. In: Cui ZF, RO membrane cleaning ranges from once in a few
Muralidhara HS (eds) Membrane technology. month up to once a year.
Butterworth-Heinemann, Oxford, pp 213254 An effective cleaning process should remove
Trgrdh G (1989) Membrane cleaning. Desalination the foulants/deposits and other contaminants from
71(3):325335
Wilson DI (2005) Challenges in cleaning: recent develop- the membrane surface and cassette feed channels
ments and future prospects. Heat Transfer Eng and restore the membrane separation performance
26(1):5159 to a predictable, consistent level. This is usually
evaluated by measurable parameters such as:
Cleaning Effectiveness Cleaning Effectiveness, Table 1 Factors and consider-

ations for selection of membrane cleaning protocol
Hongyu Li Factors Consideration for clean

UNESCO Centre for Membrane Science and Membrane material Tolerance to chemical and
(polymeric, inorganic) temperature could affect
Technology, The University of New South Wales, membrane lifetime
Sydney, Australia Feed composition and the Specic chemicals and
degree of fouling regime to specic foulants
and degree of fouling
Cleaning effectiveness refers to how useful a Frequency of cleaning
relates to costs and
cleaning regime is in regard to a specic applica- membrane lifetime
tion. An effective cleaning protocol to one appli- Membrane module Amount of cleaning
cation may not be very useful in another situation. conguration (at sheet, solution required and
An effective cleaning regime must accommodate hollow ber, spiral wound) operation conditions of
a variety of factors in order to maximize cleaning
Cleaning Efficiency 417
(a) Clean water permeability separation performance. Cleaning efciency is

This can easily be estimated by measuring commonly measured by either the membrane ux
membrane ux of clean water at dened oper- recovery or resistance removal (Li and Chen 2010).
ation conditions, for example, at xed pres- The ux recovery is estimated by the compar-
sure and temperature. Low recovery of clean ison of membrane water uxes before fouling and
water permeability indicates an insufcient cleaning and water ux after cleaning. The ux
cleaning process. recovery can be estimated as C
J wc
However, this parameter is not always reliable Flux Recovery 100 % (1)
for evaluation of cleaning effectiveness. A good J wi
clean water permeability recovery does not always where Jwc is the water ux after cleaning and Jwi is
represent a total removal of foulants from mem- the water ux before the membrane was subjected
brane. In some situations, a partial removal of the to fouling.
residual foulant does not contribute signicantly to The resistance removal can be estimated as
clean water permeability, and the separation perfor-
mance in subsequent operations could be exasper- Rr Rc
Resistance Removal 100 % (2)
ated due to the existing residual foulant. Rr
where Rr is the membrane resistance before
(b) Analysis of ush solution after cleaning can cleaning and Rc is the membrane resistance after
also be used to indicate the presence of foulant cleaning. The membrane resistance is estimated
in the system after cleaning. with clean water ux of membranes at different
stages of operation with
Apart from these methods commonly applied
in cleaning-in-place (CIP) applications that do not DP
affect continued membrane usage and function- R (3)
mJ
ing, other techniques that require removal of
membrane from module and sometimes destruc- where J is the water ux, DP is the transmembrane
tion of membrane can also be applied for diagnos- pressure of water ux measurement, and m is the
tic procedure, including: viscosity of water (at the temperature of water at
the ux measurement).
(c) Analysis of membrane surface properties, This type of assessment involves measure-
such as surface charge, contact angle, and ments of membrane uxes prior the ltration pro-
elemental analysis of species as well as SEM cess, the initial water ux (Jwi), membrane ux at
imaging on the membrane, which can be used the end of ltration process (Jwf), membrane ux
to observe the extend of the amount of resid- after rinsing (Jwr), and membrane ux after
ual foulants on the membrane. cleaning (Jwc).
Analytical methods can be used to evaluate the
changes of membrane surface properties such as
Cleaning Efficiency charge and contact angle before and after
cleaning. Infrared and X-ray photoelectron spec-
Hongyu Li troscopy of the cleaned membrane has also been
UNESCO Centre for Membrane Science and used in identifying remaining foulants on the
Technology, The University of New South Wales, cleaned membrane.
Sydney, Australia A cleaned membrane subject to extensive
extraction methods can reveal not only the amount
of residual foulant on the membrane but also the
Cleaning efciency measures how well the composition foulants left on the membrane after
cleaning process is in recovery of the membrane cleaning. This information can be very useful to
418 Cleaning-in-Place (CIP) Systems
formulate sequential cleaning strategies by using References

one agent to clean one part of the foulant and a
second, more effective, agent in removing the Li H, Chen V (2010) Chapter 10: Membrane fouling and
cleaning in food and bioprocessing. In: Cui ZF,
remaining of the residual.
Muralidhara HS (eds) Membrane technology.
One example is the cleaning of membrane used Butterworth-Heinemann, Oxford, pp 213254
in whey protein ltration. Through analysis of Norazman N, Wu W, Li H, Wasinger V, Zhang H, Chen
foulants extracted from the membranes V (2013) Evaluation of chemical cleaning of UF mem-
branes fouled with whey protein isolates via analysis of
before and after cleaning with gel electrophoresis
residual protein components on membranes surface.
(SDS-PAGE) technique for protein composition, Sep Purif Technol 103:241250
it was revealed that the high molecular
weight components in the foulant layer could
be more easily cleaned than the smaller
components. Sequential cleaning using NaOH
for removal of large protein followed by Cleaning-in-Place (CIP) Systems
HCl solution for cleaning of smaller molecules
has shown consistent performance restoration Hongyu Li
over repeated cleaning cycles (Norazman UNESCO Centre for Membrane Science and
et al. 2013). Technology, The University of New South Wales,
Sydney, Australia
Parameters Affect Cleaning Efficiency

Cleaning-in-place (CIP) is a method of cleaning
Both the properties of cleaning solution closed systems in industrial settings, such
(concentration, pH) and the operational condition as interior surfaces of pipes, vessels, heat
(pressure, temperature, velocity, and cleaning exchangers, process equipment, lters, mem-
time) could affect the outcomes of cleaning. brane systems, and their associated ttings with-
Generally speaking, conditions that favor the out disassembling.
contact between the cleaning chemical and the Opposed to cleaning-in-place is dissembled
reaction rate would be benecial for efcient cleaning, which refers to a method of cleaning
cleaning. For example, cleaning solution of higher that works by disassembling the closed systems
concentration would have a higher rate of reaction and cleaning individual components (often man-
with foulants, higher temperature improves reac- ually) before assembly for continued usage.
tion rate, and cleaning time relates to contact and Many laboratory facilities and small-scale sys-
reaction time. High circulation rates also lead to tems are commonly cleaned and disinfected
more turbulence and high shear rate which favors this way.
removal of dislodged foulants from the mem- For industry-scaled applications, with sys-
brane. Low pressure could avoid redeposition of tems that require frequent internal cleaning and
foulant. high level of hygiene, such as dairy processing,
However, practical experience suggests that beverage processing, processing of food,
for specic cleaning applications, the optimum bioprocesses, and pharmaceutical applications,
cleaning conditions should be determined on con- automated cleaning-in-place provides great ben-
siderations of limiting conditions of all those fac- ets of faster, safer, and reliable cleaning pro-
tors. For example, the integrity of membrane cesses which is very important for industrial
material and hosing material could be considerations.
compromised at temperature above their tolerance The purpose of membrane cleaning is to
level. Extensive circulation at high ow rate remove foulants and deposits on the membrane
relates to higher energy cost. surface and membrane pore structure to restore the
Cleaning-in-Place (CIP) Systems 419
membrane separation capabilities that had been Common cleaning agents used by industry for
compromised by membrane fouling. CIP include acids, bases, enzymes, surfactants,
Many membrane foulants behave as cohesive disinfectants, and steam and gas sterilization
solids ranging from soft macromolecular gels to (Mulder 1996), depending on the feed composi-
hard mineral scales on the surface. Breakdown of tion and severity of fouling.
these foulants may include dissolution or solubi- Alkaline cleaners, particularly sodium hydrox-
lization, disruption of cohesive interactions ide and potassium hydroxide, with their ability to C
between aggregates within the deposit, or disrup- dissolve protein molecules, fats, and other precip-
tion of the adhesive forces that bind foulant parti- itated materials, have been used to clean mem-
cles/molecules to the surface. branes fouled with proteins.
The common practice of cleaning-in-place of Acidic cleaners include nitric acid,
membrane system often involves a combination hydrochloric acid, and phosphoric acid, and better
of hydraulic and chemical cleaning in three dis- cleaning results with HCl for the
tinctive steps: membrane fouled with milk components have
been reported. Acidic cleaners have potential for
1. Draining before cleaning. The membrane opera- dissolving precipitates formed during the cleaning
tion (e.g., ltration of milk) is stopped and the procedure.
transmembrane pressure is released to allow Enzymatic cleaners have the unique benets of
relaxation and redistribution of the foulant into being biodegradable and more environmentally
the feed stream. The feed solution is then drained. friendly. Enzymatic cleaners possess great
2. Rinsing. This involves replacing feed with potential to remove proteinaceous matter from
clean water in circulation at low pressure to the membrane. Proteases are the most commonly
ush and remove loose foulants from the mem- used enzymatic cleaners with their function of
brane surface. hydrolyzing peptide bonds in the proteins
3. Chemical cleaning. This introduces the which leads to the disassembly of protein
cleaning solution into the membrane system, (Petrus et al 2008). However, a major
allowing circulation at low pressure in order to concern associated with enzymatic cleaning is
break down foulant structure and increasing the possibility of membrane fouling by the
solubilization of foulant species from the mem- enzyme molecules employed in the cleaning
brane surface. process.
Surfactants can be used to improve the
Selection of chemical cleaning agent should contact between the cleaning chemicals and the
consider: foulant, thus improving the reaction rate
(Trgrdh 1989).
Compatibility between the cleaning chemical Quite often, better cleaning performance can
and the membrane material and tting be achieved by combining cleaning agents or
components. sequential cleaning.
The cleaning agent should be safe and does not
pose as potential contaminants to the product.
Disposal of the cleaning chemicals.
References
Selection of cleaning process operation condi-
tion should consider: ogy, 2nd edn. Kluwer, Dordrecht
Petrus HB, Li H, Chen V, Norazman N (2008) Enzymatic
Overall downtime from production cleaning of ultraltration membranes fouled by protein
mixture solutions. J Membr Sci 325:783792
Temperature tolerance of membrane material
Trgrdh G (1989) Membrane cleaning. Desalination
Cost of energy for cleaning 71:325335
420 Closed-Loop Cascade Diafiltration
Closed-Loop Cascade Diafiltration CO Inhibition to Pd-Based Membrane
Greg Foley Inhibition of Hydrogen Permeation Through

School of Biotechnology, Dublin City University, Pd-Based Membranes by CO
Dublin, Ireland
Closed-loop cascade dialtration (Charcosset 2012)

CO Selective Oxidation
is used when the product to be recovered appears in
the permeate of the rst stage of a two-stage process.
Giuseppe Barbieri
It is similar to open-loop cascade dialtration but
differs in that the permeate from the second stage is
recycled and used as the diluant of the rst stage.
Italy
A schematic is shown below (Fig. 1).
The obvious advantage of this approach over the
open-loop conguration is the reduced requirement
Hydrogen purication is a critical technology for
for a fresh diluant. In principle, the product should
many chemical and petrochemical processes and
have a low rejection coefcient in the rst stage and
power generation by PEM-FCs (proton exchange
a high rejection coefcient in the second stage.
membrane fuel cells) as well. Carbon monoxide
is, in fact, a poisoning or inhibitor for most cata-
References
lysts such as the Fe3O4 one involved in ammonia
Charcosset C (2012) Membrane processes in biotechnol-
synthesis or the platinum-based one used at a low
ogy and pharmaceutics, 1st edn. Elsevier, Boston temperature in PEM-FCs. In the latter case, in
order to avoid cell performance degradation, car-
bon monoxide concentration must be kept below
Co (II) Ions Separation by Supported 10 ppm in the hydrogen feed streams.
Liquid Membranes CO selective or preferential oxidation is the
primary method used for hydrogen deep purica-
Cobalt Removal and Recovery by Supported tion (Avgouropoulos and Ioannides 2003), and
Liquid Membranes different research groups (Manasilp and Gulari
Closed-Loop Cascade
Diafiltration,
Fig. 1 Closed-Loop
Cascade Dialtration
Permeate Permeate
Diafiltration Ultrafiltration
CO Selective Oxidation 421
2002; Souza et al. 2007; Bissett and Oh 2005; increased heat transfer. Interesting new engineer-
Wootsch et al. 2004; Sirijaruphan et al. 2004; ing devices such as catalytic membranes and
Kotobuki et al. 2005) work on this topic. The catalytic membrane reactors were investigated
involved reactions are: in a few papers for CO selective oxidation
(Hasegawa et al. 2001, 2002; Bernardo
2CO O2 ! 2CO2 et al. 2006). The concept of using Pt in zeolite
2H2 O2 ! 2H2 O catalysts was applied for this reaction by C
Watanabe et al. (1995; Igarashi et al. 1997). In a
Therefore, a catalyst active for CO oxidation recent paper (Kotobuki et al. 2005), a signicant
in the presence of high concentrations of H2, improvement of CO conversion and selectivity
CO2, and steam is required to avoid H2 con- with high Pt dispersion in the zeolite (ZSM-5)
sumption, nor should it have a strong preference to pores was demonstrated. Zeolites play host to
the water-gas shift reaction since in these condi- noble-metal clusters for numerous catalytic appli-
tions of temperature and gas composition, the cations, and well-dened clusters of narrow size
equilibrium gives ca. 0.52 %mole of CO. distribution down to 1 nm have been obtained
Both alumina-supported Pt and Au catalysts under suitable conditions (Pan et al. 1990; de
have been found to be active for CO Selox Graaf et al. 2001). Investigations on the use of
(Luengnaruemitchai et al. 2004). Pt-based catalysts large-pore diameter zeolite such as Pt-loaded
require high temperatures (about 170 C) and a FAU-type membranes (Bernardo et al. 2006)
high feed molar ratio O2/CO (about 1.5) for com- show also the possibility of collocating a mem-
plete CO depletion, with a corresponding loss of brane reactor CO selective oxidation stage
selectivity (typically 34 %). Water vapor has a (operating at 200220 C and with a feed ratio
positive effect on these catalysts, increasing the O2/CO = 1) after the low-temperature water-gas
conversion without changing the selectivity; shift reactor in a reforming plant. One of the nd-
instead CO2 has a detrimental effect ings of this study was the possibility of improving
(Avgouropoulos et al. 2002). Au-based catalysts CO removal by increasing the operating pressure.
are more active than platinum catalysts at relatively A zeolite catalytic membrane provides catalyti-
low temperatures (<120 C) but not so resistant cally active and nano-sized particles entrapped in a
toward deactivation by CO2 and H2O. Promoted thin zeolite layer (a few microns thick), resulting in:
Pt/Al2O3 catalysts (e.g., with Ru or Rh) have also
been investigated. Values of ca. 15 ppm of CO in Efcient reactant/catalytic phase contact,
the reactor outlet are reported for a microreactor reducing bypassing and misdistribution gener-
(PtRu/a-Al2O3 catalyst) operated with a O2/CO ally shown in a packed bed
feed molar ratio of 1.5 and 4 (Delsman et al. 2004; Short and narrow distributed contact time
Cominos et al. 2005). Whereas, values of CO less Hot spot control that is benecial since a local
than 10 ppm were reported for a xed bed reactor temperature increase can lead to promotion of
using a O2/CO feed molar ratio of 2.5 (Dudeld reverse water-gas shift reaction (Ouyang and
et al. 2000). Other interesting active and very selec- Besser 2005; Ouyang et al. 2005)
tive catalysts are CuOCeO2 mixed oxides
(Avgouropoulos and Ioannides 2006; Zou With this aim, catalytic Pt-loaded zeolite mem-
et al. 2006); however, the presence of H2O and branes were utilized as an effective device for CO
CO2 in the feed causes a signicant decrease in selective oxidation in a single-stage membrane
their activity (Avgouropoulos et al. 2005). reactor operating in a continuous ow-through
Apart from the improvement of the catalyst, conguration at ca. 200 C and a low feed molar
new reactor concepts are being studied. Mono- ratio O2/CO = 1 and 1.5 (Bernardo et al. 2008).
liths present different advantages over particulate This new model reactor design leads to improved
catalysts, including pressure drop reduction, ease and really interesting performance in terms of CO
of handling, structural robustness, and possibly conversion and limited hydrogen consumption.
422 Coacervation
References selective oxidation over platinum catalyst supported

on zeolite. Appl Catal A 159:159169
Avgouropoulos G, Ioannides T (2003) Selective CO oxi- Kotobuki M, Watanabe A, Uchida H, Yamashita H,
dation over CuO-CeO2 catalysts prepared via the Watanabe M (2005) Development of Pt/ZSM5 catalyst
ureanitrate combustion method. Appl Catal A with high CO selectivity for preferential oxidation of
224(1):155167 carbon monoxide in a reformed gas. Chem Lett
Avgouropoulos G, Ioannides T (2006) Effect of synthesis 34(6):866967
parameters on catalytic properties of CuO-CeO2. Appl Luengnaruemitchai A, Osuwan S, Gulari E (2004) Selec-
Catal B 67:111 tive catalytic oxidation of CO in the presence of H2
Avgouropoulos G, Ioannides T, Papadopoulou C, Batista J, over gold catalyst. Int J Hydrog Energy 29:429435
Hocevar S, Matralis HK (2002) A comparative study of Manasilp A, Gulari E (2002) Selective CO oxidation over
Pt/g-Al2O3, Au/-Fe2O3 and CuO-CeO2 catalysts for the Pt/Al2O3 catalyst. Appl Catal Environ 37:1725
selective oxidation of carbon monoxide in excess Ouyang X, Besser RS (2005) Effect of reactor heat transfer
hydrogen. Catal Today 75(14):157167 limitations on CO preferential oxidation. J Power
Avgouropoulos G, Ioannides T, Matralis H (2005) Inu- Sources 141:3946
ence of the preparation method on the performance of Ouyang X, Bednarova L, Besser RS, Ho P (2005) Prefer-
CuO-CeO2 catalysts for the selective oxidation of ential oxidation (PrOx) in a thin-lm catalytic
CO. Appl Catal B 56:8793 microreactor: advantages and limitations. AIChE
Bernardo P, Algieri C, Barbieri G, Drioli E (2006) Catalytic J 51(6):17581771
(PtY) membranes for the purication of H2rich Pan M, Cowley JM, Chan IY (1990) HREM imaging of
streams. Catal Today 118:9097 small pt clusters dispersed in Y-zeolites. Catal Lett
Bernardo P, Algieri C, Barbieri G, Drioli E (2008) Hydro- 5:112
gen purication from carbon monoxide by means of Sirijaruphan A, Goodwin JG Jr, Rice RW (2004) Investiga-
selective oxidation using zeolite membranes. Sep Purif tion of the initial rapid deactivation of platinum cata-
Technol 62:629635. doi:10.1016/j.seppur.2008.03. lysts during the selective oxidation of carbon
024 monoxide. J Catal 221:288293
Bissett EJ, Oh SH (2005) PrOx reactor model for fuel cell Souza MMVM, Ribeiro NFP, Schmal M (2007) Inuence
feedstream processing. Chem Eng Sci 60:47224735 of the support in selective CO oxidation on Pt catalysts
Cominos V, Hessel V, Hofmann C, Kolb G, Zapf R, for fuel cell applications. Int J Hydrog Energy
Ziogas A, Delsman ER, Schouten JC (2005) Selective 32:425429
oxidation of carbon monoxide in a hydrogenrich fuel Watanabe M, Uchida H, Igarashi H, Suzuki M (1995) Pt
cell feed using a catalyst coated microstructured reac- catalyst supported on zeolite for selective oxidation of
tor. Catal Today 110:140153 CO in reformed gases. Chem Lett 1:2122
de Graaf J, van Dillen AJ, de Jong KP, Koningsberger DC Wootsch A, Descorme C, Duprez D (2004) Preferential
(2001) Preparation of highly dispersed Pt particles in oxidation of carbon monoxide in the presence of hydro-
zeolite Y with a narrow particle size distribution: char- gen (PROX) over ceriazirconia and alumina-
acterization by hydrogen chemisorption, TEM, EXAFS supported Pt catalysts. J Catal 225:259266
spectroscopy, and particle modelling. J Catal Zou H, Dong X, Lin W (2006) Selective CO oxidation in
203:307321 hydrogen-rich gas over CuO/CeO2 catalysts. Appl Surf
Delsman ER, De Croon MHJM, Pierik A, Kramer GJ, Sci 253:28932898
Cobden PD, Hofmann CH, Cominos V, Schouten JC
(2004) Design and operation of a preferential oxidation
microdevice for a portable fuel processor. Chem Eng
Sci 59(2004):47954802
Dudeld CD, Chen R, Adcock PL (2000) A compact CO Coacervation
selective oxidation reactor for solid polymer fuel cell
powered vehicle application. J Power Sources
86:214222 Emma Piacentini
Hasegawa Y, Kusakabe K, Morooka S (2001) Selective Institute on Membrane Technology, National
oxidation of carbon monoxide in hydrogenrich mix- Research Council, ITM-CNR, Rende, Italy
tures by permeation through a platinumloaded Ytype
zeolite membrane. J Membr Sci 190:18
Hasegawa Y, Ueda A, Kusakabe K, Morooka S (2002)
Oxidation of CO in hydrogenrich gas using a novel Coacervation is a chemical method for producing
membrane combined with a microporous SiO2 layer polymer droplets in suspension based on the sepa-
and a metalloaded Al2O3 layer. Appl Catal
ration of two liquid phases into one concentrated
A 225:109115
Igarashi H, Uchida H, Suzuki M, Sasaki Y, Watanabe colloidal phase, being the coacervate, and another
M (1997) Removal of CO from H2-rich fuels by highly dilute colloidal phase (Fig. 1). The phase
Coacervation 423
Coacervation, Fig. 1 Coacervate formation. (a) Intrapolymer complex, (b) soluble aggregate of intrapolymer com-
plexes, (c) coacervate with dense and dilute domains, (d) coacevates picture
Coacervation, Fig. 2 Complex coacervate gels formed by mixing a polyanion polymer with a polycation polymer
separation of a single polyelectrolyte or a mixture of

polyelectrolytes from a solution and deposition of
the agglomerated colloidal particles (i.e., the matrix
material) on an immiscible active core results in the
formation of a simple coacervate or a complex
coacervate, respectively (De Kruif et al. 2004).
Phase separation in simple coacervation is
brought about by addition of a salt, pH, or tem-
perature change in the polymeric solution (such
as alginate and calcium) while in complex coac-
ervation is brought about by anioncation inter-
actions (such as gelatine and gum Arabic or
chitosan and alginate). Complex coacervate gels
can be formed by mixing of a polyanion with a
polycation (Fig. 2). The underlying principle of Coacervation, Fig. 3 Microscopic view of the microcap-
sules obtained by complex coacervation
this method is that polymers with opposite
charges stick together and form soluble and
insoluble complexes depending on the concen- Proteins below its isoelectric point are positively
tration and pH of the respective solutions. One charged and likely to associate with anionic
such example is coacervating proteins with poly- hydrocolloids and form polyion complex hydro-
saccharides (Schmitt and Turgeon 2011). gel (complex coacervate).
424 Coagulation Bath
Coacervation uses the principle of difference compared to non-gellable hydrocolloids.

in ionic forces to cause the polymer(s) to form Gelication, in turn, is generally brought about
droplets and drop out of solution. A key to this is by lowering the temperature of the reaction mix-
knowledge of the isoelectric point (PI) of the ture below the gelling point of the gellable hydro-
polymers and adjusting the formulation accord- colloid. The emulsication step can be carried out
ingly. Electrostatic effects and other weak energy by homogenizing the oil with an aqueous sol of
interactions, especially hydrogen bonding, play a one colloid and mixing the emulsion with an
very important role during complex formation/ aqueous sol of another colloid, or the two sols
coacervation between proteins and polysaccha- may be made and mixed and the oil emulsied
rides. Hydrophobic interactions can also make a therein. Complex coacervation occurred by low-
signicant contribution to formation of com- ering the pH until a pH is low enough to enough
plexes and coacervates between oppositely opposite electrical charges and form polymer
charged biopolymers. Several physicochemical complex. The coacervate phase is uid and dis-
parameters inuencing the overall and local tributes over the oil surface forming the shell. The
charge of the protein and the polysaccharide wall hardening renders the process irreversible
play an important role in the control of the phe- and making the resulting microcapsules insoluble
nomenon such as pH, ionic strength, protein to in water, resistant to mechanical stress and to heat
polysaccharide ratio, and total biopolymer con- exposure.
centration. Some other parameters such as the
biopolymer molecular weight and exibility, the
charge density, the stirring, the pressure, or the Cross-References
temperature have been shown to also inuence
complex formation. Membrane Emulsication in Phase Separation
Complex coacervation between oppositely for Microcapsule Preparation
charged proteins and polysaccharides is a well-
known oil encapsulation technology (Fig. 3).
References
Typical steps of encapsulation by complex
coacervation generally involve: De Kruif CG, Weinbreck F, De Vries R (2004) Complex
coacervation of proteins and anionic polysaccharides.
1. Emulsication (formation of the core) of a Curr Opin Colloid Interface Sci 9(5):340349
generally hydrophobic material in a solution Schmitt C, Turgeon SL (2011) Protein/polysaccharide
complexes and coacervates in food systems. Adv Col-
comprising hydrocolloids loid Interface Sci 167(12):6370
2. Coacervation (formation of the shell) implying
the formation of a coacervate phase
3. Wall hardening (usually achieved by cross-
linking the hydrocolloid forming the wall)
Coagulation Bath
The step of wall formation is generally driven
by the surface tension difference between the Silvia Simone
coacervate phase, the water, and the hydrophobic Institute on Membrane Technology, National
material. In most industrial coacervation pro- Research Council, ITM-CNR, Rende, Italy
cesses, one of the hydrocolloids used in coacerva-
tion processes is selected from gellable proteins.
These are easier to use and less prone to aggrega- Phase Separation (PS) or Phase Inversion (PI) is
tion after the formation of the wall when the among the principal techniques for membrane
temperature is below the gelling temperature, if preparation. The membrane matrix and the
Coagulation Bath 425
membrane pores are formed, respectively, from often described in terms of Hildebrands solubility
the polymer-rich and the polymer-lean phases parameters (d). In general, it is known that the
originated by the phase separation of an initially closer the values of the solubility parameters of
homogeneous polymeric dope. Immersion of a two species, the higher the afnity between the
cast (or spun) polymeric dope in a coagulation two. Therefore, the difference between the S and
bath is normally used in nonsolvent the NS solubility parameters (dS-NS) can be used
(or diffusion)-induced phase separation (NIPS or as a reference to estimate the S/NS exchange rate C
DIPS), also called immersion precipitation (IP), to during coagulation: it will be faster if they have
achieve such separation, often referred to as high mutual afnity, indicated by a small dS-NS.
demixing or precipitation as well. Upon immer- Furthermore, the difference between the P and NS
sion in the coagulation bath, the initial composi- parameters (dP-NS) can be used to infer about the
tion of the dope changes as the solvent diffuses in coagulation power of a nonsolvent: strong, or
the bath and is gradually replaced by the harsh, coagulants are characterized by a higher
nonsolvent. Polymer precipitation due to sol- dP-NS, while soft, or gentle, nonsolvents usually
vent/nonsolvent (S/NS) exchange is exploited display a smaller dP-NS. The Hildebrand solubility
both in at sheet and hollow ber preparation parameter can be calculated, according to the
via NIPS. However, for at sheet membranes, Hansens theory (Hansen 1967), as the square
the cast polymer lm is immersed in the coagula- root of the sum of the squares of three compo-
tion bath and phase inversion starts at the top nents: dd, representing the energy from dispersion
surface of the lm. Regarding hollow bers, pre- bonds; dp, the energy from dipolar intermolecular
pared via wet or dry/wet spinning, phase inversion forces; and dh, the energy from hydrogen bonds
takes place both at the inner and at the outer between molecules. The strength of different NS
surfaces. While the term coagulation bath is can be also evaluated and, eventually, compared,
normally used when referring to the media for by looking at the position of the binodal curve in
the coagulation of the outer surface, the inner the ternary phase diagram. In general, the curve
coagulant is often referred as bore uid. In any moves towards higher NS percentages when using
case, the solvent/nonsolvent exchange is at the a softer coagulant. On the contrary, the binodal
basis of the membrane preparation mechanism curve moves closer to the P-S axis, indicating less
via immersion precipitation. Phase separation NS tolerance, when using a strong coagulant. The
will occur only when the system composition, in temperature of the coagulation bath will inuence
terms of polymer/solvent/nonsolvent (P/S/NS) the interdiffusion rate between solvent and
concentrations, reaches the miscibility gap, nonsolvent, which will be faster at higher temper-
surrounded by the spinodal curve in the ternary atures. Polymer solubility in the S/NS mixture at
phase diagram (at a selected temperature). Mem- any time during phase inversion depends on tem-
brane morphology will be strongly affected by the perature, which affects the miscibility gap bound-
coagulation bath composition and temperature, ary position in the ternary phase diagram; this will
since they both inuence the solvent/nonsolvent mostly depend on the initial temperature of the
exchange rate and the polymer precipitation. In casting solution but is affected by the coagulation
particular, the exchange rate depends on the bath temperature as well. Another factor that is
mutual afnity between the solvent contained in known to affect the S/NS exchange rate is the
the dope and the nonsolvent in the bath as well as molecular weight, since bigger size normally hin-
on temperature. On the other hand, diverse ders diffusion. Solvent diffusion in the coagula-
nonsolvents usually display different coagulation tion bath can be delayed by acting on the
power; polymer precipitation is affected by the concentration gradient, i.e., by adding some sol-
nonsolvent nature and, obviously, by temperature. vent in the bath composition; this will also reduce
The mutual afnities between S/NS and P/NS are the difference between the solubility parameters
426 Coagulation Medium
of the polymer and the coagulant, which is not Definition

pure nonsolvent. The proper choice of the coagu-
lation conditions allows to tailor membrane mor- It is a liquid bath providing the precipitation of
phology. In general, it is known that membranes an extruded or cast polymer solution with
with open and porous structures can be obtained the resulting formation of a solid membrane
by fast demixing, while delayed demixing often matrix.
results in the formation of a dense skin, with
sponge-like or closed-cell morphology in the Coagulation medium is responsible for the phase
sublayer. Furthermore, fast demixing is one of inversion process, and most of the time, it is
the factors inducing macrovoids formation. For represented by water. Some organic solvents
instance, when the as-cast polymer lm is (such as ethanol or isopropanol) can be also
immersed in the coagulation bath, the polymer used to induce polymer precipitation, and even
precipitation in the top layer will be faster with a vapors and supercritical uids can be considered
strong coagulant, resulting in a porous skin; on the as coagulation media for the phase inversion
contrary, delayed demixing may result in the for- process.
mation of a dense skin. However, the S/NS In non-solvent-induced phase separation
exchange can be properly delayed, by introducing (NIPS), the polymer solution is cast upon a sup-
some solvent in the coagulation bath or by using a port and subsequently immersed into a coagula-
soft coagulant, as a viable strategy to reduce tion bath (generally water) where polymer
macrovoids formation; in fact, it is accepted that precipitation occurs. The choice of the appropriate
the nonsolvent inux is responsible for voids coagulation medium represents a signicant
growth during membrane formation. aspect in membrane preparation. The miscibility
between solvent (in which the polymer is
dissolved) and non-solvent can affect, in fact, the
nal structure of the membrane. The solvent
References contained in the cast lm exchanges with the
non-solvent contained in the coagulation medium
Hansen CM (1967) The three dimensional solubility-key to inducing polymer precipitation.
paint component afnities. 1. Solvent, plasticizers,
Variable amounts of solvents can be also added
polymers and resins. J Paint Technol 39:104
to the coagulation medium giving two contrary
effects:
Slow down the phase separation process leading

Coagulation Medium to the formation of membranes with a less
porous structure.
Francesco Galiano Decreasing polymer concentration at lm inter-
Institute on Membrane Technology, National face leading to a membrane with a more open
Research Council, ITM-CNR, Rende, Italy structure (Drioli and Giorno 2009).
Final membrane structure will, thus, depend on

Synonyms the miscibility between the two liquids and, there-
fore, on the mutual interactions established
Coagulation medium (also known as nonsolvent between them.
bath, coagulation bath, phase-inversion medium, Even the temperature of the coagulation
precipitating medium) medium, used to precipitate the casting solution,
Cobalt Removal and Recovery by Supported Liquid Membranes 427
plays an important role. Generally low tempera-

tures lead to the formation of membranes with a Cobalt Removal and Recovery by
more dense structure (delay demixing) that nd Supported Liquid Membranes
application in nanoltration, gas separation, or
reverse osmosis. Carlos Moreno
In vapor-induced phase separation (VIPS), the Department of Analytical Chemistry, Faculty of
coagulation medium is represented by a non- Marine and Environmental Sciences, University C
solvent in a vapor phase. The vapor is generally of Cdiz, Puerto Real, Spain
water, and the precipitation occurs due to the
afnity of the solvent (contained in the polymeric Synonyms
solution) for vapor (contained in the surrounding
environment). VIPS technique leads to the Co (II) ions separation by supported liquid
formation of membranes with a porous structure membranes
usually applied in microltration. Relative humid-
ity, temperature, and exposure time are the main Naturally found in some minerals as cobaltite,
variables affecting membrane morphology. erythrite, or smaltite, cobalt is an element with
Other possibilities are supercritical uids such different industrial applications, mainly in, i.e.,
as supercritical CO2 that can be used as coagula- electroplating and magnets or, as alloying ele-
tion medium. Supercritical CO2 can be used, in ment, for the fabrication of engines with high
fact, as environmental friendly non- resistance to temperature.
solvent leading to the formation of asymmetric Some cobalt salts have the ability of giving
membranes with the advantage of avoiding the blue color, and, then, they have been extensively
collapse of membrane structure due to absence used as pigment for ceramics, glass, inks, etc.
of a liquidliquid or a liquidvapor interface Supported liquid membranes have been used as a
(Reverchon and Cardea 2004). tool for the recovery and removal of cobalt from
different waste aqueous solutions. Typically, these
solutions contain a mixture of metals that may com-
plicate the selective separation of cobalt. Thus, in
References many cases, separation of cobalt by supported liquid
membranes has to be performed in the presence of
Drioli E, Giorno L (2009) Membrane operations: innova- metals such as copper, cadmium, chromium, and
tive separations and transformations. Wiley-VCH,
Weinheim, p 30, Chapter 2
specially nickel, which exhibits similar properties
Reverchon E, Cardea S (2004) Formation of cellulose than cobalt. For a good and selective separation of
acetate membranes using a supercritical uid assisted cobalt ions from the rest of metallic element sepa-
process. J Membr Sci 240:187195 ration, it is essential to control the chemical condi-
tions of the membrane separation process, mainly
the correct selection of the organic carrier and the
acidity of the aqueous source phase.
The mechanism of the separation process of
Coagulation Medium (also Known as cobalt by supported liquid membranes is based on
Nonsolvent Bath, Coagulation Bath, the complexation and de-complexation reactions of
Phase-Inversion Medium, cobalt ions present in the aqueous solution, with an
Precipitating Medium) extractant agent present in the organic liquid mem-
brane. Although many organic reagents have been
Coagulation Medium studied and applied at laboratory scale, the best
428 Cobalt Removal and Recovery with Strip Dispersion
P
O O CH3
H3C OH
H3C CH3
CH3 O CH3
P
H3C CH3
OH
Cobalt Removal and Recovery by Supported Liquid Membranes, Fig. 1 Chemical structure of bis(2-ethylhexyl)
hydrogen phosphate (1) and bis(2,4,4-trimethylpentyl) phosphinic acid (2)
results have been obtained when using organophos- dispersion (PEHFSD), dispersion supported liq-
phorus acidic compounds, such as bis(2-ethylhexyl) uid membrane (DSLM), or emulsion pertraction
hydrogen phosphate (1) or bis(2,4,4- (EP) due to the different congurations that have
trimethylpentyl) phosphinic acid (2) (Fig. 1). been used, mainly at-sheet and hollow-ber
supported liquid membranes congurations. It
has been applied to the removal and recovery of
different organic species (organic acids, phenols,
Cobalt Removal and Recovery antibiotics, and alkaloids) and different metals
with Strip Dispersion (Co(II), Cr(VI), Cr(III), Cu(II), Au(III), etc.).
Figure 1 shows a schematic representation of the
Gerardo Len
technique.
Departamento de Ingeniera Qumica y
As shown in this gure, the aqueous product
Ambiental, Universidad Politcnica de Cartagena,
phase, containing the stripping agent, is dispersed
Cartagena, Spain
in the membrane organic phase, containing the
carrier, and the water in oil dispersion formed is
The commercial success of supported liquid added (at-sheet conguration) or is passed
membrane has been seen to be limited due to the (hollow-ber conguration) onto one side of the
lack of long-term stability. Six mechanisms have membrane support. The organic phase of the dis-
been identied for supported liquid membrane persion becomes imbedded in the pores of the
instability (Kemperman et al. 1998). The two microporous hydrophobic support of the at or
considered most important are emulsication of hollow-ber conguration, forming a stable SL-
the liquid membrane phase and osmotic M. The constant supply of organic membrane
pressure difference over the membrane. In order solution into the pores of the membrane support
to increase the stability of a supported liquid ensures stable and continuous operation. The
membrane (increasing simultaneously the ef- aqueous feed solution is added or passed on the
ciency of the separation process) by preventing other side of the membrane support, and the target
the loss of the liquid membrane process, the use of specie is extracted into the organic solution by the
the supported liquid membrane with strip disper- selective carrier and then stripped by the aqueous
sion has been described. This technique, rst strip solution (Fig. 1). For nal recovery of the
described by Ho (2001a), incorporates concepts target specie, the mixer is turned off or a settler is
of supported liquid membranes and of emulsion used for the loaded strip dispersion, and the dis-
liquid membranes. It is also known as persion quickly separates. Supported liquid mem-
pseudoemulsion-based hollow-ber strip branes with strip dispersion processes offer
Cold Plasma 429
Feed Phase Membrane phase Dispersion phase
Membrane phase
+2 CoR2(HR)2 2H+ 2H+
Co
2H+
Co+2
2(HR)2 Co+2 Co+2
Stripping phase C
+ (HR)2 loss
2H 2H+
Co+2
2H+
(HR)2 renewal
Co+2 Co+2 2H+ Co+2
+2
H+ H+ Co
Cobalt Removal and Recovery with Strip Dispersion, membranes with strip dispersion using an organophospho-
Fig. 1 Schematic representation of the carrier mediate rus acid as carrier
counter-transport of cobalt (II) through supported liquid
several advantages compared with other liquid References

membrane congurations, including the follow-
ing (Klaasen et al. 2005): no emulsion formation Alguacil FJ, Garcia-Diaz I, Lopez F, Sastre AM (2011) Cobalt
(II) membrane-extraction by DP-8R/Exxsol D100 using
in the water phase, exible process parameters,
pseudo-emulsion based hollow ber strip dispersion
phase separation is not necessary, compact modular (PEHFSD) processing. Sep Purif Tech 80:467472
equipment, low energy consumption, and large Ho SWS (2001a) Combined supported liquid membrane/
specic surface areas due to hollow-ber strip dispersion process for the removal and recovery of
radionuclides and metals. US Patent 6,328,782
contactors. This technique has been used for cobalt
Ho SWS (2001b) Combined supported liquid membrane/
removal and recovery (Ho 2001b; Alguacil strip dispersion process for the removal and recovery of
et al. 2011). Ho described several experiments metals. US Patent 6,350,419
using polypropylene as polymeric microporous Kemperman AJB, Rolevink HHM, Van Den Boomgaard T,
Strathmann H (1998) Stabilization of supported liquid
support, in both at-sheet and hollow-ber cong-
membranes by interfacial polymerization top layers.
urations, di(2,4,4-trimethylpentyl)dithiophosphinic J Membr Sci 138:4355
acid (Cyanex 301) as carrier, Isopar L and Klaassen R, Feron PHM, Jansen AE (2005) Membrane
dodecanol as organic diluent, and sulfuric acid as contactors in industrial applications. Chem Eng Res
Des 83:234246
stripping agent. Alguacil et al. removed cobalt
(II) from acidic sulfate solutions by using polypro-
pylene as polymeric microporous support in
hollow-ber conguration, di(2-ethylhexyl) Cold Plasma
phosphoric acid (DP-8R) as carrier, Exxsol D100
as organic diluent, and sulfuric acid as stripping Antonella Milella and Fabio Palumbo
agent. Department of Chemistry, University of Bari
Aldo Moro, Institute of Nanotechnology
(CNR-NANOTEC), Bari, Italy
Cross-References
Plasma can be dened as a partially ionized gas
Emulsion Liquid Membrane (ELM) which contains approximately equal numbers of
Hollow Fiber positive and negative particles (Chapman 1980;
430 Cold Plasma
Chen 1974; Grill 1994; Thornthon and Greene communications. At this frequency, only electrons
1994; dAgostino 1990; Milella 2008). can follow the instantaneous variations of the
Plasmas can be roughly classied into hot applied electric eld, while ions are driven by an
plasmas (also dened as near-equilibrium average potential.
plasmas) and cold plasmas (or nonequilibrium At the steady state, the total ion and electron
plasmas). Hot plasmas are characterized by very charge ow to a given electrode during an
high temperatures of electrons and heavy parti- RF cycle must balance to zero and a self-bias
cles (atoms, molecules, or ions) and they are that is negative with respect to the plasma poten-
almost fully ionized. In cold plasmas, instead, tial develops. If the surface is electrically isolated,
the electron temperature is relatively high the potential at the steady state is named oating
(110 eV), while the translational energy of potential and it is also negative with respect to the
heavy particles remains very low, with tempera- plasma potential. As a consequence, each surface
ture close to the room one and the ionization immersed in a plasma will be bombarded by ions.
degree is typically low (around 104 106). It should be also highlighted that cold
When an electric eld is applied to a gas, energy plasmas can be also ignited at atmospheric pressure,
is transferred more efciently to the free elec- and since the late 1980s, an increasing interest has
trons naturally present in the gas than to the grown in the application of such plasmas (e.g.,
ions. The accelerated electrons, then, transfer dielectric barrier discharge, corona, and atmospheric
the energy to heavier particles through elastic jet) to material processing (Kogelschatz 2003).
and inelastic collisions. At low pressure the colli- Cold plasmas can modify the surface of mate-
sion frequency is very low; moreover, because of rial by depositing thin lms (plasma-enhanced
the large mass difference between electrons and chemical vapor deposition), by grafting specic
heavy particles, the transfer of kinetic energy is chemical functionalities (plasma treatment), or by
inefcient. Hence, in this nonequilibrium state, ablating materials (plasma etching).
the gas temperature cannot increase, while elec- Membranes can be processed by means of cold
trons gain enough kinetic energy to promote exci- plasmas to activate the surface by introducing
tation, ionization, fragmentation, and formation of specic chemical groups such as hydroxyl, car-
reactive species by inelastic collisions with heavy boxyl, and amino moieties. This can be accom-
particles. In this scenario, energetic electrons can plished by feeding the plasma non-polymerizable
produce high temperature chemistry in a gas at low gases/vapor such as O2, H2O, NH3, and N2/H2.
temperature, and this allows the treatment of ther- Alternatively, functionalized coatings can be
mally sensitive materials like polymers. deposited for the same purposes (e.g., acrylic
Since electrons are much faster than other par- acid or allylamine based). These treatments
ticles, every surface immersed in the plasma allow to tune surface properties of membranes,
immediately develops a negative charge which notably wettability and molecular adsorption, or
accelerates ions. As a matter of fact, the plasma to serve as anchor for specic biomolecules or
potential is always the most positive one. catalysts (Favia et al. 2008; Fontananova
A low-pressure cold plasma apparatus typically et al. 2006; Roualdes et al. 2010).
consists of a vacuum chamber, a pumping unit, a
gas feeding system and gas controllers, pressure
gauge, a power supply, and a power transfer device.
Electrodes can be driven both by AC and by DC References
power supply. AC discharges, commonly run in the
radiofrequency (RF) and microwave (MW) range, Chapman B (1980) Glow discharge processes. Sputtering
are more stable and suitable for insulators (e.g., and plasma etching. Wiley, New York
Chen FF (1974) Introduction to plasma physics. Plenum
ceramic and polymers) processing. In case of RF
Press, New York
glow discharges, the operating frequency is gener- dAgostino R (1990) Plasma deposition, treatments and
ally 13.56 MHz in order to avoid interferences with etching of polymers. Academic, San Diego
ComplexationUltrafiltration Process for Heavy Metal Removal from Effluents 431
Favia P, Sardella E, Lopez LC, Laera S, Milella A,

Pistillo B, Intranuovo F, Nardulli M, Gristina R, ComplexationUltrafiltration
dAgostino R (2008) Plasma assisted surface modica-
tion processes for biomedical materials and devices. In: Process for Heavy Metal Removal
Gueri S, Fridman A (eds) Plasma assisted decontam- from Effluents
ination of biological and chemical agents. Springer,
Dordrecht, pp 203226 Mohamed Abou El-Fetouh Barakat
Fontananova E, Donato L, Drioli E, Lopez LC, Favia P,
dAgostino R (2006) Heterogenization of Department of Environmental Sciences, Faculty C
polyoxometalates on the surface of plasma- of Meteorology, Environment and Arid Land
modied polymeric membranes. Chem Mater Agriculture, King Abdulaziz University, Jeddah,
18:15611568 Saudi Arabia
Grill A (1994) Cold plasma in materials fabrication-
from fundamentals to applications. IEEE Press, Central Metallurgical R & D Institute, Helwan,
New York Cairo, Egypt
Kogelschatz U (2003) Dielectric-barrier discharges: their
history, discharge physics, and industrial applications.
Plasma Chem Plasma Proc 3:146
Milella A (2008) Plasma processing of polymers. In: Metal rejection in aqueous solution containing Ni
Encyclopedia of polymer science and technology. (II), Cu(II), and Cr(III) ions was evaluated
Wiley (Barakat and Schmidt 2010) using complexatio-
Roualdes S, Rouessac V, Durand J (2010) Plasma membranes. nultraltration. The ltration experiments
In: Drioli E, Giorno L (eds) Comprehensive membrane
science and engineering. Elsevier, Kidlington, pp 159197 were performed with ultraltration membrane
Thornthon JA, Greene JE (1994) Plasmas in deposition system, equipped with a polyethersulfone
processes. In: Bunshah RF (ed) Handbook of deposi- membrane (Figs. 1 and 2) with a 10,000 Da
tion technologies for lms and coatings. Science tech- cutoff. The pressure was xed at 1 bar with a
nology and applications. Noeys Publications, New
Jersey, pp 55107 permeate ow rate of 7.5 L/h, while CMC was
used as a complexing agent. The process is
pH dependent, and the metal rejection was
Combined Fouling Index (CFI) more efcient at neutral and alkaline conditions
than at acidic one. Metal rejection
Fouling Index efciency values were 97.6 %, 99.5 %, and
feed solution
complexation of the
target ions
solution of the
ultrafiltration process
polychelatogens
membrane
ComplexationUltrafiltration Process for Heavy Metal Removal from Effluents, Fig. 1 Principles of complexa-
tionultraltration process (Rether and Schuster (2003))
432 Composite Membrane with Inorganic Fillers: Electrolyser Application
ComplexationUltrafiltration Process for Heavy Metal Removal from Effluents, Fig. 2 Complexationultraltration
system
99.1 % for Cu(II), Cr(III), and Ni(II) ions,

respectively, at pH 7. The membrane had Composite Membrane with Inorganic
worked efciently on a wide range of concentra- Fillers: Electrolyser Application
tion up to 100 mg/l for both of Cu (II) and Cr (III)
ions, while the Ni(II) ion rejection Antonino Salvatore Arico
efciency decreased to 57 %. The process is char- CNR-ITAE Institute, Messina, Italy
acterized by low energy requirements involved in
the ultraltration, the very fast reaction kinetics,
and the high selectivity for the metal ion Composite Membrane with Inorganic
separation. Fillers: Electrolyzer Application
Composite recast Naon membranes containing

inorganic llers have been primarily employed in
References fuel cells for high-temperature operation (Aric
et al. 1998) and self-humidication purposes
Barakat MA, Schmidt E (2010) Polymer-enhanced ultra- (Watanabe et al. 1996). Composite membrane
ltration process for heavy metals removal from indus- with inorganic llers for electrolyzer application
trial wastewater. Desalination 256:9093 has been developed in order to extend the oper-
Rether A, Schuster M (2003) Selective separation and
recovery of heavy metal ions using water-soluble ating temperature range of polymer electrolyte
N-benzoylthiourea modied PAMAM polymers. membrane (PEM) electrolyzer and to reduce gas
React Funct Polym 57:1321 crossover effects (Antonucci et al. 2008).
Composite Membrane with Inorganic Fillers: Electrolyser Application 433
Generally, Naon membrane is used as llers that require water for proton conduction
conducting polymer electrolyte in PEM appear suitable for this application especially in
electrolyzer systems. An increase of the opera- the light of the high operating pressure of PEM
tion temperature of an electrolyzer should electrolyzers (up to 100 bars) that allows to main-
enhance the oxygen evolution reaction rate that tain a good fraction of liquid water even at tem-
is the rate-determining step of this process peratures above 100 C. The inorganic llers
allowing to obtain high current and high conver- enhance the water retention inside the composite C
sion efciency. However, commercial Naon membrane allowing to operate properly at high
membranes loose conductivity at temperature temperatures. A composite NaonSiO2 mem-
above 100 C due to membrane dehydration. brane for SPE electrolyzers has shown promising
Peruorosulfonic acid (PFSA) composite mem- properties for high-temperature operation
branes containing hygroscopic ceramic oxide allowing to achieve signicantly higher
Composite Membrane a 2
with Inorganic Fillers:
Electrolyser Application,
Fig. 1 Polarization
1.8
measurements for a PEM
Terminal voltage / V
water electrolysis cell based 80C

on conventional Naon 90C
115 and composite 1.6 100C
PFSASiO2 membrane at 110C
Nafion 115 120C
various temperature and
3.0 bar abs pressure 1.4
1.2
1
0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4
-2
Current density / A cm
b 2
1.8
Terminal voltage / V
80C
90C
1.6
100C
110C
1.4 Composite Membrane 120C
1.2
1
0.2 0.2 0.6 1 1.4 1.8 2.2
Current density / A cm-2
434 Composite Membrane with Inorganic Fillers: Fuel Cell Application
performances with respect to a bare commercial originally developed for reduced humidication
Naon. This effect is mainly due to a signi- operation in polymer electrolyte fuel cells
cantly better water retention than the bare (Watanabe et al.1996) due to the enhanced water
peruorosulfonic membrane and lower gas retention inside the membrane by the effect of the
crossover as a result by the increased tortuosity inorganic ller (Aric et al. 1998). A further
effect produced by the inorganic ller inside the advantage of composite membranes relies in the
membrane. The performance of the electrolyzer barrier effect given by the inorganic ller for
based on NaonSiO2 membrane increased as a methanol cross over (Ren et al. 1996) which is
function of the temperature up to 120 C and of particular relevance at high temperature. It is
pressure. A maximum current density of about well known that the physical adsorption of water
2.1 A cm2 versus 0.7 A cm2 at 1.9 V, 120 C, by materials such as silica (one of the most used
and 3 bar abs was recorded for the composite inorganic llers) is mainly determined by their
membrane compared to Naon 115 (Fig. 1). An surface properties; similar considerations can be
increase of electrical efciency was recorded at made for other hygroscopic inorganic oxides such
low current densities for the high-temperature as alumina. Functional groups on the surface of
SPE electrolyzer compared to conventional these oxides are believed to act as water coordi-
membrane-based devices (Antonucci nation centers (Aric et al. 2003). FTIR analysis
et al. 2008). of various silica materials suggests that oxygen
surface functionalities play a prevailing role in the
References adsorption of water. The surface characteristics of
an inorganic oxide can be modied by thermal
Antonucci V, Di Blasi A, Baglio V, Ornelas R, Matteucci F, treatments in inert or oxidizing atmosphere
Ledesma-Garcia J, Arriaga LG, Aric AS (2008) High through reactions with strong inorganic acids
temperature operation of a composite membrane-based
(Aric et al. 2003). Most of the works on com-
solid polymer electrolyte water electrolyser.
Electrochim Acta 53:7350 posite membranes have addressed the technical
Aric AS, Cret P, Antonucci PL, Antonucci V (1998) aspects related to the use of these materials as
Comparison of ethanol and methanol oxidation in electrolytes in high-temperature fuel cells.
a liquid-feed solid polymer electrolyte fuel cell at
Accordingly, performance, conductivity, and sta-
high temperature. Electrochem Solid-State Lett
1:6668 bility characteristics have been investigated
Watanabe M, Uchida H, Seki Y, Emori M, Stonehart P in-depth. Parallel work has been concerned with
(1996) SelfHumidifying Polymer Electrolyte the investigation of relevant ller properties for
Membranes for Fuel Cells. J Electrochem Soc
application in composite membranes such as sur-
143:3847
face area analysis, surface chemistry studies, and
surface acid-base investigations of the llers. The
conductivity of peruorosulfonic acid (PFSA)
Composite Membrane with Inorganic composite membranes and fuel cell power density
Fillers: Fuel Cell Application at high temperature have been found to be related
to the characteristics of the water adsorbed on the
Antonino Salvatore Arico ller particles. Inorganic llers characterized by
CNR-ITAE Institute, Messina, Italy acidic properties undergo a strong interaction with
water and enhance the DMFC performance at
high temperature. Appropriate selection of the
Composite recast Naon membranes containing surface properties for the inorganic llers thus
inorganic llers have been employed in high- allows to enhance proton conductivity and fuel
temperature (~150 C) direct alcohol (Aric cell performance and extends the operating tem-
et al. 1998) and H2-air fuel cells (Watanabe perature range of composite membranes (Aric
et al.1996). These composite membranes were et al. 2003; Fig. 1).
Composite Membrane with Inorganic Fillers: Fuel Cell Application 435
a
3300 3550
surface OH groups stretching/ cm-1
H2O stretching vibration / cm-1

3200 3500
C
3100 3450
3000 3400
2 3 4 5 6 7 8 9
pH filler slurry
b 1650
400
H2O bending vibration / cm-1
Power Density / mW cm-2

1640
300
200
1630
100
1620 0
0 2 4 6 8 10
pH filler slurry
Composite Membrane with Inorganic Fillers: Fuel corresponding performance of a composite membrane-
Cell Application, Fig. 1 Relationships between water based direct methanol fuel cell at 145 C
adsorption characteristics of inorganic llers and
References characteristics of inorganic llers on the high tempera-

ture performance of composite membranes in direct
Aric AS, Cret P, Antonucci PL, Antonucci V (1998) methanol fuel cells. Solid State Ion 161:251265
Comparison of ethanol and methanol oxidation Ren X, Wilson MS, Gottesfeld S (1996) High performance
in a liquid-feed solid polymer electrolyte fuel cell at direct methanol polymer electrolyte fuel cells. J
high temperature. Electrochem Solid-State Lett Electrochem Soc 143:L12
1:6668 Watanabe M, Uchida H, Seki Y, Emori M, Stonehart
Aric AS, Baglio V, Di Blasi A, Creti P, Antonucci PL, P (1996) SelfHumidifying Polymer Electrolyte Mem-
Antonucci V (2003) Inuence of the acid-base branes for Fuel Cells. J Electrochem Soc 143:3847
436 Computational Fluid Dynamics (CFD) and Membranes
turbulence models in numerical simulations, and

Computational Fluid Dynamics (CFD) their comparison with experimental results in the
and Membranes case of membrane processes is not excellent,
which limits the use of CFD in turbulent regime.
Philippe Moulin Fouling remains a major problem in membrane
Laboratoire de Mcanique, Modlisation et processes: this phenomenon limits the process
Procds Propres (M2P2-UMR 7340), Equipe efciency and is difcult to predict and anticipate.
Procds Membranaires (EPM), Aix Marseille These difculties are linked to the complexity of
Universit, AIX en PROVENCE, France this phenomenon which implies different
interdependent mechanisms occurring at pore
scale. The relative importance of each fouling
The increasing number of CFD membrane studies mechanism has been determined according to
is clearly related to the recent developments in the particle size, the pore size, and the surface
computer power and to the use of ner grid density of pores. By CFD progress it is now pos-
meshes in the vicinity of the membrane. Two sible (a) to pursue the description of the different
approaches have been particularly considered: fouling mechanisms by integrating the complex-
the comprehension of the hydrodynamics and of ity of the real membrane structure in the numerical
the mass transfer. simulation and (b) to simulate different kinds of
The hydrodynamics allows the increase of the experimental deposit structure. The limitation is
shear stress near the wall or the transmembrane the membrane reconstruction technique limiting
pressure thus allowing the enhancement of per- this description at the MF and UF membranes.
meate ux and the membrane processes. CFD The ltration in a cylindrical pore (i.e.,
allows determining the hydrodynamics, i.e., the microchannels) can be simulated for different
pressure and velocity elds, taking into account sizes of particles or pores, tortuosities, and hydro-
the geometry of the module and the membranes, dynamic conditions. For these CFD studies, the
the membrane permeability and compactness, as difculties are to well describe particle/particle
well as the operating entry values such as ltration colloidal interactions and resuspension of parti-
or backwash pressures, ltration mode, and grav- cles after capture.
ity. For example, it is possible to determine the Thus, CFD is without any doubt an important
pressure and velocity elds in (i) a hollow ber tool for understanding mass transfer in membrane
module containing more than 40,000 bers (small processes, and opportunities for the development
diameter (di = 0.93 mm) and large ow rate of new membrane geometries are numerous.
(50 m3.h1) or in (ii) ceramic multichannel mem-
branes or module to obtain information about the
wall shear stress distribution at the surface of the
membrane. The study of hydrodynamics reveals Computer-Aided Methods and Tools
the optimum operating conditions and the most
suitable geometry characteristics to determine a Giorgio De Luca
compromise between membrane area, channel Institute on Membrane Technology, National
(geometry and number), and energy consumption Research Council, ITM-CNR, Rende, Italy
to optimize membrane processes. A large number
of studies relate to the diphasic ows, to the tur-
bulence promoters, and to the geometries of mem- Computational methods can be divided into
branes capable of generating secondary ows. approaches using adjustable or empirical parame-
Usually a good agreement is obtained between ters and those which do not use them. Concerning
CFD and experimental data obtained for the trans- the rst methods, several procedures have been
fer of solvent. Some recent studies consider tur- developed for the optimization of the adjustable
bulent ows and the signicant number of parameters. For example, design of experiments is
Computer-Aided Methods and Tools 437
Time, s D,, J,, eq. FEM, Lattice Bolztmanm

Macro
103
Phase distri.,
Morphologies, C
Micro fluids. Micro
Lattice Bolztmanm
109
FFV,FAV, d-spa,Tg, Conformationa analysis,
Molecular and sorption diffusion in large
Nano system Semiempirical, Monte Carlo
1015 Bond energies ,Dipoles, Electr. Charges,

Sub Accurate geometries, Spectroscopic analysis.
Nano Quantum Mechanics
109 107 104 100
Space, m
Computer-Aided Methods and Tools, Fig. 1 Computational methods and achievable properties
an approach dening the minimum number of force eld (FF), used in the MD simulations is
experiments required to obtain the tting parame- dened by QM or parameterized over measure-
ters. Instead, in the ab initio methods, these quan- ments. The temperature, volume, pressure, and
tities are obtained by direct experimental number of particles dene the statistical ensemble
measurements or from other simulations carried in which the MD can be performed. The classic
out in smaller time-space scale (Steinhauser MC method is generally based on the von Neu-
2008), until the sub-nanometer scale is reached. mann, Metropolis, and Ulam algorithm
The choice of the computational approaches (Metropolis and Ulam 1949; Wood 1986). Later,
strictly depend on the proprieties to be studied. the method has grown to the point where it leads
Figure 1 shows the various computational to several methods all belonging to the MC fam-
methods and properties that can be obtained. ily. In the grand canonical MC ensemble, the
Binding energies, molecular electrostatic prop- simulations are performed at chemical potential,
erties must be evaluated by quantum mechanics volume, and temperature constant, whereas the
(QM) approaches (Veszprmi and Fehr 1999). number of molecules varies. The potential surface
These generally require signicant computational energy is always dened by means of a specic
resources, although the development of parallel FF. Interesting MC approaches are based on the
supercomputers and efcient algorithms have quantum MC methodology. The semiempirical
allowed to carry out QM calculations unthinkable methods use Hamiltonians to describe the system,
few years ago. The quantities, obtained at QM but some contributions of these operators are
level, can be used in subsequent molecular obtained empirically (Clementi and Corongiu
dynamics (MD), Monte Carlo (MC), or semiem- 1995). These approaches allow to evaluate the
pirical calculations to describe properties related target properties more quickly than the QM
to a huge number of molecules and atoms, like methods. In Table 1, the illustrated computational
polymers. MD methodologies are based on the methods are summarized with their advantages
simultaneous solution of Newtons equation of and disadvantages.
motion referred to the atoms of the physical sys- The information provided by smaller scales
tem (Allen and Tildesley 2003). The potential can be utilized in mesoscale calculations: coarse
energy surface (interaction potential), also called grain MD or Lattice Boltzmann (LB) (Swift
438 Computer-Aided Models
Computer-Aided Methods and Tools, separately and others in which both the method-
Table 1 Advantages and disadvantages of some compu- ologies are present. Particular attention should be
tational methods
given to parallel algorithms since parallel super-
Method Advantages Disadvantages computers allow the use of huge number of pro-
Molecular Systems of Experimental cessors (or computers interconnected by the net).
mechanics and thousands of data or values
Monte Carlo atoms, low from quantum These possibilities allow to perform complex cal-
dynamics computational mechanics, culations in short computational time.
costs absence of bond
breaking/
forming.
Noncovalent References
bonds are not
well described, Allen MP, Tildesley DJ (2003) Computer simulation of
less general liquids. Oxford University Press, New York
Semiempirical Less computational Experimental Clementi E, Corongiu G (eds) (1995) Methods and tech-
demand than data or values niques in computational chemistry METECC-95.
quantum from quantum STEF, Cagliari. ISBN 88-86327-02-1, Club Europeen
mechanics mechanics, less MOTECC
approaches, general De Luca G, Bisignano F, Paone F, Curcio S (2014) Multi-
systems of scale modeling of protein fouling in ultraltration pro-
hundreds of atoms cess. J Membr Sci 452:400414
Metropolis N, Ulam S (1949) The Monte Carlo method.
Ab initio, Do not depend on Computationally
J Am Stat Assoc 44:335341
density experimental data, very expensive
Steinhauser MO (2008) Computational multiscale model-
functional useful for a broad Small systems
ing of uids and solids. Springer, Heidelberg
theory quantum range of molecules
Swift MR, Orlandini E, Osborn WR, Yeomans JM
mechanics without available
(1996) Lattice Boltzmann simulations of liquid-gas
experimental data,
and binary uid systems. Phys Rev E 54(5):50415052
general
Veszprmi T, Fehr M (1999) Quantum chemistry. Kluwer
Academic/Plenum Publishing, New York
Wood WW (1986) Early history of computer simulations
et al. 1996). In particular, LB is a powerful method in statistical mechanics molecular dynamics simulation
for simulating uid conned in microsystems, as of statistical systems. Proceedings of the Enrico Fermi
Summer School. Varenna, pp 313
may be precisely a membrane. This method
allows to solve the Navier-Stokes equations in a
simple way and with a notable reduction of com-
putational time. LB allows an easy description of
the interfaces and especially without the use of Computer-Aided Models
adjustable parameters as made by the conven-
tional nite elements methods (FEM). Finally, Giorgio De Luca
the description of the properties of macrosystems Institute on Membrane Technology, National
can be carried out using methods based on classic Research Council, ITM-CNR, Rende, Italy
mechanics or dynamics. The differential equa-
tions describing these physical systems are eval-
uated numerically by means of FEM or nite Models can be divided into mathematical and
volume procedures (De Luca et al. 2014). There structural (chemical models), both connected to
is a notable number of codes implemented for each other. Mathematical models and procedures
each method just described. These are divided can be numerical algorithms, in which each single
into programs in which a single methodology is step is simple arithmetic and logical relations, or
implemented or those in which several methods, they can be dened by analytical relationships,
radically different, can be found. Therefore, it is regardless of how they are evaluated. In either
possible to nd codes in which only QM (based case, mathematical models and procedures are
on different theories) or MD are implemented the result of some assumptions or approximations
Computer-Aided Models 439
Computer-Aided Models, Fig. 1 CNT structural models
based on structural or chemical models (Allen and crucial. These structural models, also called ana-
Tildesley 2003; De Luca et al. 2006, 2008). For logues, inevitably lead to neglect some aspects;
example, the description of a droplet formation nevertheless in some cases these may be irrelevant
during membrane emulsication by means of ana- if the choice of the analogues is done correctly. In
lytical force-balance relationships requires some fact, albeit molecular models certainly introduce
approximations about the droplet shape and its approximations in the evaluation of the macromo-
evolution along the membrane membrane pore. lecular or biological proprieties, some functions
Moreover, the shape of the membrane pores of these only depend on a limited part of the whole
should be also modeled. Therefore, any mathe- structures. Thus, chemical models can mimic very
matical model can be correlated with structural or well the function of complex systems (Gademann
chemical models. By using computational proce- et al. 2007; Zurcher et al. 2006; Saxer et al. 2010).
dures, starting from quantum mechanical calcula- Some examples, concerning models of carbon
tions, molecular dynamics (MD) or Monte Carlo nanotubes (CNT), have been presented in Fig. 1.
(MC), in fact particular attention should be paid to In summary, mathematical models, computational
the molecular (chemical) models. Molecular methods, and structural (chemical) models care-
models are closely dependent on the target proper- fully tune as a function of the type of calculation
ties to be assessed as well as on the computational which is required to be done. Huge literature
time required to get these properties. Computa- exists about studies on any kind of chemical
tional time, in turn, depends on the level of theory models (periodic surfaces, slabs, or clusters) and
used, that is, the mathematical approaches. correlated computational procedures (periodic
For example, at the moment, quantum mechan- calculations, embedded clusters, and quantum
ical methods are not applicable to optimize the mechanics/molecular mechanics methods, etc.).
geometry of macromolecules as polymers or sys-
tems containing thousands of atoms like biologi-
cal systems. Molecular dynamics methodologies References
or coarse grain Monte Carlo can be used in these
cases. However, molecular dynamics approaches Allen MP, Tildesley DJ (2003) Computer simulation of
cannot be used to study systems in which the liquids. Oxford University Press, New York
breaking and formation of bonds or noncovalent Gademann K, Bethuel Y, Locher HH, Hubschwerlen
C (2007) Biomimetic total synthesis and
bonds are decisive. Thus, in the latter case, the antimicrobial evaluation of anachelin H. J Org Chem
choice of a chemical model of macromolecules is 72:83618370
440 Concentration Polarization Coefficient (CPC)
De Luca G, Drioli E (2006) Force balance conditions for concentration polarization coefcient is a measure
droplet formation in cross-ow membrane emulsications. of the concentration polarization extent in a
Journal of Colloid and Interface Science 294:436448
De Luca G, Di Maio FP, Di Renzo A, Drioli E (2008) Droplet membrane-aided separation system. In general,
detachment in cross-ow membrane emulsication: the concentration polarization coefcient indi-
Comparison among torque- and force-based models. cates the degree of permeation driving force lost
Chemical Engineering and Processing 47:11501158 with respect to the total driving force between the
Saxer S, Portmann C, Tosatti S, Gademann K, Zurcher S,
Textor M (2010) Surface assembly of catechol- bulks of the uid phase of feed and permeation
functionalized poly(L-lysine)-graft poly(ethylene side (Fig. 1).
glycol) copolymer on titanium exploiting combined In the open literature, it is possible to nd sev-
electrostatically driven self-organization and biomimetic eral different denitions of CPC (see Table 1 for
strong adhesion. Macromolecules 43:10501060
Z
urcher S, Wackerlin D, Bethuel Y, Malisova B, Textor M, the most used), leading to different quantications
Tosatti S, Gademann K (2006) Biomimetic surface of this phenomenon. All denitions refer to a key
modications based on the cyanobacterial iron chelator species permeating through the membrane. In par-
anachelin. J Am Chem Soc 128:10641065 ticular, for constant total pressure in feed and per-
meate side, the denitions of Wijmann
et al. (1996), Yeom et al. (2002), and Zhao
et al. (2008), who used a coefcient per each mem-
Concentration Polarization brane side in order to take into account also the
Coefficient (CPC) polarization in the downstream side, are in fact
coincident. Based on these denitions, CPC tends
Alessio Caravella to zero for increasing polarization, going toward
National Institute of Advanced Industrial Science the unity in the opposite situation (no polarization).
and Technology Research Institute for On the contrary, Wang et al. (2002) dened the
Innovation in Sustainable Chemistry, ISC-AIST, CPC in such a way that it be close to the unity when
Tsukuba, Ibaraki, Japan concentration polarization is signicant. A more
useful denition is provided by Haraya
et al. (1987) and Takaba and Nakao (2005), who
Concentration polarization coefcient considered a unique coefcient for accounting the
(indicated in the literature as PC or CPC). The polarization on both the membrane sides (Table 1).
Concentration
Polarization Coefficient
(CPC), Fig. 1 Schematic Mixture Pure H2
Membrane
representation of the driving Side Side
force (DF) drops along the
permeation direction in case
of hydrogen permeation
~ DF Membrane
~ DF Bulk
through Pd-based
membranes. The solid and H2 Flux
dashed lines indicate the
Polarization
forward and backward

phenomena
permeation case,
respectively. The proles
are expressed in terms of
equivalent partial pressure
(Adapted from Caravella
et al. (2009))
H2 Flux
Bulk Film
Concentration Polarization in Gas Separation 441
Concentration CPC denition Range References

Polarization Coefficient
Feed
cFilm
(0, 1) Wijmann et al. (1996)
(CPC), Table 1 Most used i
definitions of concentration cBulk

Yeom et al. (2002)
i
polarization coefficient in
Feed, Permeate
xFilm
(0, 1) Zhao et al. (2008)
the literature i
xi
Bulk

Feed

(0, 1) Wang et al. (2002)
xFilm
1 Bulk i
C
xi
N1 (0, 1) Zhang et al. (2006)

, for single-species permeation
N 01 (0, 1)
N1
, for multispecies permeation
N2

Feed
Permeate
xFilm
xBulk
(0, 1) Haraya et al. (1987)
i i

Feed
Permeate Takaba and Nakao (2005)
xi
Bulk xi
Bulk
p p
D pH2 , Film D pH2 , Membrane (0, 1) Caravella et al. (2009)
p 1 p
D pH2 , Bulk D pH2 , Bulk
c concentration, x molar fraction, Ni actual ux of the ith species, Ni0 ux of the i-th species
without concentration polarization
In order to link the polarization effect to the Wijmann JG, Athayde AL, Daniels R, Ly JH, Kamaruddin
selectivity of a membrane, Zhang et al. (2006) used HD, Pinnau I (1996) The role of boundary layers in the
removal volatile organic compounds from water by
the permeating uxes in the CPC denition without pervaporation. J Membr Sci 109:135146
specifying any explicit form of driving force. Yeom CK, Lee SH, Lee JM, Song HY
In general, for CPC to provide an appropriate (2002) A characterization of concentration polarization
quantication of the driving force loss, it is advis- in a boundary layer in the permeation of VOCs/N2
mixtures through PDMS membrane. J Membr Sci
able to dene it using the driving force character- 205:155174
istic of the permeation mechanism that takes Zhang J, Liu D, He M, Xu H, Li W (2006) Experimental
place. By doing this for hydrogen permeation and simulation studies on concentration polarization in
through Pd-based membranes, Caravella H2 enrichment by highly permeable and selective Pd
membranes. J Membr Sci 274:8391
et al. (2009) dened the CPC using the Sieverts Zhao S, Li Z, Liu Y, Wang L (2008) Simulation of binary
characteristic driving force, as indicated in gas separation in hollow ber membrane-acetylene
Table 1. In this case, the CPC is equal to zero in dehydration. Desalination 233:310318
systems without polarization, tending to the unity
for maximum polarization (external mass transfer
controlling the overall permeation process).
Concentration Polarization in Gas
References Separation
Caravella A, Barbieri G, Drioli E (2009) Concentration Alessio Caravella

polarization analysis in self-supported Pd-based mem- National Institute of Advanced Industrial Science
branes. Sep Purif Technol 66:613624 and Technology Research Institute for
Haraya K, Hakuta T, Yoshitome H (1987) A study of concen-
Innovation in Sustainable Chemistry, ISC-AIST,
tration polarization phenomenon on the surface of a gas
separation membrane. Sep Sci Technol 22:14251438 Tsukuba, Ibaraki, Japan
Takaba H, Nakao S (2005) Computational uid dynamics
study on concentration polarization in H2/CO separa-
tion membranes. J Membr Sci 249:8388
In a separation process using membranes, concen-
Wang R, Liu SL, Lin TT, Chung TS (2002) Characteriza-
tion of hollow ber membranes in a permeator using tration polarization (see also bed-to-wall mass
binary gas mixtures. Chem Eng Sci 57:967976 transfer limitations) consists in decrease of the
442 Concentration Polarization in Gas Separation
Concentration
Polarization in Gas Feed Side
Separation,
Fig. 1 Sketch of the A A A A A A A
concentration polarization
phenomenon. The presence A B A A B A A
of the less-permeating
Concentration
species (b) generate a A B A B A B B
resistance to the ux of the Polarization
Polarization
more permeable species (a) B A B B B A B
Permeation through
Membrane
A A
B
A
Permeation Side
permeation driving force of more permeable spe- reverse osmosis. On the contrary, in the same
cies through the membrane because of the pres- period, concentration polarization had been con-
ence of less permeable species, which act as an sidered negligible in gas separation, both because
additional resistance to permeating ux. The of the low the membrane performance in terms of
scheme reported in Fig. 1 provides the physical permeation ux and because gases have diffusiv-
situation leading to concentration polarization. In ity of four to ve orders of magnitude higher than
particular, the species B is pushed toward the liquids (He et al. 1999).
membrane surface by the effect of the ux of the However, the progresses obtained in the last
species A. In these conditions, there is also a split decades in the eld of membrane preparation for
in the respective concentration of the two species, gas separation allowed increasing signicantly the
which means that the concentration of the more transmembrane ux of the key species to separate.
permeable species A tends to be higher in the bulk The higher permeation ux causes the less-
of the feed uid phase, while at the same time, the permeating species to concentrate (polarize)
concentration of the species B tends to be higher around the membrane surface, generating a phe-
in the proximity of the surface. From this point of nomenon similar to the electrical polarization,
view, it is possible to say that the concentration from whose analogy concentration polarization
polarizes, analogously to what occurs in electrical takes its name.
polarization, where electric charges accumulate In these conditions, the inuence of the selec-
toward the respective electrodes. tive layer decreases with respect to the resistance
In practice, concentration polarization occurs of the external mass transfer can affect seriously
in every membrane-aided device and cannot be the membrane performances up to become one of
completely removed, but rather minimized. Fur- the slowest elementary steps of the overall perme-
thermore, the higher the membrane selectivity ation process (Caravella et al. 2009).
(or separation factor), the higher the effect of
concentration polarization, because the removal
of the non-permeating species from membrane References
surface becomes more difcult.
Historically speaking, the rst eld where Caravella A, Barbieri G, Drioli E (2009) Concentration
causes and effects of concentration polarization polarization analysis in self-supported Pd-based mem-
branes. Sep Pur Technol 66:613624
were analyzed in detail (since the beginning of the He G, Mi Y, Yue PL, Chen G (1999) Theoretical study on
1960s) was the liquid separation processes concentration polarization in gas separation membrane
through membranes such as ultraltration and processes. J Membr Sci 153:243258
Constant-Volume Diafiltration 443
permeating ux can be mathematically expressed

Concentration Profile as follows:
Alessio Caravella dmi dlngi xi dP

RT ui (4)
National Institute of Advanced Industrial Science dy dy dy
and Technology Research Institute for
Innovation in Sustainable Chemistry, ISC-AIST, where P [Pa] is the total pressure, gi is the activity C
Tsukuba, Ibaraki, Japan coefcient linking the molar fraction xi [mol
mol1] to the activity ai [], and ui [m3 mol1] is
The concentration of a chemical species that is the partial molar volume. Based on the different
present in a certain volume of space V with a permeation models to apply to different membrane
certain amount of mass is dened as the ratio types (dense or porous), different concentration
between this amount of mass and the volume V. proles are established through the membrane.
It can be expressed in terms of moles (i.e., molar The most common permeation models for dense
concentration ci [mol m3], Eq. 1) or mass (mass and porous membranes are the solution-diffusion
concentration or mass density ri [kg m3], Eq. 2). model and the pore-ow model. The former assumes
ni that the pressure within a membrane is uniform and
ci (1) that the chemical potential gradient across the mem-
V
brane is expressed only as a concentration gradient.
mi
ri (2) The latter assumes that the concentrations of solvent
V
and solute within a membrane are uniform and that
In general, a (nonconstant) concentration pro- the chemical potential gradient across the membrane
le can be established because of the presence of is expressed only as a pressure gradient.
sources/sinks (distributed and or concentrated) of
momentum, mass, and/or energy in some loca-
tions within the physical domain considered. References
In membrane technology, researchers are
mostly interested in the evaluation of concentra- Bachorczyk Nagy R, Danielewski M, Filipek R (2003)
Interdiffusion under the Chemical Potential Gradient;
tion proles both through membrane and in mem- Comparison of Onsager and Darken Models. Defect
brane device, including the uid phase in feed and and Diffusion Forum, 216217:141148.
permeate side. As for the former, the transmem- Hara S, Caravella A, Ishitsuka M, Suda H, Mukaida M,
brane ux Ji [mol s1 m2] and, thus, the concen- Haraya K, Shimano E, Tsuji T (2012) Hydrogen diffu-
sion coefcient and mobility in palladium as a function
tration prole of the ith species through of equilibrium pressure evaluated by permeation mea-
membranes are promoted by a difference of its surement. J Membr Sci 421422:355360
chemical potential mi [J mol1] between feed and Wijmans JG, Baker RW (1995) The solution-diffusion
permeate side (Eq. 3) (Wijmans and Baker 1995; model: a review. J Membr Sci 107:121
Bachorczyk et al. 2003; Hara et al. 2012).
dmi dm Constant-Volume Diafiltration

J i Li ci Bi i (3)
dy dy
Zoltn Kovcs
In Eq. 3, Li is a proportionality factor (not gener- Department of Food Engineering, Institue of
ally constant), Bi is the mobility (i.e., a measure of Bioengineering and Process Engineering, Szent
the aptitude of the ith species to pass through Istvan University, Budapest, Hungary
membrane), and y indicates the permeation direc-
tion (normal to membrane surface).
If the permeation driving force is mainly gen- Constant-volume dialtration (also called
erated by partial concentration and pressure, the constant-volume dilution mode) is a batch
444 Constant-Volume Diafiltration
dialtration process for separating microsolutes i in the feed tank gives the following initial value
from macrosolutes, in which the volume of the problem for its concentration cf,i(t):
process liquor is kept constant during ltration by
8
continuously adding a diluant into the feed tank at a < dcf , i t cf , i tqtRi t 1
rate equal to the permeation rate. The schematic dt Vf
:
representation of its conguration is shown in 3 mmcf , i 0 c0f , i
Fig. 1. The constant tank volume can be maintained
by the use of a ball oat valve or by means of liquid where q(t) and Ri(t) are the permeate ow and the
level controller (Beaton and Klinkowski 1983). solute rejection that are subject to change during
As ltration progresses, the concentration of operation. The constants Vf and cf,i denote the
membrane-permeating microsolutes in the feed volume of the feed tank and the initial feed con-
tank continuously decreases, while that of the centration of component i. The equation describes
macrosolutes remains ideally unchanged (or close the evolution in time of the feed concentration cf,
to constant in case of incomplete rejection). i(t) assuming that the diluant consists of no com-
Note that some literature sources misleadingly ponent i and the feed tank is well mixed.
refer to constant-volume dialtration as con- In many applications, the ux and the rejections
tinuous dialtration. Although the addition of are concentration-(inter)dependent quantities
diluant is performed in a continuous manner, (Kovcs et al. 2009) and may vary with operating
constant-volume dialtration should not be con- conditions such as temperature, applied pressure,
sidered as a continuous process. It is a true batch and hydrodynamics. In such cases, no closed form
process. It is advised, however, to use the term solution of the set of resulting complex differential
constant-volume dilution mode which is less algebraic equations exists; thus, numerical tech-
common in the literature. The term dilution niques are required to solve the model equations.
mode reveals that this technique is actually an Under the simple assumptions of constant perme-
operational mode of a membrane ltration process ate ow and rejections, however, the problem can
in which the process liquor is diluted with pure be reduced to the following algebraic expression:
solvent.

Depending on the technological goal, either the cf , i t
ln DRi 1% for i 1, 2, N
permeate or the retentate represents the phase of cf , i 0
primary economic importance. Thus, the product
of the operation is either formed in the feed tank as where D is the dialtration factor (also called
a puried mixture of macrosolutes or collected in diavolume) that is dened as the ratio of applied
the permeate tank where the microsolutes are volume of diluant to feed volume. For applica-
accumulated. Both treatments pose a dynamic tions where the objective is to reduce the
modeling problem. A mass balance on component microsolute concentration by a xed amount, the
Constant-Volume Retentate
Diafiltration, Diluant u(t)
Fig. 1 Schematic
representation of constant- Permeate q(t)
Membrane
volume dialtration settings module
Feed
diafiltration level
Feed tank
Contactor-Type Catalytic Membrane Reactor 445
Constant-Volume 0
Diafiltration,
Microsolute reduction in percentage [%]

Fig. 2 Dependence of
R = 0.9
microsolute removal on
applied diavolumes in
constant-volume R = 0.8
dialtration
R = 0.7 C
90
R = 0.6
R = 0.5
R = 0.4
R = 0.3
R = 0.2
R = 0 R = 0.1
99
0 1 2 3 4 5 6 7 8 9 10
Diavolumes [-]
necessary diavolumes can be determined based on The term catalytic membrane contactor refers to
Eq.2 as illustrated in Fig. 2. a device in which a membrane containing a cata-
In practice, constant-volume dilution mode lytically active phase is used to provide the reac-
is frequently preceded and/or followed by batch tion zone for conversion of one or more reactants
concentration operational steps in order to reduce from one or more uid phases (Dittmeyer and
the initial feed volume to a desired level and, thus, Caro 2008). The membrane not always has a
to concentrate the macrosolutes (for further separation function, it provides the surface-rich
details, see entry on traditional dialtration). reaction zone for a gas-liquid, but also for a
gas-gas and a liquid-liquid reaction. Applications
References of membranes with built-in catalysts for
gas-liquid reactions have been also reviewed in
Beaton NC, Klinkowski PR (1983) Industrial ultraltration Dittmeyer et al. (2004).
design and application of dialtration processes. J Sep The catalytic diffuser concept can be uti-
Process Technol 4(2):110
Kovcs Z, Discacciati M, Samhaber W (2009) Modeling of lized for solid-catalyzed gas/liquid reactions
batch and semi-batch membrane ltration processes. in various ways. If the membrane is not wetted,
J Membr Sci 327(12):164173 the operation principle is similar to that of a
membrane contactor. The gas is on the support
side, and the liquid is pumped through the shell
Contactor-Type Catalytic Membrane side. The gaseous reactants enter the liquid
Reactor phase at the gas/liquid contact plane which is
established at the pore mouth towards the
Juergen Caro external membrane surface. The catalyst is
Institute of Physical Chemistry and deposited on this surface and gets in contact
Electrochemistry, Leibniz University, Hannover, with the reactants. On the contrary, if the mem-
Germany brane is wetted the gas and not the liquid is at
overpressure as described in Dittmeyer and
Caro (2008).
Synonyms Gasliquid contactors without catalytic func-
tion can be used in gas adsorption. Examples are
Catalytic diffuser; Pore-through-ow catalytic carbon dioxide removal from gas mixtures using
membrane monoethanolamine (Simons et al. 2009) or ionic
446 Continuous Diafiltration: Cocurrent and Countercurrent Modes
liquids (Albo et al. 2010) as carbon dioxide References

absorbing solutions.
For liquid-liquid and gas-gas catalytic mem- Albo J, Luis P, Irabien A (2010) Ind Eng Chem Res
49:1104511051
brane contactors, see Dittmeyer and Caro (2008)
Dittmeyer R, Caro J (2008) Catalytic membrane reactors.
(Fig. 1). In: Ertl G, Knzinger H, Sch
uth F, Weitkamp J (eds)
Catalytic diffuser with wettable membrane: Handbook of heterogeneous catalysis. Wiley-VCH,
The active material is placed solely into the sur- Weinheim, pp 21982248
Dittmeyer R, Svajda K, Reif M (2004) A review of cata-
face layer of an asymmetric membrane. By
lytic membrane layers for gas/liquid reactions. Top
applying overpressure on the gas side, above Catal 29:327
the bubble point pressure of the intermediate Simons K, Nijmeijer K, Wessling M (2009) J Membr Sci
layer but below that of the surface layer, the 340:214220
gas/liquid contacting plane is established inside
the membrane close to the surface layer. In this
way, a short diffusion path for the liquid and for
the gas reactant is achieved (after Dittmeyer and
Caro (2008)). Continuous Diafiltration: Cocurrent
and Countercurrent Modes
Zoltn Kovcs
Department of Food Engineering, Institue of
Bioengineering and Process Engineering, Szent
Istvan University, Budapest, Hungary
Continuous dialtration is a type of dialtration

that accomplishes the selective separation of
multiple solutes in a continuous fashion by
mixing the feed stream with diluant and pumping
it across the membrane with permeate and
retentate being removed. The role of diluant
addition is to facilitate the separation of
membrane-permeating microsolutes from the
retained macrosolutes. This strategy is generally
applicable to a variety of microltration, ultral-
tration, nanoltration, and reverse osmosis sys-
tems and commonly employed in plants that
Contactor-Type Catalytic Membrane Reactor, require large throughputs. Its simplest congura-
Fig. 1 Catalytic diffuser with wettable membrane as an tion, known as single-pass dialtration, is
example for a Contactor-Type Catalytic Membrane Reac- schematically shown in Fig. 1a. Continuous
tor for a solid-catalyzed gas-liquid reaction. By applying
overpressure on the gas side, the gas-liquid contacting
dialtration, as opposed to batch dialtration,
plane is established inside the membrane close to the realizes a steady-state separation in which the
surface. In this way, a hort diffusion path for the liquid product (i.e., the nal retentate) is not formed
and for the gas is achieved. Reproduced from Roland in the feed vessel as ltration progresses, but being
Dittmeyer, Karel Svajda, Martin ReifA review of catalytic
continuously withdrawn from the system during the
membrane layers for gas/liquid reactionsTopics in Cataly-
sis, 29 (2004), 3-27 reprinted with permission from entire course of ltration. In practice, continuous
Elsevier) processing employs feed-and-bleed conguration
Continuous Diafiltration: Cocurrent and Countercurrent Modes 447
a b
Retentate Retentate
Diluant Diluant
Booster
pump
Feed Permeate Permeate

Feed C
High-pressure
pump
Continuous Diafiltration: Cocurrent and Countercurrent Modes, Fig. 1 (a) Single-pass continuous dialtration
and (b) continuous dialtration with partial retentate recycle (feed and bleed)
Diluant
Feed Retentate
Stage 1 Stage 2 Stage i Stage N
Permeate
Diluant
Feed Retentate
Stage 1 Stage 2 Stage i Stage N
Permeate
Continuous Diafiltration: Cocurrent and Countercurrent Modes, Fig. 2 Simplied ow diagram of multistage
cocurrent (top) and countercurrent (bottom) continuous dialtration
to enhance separation as shown in Fig. 1b. Here, the arranged in feed-and-bleed conguration. The
retentate is split, and a portion of it is recycled back two basic concepts of multistage arrangement
to the cartridge inlet. are the cocurrent (also called crosscurrent) and
Large-scale continuous dialtration systems the countercurrent congurations (Dutr and
operate in multistage arrangement in which a Trgrdh 1994). The former applies fresh diluant
portion of the retentate generated in one stage is at each stage, while the latter uses recycled per-
used as the feed to the next. These systems can meate as diluant as shown in Fig. 2.
reach higher system recoveries, without exceed- Continuous membrane plants are usually
ing the single element recovery limits. Each stage designed for both concentration and fraction-
may consist of several membrane elements ation purposes. To achieve the twin objectives
448 Continuous Electrodeionization
of concentrating and purifying a multicomponent Yee KWK, Wiley DE, Bao J (2007) Whey protein
feed, the combined application of both concen- concentrate production by continuous ultraltration:
operability under constant operating conditions. J
tration mode and dialtration mode operation Membr Sci 290:125137. doi:10.1016/j.
is required. Traditional continuous membrane memsci.2006.12.026
systems consist of a number of stages in series
in which consecutive stages are assigned to per-
form the tasks of pre-concentration, dialtration,
and nal concentration (Madsen 2001). Many
design variations exist in practice that are pri-
Continuous Electrodeionization
marily achieved by the manipulation of the
arrangement of stages within the plant, the num- Karel Bouzek
Faculty of Chemical Technology, University of
ber of stages and ltration areas assigned to the
various tasks, the amount and the place of fresh Chemistry and Technology Prague, Technick 5,
diluant addition in the process train, the intro- Prague 6, Czech Republic
duction of bypasses and manifolds, the alteration
of permeate recycle ratios, and the appointment
of the stages in which permeate is withdrawn Continuous electrodeionization is an alternative
term for electrodeionization; however, it empha-
from and recycled to (Morison and She 2003,
Yee et al. 2007; Mohanty and Ghosh 2008). The sizes the fact that, in contrast to classical deioni-
selection of optimal dialtration strategy in con- zation by ionic exchange, this process is
continuous.
tinuous processing is very case specic. In gen-
eral, a reduced consumption of fresh diluant but
a larger membrane area is required for the coun-
tercurrent approach (i.e., reintroduction of per-
meate as diluant in the process train) as Continuous Membrane Fermentor
compared to cocurrent process if the same degree (CMF)
of purication is to be attained (Lipnizki
et al. 2002). Cell Recycle Membrane Fermentor
References
Continuous Stirred Tank Membrane
Dutr B, Trgrdh G (1994) Macrosolute-microsolute
separation by ultraltration: a review of dialtra-
Reactor (CST-MR)
tion processes and applications. Desalination
95:227267 Giuseppe Barbieri
Lipnizki F, Boelsmand J, Madsen RF (2002) Concepts of Institute on Membrane Technology, National
industrial-scale dialtration systems. Desalination
144:179184
Madsen RF (2001) Design of sanitary and sterile UF-
and dialtration plants. Sep Purif Technol 2223:
7987 Before introducing the continuous stirred tank
Mohanty K, Ghosh R (2008) Novel tangential-ow coun-
tercurrent cascade ultraltration conguration for con-
membrane reactor (CST-MR), it is useful/helpful
tinuous purication of humanized monoclonal to report about the continuous stirred tank reactor
antibody. J Membr Sci 307:117125. doi:10.1016/j. (CSTR).
memsci.2007.09.010 The continuous stirred tank reactor (CSTR) is
Morison KR, She X (2003) Optimisation and graphical
representation of multi-stage membrane plants.
an ideal reactor model assuming perfect mixing,
J Membr Sci 211:5970. doi:10.1016/S0376-7388(02) with no spatial gradients of any variable such as
00375-7 species concentration, temperature, pressure,
Continuous Stirred Tank Membrane Reactor (CST-MR) 449
Continuous Stirred Tank

Membrane Reactor
(CST-MR),
Fig. 1 Continuous stirred
tank reactor (CSTR)
scheme
etc. The effect of the perfect mixing is the same

value for, e.g., concentration, temperature, etc., in
any point of the whole reactor volume; and their
values are equal to those of the stream exiting the
reactor. In addition, the reaction rate has the same
value in any point of the reactor. Therefore, this
reactor model operating at the lowest reactant
concentration and the highest product concentra-
tion results in the lowest reaction rate. However,
an important effect of perfect mixing is an easy
temperature and reaction rate control, which
results quite simply. Figure 1 shows a CSTR
scheme.
The generic balance equation on the reactor
volume will be:
Continuous Stirred Tank Membrane Reactor
(CST-MR), Fig. 2 Continuous stirred tank membrane
Inlet Outlet Generation reactor (CST-MR)
Accumulation (1)
In steady-state condition, the mass balance
equations fall into algebraic equations:
The mass balance for the component i-th consid-
ering a single reaction will be: NX
reactions
FInlet
i FOutlet
i ni, j r j V Reaction
CSTR 0 (3)
NX
reactions j1
dN i
FInlet
i FOutlet
i ni, j r j V Reaction
CSTR (2)
j1
dt Equation 3 provides the design equation Eq. 4 for
a CSTR in steady-state conditions:
Mass balances provide a set of ordinary differ-
ential equations (ODEs), in which the number FInlet FOutlet
of moles of component i-th is time dependent. V Reaction
CSTR i i
(4)
NX
reactions
In Eq. 2, FiInlet and FiOutlet are, respectively, ni, j r j

the inlet and outlet molar ow of the compo- j1
nent i-th, ri is the reaction rate, Ni is the num-
ber of moles, and V is the reactor/reaction An increase in reaction rate ri causes a reduction
volume. of reactor volume (Eq. 4). The design equation
450 Continuous Stirred Tank Membrane Reactor (CST-MR)
also provides the reactor volume necessary for for the CSTR (perfect mixing that has no spatial
obtaining the exit ow rate FiOutlet, from the feed gradients of species concentration, temperature,
conditions FiInlet, and the reaction rate knowledge. pressure, etc.) on both reaction and separation
The energy balance, coupled to mass balance, sides (2) by the selective removal of reaction
is given by Eq. 5: product by the membrane allowing improved per-
formance (see later on) with respect to a CSTR.
Conversion The generic mass balance (Eq. 1) is still valid

Cp T Outlet T Inlet UA T Outlet T External and has to be written down for both reaction and
DH Reaction permeation even though another term has to be
included: the one taking into account the perme-
(5)
where Cp is the mean heat capacity, TExternal is the ation through the membrane. Therefore, Eqs. 6
external temperature, U is the overall heat transfer and 7 include the permeating ow rate
coefcient, A is the heat exchange area, and DH is (AMembrane JPermeating, the product of the mem-
the enthalpy of the chemical reaction. brane area and permeating ux). This term is
This equipment is very commonly used in con- negative (being mass leaving the reaction volume)
tinuous industrial processes as well as in the plug on the retentate side and positive (entering the
ow reactor. For reactions of greater than zero permeation volume) on the permeate one.
order, CSTR always requires a volume larger than Reaction side:
that of a plug ow reactor to achieve the same
conversion; this is owing to the lower reactant FFeed FRetentate AMembrane J i Permeating
i i
concentrations, as said, at which it operates. There-
NX
reactions
fore, the use of CSTR is recommended when the
ni, j r j V Reaction
CSTMR
desired reaction rate is smaller than that of the side j1
reaction one in order to limit by-product formation.
dN Retentate
i
(6)
dt
CST Membrane Reactor
Permeation side:
A membrane reactor is a device that combines in one
unit a chemical reaction with selective product sepa- FSweep
i FPermeate
i AMembrane J i Permeating
ration by means of a permselective membrane. The
dN Permeate
selective removal of products from the reaction side i
(7)
to the permeate side also provides an increase, e.g., of dt
the reaction rate and equilibrium conversion (Fig. 2).
A continuous stirred tank membrane reactor The mass balance equations for steady state have
[1, 2] (CST-MR) is a device characterized (1) in an algebraic form:
addition to the same properties described above Reaction side:
NX
reactions
FFeed
i FRetentate
i A Membrane
Ji Permeating
CSTMR 0
ni, j r j V Reaction
j1

FFeed FRetentate AMembrane J i Permeating
CSTMR
i i
V Reaction (8)
NX
reactions
ni , j r j
j1
Controlled Delivery 451
Permeation side: small fraction of the drug is delivered to the

target site. Accumulation in tissues not associated
with drug causes side effects. Controlled delivery
FSweep
i FPermeate
i AMembrane J i Permeating 0 to the target tissue eliminates side effects and
AMembrane J i Permeating FPermeate
i FSweep
i allows to reduce drug dosing to reach the desired
(9) effect. The need for controlled targeted delivery is
obvious in chemotherapy, for example, where pain- C
ful side effects of the drugs on healthy cells can be
Equation 8 also identies the reaction volume. It
eliminated by directly delivering the drug to the
results lower than the CSTR reaction volume
target cells. Other examples include drug delivery
owing to the permeation through the membrane.
to the eye, lung, brain, and coronary arteries and
Therefore, a continuous stirred tank membrane
release of growth and angiogenesis factors from
reactor (CST-MR) has a higher performance of a
surgically implanted tissue-engineered constructs.
continuous stirred tank reactor (CSTR), speci-
The eye cavity is used to deliver antibiotics and
cally, for the following points:
drugs. Ophthalmic inserts made from soluble,
insoluble, and bioerodible polymers are currently
1. Higher conversion owing to the selective per-
available. Insoluble inserts can be categorized as
meation through the membrane
diffusional, osmotic, and contact lens. Polyvinyl
2. Higher reaction rate owing to the lower con-
alcohol, hydroxypropylcellulose, polyvinylpyr-
centration of the products removed by the
rolidone, and hyaluronic acid are commonly
selective membrane
used for constructing degradable inserts, while
3. Lower reaction volume at the same productiv-
nondegradable inserts are manufactured from
ity owing to the species permeation
ethylene-vinyl acetate copolymers and styrene-
isoprene-styrene block copolymers. The contact
lenses are not only used for correcting vision but
References
also act as potential drug delivery device by
Barbieri G.; Scura F.; Brunetti A. Series Membrane Sci- presoaking them in drug solutions. Conventional
ence and Technology, Volume 13 Inorganic Mem- eye drops are not efcient since most of it is lost
branes: synthesis, Characterization and Applications; due to overow. To increase drug retention in the
Chapter 9-Mathematical modelling of pdalloy mem- cavity and bioavailability, drug delivery product
brane reactors, 325400
Raich BA, Foley HC (1995) Supra-equilibrium conversion Ocusert was developed. In this system, a reser-
in palladium membrane reactors: Kinetic sensitivity voir lled with a saturated concentration of drug
and time dependence. Applied Catalysis A.: General was placed between two membranes that can con-
129:167188 trol release of drug by the partition/diffusion
mechanism. This product was placed under the
lower eyelid and released drug at constant rate
Controlled Delivery into the tear uid for up to 1 week. Another
commercial product, Lacrisert is a soluble insert
Sacide Alsoy Altinkaya used for the treatment of dry eye syndrome. The
Department of Chemical Engineering, Izmir device is composed of hydroxypropyl cellulose
Institute of Technology, Urla, Izmir, Turkey and slowly dissolves over 68 h to stabilize and
thicken the tear lm.
In addition to the eye, inserts are also used for
The benet of a drug can be enhanced when targeted controlled delivery of drugs to various
controlled delivery of a drug to a targeted tissue tissues. ATRIDOX is an FDA-approved product
becomes possible. After release, the drug enters designed for the treatment of periodontal disease.
the systemic circulation and distributes according Upon injection into the periodontal cavity, drug
to its relative afnities to all tissues, and only a delivery depot is formed and delivers the
452 Controlled Drug Release
antibiotic to the cavity. DUROS implants are (2) solutiondiffusion mechanism. Pills, implants,
nondegradable, osmotically driven systems and and patches are commonly used dosage forms to
used for the delivery of small drugs, peptides, deliver drugs where drug release is controlled by
proteins, and DNA for up to 1 year. Viadur is diffusion through a rate-limiting membrane. Pills
another commercial implant used for treatment of are prepared by pressing the drug into a tablet and
advanced prostate cancer. then coating with a water-insoluble or water-
Drug-eluting stents were developed to prevent soluble polymeric membrane, such as
restenosis or reclosing of coronary arteries. In hydroxylpropyl methylcellulose, ethylcellulose,
these systems, the implantable metal stents are povidone, or polyethylene glycol. Tablets are
coated with small amounts of anti-inammatory also coated with a membrane to protect drugs
and antiproliferative drugs that are released against decomposition in the acid environment
directly into the arterial tissue by dissolution, of the stomach. Commonly used polymers for
partitioning, and diffusion. The release rate is this purpose are cellulose acetate trimellitate
inuenced by the thickness of the coating, type (CAT), hydroxypropylmethylcellulose phthalate
of the coating material, and diffusivity and solu- (HPMCP), polyvinyl acetate phthalate (PVAP),
bility of drug in the coating material. and cellulose acetate phthalate (CAP). Implants
Nanocarriers are alternative systems for targeted are used to deliver drug at a particular site and
controlled drug delivery since they are less cleared contain drug in liquid or powder form in a reser-
than free drugs and they may also have favorable voir surrounded by a semipermeable membrane to
distribution properties into target tissues. Targeting allow slow release of the drug. The membrane can
ligands can be incorporated into the nanoparticles, be made from a degradable or nondegradable
and they permit particles to bind specic cells and polymer, and the release is controlled by rate of
promote the cell uptake and release of drug into the diffusion and degradation. Patches are typical
cells. Drug-loaded nanoparticles can easily reach to membrane-controlled reservoir systems and fre-
target tumor tissues since tumors have leaky capil- quently used for controlled drug delivery. In ocu-
laries that permit the passage of nanoparticulates. lar patches, a drug is enclosed by two release-
Microemulsions, liposomes, dendrimers, block controlling polymer membranes, and through dif-
polymer micelles, and solid lipid and polymer fusion, drug is delivered to the target area. Trans-
nanoparticles are typical examples of nanocarriers. dermal patches are used to deliver drugs with low
solubility since the skin has a large surface area.
The simplest patch consists of a drug-containing
reservoir placed between an impermeable backing
Controlled Drug Release layer and a membrane. The total resistance to drug
permeation is controlled by sum of the resistances
Sacide Alsoy Altinkaya in the membrane and the skin. In the case of
Department of Chemical Engineering, Izmir passive drug transport where the drug concentra-
Institute of Technology, Urla, Izmir, Turkey tion difference between the patch and the skin is
main driving force, the release rate of the drug
may vary signicantly from patient to patient due
Controlled drug release is dened as delivery of a to strong variations in skin permeability. In order
drug to a specic site, in specic time and release to provide an effective rate control and to ensure
pattern. Controlled drug release eliminates over- that drug delivery does not depend on patient or
or underdosing, maintains drug levels in desired patch position, the permeability of membrane
range, needs less dosing, prevents side effects, and should be less than that of the skin, i.e., the release
increases patient compliance. Most of the com- rate should be controlled by the membrane. To
mercially available systems for controlled release expand the spectrum of transdermally deliverable
of drugs rely on membranes where drug release is drugs, chemical permeation enhancers,
generally controlled by (1) osmotic pumping or microneedles, ultrasound, heat, or short- or high-
Controlled Release 453
voltage bursts of electricity are temporarily applied efcacy, convenience, and improved safety by
to the skin to disrupt its barrier function. Alterna- maintaining optimal concentrations of the active
tively, drugs can be delivered by iontophoresis, in ingredient. Many controlled-release technologies
which small amounts of electrical current are operate by diffusion of the active ingredient or
applied through the skin to allow delivery of water through a rate-controlling membrane and
charged drug molecules into the body. The device can be classied into four categories: (1) reservoir
consists of two patches: one of them contains a types, (2) matrix systems, (3) swelling-controlled C
drug reservoir, while the other patch contains ref- systems, and (4) osmotic systems.
erence electrolyte or gel. The drug placed in the In reservoir (membrane) systems, the bioactive
patch electrode which has the same charge as the agent is either encapsulated by two semiperme-
drug is delivered into the skin by electrostatic able membranes or core containing the active
repulsion when electric current is applied. The ingredient is surrounded by a semipermeable
rate of transport of a charged drug during iontopho- membrane. Diffusivity and solubility of the active
resis is controlled by passive diffusion, migration ingredient in the membrane are two key parame-
due to electric current applied, and electroosmosis ters that inuence the rate of diffusion through the
due to bulk uid ow in the same direction as the membrane. The size and shape of the active ingre-
ow of the counter ions. In addition to classical dient, the packing density, and the exibility and
polymers, intelligent, stimuli-responsive polymers crystallinity of polymer chains affect the diffusiv-
are also used as membrane material for controlling ity in the membrane. Solubility of the active ingre-
the release rate of drugs through the membrane. For dient in the polymer depends on its interaction
example, hydrogel membranes manufactured from with the membrane material.
glucose-containing copolymers could respond to In matrix (monolithic) systems, the bioactive
the dynamic changes in external glucose concen- agent is incorporated in a polymer either in
tration, and this response can be used to control the dissolved or dispersed form. When the active
release of insulin through the hydrogel membrane ingredients are in liquid form, they are typically
(Obaidat and Park 1997). dissolved in polymer matrices, while solid ingre-
dients are dispersed in the polymer. Release of the
agent occurs by diffusion through the polymer
References matrix and solubility of the agent becomes a con-
trolling factor.
Obaidat AA, Park K (1997) Characterization of protein
release through glucose-sensitive hydrogel mem- Osmotic controlled-release systems are
branes. Biomaterials 18:801 reservoir-type systems except that the release of
the active agent depends on the diffusion of water
into the membrane. In these systems, a semiper-
meable membrane that is much more permeable to
Controlled Release water than to the active agent separates a core with
high osmotic pressure from a solution with low
Sacide Alsoy Altinkaya osmotic pressure. Water penetration into the core
Department of Chemical Engineering, Izmir generates a hydrostatic pressure within the core
Institute of Technology, Urla, Izmir, Turkey that forces the solution containing the active
ingredient out through the membrane. Water per-
meability of the rate-controlling membrane deter-
Controlled-release technologies are developed to mines the release rate of the active agent.
deliver drugs, proteins, pesticides, fragrances, and Swelling-controlled-release systems are simi-
other bioactive agents at a controlled rate for a lar to the design of matrix systems and usually
desired time period to maintain a constant and consist of an active ingredient dispersed through-
effective level of an active ingredient. out a swellable polymer. In these systems, the
Controlled-release formulations provide greater active ingredient is not released until the polymer
454 Controlled Release by Membrane
matrix is swollen since the diffusivity of the active molecules rst dissolve in the membrane and then
ingredient in the swollen matrix is several orders diffuse from interior part of the membrane to its
of magnitude higher than its diffusivity in the surface. The diffusive ux through the membrane
unswollen polymer. is described by Ficks rst law of diffusion.
Controlled-release systems that are not depen-
dent on diffusion have also been developed. For @Cm
J D (1)
example, in the case of bioerodible systems, the @x
release rate is controlled by either polymer disso- where D is the diffusivity of drug in the membrane,
lution or a chemical reaction at the polymer sur- J is the ux of the drug, and Cm is the concentration
face; therefore, geometric shape of the release of drug in the membrane. At the membrane surface,
system is important. In pendant chain systems, drug molecules choose to partition into either
the active agent is bound to a polymer backbone membrane or surrounding medium. The partition
chain chemically and is released by hydrolytic or coefcient gives a measure of the preference of
enzymatic cleavage. The agent can be attached drug molecules between two phases. Dening the
directly or via a spacer group. The spacer group partition coefcient, K, as the ratio of the concen-
can be used to affect the release rate and hydro- tration in the membrane to the concentration in the
philicity of the system. In a magnetically solution, ux in Eq. 1 is rewritten as follows:
controlled-release system, the active agent is uni-
formly dispersed within a polymer matrix. The @C
J DK (2)
system releases the agent in a way typical of @x
diffusion-controlled matrix systems upon expo-
sure to aqueous media. On other hand, when the where C is the concentration of drug in the sur-
system is exposed to an oscillating external mag- rounding medium. Under steady-state conditions,
netic eld, the release rate becomes much higher. ux through the membrane is obtained by inte-
Various types of polymers have been used in grating Eq. 2.
the design of controlled-release systems. The DC
choice of polymer can affect diffusion and parti- J DK (3)
L
tion coefcients in diffusion-controlled systems,
the rate of erosion or hydrolysis in chemically- Controlled-release systems based on membranes
controlled systems, permeation rates of solvents usually provide constant release rate as long as
in swelling-controlled systems, and mechanical the concentration of drug within core is at satura-
forces in magnetically controlled systems. tion concentration. Deviations from constant
release kinetics occur depending on the storage
history of the release device. Time lag is observed
when the system is used shortly after being
Controlled Release by Membrane manufactured while migration of the agent from
the core into the membrane due to long-term stor-
Sacide Alsoy Altinkaya age causes initial burst effect. Equation 3 clearly
Department of Chemical Engineering, Izmir shows that the release rate of drug is controlled by
Institute of Technology, Urla, Izmir, Turkey diffusivity and solubility of drug in the membrane
as well as the thickness of the membrane. Diffusiv-
ity of drug in the membrane is inuenced by the
Many controlled-release systems commercially size and shape of the drug molecules, exibility,
available now are based on membrane technol- packing density, or crystallinity of polymer chains.
ogy. These systems consist of a core containing Flexibility of polymer chains depends on glass
the drug surrounded by a thin polymer membrane transition temperature and degree of crosslinking
and transport of drug through membrane of the polymer. Diffusivities are higher in rubbery
occurs by a solution-diffusion mechanism. Drug polymers than those in glassy polymers and
Controlled Release Dose 455
decrease as crosslinking density in the polymer

increases. Diffusion takes place in the amorphous Controlled Release Dose
phase of the polymers; therefore, diffusivities are
lower in more-crystalline polymers than in Sacide Alsoy Altinkaya
less-crystalline ones. Solubility of drugs in the mem- Department of Chemical Engineering, Izmir
brane depends on the interaction between the poly- Institute of Technology, Urla, Izmir, Turkey
mer and drug. Simply, solubility of drug is higher in C
polymers containing similar functional groups with
those of drug, in addition, solubility decreases with Controlled release dosage forms are designed to
an increase in crystallinity of the polymer. Design of deliver optimum dose of drug at the right time and
membrane-based controlled-release system for a right place. For an efcient and nontoxic therapy,
particular drug requires to choose an appropriate the drug concentration in plasma should lie within
polymer as a membrane material and appropriate the therapeutic range which is bounded below by
thickness to adjust the release rate at a desired level. the minimum effective concentration (MEC) and
Diffusion-controlled release systems are com- above by the minimum toxic concentration
monly used in commercial products and the most (MTC) as shown in Fig. 1. Conventional doses
successful example is in transdermal applications. lead to a rapid rise and fall in the drug concentra-
These devices contain the drug either in a reser- tion; therefore, several doses at regular intervals
voir surrounded by a membrane or dispersed in a are required to maintain effective levels. After an
polymeric matrix and deliver drugs directly initial rise, controlled release dosage forms main-
through the skin and into the bloodstream. The tain a constant concentration in plasma between
ux of the drug through the skin is enhanced by MEC and MTC. As a result, several benets are
using penetration enhancers or electric current. achieved with these systems including a
Osmotic controlled-release systems also depend prolonged therapeutic effect, increasing short
on diffusion of water across a semipermeable half-life of drugs, reduction in dosing frequency
membrane. These systems for oral delivery of and side effects, and improved patient compli-
drugs are prepared by compressing an osmotically ance. As shown in Fig. 1, zero-order release
active agent into a tablet, coating the tablet with a allows the best control of plasma concentration
semipermeable membrane, and drilling a small and controlled release dosage forms following
hole through the coating. As water permeates zero-order release kinetics are usually achieved
through the membrane into the tablet, a solution by using a rate-limiting membrane that is perme-
of the agent is pumped out of the hole. Such a able to both drug and water. Osmotic pumping
device provides a constant rate of release as drug and solution-diffusion mechanism control drug
traverses the gastrointestinal tract. release from membrane systems. In an elementary
Controlled Release Dose,

Fig. 1 The change in drug
concentration in blood
during drug delivery
456 Controlled Release Kinetics
osmotic pump, a table or capsule containing drug

in the core is surrounded by a membrane that is Controlled Release Kinetics
permeable to water but not drug. After
swallowing, water permeated into the core Sacide Alsoy Altinkaya
through the membrane dissolves drug, and a con- Department of Chemical Engineering, Izmir
stant osmotic pressure gradient established pumps Institute of Technology, Urla, Izmir, Turkey
the drug through a hole drilled into the membrane.
Zero-order release kinetics are observed so long as
a constant osmotic pressure gradient between core The release of an active agent from controlled
and surrounding medium is maintained. When release systems can be categorized into three
drug concentration falls below its solubility, then types: (1) zero-order release kinetics, (2) t1/2
rate of drug release starts to decay. To improve the kinetics, and (3) rst-order kinetics. In reservoir
efciency of osmotic systems, excipients such as (membrane) systems, the active ingredient is usu-
water-soluble polymers are added into core, or ally encapsulated between two semipermeable
drug formulation is placed between the water- membrane, and the release rate is controlled by
soluble polymer and the exit orice. A majority diffusion through the membrane. Reservoir sys-
of the coated dosage forms are controlled by tems generally provide a zero-order release where
solution-diffusion mechanism where the driving the release rate remains constant until all of the
force for the diffusional drug release is the drug active ingredient has been delivered. The release
concentration gradient across the rate-controlling kinetics from reservoir systems in the form of a
membrane. The drug release rate can be manipu- at sheet is described by the following equation:
lated by varying the membrane thickness and
membrane permeability. Permeability can be dMt DC
ADK (1)
modied by incorporating soluble additives or dt L
pore-forming agents in the membrane material
which allows the drug release not only by diffu- where
sion but also by osmotic pumping.
To develop a controlled release dosage form, Mt = mass of the active agent released up to time t
relevant pharmacokinetic parameters for the A = surface area
absorption, distribution, and elimination of a drug L = thickness of the rate-controlling membrane
as well as its in vivo release prole from the dosage D = diffusivity of the active agent through the
form should be considered. The dosage form membrane
should deliver the drug within 24 h which corre- K = partition coefcient of the active agent
sponds to the total gastrointestinal tract (GI) transit through the membrane
time in humans. To protect drug from hydrolysis in
acidic environment of the stomach, dosage forms When the concentration of the active ingredi-
are prepared with membranes made from ent within the reservoir is less than its saturation
pH-sensitive polymers that are insoluble in acid value, then the concentration difference through
but dissolve in the neutral or slightly alkaline the membrane decreases with time, and rst-order
environment of the gut. Drug is released rapidly if release kinetics is observed which can be
the molecular weight of the polymer is low. If a expressed as follows:
high molecular weight, water-soluble polymer is
used for coating, initially it swells into a gel dMt Cs ADK ADKt
exp (2)
through which drug diffuses. Following ingestion, dt L VL
the rate of swelling and diffusion through the
membrane affect the time for passage of the where Cs is the saturation concentration of the
dosage form through the stomach to the small agent and V is the volume of the reservoir and
intestine. t is time.
Convective Transport 457
In matrix systems, active ingredients are release controlled by relaxation follows zero-
dissolved or dispersed through a polymer and order release kinetics (Alfrey et al. 1966).
release rates from these systems decrease with
time. Release kinetics from matrix systems
containing dissolved active agents is described References
by a combination of rst-order and t1/2 kinetics
as follows (Baker and Londsdale 1974): Alfrey T, Gurnee EF, Lloyd WG (1966) Diffusion in glassy C
polymers. J Polym Sci C 12:249
Baker RW, Londsdale HK (1974) Controlled release:
Mt dMt D 1=2 mechanisms and rates. In: Tanquary AC, Lacey RE
0:6 2M1 (eds) Controlled release of biologically active agents.
M1 dt pL2 t Plenum Press, New York, pp 1572
2 Theeuwes F, Swanson D, Wong P (1983) Elementary
Mt dMt 8DM1 p Dt
0:6 exp 2 (3) osmotic pump for ndomethacin. J Pharm Sci 72:253
M1 dt L2 L
where M1 is the total amount of active ingredient

initially loaded into the matrix. When active agent
is dispersed throughout the polymer, t1/2 kinetics Convective Transport
is observed and the release rate is predicted by
Eq. 4 (Baker and Londsdale 1974): Boguslaw Kruczek
Department of Chemical and Biological

dMt A DCms 2C1 Cms 1=2 Engineering, University of Ottawa, Ottawa, ON,
(4) Canada
dt 2 t
where Cms is the saturation concentration of the

active ingredient in the polymer and C1 is the Introduction
concentration of active ingredient initially loaded
into polymer matrix. Convective transport (convective ow, bulk ow,
Osmotic-controlled release which is usually convection) may refer to either heat or mass trans-
reservoir-type systems follow zero-order release fer. Convective transport takes place both by dif-
kinetics as long as the osmotic pressure gradient fusion, i.e., the random Brownian motion of
through the membrane is constant. The release individual particles in the uid, and by advection,
rate of the active agent from these types of sys- in which matter or heat is transported by the
tems is dened with the following equation larger-scale motion of currents in the uid. Con-
(Theeuwes et al. 1983): vective transport is used to refer to the sum of
advection and diffusion (Bergman et al. 2011).
dMt APw DpCsa Convective mass transfer cannot take place in
(5) solids, because neither bulk current ows nor
dt L
signicant diffusion can take place in solids, but
where Pw is the permeability of water through the it can take place in porous solids. Diffusion of heat
membrane, Dp is the osmotic pressure difference, can take place in solids and is referred to as heat
and Csa is the solubility of the active agent in the conduction. In the context of membranes, convec-
membrane. tive ow refers to the transport of uid (pure
In swelling-controlled release systems, the solvent or solution) across a porous membrane,
release rate depends on the degree of swelling of which is driven by the applied pressure. Convec-
the matrix which is inuenced both by the diffu- tive transport is the main mode of transport in
sion of solvent into the polymer and relaxation of microltration and ultraltration processes; it
the polymer chains. When diffusion is the rate- may also contribute to the transport in diffusion
controlling step, t1/2 kinetics is observed and the dialysis if the pressure is applied. Convective
458 Convective Transport
transport may occur in gas separation membranes are then expressed in terms of the specic surface
if pinholes (defects) are present in the membrane area, av, and the porosity of the membrane, e,
structure, which is undesirable because it dimin- (Seader et al. 2011). Equation 2 is obtained by
ishes the selectivity of the membrane. substituting Eq. 3 into Eq. 1.
Equation 2 is sometimes presented in an alter-
native form (Mulder 1996):
Microfiltration and Ultrafiltration
e3 r
Ultraltration and microltration membranes are N Dp (4)
both considered as porous membranes in which K 1 e2 a2v mL
rejection is mainly determined by the size and
shape of solute relative to the pore size of the mem- where K is the Carman-Kozeny constant, which
brane and where the transport of solvent is directly depends on the shape of the pores and the tortu-
proportional to the applied pressure. For laminar osity. The commonly used value K = 5 implies
convective ow through a porous membrane, both the membrane tortuosity t = 2.5.
the Hagen-Poiseuille and the Carman-Kozeny equa-
tions can be applied (Mulder 1996). If the membrane
consists of straight capillaries, the Hagen-Poiseuille Gas Permeation
relationship is applicable in which the mass velocity
of uid (N) through the membrane is given by Gas separation requires membranes to be
nonporous (e.g., polymeric membranes) or
erD2 ultramicroporous, i.e., with pores smaller than
N Dp (1)
32mLt 0.7 nm diameter (e.g., zeolite and carbon mem-
branes). The pores greater than 2 nm are consid-
where r and m are the density and viscosity of the ered to be defects, because they offer limited
uid and e, D, L, and t are the porosity, pore (Knudsen ow) or no selectivity (convective
diameter, thickness, and tortuosity of the mem- ow) for gas separation. The actual mechanism
brane, respectively. For a cylindrical perpendicu- of gas transport through the defects depends on
lar pore, t = 1. If the pore is not straight, the the ratio of the pore diameter (D) and the mean
product Lt represents the actual length of the free path of gas molecules (l):
pore. When membrane has a nodular structure,
i.e., it is an assembly of spherical particles,
3m pRT 1=2
the pores are not cylindrical and straight; l (5)
Carman-Kozeny equation is then employed 2p 2M
(Seader et al. 2011):
where T is the absolute temperature, R is the
3
e r universal gas constant, P is the pressure, and M is
N Dp (2) the molecular weight of the gas. If D/l 1, i.e., the
21 e2 a2v mtL
collisions between gas molecules are more frequent
than the collisions with the pore walls, and there is a
Equation 2 is obtained by replacing the pore diam-
pressure gradient across the membrane, the con-
eter in Eq. 1 by the hydraulic diameter (DH):
vective transport dominates. Since gases are com-
pressible, Eqs. 1, 2, and 4 are applicable in a
4Ap 4e differential form. Using the ideal gas law, gas den-
DH (3)
P av 1 e sity can be expressed by
where Ap is the cross-sectional area of the pore pM

r (6)
and P is the wetted perimeter of the pore, which RT
Conventional Pretreatment of Water 459
Assuming defects to be straight cylindrical pores, References

Eq. 1 becomes:
Bergman TL, Lavine AS, Incropera FP, Dewitt DP
2 (2011) Fundamentals of heat and mass transfer,
eD Mp dp
N (7) 7th edn. Wiley, New York
32mRT dx Kessler SB, Klein E (1992) Dialysis. In: Ho WSW, Sirkar
KK (eds) Membrane handbook. Van Nostrand
At steady state the ux is constant; integrating Reinhold, New York, pp 161215 C
Kesting RE, Fritzsche AK (1993) Polymeric gas separation
Eq. 7 from the feed side (x = 0, p = po) to the membranes. Wiley, New York
permeate side (x = L, p = pL) leads, after Mulder M (1996) Basic principles of membrane technol-
rearrangements, (Kesting and Fritzsche 1993) to: ogy. Kluwer, Dordrecht
Seader JD, Henley EJ, Roper DK (2011) Separation pro-
cess principles. Wiley, Hoboken
eD2 M po pL po pL
N (8)
32mRT 2 L
where po and pL are the feed and permeate pres-

sures, respectively. Conventional Ball Milling and
Calcination
Dialysis Perovskite Powder Preparation Methods
Diffusion dialysis is a rate-governed membrane

process in which solutes are driven across a porous
membrane by means of a concentration gradient.
When a pressure gradient is applied, in addition to Conventional Pretreatment of Water
diffusive transport, the solutes may also be
transported by the convective ow. The pressure A. Moura Bernardes
gradient may be applied in order to concentrate the Universidade Federal do Rio Grande do Sul
feed, or may be obligatory as a consequence of the (PPGE3M-UFRGS), RS, Porto Alegre, Brazil
geometry of the dialysis device and the desired feed
ow rate. On the other hand, in the case of highly
concentrated feed, the convective ow may occur Conventional pretreatment of water is the process
from diluted dialysate (permeate) to the feed in and operation applied to the treatment of water or
response to an osmotic pressure gradient. At steady wastewater prior to the feeding of the membrane
state, the solute ux (Ns) at a point on the mem- separation process (MSP).
brane consists of the sum of a diffusive and con- In membrane processes, the nature of the feed
vective component (Kessler and Klein 1992): stream is very important, since the presence of
colloidal material leads to clogging and adsorp-
tion phenomena that alter the mechanisms of
N s Pm Dcs N v 1 scs (9) operation (Minhalma 2001). Many feed solutions
used for the production of potable or industrial
where Pm is the diffusive permeability of solute, process water contain suspended and colloidal
Dcs is the concentration gradient of solute across matter or soluble salts close to their saturation
the membrane, Nv is the volumetric ux level. These components cause severe membrane
(supercial velocity) of solvent, s is the reection fouling and poisoning. A pretreatment of the feed
coefcient, and cs is the average solute concentra- solution, which guarantees the removal of these
tion within the membrane. The second term of materials, is therefore mandatory for a trouble-
Eq. 9, N v 1 scs , represents the convective free operation of membrane separation processes
component of the total solute transfer. (Strathmann 2004).
460 Conventional Pretreatment of Water
Water pretreatment processes are chosen pretreatment will also depend on the feed water
mainly based on the initial type (seawater, brack- composition and its variability, but will always
ish water, surface water, wastewater) and quality comprise a step of solids and turbidity removal
of the water with respect to the type of foulants and other step of organic compounds removal.
(suspended solids, bacteria, organics). Typical Most often, it uses physicochemical processes to
methods of pretreatment are the addition of an remove solids and activated carbon lters to
oxidant (to remove bacteria, algae), coagulation/ remove organics (Bernardes et al. 2014).
occulation (to remove suspended solids), pH By NF and RO processes, it is important to
adjustment (to prevent scaling), adsorption on prevent scaling by silica and sparingly soluble
activated carbon, and all types of lters such as salts such as calcium carbonate and calcium sul-
multilayers and cartridge lters (Nobles and Stern fate. Besides silica, calcium carbonate, and cal-
1995). The integration of otation processes to cium sulfate, the other salts that most commonly
coagulation/occulation steps is also envisaged form scale are barium sulfate, strontium sulfate,
(Geraldes et al. 2008). Figure 1 presents a sum- and calcium uoride (Rautenbach and Melin
mary of the pretreatment processes applied to 2004; Baker 2012). Scale control consists of pH
MSP. adjustment and a softening process. After soften-
Besides the feed quality, a classication must ing, the efuent is usually further claried, for
also be made among different membrane separa- example, by sand ltration (Scott 1995). Filtration
tion processes. Each desalting technology has dif- is used to remove particulate matter that will clog
ferent requirements for the quality and condition the feed channels or accumulate on the membrane
of water entering the process. For membrane pro- surface. The minimum ltration requirement
cesses, the major concerns are membrane fouling regardless of the water source is a cartridge lter
and scaling, suspended solid plugging, biological rated in the 125 mm range with a typical rating of
fouling or attack, membrane degradation by oxi- 5 mm (Davis 2011). The RO units are tted with a
dation, or other means (Watson et al. 2003). 0.45 mm cartridge lter, and a sand lter, some-
To lower pressure processes as microltration times supplemented by addition of a occulating
(MF) and ultraltration (UF), pretreatment may chemical such as alum or a cationic polymer, may
include chemical coagulation, chlorination, be required (Baker 2012). Disinfection by chlori-
screening, and ow equalization. If the raw nation, ozonation, or UV irradiation may also be
water turbidity and/or NOM (natural organic mat- required to prevent biofouling (Davis 2011). If
ter) concentration is high, pretreatment will chlorine is used, residual chlorine must be
include coagulation, occulation, and sedimenta- removed when polyamide and interfacial compos-
tion. The effect of coagulation is site specic due ite membranes are used, since these are chlorine
to interactions between the coagulants, raw water sensitive. The development of low-cost MF/UF
components, and the membrane. In some cases, membrane processes that remove particulate and
especially for UF processes, low coagulant doses bacteria has encouraged the use of these mem-
may cause greater fouling than no coagulation, branes as a pretreatment step for RO (Baker
but higher doses for enhanced coagulation fre- 2012).
quently reduce fouling. When iron and manga- Although for electrodialysis the water
nese are prevalent in the raw water, oxidation pretreatment is also an important task, comparing
may be performed to form a precipitate that can to RO, electrodialysis (ED) demands less
be removed by coagulation. Nevertheless, the use pretreatment operations. ED only removes ions.
of oxidants requires careful selection of the mem- Therefore, any bacteria, colloidal material, or sil-
brane material, as well as precautions to remove ica present in the feed water stream will remain in
residual oxidant before the membrane treatment the product stream. This can be a disadvantage,
step (Davis 2011). especially for the production of drinking water,
In nanoltration (NF), reverse osmosis (RO), because as only ions are removed, uncharged
and electrodialysis (ED), the selection of the components such as microorganisms or organic
Conventional Pretreatment of Water 461
Conventional
Conventional
Pretreatment of Water, Raw Water
Membrane
pre-treatment
Fig. 1 Summary of the processes
Separation Processes
pretreatment processes
applied to MSP
Turbidity and/or Physicochemical processes: pH
Natural Organic adjustment, coagulation, flocculation,
Matter sedimentation, flotation
C
Chemical oxidation followed by
Iron and Manganese coagulation, flocculation,
sedimentation
Screening (strainers) and Filtration

Suspended Solids
(granular filtration, bag filtration)
Oxidation by chlorine, ozone or UV

Microbial Population radiation
Flow equalization
Microfiltration
Ultrafiltration
Oxidation by chlorine, ozone or UV

Microbial Population
radiation
Organic Matter,
Excess of Chlorine, Adsorption (activated carbon)
Metals
Suspended Solids Filtration (cartridge filters)
Nanofiltration
Reverse Osmosis
Electrodialysis
contaminants are not eliminated (Bernardes ED plant (Watson et al. 2003). By ED, there is also
et al. 2014). On the other hand, it brings benets a possibility of fouling minimization by reversing
when, for example, silica is present. Silica occurs the polarity of the system several times per hour
in varying concentrations in all natural waters. RO with electrodialysis reversal (EDR). By reversing
membranes reject silica, but silica is largely unaf- the polarity (and the solutions ow direction),
fected by passage through an ED system. Conse- ions move in the opposite direction through the
quently, silica precipitation is not of concern in membranes, minimizing buildup. This generally
462 Conventional Pretreatment of Water
minimizes membrane fouling, decreasing constituent of groundwater along the coast or at

pretreatment requirements in comparison to RO other locations where wetlands have existed.
(Bernardes et al. 2014). Since microorganisms are Hydrogen sulde is a by-product of the life cycle
not removed from a feed stream by electrodialy- of sulfate-reducing bacteria (SRB) that are com-
sis, most commonly a chlorination of the product mon in many groundwater systems. Like all gases,
water is required. When the concentrate stream H2S is passed by RO membranes and appears in
has been acidied, the product water pH value both the permeate and the concentrate, not being
must be readjusted. Other measurements may be of concern in RO plants. On the other hand, H2S
necessary depending on the feed water composi- must be removed from the feed water to ED/EDR
tion and required product water quality plants to prevent oxidation of H2S in the stack and
(Strathmann 2004). Feed waters containing sig- the resultant membrane fouling by colloidal sul-
nicant quantities of calcium carbonates and fur. Oxidation occurs by reaction with both the
bicarbonates are likely to cause precipitation in chlorine and oxygen produced by electrolysis in
the concentrate stream at the anion-exchange the process. The most common treatment is air
membrane surface due to pH changes resulting stripping, followed by (usually) chlorination to
from water splitting which is obtained when the complete the oxidation of the remaining sulde,
limiting current density is exceeded locally and granular ltration (Watson et al. 2003).
(Strathmann 2004). Organic fouling of ion
exchange membranes is one of the major prob-
lems in electrodialysis. It is caused by the precip-
References
itation of colloids on the membranes, and since
most of the colloids present in natural water are Baker RW (2012) Membrane technology and applications.
negatively charged, it is usually the anion- Wiley, Chichester
exchange membranes which are affected. Certain Bernardes AM, Rodrigues MAS, Ferreira JZ (eds)
detergents are also the cause of this type of poi- (2014) Electrodialysis and water reuse. Springer, Berlin
Davis ML (2011) Water and wastewater engineering:
soning, which is very difcult to deal with and can design principles and practice. McGraw Hill Interna-
best be avoided by a proper pretreatment of the tional, New York
feed solution (Strathmann 2004; Baker 2012). Geraldes V, Anil A, De Pinho MN, Duarte E (2008)
The usual method for removing colloidal material Dissolved air otation of surface water for spiral-
wound module nanoltration pre-treatment. Desalina-
is the pretreatment by coagulation-occulation tion 228:191199
and ltration. In this step, some chemicals are Minhalma M (2001) Synthesis and optimization of pro-
used to assist in removal of solids and turbidity cesses for the recovery of industrial wastewaters with
as coagulants, occulants, acids, or bases to adjust ultraltration and nanoltration. Ph D thesis, IST, Tech-
nical University of Lisbon
pH and oxidizing agent for organic matter degra- Nobles RD, Stern SA (1995) Membrane separations
dation. When using membrane process for solid technology: principles and applications,
removal, a preliminary step of coagulation may be vol 2, Membrane science and technology series.
needed, to increase the efciency of removal of Elsevier, Amsterdam
Rautenbach R, Melin T (2004) Membranverfahren:
colloids and/or oxidation. Organic compounds grundlagen der Modul- und Anlagenauslegung.
that were not degraded in biological treatment Springer, Berlin
can be removed in activated carbon lters. In Scott K (1995) Handbook of industrial membranes.
this step, it is important to control the generation Elsevier, Oxford
Strathmann H (2004) Ion exchange membrane separation
of ne carbon and, if necessary, adopt operational processes, vol 9, Membrane science and technology
changes in order to prevent clogging of the car- series. Elsevier, Amsterdam
tridge lter and passage into the membrane mod- Watson IC, Morin OJ, Henthorne L (2003) Desalting hand-
ules of the electrodialysis. Cartridge lters are also book for planners, desalination and water purication,
3rd edn. Research and Development Program, report
used upstream of electrodialysis as the ultimate no 72. United States Department of the Interior. Bureau
protection against suspended solids (Bernardes of Reclamation. Technical Service Center. Water Treat-
et al. 2014). Hydrogen sulde, H2S, is a frequent ment Engineering and Research Group
Copolyimide Precursors 463
such as in electronics, aircraft, automobile, medi-

Copolyimide Precursors cal, machining, gas purication, aerospace, and
military applications.
Tauqir A. Sherazi For application in electronic and microelec-
Department of Chemistry, COMSATS Institute of tronics industry, uorine may be introduced to
Information Technology, Abbottabad, Pakistan polyimides which reduced the dielectric constant
along with retaining the other characteristics men- C
tioned above (Hermciuc et al. 2000). Formation of
Polyimide Precursors nanofoam polyimides is a novel approach for
reducing the dielectric constant. Basically in
Polyimides are a class of thermally stable poly- nanofoam formation, polyimide polymer replaces
mers that are often based on stiff aromatic back- with air which reduces the dielectric constant
bones derived from aromatic dianhydrides and (Hedrick and Charlier 1994). The membrane
aromatic diamines. Polyimides due to their developed by polyimide containing functional or
unusual properties are nding a wide range of pendant groups in the backbone of the aromatic
applications and thus used as precursor dianhydride and/or diamine exhibits good as well
(a precursor is a compound that participates in as selective gas permeability.
the chemical reaction that produces another com- Hyperbranched polyimide (HBPI) precursors
pound) to develop materials for various applica- were also studied to develop high-performance
tions. The chemistry of polyimides is in itself a carbon molecular sieve membranes for improved
vast area with a large variety of monomers avail- gas separation applications. It was found that
able and several methodologies available for the unique hyperbranched network structure
synthesis. found in HBPI possesses great potential to pro-
The general formula and chemical structure of duce carbon molecular sieve membranes with
polyimide is shown in Fig. 1. superior performance (Sim et al. 2013). It is also
Few of the specialty properties of polyimides reported that the membrane produced from
include: the pyrolysis of a hollow-ber polyimide
precursor under suitable conditions has good sep-
aration properties when applied to mixed gas
High thermal and thermo-oxidative stability up to
pairs, such as O2/N2, CO2/CH4, and H2/CH4
400 C (750 F)
(Jones and Koros 1994). Polyimide-based mem-
Excellent mechanical properties, both at room
branes have been extensively studied for removal
temperature and elevated temperatures
of CO2 particularly from natural gas and found
Film- and ber-forming ability
few of them quite efcient (Xiao et al. 2009).
Excellent adhesive properties, both at room tem-
Soluble and optically transparent uorine-
perature and elevated temperature
containing photoreactive polyimide precursors
Nonammability will not support combustion
were developed. These precursors offer high-
resolution patterns with aspect ratio of more than
Applications of Polyimide Precursors 2.0. In these polyimide precursors, the polymers
which have a benzophenone segment in the poly-
Owing to its excellent chemical and thermal sta- mer backbone are self-sensitized and show inter-
bility, polyimides have wide range of applications esting photochemical reactions (Omote
et al. 1989).
Copolyimide Precursors, Polyimide processability is one of the issues
Fig. 1 Chemical structure associated which cause hurdle to manufacture
of polyimide polyimide parts, such as composites, at costs com-
petitive to other metal parts. Improvement in
polyimide processability is essential which could
464 Copper Removal and Recovery by Supported Liquid Membranes
be achieved by reduction in its melt viscosity.

Low-molecular-weight end-capped oligomers Copper Removal and Recovery by
are potential candidates for the development of Supported Liquid Membranes
polyimide oligomers for improved processability
(Smith and Connell 2000; Simone and Scola Carlos Moreno
2000). Department of Analytical Chemistry, Faculty of
Marine and Environmental Sciences, University
of Cdiz, Puerto Real, Spain
References
Hedrick JL, Charlier Y (1994) High temperature polyimide Copper is an element that occurs naturally as a
nanofoams. Polym Prepr 35:245346
metal or as part of sulde minerals, such as chal-
Hermciuc S, Hamciuc E, Sava I, Diaconu I, Bruma
M (2000) New uorinated poly(imide-ether-amide)s. copyrite or bornite, as well as oxides and
High Perform Polym 12:205276 carbonate ores.
Jones CW, Koros WJ (1994) Carbon molecular sieve gas It has been used by human being from the
separation membranes-I. Preparation and characteriza-
tion based on polyimide precursors. Carbon
ancient age and today is still extensively applied
32:14191425 in many areas that cover a wide variety of differ-
Omote T, Koseki K, Yamaoka T (1989) Soluble and opti- ent disciplines. It is used mostly as copper metal
cally transparent uorine-containing photoreactive or copper alloys (electrical devices and wires,
polyimide precursors: Spectral sensitization by organic
coins, pipes, automotive industry, etc.) and also
peroxide and organic dye combination. Polym Eng Sci
29(14):945949 as copper salts (agriculture, water purication,
Sim YH, Wang H, Li FY, Chua ML, Chung T-S, Toriida M, etc).
Tamai S (2013) High performance carbon molecular This extensive use produces a huge amount of
sieve membranes derived from hyperbranched
residual wastes containing copper, which may
polyimide precursors for improved gas separation
applications. Carbon 53:101111 lead to environmental pollution and then justify-
Simone C, Scola DA (2000) Novel uorinated ing a previous treatment to remove and recover
polyimidess. Proc Fluoro Polym Am Chem Soc 1518 copper ions.
Smith JG Jr, Connell JW (2000) Chemistry and properties
of imide oligomers fom phenylethynyl containing
Although in general the stability of supported
diamines. High Perform Polym 12:213223 liquid membranes may represent an important
Xiao Y, Low BT, Hosseini SS, Chung TS, Paul DR difculty for large-scale operations, recovery of
(2009) The strategies of molecular architecture and copper was successfully conducted at pilot scale
modication of polyimide-based membranes for CO2
on a hollow-ber membrane contactor with a sur-
removal from natural gas a review. Prog Polym Sci
34:561580 face area of 130 m2 (Yan and Kocherginsky
2006).
Copper Removal
and Recovery by
Supported Liquid
Membranes,
Fig. 1 Chemical structure
of a ketoxime (1) and
5-nonylsalicylaldoxime (2)
Corrosion 465
On the other hand, copper is an essential ele- of iron and steel. Other metals do corrode. The
ment that takes part in several biogeochemical oxide formed by oxidation does not rmly adhere
processes, most of them at very low concentra- to the surface of the metal and akes off easily
tions. This fact has led to the growing develop- causing pitting. Continued pitting causes struc-
ment of hollow-ber supported liquid membrane tural weakness and disintegration of the metal.
systems for analytical sample treatment to the Metals such as aluminum form a tough oxide
determination of copper traces in environmental coating onto the surface to prevent corrosion. C
samples. In some circles, corrosion has been described
Copper separation though supported liquid as the tendency of a metal to revert to its natural
membranes may be performed by using a variety stable state as an ore. The process may involve a
of carriers in the organic membrane, mainly second step in which an oxide, hydroxide, or
organophosphorus acidic compounds (see Cobalt carbonate of the metal may form and deposit at
removal and recovery by supported liquid mem- the corrosion site. Thus, corrosion takes place
branes) or commercial mixtures of oximes, when a metal dissolves in or is disintegrated by
mostly ketoximes (1) and salicylaldoximes, such water.
as 5-nonylsalicylaldoxime (2), which form com- Different metals have different tendencies to
plexes with copper ions (Fig. 1). corrode or not to corrode in the same water.
A certain type of water may corrode one metal
and not another, but the reverse may be true for
References another type of water. An excellent method for
corrosion control in water heaters is cathodic pro-
Yang Q, Kocherginsky NM (2006) Copper recovery and tection involving the use of a sacricial anode,
spent ammoniacal etchant regeneration based on hol-
usually composed of magnesium or aluminum.
low ber supported liquid membrane technology: from
bench-scale to pilot-scale tests. J Membr Sci Chemical control of corrosion attempts to retard
286:301309 electrode reactions.
The Mechanism and Chemistry of Corrosion

Reactions
Core Shell Particles Corrosion reactions are electrochemical in nature.
They involve the transfer of charged ions across
Membrane Emulsication in Phase Separation the surface between a metal and the electrolyte in
for Microcapsule Preparation which the metal is immersed. There are two types
of electrode reaction occurring at the metal sur-
face: anodic and cathodic.
Corrosion occurs in the presence of moisture.
Corrosion When iron is exposed to moist air, it reacts with
oxygen to form rust. The amount of water http://
C. V. Vedavyasan www.corrosion.com, precisely the number of
Department of Technology, Annamalai molecules of water, present determines the color
University, Chidambaram, Tamil Nadu, India of rust varying from black to yellow to orange
brown. The formation of rust is a very complex
process which begins with the oxidation of iron to
Introduction ferrous (+2) ions.
In an anodic reaction, electrons appear on the
Corrosion is the gradual process of wearing, right-hand side. An example is shown below:
weaning away, and thinning down the metal sur-
face by the uid with which it is in contact. Mil- Fe > Fe2 2e :
lions of dollars are lost every year due to corrosion
466 Corrosion
Both water and oxygen are required for the involvement of water accounts for the fact that
sequence of reactions. Iron (+2) ions are further rusting occurs much more rapidly in moist envi-
oxidized to form ferric ions (iron+3) ions: ronment as compared to dry one. The mechanism
of corrosion is shown in Fig. 1 below.
Fe2 > Fe3 1e
Influence of Salt on Corrosion
The electrons provided from both oxidation steps The presence of salt enhances the rate of corro-
are used to reduce oxygen as shown. Cathodic sion. As the dissolved salt increases, the conduc-
reactions involve electrochemical reduction tivity of the aqueous solution at the metal surface
wherein electrons appear on the left side: enhances the rate of electrochemical corrosion.
Hence iron and steel corrode more rapidly when
O2 g 2 H2 O 4e_ > 4 OH exposed to salt or moist salty air near the ocean.
The ferric ions then combine with oxygen to form Factors Associated Mainly with the Metal
ferric oxide [iron (III) oxide] which is then Effective electrode potential of a metal in a
hydrated with varying amounts of water. The solution
reduction of oxygen at an electrode will cause a Overvoltage of hydrogen on the metal
rise in pH due to production of hydroxide ion. The Chemical and physical homogenity of the
equation for rust formation is metal surface
Inherent ability to form an insoluble protective
4Fe2 aq O2 g 4 2 H2 Ol lm
> 2Fe2 O3 H2 O 8 H aq :
Factors Varying with the Environment
The formation of rust can occur at some distance pH of the solution
from the actual pitting or erosion of iron. The Inuence of oxygen in solution adjacent to the
electrons produced via the initial oxidation of iron metal
can be conducted through the metal, and the iron
ions can diffuse through the water layer to another Forms of Corrosion
point on the metal surface where oxygen is avail-
able. This process results in an electrochemical cell Galvanic Corrosion
in which iron serves as anode, oxygen gas as the This is an electrochemical action of two dissimilar
cathode, and iron solution as salt bridge. The metals in the presence of an electrolyte and an
electron conductive path. It occurs when
Corrosion, Fig. 1 The

mechanism of corrosion
Corrosion 467
Red brass
Aluminum bronze
Admiralty brass
Yellow brass
76 Ni-16 Cr-7 Fe alloy (active)
Nickel (active)
Naval brass C
Manganese bronze
Muntz
Tin
Lead
18-8-3 stainless steel, type 316 (active)
18-8 stainless steel, type 304 (active)
Thirteen percent chrome stainless steel, type
410 (active)
Cast iron
Mild steel
Corrosion, Fig. 2 Galvanic corrosion Aluminum 2024
Cadmium
Aluminum 6053
dissimilar metals are in contact. It is recognizable Galvanized steel
by the presence of a buildup of corrosion at the Zinc
joint between the dissimilar metals. When alumi- Magnesium alloys
num or magnesium alloys are in contact with Magnesium
carbon or stainless steel, galvanic corrosion
occurs and accelerates the corrosion of aluminum Anodic (Most Active) The natural differences in
or magnesium (Fig. 2). metal potentials produce galvanic differences, such
as the galvanic series in sea water. If electrical
Galvanic Series in Sea Water contact is made between any two of these materials
(Least Active) in the presence of an electrolyte, current must ow
Gold between them. The farther apart the metals are in
Graphite the galvanic series, the greater the galvanic corro-
Silver sion effect or the rate. Metals or alloys at the upper
18-8-3 stainless steel, type 316 (passive) end are noble while those at the lower end are
18-8 stainless steel, type 304 (passive) active. The more active metal is the anode or the
Titanium one that will corrode. The anode must be chosen
Thirteen percent Cr stainless steel, type from the above material list which is lower on the
410 (passive) list than the material to be protected.
7 Ni- 33 Cu alloy Control of galvanic corrosion is achieved by
75Ni 33 Cu alloy using metals closer to each other in the galvanic
75Ni 16 Cr 7 Fe alloy (passive) series or by electrically isolating metals from each
Nickel (passive) other. Cathodic protection can be used to control
Silver solder galvanic corrosion effects.
M-Bronze The following practices are recommended to
G- Bronze keep galvanic corrosion to a minimum.
70-30 cupro-nickel
Silicon bronze Avoid the use of widely dissimilar metals in
Copper direct contact.
468 Corrosion
Corrosion,
Fig. 3 Isolation anges to
prevent galvanic
When dissimilar metals must come into con- stainless steel resist corrosive media and can per-
tact, they should be separated by using form well over long periods of time. However, if
nonconductive barrier materials, a paint coat- corrosion does occur, it forms at random pits.
ing, or by plating. Pitting may be a serious type of corrosion because
The anode should be as large as feasible in it tends to penetrate rapidly into the metal section.
relation to cathode. It is most likely to occur in the presence of chlo-
Coat both the anode and the cathode with the ride ions, combined with such depolarizers as
same material. oxygen or oxidizing salts. Methods that can be
Install fasteners that have been dipped in epoxy used to control pitting include maintaining clean
mastic coatings. surfaces, application of a protective coating, and
Seal threaded inserts with epoxy mastic coat- use of inhibitors or cathodic protection for immer-
ings prior to insertion into castings. sion service.
Avoid the use of lock or toothed washers over Sometimes pitting corrosion can be quite small
plated or anodized surfaces. on the surface and very large below the surface.
The gure below left shows this effect common
The scuba tank above suffered galvanic corro- on stainless steel and other lm-protected metals.
sion when the brass valve and the steel tank were Pitting shown on the right (white arrow) led to the
wetted by condensation. Electrical isolation stress corrosion fracture shown by the black
anges (Fig. 3) like those shown on the right are arrows (Fig. 4).
used to prevent galvanic corrosion. Insulating Pitting corrosion can lead to unexpected cata-
polymeric gaskets are inserted between the strophic system failure. The split tubing below left
anges, and insulating sleeves and washers isolate was caused by pitting corrosion of stainless steel.
the bolted connections. A typical pit on this tubing is shown below right
(Fig. 5).
Pitting Corrosion
The most common effect of corrosion on alumi- Stress Corrosion Cracking
num and magnesium alloys is pitting. First is Stress corrosion cracking (SCC) is caused by the
noticeable as a white or gray powder, similar to simultaneous effects of tensile stress and corro-
dust, which blotches the surface. When the sion. Stress may be internally or externally
deposit is cleaned away, tiny pits or holes can be applied. Internal stresses are produced by
seen in the surface. Passive metals such as nonuniform deformation during cold working,
Corrosion 469
Corrosion, Fig. 4 Same images for pitting corrosion
Corrosion, Fig. 5 Pitting corrosion for stainless steel tube
by an unequal cooling from high temperatures, Crevice Corrosion

and by internal structural rearrangement involving Crevice or contact corrosion is produced at the
volume changes. Stresses induced when a piece is region of contact of metals with metals or metals
deformed, those induced by press and shrink ts, with nonmetals. It may occur at washers, under
and those in rivets and bolts are internal stresses. barnacles, at sand grains, under applied protective
Concealed stress is more important than design lms, and at pockets formed by threaded joints.
stress, especially because stress corrosion is dif- Whether or not stainless steels are free of pit
cult to recognize before it has overcome the nuclei, they are always susceptible to this kind of
design factor. Few guidelines in avoiding the corrosion because a nucleus is not necessary.
problem are: Crevice corrosion may begin through the action
of an oxygen concentration cell and continue to
Use metal alloys at no greater than 75 % of form pitting. Cleanliness, the proper use of seal-
their yield strength and use exotic materials ants, and protective coatings are effective means
only where they are actually required. of controlling this problem.
Avoid assemblies where high-tensile loads are
concentrated in a small area. Intergranular Corrosion
Remove stress risers from counter bores, This is an attack on the grain boundaries of a metal
grooves, etc. or alloy. A highly magnied cross section of any
470 Corrosion
commercial alloy will show its granular structure. appropriate. Failure to adequately clean all resi-
This structure consists of quantities of individual dues will permit corrosion to continue. Light cor-
grains, and each of these tiny grains has a clearly rosion may be removed from thin
dened boundary that chemically differs from the members ducts and tubing with a nonwoven,
metal within the grain center. One example of this nonmetallic abrasive mat in accordance with
type of corrosion is in unstabilized 300-series MIL-A-9962 or number 400 aluminum oxide or
stainless steels sensitized by welding or brazing silicon carbide grit abrasive paper or cloth. Do not
and subsequently subjected to a severe corrosion use steel wool.
environment. Another example of intergranular or
grain boundary corrosion is that which occurs Chemical Method
when aluminum alloys are in contact with steel Chemical corrosion-removal methods can be used
in the presence of an electrolyte. The aluminum when no danger exists that the chemical used will
alloy grain boundaries are anodic to both the become entrapped in recesses and when there is
aluminum alloy grain and the steel. In the later no danger that adjacent material will be attacked.
case, intergranular corrosion of the aluminum The chemical method should be used on complex
alloy occurs. Some austenitic steels are unstable shapes and machined surfaces. Chemical rust
when heated in the temperature range removers are of two types: acid or alkaline. The
470915 C. acid type is used in removing rust and black oxide
Decreased corrosion resistance in austenitic by immersion or brush application. This is a phos-
stainless steels is due to depletion of chromium phoric acid-type remover and must not be used on
in the area near the grain boundaries, caused by high-strength steel heat treated above 1.24
the precipitation of chromium carbide. This con- gigapascals (GPa) tensile strength due to possible
dition can be eliminated by the use of stabilized stress corrosion or hydrogen embrittlement prob-
stainless steel, such as columbium-, tantalum-, or lems. The alkaline type (sodium hydroxide base)
titanium-stabilized stainless steels (types is suitable for use by immersion only. Use on
321/347), or by the use of low-carbon stainless machined surfaces where a dimensional change
steels. Molybdenum additions as in type 316 stain- would be objectionable.
less steels decrease the sensitivity to and the sever- Scale conditioners are used as necessary to
ity of the intergranular attack. facilitate oxide scale removal by acid cleaning.
Intergranular corrosion can be prevented by: The use of scale conditioners shall not cause
pitting, intergranular attack, or reduction of
Select an alloy type that is resistant to mechanical properties below the minimum
intergranular corrosion. values as specied in the applicable material spec-
Avoid heat treatments or service exposure that ication for the alloy, gage, and heat-treat
makes a material susceptible. Normally this condition.
occurs with austenitic stainless steels when When acid cleaning is used to remove heat-
they are held for some time in the sensitizing treat scale, ux, corrosive media, stains, and other

temperature range of 470915 C contaminants, it shall be within the limits speci-

(8001600 F). ed. Acid cleaning shall not result in intergranular
Apply a protective coating. attack that would be detrimental to the fabrication
or use of the material or part. Acid cleaning shall
Corrosion Removal not result in pitting or smutting, which will not be
Abrasive blasting is the preferred method of readily removed by subsequent processing.
removing corrosion; other mechanical methods Cleaning shall be accomplished in the following
used are grinding, chipping, sanding, or wire bath:
brushing.
Remove corrosion by mechanical method such 1. Nitric acid (HNO3) (42 Baume)
as wire brushing or abrasive blasting as 225375 kg/m3
Corrosion 471
2. Hydrouoric acid (HF) or NH4HF2 preparation prior to application of a protective

952 kg/m3 coating. A brush coat of primer prior to spray
3. Temperature: room 60 C (140 F) will ensure adequate coverage.
4. Metal content, HF ratio = 1: 1.8 (replenish Tube clamps. Carbon steel clamps for interior
bath when the metal concentration >1 part of applications are either zinc plated or painted.
metal to 1.8 parts of HF) Stainless clamps shall be used for all exterior
applications. Corrosion at the interface C
Stainless Steel Alloys between the stainless steel tubing and the
Stainless steels owe their inherent corrosion resis- clamp can occur due to dissimilar metal and
tance to a condition known as passivity as a crevice corrosion. Corrosion at the interface
result of the presence of their oxide lms called between tubing and clamps is controlled by
passive lms. Under favorable conditions, such application of protective coatings.
lms are protective; however, unfavorable condi- Unistrut channels. Use of unistrut channels
tions decient in oxygen will destroy the lms and should be avoided in exterior locations. When
leave the surface in an active state with corro- the exterior use of unistrut cannot be avoided,
sion resistance comparable to carbon steel. The selection of appropriate material shall be
presence of hygroscopic salt deposits, dirt, dust, utilized, such as stainless steel or ber
and other foreign matter all serve to destroy pas- glass. Where corrosion is noted, mechanically
sivity. Underground exposure of bare stainless clean the member to remove corrosion prod-
steel will result in unacceptable quality. Where ucts followed by application of a zinc-rich
localized corrosion occurs, rapid penetration coating.
(pitting) at the point of initiation can occur as an
electrolytic cell is formed between the large Stainless Steel Components
cathodic (passive) area and the small anodic area
under attack. This attack is particularly severe in General Stainless steel, although very corrosion
the presence of halide salts. Localized attack will resistant, is susceptible to localized corrosion
also occur in crevices, such as under sleeves on (e.g., pitting, crevice corrosion, etc.) when
tube ttings. exposed to marine environment. A mistake fre-
Superior resistance to pitting is attainable with quently made is to conclude that the corrosion
type 904 L unied numbering system (UNS) noted on stainless steel tubing and bellows is
N08904 stainless steel over other commonly only supercial. This conclusion is improperly
used steels. However, this is only a matter of reached when removal of external corrosion prod-
degree and localized attack can still occur. ucts leaves the surface in a condition that appears
Maintaining clean surfaces will greatly reduce almost like new except for what appears to be a
the opportunity for corrosion, regardless of the very tiny pit. A cross section taken through such a
alloy employed. typical pit frequently discloses a void consider-
ably greater in diameter than the surface pit diam-
Typical Problem Areas eter. Failure to arrest the apparent supercial
Sharp edges. Sharp edges of metal structures corrosion will result in ultimate penetration of
will often be decient of proper coating thick- thin wall members.
ness; sharp edges should be rounded when
possible with the National Association of Cor- Application of Protective Coatings
rosion Engineers (NACE). A stripe coat or Stainless steel tubings shall be treated in the fol-
brush coat of primer prior to spray will assist lowing way:
in getting adequate coverage.
Nuts and bolts. Premature coating failure and Accumulated dirt and oil shall be removed by
corrosion on nuts and bolt heads are common rinsing with water followed by rinsing with
and can be reduced by conscientious surface methyl ethyl ketone (MEK).
472 Corrosion
Remove corrosion products by mechanical Cathodic protection can, in some cases, pre-
means, such as power tool cleaning or hand- vent stress corrosion cracking.
tool cleaning.
Clean surfaces with methyl ethyl ketone using Applications Pipelines are routinely protected
clean rags. by a coating supplemented with cathodic protec-
Apply by spraying, brushing, or dipping 75 mm tion. An ICCP impressed current cathodic
(3 mils) minimum of the coating Aerocoat protection system for a pipeline would consist
AR-7 manufactured by B.F.Goodrich. of a DC power source, which is often an
AC-powered rectier and an anode, or array of
Tubing assemblies shall be abrasive blasted. anodes buried in the ground (the anode ground
When tubing assemblies are in close proximity head)
to carbon steel structural members that are to be Ships: cathodic protection on ships is often
abrasive blasted and coated with inorganic zinc- implemented by galvanic anodes attached to the
rich primer, the tubing assemblies shall be simi- hull, rather than using ICCP. Since ships are reg-
larly treated. ularly removed from the water for inspections and
maintenance, it is a simple task to replace the
Using clean rags, accumulated dirt and oil shall galvanic anodes. Galvanic anodes are generally
be removed with water followed by wiping shaped to reduced drag in the water and tted
with methyl ethyl ketone. ush to the hull to also try to minimize drag.
Apply a zinc-rich coating in accordance
with the manufacturers recommendations to How to Prevent Metal Corrosion
a DFT (dry lm thickness) of 100150 mm Choose products that are made of noncorrosive
(46 mils). metals like stainless steel and aluminum.
Maintain a dry environment using appropriate
Cathodic Protection moisture barriers.
Cathodic protection (CP) is a technique used to Ensure the electrical connections are clean.
control the corrosion of a metal surface by making On a car or truck, apply a thin coating of
it the cathode of an electrochemical cell. CP is petroleum jelly after cleaning the terminal.
carried out by connecting the metal to be protected Coat metals with oil, paint, grease, or varnish
with a piece of another more easily corroded sac- because it can prevent corrosion.
ricial metal to act as the anode of the electro- Make use of cleaning agents like soaps, sol-
chemical cell. The sacricial metal then corrodes vents, emulsion compounds, and chemicals to
instead of the protected metal. For structures efciently get rid of oil, grease, dirt, and other
where passive galvanic CP is not adequate, like unwanted foreign deposits and follow the cor-
in long pipelines, an external DC electrical current rect procedures in applying them.
is applied. Cathodic protection systems are used A mixture of cola and baking soda paste will
to protect a wide range of metallic structures in remove metal corrosion on car batteries.
various environments. Common applications are To prevent soil corrosion, install correctly cop-
steel water or fuel pipelines and storage tanks such per or copper alloy plumbing underground.
as home water heaters, steel pier piles, ship and The main causes of copper corrosion are poor
boat hulls, offshore oil platforms and onshore oil drainage and moisture. A loose layer of backll
well casings, and metal reinforcement bars in such as limestone or pea level must be put
concrete buildings and structures. A common down in the trench before laying copper pipes.
application is in galvanized steel, in which a sac- Galvanizing also provides metal corrosion pro-
ricial coating of zinc on steel protects them tection. This is the process of giving a thin
from rust. coating of zinc or steel material by immersing
Co-sintering Method for Ceramic Membrane Preparation 473
the object in a bath primarily of molten

zinc. Galvanizing is an efcient way to protect Co-sintering Method for Ceramic
steel because even if the surface is scratched, Membrane Preparation
the zinc still protects the underlying layer. This
process is widely used in the auto industry. Zhaoliang Cui
State Key Laboratory of Materials-Oriented
Chemical Engineering, College of Chemistry and C
Summary Chemical Engineering, Nanjing Tech University,
Nanjing, China
Corrosion is effectively controlled by cathodic
protection or by appropriate selection of inhibi-
tors, provided the chemical and electrical condi- Co-sintering method for ceramic membrane prep-
tions are scientically monitored. Understanding aration. Conventionally, a multilayer ceramic
of the mechanism is the key in handling the con- membrane can be obtained by repeating the coat-
dition of the metals and structures. Every scenario ing procedure. But the complicated, time- and
is site specic and needs to be addressed on its energy-consuming production process, which
associated factors of pH of the medium, environ- generally comprises several sequential steps,
ment, and temperature. Accordingly adopt the induces high price. Obviously, reducing or com-
right protection techniques out of the available: bining steps is expected to cut production time
and costs and thereby ceramic membrane price.
Removal of oxidizing agent As a method for fabrication of multilayer struc-
Prevention or inhibition of surface reaction tures, the co-sintering process has attracted
Application of protective coatings organic/ increasing attention due to simple preparation
metallic/nonmetallic process and low production cost, including
Modication of the metal or the surface ceramic membranes (Lindqvist et al.) (Lindqvist
conditions and Liden 1997).
As shown in Fig. 1, in co-sintering process, a
drying and a sintering step are eliminated,
References
resulting in a faster and cheaper preparation.
Corrosion Technology Laboratory Galvanic corrosion. Another benet of this process is the excellent
http://corrosion.ksc.nasa.gov/galcorr.htm. http://corro linkage between the support and the intermediate
sion.ksc.nasa.gov/pitcor.htm. http://corrosion.ksc. layer, as well as the uniform and controlled thick-
nasa.gov/stresscor.htm. http://corrosion.ksc.nasa.gov/ ness of the intermediate layer (Lindqvist et al.)
crevcor.htm
TM 584 Revision C. http://www.corrosioncontrol.com. (Lindqvist and Liden 1997). However, the ther-
Accessed Nov 1994 mal mismatch of each layer is a common and key
VIII- Met J Corros Prot problem during the co-sintering process. In a
Co-sintering Method Top-layer

for Ceramic Membrane membrane tension
Preparation,
Fig. 1 Co-sintering Green top-layer
membrane Co-sintering compression
process (Feng et al. 2007a)
Green sub-layer Sub-layer
membrane tension
membrane
Support compression
Support
474 Cost-Effective Gas Separations
multilayer ceramic membrane, each membrane separation membranes can be used for a wide
layer, which is made from ceramic powders with range of applications, they have generally found
different compositions and particle sizes, exhibits their applications only in a few niche markets.
different sintering behavior. Therefore, it is impor- These applications are listed in Table 1. In these
tant to ensure the sufcient sintering of each mem- markets membranes may not be the most cost-
brane layer and also ensure the high bonding effective gas separation process but offer other
strength of each membrane layer in an asymmetric advantages which make a compelling case for
ceramic membrane (de Jong et al. 2004). For their use. These non-tangible benets of the mem-
a-alumina membranes, too large shrinkage brane gas separation processes are listed in
mismatch between two membrane layers, which Table 2.
occurs when co-sintering at 1,450 C, causes Two necessary but not sufcient properties of
many defects such as cracks and big pores. Mod- the gas separation membranes are its selectivity
erate shrinkage mismatch existed when for a given pair of gases and the gas permeation
co-sintering at 1,300 C, which can bring com- rate of the faster gas. The membrane selectivity is
pressive stress in the sub-layer membrane, is ben- associated with the energy usage and/or product
ecial to the whole co-sintering process (Feng loss (energy cost), and the gas permeation rate is a
et al. 2007a, b). measure of the membrane area needed (capital
cost) for a given separation. Thus, these two fac-
tors have a major impact on the economic viability
of a given membrane for a separation of interest.
References
In addition, these two properties dictate the geom-
de Jong J, Benes NE, Koops GH, Wessling M (2004) etry of the membrane (tubular, hollow ber of at
Towards single step production of multi-layer inorganic sheet), design of the membrane module, congu-
hollow bers. J Membr Sci 239:265269 ration of the membrane cascade (single stage or
Feng J, Fan Y, Qi H, Xu N (2007a) Co-sintering synthesis
multiple stages), and process ow schemes
of tubular bilayer a-alumina membrane. J Membr Sci
288:2027 (parallel, series, recycle, sweep, etc.) (Agrawal
Feng J, Qiu M, Fan Y, Xu N (2007b) The effect of mem-
brane thickness on the co-sintering process of bi-layer
ZrO2/Al2O3 membrane. J Membr Sci 305:2026
Cost-Effective Gas Separations, Table 1 Commercial
Lindqvist K, Liden E (1997) Preparation of alumina mem-
applications of gas separation membranes
branes by tape casting and dip coating. J Eur Ceram Soc
17:359366 Nitrogen production from air
Air-drying
H2/CO adjustment in synthesis gas
H2 recovery in ammonia synthesis
Cost-Effective Gas Separations H2 recovery in petrochemical processes
Acid gas removal from natural gas
Recovery and recycle of olens in olen polymerization
Pushpinder S. Puri
process
PuriMem, LLC, Emmaus, PA, USA
The separation of gases is an essential unit oper- Cost-Effective Gas Separations, Table 2 Benefits of
ation for a large number of chemical processes. membrane gas separation process
The membrane-based gas separation processes are Lower foot print and weight
one of the several options available to the process No moving parts (except for compressors)
engineers. Therefore, a membrane gas separation Remote operation
process has to compete with the more established Easy turnup/turndown
gas separation processes such as cryogenic distil- Portability
lation, absorption, and adsorption. Although gas No open ames
Critical Flux 475
and Xu 1996). They also determine the opera- References

tional mode of the membrane such as pressurized
feed gas with or without intermediate compres- Agrawal R, Xu J (1996) Gas-separation membrane cas-
cades utilizing limited numbers of compressors. AIChE
sion or membrane module operation with vacuum
J 42:2141
on the permeate side while keeping the feed gas a Baker RW, Lokhandwala K (2008) Natural gas processing
little over the atmospheric pressure. All these fac- with membranes: an overview. Ind Eng Chem Res
tors contribute to the bottom line economics of the 47(7):21092121 C
Choe JC et al (1987) Process for separating components of
gas separation process.
a gas stream, US Patent 4,701,187
The life of the membrane and the replacement
cost of the membrane module are a major contrib-
Further Reading
utor to the cost effectiveness of the membrane gas Drioli E, Barbieri G (eds) (2011). Membrane engineering
separation. For commercial membrane gas sepa- for the treatment of gases, volume 1: gas-separation
rations, the membrane life varies from application problems with membranes. RSC Publishing, Cam-
bridge, UK
to application. A major contributor to the life of
Paul DR, Yampolskii YP (1994) Polymer gas separation
the membrane is its chemical compatibility with membranes. CRC Press, Boca Raton, Fl, USA
the feed gases. A membrane material may be
chemically inert to the major gases in the feed
gas, but the trace impurities present in it may
damage the membrane over a period of time by
causing compaction or other defects. In those Critical Flux
cases, the feed gas may be pretreated to remove
trace impurities before the gases are fed to the Xianhui Li and Jianxin Li
membrane separator. The type and extent of State Key Laboratory of Separation Membranes
cleaning of the feed gases has a major bearing and Membrane Processes, School of Materials
on the overall cost of gas separation by Science and Engineering, Tianjin Polytechnic
membranes. University, Tianjin, China
Hybrid processes consisting of a membrane
separation process with a conventional gas sepa-
ration process sometimes offer a more cost- The reduction in the permeate ux with time is a
effective solution for gas separation. Two exam- major problem in many pressure-driven mem-
ples of these processes which are commercially brane processes due to membrane fouling. This
used are in production of high purity nitrogen problem should be solved before the membrane
and in the treatment of natural gas. Air separation processes are economically or technically viable.
membranes have selectivity limitations; To control membrane fouling and maintain sus-
therefore, they are not best suited to produce tainable operation, the concept of critical ux was
high purity nitrogen. However, when a membrane rst introduced in 1995 (Field et al. 1995). The
air separation process is combined with an critical ux is dened as the ux below which ux
adsorption process, membrane separation pro- decline with time does not occur, while fouling is
duces 9599 % nitrogen from which nal traces of observed above it. The supercritical ux fouling is
oxygen is removed by adsorption process ascribed to the fact that the convection of foulants
(Choe et al. 1987). The second application of toward the membrane by the permeate drag ow
hybrid process is for the removal of acid exceeds the back transport velocity of material
gases from the natural gas. Here, a hybrid from the membrane induced by shear and
process consisting of acid gas selective Brownian diffusion. Since then, the critical ux
membrane followed by an amine scrubbing pro- has been extensively applied to membrane pro-
cess offers a more economical process for cesses from microltration (MF) to reverse osmo-
natural gas upgrading (Baker and Lokhandwala sis (RO). Nowadays, the critical ux concept is
2008). well accepted by both scientists and engineers.
476 Critical Flux
There are two forms of the critical ux: strong In the alternative weak form, the uxTMP rela-
and weak form (Fig. 1a). In the strong form, the tionship is below that of the pure water line, but
ux obtained during subcritical ux is equated to increases linearly with pressure. The weak form of
the clean water ux obtained under the same critical ux is the point at which this line becomes
conditions. The strong form of critical ux is the nonlinear. The difference between the two forms
point, at which the transmembrane pressure is that the strong and weak forms of critical ux
(TMP) starts to derive from the pure water line. are mainly caused by reversible and irreversible
fouling, respectively. Therefore, the weak form of
the critical ux is typically observed with the real
a efuents which contain many solutes of various
molar masses.
However, the main drawback of this concept is
Pure water that the determination of critical ux values
cannot be theoretically predicted, but only exper-
Flux Strong form imentally measured by time-consuming experi-
ments. The reason is that critical ux value is
dependent on various factors, such as the charac-
teristics of the membrane (pore diameter, porosity,
Weak form
and material), suspension system (nature, particle
size distribution in relation to pore size distribu-
TMP
b tion, and concentration), and the hydrodynamic
6.0
TMP
35
and ltration conditions. Thus, a number of mea-
5.0 Flux 30 surement methods for critical ux has been pro-
Flux (L/m h)
25 posed (Table 1). The differences among these

TMP (kPa)
4.0 various methods are the quality (accuracy and

20
3.0 reliability) and cost (length of experiment and
15 complexity). The critical ux has mainly been
2.0 10 obtained from uxTMP measurements often by
1.0 5 ux or pressure stepping (Fig. 1b). The mass
0 30 60 90 120 150 balance method could be a useful complement
Time (min) for these methods with a determination of depos-
ited mass. However, for complex suspensions, the
Critical Flux, Fig. 1 The strong and weak forms of crit-
ical ux (a) (Field et al. 1995) and critical ux measure- analysis of the criticality via fouling rate has
ment by the ux-step method (b) (Li et al. 2013) become increasingly common. Also the critical
Critical Flux, Table 1 Measurement methods for critical flux (Bacchin et al. 2006)
Methods Advantages Disadvantages Suitability
Fluxpressure prole: Simplicity Subjective and no link with Feeds with low
deviation from linearity reversibility osmotic pressure
Flux or pressure vs. time: With up and down steps, fouling Points of transition to Feeds with low
ux stepping hysteresis found irreversibility can be missed osmotic pressure
Flux or pressure vs. time: Rigorous when allowance made Time consuming and All feeds
ux cycling for osmotic pressure complex
Direct observation through Direct observation of ux giving Soluble deposition not Particulate feeds
membrane (DOTM) deposition visible
Mass balance Linked to complementary Soluble deposition not Particulate feeds
parameter, the deposited mass visible
Determination by fouling Identify a point of sustainable Subjective and no link with All feeds
rate analysis ux reversibility
Cross-Flow Filtration 477
ux can be deduced from the observation of par-

ticle deposition (or their absence) on the mem- Cross-Flow Filtration
brane surface by direct observation through the
membrane (DOTM). However, the industry sys- Xianhui Li and Jianxin Li
tem often operates at an acceptable fouling rate to State Key Laboratory of Separation Membranes
achieve a higher permeability. The industrially and Membrane Processes, School of Materials
acceptable ux is determined from economic con- Science and Engineering, Tianjin Polytechnic C
siderations, and then its value would change as the University, Tianjin, China
balance between capital and running costs. Over-
all, the critical ux provides an important guide
for the appropriate operating ux. The conventional pressure-driven membrane pro-
cesses in liquid separation can be divided into two
modes: dead-end ltration and cross-ow ltra-
tion. If there is no retentate stream, then the oper-
References ation is termed dead-end ltration. If the
retentate continuously ows from the module out-
Bacchin P, Aimar P, Field RW (2006) Critical and sustain- let, then the operation is termed cross-ow ltra-
able uxes: theory, experiments and applications.
tion. In the cross-ow ltration, the uid feed
Field RW, Wu D, Howell JA, Gupta BB (1995) Critical stream runs tangential to the membrane surface,
ux concept for microltration fouling. J Membr Sci leading to a pressure differential across the mem-
100:259272 brane. This causes some of the particles to pass
Li XH, Li JX, Wang J, Wang H, He BQ, Zhang HW
(2013) Ultrasonic visualization of sub-critical ux foul-
through the membrane. And the remaining parti-
ing in the double-end submerged hollow ber mem- cles continue to ow across the membrane. In
brane module. J Membr Sci 444:394401 contrast to the dead-end ltration technique, the
Cross-Flow Filtration, Pressure

Fig. 1 Schematic diagrams a
of cross-ow ltration (a)
and representation of the
steady ltration ow (b) Concentrate
Feed
Membrane
Permeate
b
Flux
Layer thickness (mm)
Steady Layer
filtration thickness
Flux (L/m2 h)
Time
478 Cross-Flow Membrane Bioreactor
shearing effect caused by cross ow can sweep the particle size, interactions, etc.) will strongly
deposited particles toward the concentrate or impact the physicochemical interactions that
retentate side so that the cake layer remains rela- occur between the uid constituents and the mem-
tively thin (Fig. 1a). In the case of cross-ow brane surface as well as between these constitu-
ltration, a cake layer will be built up gradually. ents. Therefore, the goal of membrane
After some time, a steady ltration state is technologists is to use appropriate membrane
obtained for a long time as shown in Fig. 1a. material, module conguration, system design,
This fact is due to the equilibrium between the and operating conditions to achieve the most eco-
transport of particles to the cake layer and the back nomical process.
transport of particles into the feed stream. Conse-
quently, the cross-ow ltration is commonly
employed for the large-scale industrial membrane References
ltration process.
The permeate ux is governed by the general El Rayess Y, Albasi C, Bacchin P, Taillandier P, Raynal J,
Mietton-Peuchot M, Devatinee A (2011) Cross-ow
ltration equation (Darcys law) given as
microltration applied to oenology: a review.
J Membr Sci 382:119
DP
J (1)
m:Rh
where J (m3/m2/s) is the permeate ux, DP (Pa) is Cross-Flow Membrane Bioreactor

the applied pressure, m (Pa.s) is the solvent vis-
cosity, and Rh (m1) is the hydraulic resistance. Aerobic Membrane Bioreactor
Certainly, many parameters of cross-ow l-
tration play a key role in ltration performance.
As summarized in Table 1, the parameters
inuencing ltration performance cover hydrody-
namics and operating conditions, membrane char- Cross-Flow Membrane Module
acteristics, as well as the uid characteristics. For
example, on the basis of Darcys law, the operat- Joerg Balster
ing conditions (transmembrane pressure, cross Application Technology, Evonik Fibres GmbH,
ow, and temperature) are directly related to the Schoering, Austria
permeate ux. Membrane characteristics (pore
size, surface free energy, hydrophilicity/hydro-
phobicity, etc.) and uid characteristics (pH, Synonyms
Cross-Flow Filtration, Table 1 Summary of the param- Flat sheet membrane module; Plate and frame
eters influencing cross-flow filtration performance (Rayess module
et al. 2011)
Parameters In order to apply membranes in a technical scale,
Operating Transmembrane pressure, permeate large membrane areas normally are required. The
conditions ux, cross-ow velocity, and smallest unit into which a membrane area is
temperature
packed is called a module (Mulder 1997). The
Membrane Pore size, porosity, hydrophilicity/
characteristics hydrophobicity character, and earliest designs were based on simple ltration
surface free energy technology and consisted of at sheets of mem-
Fluid Physicochemical interactions brane held in a type of lter press (Baker 2004).
characteristics (electrostatic, polar, etc.), pH, ionic Modules containing only one membrane, which
strength, suspension concentration, are mostly used in laboratory scale, are cross-ow
and suspension size
membrane modules. If several membranes are
Cross-Link 479
used, these modules are called plate and frame In the membrane fabrication process, cross-
modules. link is also mainly divided into two types includ-
ing chemical cross-link and physical cross-link.
Hydrogen bond cross-link is a special physical
References cross-link. Different from other cross-links, the
functional groups of the polymer will be not
Baker (2004) Membrane technology and applications, depletion after the hydrogen bond cross-link C
(Basu et al. 2010). Compared with the covalent
ogy, 2nd edn. Kluwer, Dordrecht bond, the hydrogen bond energy is smaller, and
the bond length is longer, which is a weak link.
However, a large amount of hydrogen bonds can
stably exist in the membrane at low pressure due
to the strong physical interactions (Basu
Cross-Link et al. 2010). In the membrane fabrication process,
chemical cross-link and physical cross-link often
Zhi Wang make the gap space between the polymer chain
Chemical Engineering Research Center, School segments smaller and limit the vibration of the
of Chemical Engineering and Technology, Tianjin polymer chain segment, which are main ways to
University, Tianjin, China improve the selectivity (rejection) of the permeate
(Luis et al. 2012; Zhao et al. 2006; Basu
et al. 2010; Qiao et al. 2013; Kim et al. 2000).
The reticulate polymer formation processes by the However, most of the cross-linking processes will
interactions of linear or branched polymer chains make the membrane structure densication and
using the chemical modication or physical modi- may be not favorable for the permeance (ux)
cation are dened as cross-link. Cross-link is increase of the permeate in the membrane process
mainly divided into two types including chemical (Luis et al. 2012; Basu et al. 2010; Zhao
cross-link such as polycondensation or polyaddition et al. 2006). Fortunately, in the recent studies,
cross-link (Beryozkina et al. 2009; Luis et al. 2012) the cross-linking agents with functional groups
and physical cross-link such as heat or photo cross- such as amine group used to modify the polymer
link (Chujo 2010; Farrell et al. 2005). Chemical could simultaneously improve the selectivity and
cross-link is dened as the interaction process of permeance of the permeate in the gas separation
linear or branched polymer chains by using the membrane process (Qiao et al. 2013).
covalent bond among the functional groups of the
polymers and cross-linking agents. Most of the
chemical cross-linking agents are small molecular References
compounds with the molecular weight of 200400,
which have two or more functional groups such as Basu S, Khan AL, Cano-Odena A, Liu C, Vankelecom IFJ
(2010) Membrane-based technologies for biogas sepa-
amine group (Bernardo et al. 2009). Physical cross-
ration. Chem Soc Rev 39:750768
link is the interaction process of linear or branched Bernardo P, Drioli E, Golemme G (2009) Membrane gas
polymer chains by using the physical interaction separation: a review/state of the art: a review. Ind Eng
such as van der Waals forces and hydrogen bond Chem Res 48:46384663
Beryozkina T, Boyko K, Khanduyeva N, Senkovskyy V,
among the functional groups of the polymers and Horecha M, Oertel U, Simon F, Stamm M, Kiriy
cross-linking agents. After cross-linking, polymer A (2009) Grafting of polyuorene by surface-initiated
properties such as the mechanical strength, elastic- Suzuki polycondensation. Angew Chem Int Ed
ity, solvent resistance, and permeate selectivity 48:26952698
Chujo Y (2010) Conjugated polymer synthesis: methods
could be improved (Bernardo et al. 2009;
and reactions. Wiley-WCH, Weinheim, pp 38,112113
Beryozkina et al. 2009; Chujo 2010; Farrell Farrell IS, Toroney R, Hazen JL, Mehl RA, Chin JW
et al. 2005; Luis et al. 2012). (2005) Photo-cross-linking interacting proteins with a
480 Cross-Linkable Copolymer Membranes
genetically encoded benzophenone. Nat Methods light-emitting diodes (Segalman et al. 2009).
2:377384 Moreover, cross-linkable copolymers are the
Kim CK, Kim JH, Roh IJ, Kim JJ (2000) The changes of
membrane performance with polyamide molecular favorable polymeric membrane materials owing
structure in the reverse osmosis process. J Membr Sci to their diversity.
165:189199 Cross-linkable copolymer membranes exhibit
Luis P, Gerven TV, Bruggen BVD (2012) Recent develop- the characteristics of structure control and prop-
ments in membrane-based technologies for CO2 cap-
ture. Prog Energy Combust 38:419448 erty tunability due to the versatility of comono-
Qiao Z, Wang Z, Zhang C, Yuan S, Zhu Y, Wang J, Wang mers and cross-linkable groups of cross-linkable
S (2013) PVAm-PIP/PS composite membrane with copolymers. Cross-linkable copolyimide is a class
high performance for CO2/N2 separation. AIChE of ne membrane materials for CO2 separation.
J 59:215228
Zhao J, Wang Z, Wang J, Wang S (2006) Inuence of heat- Most of the glassy polymers suffer from CO2-
treatment on CO2 separation performance of novel induced plasticization at high CO2 partial pressure
xed carrier composite membranes prepared by inter- (Bos et al. 1999). However, cross-linked
facial polymerization. J Membr Sci 283:346356 copolyimide reveals excellent CO2 plasticization
resistance due to the formation of rigid polymer
networks (Vanherk et al. 2013). End-group cross-
linkable sulfonated copolymers (poly(arylene
Cross-Linkable Copolymer ether ketone, poly(ether ether ketone), etc.) are
Membranes the predominant proton exchange membrane
materials for fuel cell, which often provide
Zhi Wang improved proton conductivity and reduced water
Chemical Engineering Research Center, School uptake via proper cross-linking (Jones and
of Chemical Engineering and Technology, Tianjin Roziere 2009). Cross-linkable copolymers with
University, Tianjin, China different functional comonomers are the primary
sources of stimuli-responsive membrane. Stimuli-
responsive membranes have an extensive applica-
Cross-linkable copolymer membrane is one of the tion of drug controlled release and sensors/actua-
most extensively used polymeric membranes, tors for different stimuli (temperature, light,
such as gas separation membrane, pervaporation thermal, pH, etc.) (Qiu and Park 2012). Further-
membrane, proton exchange membrane for fuel more, another decided advantage of cross-
cell, and stimuli-responsive membrane. linkable copolymer membrane is their exceptional
Cross-linkable copolymer is a kind of copoly- thermal, mechanical, and chemical stability after
mer bearing the cross-linkable groups in the chain cross-linking reaction, which provides a solid fun-
end or the side-chain of one specic unit. Amino dament for practical applications in harsh
group, epoxy group, carboxyl group, silane group, environment.
and vinyl group are the common cross-linkable An increasing number of novel cross-linkable
groups. Furthermore, some specic cross-linkable copolymers will be developed with the progress of
groups could be incorporated in the copolymer polymer chemistry. Therefore, they will have
matrix in terms of actual demands. Cross-linking broad application prospects in membrane science
reaction can be carried out through thermal/photo and technology.
initiation by oneself or chemical reaction with
other cross-linking reagents. Various properties
of copolymer such as thermal stability, mechani- References
cal strength, and chemical resistance could be
improved after cross-linking. As a novel poly- Bos A, Punt IG, Wessling M, Strathman H (1999) CO2-
meric material, cross-linkable copolymers are induced plasticization phenomena in glassy polymers.
widely used for fabrication of stimuli-responsive Jones DJ, Roziere J (2009) Fuel cells-proton-exchange
polymer vesicles (Li and Keller 2009) and organic membrane fuel cells membrane: ambient temperature.
Cross-Linked Aromatic Polyamide Membranes 481
In: Garche J (ed) Encyclopedia of electrochemical condensation reaction of the aromatic amino
power sources. Elsevier, Amsterdam, pp 667679 groups and the aromatic carboxylic acid or acid
Li MH, Keller P (2009) Stimuli-responsive polymer vesi-
cles. Soft Matter 5:927937 chloride groups (Palmer 2002). Aromatic polyam-
Qiu Y, Park K (2012) Environment-sensitive hydrogels for ide bers, which have high resistance to abrasion,
drug delivery. Adv Drug Deliv Rev 64:4960 low ammability, and excellent mechanical prop-
Segalman RA, McCulloch B, Kirmayer S, Urban JJ erties, are widely used in aerospace and military
(2009) Block copolymers for organic optoelectronics.
Macromolecules 42:92059216 applications. Moreover, aromatic polyamide is C
Vanherk K, Koeckelberghs J, Vankelecom FJ also one of the best membrane materials, which
(2013) Crosslinking polyimide for membrane applica- can form dense cross-linked aromatic polyamide
tion: a review. Prog Polym Sci 38:874896 membranes via interfacial polymerization
processes.
The cross-linked polyamide membranes are,
usually, used as brackish water or seawater desa-
Cross-Linked Aromatic Polyamide lination membranes, such as reverse osmosis
Membranes (RO), forward osmosis (FO), and nanoltration
membranes. Cross-linked aromatic polyamide
Zhi Wang RO membranes, derived from metaphenylene
Chemical Engineering Research Center, School diamine (MPD) and trimesoyl chloride (TMC),
of Chemical Engineering and Technology, Tianjin are the most widely used commercial reverse
University, Tianjin, China osmosis membranes. The structure of the mem-
brane is shown in Fig. 1. This partially cross-
linked chemical structure consists of the cross-
Cross-linked aromatic polyamide membranes are, linked portions (C18H12N3O3) and the linear por-
typically, the aromatic polyamide-based thin lm tions (C15H10N2O4) (Wei et al. 2010). The com-
composite membranes. Aromatic polyamides are mercial cross-linked aromatic polyamide RO
macromolecules with repeating units linked by membranes always exhibit superior water ux
amide bonds, which are produced from the and salt rejection, resistance to pressure
Cross-Linked Aromatic Polyamide Membranes, Fig. 1 The barrier layer structure of the cross-linked polyamide RO
membrane
482 Cross-Linked Layer-by-Layer Membranes
compaction, wider operating temperature range

and pH range, and higher stability to biological Cross-Linked Layer-by-Layer
attack, compared with cellulose acetate RO mem- Membranes
branes (Li and Wang 2010). Currently, there are
kinds of commercial cross-linked polyamide RO Saren Qi1 and Chuyang Y. Tang2
1
membranes, produced by some leading mem- Singapore Membrane Technology Centre,
brane manufacturing companies, such as Dow Nanyang Technological University, Singapore,
(Filmtec), Hydranautics, Toray, GE Osmonics Singapore
2
(Desal), and Koch Membrane Systems (Fluid Department of Civil Engineering
Systems). The University of Hong Kong, Hong Kong,
However, the cross-linked aromatic polyamide Hong Kong SAR
membranes, including RO, FO, and nanoltration
membranes, have some problems to be solved.
The rst is the membrane fouling due to its high A layer-by-layer (LbL) membrane is formed by
pollutant deposition tendency (Subramani and coating a porous substrate with polyanions and
Hoek 2010). Membrane fouling, which is a pro- polycations in an alternative sequence (Decher
cess that the solutes, particles, and bacterium 1997). Because of the electrostatic forces between
deposit on the membrane surfaces, can cause the charged polymers, the polyanions and
ux decline and affect the quality of the water polycations are self-assembled into thin layers.
produced (Kang and Cao 2012). The second prob- Advantages of LbL membranes include its con-
lem is that the chlorine resistance of the mem- trollable structure and composition as well as fac-
branes needs to be improved compared with ile process of fabrication and modication. LbL
cellulose acetate RO membranes (Kang membranes are able to reject a variety of contam-
et al. 2007). This is because the amide bonds can inants including divalent ions. Their rejection of
be easily halogenated by the free chlorine in the monovalent ions tends to be relatively low, but
feed solution, resulting in the degradation of the can be potentially improved by posttreatment.
membranes. Cross-linking is commonly used to treat LbL
membranes for improved rejection and enhanced
membrane stability (Qiu et al. 2011). The selective
References multilayers (the rejection layers) of LbL membrane
are mainly formed by the electrostatic interaction that
Kang GD, Cao YM (2012) Development of antifouling is weaker than covalent bonds. Therefore, the stability
reverse osmosis membranes for water treatment: a
of LbL membranes is a key concern in their applica-
review. Water Res 46(3):584600
Kang GD, Gao CJ, Chen WD, Jie XM, Cao YM, Yuan tions, particularly when the feedwater has high ionic
Q (2007) Study on hypochlorite degradation of aro- strength that weakens the ionic bonds. The cross-
matic polyamide reverse osmosis membrane. linking treatment offers an effective solution to this
J Membr Sci 300(12):165171
problem. The commonly used cross-linking method
Li D, Wang H (2010) Recent developments in reverse
osmosis desalination membranes. J Mater Chem includes chemical cross-linking (with the help of
20(22):45514566 cross-linking agents such as terephthalaldehyde and
Palmer RJ (2002) Polyamides, plastics. Encyclopedia of glutaraldehyde) (Qi et al. 2012), UV-assisted cross-
polymer science and technology. John Wiley & Sons,
linking (Duong et al. 2013), and heat-induced cross-
Inc.
Subramani A, Hoek E (2010) Biolm formation, cleaning, linking (Dai et al. 2001). The resultant cross-linked
re-formation on polyamide composite membranes. LbL membranes exhibit improved rejection, good
Desalination 257(1):7379 chemical resistance, mechanical strength, and long-
Wei XY, Wang Z, Chen J, Wang JX, Wang SC
(2010) A novel method of surface modication on
term stability.
thin-lm-composite reverse osmosis membrane by LbL membranes can be classied according to
grafting hydantoin derivative. J Membr Sci their applications. The common application elds
346(1):152162 include water treatment (Bruening et al. 2008),
Cross-Linked Layer-by-Layer Membranes 483
Cross-Linked Layer-by-
Layer Membranes,
LbL membrane
of applications for LbL
membrane (Note: NF
nanoltration, FO forward
osmosis, UF ultraltration,
MF microltration)
pervaporation
C
Water treatment Fuel cell
NF FO Dialysis UF or MF
pervaporation (Zhao et al. 2011), and fuel cell References

(Zhang and Shen 2012) (Fig. 1). In water treatment
process, LbL membranes present high water per- Bruening ML, Dotzauer DM, Jain P, Ouyang L, Baker GL
(2008) Creation of functional membranes using poly-
meability and relatively good salt rejection. It is
electrolyte multilayers and polymer brushes. Langmuir
because ultrathin multilayers can achieve high 24(15):76637673
water permeability and dense multilayers can main- Dai JH, Jensen AW, Mohanty DK, Erndt J, Bruening ML
tain membrane selectivity. On the other hand, LbL (2001) Controlling the permeability of
multilayered polyelectrolyte lms through derivatiza-
membranes usually bring charges that may enhance
tion, cross-linking, and hydrolysis. Langmuir
the rejection due to Donnan exclusion. Addition- 17(3):931937
ally, due to the availability of functional groups in Decher G (1997) Fuzzy nanoassemblies: toward layered
polyelectrolyte molecules, LbL method is easily polymeric multicomposites. Science
277(5330):12321237
incorporated with other technologies to produce
Duong PHH, Zuo J, Chung T-S (2013) Highly
new functional membrane. For example, Ag crosslinked layer-by-layer polyelectrolyte FO mem-
nanoparticles can be incorporated into LbL FO branes: understanding effects of salt concentration and
membrane to produce a novel FO membrane for deposition time on FO performance. J Membr Sci
427:411421
antibacterials (Liu et al. 2013). In pervaporation,
Liu X, Qi S, Li Y, Yang L, Cao B, Tang CY (2013) Syn-
LbL membranes used can produce high permeate thesis and characterization of novel antibacterial silver
ux due to its ultrathin layers, and the selective nanocomposite nanoltration and forward osmosis
layer is resistant to organic solvent. In fuel cell, membranes based on layer-by-layer assembly. Water
Res 47(9):30813092
compared with traditional process of fabrication
Qi S, Qiu CQ, Zhao Y, Tang CY (2012) Double-skinned
(i.e., spraying, paining, ink-jet printing, and so forward osmosis membranes based on layer-by-layer
on), more efcient process could be achieved by assembly-FO performance and fouling behavior.
using LbL membrane for proton transfer. Addition- J Membr Sci 405406:2029
Qiu C, Qi S, Tang CY (2011) Synthesis of high ux
ally, while incorporated with other materials, LbL
forward osmosis membranes by chemically crosslinked
membranes can increase the catalyst utilization rate layer-by-layer polyelectrolytes. J Membr Sci
so as to enhance the fuel cell performance. 381(12):7480
Zhang H, Shen PK (2012) Recent development of polymer
electrolyte membranes for fuel cells. Chem Rev
112(5):27802832
Zhao Q, An QF, Ji Y, Qian J, Gao C (2011) Polyelectrolyte
Cross-References complex membranes for pervaporation,
nanoltration and fuel cell applications. J Membr Sci
Layer-by-Layer (LBL) Method 379(12):1945
484 Cross-Linked Poly(ethylene oxide) Membranes
PEO or by reactions of end groups such as

Cross-Linked Poly(ethylene oxide) hydroxyl or vinyl groups. Hirayama
Membranes et al. prepared cross-linked PEO from mixtures
of poly(ethylene glycol) methacrylate
Haiqing Lin (a monomer containing nine EO units) and poly
Department of Chemical and Biological (ethylene glycol) dimethacrylate (a cross-linker
Engineering, University at Buffalo, The State containing 14 EO units) by plasma irradiation
University of New York, Buffalo, NY, USA (Hirayama et al. 1999). They reported that CO2
permeability increases as monomer content
increases while CO2/N2 selectivity remains
Poly(ethylene oxide) (PEO) and its containing almost unchanged. They prepared a polymer
polymers have been extensively explored as containing 70 wt% monomer; it exhibited a CO2
membrane materials for CO2 removal from the permeability of 260 Barrers and a CO2/N2 selec-
mixtures with light gases such as hydrogen, nitro- tivity of about 48 at 35 C and 1 atm (Hirayama
gen, and methane (Blume and Pinnau 1990; Li et al. 1999).
et al. 1995; Okamoto et al. 1995; Bondar Several series of cross-linked PEO have been
et al. 1999, 2000; Hirayama et al. 1999; Yoshino prepared by UV photopolymerization of poly(eth-
et al. 2000; Kim et al. 2001; Patel et al. 2003, ylene glycol) diacrylate (PEGDA) containing
2004; Lin and Freeman 2004, 2005; Metz 14 EO units and poly(ethylene glycol) methyl
et al. 2004; Lin et al. 2006a, b, c; Car ether acrylate (PEGMEA) containing about 8.5
et al. 2008; Reijerkerk et al. 2010; Yave EO units (Lin and Freeman 2005; Lin
et al. 2010; Lau et al. 2011; Liu et al. 2013). The et al. 2006a, c). The resulting chemical structure
polar ether oxygen has strong afnity with CO2 is shown in Fig. 1 (Kalakkunnath et al. 2005,
with quadrupole moment, but not with light gases, 2006). PEO crystallinity was completely absent
thus providing high CO2 solubility and high sol- in these copolymers at temperatures as low as
ubility selectivity of CO2/light gas (Okamoto 90 C (the lowest limit of the calorimeter used)
et al. 1995; Bondar et al. 1999; Lin and Freeman due, presumably, to the short nature of the EO
2004). On the other hand, unlike any other branches in the side chains of these materials
polar groups which would inevitably increase and to the frustration of crystallization by cross-
the chain rigidity of the polymer and decrease linking. The average number of EO units per
CO2 diffusivity and permeability, ether oxygens acrylate group is approximately seven, which is
appear to be the only well-known groups which the minimum number of monomers required for
improve CO2 diffusivity and permeability (Lin the unit cell of a PEO crystal (Brandrup
and Freeman 2005; Lin et al. 2006c). Ethylene et al. 1999). Therefore, using these starting mate-
oxide units have a high concentration of ether rials, network polymers can be prepared which are
oxygens, which makes the ideal as building noncrystalline and have a high concentration of
blocks for membrane materials for CO2/light gas EO units. The use of acrylate groups instead of
separation. However, PEO has high crystallinity methacrylate groups provides higher chain exi-
which is deleterious for gas permeability (Lin and bility (or lower glass transition temperature),
Freeman 2005). which can increase gas diffusion coefcients
Cross-linking is an effective way to inhibit and, in turn, permeability (van Amerongen 1964;
polymer chain crystallization (Hirayama Lin et al. 2005).
et al. 1999). Graham proposed empirically that The inclusion of methyl ether chain end groups
signicant crystallinity in cross-linked PEO is in the monomer (as opposed to hydroxyl end
not evident when the molecular weight between groups) increases free volume and improves
cross-links is lower than 1,500 (Graham 1987). CO2/H2 separation performance of copolymers
Cross-linking can be achieved by radiation or of PEGMEA and PEGDA. As PEGMEA mono-
radical cross-linking of high molecular weight mer content increases from zero to 99 wt%, CO2
Cross-Linked Poly(ethylene oxide) Membranes 485
Cross-Linked Poly
(ethylene oxide)
Membranes,
Fig. 1 Schematic
representation of PEGDA/
PEGMEA copolymer
network (Kalakkunnath
et al. 2005, 2006). Italicized
and bolded parts of the
C
network derive from the
cross-linker. R1 is CO
(OCH2CH2)8OCH3 from
PEGMEA; R2 is COO
(CH2CH2O)14OC from
PEGDA
Cross-Linked Poly 102

(ethylene oxide)
Membranes, -20C
Fig. 2 Comparison of
cross-linked poly(ethylene
oxide) membrane with other
polymeric membranes in 10C
101
the literature used for CO2/ 35C
H2 separation (Lin
2/H2
et al. 2006a) Upper Bound

a CO
100
10-1
10-2 10-1 100 101 102 103 104
CO2 Permeability [Barrer]
permeability increases about fourfold, reaching performance, compared with other

570 Barrers, and CO2/H2 selectivity increases by PEG-containing polymers such as block copoly-
50 %, up to 12, at 35 C and innite dilution (Lin mers (Lin and Freeman 2006).
et al. 2006). These materials show good CO2/H2
separation properties, as shown in Fig. 2. If the 1. Cross-linking ensures good chemical resis-
OCH3 end groups are replaced by OH end tance, simply because cross-linked networks
groups, CO2 permeability and CO2/H2 selectivity are not soluble. It is straightforward to incor-
remain unchanged as the OH end group content porate more than 80 wt% PEO in cross-linked
changes across the entire composition window polymers from acrylate monomers and/or
(Lin et al. 2006c). cross-linkers, which is higher than the maxi-
In general, the cross-linked network materials mum value (about 60 %) reported for block
formed from PEG acrylates and diacrylates pro- copolymers and blends (Okamoto et al. 1995;
vide advantages for CO2/light gas separation Yoshino et al. 2000).
486 Cross-Linked Poly(ethylene oxide) Membranes
2. Cross-linked PEO can be further modied to Lau CH, Liu S et al (2011) Silica nanohybrid membranes
give much higher CO2 permeability and CO2/ with high CO2 afnity for green hydrogen purication.
Adv Energy Mater 1(4):634642
H2 selectivity than block copolymers or blends Li J, Nagai K et al (1995) Preparation of polyethyle-
by introducing methyl ether chain end groups neglycol (PEG) and cellulose acetate (CA) blend mem-
(Hirayama et al. 1999; Lin et al. 2006a, c). branes and their gas permeabilities. J Appl Polym Sci
3. Finally, cross-linking could completely sup- 58:14551463
Lin H, Freeman BD (2004) Gas solubility, diffusivity and
press crystallization at all temperatures of prac- permeability in poly(ethylene oxide). J Membr Sci
tical interest (Lin et al. 2006c), while block 239:105117
copolymers or blends would still exhibit a Lin H, Freeman BD (2005) Materials selection guidelines
melting temperature at temperatures near for membranes that remove CO2 from gas mixtures.
J Mol Struct 739(13):5774
ambient (Bondar et al. 1999). Consequently, Lin H, Freeman BD (2006) Gas permeation and diffusion
with the cross-linked PEO materials, tempera- in crosslinked poly(ethylene glycol Diacrylate). Mac-
ture can be lowered to optimize CO2 separation romolecules 39(10):35683580
performance because lower temperatures favor Lin H, Kai T et al (2005) The effect of cross-linking on gas
permeability in crosslinked poly(ethylene glycol
CO2/light gas solubility selectivity (Lin diacrylate). Macromolecules 38:83818393
et al. 2006a, c). Lin H, Van Wagner E et al (2006a) Plasticization-enhanced
H2 purication using polymeric membranes. Science
311(5761):639642
Lin H, Van Wagner E et al (2006b) High performance
References polymer membranes for natural gas sweetening. Adv
Mater 18(18):3944
Blume I, Pinnau I (1990) Composite membrane, method of Lin H, Van Wagner E et al (2006c) Transport and structural
preparation and use. US Patent 4,963,165 characteristics of crosslinked poly(ethylene oxide) rub-
Bondar VI, Freeman BD et al (1999) Gas sorption and bers. J Membr Sci 276(12):145161
characterization of poly(ether-b-amide) segmented Liu SL, Shao L et al (2013) Recent progress in the design
block copolymers. J Polym Sci Part B Polym Phys of advanced PEO-containing membranes for CO2
37:24632475 removal. Prog Polym Sci 38(7):10891120
Bondar VI, Freeman BD et al (2000) Gas transport prop- Metz SJ, Mulder MHV et al (2004) Gas-permeation
erties of poly(ether-b-amide) segmented block copoly- properties of poly(ethylene oxide) poly(butylene tere-
mers. J Polym Sci Part B Polym Phys 38:20512062 phthalate) block copolymers. Macromolecules
Brandrup J, Immergut EH et al (1999) Polymer handbook. 37:45904597
Wiley, New York Okamoto K-I, Fuji M et al (1995) Gas permeation proper-
Car A, Stropnik C et al (2008) PEG modied poly(amide- ties of poly(ether imide) segmented copolymers. Mac-
b-ethylene oxide) membranes for CO2 separation. romolecules 28:69506956
J Membr Sci 307:8895 Patel NP, Miller AC et al (2003) Highly CO2-permeable
Graham NB (1987) Poly(ethylene oxide) and related and selective polymer nanocomposite membranes. Adv
hydrogels. In: Peppas NA (ed) Hydrogels in medicine Mater 15(9):729733
and pharmacy. CRC Press, Boca Raton, pp 95113, II: Patel NP, Miller AC et al (2004) Highly CO2-permeable
Polymers and -selective membranes derived from crosslinked
Hirayama Y, Kase Y et al (1999) Permeation properties to poly(ethylene glycol) and its nanocomposites. Adv
CO2 and N2 of poly(ethylene oxide)-containing and Funct Mater 14(7):699707
crosslinked polymer lms. J Membr Sci 160:8799 Reijerkerk SR, Knoef MH et al (2010) Poly(ethylene gly-
Kalakkunnath S, Kalika DS et al (2005) Segmental relax- col) and poly(dimethyl siloxane): combining their
ation characteristics of crosslinked poly(ethylene advantages into efcient CO2 gas separation mem-
oxide) copolymer networks. Macromolecules branes. J Membr Sci 352:126135
38(23):96799687 van Amerongen GJ (1964) Diffusion in elastomers. Rubber
Kalakkunnath S, Kalika DS et al (2006) Viscoelastic char- Chem Technol 37:10651152
acteristics of U.V. polymerized poly(ethylene glycol) Yave W, Car A et al (2010) Nanostructured membrane
diacrylate networks with varying extents of material designed for carbon dioxide separation.
crosslinking. J Polym Sci Part B Polym Phys J Membr Sci 350:124129
44(15):20582070 Yoshino M, Ito K et al (2000) Effects of hard-segment
Kim J, Ha S et al (2001) Gas permeation of poly(amide-6- polymers on CO2/N2 gas separation properties of poly
b-ethylene oxide) copolymer. J Membr Sci (ethylene oxide)-segmented copolymers. J Polym Sci
190:179193 Part B Polym Phys 38:17071715
Cross-Linked Polyamide Membranes 487
membranes for different applications, for

Cross-Linked Polyamide Membranes example, piperazine (PIP) for NF;
m-phenylenediamine (MPD) for RO, FO, and
Zhi Wang PV; and 3,3-diamino-N-methyldipropylamine
Chemical Engineering Research Center, School (DNMDAm) for GS (Petersen 1993; Chapman
of Chemical Engineering and Technology, Tianjin et al. 2008; Lau et al. 2012; Zhao et al. 2012; Li
University, Tianjin, China et al. 2012). Trimesoyl chloride (TMC) is the most C
widely used acyl chloride in IP. The three reactive
acyl chloride groups of TMC could form the
The cross-linked polyamide membrane is one of cross-linked structure that potentially enhances
the most commonly used polymeric membranes both separation performance and mechanical
for nanoltration (NF), reverse osmosis (RO), strength of the membrane. The reaction between
forward osmosis (FO), pervaporation (PV), and these two reactive monomers, i.e., diamine and
gas separation (GS). TMC, is shown in Fig. 1. The formed polyamide
The polyamide could be produced either natu- contains both cross-linked structure and linear
rally or articially. The natural polyamides structure. The linear structure is caused by the
include proteins and peptides. The articial poly- hydrolysis of acyl chloride group. The ratio of
amides, such as nylons and aramids, could be cross-linked part to linear part of the polyamide,
produced by polymerization. Polyamides are as well as the surface characteristic and the thick-
widely used in textiles, automotives, and carpet ness of polyamide layer, could be nely adjusted
due to their extreme durability and strength. by changing technological parameters of IP pro-
Moreover, polyamides possess excellent separa- cess, such as solvent type, monomer concentra-
tion property and are promising materials for sep- tions, reaction time, and posttreatment conditions.
aration membranes. The performance of the membrane therefore could
The cross-linked polyamide membrane is usu- be improved by optimizing the structures.
ally thin-lm composite (TFC) membrane The cross-linked polyamide membranes for
consisting of a dense top layer of cross-linked NF and RO in applications of desalination have
polyamide with a thickness of 0.10.3 mm been successfully commercialized for decades.
supported by a sublayer with a thickness of There are several membrane suppliers providing
50150 mm. The membrane is generally fabri- NF and RO membranes for large-scale desalina-
cated by interfacial polymerization (IP) of di- or tion plants, such as DOW, Toray, Hydranautics,
multi-amines in aqueous phase and di- or multi- and Toyobo. The salt rejection and water perme-
acyl chloride in organic phase on the sublayer. ability of NF and RO membranes have been dras-
Various diamines have been used to fabricate tically improved during these years but the
O O O O O O
HN R1 NH + Cl C C Cl HCl + N R1 N C C N R1 N C C
m 1m
R2 R3 R2 R3 R2 R3
C O C O COOH
Cl N R2
R1
N R3
Cross-Linked Polyamide Membranes, Fig. 1 Reaction scheme for preparation of cross-linked polyamide membranes
by IP
488 Crown Ethers
antifouling and chlorine resistant properties of the stable complexes with alkali/alkaline earth metal
membranes still have to be enhanced ions and displayed a unique selectivity based on
(Xu et al. 2013). The cross-linked polyamide the size compatibility of the ligand cavity size and
membranes for FO in applications of desalination the ionic size of these metal ions. The crown
are still in laboratory research due to the chal- ethers mimic the biological receptors such as
lenges of concentration polarization, fouling, and valinomycin (see Fig. 1b) which selectively trans-
reverse solute diffusion (Zhao et al. 2012). The ports K+ ion as compared to Na+ ion (by a factor of
cross-linked polyamide membranes for PV in 105) across cell membranes. The transport prop-
applications of dehydration and for GS in appli- erties of the crown ethers depend on the size of the
cations of CO2 separation are fabricated in labo- crown ring, number of donor atoms, nature of
ratory scale at present, because the permeance and donor atoms, nature of lipophilic/ionizable side
selectivity of membranes have to be improved to arms, etc. Crown ethers have been overwhelm-
fulll technical and economical requirements of ingly used for the transport of alkali metal/alkaline
practical applications. The improvement of sepa- earth ions, though there have been quite a large
ration performance of membranes for GS could be number of reports on the transport of Ag+, Hg2+
achieved by combining multiple separation mech- (with thia-crown ethers), and Pb2+ which are pro-
anisms into a single membrane through selecting posed for heavy metal removal from wastewaters
proper monomers in IP process (Li et al. 2012). (Lamb et al. 1980). There have also been several
reports on the use of crown ethers for the selective
transport of lanthanides and actinides though size
References selective factors are far less pronounced in such
cases. Another major application of crown ethers
Chapman PD, Oliveira T, Livingston AG, Li K (2008) for metal ion transport includes the recovery of
Membranes for the dehydration of solvents by
radio-cesium and radio-strontium from radioac-
pervaporation. J Membr Sci 318:537
Lau WJ, Ismail AF, Misdan N, Kassim MA tive wastes which can signicantly reduce the
(2012) A recent progress in thin lm composite mem- load on the vitried glass blocks (Dozol
brane: a review. Desalination 287:190199 et al. 1995; Dozol and Casas 1994). Crown
Li S, Wang Z, Yu X, Wang J, Wang S (2012) High-
ether-based supported liquid membranes have
Performance membranes with multi-permselectivity
for CO2 separation. Adv Mater 24:31963200 shown large amounts of acid cotransport when
Petersen RJ (1993) Composite reverse osmosis and attempts have been made for the transport of
nanoltration membranes. J Membr Sci 83:81150 Cs-137 and Sr-90 from nitric acid feeds. This
Xu J, Wang Z, Yu L, Wang J, Wang S (2013) A novel reverse
has affected the overall efciency of the transport
osmosis membrane with regenerable anti-biofouling and
chlorine resistant properties. J Membr Sci 435:8091 process, and appropriate selection of the diluent
Zhao S, Zou L, Tang CY, Mulcahy D (2012) Recent devel- has helped in overcoming this issue (Raut
opments in forward osmosis: opportunities and chal- et al. 2012). Though crown ethers have been
lenges. J Membr Sci 396:121
employed for the transport of Cs(I) in
laboratory-scale studies, calix crowns (where
crown ether structure has been appended to a
calix[4]arene) have been proposed as one of the
Crown Ethers most efcient carrier extractants for the effective
transport of radio-cesium (Casnati et al. 1995).
P. K. Mohapatra Transport of organic/biological receptors like
Radiochemistry Division, Bhabha Atomic cytochrome C (Paul et al. 2003) and amino acids
Research Centre, Mumbai, Maharashtra, India (Yamaguchi et al. 1988) has also been facilitated
by crown ethers.
Usually, the neutral macrocyclic ionophores
Pedersen (1967) synthesized the crown ethers (see like crown ether, cryptand, calixarenes, cavitands,
Fig. 1a) for the rst time in 1967 which formed etc. require a large lipophilic counter anion like
Crown Ethers 489
O
H
O O N
H
O
OCH2COH
HN O
O O O O
NH
O O
O O O
O O
O O O O
O
HN
O O O
NH
O O C
O
O
N O
H O
Crown Ethers, Fig. 1 Structural formulae of (a) a typical crown ether, (b) valinomycin, and (c) an ionizable lariat crown
ether
picrate, tetraphenyl borate, etc. for the effective arene-crown-6 conformers: a new class of cesium
transport of the cationic species from the aqueous selective ionophores. J Am Chem Soc 117:27672777
Dozol JF, Casas J (1994) Inuence of the extractant on
phase to the membrane phase. The ion pairs are strontium transport from reprocessing concentrate solu-
subsequently transported across the membrane tions through at-sheet supported liquid membranes.
phase which usually contains a polar diluent Sep Sci Technol 29:19992018
such as chloroform, long-chain alcohols, ethers, Dozol JF, Casas J, Sastre AM (1995) Transport of cesium
from reprocessing concentrate solutions through at-
etc. for the stabilization of the charged metal- sheet supported liquid membranes: inuence of the
carrier (crown ether) complex. However, the extractant. Sep Sci Technol 30:435448
large molar volume counter anions can make the Lamb JD, Izatt RM, Robertson PA, Christensen JJ
diffusion of the ion pair rather slow. Appending an (1980) Highly selective membrane transport of Pb2+
from aqueous metal ion mixtures using macrocyclic
ionizable pendent arm to the crown ether ring (the carriers. J Am Chem Soc 102:24522454
ligand is called an ionizable lariat ether; see Paul D, Suzumura A, Sugimoto H, Teraoka J, Shinoda S,
Fig. 1c) can eliminate the need for the counter Tsukube H (2003) Chemical activation of cytochrome c
anions making it a more efcient transport system proteins via crown ether complexation: cold-active
synzymes for enantiomer-selective sulfoxide oxidation
(Strzelbicki et al. 1989). Other types of functiona- in methanol. J Am Chem Soc 2003(125):1147811479
lization have been found to impart exotic proper- Pedersen CJ (1967) Cyclic polyethers and their complexes
ties to the receptors. For example, crown ethers with metal salts. J Am Chem Soc 20:70177022
with suitable functional groups based on redox- Raut DR, Mohapatra PK, Manchanda VK (2012) A highly
efcient supported liquid membrane system for selec-
switched (Shinkai et al. 1985a), thermosensitive tive strontium separation leading to radioactive waste
(Shinkai et al. 1985b), and photoresponsive remediation. J. Membr. Sci. 390391: 7683
(Shinkai et al. 1981) properties have also been Shinkai S, Nakaji T, Ogawa T, Shigematsu K, Manabe
used for the transport of receptors. Functionalized O (1981) Photoresponsive crown ethers. 2.
Photocontrol of ion extraction and ion transport by a
crown ether-type ligands with carboxylate groups bis(crown ether) with a buttery-like motion. J Am
have been suggested for the recovery of U from Chem Soc 103:111115
seawater which is yet another exotic application Shinkai S, Inuzuka K, Miyazaki O, Manabe O (1985a)
of the crown ether-based membrane transport Redox-switchable crown ethers. 3. Cyclic-acyclic
interconversion coupled with redox between dithiol
systems. and disulde and its application to membrane transport.
J Am Chem Soc 107:39503955
Shinkai S, Nakamura S, Tachiki S, Manabe O, Kajiyama
References T (1985b) Thermocontrol of ion permeation through
ternary composition membranes composed of polymer/
liquid crystal/amphiphilic crown ethers. J Am Chem
Casnati A, Pochini A, Ungaro R, Ugozzoli F, Arnaud F, Soc 107:33633365
Fanni S, Schwing MJ, Egberink RJM, De Jong F, Strzelbicki J, Charewicz WA, Liu Y, Bartsch RA
Reinhoudt DN (1995) Synthesis, complexation and (1989) Solvent extraction and bulk liquid membrane
membrane transport studies of 1,3-alternate calix[4]
490 Cryogels
transport of Co(II) and Ni(II) ammine cations by polymeric cryogels, are found to have an increas-
proton-ionizable crown ethers. J Incl Phenom Mol ing use in the separation science due to their easy
Recognit Chem 7:349361
Yamaguchi T, Nishimura K, Shinbo T, Sugiura M (1988) preparations, excellent ow properties, and high
Amino acid transport through supported liquid mem- performances compared with the conventional
branes: mechanism and its application to enantiomeric beads. Cryogels are mega-porous three-
resolution. Bioelectrochem Bioenerg 20:109123 dimensional networks formed under freezing con-
ditions. The pore size of the cryogels varies from
10 to 250 mm (Fig. 1), which can be changed by
optimizing the freezing regime and type and con-
Cryogels centrations of polymerization precursors. The
unique properties of cryogels like osmotic, chem-
Nilay Bereli, Handan Yavuz and Adil Denizli ical, and mechanical stability, large pores, short
Department of Chemistry, Biochemistry Division, diffusion path, low-pressure drop (Fig. 2), and
Hacettepe University, Ankara, Turkey short residence time for both binding and elution
stages make them attractive matrices for afnity
chromatography of large molecules such as pro-
Chromatography is the most powerful technology teins, plasmids even whole cells, as well as small
in the downstream applications for the separation molecules (Lozinsky et al. 2001; Stela and
of proteins both in the analytical and large scale. Valentina 2013; Bereli et al. 2008, 2010;
Conventional packed bed columns have been Tamahkar et al. 2011; Derazshamshir
used for many applications; however, they have et al. 2010). Therefore, cryogels can be used in
some important drawbacks such as the slow dif- the various afnity chromatography applications
fusional mass transfer and the large void volume such as protein A afnity chromatography, histi-
between the beads (Gunko et al. 2013). In order dine afnity chromatography, thiophilic afnity
to resolve these problems, nonporous beads and chromatography, metal-chelate afnity chroma-
perfusion chromatography packing have been tography, dye afnity chromatography,
designed and used as a carrier, but these adsor- ion-exchange chromatography, DNA afnity
bents are not sufcient to resolve these limitations chromatography, cell afnity chromatography,
in essence. New-generation stationary phases i.e., molecular imprinting technique, etc.
Cryogels, Fig. 1 SEM

images of the PHEMA
cryogel
Crystal Grow 491
0.3 new atoms, ions, molecules, or polymer strings

on preformed crystalline nuclei coming from the
nucleation stage. The result of the crystal growth
is the formation of a crystalline solid, whose
Pressure (MPa)
0.2 atoms are reciprocally xed in the space

according to the symmetry rules characteristic of
the space group in which the growth takes place. C
Similarly to crystal nucleation, crystal growth
0.1
requires that the concentration of the solute is in
the supersaturation range (metastable zone): when
supersaturation is reached, the number of mole-
0.0 cules that attach the crystal surface exceeds that of
0 100 200 300 400 the molecules that detach, leading to an increase
Flow rate (cm/h)
of crystal dimension. The strength of bonding
Cryogels, Fig. 2 Pressure drop at different ow rates formed between the crystal surface and the new
molecules depends on the temperature rather than
References the concentration of the solute. Accordingly, crys-
tal growth requires a lower supersaturation level
Bereli N, Andac M, Baydemir G, Say R, Galaev IY, Denizli than that required for nucleation stage; however,
A (2008) Protein recognition via ion coordinated too low supersaturation can result in the crystal
molecularly imprinted supermacroporous cryogels.
dissolution. Traditionally, crystal growth is
Bereli N, ener G, Altinta EB, Yavuz H, Denizli A (2010) divided in two distinct phases: (a) transfer of
Poly(glycidyl methacrylate) beads embedded cryogels mass from the crystallizing solution to the crys-
for pseudo-specic afnity depletion of albumin and tal/solution interface, which is governed by diffu-
immunoglobulin G. Mater Sci Eng 30:323
sion and convection mechanisms, and (b) the
Derazshamshir A, Baydemir G, Anda M, Say R, Galaev
IY, Denizli A (2010) Molecularly imprinted PHEMA- attachment of molecules on the preformed crystal.
based cryogel for depletion of hemoglobin from human Crystal surface consists of at regions (terraces),
blood. Macromol Chem Phys 211:657 raised layers (steps), islands, and kinks (Fig. 1a),
Gunko VM, Savina IN, Mikhalovsky SV (2013) Cryogels:
each having different abilities to interact with new
morphological, structural and adsorption characteriza-
tion. Adv Colloid Interface Sci 187188:1 molecules from the bulk of the solution. Differ-
Lozinsky VI, Plieva FM, Galaev IY, Mattiasson B (2001) ently from the at surfaces, kinks are the most
The potential of polymeric cryogels in bioseparation. suitable sites to promote crystal growth, since the
Bioseparation 10:163
presence of adjacent walls to which new mole-
Stela DE, Valentina DM (2013) Design, synthesis and
interaction with Cu2+ ions of ice templated composite cules can attach allows reducing the activation
hydrogels. Res J Chem Environ 17:4 energy to incorporate new molecules (Chernov
Tamahkar E, Bereli N, Say R, Denizli A (2011) Molecu- 1984). In the presence of high density of kinks,
larly imprinted supermacroporous cryogels for cyto-
crystal growth occurs in spiral fashion around the
chrome c recognition. J Sep Sci 34:3433
single kinks and expands by rotating around the
emergence of the screw dislocation (screw dislo-
cation mechanism) (Fig. 1b). Differently, in the
Crystal Grow
case of crystals with perfectly at faces, the mol-
ecules coming from solution tend to form
Benny Danilo Belviso
two-dimensional layers on the surface due to the
Istituto di Cristallograa, Consiglio Nazionale
lack of suitable attachment points to bind the
delle Ricerche (C.N.R), Bari, Italy
crystal surface. The mechanism is slow and
requires high supersaturation. In both cases, the
Crystal growth is the second stage of the crystal- increase of crystal dimension stops when the den-
lization process and consists in the addition of sity of impurities on the surface of the crystal
492 Crystal Nucleation
Crystal Grow, Fig. 1 Crystal growth process. (a) The formation of the spiral structure (growth direction in red
molecule of solute binds to the kink site after adsorbing line), since the new segment of the step cannot move until
and diffusing across the terraces. (b) Geometry of a screw it reaches that size
dislocation. The existence of a critical size leads to the
reaches a critical value able to poison the process Cross-References

(Durbin and Feher 1990). Quality of the crystals
strictly depends on the supersaturation level and, Crystal Nucleation
hence, on the ratio between the rate of nucleation
and crystal growth: while a low level of supersat-
References
uration favors the crystal growth and slowly leads
to few large crystals usually having high quality Chernov AA (1984) Modern crystallography III: crystal
(Frank 1949), higher supersaturation supports an growth. Springer, Berlin
excess of nucleation and leads to a rain of small Curcio E, Di Proo G, Drioli E (2003) A new membrane-
crystals. The use of membranes as dosing devices based crystallization technique: tests on lysozyme.
J Cryst Growth 247:166176
to modulate the solvent removal from the crystal- Durbin SD, Feher G (1990) Studies of crystal growth
lizing solution by varying process parameters and mechanisms of proteins by electron microscopy.
membrane features makes possible the control of J Mol Biol 212:763774
the extent both of nucleation and of crystal Frank FC (1949) The inuence of dislocations on crystal
growth. Discus Faraday Soc 5:4854
growth: variations able to increase the supersatu- Zhang X, Zhang P, Wei K, Wang Y, Ma R (2008) The study
ration stimulate the nucleation at the expenses of of continuous membrane crystallization on lysozyme.
crystal growth, while variations inuencing the Desalination 219:101117
transmembrane ux can slow the achieving of
the supersaturation, encouraging the crystal
growth over nucleation. It follows that crystals
appear after more time, but they are larger and Crystal Nucleation
usually of high quality (Curcio et al. 2003; Zhang
et al. 2008). In addition, the use of membrane Benny Danilo Belviso
devices as active supports of crystallization allows Istituto di Cristallograa, Consiglio Nazionale
to induce the nucleation and crystal growth stages delle Ricerche (C.N.R), Bari, Italy
in separate sites. This reduces the risk of mem-
brane fouling, even when the same membrane
supports heterogeneous nucleation, and allows to Crystal nucleation is the rst stage of a crystal-
fully exploit the advantages coming from hetero- lization process and leads to the formation of
geneous nucleation and growth at low supersatu- distinct crystal nuclei. It is a local phenomenon
ration, leading to a great production of large of aggregation, and it is said to be homogeneous
crystals. when it occurs in the bulk of the solution and
Crystal Nucleation 493
Crystal Nucleation,
Fig. 1 Effect of the
nucleus radius (r) on the
free energy of germination
DGg (in black color).
Interfacial free energy and
free energy of aggregation
are reported in red and
green color, respectively
C
heterogeneous on very small solid particles, such place. The mechanism and the rate of nucleation
as amorphous clusters or crystalline particles. In strongly inuence quality and quantity of the pro-
both cases, the formation of crystal nuclei requires duced crystals; therefore, many efforts are
well-oriented collisions between molecules to devoted to control the nucleation in order to obtain
occur. The activation energy necessary to induce crystals having high quality. One of the main
nucleation is represented by the free energy of problems in industrial or laboratory crystallization
germination D(Gg) shown in Eq. 1 (Mullin processes is the excess of nucleation that usually
1972), consisting of a positive term called inter- leads to a rain of tiny poor-quality crystals. The
facial free energy (in red color in Fig. 1) and a use of ltered solutions is the solution commonly
negative term called free energy of aggregation adopted to overcome this problem; however, in
(in green color in Fig. 1): the cases where nucleation is welcome, the pres-
ence of dust or small foreign particles, such as
3
DGg 4prr 2 4pr SkT =3V (1) mica or silica able to promote heterogeneous
nucleation, can represent an advantage (Falini
In the Eq. 1, V is the volume of the molecule to et al. 2002; Saridakis and Chayen 2009). These
crystallize, r is the interfacial free energy, T is the particles, also called crystallization seeds, act as
temperature, k is the Boltzmann factor, and r is the nucleant and usually increase the chances of crys-
average radius of the crystal nuclei. DGg reaches a tallization by facilitating the proper molecular
maximum value for a given nucleus radius orientation during the formation of the crystal
(Fig. 1); therefore, only crystal nuclei that grow nuclei (Bolanos-Garcia and Chayen 2009). The
up to this dimension are stable, while smaller result is the reduction of the activation energy
nuclei dissolve before to pass the next crystal and, consequently, the decrease of the supersatu-
growth. In order to reduce the energy of activa- ration level required to induce the formation of
tion, high S value, that is an estimation of the crystal nuclei with respect to that required in the
supersaturation level of the solution, is required. homogeneous nucleation. In these conditions,
Supersaturation is an essential condition required nucleation could even occur in the metastable
during each stage of the crystallization process; zone where crystals grow larger and better ordered
however, the nucleation, being the most energy than those grown at higher supersaturation level.
expensive stage of the entire crystallization pro- The direct contact between crystallizing solution
cess, requires higher supersaturation level to take and an irregular surface can induce heterogeneous
494 Crystal Nucleation
nucleation, as occurs in membrane crystallizers formation of crystal nuclei also in the presence of
(Di Proo et al. 2003, 2005). The local entrapping lower amount of initial substance. For this reason,
of the solute molecules in the pores of the mem- membrane crystallizers are particularly attractive
brane causes an increase of the local level of for crystallization processes where the solute is
supersaturation (Curcio et al. 2006), inducing the expensive, such as in the case of proteins. In
addition, a ne regulation of the nucleation rate
can be achieved by managing the transmembrane
ux of the solvent from the crystallizing solution
toward the stripping solution located on the other
side of the membrane. Nucleation mechanism on
the membrane (Fig. 2) consists of a rst phase
where solute molecules are adsorbed on the sur-
face by means of attractive interactions able to
properly orient the molecule and that usually
drive the growth of the crystal toward a specic
crystal structure (Di Proo et al. 2009). The effec-
tive control of the nucleation process by means of
an appropriate choice of the porosity and hydro-
phobicity of the membrane and by managing the
transmembrane ux makes membrane crystal-
lizers particularly suitable for industrial
applications.
References
Bolanos-Garcia VM, Chayen NE (2009) New directions in

conventional methods of protein crystallization. Prog
Biophys Mol Biol 101:2637
Curcio E, Fontananova E, Di Proo G, Drioli E (2006)
Inuence of the structural properties of poly(vinylidene
uoride) membranes on the heterogeneous nucleation
rate of protein crystals. J Phys Chem
B 110:1243812445
Di Proo G, Curcio E, Cassetta A, Lamba D, Drioli
E (2003) Membrane crystallization of lysozyme:
kinetic aspects. J Cryst Growth 257:359369
Di Proo G, Perrone G, Curcio E, Cassetta A, Lamba D,
Drioli E (2005) Preparation of enzyme crystals with
tunable morphology in membrane crystallizers. Ind
Eng Chem Res 44:1000510012
Di Proo G, Curcio E, Ferraro S, Stabile C, Drioli E (2009)
Effect of supersaturation control and heterogeneous
nucleation on porous membrane surfaces in the crystal-
lization of L-glutamic acid polymorphs. Cryst Growth
Des 9:21792186
Falini G, Fermani S, Conforti G, Ripamonti A (2002)
Protein crystallisation on chemically modied mica
Crystal Nucleation, Fig. 2 Mechanism of crystal nucle- surfaces. Acta Crystallogr D 58:16491652
ation on membrane. Solute molecules (red spheres) are Mullin JW (1972) Crystallization. Butterworths, London
adsorbed on the surface and are trapped inside the pores Saridakis E, Chayen NE (2009) Towards a universal
of the membrane, resulting in an increase of local nucleant for protein crystallization. Trends Biotechnol
concentration 27:99106
Crystal Polymorphism 495
phenomenon, which refers to pure elements.

Crystal Polymorphism Crystal forms that differ for packing or for molec-
ular conformation are said packing or conforma-
Polymorphism tional polymorphs, respectively. From a
thermodynamic viewpoint, polymorphism can be
Benny Danilo Belviso of two kinds: enantiotropic (Fig. 1a), in the case of
Istituto di Cristallograa, Consiglio Nazionale crystal forms stable in a precise range of temper- C
delle Ricerche (C.N.R), Bari, Italy ature and pressure (e.g., the system rhombic
sulfur-monoclinic sulfur in Fig. 1b), and
monotropic (Fig. 1c), in the case of crystallization
The term polymorphism comes from the Greek process that spontaneously evolves toward a crys-
word polus = many and morph = shape, and it is tal form characterized by lower internal energy
the propriety of the chemical substances to exist in with respect to the other forms (metastable
more crystal forms, each having a specic forms). Generally, monotropes and enantiotropes
arrangement of atoms or ions in the crystal lattice. can be recognized by their heats of fusion, since
Crystal polymorphism includes the allotropy monotropes come from exothermic processes and
a b
Pressure
Free energy
Monoclinic form 153C

1420 atm
95C Liquid
1 atm
A form
Rhombic 445C
form
115C
10-5 atm Gas
115C
B form 10-5 atm
Tx TmB TmA temperature temperature
c
Free energy
A form
B form
TmB TmA temperature
Crystal Polymorphism, Fig. 1 Enantiotropic and enantiotropic system rhombic sulfur-monoclinic sulfur. (c)
monotropic systems. (a) Variation of free energy with Variation of free energy with temperature for enantiotropic
temperature for enantiotropic substances. The temperature substances. No crossing of free energy curves belonging to
Tx, which marks the boundary between stability zones for the A and B forms can be observed below of their melting
A and B forms, is below the melting temperatures of the temperatures (TmA and TmB)
two substances (TmA and TmB). (b) Phase diagram for the
496 Crystal Polymorphism
enantiotropes from endothermic processes. What- managing of the transmembrane ux and an accu-
ever is the kind of polymorphism, under static rate choose of the membrane properties (Di Proo
crystallization conditions, the probability of car- et al. 2007, 2009, 2013). So crystallization
rying out crystallization precisely at the boundary assisted by membrane can occur both under ther-
that separates the polymorphic forms is small; modynamic and kinetic control: low evaporation
therefore only one crystal form appears at the rates through the membrane allow the growth of
end of the crystal growth. The mechanism that the more stable nuclei at the expense of the less
explains the polymorphism involves the stage of stable forms (thermodynamic control), while the
crystal nuclei formation. During the nucleation induction of a more rapid nucleation increases in
stage, some molecules can leave the preformed the presence of higher evaporation rates and
aggregates, leading to their complete dissolution. increases the conversion rate from the stable to
The surviving aggregates provide the new differ- the metastable form (kinetic control). The relative
ently arranged templates for the subsequent amount of the two forms depends on the ratio
growth stage, and the molecular arrangement in between the conversion and crystallization rate.
largest amount will form the nuclei that will grow Similarly, the production of a specic polymorph
in a particular crystal form. The critical dimension can be obtained by using antisolvent membrane
that the aggregates must reach to survive depends crystallization processes (Di Proo et al. 2010).
of the supersaturation level; therefore, the entire
process is under kinetic control rather than ther-
modynamic (Volmer 1939; Ostwald 1987).
References
Today, the interest for the phenomenon of poly-
morphism is increasing, particularly for pharma- Arnesano F, Belviso BD, Caliandro R, Falini G, Fermani S,
ceutical industry where its study is became Natile G, Siliqi D (2011) Crystallographic analysis of
integral part of the drug optimization process metal-ion binding to human ubiquitin. Chemistry
(Brittain 1999). Paracetamol (Beyer et al. 2001), 17:15691578
Beyer T, Day GM, Price SL (2001) The prediction, mor-
famotidine (Lu et al. 2007), and piroxicam phology, and mechanical properties of the polymorphs
(Giordano et al. 1998) are the classical examples of paracetamol. J Am Chem Soc 123(21):508694
of drugs crystallized in several forms, each having Brittain HG (1999) Polymorphism in pharmaceutical
different chemical and pharmaceutical proprieties. solids. Marcel Dekker, New York
Di Proo G, Tucci S, Curcio E, Drioli E (2007) Selective
Usually, pharmaceutical industry prefers the most glycine polymorph crystallization by using micropo-
stable form with respect to that metastable, rous membranes. Cryst Growth Des 7:526530
although, at the same concentration, the latter is Di Proo G, Curcio E, Ferraro S, Stabile C, Drioli E (2009)
more available in the organism due to its higher Effect of supersaturation control and heterogeneous
nucleation on porous membrane surfaces in the crystal-
solubility. Also protein molecules can crystallize lization of l-glutamic acid polymorphs. Cryst Growth
in several crystal forms, as a consequence of the Des 9:21792186
different conformations assumed by peptide chain Di Proo G, Caridi A, Caliandro R, Guagliardi A,
in a given condition. A classic example of poly- Curcio E, Drioli E (2010) Fine dosage of antisolvent
in the crystallization of l-histidine: effect on polymor-
morphism for crystal of proteins is the lysozyme phism. Cryst Growth Des 10:449455
that crystallizes in six different crystalline forms. Di Proo G, Reijonen MT, Caliandro R, Guagliardi A,
Another important example is the ubiquitin, Curcio E, Drioli E (2013) Insights into the polymor-
whose crystal form depends on the nature and phism of glycine: membrane crystallization in an elec-
tric eld. Phys Chem Chem Phys 15:92719280
the amount of the metal ion used during the crys- Falini G, Fermani S, Tosi G, Arnesano F, Natile G (2009)
tallization (Arnesano et al. 2011; Falini Structural probing of Zn(II), Cd(II) and Hg(II) binding
et al. 2009). The technology of membrane crys- to human ubiquitin. Chem Commun (Camb)
tallizers gives the opportunity to drive the crystal- 45:59605962
Giordano F, Gazzaniga A, Moyano JR, Ventura P, Zanol
lization toward a specic polymorph. This is M, Peveri T, Carima L (1998) Crystal forms of
possible by regulating the degree and the rate to piroxicam pivalate: preparation and characterization
which of supersaturation is reached through the of two polymorphs. J Pharm Sci 87(3):333337
Crystal Purity 497
Lu J, Wang XJ, Yang X, Ching CB (2007) Cryst Growth & may be carried out to increase crystal purity by
Des 2007(7):15901598 removing many of the surface impurities or those
Ostwald WZ (1987) Studien uber die Bildung und
Umwandlung fester Korper. Z Phys Chem 22:289330 specically attached inside the crystal. However,
Volmer M (1939) Kinetik der Phasenbildung. Steinkopf, the technique is quite risky, since the washing
Leipzig solvent should be able to remove the impurities
without dissolving the crystal. Although contam-
ination is usually the rst cause of unsuccessful C
crystallization, the alone presence of impurities is
Crystal Purity not sufcient to frustrate the crystallization
attempts. Crystal features, such as crystal lattices
Benny Danilo Belviso having weak interactions between atoms or
Istituto di Cristallograa, Consiglio Nazionale containing large empty spaces where foreign mol-
delle Ricerche (C.N.R), Bari, Italy ecules have possibility to t, should be avoided to
reduce the probability of crystal contamination.
Impurities are included in a crystal as a result of
Crystal purity is the parameter used to indicate the the processes that occur concomitantly to the crys-
contamination level of a crystal and represents tallization (in particular during the nucleation) or
one of the most important properties of the crys- after that crystal growth has nished. Particularly,
talline materials. Crystal purity is also an index of the degree of crystal purity depends on the ratio
the heterogeneity of the substance to crystallize, in between the rates of these inclusion processes
particular in the presence of proteins or nucleic with that of the crystallization process that, being
acids, whose uncontrolled modications may a purication process, pushes toward the recovery
increase the number of species in the crystallizing of the desired product without contaminations
solution. Contaminants may distort the crystal (Burton et al. 1953): usually, slow crystallizations
lattice, affect the crystal features, such as diffrac- allow the substance to crystallize to properly t in
tion quality, and change the crystal morphology in the lattice, differently from more rapid precipita-
particular when the impurities are incorporated in tions, where the insertion of the impurities results
the crystal at different rates into adjacent steps on facilitate. In addition, high supersaturation often
the crystal surface. However, the main feature of a leads to contaminant inclusions that reduce the
low-purity crystal is its higher internal energy crystal purity. At the present, membrane technol-
with respect to that in absence of impurities (van ogy represents one of more promising tools to
Enckevort and van der Berg 1998; Davis produce high-purity crystal. By controlling the
et al. 2004), which causes an increase of the crys- transmembrane ux, membrane crystallizer can
tal solubility and a decrease of the effective super- limit the maximum level of supersaturation and
saturation level of the solution containing the reduce the growth rate to reach the critical thresh-
substance to crystallize: the higher the contamina- old, beyond which inclusion of impurities occurs
tion, the more difcult the crystallization because (Weckesser and Konig 2008; Mullin 1993).
the dissolution probability of the crystals Today, the advances made in the development of
increases. Given this, many efforts should be membrane-assisted crystallization allow produc-
devoted to adequately purify the substance to ing crystalline materials having the purities
crystallize, by paying attention to the purity required for their commercialization as puried
grade of the chemicals employed in each step. product. One of the most important examples is
Particular attention should be paid for precipitant represented by the production of sodium carbon-
agents, since they are used in very high concen- ate crystals having purity grade that reaches
trations: accordingly, salts and polymeric precip- ca. 99.5 % also starting from crystallizing solu-
itants, such as polyethylene glycol or alcohols, tions containing concentrations ranging from 0.2
should be recrystallized or puried, respectively. to 0.6 M of impurities, such as sodium nitrate,
Once the crystal is formed, washing procedures sodium sulfate, and sodium chloride
498 Crystal Quartz Microbalance
(Ye et al. 2013). In addition, membrane devices (in g/cm2), and Cf is the sensitivity factor for the
are also used to produce high-purity cocrystal, a crystal used (in Hz.cm2.ng1). This sensitivity is
challenge particularly interesting for pharmaceu- proportional to the square of the resonant fre-
tical industry. quency, according to the expression
Df 2 f 2n
References Cf
Dm rv n
Burton JA, Prim RC, Slichter WP (1953) The distribution where r is the quartz density, v is the propagation
of solute in crystals grown from the melt. Part
I. Theoretical. J Chem Phys 21:19871991 velocity of the wave in the crystal, fn is the fre-
Davis KJ, Dove PM, Wasylenki LE, Deyoreo JJ quency of the selected harmonic resonant mode,
(2004) Morphological consequences of differential and n is the harmonic number (n = 1 for the
Mg2+ incorporation at structurally distinct steps on fundamental mode). By way of comparison, the
calcite. Am Mineral 89:714720
Mullin JW (1993) Crystallization, 3rd edn. Butterworth- typical balances used in research labs (torsion,
Heinemann, Oxford cantilever, or spring balances) are able to detect
van Enckevort WJP, van der Berg (1998) Impurity mass until 107 g, while QCM can detect up to
blocking of crystal growth: a Monte Carlo study. 1013 kg (Mecea 2005). For many years QCMs
J Cryst Growth 183:441455
Weckesser D, Konig A (2008) Particle shape and purity in were only applied as gas-phase mass detectors in
membrane based crystallization. Chem Eng Technol physical vapor deposition systems. Their applica-
31:157162 tion has been extended recently, with other uses
Ye W, Lin L, Shen J, Luis P, Van der Bruggen B (2013) beyond mass deposition, including the monitoring
Cryst Growth Des 13:23622372
of adsorption kinetics for biomolecules and sen-
sors based on QCM detection.
Crystal Quartz Microbalance References

Osvaldo N. Oliveira Jr. and Diogo Volpati Mecea VM (2005) From quartz crystal microbalance to
Sao Carlos Institute of Physics IFSC, University fundamental principles of mass measurements. Anal
of Sao Paulo USP, Sao Paulo, Brazil Lett 38:753767
Sauerbrey G (1959) Verwendung Von Schwingquarzen
Zur Wagung Dunner Schichten Und Zur
Mikrowagung. Z Phys 155:206222
The quartz crystal microbalance (QCM) is a mass
sensor able to measure mass below the nanogram
scale range. This device was rst introduced by
Sauerbrey in 1959 (Sauerbrey 1959), and its oper- Crystal Quartz Microbalance
ation is based on a piezoelectric plate of quartz for Protein Adsorption Measurement
with two electrodes xed to the plate. The small
mass absorbed at the crystal is detected by Osvaldo N. Oliveira Jr. and Diogo Volpati
decreasing the resonant frequency of the crystal. Sao Carlos Institute of Physics IFSC, University
The Sauerbreys equation is used to associate the of Sao Paulo USP, Sao Paulo, Brazil
mass adsorbed at the QCM crystal surface with
the change in oscillation frequency of the crystal:
With the advent of biosensors, techniques able to
Df Cf Dm monitor the immobilization of active molecules
on electrodes become essential, especially those
where Df is the experimental frequency shift able to provide molecular-level information for
(in Hz), Dm is the change in mass per unit area the adsorption process. Proteins represent an
Crystal Seeding 499
example of molecules largely used in biosensors,

and their immobilization process onto electrode Crystal Seeding
surfaces has been monitored by techniques such
as uorescence labeling (Lee 1997), microarray Benny Danilo Belviso
techniques (Park et al. 2008), surface plasmon Istituto di Cristallograa, Consiglio Nazionale
resonance (SPR) (Smith et al. 2003), and quartz delle Ricerche (C.N.R), Bari, Italy
crystal microbalance (QCM) (Cooper and Single- C
ton 2007). The latter has become the most used
because it allows for determining the amount of Synonyms
bound molecules at the electrode surface and eval-
uation of the adsorption kinetics and of the afnity Seeding
between ligands and analytes in sensors.
A quartz crystal microbalance (QCM) is a device Crystal seeding is a technique widely employed
often used to monitor the material adsorption, being to provide preformed templates (seeds) on which
capable to detect mass change of the order of 109 g new molecules can converge to form crystals.
in commercial systems. The QCM principle of Crystal seeding is said homogeneous if the seeds
operation is based on changing the resonance fre- are made of the same molecule to crystallize, as
quency of a quartz crystal by adsorbing mass on its occurs in macroseeding, microseeding, and streak
surface (Montagut et al. 2011). For monitoring pro- seeding, and heterogeneous if it foresees the use
tein adsorption, usually a ligand is immobilized as a of foreign material, such as porous silicon and
monolayer on a gold surface. Since gold reacts with mica (Saridakis and Chayen 2009), as in the case
thiols producing active sites at the surface, proteins of cross seeding and epitaxial growth. Whatever
can be immobilized by the thiol groups from their the kind of used seeds, their presence generally
cysteine residues or can be activated by using a makes less energetically expensive the entire crys-
thiol-containing bifunctional linker (Bieri and tallization process, because it allows the passing
Burgi 2006). directly to the crystal growth stage without deal-
ing with the intrinsic uncertainties of the nucle-
ation. However, the usage of seeds might have the
References disadvantage to affect the crystal growth and the
forming crystal lattice (Hermes et al. 2011; Liu
Bieri M, Burgi T (2006) Adsorption kinetics of et al. 2007). The rst step for much of seeding
L-glutathione on gold and structural changes during techniques is the production of seeds: during this
self-assembly: an in situ ATR-IR and QCM study. phase, preformed crystals are washed to remove
Phys Chem Chem Phys 8:513520
Cooper MA, Singleton VT (2007) A survey of the 2001 to defects and impurities that could block the crystal
2005 quartz crystal microbalance biosensor literature: growth, stabilized in an appropriate solution, and
applications of acoustic physics to the analysis of bio- crushed by using a glass homogenizer or by using
molecular interactions. J Mol Recog 20:154184 spherical beads. Once the seed stock has been
Lee YC (1997) Fluorescence spectrometry in studies of
carbohydrate-protein interactions. J Biochem obtained, pre-equilibrium conditions (e.g., tem-
121:818825 perature, precipitant agents, additives, etc.) of
Montagut Y, Narbon JG, Jimenez Y, March C, Montoya A, the solution to seed are carefully determined
Arnau A (2011) QCM technology in biosensors. In: together with the appropriate level of supersatura-
Serra PA (ed) Biosensors emerging materials and
applications. InTech, Rijeka, pp 153178 tion, in order to avoid the trigger of a self-
Park S, Lee MR, Shin I (2008) Carbohydrate microarrays nucleation that would frustrate all following
as powerful tools in studies of carbohydrate-mediated efforts. In the absence of such information, the
biological processes. Chem Commun 37:43894399 seeding process could result in a dissolving of the
Smith EA, Thomas WD, Kiessling LL, Corn RM
(2003) Surface plasmon resonance imaging studies of seeds or in a rain of tiny and poor-quality crystals.
protein-carbohydrate interactions. J Am Chem Soc In order to assess seeding conditions, protein crys-
125:61406148 tallographers usually exploit streak seeding
500 Crystal Seeding
technique, where a synthetic or natural ber procedures, such as cross-ow ltration and
(Bergfors 2003) is used to crush existing crystals dead-end ltration (Pera-Titus et al. 2005; Huang
and to introduce the resulting small crystalline et al. 2004), based on the seed deposition on the
particles into different pre-equilibrated solutions: membrane support under the action of a difference
mode and timing of seed growth allow to of pressure applied, have been also developed.
determine both the appropriate range of
pre-equilibration conditions and the dilution
of the seed stock. Although the techniques of
References
seeding were initially developed with the aim to
optimize and make more reproducible the crystal- Bergfors T (2003) Seeds to crystals. J Struct Biol 142:6676
lization processes performed in laboratory, today Bernal MP, Xomeritakis G, Tsapatsis M (2001) Tubular
they are widely employed also in high-throughput MFI zeolite membranes made by secondary (seeded)
processes that range from the industrial crystalli- growth. Catal Today 67:101107
Caro J, Noack M, Kolsch P, Schafer R (2000) Zeolite
zations to the preparation of devices employed in membranes-state of their development and perspective.
membrane reactors or in separation processes, Micropor Mesopor Mater 38:324
such as membranes of zeolite (Li et al. 2005) or Hasegawa Y, Sotowa KI, Kusakabe K, Morooka S (2002)
silicate crystals (Hasegawa et al. 2006). Prepara- The inuence of feed composition on CO oxidation
using zeolite membranes loaded with metal catalysts.
tion of these membrane devices is very challeng- Micropor Mesopor Mater 53:3743
ing because it requires a strong control of the Hasegawa Y, Ikeda T, Nagase T, Kiyozumi Y, Hanaoka T,
seeding in order to obtain high-quality mem- Mizukami F (2006) Preparation and characterization of
branes having determined features, such as a spe- silicalite-1 membranes prepared by secondary growth
of seeds with different crystal sizes. J Membr Sci
cic crystalline orientation (Caro et al. 2000). 280:397405
Several seeding techniques have been developed Hedlund J, Noack M, Kolsch P, Creaser D, Caro J, Sterte
to produce a membrane of zeolite crystal J (1999) ZSM-5 membranes synthesized without
(Table 1), and the secondary growth together organic templates using a seeding technique. J Membr
Sci 159:263273
with dip coating is considered among those most Hermes M, Vermolen ECM, Leunissen ME, Vossen DLJ,
promising: secondary growth exploits a hydro- van Oostrum PDJ, Dijkstra M, van Blaaderen A (2011)
thermal treatment of the membrane support that Nucleation of colloidal crystals on congurable seed
has been previously covered with layers of seeds structures. Soft Matter 7:46234628
Huang A, Lin YS, Yang W (2004) Synthesis and properties
(Lovallo and Tsapatsis 1996), while dip coating of A-type zeolite membranes by secondary growth
takes advantage from the capillary force to aid the method with vacuum seeding. J Membr Sci 245:4151
deposition of crystal seeds (Lovallo et al. 1998; Kusakabe K, Kuroda T, Morooka S (1998) Separation of
Bernal et al. 2001). In addition, ltration carbon dioxide from nitrogen using ion-exchanged
faujasite-type zeolite membranes formed on porous
support tubes. J Membr Sci 148:1323
Li G, Lin RS, Kikuchi E, Matsukata M (2005) Growth
Crystal Seeding, Table 1 Seeding procedures for zeolite mechanism of a preferentially oriented mordenite
membrane preparation membrane. J Zhejiang Univ Sci B 6:369372
Seeding procedure References Liu YX, Wang XJ, Lu J, Ching CB (2007) Inuence of the
Secondary growth Lovallo and Tsapatsis (1996) roughness, topography, and physicochemical proper-
ties of chemically modied surfaces on the heteroge-
Dip coating (Lovallo et al. 1998; Bernal
neous nucleation of protein crystals. J Phys Chem
et al. 2001)
B 111:1397113978
Spin coating Mintova and Bein (2001) Lovallo MC, Tsapatsis M (1996) Preferentially oriented
Rubbing (Kusakabe et al. 1998; submicron silicalite membranes. AIChE
Hasegawa et al. 2002) J 42:30203029
Cationic polymer use (Hedlund et al. 1999; Weh Lovallo MC, Gouzinis A, Tsapatsis M (1998) Synthesis
et al. 2002) and characterization of oriented MFI membranes pre-
Cross-ow ltration Pera-Titus et al. (2005) pared by secondary growth. AIChE J 44:19031913
Mintova S, Bein T (2001) Microporous lms prepared by
Dead-end ltration Huang et al. (2004)
spin-coating stable colloidal suspensions of zeolites.
(vacuum seeding)
Adv Mater 13:18801883
Crystallization 501
Pera-Titus M, Llorens J, Cunill F, Mallada R, Santamara three-dimensional lattice. Differently, in the

J (2005) Preparation of zeolite NaA membranes on the absence of orderly arrangement, solidication
inner side of tubular supports by means of a controlled
seeding technique. Catal Today 104:281287 process leads to amorphous precipitates or to
Saridakis E, Chayen NE (2009) Towards a universal solids consisting of a mix of crystalline and amor-
nucleant for protein crystallization. Trends Biotechnol phous areas (as it happens for some polymers).
27:99106 Crystallization takes place when the solution
Weh K, Noack M, Sieber I, Caro J (2002) Permeation of
single gases and gas mixtures through faujasite-type containing the substance to crystallize reaches C
molecular sieve membranes. Micropor Mesopor the supersaturation state, namely, when the con-
Mater 54:2736 centration of this substance overcomes the solu-
bility limit while remaining below the threshold of
precipitation. The kinetics and the thermodynam-
ics of the crystallization depend on the supersatu-
Crystallization ration level that, hence, represents the driving
force of the entire process. Crystallization consists
Benny Danilo Belviso of two major events: nucleation and crystal
Istituto di Cristallograa, Consiglio Nazionale growth. Nucleation occurs in the metastable
delle Ricerche (C.N.R), Bari, Italy zone (Fig. 1, in yellow) as a consequence of the
rapid local molecular collisions characterizing the
homogeneous phase (solution or gas) in condition
Crystallization is a phase transition that takes of supersaturation. Crystalline nuclei, the smallest
place from the liquid toward the solid state molecular aggregates having a structure ordered
resulting in the formation of crystals; less fre- according to periodic rules, can be formed only
quently, the deposition of crystals can also occur when well-oriented collisions between particles
starting from the gas state. In any case, crystalli- take place. The smallest nuclei dissolve, while
zation process represents a particular kind of crystal nuclei having critical dimensions pass to
solidication in which the molecules or ions of the next step, where other molecules of solute can
a given substance arrange themselves in an converge on the preformed aggregates. This stage
orderly fashion according to a substance-specic is called crystal growth and takes place in the
Crystallization,
Fig. 1 Crystallization
phase diagram. Crystal
nucleation takes place in
metastable zone (in yellow)
and crystals growth in labile
zone (in pink), both
belonging to the
supersaturation zone. The
condition of supersaturation
is reached when solute
concentration is slightly
higher than the solubility
(gray line); differently, a
much higher solute
concentration leads to
amorphous precipitation
(light cyan area)
502 Crystallizing Solution
labile zone (Fig. 1, in pink). During the crystal References

accretion, also the concentration of contaminants
on crystal surface increases causing a slowdown Curcio E, Simone S, Di Proo G, Drioli E, Cassetta A,
Lamba D (2005) Membrane crystallization of lyso-
of the growth up to its complete stop. The adding
zyme under forced solution ow. J Membr Sci
of small particles, such as small crystals or amor- 257:134143
phous solids (Saridakis and Chayen 2009), can Di Proo G, Curcio E, Drioli E (2005) Trypsin crystalliza-
promote the crystallization process by acting as tion by membrane-based techniques. J Struct Biol
150:4149
nucleation centers. Although the rates of nucle-
Di Proo G, Curcio E, Drioli E (2010a) Supersaturation
ation and crystal growth increase with the super- control and heterogeneous nucleation in membrane
saturation level, fast crystallization processes at crystallizers: facts and perspectives. Ind Eng Chem
low supersaturation are possible in the presence of Res 49:1187811889
Di Proo G, Caridi A, Caliandro R, Guagliardi A,
crystal defects. Today, advanced crystallization
Curcio E, Drioli E (2010b) Fine dosage of antisolvent
techniques allow getting crystals with few defects in the crystallization of L-histidine: effect on polymor-
even starting from lower levels of supersaturation. phism. Cryst Growth Des 10:449455
This is possible by means of an effective control Di Proo G, Reijonen MT, Caliandro R, Guagliardi A,
Curcio E, Drioli E (2013) Insights into the polymor-
of the nucleation and crystal growth steps, as
phism of glycine: membrane crystallization in an elec-
occurs in the case of the membrane-assisted crys- tric eld. Phys Chem Chem Phys 15:92719280
tallization (Drioli et al. 2012), where the entire Drioli E, Di Proo G, Curcio E (2012) Progress in mem-
crystallization process takes place on a support brane crystallization. Curr Opin Chem Eng 1:178182
Saridakis E, Chayen NE (2009) Towards a universal
consisting of a polymeric membrane that acts
nucleant for protein crystallization. Trends Biotechnol
also as active surface for heterogeneous nucle- 27:99106
ation (Di Proo et al. 2010a). On the other side
of the membrane (distillate side), a stripping solu-
tion allows the removal of the solvent from the
crystallizing solution. The high hydrophobicity of
the membrane prevents its wetting, allowing that Crystallizing Solution
the mass transfer between crystallizing solution
and stripping solution takes place only by vapor Benny Danilo Belviso
diffusion mechanism. Since the solute species, Istituto di Cristallograa, Consiglio Nazionale
such as macromolecules and electrolytes, have delle Ricerche (C.N.R), Bari, Italy
lower volatility than the solvent, only the latter
passes from the crystallizing solution to the strip-
ping solution through the membrane, by resulting Crystallizing solution (also called crystallizing
in an increase of solute concentration. Therefore, cocktail) is the ensemble of chemicals able to
the driving force of the membrane-assisted crys- promote the crystallization of a substance in a
tallization is represented by the vapor pressure given physical condition (e.g., temperature, vibra-
gradient at both sides of the membrane. Kinetics tions, surface of crystallization, etc.). Crystalliza-
and thermodynamics of the nucleation process tion requires that the solution containing the
can be modulated according to the transmembrane substance to crystallize is supersaturated with
ux and the features of the membrane, offering the respect to the latter, namely, that the solute con-
opportunity of a more rationale design of the centration is higher than its solubility limit but
crystallization process. Crystallization on mem- lower than the precipitation threshold: an appro-
brane has been extensively tested both on small priately formulated crystallizing solution reduces
molecules (Di Proo et al. 2010b, 2013) and on the solubility of the substance (both in the case of
large macromolecules, such as proteins (Di Proo small inorganic/organic molecules and biological
et al. 2005; Curcio et al. 2005). macromolecules, such as proteins), enabling a
Crystallizing Solution 503
Crystallizing Solution, Fig. 1 Vapor phase transfer in antisolvent membrane crystallizers, the solubility of the
membrane crystallizers. (a) On the crystallizing solution solute decreases due to the increase of the antisolvent
side, the solute concentration increases due to the solvent concentration inside the crystallizing solution usually gen-
removal triggered by the gradient of vapor pressure erated by a temperature difference between the two sides of
between the two sides of the membrane. (b) In the case of the membranes
more easily reaching of the crystallization range, promoting weak interactions is usually very
where spontaneous formation of the rst crystal- small, it is necessary to explore several formula-
line nuclei and their next growth can occur. Usu- tions before obtaining the best one. One of the
ally, crystallizing solution consists of a most used methods to deal with the crystallizing
precipitating agent, a substance whose concentra- solution formulation problem is the multivariate
tion is inversely correlated to the solubility of the analysis that allows testing several successful
molecule to crystallize, and a buffer salt able to events of crystallization against a high-
stabilize the pH during the crystallization process. dimensional space represented by the crystalliza-
In order to optimize the crystallization, additives, tion conditions (Segelke 2001). Depending on the
whose concentration is lower than that of the main crystallization method, the concentration of each
components, can be added. In the case of crystal- component of the crystallizing solution can vary
lization of proteins, ligands and cofactors can be or keep unchanged during the crystallization pro-
exploited to stabilize the protein or to increase its cess: no change occurs in the case of the batch
conformational purity. Notwithstanding it is com- method where the level of supersaturation is
mon to classify the components of a crystallizing directly adjusted on the crystallizing solution.
solution according to these three categories, a Conversely, the component concentrations
sharp distinction is not possible because a sub- change when the supersaturation level is reached
stance might act as precipitating agent in a given in the time by using vapor or liquid diffusion
cocktail and be used as an additive in a different techniques. Techniques that exploit membranes
case. Besides to affect the solubility, the compo- to put in contact a crystallizing solution
nents forming the crystallizing solution have to containing the molecule to crystallize and a strip-
weakly interact with the molecule of the substance ping hypertonic solution or an antisolvent on the
to crystallize (Wilson 2003). Differently, stronger other side of the membrane (distillate side) belong
interactions make more soluble the substance to to the latter category. Due to its hydrophobicity,
crystallize, while the absence of interactions dras- membrane is not wet by the two liquids on the two
tically reduces the solubility, resulting in an amor- sides and the ux of solvent molecules through
phous precipitation. As the space of the conditions the membrane can occur by vapor diffusion
(nature and concentration of each component) (Fig. 1a). Accordingly, it is possible to achieve a
504 Cutoff Membrane
ne regulation to the rate to which solute increases are suspended in a liquid or mixed in the gas. The
its concentration in the crystallizing solution by membrane is a physical barrier that allows certain
managing the transmembrane ux. Similarly, in compounds to pass through, depending on their
the case of antisolvent membrane devices, vapor physical and/or chemical properties. Therefore,
diffusion triggers the increase of the antisolvent the cutoff property of membrane is a signicant
concentration inside the crystallizing solution parameter.
(Fig. 1b), resulting in a decrease of the solute In the liquid separation process, a membrane
solubility (Di Proo et al. 2009). By using hydro- lter comprises a membrane with a pore size
philic membranes, it is possible to exploit liquid- between 5 nm and 50 mm. On the basis of the
phase transfer: in this case, crystallizing solution is cutoff membrane rate, the membrane technologies
directly pressed through a porous membrane in an can be classied into microltration (MF), ultra-
antisolvent (or vice versa). This technique has been ltration (UF), nanoltration (NF), and reverse
used to produce crystal of L-asparagine with nar- osmosis (RO) (Table 1). The MF membrane has
row size distribution (Zarkadas and Sirkar 2006). pore sizes in the range of 0.15 mm and is capable
of removing suspended particles like blood cell,
microorganism, and latex emulsions. The UF
Cross-References
membranes with pore sizes in the range of
2100 nm can remove large molecules like albu-
Crystallization
min, bacteria, or pepsin. The cutoff membrane
Hydrophobicity
rate of UF can also be characterized in terms of
Multivariate Analyses
the so-called molecular weight cutoff (MWCO),
Solubility
which is generally expressed in standard dalton
(Da) or kilodalton (kDa) units. While this classi-
References cation method is inherently decient since the
atomic mass does nothing to describe the actual
Di Proo G, Stabile C, Caridi A, Curcio E, Drioli E (2009) size or geometry of a molecule, it remains the
Antisolvent membrane crystallization of pharmaceuti-
standard convention for UF membrane classica-
cal compounds. J Pharm Sci 98:49024913
Segelke BW (2001) Efciency analysis of sampling pro- tion. NF membrane has a pore size less than 1 nm.
tocols used in protein crystallization screening. J Cryst It can separate small molecules like divalent salts,
Growth 232:553562 dissociated acids, and sugar. NF membranes are
Wilson WW (2003) Light scattering as a diagnostic for
generally characterized by their ability to retain a
protein crystal growth-a practical approach. J Struct
Biol 142:5665 divalent ionic species such as magnesium sulfate
Zarkadas DM, Sirkar KK (2006) Antisolvent crystalliza- (MgSO4) or calcium chloride (CaCl2). The typical
tion in porous hollow ber devices. Chem Eng Sci rejection rate of MgSO4 might range from around
61:50305048
80 % to 98 %. RO membrane with pore sizes in
the range of few angstroms can separate water
molecules or ions like sodium and chloride on
Cutoff Membrane the molecular or ion level. The typical rejection
rate of RO for NaCl can be on the range of
Xianhui Li and Jianxin Li 9899.5 %, which is inuenced by given pressure,
State Key Laboratory of Separation Membranes temperature, and concentration conditions. How-
and Membrane Processes, School of Materials ever, the gas separation requires membrane with
Science and Engineering, Tianjin Polytechnic small pore size.
University, Tianjin, China Many characterization methods like
permporometry, thermoporometry, mercury
porometry, gas adsorptiondesorption, nuclear
The membrane ltration is a very efcient and magnetic resonance (NMR), and liquidliquid
economical way for separating components that porometry, along with several microscopic
Cyclodextrin in Membranes 505
Cutoff Membrane, Table 1 Cutoffs of different liquid filtration techniques
Micrometer 0.001 0.01 0.1 1 10 100 1000
Angstroms 1 10 100 1000 104 105 106 107
Molecular C
weight 0.5 50 7000
Solved Viruses Bacteria Yeast Sand

Substances salts
to be
separated Red blood
Sugar Albumin cells
Reverse Ultra
osmosis filtration Particle filtration
Process
Nano Microfiltration
filtration
techniques, have been used to characterize the characteristics such as pore size distribution (and
membrane pore and pore size distribution. Each mean pore size), porosity (void volume), and tor-
of these methods has different characteristics and tuosity (or the relative length of the path of diffu-
relies on different theoretical considerations to be sion across the membrane material).
taken into account to convert the direct results into
pore sizes. Among these techniques, the bubble
point test has gained enormous relevance for the
characterization of MF membranes. However, it Cyclodextrin in Membranes
cannot be properly applied to UF membranes due
to the high pressure (more than 10 bar) necessary Francesco Trotta
to evaluate pore with sizes below 0.1 mm. On the Department of Chemistry, University of Torino,
contrary, the liquidliquid displacement Torino, Italy
porometry (LLDP) is very suitable for character-
izing UF membranes at relatively low applied
pressures. However, for the dense nonporous Cyclodextrins have the unsurpassed ability to
membrane such as NF, RO, and gas separation form inclusion compounds with suitable guest
membranes, some experimental techniques such molecules. Interaction of native and modied
as positron annihilation lifetime spectroscopy, cyclodextrins with cellular membrane was inten-
inverse gas chromatography, and Xe NMR are sively studied in view of their large use as drug
available for determining the average radius delivery systems. Moreover, cyclodextrins have
(110 ) and the size distribution of the free- found many potential applications in membrane
volume elements. And atomistic modeling can science. In this latter case, in addition to the ability
only explain the topology and structure of free- to form inclusion complexes, also the intrinsic
volume elements in polymers. chirality of the cyclodextrins and their enzyme-
Generally, the selection of a membrane cutoff like behavior were exploited in membrane appli-
rating is 0.20.3 times the value of the molecular cations. Almost all the membrane processes are
weight targeted for retention. Thus, prior to involved independently from their driving forces
use, it is necessary to characterize physical applied (Kozlowski and Sliwa 2008). Native
506 Cyclodextrin in Membranes
cyclodextrins, but also dimers, modied cyclo- propanol isomers, xylene isomers, and water/
dextrins, polymers, and specially cross-linked ethanol and water/1-propanol or 2-propanol
cyclodextrins were used. Cyclodextrin can be were studied at length as were permeabilities
grafted to polymer chains, dispersed in casting greatly inuenced by cyclodextrin contents.
solutions to form composite membranes, used as Pervaporation membranes were also produced
carriers in liquid membranes, bonded to ceramic by using polyurethanes polymers hosting cyclo-
membranes, or cross-linked alone or with suitable dextrins. This system was used essentially to sep-
organic polymers. In this latter case, formalde- arate phenol/water mixtures. Polyimide-based
hyde, glutaraldehyde, citric acid, hexamethylene cyclodextrin-containing membranes found appli-
diisocyanate, tetramethyl orthosilicate, as well as cation in isopropanol dehydration and butanol
ethylene glycol diglycidyl ether were used isomer separation. Cyclodextrin was used in
as cross-linking agents. The main organic membrane reactors as an additive to obtain enan-
polymers used were PVDF, polyvinyl alcohol, tiomeric separation, being the reaction catalyzed
polydimethylsiloxane, silicone rubber, by enzyme or exploiting the enzyme-like behavior
polypyrrole, polyurethane, poly(N-iso- of water-insoluble modied cyclodextrin dis-
propylacrylamide), polyamide-imide, poly- persed in PEEK-WC membranes.
sulfone, PVC, polyacrylonitrile, PEEK-WC, as Polyimide polymer membranes added to
well as chitosan, cellulose, and cellulose cyclodextrins alone or grafted onto carbon nano-
diacetate. Flat sheet and hollow ber membrane tube surfaces have a relevant inuence on
congurations were reported. Phase inversion CO2/CH4 as well as propane/propylene gas sep-
technique and casting followed by a cross-linking aration. The characteristics of cyclodextrins t
reaction were the most used procedures to get well with the requirements for membrane sensor-
cyclodextrin-containing membrane; however, making. Depending on the cyclodextrins and
other techniques were also employed, i.e., membranes used, excellent selectivity toward glu-
electrospinning. As carriers in liquid membranes, cose, tetracycline, several heavy metal cations,
cyclodextrins allow the selective transportation of and even an enantioselective determination of
saccharides and to achieve good results in this R-clenbuterol was reported. In this latter case, a
difcult separation. Organic structural and detection limit of 2.99 10exp-7 mol/L was
geometrical isomer enrichments were obtained reached. The introduction of cyclodextrins into
by using water-based liquid membranes. Liquid- the polysulfone membrane structure greatly
liquid extraction was also used to separate racemic inuenced the rejection of humic acid and natural
mixtures because the cyclodextrin nanocavity organic matter from water. Again polysulfone/CD
allowed the formation of diastereomeric com- membrane found uses in endocrine disruptors
plexes. Enantiomeric discrimination of native aro- removal from water. Alternatively, acid dyes can
matic amino acids was also performed. Chiral be removed from wastewater by using a chitosan
enrichment of racemic chlorthalidone and pro- membrane modied with cyclodextrin. Ceramic
pranolol was achieved using bulk liquid mem- membranes impregnated with cyclodextrin poly-
brane containing cyclodextrin as a mobile mers are effective in facilitating polycyclic
enantioselective carrier. Because of their different aromatic hydrocarbons removal. In addition,
ability to complex isomers, immobilized cyclo- the polyacrylonitrile membrane with
dextrins also largely inuenced permeation in triacetyl-b-cyclodextrin is able to reduce the
dialysis membranes. In this way, orto, para, and amount of pesticides in citrus essential oil.
meta isomers of aromatic compounds were suc- Both hydrophobic b-cyclodextrin polymers in tri-
cessfully achieved as well as the enantiomeric acetate membrane and polypyrrole polymer
separation of aromatic amino acids. Pervaporation membrane doped with sulfated a-cyclodextrin
processes were also greatly inuenced by the facilitate the transport of heavy metal cations
addition of cyclodextrins or cyclodextrin deriva- from aqueous solutions. The high versatility of
tives prevalently in PVA membranes. Thus, membrane processes allows the preparation of
Cyclodextrins 507
composite Naon membranes with sulfated- lipophilic cavity surrounded by a hydrophilic rim
b-cyclodextrin. This induces a remarkable being the OH groups located on the exterior edge.
increase on proton conductivity with a very slight Primary hydroxyl groups are all located on the
increase in methanol permeability. Even better narrow edge of the truncated cone being the sec-
results were achieved by using sulfonated PEEK ondary positioned on the larger edge. The reactiv-
membranes. Anion exchange membrane perfor- ity of the hydroxyl groups differs enough to be
mances in electrodialysis processes are inuenced used in selective modication of the native cyclo- C
by the introduction of cyclodextrins in the poly- dextrins. Cyclodextrins were rst discovered by
mer network. Due to the changes in the hydrophi- French chemist A. Villiers in 1891 as a by-product
licity of the environment, transport of less in the production of dextrins. A few years later,
hydrated anions decreased, while transport of Austrian chemist F. Schardinger isolated the strain
strongly hydrated anions increased by enhancing on bacteria (Bacillus macerans) responsible for
the content of cyclodextrin. Minor application of the synthesis of cyclodextrins. However, it was
cyclodextrins was also found in UF, NF, RO, only through the work of Freudenberg in 1935
temperature-controlled membrane processes, and that the cyclic nature of cyclodextrin was discov-
a new drug delivery system. ered. The toroidal structure of cyclodextrins
allows them to form inclusion compounds with a
great set of molecules with suitable polarity and
References size. The polarity of the inner nanocavity is
believed to be similar to an aqueous ethanolic
Kozlowski CA, Sliwa W (2008) The use of membranes solution. The inner cavities have a diameter in
with cyclodextrin units in separation processes: recent
the range of 0.470.53, 0.600.65, and
advances. Carbohydr Polym 74:19
0.750.88 nm, respectively. On the other hand,
the height (0.79 nm) is the same for all three native
types. Generally speaking, the internal diameter of
the cavity mostly accounts for the encapsulation
Cyclodextrins of the guest molecule. The smaller a-cyclodextrin
complex compounds with linear aliphatic chains,
Francesco Trotta b-cyclodextrin includes well aromatic structures,
Department of Chemistry, University of Torino, and g-cyclodextrin accommodates larger mole-
Torino, Italy cules. All native cyclodextrins are quite safe and
can be orally administered without toxic effects
because they are not absorbed by the gastrointes-
Cyclodextrins (Astray et al. 2009; Loftsson and tinal tract. On the other hand, both parent a- and
Duchene 2007; Martin Del Valle 2004; Szejtli b-cyclodextrins are not suitable for parenteral
1998; Szente and Szeman 2013) (Schardinger administration. a- and g-cyclodextrins have
dextrins, cycloamylose, cyclomaltose) (Fig. 1) good water solubility, but b-cyclodextrins have
are cyclic nonreducing oligosaccharides obtained limited solubility in water, i.e., 1.85 wt.% at
through the enzymatic degradation (cyclomal- room temperature. In any case, the water solubil-
todextrin glucanotransferase) of starch. They are ity of the parent cyclodextrins is much lower than
made up of six, seven, or eight glucose units corresponding linear oligosaccharides. Solubility
linked by 14 a glucoside bonds and have a increases to a large extent by increasing the
typical truncated cone-like shape. They are temperature and/or pH. Inclusion compounds
named a-, b-, and g-cyclodextrin, respectively. can be produced following several techniques
Larger cyclodextrins, i.e., d and e, are also such as grinding, freeze-drying, spray-drying,
known, but they are much more expensive, have kneading, coprecipitation, melting, evaporation,
a twisted structure, are less stable, and are of etc. Hydrophobic molecules are incorporated
limited interest. The position of the atoms form a into the cavity of cyclodextrins by displacing
508 Cyclodextrins
Cyclodextrins, Fig. 1 Molecular structure of native cyclodextrins
water. This effectively encapsulates the molecule reported in the literature, only few have
of interest within the cyclodextrin, rendering the found industrial application due to their superior
molecule water soluble. When the water-soluble performances and safe prole: hydroxypropyl
complex is diluted in a much larger volume of b-cyclodextrin and sulfobutyl ether
aqueous solvent, the process is reversed, thereby b-cyclodextrin. Random methylated b-cyclo-
releasing the molecule of interest into the solution. dextrin is also of interest but it cannot be used in
The encapsulation of the guest molecules inside drug delivery due to its toxic effects. In order to
the cyclodextrin cavity leads to many changes in increase the stability constant of the inclusion
their physical-chemical properties. In particular, complexes, dimer and trimer of cyclo-
cyclodextrins can: dextrin were synthesized as well as linear and
cross-linked polymers. The latter can act as
1. Increase the solubility of poor water-soluble nanosponges, nanocontainers, or nanocarriers.
molecules and drugs
2. Transform liquid in solid powder
3. Host volatile molecules in solid form References
4. Mask smell and taste
5. Protect substances by light and oxidation Astray G, Gonzalez-Barreiro G, Mejuto JC, Rial-Otero R,
Simal-Gandara J (2009) A review on the use of cyclo-
6. Act as enzyme-like catalysts
dextrin in foods. Food Hydrocoll 23:16311640
7. Act as inverse phase catalysts Loftsson T, Duchene D (2007) Cyclodextrins and their
8. Change membrane permeabilities pharmaceutical applications. Int J Pharm 329:111
Martin Del Valle EM (2004) Cyclodextrins and their uses:
a review. Process Biochem 39:10331046
The abovementioned properties allow cyclo-
Szejtli J (1998) Introduction and general overview of
dextrin to nd many applications in cosmetics, cyclodextrin chemistry. Chem Rev 98:17431753
separation, laundry, catalysis, drug delivery, ana- Szente L, Szeman J (2013) Cyclodextrins in analytical
lytical chemistry, food additive, agriculture, and chemistry: hostguest type molecular recognition.
Anal Chem 85:80248030
chemical industries. Although several
hundred cyclodextrin derivatives were already
D
Dead-End Filtration with constant ux ensures that the permeate ux

through the membrane remains constant. This
Xianhui Li and Jianxin Li ltration can be achieved by positive displace-
State Key Laboratory of Separation Membranes ment pump. As the cake buildup increases with
and Membrane Processes, School of Materials time, the pressure drop would increase so as to
Science and Engineering, Tianjin Polytechnic maintain constant ux. The constant ux strategy
University, Tianjin, China is usually applied in the industrial applications.
Since the produced volume is a main operating
goal. In another dead-end ltration mode of con-
stant pressure, the permeate ux decreases as the
The dead-end ltration is the one where the ow cake build-up with the time. The constant pressure
of water is perpendicular to the membrane surface strategy is commonly chosen in the laboratory
(Fig. 1a). And the water is pushed through the scale installations, because this is easy to imple-
membrane by pressure. All the water that is intro- ment on a small scale.
duced in the dead-end cell passes through the However, the dead-end ltration performance
membrane. In other words, there is no retentate. is often limited by fouling phenomenon owning to
In dead-end ltration the retained particles build its ltration characteristics. Some strategies are
up with time on the membrane surface or within employed to control membrane fouling, such as
the membrane. In either case, the building particle module design, pretreatment, or hydrophilic mod-
would result in the increase of layer thickness and ication of membrane materials. Four distinct
cause the permeate ux to decline (Fig. 1b). Thus, models based on Darcy equation can be used to
the dead-end ltration requires the stopping of describe the dead-end fouling behavior (Table 1).
ltration to clean or replace the membrane. There- The relationships described were the complete
fore, this type of ltration is also called batch blocking model, which assumes that each particle
ltration. that deposits on the membrane will seal a mem-
There are two types of ltration which can be brane pore; the intermediate blocking model,
employed in a dead-end ltration unit: constant which assumes that only a portion of the deposited
ux and constant pressure. The dead-end ltration particles will seal a membrane pore; the standard

DOI 10.1007/978-3-662-44324-8
510 Dead-End Filtration
a Feed blocking model, which assumes that the pore vol-

ume will decrease proportionally to the ltrate
volume by deposition of solids on the pore
walls; and the cake ltration model, which
Cake layer assumes that the cake resistance is proportional
to the ltrate volume by deposition of solids on
Membrane
the membrane surface.
The mathematical expressions for these laws
are summarized in Table 1. For a constant pressure
Permeate
ltration mode, they have been reformulated in a
b Flux common frame of power-law relationships
Layer (Hermia 1982):
Layer thickness (m)

thickness
Flux (L/m2h)
n
d2 t dt
2
k (1)
dV dV
Table 1 also presents the physical basis for each

fouling mechanism and the corresponding
n-value. Each of these mechanisms could be
Time alone or in combinations, to explain experimental
observations. Overall, the membrane fouling con-
Dead-End Filtration, Fig. 1 Schematic diagram of dead-
end ltration (a) and representation of ux rates and layer trol is a key problem for the large-scale
thickness (b) application.
Dead-End Filtration, Table 1 Fouling mechanisms and their corresponding physical basis (Grenier et al. 2008)
Corresponding linear
Fouling mechanism n form Physical concept
Cake ltration 0 dt
dV Q1 f V Surface deposit
Intermediate 1 dt
dV Q1 f t Pore blocking and surface
blocking deposit
Standard blocking 1.5 dV 1=2 Pore constriction

dt Q1=2 f V
Complete blocking 2 dV
dt Q f V Pore blocking
Dealcoholization by Pervaporation 511
References consumed in distillation. Thus, dealcoholization

by PV is efcient because of energy conservation
Grenier A, Meireles M, Aimar P, Carvin P (2008) and separation of an azeotropic mixture, which
Analysing ux decline in dead-end ltration. Chem
cannot be separated by distillation. In PV, a liquid
Eng Res Des 86:12811293
Hermia J (1982) Constant pressure blocking ltration mixture is supplied to a membrane on one side.
laws application to power-law non-Newtonian uids. The permeated components are evaporated from
Trans Inst Chem Eng 60:183187 the other side (Fig. 1). Thus, the driving force of
this process is a gradient of partial pressure or
chemical potential located across a membrane. D
To enhance the ux or the total amount of the
Dealcoholization by Pervaporation product, the permeated side is always maintained
at a pressure below the saturated pressure of a
Shota Ando, Shuichi Sato and Kazukiyo Nagai given component, either under a vacuum or by
Department of Applied Chemistry, Meiji sweeping the vapor using a carrier gas. The vapor
University, Tama-ku, Kawasaki, Japan at the permeated side is then condensed and recov-
ered as liquid (Nagai 2010). The features of this
method are: (1) increased separation performance
Dealcoholization by pervaporation (PV) is one of from solution, diffusion, and vaporization; (2) the
the most innovative processes for the selective driving force of permeation does not depend on
separation and condensation of ethanol from a the pressure and is applicable to the organic liquid
dilute aqueous solution of ethanol obtained from solution with high osmotic pressure; (3) PV does
biomass fermentation (about 10 wt%) using PV not need pressurization; (4) this method can be
method. The energy for bioethanol that is pro- applied to separate the mixture solution, which
duced using distillation requires approximately cannot be done by distillation, for example, the
1.5 times than that obtained from ethanol fuel. separation of the thermal denaturation/thermally
Among them, about 70 % is estimated to be decomposable mixture, azeotropic mixture, near-
Dealcoholization by
Pervaporation,
Fig. 1 Dealcoholyzation
by a pervaporation process
512 Dearomatization by Membrane Operations
boiling point mixture, and structural/optical iso- blend, and three dimensional structures by cross-
mers can be concentrated; (5) the reaction can be link. In each case, there is a limit to improvement
promoted because of the movement of the equi- in separation performance by previous method
librium reaction; and (6) the method does not because there is a discrepancy between afnity
pressurize, and no consolidation of the membrane and swelling.
occurs. The term PV is a compound word of
Permeation and Evaporation. Moreover, it has
been originated from the article of Kober (1917). References
There are two kinds of method for alcohol con-
centration by PV. The alcohol-selective method Kober PA (1917) Pervaporation, perstillation and
percrystallization. J Am Chem Soc 39:944948
removes alcohol from the dilute alcohol aqueous
Nagai K (2010) Fundamentals and perspectives for
solution of concentrated alcohol, whereas pervaporation. In: Enrico D, Lidietta G (eds) Compre-
the water-selective method removes water at hensive membrane science and engineering, 1st edn.
a high-concentrated alcohol solution. The mate- Elsevier, Amsterdam, pp 243271
rials differ according to their use. The alcohol-
selective method uses silicone gum and poly
(1-trimethylsilyl-1-propyne). By contrast, the
water-selective method uses cellulose acetate Dearomatization by Membrane
and poly(phenylene oxide), which yield good Operations
results. Generally, the alcohol-selective method
is called dealcoholization by PV. Researches on Aromatic and Nonaromatic Separation by
alcohol-selective permeable membranes are fewer Membrane Operations
than those on water-selective permeable mem-
branes. In addition, membranes with performance
that can be set in bench-scale plants have not been
developed. Thus, the developments of alcohol- Decatungstate, Catalytic Membrane
selective permeable membranes are delayed com- Containing
pared with that of water-selective membranes.
The cause is due to the larger molecular size of Enrica Fontananova
alcohol molecules, which are permeated preferen- Institute on Membrane Technology, National
tially, compared with that of water molecules. Research Council of Italy, ITM-CNR, Rende,
When alcohol from alcohol aqueous solution is Italy
permeated preferentially, separation on the diffu-
sion process cannot be expected in PV method Catalytic membrane containing decatungstate is a
based on a solution-diffusion mechanism. Thus, polymeric catalytic membrane applied on a lab
separation can be performed using the difference scale in liquid phase photooxidation reactions
in the only solution process of both components of (Drioli and Fontananova 2010).
the membrane. However, the membrane is exces- The decatungstate (W10O324) is a photo-
sively swelled by very high afnity to alcohol catalytic polyoxometalate (polyanionic metal
because it is in direct contact with the mixture oxide cluster of early transition metals) with
solution to be separated in PV method. Thus, remarkable properties for application in oxidation
permeability increases. However, the small catalysis for ne chemistry and wastewater
molecular size of water enters the membrane by treatments.
the plasticizing effect because of membrane In particular, decatungstate exhibits interesting
swelling, which signicantly reduces the selectiv- properties for the photocatalytic detoxication of
ity to alcohol. To solve this problem, the balance wastewater since its absorption spectrum (lmax
of hydrophobicity and afnity of the membrane is = 324 nm) partially overlaps the UV solar emis-
examined by graft, copolymers, multiblock, sion spectrum, opening the potential route for an
Decatungstate, Catalytic Membrane Containing 513
Decatungstate, Catalytic F F
Membrane Containing,
Fig. 1 Chemical formula F F
O O
of the polymers used to
prepare the catalytic
membranes entrapping
decatungstate CH2 CF2 O F F F
n CF3
n m
PVDF Hyflon
D
CH3 CH3 CH3 CH3 CH3

Si H2C H2C Si O Si O Si O Si CH2 CH2 Si
CH3 H CH3
CH3 CH3
j
PDMS
environmentally benign solar-photoassisted appli- These polymers dont absorb in the region of
cation (Texier et al. 1999). interest of the catalyst; moreover, they are charac-
However, decatungstate has also some relevant terized by a high chemical, thermal, and UV
limitations: low quantum yield, small surface stability.
area, poor selectivity, and limited stability at pH In order to improve the catalyst/polymer inter-
higher than 2.5 (Mylonas et al. 1996). actions and to avoid catalyst leaching out from
Membrane technology can offer the possibility the membrane, lipophilic and insoluble-in-water
to overcome these limitations by the multi- derivatives of the decatungstate were employed:
turnover recycling associated to heterogeneous the tetrabutylammonium salt ((n-C4H9N)4
supports, the selectivity tuning as a function of W10O32) and a uorous-tagged decatungstate
the substrate afnity toward the membrane phase, ([CF3(CF2)7(CH2)3]3CH3N)4W10O32) (Bonchio
and the effect of the polymeric microenvironment et al. 2003; Carraro et al. 2006).
on catalyst stability and activity. A membrane-induced structure-reactivity
In this respect, the design of alternative hetero- trend, which may be exploited to achieve selective
geneous photooxygenation systems able to employ processes, was observed in polymeric catalytic
visible light, oxygen, mild temperatures, and solvent membranes prepared embedding decatungstate
with a low environmental impact (water or neat within membranes made of PVDF (PVDF-
reaction) was investigated by the immobilization of W10), PDMS (PDMS-W10), or Hyon (Hyon-
the decatungstate in polymeric membranes W10).
(Bonchio et al. 2003; Carraro et al. 2006; The polymeric catalytic membranes were pre-
Fontananova et al. 2006; Drioli et al. 2008). pared by solvent evaporation (to obtain dense
The successful heterogenization was membrane) or nonsolvent-induced phase separa-
guaranteed by a proper choice of both the catalyst tion (to obtain porous membranes). The
functionalization and the polymer material. decatungstate was solubilized or dispersed in the
Considering the interest toward the photooxida- casting solution and heterogenized in the polymer
tion reactions of organic substrates, principally in matrix during the membrane formation process.
aqueous phase, hydrophobic polymer materials One of the crucial aspects during catalyst
were selected for membrane preparation: heterogenization is the maintaining of its struc-
polyvinylidene uoride (PVDF), Hyon (a peruoro tural integrity. Solid-state characterization tech-
amorphous polymer), and polydimethylsiloxane niques, including Fourier transform infrared
(PDMS) (Fig. 1). (FT-IR) analysis and UVvis spectroscopy in
514 Decatungstate, Catalytic Membrane Containing
diffuse reectance, conrmed that the catalyst precipitate toward the down surface (Carraro
structure and spectroscopic properties were pre- et al. 2006).
served within the membranes. On the contrary, in the case of the uorous-
The PVDF-W10 and PDMS-W10 photo- tagged decatungstate, SEM images of the mem-
catalytic systems were used for the selective brane surface and cross section highlighted a
photooxidation of water-soluble alcohols (Bonchio homogeneous distribution of the catalyst domains
et al. 2003). A membrane-induced discrimination which appear as spherical particles with
of the substrate results from the oxidations of a uniform size.
series of alcohols with different polarity, through The dimensions of these clusters and, as a
comparison with the homogeneous reactions. consequence, the surface area and catalytic activ-
The PVDF-W10 membranes were also suc- ity of the decatungstate were modulated acting on
cessfully employed in the aerobic photooxidation the membrane preparation conditions and catalyst
of phenol in water carried out in a membrane loading.
reactor operating with ow through the membrane The catalytic Hyon-based membranes were
(Drioli et al. 2008). tested in batch solvent-free oxygenation of ben-
Polymeric catalytic membranes were also pre- zylic CH bonds of the ethylbenzene at 25 C
pared by immobilizing sodium decatungstate under O2 atm.
(Na4W10O32) (Lopez et al. 2006; Lopez The key observation was provided by an
et al. 2007) and phosphotungstic acid increase of the selectivity toward the alcohol
(H3PW12O40) (Fontananova et al. 2006), on the (product of interest) and a higher turnover number
surface of PVDF membranes modied by Ar/NH3 (TON) when the catalyst was heterogenized in
plasma discharges. Hyon membrane in comparison with homoge-
The groups grafted on the surface were able to neous catalysts or the catalyst heterogenized
bind the catalysts dissolved in aqueous solution, in PVDF.
forming charge-transfer complexes.
Surface diagnostic techniques such as X-ray
photoelectron spectroscopy (XPS), contact angle
measurements (CA), and RX maps were used to References
attest the surface modication.
The catalytic membranes showed superior per- Bonchio M, Carraro M, Scorrano G, Fontananova E,
formances (higher reaction rate) compared to the Drioli E (2003) Heterogeneous photooxidation of
alcohols in water by Photocatalytic Membranes
corresponding homogeneous catalysts, in the pho- Incorporating Decatungstate. Adv Synth Catal
tooxidation of phenol. 345:11191126
The decatungstate in the form of tetrabuty- Carraro M, Gardan M, Scorrano G, Drioli E,
lammonium salt was also heterogenized in Fontananova E, Bonchio M (2006) Solvent-free, het-
erogeneous photooxygenation of hydrocarbons by
Hyon membranes by dispersion in the poly- Hyon membranes embedding a uorous-tagged
meric solution obtained using Galden HT55 decatungstate: the importance of being uorous.
(a mixture of peruoro-hydrocarbons) as sol- Chem Commun 43:45334535
vent and successive solvent evaporation. How- Drioli E, Fontananova E (2010) Catalytic membranes
embedding selective catalysts: preparation and applica-
ever, the insolubility of catalyst in this solvent tions. In: Heterogenized homogeneous catalysts for
and the low afnity between the catalyst and ne chemicals production: materials and processes.
the polymeric matrix induced the formation of Springer, Berlin
irregular catalyst aggregates, not well dis- Drioli E, Fontananova E, Bonchio M, Carraro M,
Gardan M, Scorrano G (2008) Catalytic membranes
persed in the polymeric matrix, which tend to and membrane reactors: an integrated approach to
Dechlorination Process 515
catalytic process with a high efciency and a low envi- refrigerant, agricultural fumigant, and solvent
ronmental impact. Chin J Catal 29(11):1152 for metal degreasing and production of semicon-
Fontananova E, Donato L, Drioli E, Lopez L, Favia P,
dAgostino R (2006) Heterogenization of ductors in industrial processes for several decades
polyoxometalates on the surface of plasma modied (Kyunghoon and Woojin 2009). Chlorination has
polymeric membranes. Chem Mater 18:15611568 been used widely to disinfect wastewater prior to
Lopez LC, Buonomenna MG, Fontananova E, discharge. For these reasons, the chlorinated com-
Iacoviello G, Drioli E, dAgostino R, Favia P (2006)
New generation of catalytic PVDF membranes: cou- pounds and residual chlorine after disinfection are
pling plasma treatments with chemical immobilization both contained in wastewater. However, the resid-
of W-based catalysts. Adv FuncT Mater 16:14171424 ual chlorine is toxic to many kinds of aquatic life, D
Lopez LC, Buonomenna MG, Fontananova E, Drioli E, and the residual chlorine may also react with
Favia P, dAgostino R (2007) Heterogenization of tung-
sten catalysts on plasma modied membranes. Plasma organic materials in the water forming chlorinated
Processes Polym 4:326333 compounds, which have attracted an attention due
Mylonas A, Roussis V, Papaconstantinou E (1996) to their carcinogenic and mutagenic characteris-
Photocatalytic degradation of phenol and p-cresol by tics. In membrane process, many kinds of mem-
Polyoxotungstates Mechanistic implications. Polyhe-
dron 15:32113217 branes are also sensitive to residual chlorine,
Texier I, Giannotti C, Malato S, Richter C, Delaire J (1999) such as reverse osmosis (RO), nanoltration
Solar photodegradation of pesticides in water by (NF), and forward osmosis (FO). Consequently,
sodium decatungstate. Catal Tod 54:297307 dechlorination process is needed to remove the
residual chlorine and degrade the chlorinated
compounds.
The chlorinated compounds can be degraded
by microbial and physicalchemical processes
Decentralized Wastewater such as pump-and-treat, air sparging, and zero-
Treatment System valent iron passive barriers (Ma et al. 2003). The
residual chlorine can be removed by absorption,
Membrane Bioreactor for Decentralized Waste- chemical, and ultraviolet irradiation methods.
water Treatment: A Case Study
In membrane process, the feed must be
dechlorinated to remove the residual chlorine
and prevent the serious degradation of membrane
chemical structure. Absorption and chemical
Dechlorination Process methods are quite commonly used to dechlorinate
in membrane process. The materials that have
Zhi Wang been proposed for dechlorination are granulated
Chemical Engineering Research Center, School activated carbon (GAC), hydrogen peroxide
of Chemical Engineering and Technology, Tianjin (H2O2), ammonia (NH3), sodium thiosulfate
University, Tianjin, China
(Na2S2O3), and the sulfur(IV) species: sulfur
dioxide (SO2), sodium bisulfate (NaHSO3),
sodium sulte (Na2SO3), and sodium
Dechlorination process is the process to remove
metabisulte (Na2S2O5) (George and Lynn
the chlorine from chlorinated compounds or water 1984). Some of the important reactions of sulte
that has been chlorinated. with residual chlorine species are:
Dechlorination process has been widely used
in water treatment, especially in the eld of waste-
water treatment and membrane process. Chlori- SO3 2 OC1 ! SO4 2 C1 (1)
nated compounds have been widely used as
516 Definition of Various Membrane Processes
SO3 2 NH2 Cl H2 O
Definition of Various Membrane
! NH4 2 SO4 2 C1 (2)
Processes
SO3 2 RNHCl H2 O Lidietta Giorno1, Heiner Strathmann5 and
! RNH2 SO4 2 C1 H (3) Enrico Drioli1,2,3,4
1
The residual chlorine in membrane process Italy
2
should be controlled in a certain range, which should Department of Environmental and Chemical
rely on the dechlorination process. At present, few Engineering, University of Calabria, Rende,
options exist for reliable close-to-zero levels of resid- Italy
3
ual chlorine in the water, and the residual chlorine is WCU Energy Engineering Department,
needed to prevent the membrane from fouling as Hanyang University, Seoul, S. Korea
4
well (Kim et al. 2009). The rate of chlorine attack Center of Excellence in Desalination
to the membrane depends on various feedwater char- Technology, King Abdulaziz University, Jeddah,
acteristics, like pH and temperature, and different Saudi Arabia
5
membranes also have not alike chlorine tolerance Universitt Stuttgart, Institute of Chemical
therefore, the different kinds of feedwater and mem- Process Engineering Stuttgart, Baden-
branes decide the residual chlorine concentration Wurttemberg, Germany
needed should be different. In addition, the excess
of oxidizing agents, like chlorine, not only oxidizes
microbiology but also oxidizes other organic mole- Denition of Various Membrane Processes can be
cules into organic pieces and is converted into assim- grouped according to the applied driving forces
ilable organic carbon (AOC), which is good food for into: (1) hydrostatic pressure-driven processes
the bacteria (Redondo and Lomax 2001). Conse- such as reverse osmosis, nano-, ultra-, and
quently, the residual chlorine should be monitored microltration, membrane emulsication or gas
and adjusted in dechlorination process. separation, and pervaporation; (2) concentration
gradient or chemical potential-driven processes
such as dialysis, Donnan dialysis, pervaporation,
References and membrane contactors, such as membrane-
based solvent extraction, membrane scrubbers
George RH, Lynn KC (1984) Dechlorination of wasterwater and strippers, and osmotic distillation; (3) electri-
and cooling water. Environ Sci Technol 18(2):4855 cal potential-driven processes such as electrodial-
Kim D, Jung S, Sohn JB (2009) Application for controlling
biofouling of SWRO membranes an overview. Desa-
ysis; and (4) temperature difference-driven
lination 238:4352 membrane processes such as membrane
Kyunghoon C, Woojin L (2009) Reductive dechlorination of distillation.
carbon tetrachloride in acidic soil manipulated with iron The molecular mixture which will be separated
(II) and bisulde ion. Desalination 172:623630
Ma X, Novak PJ, Clapp WL, Semmens MJ, Hozalski RM
is referred to as feed, the mixture containing the
(2003) Evaluation of polyethylene hollow-ber mem- components retained by the membrane is called
branes for hydrogen delivery to support reductive dechlo- the retentate, and the mixture composed of the
rination in a soil column. Water Res 37:29052918 components that have permeated the membrane
Redondo JA, Lomax I (2001) Y2K generation FILMTEC RO
membranes combined with new pretreatment techniques
is referred to as permeate (or ltrate in micro- and
to treat raw water with high fouling potential: summary of ultraltration). Table 1 lists the basic properties of
experience. Desalination 136:287306 membrane operations.
Definition of Various Membrane Processes 517
Definition of Various Membrane Processes, Table 1 Basic concepts of membrane operations

Mode of
Process Concept Driving force transport Species passed Species retained
Microltration Feed Retentate Pressure Size Solvent (water) Suspended
(MF) difference exclusion, and dissolved solids, ne
Solvent 100500 kPa convection solutes particulars,
Porous MF membrane
some colloids
Ultraltration Feed Retentate Pressure Size Solvent (water) Macrosolutes D

(UF) difference exclusion, and and colloids
Solvent 100800 kPa convection Low-molecular-
UF membranes weight solutes
(<1,000 Da)
Nanoltration Feed Retentate Pressure Size Solvent (water), Molecular-
(NF) difference exclusion, low- molecular- weight
Solvent 0.33 MPa solution weight solutes, compounds
NF membranes diffusion, monovalent >200 Da
Donnan ions multivalent ions
exclusion
Reverse Feed Retentate Pressure Solution Solvent (water) Dissolved and
osmosis (RO) difference diffusion suspended solids
Solvent 110 MPa
RO membranes
Gas separation Feed Pressure Molecular Gas molecules Gas molecules

(GS) difference sieving, having having high-
Permeate 0.110 MPa solution low-molecular- molecular-
Dense/Porous
membrane diffusion weight or high weight or low
solubility solubility
diffusivity diffusivity
Pervaporation Feed Retentate Chemical Solution High permeable Less permeable
Liquid
(PV) potential or diffusion solute or solute or
Vapour
Permeate concentration solvents solvents
Dense membrane difference
Electrodialysis Concentrate Product

Electrical Donnan Solutes (ions) Nonionic and
+ + +
(ED) +
-
- potential exclusion small quantity macromolecular
- -
+ -
difference of solvent species
Feed Ion exchange membrane 12 V/cell pair
Dialysis (D) Feed Purified stream Concentration Diffusion Solute (ions and Dissolved and
gradient low MW suspended solids
Impurities Dialysate feed organics) small with MW
Dialysis membrane solvent quantity >1,000 Da
Membrane Retentate Chemical Evaporation, Compounds Compounds

Feed
contactors potential, diffusion soluble in the non-soluble in
(MC) concentration extraction the extraction
Permeate Sweep
difference, solvent; solvent;
Porous membrane
temperature volatiles nonvolatiles
gradient
(continued)
518 Degenotoxification
Definition of Various Membrane Processes, Table 1 (continued)

Mode of
Process Concept Driving force transport Species passed Species retained
Membrane- Feed
Retentate Chemical Diffusion Compounds Compounds
based solvent potential, partition soluble in the non-soluble in
extraction Permeate Sweep concentration extraction the extraction
(MBSX) Porous membrane
gradient solvent solvent
Membrane Warm feed Liquid Warm concentrate Vapor pressure Vapor Water, solvent Salts,
distillation vapour
gradient, transport nonvolatiles
Liquid
(MD) Cool Distillate Cool stream
temperature
Porous membrane
difference
Osmotic Feed
Retentate Vapor pressure Vapor Water, solvent Salts,
distillation gradient, transport nonvolatiles
(OD) hypertonic salt
Concentrated
soln
Porous membrane solution
Forward Feed
Retentate Osmotic Diffusion Water, solvent Molecules, salts,
osmosis (FO) pressure ions
Draw soln
Dense membrane
Supported Product A Concentration Diffusion Ions, low MW Ions, less

liquid Feed gradient organics permeable
membranes Product B organics
(SLM) Supported liquid membrane
Membrane Feed Reagents Various Various Permeable Non-permeable

reactors (MR) product/ reagents
Product Product controlled supply
Reactive membrane
of reagent
References
Dehumidification of Atmospheric Air
Cheryan M, Ultraltration and microltration handbook, by Membrane Technology
Technomic Publishing, Lancaster, PA, 1998 R. W.
Baker, Membrane technology and applications, Wiley, Wei Liu
2012, ISBN: 978-0-470-74372-0
Strathmann H, Giorno L, Drioli E (2006) An Introduction Energy and Environment Technology Directorate,
to Membrane Science and Technology, CNR Publisher, Pacic Northwest National Laboratory, Richland,
Roma, ISBN 88-8080-063-9 WA, USA
Degenotoxification
Synonyms
Genotoxin Removal
Membrane dehumidication of atmospheric air
Dehumidication of atmospheric air represents

Degree of Mixedness one large application opportunity for advanced
membrane technologies. Air is used to supply
Mixing Index oxygen for a number of industrial processes.
Dehumidification of Atmospheric Air by Membrane Technology 519
Dehumidification of Atmospheric Air by Membrane Technology, Table 1 Comparison of different air dehumid-
ification technologies
Technology Working principle Feature
Membrane Removal of moisture through A steady-state, continuous process
dehydration a selective membrane with A selective membrane is needed
minimal changes to air A driving force needs to be provided for moisture to transport across the
temperature and pressure membrane
Water Cooling of humid air below Cooling of whole air stream is necessary
condensation dew point for moisture to Condensation adds latent cooling duty D
condense as liquid water or ice
Adsorption Capturing moisture on a solid Solid materials are durable and do not generate any environmental
or solid adsorbent emission
desiccating Heat of adsorption is released during capture
Saturated adsorbent has to be regenerated by increasing temperature
and/or reducing partial pressure of water vapor, and both sensible and
latent heat need to be supplied for desorption
Regenerated adsorbent needs to be cooled to capture temperature and heat
of adsorption is released during capture
Absorption Capturing moisture in liquid Fluid can be readily moved around
or liquid uid Same heating and cooling cycles as adsorption are needed for
desiccating regeneration of saturated uid
Moisture often needs to be removed when its processes. A variety of solid adsorbents (e.g., sil-
presence presents a problem to downstream pro- ica gel, polymers, and zeolite) and liquid sorbents
cesses. For example, in oxygen production via (e.g., LiCl, CaCl2, and ethylene glycol) are avail-
cryogenic air separation, feed air needs to be able for capturing moisture at various tempera-
adequately dehydrated to prevent formation of tures and humidity levels. However, the
ice in the cooling equipment. Air dehumidica- desiccating material needs to be periodically
tion and conditioning in hot and humid climate are regenerated by heating and cooling. The saturated
required to maintain comfort in buildings. desiccant is typically heated up to a much higher
There are several commercial technologies for temperature than the capture to release the water
air dehydration, such as water condensation, solid as water vapor. Thermal energy has to be provided
adsorption, and liquid absorption. Their working for the sensible heating and for compensating heat
processes and salient features are compared to of desorption. After regeneration, the desiccant
membrane dehydration in Table 1. needs to be cooled down to the working temper-
As discussed above, the membrane dehydra- ature. During the capture process, heat of the
tion is a steady-state process that works under feed adsorption (or absorption) is released to the dried
air pressure and temperature. If humid air is air, which adds the sensible cooling duty of the air.
cooled down or pressurized, moisture can be con- Because of process complexity and signicant
densed out of the air as water. Cryogenic cooling thermal energy requirement, the adsorption or
of atmospheric air is often used in the industrial absorption method has found very limited appli-
processes and in current building air conditioners cations in buildings.
for air dehydration. The main disadvantages of Increasing energy demand in the building sec-
this method are the following: (i) the whole air tor worldwide and growing concerns of indoor air
stream has to be cooled down below water dew quality have driven the development of new
point and (ii) heat of condensation adds signicant energy-efcient air dehumidication and condi-
cooling duty. As a result, the energy efciency is tioning technologies. Membrane dehumidica-
low. Solid adsorption and liquid absorption are tion appears to be an attractive approach due to
widely used in the industrial gas and air drying the following considerations. Different from
520 Dehumidification of Atmospheric Air by Membrane Technology
Dehumidification of Atmospheric Air by Membrane Technology, Fig. 1 Process scheme of dehumidication

system
process industries, air in buildings does not need membrane has to be pumped out of the water
to get too dry, and high-degree dehumidication is condenser into environment, while majority of
not necessary. Membrane generally is effective for the permeated water vapor is condensed. Thus,
bulk separation. Membrane dehumidication compression ratio for the leaked air is much
does not generate any environmental pollution greater than the permeated moisture, and air leak-
and is a green process. Lastly, membrane dehu- age would drastically increase power consump-
midication has high thermodynamic energy tion of the vacuum pump. The previous study
efciency. suggests that a H2O/air separation factor above
One key challenge is the selective membrane. 200 be required for about 80 % dehumidication
Due to handling of large gas ows and low partial of 32 C and 90 % RH air (Xing et al. 2013).
pressures of water vapor in atmospheric air, the Other important factors about the membrane
membrane has to be highly permeable to mois- are durability and cost. The membrane has to be
ture. As expressed in the following equation, for a robust enough to be resistant to weathering, air
given dehydration rate (nw) and partial pressure contamination, and mechanical erosion in the
gradient of water vapor (Dpw), the required mem- duct. Also, the membrane surface has to be resis-
brane area decreases with increasing permeance in tant to bacteria attachment and growth in warm
a reverse rst order. and humid air environment. The membrane has to
be cost-effective to penetrate the commercial
nw building market.
SAm
Permeance Dpw Compared to extensive studies on membrane
dehydration of water/alcohol mixtures, scientic
Reducing usage of membrane area is not only publications on air dehumidication membrane
benecial to reduction of membrane cost but also development and tests are limited. Performance
enables a compact membrane module with mini- characteristics of a few membrane materials are
mal pressure drops for air to ow through. compared in Table 2. The membrane permeance is
The membrane also needs to have sufciently given in both SI unit (mol/m2/s/Pa) and gas per-
high selectivity toward moisture over air. Air is meation unit (GPU) for convenience of compari-
non-condensable gas. As shown in the process son. From an application point of view, the
ow of Fig. 1, any air leaked through the separation factor measured with actual humid air
Dehumidification of Atmospheric Air by Membrane Technology 521
Dehumidification of Atmospheric Air by Membrane Technology, Table 2 Air dehumidification performance

characteristics of a few membrane materials reported in the literature
Testing conditions Permeance
Permeate H2O/N2
T pressure Feed separation factor
Membrane material ( C) (103 Pa) gas mol/(m2 Pa s) GPU (selectivity)
NaA/thin metal sheet (Xing 32 0.31 Air+ 6.8 106 20,042 178
et al. 2013) H2O
Polysulfone hollow ber (Auvil 1.8 107
et al. 1993)
32 4.7 N2
+H2O
529 50
D
Hollow ber of stabilized liquid ~27 0.54 Air N/A N/A N/A
membrane (Bonne et al. 1990) +H2O
Ionic liquid [emim] [Tf2N] 31 0.69 Pure 2.1 107 635 (3,843)a
(Scovazzo 2010) gas
Stabilized liquid triethylene 15 N/A Pure 5.7 108 171256 (1,7002,500)a
glycol (Sijbesma et al. 2008) ~30 gas 8.5 108
Stabilized liquid polyethylene 22 N/A Pure ~5.0 108 149 (~2,000)a
glycol (M.W. = 400) (Sijbesma gas
et al. 2008)
Sulfonated poly(ether ether 30 <0.37 Pure ~5.0 107 1,500 (300,000)a
ketone) (SPEEK) (Sijbesma gas
et al. 2008)
~2.5 107

PEBAX 1074 (Sijbesma 30 <0.37 Pure 750 (104,167)a
et al. 2008) gas
Zeolite 3A lled triethylene 22 0.44 Air+ 3.8 109 12 25
glycol (Ito et al. 1998) ~25 H2O
a
The selectivity was calculated as a ratio of relative permeability of single gas
is a meaningful parameter to quantify the mem- increased by reducing effective zeolite pore dif-
brane selectivity and calculate energy consump- fusion length without affecting the selectivity.
tion. The separation factor was not reported in The NaA framework is stable in air and various
some of the literature. Instead, ratio of permeances gas environments and is expected to possess
measured with single gases was used to charac- excellent durability. It was reported that this
terize the membrane selectivity. It should be kind of membrane enables development of com-
alerted that such selectivity can differ drastically pact membrane dehumidiers with energy ef-
from the separation factor, because membrane ciency signicantly higher than other
structures and separation mechanisms under dehumidication means at a competitive cost
mixed gases could be substantially different (Xing et al. 2013).
from the single gas.
A NaA zeolite membrane supported on a thin
porous metal sheet shows the highest H2O
permeance and excellent H2O/air separation fac- References
tor. These performance attributes of the NaA
Auvil SR, Choe JS, Kellogg LJ (1993) Use of membrane
membrane can be explained by its microporous separation to dry gas streams containing water vapor.
structure and molecular-sieving function. High US Patent no. 5,259,869
selectivity is obtainable by eliminating or mini- Bonne U, Deetz DW, Lai JH, Odde DJ, Zook JD
mizing non-zeolite defects on the membrane. (1990) Membrane dehumidication. US Patent
no. 4,900,448
Since the selectivity results from selective H2O
Ito A, Sasaki H, Yonekura M (1998) Dehumidication of
adsorption at the entrance of zeolite pores, the air by zeolite-lled polymer membrane. Sekiyu
membrane permeance can be proportionally Gakkaishi 41:216221
522 Dehydration
Scovazzo P (2010) Testing and evaluation of room tem- pervaporation, water molecules are removed from
perature ionic liquid (RTIL) membranes for gas dehu- water-containing liquid as water vapor and thus, a
midication. J Membr Sci 355:717
Sijbesma H, Nymeijer K, van Marwijk R, Heijboer R, phase change of water occurs during the process.
Potreck J, Wessling M (2008) Flue gas dehydration Due to signicant heat of vaporization, the liquid
using polymer membranes. J Membr Sci 313:263276 feed temperature will be reduced without thermal
Xing R, Rao Y, TeGrotenhuis W, Caneld N, Zheng F, energy inputs. Continuous supply of thermal
Winiarski DW, Liu W (2013) Advanced thin zeolite/
metal at sheet membrane for energy efcient air dehu- energy is necessary to conduct pervaporation
midication and conditioning. Chem Eng Sci under a constant temperature. In the vapor-phase
104:596609 separation process, water vapor is separated out of
a water vapor-containing uid, and no phase
changes are involved. Vapor-phase membrane
separation is nearly an isothermal process.
The driving force for water molecules to move
Dehydration across the membrane is typically partial pressure
gradient of water vapor. For a given feed uid,
Wei Liu partial pressure of water vapor in the permeate
Energy and Environmental Technology side can be lowered by pulling vacuum and/or
Directorate, Pacic Northwest National introducing a sweep gas stream. Water transport
Laboratory, Richland, WA, USA across the membrane can also be driven by chem-
ical potential gradient of water. For example, the
forward osmosis process involves water transport
from water-containing liquid of a lower solute
Dehydration is a common terminology but often concentration to liquid of a higher solute concen-
has different meanings in different elds. In chem- tration, and a membrane gas/liquid contactor for
istry, dehydration is a chemical reaction process gas drying involves transport of water from water
about conversion of one molecule into another vapor-containing gases into water-absorbing liq-
one by removing H and O atoms as a water mol- uid. However, those membrane processes are
ecule, such as conversion of ethanol (C2H5OH) viewed as different technologies from membrane
into ethylene (C2H4). In physiology and medicine, dehydration.
dehydration means the excessive loss of body There are a variety of applications for mem-
water. Dehydration in food processing involves brane dehydration. Dehydration is necessary for
removal of water from various types of food for production of pure or anhydrous alcohols,
long-term preservation. because a water-alcohol mixture is often produced
In the membrane eld, dehydration generally by fermentation or catalytic reactions from feed-
refers to removal of water molecules from a stock of sugars, corn, cellulose, or syngas. Ethanol
water-containing uid or mixture, i.e., it is a fuel production represents one major application
physical process. Two common membrane of dehydration technologies, and its worldwide
dehydration processes are illustrated in Fig. 1. In production capacity reaches about 85 billion
a b T2 T1 Raffinate
T2 <T1 Liquid-phase
raffinate Water vapor-
Water- containing fluid
containing liquid
Membrane
T1
T1
Water vapor
Water vapor
Dehydration, Fig. 1 Common membrane dehydration processes. (a) Pervaporation. (b) Vapor-phase separation
Dehydration 523
Dehydration,
Fig. 2 Separation
mechanisms of dehydration
membrane. (a) Solute
diffusion in dense
membrane. (b) Molecular
sieving in microporous
membrane
liters/year in 2012. Successful development of dehydration is considered as an energy-efcient

cellulosic ethanol technologies is expected to and compact process.
lead more growth of ethanol fuels. Methanol is A number of membrane materials have been
another commodity chemical with a large market. studied, and some of them are commercially
Water is miscible with a number of solvents and developed for the dehydration application. They
chemicals, such as acetic acid, acetone, and tetra- can be classied into two groups (Fig. 2): dense
hydrofuran. Dehydration of these solvents is and microporous. The dense membrane provides
involved in a range of industrial chemical conver- selective absorption of water over other mole-
sion processes. The presence of water in industrial cules. The separation mechanism can be described
gases is ubiquitous. Air dehumidication and nat- by the solute diffusion model. Most polymeric
ural gas dehydration are examples of signicant membrane materials fall into this category,
industrial gas drying processes. which include polyvinyl alcohol (PVA), chitosan,
One salient feature for the membrane dehydra- alginate, polysulfone (PSF), polyimides, polyam-
tion process is molecular separation functions of ides, polyaniline, and polyelectrolyte membranes
the membrane. Removal of free water from slurry, such as cation-exchanged Naon. Separation in a
i.e., separation of water from particulates, is typ- microporous membrane is mostly achieved
ically referred as dewatering processes. The mem- through molecular sieving based on molecular
brane allows permeation of water molecules while sizes and/or selective adsorption of water at
blocking other molecules. Because of the molec- entrance of the membrane pore. Microporous zeo-
ular separation function, the membrane dehydra- lite, silica, and carbon are common membrane
tion provides signicant advantages over materials. Water molecule is often the smallest
conventional evaporation and gas drying technol- one in a mixture, and selective separation can be
ogies. The vapor-phase membrane separation is a realized by making the membrane pore size small
steady-state process operating under constant enough. However, the small pore size reduces
temperature and pressure, as compared to periodic molecule diffusion rates and may have a negative
switches between adsorption (or absorption) and impact on ux. Molecular sieving based on selec-
regeneration required for the solid (or liquid) des- tive water adsorption enables achievement of both
iccating process. For separation of a water- high ux and selectivity (Zhang and Liu 2011).
miscible liquid, the membrane dehydration is not Some of zeolite frameworks, such as NaA,
limited by thermodynamic gas/liquid equilibrium ZSM-5, and X-type and Y-type zeolite, have the
such as co-boiling that often occurs in alcohol- pore size larger than that of water molecule but
water distillation processes. The membrane can be very selective toward water adsorption
524 Dehydration
over other molecules. Permeation of the other 100 C). It should be noted that ux is a strong
molecules is blocked due to adsorption of water function of testing conditions. For a given mem-
molecules at the entrance of the zeolite pore, brane, water ux tends to increase with water
although the pore can be permeable to the other content in the feed and with the separation tem-
molecules in the absence of moisture. perature. Permeance should be a better parameter
The membrane dehydration performances are to characterize a membranes throughput. How-
characterized by ux (Jw), permeance (Pw), and ever, this parameter was seldom reported in the
separation factor (Sw/i), which can be calculated literature. The permeance is estimated to vary
from experimental measurements as follows: over a wide range from 1.0E-5 to 1.0E-9
mol/m2/s/Pa among different membranes.
Selection of a suitable membrane for a given
Dnw
Jw application depends on specic properties of the
Am Dt
water-containing uid. The material has to be
Jw chemically and physically stable under the sepa-
Pw ration conditions that the membrane structure can
p w, F p w, P
be maintained intact for a long time. For example,
NaA is a commonly studied zeolite membrane for
yw =yi P
aw=i dehydration but it is not stable in acidic solutions.
xw =xi F
For dehydration of acetic acids, the more stable
ZSM-5 and T-type zeolite materials are suggested.
where Dnw is the amount of water collected in the Pure zeolite membranes were considered difcult
permeate side of a membrane during a sampling to prepare. To overcome the shortcomings with
period of time Dt, Am is the membrane surface area pure polymeric and ceramic membranes,
exposed to the feed, pw,F is partial pressure of researchers proposed mixed matrix membranes
water in feed side, pw,P is partial pressure of (MMM) comprising a polymeric base into which
water in permeate side, yw and yi are the respective an inorganic material is dispersed and locked. The
molar fraction of water and specie i in the perme- rationale is that some drawback associated with
ate side, and xw and xi are the respective molar the polymeric material can be mitigated by incor-
fraction of water and specie i in the feed side. poration of inorganic particles of durable struc-
The membrane materials and separation per- tures, and MMM products may be produced with
formances studied for the dehydration applica- the manufacturing processes similar to polymeric
tions are well summarized in recent reviewing membranes. There could be a variety of combina-
articles (Chapman et al. 2008; Wee et al. 2008; tions from two groups of materials. Researchers
Bolto et al. 2012). The ux and separation factor have been actively exploring unique performance
reported for polymeric membranes are ranged attributes of the mixed matrix membranes.
from 0.01 to 1.0 kg/m2/h and from 10 to 1000, Membrane dehydration has found some indus-
respectively, over a range of operation tempera- trial applications at relatively small scales so far. It
ture from 25 C to 80 C. The microporous silica is expected that widespread applications may
membrane has a higher ux and separation factor become possible by lowering the membrane cost
at comparable separation temperatures. The zeo- and enhancing productivity through continuing
lite membrane shows the highest ux and separa- innovations.
tion factor. The ux and separation factor reported
for separation of 10 wt.% water in ethanol with
zeolite membranes (Zhang and Liu 2011) can be
as high as 10 kg/m2/h and 10,000, respectively. References
The zeolite membrane can be used for both
Bolto B, Hoang M, Xie Z (2012) A review of water
pervaporation at low temperatures and vapor- recovery by vapour permeation through membranes.
phase separation at high temperatures (above Water Res 46:259266
Dehydrogenation Reactions 525
Chapman PD, Oliveira T, Livingston AG, Li K (2008) bioorganic solutions under mild (low tempera-
Membranes for the dehydration of solvents by ture) conditions by retaining aroma compounds.
pervaporation. J Membr Sci 318:537
Wee S, Tye C, Bhatia S (2008) Membrane separation However, the principal and practical limitation
process pervaporation through zeolite membrane. which needs to be overcome is the osmotic pres-
Sep Purif Technol 63:500516 sure limitation. For example, the minimum pres-
Zhang J, Liu W (2011) Thin porous metal sheet-supported sure to dehydrate wine (11.9 % EtOH) by an
NA zeolite membrane for water/ethanol separation.
J Membr Sci 371:197210 ideally hydrophilic barrier is 64 bar; conversely,
at least 860 bar is needed to remove pure ethanol
from wine through an organophilic barrier. D
Removing water from the ethanol-water azeo-
trope (4 wt% H2O at 78 C) by reverse osmosis
Dehydration of Solvents with would require pressures in excess of 2000 bar; the
Membranes other way around is meaningless (Bddeker
2008). As an alternative, pervaporation can be
Nora Jullok implemented. This process may involve a drastic
School of Bioprocess Engineering, Universiti reduction of the activity of the permeate by caus-
Malaysia Perlis (UniMAP), Arau, Perlis, ing it to evaporate. Pervaporation can be poten-
Malaysia tially applied in the dewatering of organic liquid
such as alcohols, ketones, and ethers among
others. Although most of the installed solvent
Purication processes are required in the chemical dehydration systems are applied for ethanol dehy-
industry in order to obtain a high purity of raw dration, dehydration of other solvents, including
materials, intermediates, or end products. Separa- isopropanol, glycol, acetone, and methylene chlo-
tion of compounds from a mixture requires ride, can be considered.
energy, and in many cases, the separation is dif-
cult to perform due to the high energy consump-
tion and high cost. The chemical or physical References
properties (e.g., molecular size, vapor pressure,
freezing point, afnity, charge, density, and the Bddeker KW (2008) Liquid separations with membranes:
an introduction to barrier interference. Springer, Berlin
chemical nature) of the target compounds are
Lee KR, Lai JY (1994) Dehydration of acetic acid/water
important factors in determining the viability of mixtures by pervaporation with a modied poly
the separation process (Lee and Lai 1994; (4-methyl-1-pentene) membrane. J Polym Res
Mulder 1996). Solvent removal from dilute 1:247254
solutions via membrane ltration effectively ogy. Kluwer, Dordrecht
leads to solute enrichment, but may just serve as
a means to purify the solvent. Membrane pro-
cesses are signicantly advantageous in some
applications (e.g., hemodialysis, azeotrope split- Dehydrogenation Reactions
ting, bioseparations, ultrapure water, fuel cells)
while competing on equal terms with other tradi- Juergen Caro
tional processes (Bddeker 2008). In membrane Institut of Physical Chemistry and
processes, the aim is to allow only one compound Electrochemistry, Leibniz University, Hannover,
to pass through the membrane matrix while hin- Germany
dering penetration of other compounds. Reverse
osmosis and pervaporation are the most common
membrane processes for dewatering of solvent. Dehydrogenations of aliphatic or cyclic hydrocar-
In bioseparation, for instance, reverse osmosis bons are under technical condition thermodynam-
membranes are capable of concentrating ically controlled, i.e., the reaction takes place near
526 Dehydrogenation Reactions
Dehydrogenation ethane
Reactions,
Fig. 1 Principle of a 2-step
oxidative ethane C2H4, H2O
C2H6 C2H4 + H2
non-consumed C2H6
dehydrogenation in an
oxygen-transporting perowskite membrane
membrane reactor with a
perovkite membrane O2-depleted O2
air
air
at the equilibrium alkane D olen + H2. There are alkane + oxygen ! olen + water (steam)
hundreds of papers describing an increase of the (Czuprat et al. 2009, 2010). The problem to be
alkane conversion and an increase of the olen solved consists in the selective combustion of
yield, if hydrogen is removed through a hydrogen- hydrogen in the presence of the alkane and olen
selective membrane from the reactors. Numerous molecules. However, because of the higher reac-
hydrogen-selective membranes have been evalu- tivity of hydrogen in comparison with the hydro-
ated in dehydrogenation reactions, most often carbons, this problem could be used, by using two
Pd-based membranes as compact foils or catalysts: one catalyst for the dehydrogenation
supported thin layer, pure Pd or as alloy with Ag and the other one for the selective hydrogen
or Cu, zeolite-based membranes, but also metal combustion.
oxide membranes prepared via the sol-gel route or On the one-hand side, hydrogen is lost by
carbon membranes prepared by pyrolysis. For combustion; on the other hand, in addition, in
details, see (Dittmeyer and Caro 2008; Caro the boosting of alkane conversion and olen
2010). Interestingly, besides alkane conversion yield, there are three advantages:
and olen yield, also the olen selectivity
increased in comparison with classical xed or 1. Hydrogen combustion brings heat for the
uidized bed reactors since the removal of hydro- endothermic dehydrogenation.
gen suppresses hydrogenolysis. 2. Steam suppresses coking of catalyst and
However, for all membrane-supported dehy- membrane.
drogenation reactions with hydrogen removal, it 3. Steam dilutes the reactants which is favorable
was found that after an initial improvement of for thermodynamic reasons (shift of the equi-
alkane conversion and olen yield, because of librium) (Fig. 1).
the hydrogen drain-off, a severe coking of the
catalyst and the membrane is observed which Principle of the oxidative ethane dehydrogena-
causes the conversion and yield to be after tion in a membrane reactor with selective in situ
30 min2 h below those of a packed bed reactor. combustion of hydrogen using an oxygen-
The development of coke-resistant catalysts and transporting perovskite hollow ber membrane
membranes could not compensate this negative (Czuprat et al. 2009) (Fig. 2).
inuence. Therefore, the scientic interest in Propene selectivity as a function of the
dehydrogenation reaction dropped dramatically propane conversion in the membrane reactor
within the last decade. with oxygen-transporting hollow ber perovskite
However, there is a new concept to support membrane (lled spheres) in comparison with the
dehydrogenation reactions by a selective best literature data for the oxidative dehydrogena-
in situ combustion of the hydrogen using tion of propane (empty spheres) (Czuprat
oxygen-transporting membranes according to et al. 2010).
Dehydroisomerization of N-Butane to Isobutene 527
100
literature
this work
80 air N2
propene selectivity / %
60
O2
40
D
C3 H8 C3 H6 + H2 C3 H6 + H2O
20
0
0 20 40 60 80 100
propane conversion / %
Dehydrogenation Reactions, Fig. 2 2-step oxidative propane dehydrogenation as described in Figure 1 for ethane
References obtained as a by-product from an oil rening

process (e.g., cracking for ethylene production)
Caro J (2010) Comprehensive membrane science and engi- or produced by converting normal butane by
neering. In: Drioli E, Giorno L (eds) Comprehensive
means of a two-stage process, isomerization and
membrane science and engineering. Elsevier, Oxford,
pp 1108 dehydrogenation (Romanow-Garcia et al. 2007;
Czuprat O, Werth S, Schirrmeister S, Schiestel T, Caro Obenaus et al. 2005; Buonomo et al. 1997).
J (2009) ChemCatChem 1:401405 Butanes can be isomerized to isobutane and then
Czuprat O, Werth S, Caro J, Schiestel T (2010) AIChE
dehydrogenated to the corresponding olen; alter-
J 56:23902396
Dittmeyer R, Caro J (2008) Catalytic membrane reactors. natively, normal butane may rst be
In: Ertl G, Knzinger H, Schuth F, Weitkamp J (eds) dehydrogenated to normal butene and then, sub-
Handbook of heterogeneous catalysis. Wiley-VCH, sequently, isomerized to isobutene. The catalysts,
Weinheim, pp 21982248
reaction conditions, and technologies needed for
each of these stages are different, setting thus an
economic penalty to the composite two-stage
process.
Dehydroisomerization of N-Butane Some alternative ways to produce isobutene
to Isobutene are by dehydration of isobutanol (also by enzy-
matic dehydration) and by fermentative route.
Ilaria Mirabelli Although biological isobutene formation is
HyGear B.V. Westervoortsedijk, AV, Arnhem, known since the 1970s, extensive metabolic engi-
The Netherlands neering is required to achieve economically viable
yields and productivities (van Leeuwen
et al. 2012).
Isobutene is a valuable intermediate in the petro- An interesting alternative to the two-stage pro-
chemical industry, used as a raw material in sev- cess is a one-step process, allowing direct conver-
eral commercial processes to produce, for sion of n-butane to isobutene by means of a
instance, organic acid (methacrolein and dehydroisomerization reaction (United States Pat-
methacrylic acid), fuel additive (ethyl tert-butyl ent 3; Li and Iglesia 2008).
and methyl tert-butyl ether), and rubbery mate- n-Butane dehydroisomerization (reaction III)
rials (polybutene, isoprene). Isobutene can be involves two equilibrium reactions: dehydrogenation
528 Dehydroisomerization of N-Butane to Isobutene
of n-butane (reaction I) and skeletal isomerization found that n-parafns can be dehydroisomerized
to isobutene (reaction II); both equilibria limit to iso-olens in a single step, by using a catalyst
the possible yield of isobutene (Pirngruber composition in which the support has low acidity.
et al. 2000). Metal dispersion affects the selectivity toward
reaction products; high dispersion in the pores of
n-butane $ n-butene + H2 DHReaction (Reaction I) the support seems to lead to a very low dehydro-
(@25 C) = 130 kJ mol1
genation activity and consequently to a lower
n-butene $ isobutene DHReaction (Reaction II)
(@25 C) = 17 kJ mol1
n-butene and isobutene formation. Chromium
n-butane $ isobutene + H2 DHReaction (Reaction III) or platinum catalysts have generally been
(@25 C) = 113.7 kJ mol1 employed for dehydrogenation. A number of
solid acids, which includes ferrierite, SAPO
It is believed that dehydroisomerization pro- (silicoaluminophosphates), alumina, and zeolites,
ceeds through a transient intermediate stage can be used as support. Several zeolite-supported
wherein normal butenes are produced which are platinum catalysts, such as Pt- ZSM5, Pt-ZSM11,
then subsequently isomered (United States and Pt-MCM22, have been explored (Derouane-
Patent 4). Due to the endothermicity of dehydro- Abd Hamid et al. 2000; Bee Derouane-Abd
genation (reaction I), dehydroisomerization has to Hamid et al. 2000; Gulin Selda et al. 2008;
be carried out at high temperatures (above 400 C). Megumu et al. 2002). Bimetallic/multimetallic
The reaction is favored at low pressure occurring systems on acid support (Pt,Re/[B]-ZSM-11,
with a mole number increase. Formation of PtSn on SAPO-11) have been also investigated
by-products occurs as a consequence of some typ- due to the fact that they may exhibit better cata-
ical reactions side. Butane protolytic cracking over lytic performances than corresponding monome-
acid sites and hydrogenolysis over metal lead to tallic samples (Gulin Selda et al. 2008; Megumu
methane, ethane/ethane, and propane/propene for- et al. 2002; Ponec and Bond 1995; Bond 1991).
mation. Most of the by-products originate from Research activity on dehydroisomerization
secondary reaction of the butenes formed, which reaction is going on, and, in the future, it can
are signicantly more reactive than butane. become a feasible alternative reaction route to
To carry out the dehydroisomerization of produce isobutene.
n-butane to isobutene, two options have been
taken into account: the rst one involves the use
of a two-bed system with a dehydrogenation cat-
alyst in combination with an isomerization cata- References
lyst (United States Patent 5; Byggningsbacka
Bee Derouane-Abd Hamid S, Lambert D, Derouane Eric
et al. 1998). The other approach uses a bifunc-
G (2000) Dehydroisomerisation of n-butane over (Pt,
tional catalyst system. In the open literature, Cu)/H-TON catalysts. Catal Today 63:237247
bifunctional noble metal(s)-modied acidic sup- Bond Geoffrey C, (1991) Supported metal catalysts: some
ports have been reported (Pirngruber et al. 1999; unsolved problems Chem. Soc. Rev. 20:441475, DOI:
10.1039/CS9912000441.
Derouane-Abd Hamid et al. 2000). It is generally
Buonomo F, Sanlippo D, Trir F (1997) Dehydrogena-
assumed that the dehydrogenation is the primary tion reactions. In: Ertl G, Knozinger H, Weitkamp
reaction step which occurs over the metal sites; J (eds) Handbook of heterogeneous catalysis,
then the n-butanes produced are successively vol 5. Wiley-VCH, Weinheim, pp 21402151
Byggningsbacka R, Kumar N, Lindfords L-E (1998)
isomerized to isobutene over the Brnsted acid
Simultaneous dehydrogenation and isomerization of
sites. These kind of bifunctional catalysts have n-butane to isobutene over ZSM-22 and zinc modied
to be optimized with respect to their dehydroge- ZSM-5 zeolites, Catal Lett 55:173
nation and isomerization activity, as well as their Inaba M, Murata K, Saito M, Takahara I, Mimura N,
Hamada H, Kurata Y (2002) Dehydrogenation and
stability toward deactivation. Concentration of
Isomerization of n-Butane or Isobutane Over Cr Cata-
acid sites, metal loading, and metal dispersion lysts Supported on Zeolites. Catal Lett
are some key factors. In particular, it has been 84(34):273279
Deionization by Membrane Operations 529
Li X, Iglesia E (2008) Catalytic dehydroisomerization of processes where during water treatment cations
n-alkanes to isoalkenes. J Catal 255:134137 move in the one direction and anions in the other
Obenaus F, Droste W, Neumeister J (2005) Butenes. In:
Ullmanns encyclopedia of industrial chemistry. Wiley- direction.
VCH, Weinheim Of course, during deionization, the total
Pirngruber GD, Seshan K, Lercher JA (1999) Dehydroi- removal rate of anions (times their charge num-
somerization of n-Butane over Pt-ZSM5 (I): Effect of ber) must equal the total removal rate of cations
the Metal Loading and Acid Site Concentration. J Catal
186:188200 (time charge), to keep the produced efuent water
Pirngruber GD, Seshan K, Lercher JA (2000) Dehydroi- electroneutral (conserve electroneutrality). The
somerization of n-Butane over Pt-ZSM5: II. Kinetic condition of electroneutrality will hold both for D
and thermodynamic aspects. J Catal 190:338351 the produced freshwater and for the produced
Ponec V, Bond GC (eds) (1995) Catalysis by metals and
alloys, studies in surface science and catalysis, vol 95. concentrate, the brine.
Elsevier, Amsterdam To make ions go in different directions, deion-
Pozan Soylu, GS, Baygin Isik, A, Boz, I (2009) Dehydroi- ization processes in the limited sense as dened
somerization of n-butane over metal promoted sulfated above make use of electrical eld effects, i.e., by
zirconia. Turk. J. Eng. Environ. Sci. 33:273279
Romanow-Garcia S, Hoffman HL (2007) Petroleum and its applying an external current to a water ow cell or
products. In: Kent JA (ed) Kent and Riegels handbook system, ions in the water are induced to move in
of industrial chemistry and biotechnology. Springer, the electric eld, with cations moving to positions
New York, pp 801842 of lower electrical potential (i.e., in the direction
United States Patent 3,756,941 Dehydroisomerization pro-
cess, Norman D. Carter, Poughkeepsie, John H. Estes, of the eld strength), while anions move in the
Wappingers Falls, and Stanley Kravitz, Wiccopee, N. opposite direction.
Y., assignors to Texaco Inc., New York. No Drawing The most well-known membrane process for
United States Patent 4,392,003 Isobutene by dehydroi- deionization is electrodialysis and the various
somerization of normal butane, Kolombos et al.,
assignors to The British Petroleum Company modications thereof, including electrodeio-
United States Patent 5,336,830 Dehydroisomerisation cat- nization, a process in which ion-exchange resin
alyst and its use in the preparation of isobutene from particles ll the water ow channels in between
N-butane, G. Bellussi, A. Giusti, L. Zanibelli, 1994, the ion-exchange membranes and electrodialysis
assigned to Eniricerche S.p.A. and Snamprogetti S.p.A
van Leeuwen BNM, van der Wulp AM, Duijnstee I, van reversal, a modication of electrodialysis in
Maris AJA, Straathof AJJ (2012) Fermentative produc- which the applied electrical current is temporarily
tion of isobutene. Appl Microbiol Biotechnol reversed during short periods of time to abate
93(4):13771387 biofouling and scaling on the membranes (Sonin
and Probstein 1968; Strathmann 2010;
Nikonenko et al. 2010; Andersen et al. 2012).
A novel deionization technology making use of
Deionization by Membrane membranes is membrane capacitive deionization in
Operations which ion-exchange membranes are placed in front
of porous electrodes (either in the form of freestand-
Maarten Biesheuvel ing membranes or as a coating on the electrode) to
Wetsus, European Centre of Excellence for capacitively remove salt from water in a cyclic pro-
Sustainable Water Technology, Leeuwarden, cess, i.e., ions are temporarily stored in the electrodes
The Netherlands in one phase of the cycle and are released again
during the phase where a brine is produced (Lee
et al. 2006; Li and Zou 2011; Biesheuvel
Deionization, or desalination, is the process in et al. 2011; Nie et al. 2011; Zhao et al. 2012).
which salt ions are removed from a polar liquid,
which is virtually always water. Water deioniza-
tion and water desalination are generally consid- Cross-References
ered as synonyms, though deionization can be
applied in a more restricted sense to denote those Membrane Capacitive Deionization
530 Dense Membranes
References a b
Andersen MB et al (2012) Current-induced membrane

discharge. Phys Rev Lett 109:108301
Biesheuvel PM, Zhao R, Porada S, van der Wal A (2011)
Dense Membranes, Fig. 1 Schematic drawing illustrat-
Theory of membrane capacitive deionization including
ing the a symmetric and b asymmetric structures of dense
the effect of the electrode pore space. J Colloid Inter-
membranes
face Sci 360:239248
Lee J-B, Park K-K, Eum H-M, Lee C-W (2006) Desalina-
tion of a thermal power plant wastewater by membrane
capacitive deionization. Desalination 196:125134 membrane determines the ux. These types of
Li HB, Zou L (2011) Ion-exchange membrane capacitive membranes are also named homogeneous
deionization: a new strategy for brackish water desali- (dense) membranes. The asymmetric membrane
nation. Desalination 275:6266
consists of thin and dense selective layer (skin)
Nie CY, Zhan YK, Pan LK, Li HB, Sun Z (2011)
Electrosorption of different cations and anions with supported on a much thicker microporous support
membrane capacitive deionization based on carbon layer that provides mechanical support.
nanotube/nanober electrodes and ion-exchange mem- The surface layer performs the separation, and
branes. Desalination Water Treat 30:266271
it is the principal barrier to the ow through the
Nikonenko VV et al (2010) Intensive current transfer in mem-
brane systems: modelling, mechanisms and application in membrane (Strathmann et al. 2006). They can be
electrodialysis. Adv Colloid Interface Sci 160:101123 prepared from a large number of different mate-
Sonin AA, Probstein RF (1968) A hydrodynamic theory of rials: inorganic (metals, ceramics) or organic
desalination by electrodialysis. Desalination
(polymers).
5:293329
Strathmann H (2010) Electrodialysis, a mature technology Inorganic materials generally possess superior
with a multitude of new applications. Desalination chemical, mechanical, and thermal stability rela-
264:268288 tive to polymeric materials. However, these mate-
Zhao R, Biesheuvel PM, Van der Wal A (2012) Energy
rials have the disadvantages of being very brittle
consumption and constant current operation in mem-
brane capacitive deionization. Energy Environ Sci and more expensive than the organic materials.
5:95209527 The dense inorganic membranes, extensively
studied in the past decade, include metallic mem-
branes, primarily palladium alloy membranes
used for hydrogen separation. Generally, there
Dense Membranes are three techniques for coating metallic thin
lms onto porous metallic or ceramic supports:
Fabio Bazzarelli, Lidietta Giorno and Emma electroless plating, chemical vapor deposition
Piacentini (CVD), and physical sputtering. Electroless plat-
Institute on Membrane Technology, National ing is a method of metal plating by autocatalytic
Research Council of Italy, ITM-CNR, Rende, chemical reduction of the corresponding metal
Italy ions with simultaneous oxidation of a reducing
agent. This technique can be applied for forming
metal coatings even on nonconductive supports
Dense membranes consist of dense structure such as porous ceramic or glass. In CVD, a vola-
presenting no detectable pore at the limits of elec- tile component of coating materials is thermally
tron microscopy. A mixture of molecules is decomposed on the surface of the heated substrate
transported through dense membranes by diffu- to form a thin lm or coating. In physical
sion under the driving force of a pressure, concen- sputtering, a thin continuous lm is deposited on
tration, or electrical potential gradient. Dense the porous substrate by the bombardment with
membranes may have a symmetric or an asym- high-energy particles on the support (Adhikari
metric structure (Fig. 1). The rst type has a and Fernando 2006).
uniform composition and structure over the entire The dense polymeric membranes are produced
cross section, and the thickness of the entire mainly by phase inversion process and coating
Dense Membranes 531
technique. The phase inversion is a process where a Dense Membranes, Table 1 Applications of processes
polymer is transformed in a controlled manner from based on dense membrane
liquid to a solid state. The process of solidication is Type of
very often initiated by the transition from one liquid process Applications Polymeric membrane
state into two liquids (liquid-liquid demixing). At a Reverse Desalination of Cellulose acetate
osmosis brackish water
certain stage during demixing, one of the liquid and seawater
phases (the high polymer concentration phase) Concentration Aromatic polyamide
will solidify so that a solid matrix is formed. The of solutions of
phase inversion can be induced by solvent evapo- food products,
Sulfonated
polysulfone
D
pharmaceutical
ration, precipitation from the vapor phase, and ther-
solutions, and
mal and immersion precipitation. Making dense chemical
membrane by coating involves a two-step process. streams
One is the preparation of a suitable porous support Wastewater Polybenzimidazole
obtained by phase inversion, and the second is the treatment
preparation of the barrier layer on the surface of the Gas Separation of Polymers of intrinsic
separation nitrogen from microporosity (PIMs)
support. Several techniques can be used to apply a air polydimethylsiloxane
thin top layer on support such as dip coating, spray Removal of
coating, spin coating, grafting, plasma polymeriza- water from air
tion, interfacial polymerization, and in situ poly- Recovery of Peruoropolymer
merization to achieve these membranes (Mulder hydrogen from Teon AF 2400
ammonia purge
1996). Coating technique is used to prepare thin
gases
and dense structure, possessing high ux.
CO2 separation Polyimides
Dense membrane is used mainly to separate from ue gas
components which are similar in size but have streams
different chemical nature in processes such as Pervaporation Removal of Polyurethane
reverse osmosis, gas and vapor separation, and volatile organic Polydimethylsiloxane
compounds
pervaporation. The transport of gas, vapor, or from
liquid through a dense membrane can be wastewater or
described in terms of a solution-diffusion model. gas streams
According to this model, permeant molecules are Recovery of Poly(vinyl alcohol)
dissolved into the entrance face of the membrane aroma and Poly(vinylidene
biofuels from uoride)
and diffuse across the membrane matrix, then fermentation
dissolved species are desorbed in the downstream broth
face of the membrane. The reverse osmosis pro- Dehydration of
cess uses a large pressure difference across the organic
solvents
membrane to separate water from salt solutions.
In pervaporation, the pressure difference across
the membrane is small, and the process is driven
by the vapor pressure difference between the feed of materials. Table 1 summarizes some applica-
liquid and the low partial pressure of the permeate
tions of processes based on dense membrane, and
vapor. Gas permeation involves transport of gases
the most used polymeric membranes for each
under a pressure or concentration gradient (Baker application are also reported.
2004). The separation of the various components
of mixture is directly related to their transport
rates within the membrane phase, which is deter- References
mined by their diffusivities and concentration in
the membrane matrix. Therefore, the performance Adhikari S, Fernando S (2006) Hydrogen membrane sep-
of membrane is determined by intrinsic properties aration techniques. Ind Eng Chem Res 45:875881
532 Density Functional Theory Modeling of Membrane Systems
Baker RW (2004) Membrane technology and applications. functional, which takes into account the differ-
Wiley, Chichester ence between the kinetic energies of the isoelec-
ogy. Kluwer, Dordrecht tronic interacting and noninteracting systems in
Strathmann H, Giorno L, Drioli E (2006) An introduction addition to the difference between the quantum
to membrane science and technology. Betmultimedia, electron-electron and Coulomb electrostatic
Rome energy. r(r) can be evaluated by the nonlinear
Kohn-Sham equations, dened by means of an
effective potential. The electron density allows to
calculate all the properties of the quantum system
Density Functional Theory Modeling in addition to the total energy. Computational
of Membrane Systems approaches, based on density functional theory,
are applied in the study of catalysts used in mem-
Giorgio De Luca branes and in the study of noncovalent interac-
Institute on Membrane Technology, National tions, such as hydrogen-bonding and London
Research Council of Italy, ITM-CNR, Rende, dispersion interactions (De Luca et al. 2009). In
Italy fact, hydrogen-bonding, electrostatic interactions
and London force are important for membranes at
the basis of fundamental properties such as molec-
Density functional theory is a new methodology ular adsorption and sorption, recognition, and
in the frame of quantum mechanics (ab initio). self-assembly. Also size, shape, and electrostatic
The assessment of material features depending features of supramolecular architectures can be
on the electron interactions or electron density studied using density functional theory. All these
polarization necessarily needs the use of quantum properties control the selectivity of the materials,
mechanics. Large molecular systems cannot be used in membranes, the permission as well as the
described with accurate quantum mechanical antifouling or anti-embrittlement features. Den-
approaches, such as correlated Hartree-Fock sity functional theory studies of the catalysis in
(Szabo and Ostlund 1994), due to the huge com- membrane reactors require the denition of struc-
putational time required by these. Instead, the tural models of the catalysts. For the most part,
density functional theory (Parr and Yang 1989) they are modeled by innite surfaces, slabs, or
allows to get results with similar precision in different types of adsorbed or absorbed atomic
relatively shorter computational time. Thus, clusters. These studies could be treated indepen-
density functional theory is a powerful tool to dently from the aforementioned analysis. How-
investigate large chemical systems like the ever, it is important to emphasize that the
nanostructures involved in the membrane prepa- merging of the results obtained by the different
ration. In density functional theory, the total investigations would be advisable. For example,
energy of an electronic system is evaluated the kinetic constants characterizing a reaction
through a total functional which depends on the path, related to a particular catalyst, evaluated by
electron density of the quantum system, r(r), and the density functional theory, should be compared
external potential v(r). It is dened as follows: with the diffusion coefcients characterizing the
permeability of reagents and products through the
EDFT rr Ts rr Jrr Exc rr membrane derived by the adsorption/absorption

of different molecules.
vrrrdr
where Ts[r(r)] is the kinetic energy of an isoelec- References

tronic noninteracting system, while J[r(r)]
describes the Coulomb electrostatic energy. The De Luca G, Gugliuzza A, Drioli E (2009) Competitive
Exc[r(r)] is the exchange and correlation hydrogen-bonding interactions in modied polymer
Design of Experiment (DOE) 533
membranes: a density functional theory investigation. the response variables are affected by varying
J Phys Chem B 113:54735477 each factor with all other factors held constant.
Parr RG, Yang W (1989) Density functional theory of
atoms and molecules. Oxford University Press, New Figure 1 shows a set of these graphs in which the
York variation of permeate ux (Jp) is studied as a
Szabo A, Ostlund NS (1994) Modern quantum chemistry. function of operating time, transmembrane pres-
Macmillan Publishing CO, New York sure (TMP), feed concentration, and temperature.
Despite that this approach has been widely
used, the major disadvantage is that it does not
Design of Experiment (DOE) consider any possible interaction between the fac- D
tors. An interaction is the failure of the one factor
Ren Ruby-Figueroa to produce the same effect on the response at
Institute on Membrane Technology, National different levels of another factor. In membrane
Research Council of Italy, ITM-CNR, Rende, processes interactions between factors are very
Italy common, and if they occur, the one-factor-at-a-
time strategy will usually produce poor results.
One-factor-at-a-time experiments are always less
Generally speaking, experiments are performed efcient than other methods based on a statistical
by investigators in virtually all applications of approach.
membrane processes, usually to discover some- In this regard, DOE can be dened as a sys-
thing about a particular process or system. More tematic approach to understanding how factors
formally, we can dene an experiment as a test or affect the response variables such as efciency,
series of tests in which purposeful changes are yield, or productivity. DOE uses a statistical meth-
made to the input variables (also called odology to analyze data and predict a response
factors) of a process or system so that we under all possible conditions within the limits
observe and identify the reasons for changes that selected for the experimental design. In addition,
may be observed in the output response (also DOE allows to generate the required information
called response variables or just response) in order to determine which factors and interac-
(Montgomery 2001). Until now the extensively tions are signicant in contributing to the response
used strategy of experimentation in membrane being measured and those factors and interactions
eld has been the one-factor-at-a-time approach. that are insignicant and do not contribute to
This method consists of selecting a starting point, either a particular response. As a matter of fact,
or baseline set of levels, for each factor, then the use of this approach allows to save time and
successively varying each factor over its range money, minimizing the amount of experimenta-
with the other factors held constant at the baseline tion (Wagner 2014).
level. After all tests have been performed, a series DOE includes different classes of designs
of graphs are usually constructed showing how divided in three main groups such as mixture
Jp Jp Jp Jp
operating time TMP Feed concentration Temperature
Design of Experiment (DOE), Fig. 1 Typical examples of the one-factor-at-a-time strategy concerning the effect of
operating parameters on the permeate ux
534 Design of Experiment (DOE)
Design of Experiment measure for each experiment). The limits used in

(DOE), Fig. 2 Steps Define the problem screening design, as well as response surface, are
involved in design of
experiment (DOE)
represented as 1 and 1 for the highest and the
lowest value, respectively, and 0 for the middle
value.
Plan of experiment In membrane processes, some common inde-
pendent factors include transmembrane pressure,
temperature, ow rate, and membrane character-
istics as well. On the other hand, permeate ux,
Run the experiment rejection, and selectivity are common dependent
variables or responses. Independent variables are
normally quantitative in nature, meaning that they
are set at a specic numerical value. However, in
Analyze the data some experimental design, the independent vari-
ables are qualitative.
The quantity of experimental runs will depend
on the quantity of selected factors as well as the
Report the results type of design chosen. Experimental design crea-
tion produces all possible combinations of maxi-
mum, minimum, and middle values. Finally the
experiments, screening design, and response sur- investigator will obtain the quantity of experimen-
face. Each design is used in specic situations to tal runs as well as the conditions for each
gather information from a particular set of inde- experiment.
pendent variables. Mixture design can be used to The experimental procedure, in which the
optimize membrane characteristics (response var- investigator performs experimental runs
iable) over different compositions of materials, according to the experimental design, is a critical
solvents, additives, etc. step in DOE, because in this phase response vari-
On the other hand, screening and response ables data are collected for all the experiments or
surface designs can be applied to study all the treatment combinations.
membrane processes, to get the main factors to Once completed all the experiments and col-
control and the optimal process conditions, as lected all the data for each response variables, data
well. Screening design is used to screen process analysis can be performed. Generally the analysis
and/or product variables, while response surface is carried out by using statistic software in which
is used to get a better description of their curvature the investigator should:
and interactions in the experimental space.
Response surface approach, unlike factorial Examine the presence of outliers and typos. In
design, includes the quadratic term in the model; this step many graphs such as response distri-
therefore it shows curvature. butions (histograms, box plots, etc.) and
DOE procedure involves several steps, as responses versus factors levels (main effect
shown in Fig. 2. mean plots and interaction plots), can be
Dening the problem may seem obvious, but constructed.
in practice, it is sometimes difcult to do. In this Evaluate by means of ANOVA analysis the
step the aim of the study should be properly and signicance of each factor as well as interac-
clearly dened before starting any experiment. tion and quadratic effect (only for response
After the problem is properly dened, the second surface designs).
step in DOE framework is to select independent Create the model to describe the experimental
variables (factors), with their limits for evaluation, data. The most common empirical models to t
as well as dependent variables (responses to the experimental data take either a linear form
Diafiltration 535
(used in factorial design) or quadratic form batch operation, the retentate stream is recycled
(applied for response surface). into the feed tank where it is mixed with the content
Check the correlation coefcient (R2) of the of the reservoir and with fresh diluant. The concen-
model, normal distribution of the residuals tration of membrane-permeable microsolutes in the
(error or noise), and autocorrelation (using feed tank decreases over time, and the nal product
Durbin-Watson statistic) as well. is obtained directly in the feed reservoir when oper-
ation is terminated. In contrast, continuous
Subsequently, the model can be used to give an processing is performed in a ow-through manner
answer to the questions xed in the experiment with both permeate and retentate being withdrawn. D
objectives as well as to determine the optimal Thus, batch dialtration is a dynamic process in
conditions to multiple responses. which the composition of uid stream fed into the
membrane module is a function of time, while
continuous dialtration is essentially a static
References (steady-state) process in which the feed stream is
of constant composition.
Montgomery DC (2001) Design and analysis of experi- The benets and disadvantages offered by the
ment. Wiley, New York
two processing modes in industrial-scale ultral-
Wagner JR Jr (2014) Design of experiments. In: Wagner JR
Jr, Mount EM III, Giles HF Jr (eds) Extrusion, 2nd edn. tration (UF) systems have been summarized by
Elsevier, Amsterdam, pp 291308 Lipnizki et al. (2002) as shown in Table 1.
Dialtration has found a vast number of indus-
trial applications in the chemical, pharmaceutical,
environmental, biotech, and food and beverage
Detergents sectors. The requirement for an effective separa-
tion is the utilization of a membrane that has a
Amphiphilic Molecules high rejection for macrosolute but allows the pas-
sage of microsolute. The mass transport, unlike in
dialysis, is driven by pressure differences across
the membrane. Depending on the particular appli-
Diafiltration cation, either the retentate or the permeate is the
valuable phase of economic importance (or both),
Zoltn Kovcs and the applied membrane can be a microltration
Department of Food Engineering, Institue of (MF), ultraltration (UF), nanoltration (NF), or
Bioengineering and Process Engineering, Szent reverse osmosis (RO) membrane. Some typical
Istvan University, Budapest, Hungary applications include:
MF: enzyme recovery from fermentation

Dialtration (also called membrane ltration broths; cell harvest and debris removal; sepa-
in dilution mode) is an operation mode of a ration of casein from serum milk proteins
pressure-driven membrane ltration process in UF: protein purication and buffer exchange;
which a diluant (pure solvent) is added into the ethanol removal from human serum albumin
process liqueur in order to facilitate the separation obtained from blood plasma; fruit juice
of macrosolutes from microsolutes. clarication; milk fractionation; Kraft black
Dialtration may be performed either in batch liquor treatment; recovery of small molecules
(discontinuous) or continuous processing mode as (e.g., amino acids, antibiotics, organic acids,
schematically shown in Fig. 1. In both processing etc.) from fermentation broth
modes, diluant is added to and permeate is with- NF: desalination of dyes and optical bright-
drawn from the system. The fundamental differ- eners; partial demineralization of whey; pro-
ence between the two congurations is that in duction of desalted and concentrated
536 Diafiltration
a
Diluant
Retentate
b
Diluant Retentate
Feed
Permeate Feed Permeate
Diafiltration, Fig. 1 Schematic ow diagram. (a) Batch dialtration. (b) Continuous dialtration
Diafiltration, Table 1 Comparison of batch and contin- applications require a simultaneous microsolute
uous diafiltration (Adopted from Lipnizki et al. 2002) removal (i.e., macrosolute purication) and
Batch Continuous macrosolute concentration (i.e., feed volume
UF plant investment Low High reduction). In many cases, the term dialtration
Batch tank investment High Low stands for a process that is designed to achieve
Control investment Low High these dual objectives. In continuous processing,
Space requirement UF Low High the different modes of operation (i.e., concentra-
Space requirement tank High Low tion and dilution modes) are assigned to different
Feed volume exibility High Low stages which are connected in series. In batch
Antibiotic/protein Variable Constant processing, the targeted twin aims of concentra-
concentration in permeate
tion and fractionation are obtained either in con-
Residence time High Low
secutive operational steps (e.g., in traditional
Temperature exibility Low High
dialtration, intermittent feed dialtration,
Module efciency Low High
sequential dilution dialtration) or applying
Energy consumption High Low
specic diluant utilization strategies (e.g.,
variable-volume dialtration, dynamic-volume
dialtration). Overall, these characteristics of
antibiotics; desalination of gelatin for dialtration put it in a favorable position in com-
improved color and whipping properties parison with other separation techniques and even
RO: solvent replacement in herb extracts; beer with a sequence of competitive unit operations
dealcoholization (Jnsson and Trgrdh 1990).
Dialtration is also called membrane ltration

in dilution mode in accordance with the terminol- Cross-References
ogy recommendation of the former European
Society of Membrane Science and Technology Dynamic-Volume Dialtration
(now known as the European Membrane Society)
(Gekas 1988). Although the latter term is less
frequently used in the literature, it better reveals References
that an operation mode of a ltration process is
Gekas V (1988) Terminology for pressure-driven membrane
considered.
operations. Desalination 68(1):7792. doi:10.1016/0011-
Membrane ltration plants are rarely designed 9164(88)80045-6. http://www.sciencedirect.com/science/
exclusively for microsolute removal. Most article/pii/0011916488800456
Dialysis 537
Jnsson AS, Trgrdh G (1990) Ultraltration applica- In discontinuous dialtration processes,

tions. Desalination 77:135179. doi:10.1016/0011- including both dilution and volume reduction
9164(90)85024-5
Lipnizki F, Boelsmand J, Madsen RF (2002) Concepts of methods, one must specify both the level of vol-
industrial-scale dialtration systems. Desalination ume reduction/dilution in each stage and also the
144:179184 number of stages so the dialtration factor is not
so easily, or meaningfully, dened in those
Further Reading instances.
Beaton NC, Klinkowski PR (1983) Industrial ultraltration
design and application of dialtration processes. J Sep D
Process Technol 4(2):110
Dutr B, Trgrdh G (1994) Macrosolute-microsolute
separation by ultraltration: a review of References
dialtration processes and applications. Desalination
95:227267 Cheryan M (1998) Ultraltration and microltration hand-
Madsen RF (2001) Design of sanitary and sterile UF- and book, 2nd edn. CRC Press, Boca Raton
dialtration plants. Sep Purif Technol 2223:7987
Dialysis
Diafiltration Factor
Emma Piacentini, Lidietta Giorno and Fabio
Greg Foley Bazzarelli
School of Biotechnology, Dublin City University, Institute on Membrane Technology, National
Dublin, Ireland Research Council of Italy, ITM-CNR, Rende,
Italy
One of the most commonly used methods of

dialtration is constant volume dialtration Dialysis is a process in which small solutes
(CVD) also known as continuous dialtration. In diffuse from a high concentration solution to a
this process, diluant is added to the retentate solu- low concentration solution across a semiperme-
tion which is undergoing batch ultraltration. The able membrane until equilibrium is reached
process is set up so that the rate of addition of the (Strathmann 2006; Mulder 1996). From a liquid
diluant exactly balances the permeate ow solution containing the solvent and molecular
rate hence the retentate volume is constant. For components of different molecular weight or
any given solute with apparent rejection coef- sizes (feed solution), the components diffuse
cient, s, the solute concentration declines during a through the membrane into the solvent referred
CVD process according to the equation (Cheryan as dialysate or stripping solution under the force
1998) of a concentration difference (Fig. 1). Dialysis
process is normally operated in a countercurrent
c c0 e1sV b =V s conguration and involves the use of nonporous
membrane.
where Vb is the volume of diluant (buffer) added Membrane materials most often used include
while Vs is the solution (i.e., retentate) volume. regenerated cellulose, cellulose acetate, polysulfone,
The ratio Vb/Vs is termed the dialtration factor, polycarbonate, polyethylene, polyolen, polypro-
sometimes referred to as the number of pylene, and polyvinylidene uoride.
diavolumes. Once the rejection coefcient is There are different forms of dialysis depending
known, this quantity completely species the on the components of the feed solution and the
reduction in solute concentration that will be structure of the membrane. The most simple form
achieved during CVD. of dialysis is the diffusion of neutral molecules;
538 Dialysis
DIALYSIS
FEED SOLUTION FLOW
FEED SOLUTION
MEMBRANE
DIALYSATE
DIALYSATE FLOW
Dialysis, Fig. 1 Schematic representation of dialysis process operated in a countercurrent conguration
instead when the electrolytes are separated with Da Dc za zc dCs

Js
neutral membrane or with charged membrane, the Da za Dc zc dx
process is called Donnan dialysis.
When the feed solution contains neutral mole- where Js is the rate of transfer of salts or ux,
cules, the transport in dialysis is described by the dCs/dx is the concentration gradient of salt s, and
Ficks law: Da and Dc are the diffusion coefcients of anion
a and cation c, respectively.
dCi The situation becomes more complex if the
J i Di
dx feed solution contains more than one electrolyte
and if the membrane carries xed charges. There
where Ji is the rate of transfer of component i or are two forms of coupled transport (Fig. 2):
ux, dCi/dx is the concentration gradient of com-
ponent i, and Di is the diffusion coefcient that is a Cocurrent or diffusion dialysis (Fig. 2a) when
measure of the mobility of the individual co- and counterions are transported in the same
molecules. direction
When feed solution contains and electrolyte Countercurrent or Donnan dialysis (Fig. 2b)
which consists of a cation and an anion with when no counterions permeate the membrane
signicantly different diffusivities, an electrical and co-ion uxes are in an opposite direction
potential gradient will be created. This potential
gradient is referred as diffusion potential, and In the cocurrent coupled transport (Fig. 2a), the
although it might be very small, it will slow feed solution, containing the component A (i.e.,
down the faster diffusing ion and speed up the Cl) as anion and B+ and C+ (i.e., Na+ and H+) as
slower ion so that both ions will diffuse with the cations, is separated by an anion-exchange mem-
same velocity. Because the membrane does not brane, from a solution containing the component
carry xed charges, all ions of the electrolyte are A as anion and only C+ as cation. If the concen-
transported in the same direction even if their tration of specie AB (i.e., NaCl) is assumed to be
diffusion coefcient is quite different. In this much higher in the feed solution than in the other
case, the transport is described by the following solution, while the concentration of AC (i.e., HCl)
equation: is supposed to be lower in the feed solution than in
Dialysis 539
Dialysis, Fig. 2 Schematic a CO-CURRENT COUPLED TRANSPORT

diagram illustrating (a)
FEED SOLUTION: AB (i.e. NaCl) + AC (i.e. HCl) AB > AC
cocurrent coupled and (b)
countercurrent transport Cl- Cl-
Cl- H+ H+ Na+ Na+
Cl-
+ + + + + + + + +
MEMBRANE
Cl- Cl- Cl- H + H+ H+

Cl- Cl- H+ H+ D
SOLUTION: AC (i.e. HCl)
b COUNTER-CURRENT TRANSPORT OR
DONNAN DIALYSIS
FEED SOLUTION: AB (i.e. NaCl) + AC (i.e. HCl)
Cl- Cl- Cl- Cl- H+ H+ Na+ Na+
Cl- Cl- Cl- Cl- H+ Na+
Cl- H + H+ Na+ Na+
Cl-
- - - - - - - - -
MEMBRANE
Cl- Cl- Na+ Na+

H+ H+
Cl- Cl-
SOLUTION: AB (i.e. NaCl) + AC (i.e. HCl)
the other solution, the anion A can pass through blood of a person with a kidney failure is treated in
the membrane from the feed solution to the other a dialysis process to remove toxic small molecular
solution as a consequence of the concentration weight metabolites such as urea, creatine, uric acid,
gradient. Because of the electroneutrality require- and others. During the treatment, the blood ows
ment, an anion can permeate the membrane only on one side of the membrane, whereas dialyzing
when accompanied by a cation. The only cation uid, which contains vital salts such as sodium,
able to permeate through the membrane by using a potassium, and calcium that may not be removed
specic carrier is C+ that is transported against from blood, ows on the other side. The small
their concentration gradient. In the countercurrent organic solutes diffuse through the membrane,
coupled transport (Fig. 2b), the feed solution, and the process is continued until the concentration
containing the component A (i.e., Cl) as anion of the toxic components has been reduced to a
and B+ and C+ (i.e., Na+ and H+) as cations, is certain level. Other applications are the recovery
separated by a cation-exchange membrane, from a of caustic soda from colloidal hemicellulose during
solution containing the same components. If the viscose manufacture and removal of alcohol from
concentration of species CA and BA (i.e., HCl and beer. Also in biotechnology and pharmaceutical
NaCl) in the feed solution is assumed to be much industry, the process is used to remove salts from
higher than in the other solution, the anion A bioproducts and for fractionation.
cannot pass through the membrane, while the
cation C+ moves from feed solution to the other
solution and B+ in the opposite direction against References
its gradient concentration.
Dialysis process is mainly used to separate low Mulder M (1996) Basic principles of membrane technol-
molecular weight components from those of high
Strathmann H, Giorno L, Drioli E (2006) An introduction
molecular weight. The major application of dialy- to membrane science and technology. Betmultimedia,
sis is the articial kidney; in this application the Rome
540 Dielectric Exclusion Model in Membranes
of neutral solutes, atomic force microscopy

Dielectric Exclusion Model in (AFM), or liquid-liquid porosimetry (Otero
Membranes et al. 2008).
3. The volumetric charge density: X. It can be
Veronica Silva inferred by the tangential streaming potential
Dow Water and Process Solution Division at The (TSP) technique, where the zeta potential and
Dow Chemical Company, Boulevard Cecilia the surface charge density on the top surface of
Grierson 355, Buenos Aires, Argentina the membrane are measured and then, assum-
ing that this charge is representative of the
selective layer of the membrane, the value of
Introduction X can be calculated (Martnez et al. 2002).
However, a concentration gradient appears
Dielectric exclusion has been recognized in the through the membrane and X is correlated
last two decades as one important mechanism for with this gradient as long as the separation
ion separation in membranes with xed charges in takes place. Thus, a charge gradient has to be
the active surface as it is the case of NF mem- considered as well. In a recent work, a
branes. They are manufactured to selectively Freundlich isotherm was included to correlate
reject a specic ion or ion groups, which was X with ion concentration (Silva et al. 2011).
achieved by the inclusion of functional groups 4. Dielectric constant inside the pores: ep. It can
(charges) in the membrane active layer. These be obtained by impedance spectroscopy tech-
charges produce an additional rejection due to nique or by mathematical modeling of
electrostatic phenomena that hinder the move- designed experiments.
ment of charges through the membrane. Conse-
quently, NF membranes allow the rejection of ions
when their size is much lower than the pore size. Dielectric Exclusion Model
The rejection of the target compounds has been
studied to occur in two main mechanisms: At the feed and permeate interfaces a thermody-
namic equilibrium takes place and a difference of
Partitioning mechanisms: steric effect, Donnan solute concentration (ci) is generated from one
equilibrium, and dielectric exclusion; which side to the other at each interface. These concen-
occur in the interfaces of the active layer. tration leaps can be described as:
Transport mechanisms: convection, diffusion,
and electrokinetic effects; which occur through ci, 0 gi , m 0

fi expzi DCexp DWi, B
the active layer thickness length. ci, m gi, 0
0

exp DWi, im
The phenomenological models available in lit-
erature always include the transport mechanisms (1)
through the Nernst-Planck simplied equation but
not all of them have been implemented with all the for the x = 0 interface and similarly for the
partitioning mechanisms recognized year to date. x = Dx interface.
In some cases, dielectric exclusion has been In this equation all partitioning effects are
neglected. NF models have four parameters: included: fi takes into account the steric effect,
the deviation from the ideal bulk solutions at both
1. The ratio (active layer thickness)/(active layer sides of the membrane is taken into account
porosity): Dx/A. It is obtained using Hagen- through the activity coefcients gi,m and gi,0; and
Poiseuille law. DC represents the Donnan potential.
2. The active layer mean pore radius: rp. It is The dielectric exclusion is accounted in the
usually obtained from experiments of retention terms: DW0i,B for the Born contribution and
Dielectric Exclusion Model in Membranes 541
DW0i,im for the effect of the image forces one of the membrane, reducing the effect of
(Yaroshchuk 2000). The expressions typically image forces within the pores while increasing
used for these dielectric effects are shown in the solvation energy barrier. In that case, Eq. 2 is
Box 1. used as an approximation to obtain the dielectric
In the Born effect, a difference between the exclusion effect.
dielectric constants of the bulk solution (eb) and
the solution inside the pore (ep) exists as a result of Box 1: Dielectric Exclusion Mathematical
connement effects. Therefore, to transfer the Model
ions from the bulk solvent to the membrane is The dielectric Born energy (DWi,B) D
necessary to make an extra work that can be

estimated using Born equation (Eq. 2) and assum- zi e2 1 1
ing that ep must be lower than eb. DW 0i, B (2)
8pkB Te0 as ep eb
The image force effects appear when two
media of different dielectric constants, like a Where as is the cavity radius dened by Rashin
membrane matrix and an electrolyte solution, are and Honig (Rashin and Honig 1985) as the dis-
in contact and the interaction of ions with the tance from the center of the ion to the point where
electrical charges induced by the ions at the inter- the dielectric constant becomes different than the
faces causes a dielectric exclusion. The interaction vacuum one, e0.
with a polarized interface is well known as image The dielectric image force energy (DWi,im)
force in analogy to the interaction with a cti- This effect depends on the geometry of the
tious image charge located at the other side of the pores,
interface, at the same distance from it than the real
charge. The sign and magnitude of the image are
1
determined by the dielectric constants of the 0 2ai K0 kK1 n bK0 nK1 k
DWi, im dk
media. Consequently, the exclusion due to image p I1 nK0 k bI0 nK1 k
0
force effect depends on the difference between the Cylindrical geometry
dielectric constant of the membrane matrix (em) (3)
and the solution inside the pores.
Equations 1, 2, 3, 4, 5, 6, 7, and 8, have to be
0 ep em
introduced the more global a complex NF model DWi, im ai ln 1 exp2m
ep em
for the resolution where ep appears as only tting Slit like geometry
parameter (or coupled with X). A resolution algo- (4)
rithm can be found in specic literature for NF
modeling (Yaroshchuk 2000). Where,
Nowadays, all authors are in agreement that
dielectric effect must be used to characterize the
zi F2
rejection of ionic solutions. But, there is still con- ai (5)
8pe0 ep RTNA rp
troversy if Born or image force effects are more
suitable for NF. The Born effect has been used by p
many authors (Bowen et al. 1997; 2002) n k2 m2 (6)
neglecting the image force effects, since the
small NF pores make the ep value close to the
v

u
uX z2 ci, m f g =g exp zi DC DW0 DW0
t i i i, m i, 0 i, B i, im (7)
m Frp i RTe0 eb
542 Dielectric Permittivity

k em presence of an applied electric eld (Cosgrove
b p (8) 2010; Hunter 2001). The ratio of the dielectric
k m ep
2 2
permittivity of a material to that of a vacuum
The K0, K1, I0, I1 are the modied Bessel func- (o = 8.85 1012 F/m) is dened as the
tions, em is the dielectric constant of the dry poly- dielectric constant (k). Water has a dielectric con-
mer of the membrane, which is usually a constant stant of 80.1 at 20 C; this large value reects the
value obtained from bibliography highly polar nature of water.
The dielectric permittivity appears in a number
of important equations describing the surface
Cross-References charge and electrostatic interactions in membrane
systems. For example, the membrane zeta poten-
Impedance Spectroscopy tial (a measure of the membrane surface charge)
can be evaluated from the measured streaming
potential (Ez) using the Helmholtz-Smoluchowski
References equation:
Bowen WR, Mohammad AW, Hilal N (1997) Characteri-

sation of nanoltration membranes for predictive Ez ez
purposes use of salts, uncharged solutes and atomic
DP Lo
force microscopy. J Membr Sci 126:91105
Bowen WR, Welfoot JS, Williams PM (2002) Linearized
transport model for nanoltration: development and
where DP is the pressure difference across the
assessment. AIChE J 48:760773
Martnez F, Martn A, Malfeito J, Palacio L, Prdanos P, membrane, is the solution viscosity, z is the
Tejerina F, Hernndez A (2002) Streaming potential zeta potential, and Lo is the solution conductivity
through and on ultraltration membranes: inuence of (Burns and Zydney 2000).
salt retention. J Membr Sci 206:431441
The distance over which the electrical potential
Otero JA, Mazarrasa O, Villasante J, Silva V, Prdanos P,
Calvo JI, Hernndez A (2008) Three independent ways varies in an electrolyte solution is described by the
to obtain information on pore size distributions of Debye length (lD), which provides a measure of
nanoltration membranes. J Membr Sci 309(12):17 the thickness of the electrical double layer (Hunter
Rashin AA, Honig B (1985) Reevaluation of the born
2001):
model of ion hydration. J Phys Chem
89(26):55885593
Silva V, Martn A, Martnez F, Malfeito J, Prdanos P, 1=2
Palacio L, Hernndez A (2011) Electrical characteriza- F2 X 2
tion of NF membranes. A modied model with charge lD Z i Ci
eRT
variation along the pores. Chem Eng Sci 66:28982911
Yaroshchuk AE (2000) Dielectric exclusion of ions from
membranes. Adv Colloid Interf Sci 85:193230 where F is Faradays constant, R is the ideal gas
constant, T is the absolute temperature, and zi and
Ci are the valence and concentration, respectively,
of each ion in solution.
Dielectric Permittivity
Andrew L. Zydney References

Department of Chemical Engineering,
The Pennsylvania State University, University Burns DB, Zydney AL (2000) Buffer effects on the zeta
Park, PA, USA potential of ultraltration membranes. J Membr Sci
172:3948
Cosgrove T (2010) Colloid science: principles, methods,
and applications. Wiley, Chichester
The dielectric permittivity () provides a measure Hunter RJ (2001) Foundations of colloid science. Oxford
of the ability of a material to be polarized in the University Press, New York
Differential Scanning Calorimetry in Membrane Characterization 543
Differential Scanning Calorimetry in Membrane

Dielectric Spectroscopy (DS) Characterization, Table 1 Thermal transitions and
related parameters measurable by DSC
Impedance Spectroscopy Phase

Event transition Energy sign Parameter
Fusion sol!liq Endothermic l, DH
Crystallization liq!sol Exothermic l, DH
Evaporation liq!gas Endothermic l, DH
Differential Scanning Calorimetry
in Membrane Characterization
Condensation gas!liq Exothermic l, DH D
Sublimation sol!gas Endothermic l, DH
Deposition gas!sol Exothermic l, DH
Annarosa Gugliuzza Glass Change of Cp
Institute on Membrane Technology, National transition the baseline
Research Council of Italy, ITM-CNR, Rende, Chemical Exothermic/ DH
Italy reactions endothermic
The differential scanning calorimetry (DSC) is a chemical reactions. For each thermal event, ther-
thermodynamic technique that measures the dif- modynamic magnitudes such as latent heat (l),
ference in heat ow between sample and related specic heat (Cp), and/or a variation of enthalpy
reference against time or temperature, while pro- (DH) can be calculated (Table 1).
gramming the sample temperature under specic The transitions studied for polymers and com-
atmosphere conditions (Haines 1995). The sample pounds at low molecular weight can also be exam-
is contained in a pan and the reference is the ined for related membranes (Gugliuzza and Drioli
related empty pan. There are two kinds of calori- 2003). DSC plays a relevant role in materials
metric techniques, including: science and membrane technology, because it per-
mits to identify the best-performing materials
Heat ux DSC, where sample and reference are through the identication of structure-transport
heated in the same furnace and a difference of relationships that form the base of the membrane
temperature (DT) is measured. Successively, performance (Gugliuzza and Drioli 2004). His-
the signal is converted to a power difference tory, molecular rearrangements, and interactions
by the calorimetric sensitivity. are some critical issues to be taken in account
Power-compensated DSC, where both the when a membrane is processed (Yang
sample and reference are heated in two inde- et al. 2000). In this respect, DSC analysis can
pendent heaters, while the electrical power provide useful indication about structural features
DP dDQ=dt, requested to maintain close of the samples enabling the design of new inter-
to zero the difference of temperature between faces with superior performance in prospect.
sample and reference, is measured in real time. Figure 1 shows typical transitions studied on
polymeric membranes, including the glass transi-
The energy ux is associated with a transition tion (Tg) that is displayed as a change in the
or reaction and is graphed as a heat ow along the baseline due to a variation of Cp of the polymer
ordinate, while the temperature or time is the forming the membrane.
abscissa. In DSC curve, the endothermic peak is All amorphous polymers exhibit a Tg; below
associated with absorption of power, while the this temperature value, the membrane becomes
exothermic peak implies release of power from glassy owing to the inability of the polymer seg-
the sample. The verse of the peaks depends on the ments to rotate and rearrange themselves. When
sign of the heat ow. the membrane temperature is increased over its
DSC allows studying different events, includ- Tg, an increase in energy and free volume is
ing phase transitions, glass transition, and observed due to a broader mobility of the chains,
544 Diffusion Coefficient
Differential Scanning
Calorimetry
in Membrane
Characterization,
Fig. 1 Main endothermic
and exothermic transitions
observable by DSC
which enters the polymer in the rubbery References

region. Heating and cooling rates as well as
chain relaxation can affect Tg value and shape, De Lorenzo L, Tocci E, Gugliuzza A, Drioli E (2012)
Assembly of nanocomposite PEBAX membranes: a
respectively.
complementary study of afnity and clusterization phe-
The fusion is another endothermic event nomena. J Membr Sci 421422:7584
involving crystalline domains distributed through Gugliuzza A, Drioli E (2003) New performance of a mod-
the membranes. The amount and extent of these ied poly (amide-12-b-ethyleneoxide). Polymer
44(7):21492157
ordered regions depends on the chemistry, molec-
Gugliuzza A, Drioli E (2004) Role of additives in the water
ular weight, and thermal history of the samples. vapor transport through block co-poly (amide/ether)
Polymorphism and crystalline features of the membranes: effects on surface and bulk polymer prop-
membranes can be examined in relation to trans- erties. Eur Pol J 40/10:23812389
Haines PJ (1995) Thermal methods of analysis, principles,
port properties and mechanical stability as well. If
applications and problems. Blackie A&P, Oxford
the polymer forming the membrane is available to Yang H, Nguyen QT, Ding YD, Long YC, Ping Z (2000)
crystallize, a release of energy is associated with a Investigation of poly(dimethyl siloxane) (PDMS)solvent
decrease in heat ow. The study of crystallization interactions by DSC. J Membr Sci 164(12):3743
is often more useful than fusion, because it allows
distinguishing peaks that normally cannot be well
visible in melting.
Phase separation, miscibility, and compatibility Diffusion Coefficient
of polymer binary blends, copolymers, and, more
generally, composites are among the most interest- Boguslaw Kruczek
ing events to examine by DSC (De Lorenzo Department of Chemical and Biological
et al. 2012). In this case, the ability of the materials Engineering, University of Ottawa, Ottawa, ON,
to interact and distribute themselves in continuous Canada
and uniformly interlaced spaces can be evaluated
under different experimental conditions. Other
interesting indication about the membranes solidity Introduction
can be achieved from the study of curing, aging, as
well as degradation and oxygen stability. All these The diffusion coefcient (or diffusivity) is a
parameters can signicantly affect the membrane measure of mobility of a species (atom, molecule,
transport properties. or ion), which depends on the frequency with
Diffusion Coefficient 545
which a species moves and the size of each move- Molecular diffusivity of solutes in liquid is
ment. The magnitude of diffusion coefcient is estimated using well-established correlations
governed by the restricting forces of the medium (Seader et al. 2011). For large solutes that can be
in which diffusion takes place. The diffusion coef- approximated as rigid spheres diffusing at innite
cient is typically associated with, but not limited dilution through a stationary solvent, Di is
to, Ficks laws of diffusion. It is generally pre- predicted from the theoretical equation of
scribed to a pair of species. For a multicomponent Stokes-Einstein. On the other hand, the diffusion
system, it is prescribed for each pair of species in coefcients of small solutes at innite dilution are
the system, and diffusion in multicomponent sys- estimated using the empirical equation of Wilke- D
tems is described by the Stefan-Maxwell equa- Chang or more modern Hayduk and Minhas cor-
tion. In a binary system either species can be relations. In the case of dissolved salt, acid, or
mobile, or one species can be mobile while the base, the diffusing entities are ions. In the absence
other immobile. The example of the latter is dif- of an electric potential, the innite-dilution diffu-
fusion in porous and nonporous membranes. The sivity of single salts in an aqueous solution is
diffusion coefcient in membrane governs the estimated from the Nernst-Haskell equation.
membrane permeability and selectivity.
Diffusion Coefficient of Gas in Porous

Diffusion Coefficient of Liquid in Porous and Microporous Membranes
Membrane
The diffusion coefcient of gas in porous mem-
In the case of liquids, the effective diffusion coef- branes depends on the mean free path of the gas
cient of solute i in membrane (Dei) is related to molecules (l):
the ordinary molecular diffusion coefcient (Di)

in liquid by (Seader et al. 2011) 3m pRT 1=2
l (4)
2P 2M
eDi
Dei K ri (1)
t where m is the gas viscosity, T is the absolute
temperature, R is the universal gas constant, P is
where e and t are membrane porosity and tortu- the pressure, and M is the molecular weight of the
osity, and Kr is the restrictive factor, which gas. When dp/l << 1, i.e., the collisions with
depends on the solute diameter (dm) and the pore pore wall are much more frequent than the colli-
diameter (dp): sions with other gas molecules, the transport is
dominated by Knudsen diffusion, and the

dm 4 dm Knudsen diffusion coefcient of species i in a
Kr 1 , 1 (2) straight cylindrical pore is given by (Seader
dp dp
et al. 2011)
When dm > dp, Kr = 0 and the solute cannot

diffuse through the pore, which is referred to as dp 8RT 1=2
size exclusion. In the absence of size exclusion DK i (5)
3 pMi
and a hydrostatic pressure gradient across the
membrane (the example of such process is diffu-
In the other extreme when dp/l >> 1, the ordi-
sive dialysis), the selectivity for the separation of
nary molecular diffusion prevails, and the
solutes i and j is
corresponding molecular diffusivity can be esti-
mated using a theoretical equation of Chapman
Di K ri
aij (3) and Enskog or a more practical semiempirical
Dj K rj
equation of Fuller, Schettler, and Giddings.
546 Diffusion Coefficient
When dp is comparable with l, the effective dif- in which the permeability coefcient is the
fusion coefcient of solute i is product of the diffusion coefcient and the sol-
ubility coefcient (Si):

e 1
Dei (6)
t 1=Di 1=DKi Pi Si D i (9)
In microporous membranes (dp <2 nm) Consequently, membrane selectivity is a product

and particularly in ultramicroporous membranes of the solubility-selectivity (aS) and the
(dp <0.7 nm), dm becomes comparable to dp, and diffusivity-selectivity (aD), where the latter is the
one of the two possible transport mechanisms is ratio of the diffusion coefcients. The diffusion
the activated Knudsen diffusion, and the coefcient along with Pi, and Si are considered to
corresponding diffusion coefcient is given by be the intrinsic property of the membrane material
(Shelekhin et al. 1995) (Kesting and Fritzsche 1993).
Nonporous membranes are generally formed
from synthetic organic polymers; they can be
d p 8RT 1=2 DEi
DAK i exp (7) also formed from metals. Diffusion in metal mem-
3 pMi RT
branes is limited to diatomic gases, which rst
where DEi is the activation energy for diffusion of dissociate and dissolve according to Sieverts
species i, which rapidly increases as dp Law and then diffuse as individual atoms through
approaches to dm. On the other hand, when dp >> the membrane (Burggraaf 1996). In the case of
dm, DEi approaches to zero and Eq. 7 simplies to polymer membranes, the solute does not dissoci-
Eq. 5. ate before dissolving in the membrane. Conse-
The DKi, Dei, and DAKi (57) are generally quently, the size of diffusing solutes in polymer
referred to as Dgas, which is related to the perme- membranes may vary considerably. In turn, since
ability coefcient (Pgas) by the diffusion coefcient strongly depends on the
molecular weight of solute, the diffusion coef-
Dgas cients in polymer membranes vary by many
Pgas (8) orders of magnitude. The effect of the molecular
RT
weight on the diffusion coefcient is stronger in
Consequently, the selectivity for the separation of glassy polymers than in rubbery polymers
gases i and j is solely determined by the ratio of (Kesting and Fritzsche 1993).
the diffusivity coefcients, which varies For a given family of polymers, the diffusion
depending on the pore size of the membrane. If coefcient of solute depends on the fractional free
molecular diffusion dominates, the membrane has volume of polymer (vf) according to (Baker 2012)
no selectivity. If Knudsen diffusion dominates,
the selectivity is limited to the square root of the B
Di Aexp (10)
ratio of the molecular weights of solutes j and i. vf
On the other hand, if the activated Knudsen dif-
fusion prevails, very high selectivities are possible where A and B are adjustable parameters. Despite
(Shelekhin et al. 1995). the fact that vf in glassy polymers, which ranges
from 15 % to 20 %, is greater than that in rubbery
polymers (1015 %), the diffusivity in rubbers is
Diffusion Coefficient in Nonporous generally greater than that in glasses. The diffu-
Membranes sion coefcients in rubbery polymers range from
1013 to 108 m2/s. The upper values of this range
The transport of species in nonporous membranes are comparable to those of molecular diffusion in
is described by solution-diffusion mechanism, liquids. The diffusion coefcients in glassy
Diffusion Dialysis 547
polymers range from 1018 to 108 m2/s. The

upper values are comparable and sometimes Diffusion Dialysis
exceed those in the rubbery polymers. This is
because of some glassy polymers; in particular Heiner Strathmann
polymers with intrinsic microporosity (PIMs) Universitt Stuttgart, Institute of Chemical
have extraordinary large vf (up to 35 %) despite a Process Engineering Stuttgart, Baden-
very rigid polymer backbone (Baker 2012). Gen- Wurttemberg, Germany
erally greater diffusivity in rubbery polymers
arises from the fact that portions of the polymer D
chains can freely move due to thermal motions, Diffusion dialysis with ion-exchange membranes
and segments of the polymer backbone can also is mainly used today to recover acids or bases
rotate around their axis. On the other hand, ther- from a mixture with salts. The process is based
mal motion in glasses is limited. Strong interac- on the fact that ion-exchange membranes gener-
tions between solute and polymer chains may ally show a high permeability for counterions
increase the chain mobility. As the concentration while being more or less impermeable for
of interactive solute increases, membrane plasti- co-ions. Exceptions are protons and hydroxide
cization may occur, leading to increase in the ions which can easily permeate both cation- and
diffusion coefcient of solute (Kesting and anion-exchange membranes. The transport mech-
Fritzsche 1993). anism in diffusion dialysis is more complex than
Diffusion in nonporous membranes is an acti- in conventional dialysis due to the electrostatic
vated process, and consequently, the diffusion interaction between positive and negative charges
coefcient in both glassy and rubbery polymers and the electroneutrality requirement (Kobuchi
increases with temperature according to et al. 1987). The process and equipment design
Arrhenius-type relation (Seader et al. 2011): is very similar to that used in conventional dialy-
sis. In diffusion dialysis, the driving forces for the

Ei transport of ions through the ion-exchange mem-
Di Ai exp (11)
RT brane are gradients in their chemical potentials
only. There is no external electrical potential
where Ai is a pre-exponential constant which applied. The principle of the process is illustrated
depends among other on the nature of the solute in 1 showing part of a diffusion dialysis stack for
and Ei is the activation energy for diffusion of the recovery of HCl from a salt mixture. It is
solute i. Eq. 11 is similar to Eq. 7. composed of anion-exchange membranes
forming cells in which in alternating series a
feed solution of an HCl/FeCl2 mixture and a strip-
References ping solution of water is introduced. Since anion-
exchange membranes are permeable for H+, Cl
Baker RW (2012) Membrane technology and applications.
Wiley, West Sussex
and H+ permeate the membranes, while the other
Burggraaf AJ (1996) Transport and separation properties cations (Fe2+) are retained (Fig. 1).
of membranes with gases and vapours. In: Burggraaf The mass transport in diffusion dialysis is
AJ, Cot L (eds) Fundamentals of inorganic membrane determined by the transport of ions and can be
science and technology. Elsevier Science, Amsterdam,
pp 331433
determined by the respective mass balance equa-
Kesting RE, Fritzsche AK (1993) Polymeric gas separation tions described in the literature. A typical appli-
membranes. Wiley, New York cation of diffusion dialysis is the recovery of acids
Seader JD, Henley EJ, Roper DK (2011) Separation pro- from spent pickling solutions used in the metal
cess principles. Wiley, Hoboken
Shelekhin B, Dixon G, Ma Y (1995) Theory of gas diffu-
industry. A typical diffusion dialysis plant is
sion and permeation in inorganic molecular-sieve shown in the process ow diagram of Fig. 2.
membranes. AIChE J 41(1):5867 A stack with 100200 cell pairs is operated with
548 Diffusion Dialysis
Diffusion Dialysis, Repeating cell unit

Fig. 1 Schematic drawing Salt + Acid Acid
Acid
illustrating the principle of
diffusion dialysis utilizing
+ + + +
anion-exchange membranes
+ HCl + FeCl + HCl + HCl +
to recover an acid from a 2
mixture with salts + + + +
+ + + +
H+ + H+ + H+ + H+ + H+
+ + + +
Cl Cl Cl Cl Cl
+ + + +
+ + + +
+ + Fe ++ + +
Water Salt + Acid Water
Diffusion Dialysis, Dialyser

Fig. 2 Flow diagram of the
diffusion dialysis process
Waste
solution
Flow
meter
Recovered
acid
Receiving Feed
solution solution
countercurrent ow of the feed and the stripping stripping solution and leaves the system as salt-
solutions, which allows high acid recovery rates. free product. The salt is retained and leaves the
The ow diagram shows a feed solution stack as waste. The main application of diffusion
containing an acid and a salt which often is the dialysis is in the treatment of spent metal surface
efuent of metal surface treatment procedure. pickling solutions (Strathmann 2010; Oh
This solution is fed by a pump from a reservoir et al. 2000).
into a dialyser stack which contains only anion-
exchange membranes. A stripping solution, which
References
generally is water, is pumped in countercurrent
ow into the stack. During the pathway through Kobuchi Y, Motomura H, Noma Y, Hanada F (1987)
the stack, the acid diffuses from the feed solution Application of ion-exchange membranes to recover
through the anion-exchange membrane into the acids by diffusion dialysis. J Membr Sci 27:173179
Diffusion in Membranes 549
Oh SJ, Moon SH, Davis T (2000) Effects of metal ions on whereas pore ow in meso- and macroporous
diffusion dialysis of inorganic acids. J Membr Sci membranes. Consequently, membrane processes
169:95105
Strathmann H (2010) Electrodialysis, a mature technology such as gas separation (GS), pervaporation (PV),
with a multitude of new applications. Desalination and reverse osmosis (RO) are governed by diffu-
264:268288 sion, or more precisely the solution-diffusion
mechanism, whereas ultraltration (UF) and
microltration (MF) by the pore ow mechanism.
Diffusion in Membranes D
Diffusion in Dense Membranes
Boguslaw Kruczek
Department of Chemical and Biological Polymeric membranes have generally a dense
Engineering, University of Ottawa, Ottawa, structure with no visible pores (they are also
ON, Canada referred to as nonporous membranes). Molecules
permeate through transient free volume elements
between polymer chains. A molecule entrapped in
Introduction a cavity between the polymer chains bounces
around the cavity walls and eventually jumps to
Diffusion is the net movement of a substance an adjacent cavity once transient opening for the
(atom, ion, or molecule) from a region of high molecule passage appears. Molecular dynamic sim-
concentration to a region of low concentration. In ulations, which allow calculations of the uctua-
the case of membranes, diffusion along with pore tions in the volume between polymer chains due to
ow is the main transport mechanism. Diffusion thermal motion, have been used to visualize diffu-
dominates in dense and microporous membranes sion of gases in polymers (Baker 2012). Figure 1
Diffusion in Membranes,
Fig. 1 Motion of a carbon
dioxide molecule in a
6FDA-4PDA polymer
Cavity
matrix (Reprinted from size
Smit et al. (1992).
Copyright Elsevier
(permission pending))
Jump length
25
From Smit et al. (1992)
Jump
20
Movement in cavity
Position ()
15 Movement in cavity
10
Jump length
5 Cavity size
0
0 20 40 60 80 100 120 140 160 180 200
Simulation time (picoseconds)
550 Diffusion in Membranes
shows motion of a carbon dioxide molecule in a where DEa is the activation energy for diffusion,
6FDA-4PDA polymer matrix. The molecule D* is a preexponential factor often related to as the
bounces around the cavity in the rst 100 ps and entropy of activation and the average jump length,
it does not move more than 5 , which is indicative and R and T are the universal gas constant and the
of the cavity size. Then, thermal motion moves a absolute temperature, respectively. Diffusion pro-
segment of the polymer chain allowing the mole- cess in which D follows Eq. 2 is referred to as an
cule to jump to an adjacent cavity, where it remains activated diffusion. The activation energy for dif-
until another movement of the polymer chain fusion in polymeric membranes is positive; thus,
allows it to jump to another cavity (Baker 2012). the greater the DEa the smaller the D; also it
The diffusion coefcient (D) in polymer mem- indicates that D increases with T, which is consis-
branes is therefore dened (Koros and Fleming tent with molecular dynamics simulations. In turn,
1993): DEa is directly proportional to a cohesive energy
density (CED), which is indicative of strength of
f l2d attractive forces between adjacent polymer
D (1) chains. The CED increases when polar and hydro-
6
gen bond groups are present in polymer chains
where f is the average jump frequency and ld is (Koros and Fleming 1993).
the average jump length. The smaller the molecule
the more frequent jumps between the cavities.
However, in addition to the size, the shape of the Diffusion in Microporous Membranes
molecule is also very important. For example,
carbon dioxide is known to be much more diffusive Microporous membranes have permanent path-
in polymer membranes than methane even though ways (pores) for transport of molecules. They
carbon dioxide has a slightly greater molar volume are generally made from inorganic materials
than methane. On the other hand, methane is a such as zeolites, carbonized polymers, and amor-
spherical molecule with a kinetic diameter of phous silica; however, organic membranes made
3.72 , while the molecule of carbon dioxide is from polymers with intrinsic microporosity
an oblate ellipsoid 5.25 long, but with the diam- (PIM), such as polytrimethylsilylpropyne
eter of only 3.03 . Consequently, as carbon (PTMSP), are also considered as microporous
dioxide molecule bounces around in a cavity it (Baker 2012).
occasionally becomes oriented in such a way that The three most common diffusion mechanisms
it can jump through a gap between the polymer in microporous membranes, shown in Fig. 2,
chains as small as 3.03 . Such a small gap include activated diffusion (small pores), surface
would not permit jumping of the methane molecule diffusion (intermediate pores), and Knudsen dif-
(Baker 2012). fusion (large pores). The exact pore size ranges for
As temperature increases, the thermal motions a given diffusion mechanism are determined
of polymer chains allowing the molecule to jump based on the difference (W) between the potential
from one cavity to the other become more fre- energy outside the pore (Eout) and the potential
quent, and also the molecule bounces faster with energy within the pore (Ein) (Thornton
a cavity, which means that once a transient open- et al. 2010).
ing of appropriate size becomes available, it will As a molecule approaches a pore opening, each
more likely jump into the adjacent cavity. Exper- atom in the gas molecule interacts with every
imentally observed dependence of D in polymeric atom from the pore wall through van der Waals
membranes on temperature follows the Arrhenius forces. The potential difference is also referred to
relationship (Burggraaf 1996): as a suction energy; because a positive
W translates to a force sucking the molecule
DEa from outside to inside of the pore (adsorption),
D D exp (2)
RT while a negative W results in a force repulsing the
Diffusion in Membranes 551
Diffusion in Membranes, Activated diffusion constriction energy barrier

Fig. 2 Diffusion
mechanisms in microporous Ea
membranes (Reprinted
from Thornton et al. 2010.
Copyright John Wiley &
Sons. (permission adsorption-site energy barrier
Surface diffusion
pending))
Es
Kundsen diffusion Direction and velocity
Diffusion in Membranes, 0.6

Fig. 3 Potential energy Activated Surface Knudsen
difference for a single 0.5 diffusion diffusion diffusion
oxygen molecule at the
entrance of a carbon 0.4
cylindrical pore (Reprinted
from Thornton et al. 2010. 0.3
Copyright John Wiley &
W (eV)
Sons. (permission 0.2

pending))
0.1
0
dmin dK
0.1
0.2
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Pore size()
molecule from the pore. The activated transport Knudsen diffusion dominates when d > dK.
takes place when W < 0, and the activation Table 1 summarizes dmin and dK for different
energy for diffusion in Eq. 2 corresponds to gases in carbon and silica pores with two different
DEa jW j (Thornton et al. 2010). pore geometries.
Figure 3 presents an example of the potential
energy difference as function of pore size (d) for a
single oxygen molecule entering a cylindrical car-
bon pore from outside environment, in which Activated Diffusion Versus Surface
Eout = 0. The pore diameter corresponding to Diffusion
W = 0 is referred to as a minimum pore size
(dmin), while the pore diameter corresponding to When transport is dominated by activated diffu-
W = RT represents Knudsen diameter (dK). Acti- sion, there is no adsorption and membrane perme-
vated diffusion occurs when d < dmin, surface ability (P) is solely governed by the diffusion
diffusion dominates when dmin < d < dK, while coefcient, i.e.:
552 Diffusion in Membranes
Diffusion in Membranes, Table 1 Minimum pore size for the activated transport (dmin) and minimum pore size for
Knudsen diffusion (dK) at room temperature (298 K). All pore sizes are given as the PAS pore sizea (Adapted from
Thornton et al. 2010)
Pore geometry
Cylindrical pore Slit-shaped pore
Carbon Silica Carbon Silica
Gas dmin () dK () dmin () dK () dmin () dK () dmin () dK ()
He 2.30 5.69 2.53 6.37 1.86 2.19 2.06 4.35
H2 2.57 8.82 2.79 9.79 2.10 6.52 2.31 7.22
CO2 2.95 12.30 3.17 13.35 2.46 9.27 2.65 10.12
O2 3.10 11.68 3.31 12.68 2.59 8.76 2.79 9.58
N2 3.27 11.49 3.48 14.46 2.75 8.60 2.94 9.39
CH4 3.49 13.74 3.69 14.84 2.95 10.43 3.13 11.32
CO 3.32 12.14 3.52 13.16 2.79 9.14 2.98 10.00
Ar 3.18 11.68 3.39 12.67 2.66 8.76 2.86 9.56
n-C5H12 5.27 23.51 5.45 25.04 4.59 18.33 4.75 19.57
C2H6 4.02 16.69 4.21 17.93 3.44 12.82 3.62 13.86
SF6 4.66 19.46 4.85 20.81 4.03 15.05 4.20 16.15
a
Positron annihilation spectroscopy (PAS) diameter d0 = d dd, where dd is the electron cloud thickness surrounding the
surface atoms (dd = 3.32 )

DEa jW j diffusion, which represents the energy barrier sep-
P P exp P exp (3) arating two adjacent sorption sites (Thornton
RT RT
et al. 2010). Since Ps = DsSs, combining Eqs. 4
where DEa is the activation energy for diffusion, and 5 leads to:
which is identical to DEa in Eq. 2, which is equal
to the absolute value of the potential difference.
1 aq
On the other hand, when transport is dominated by Ps Ps exp (6)
surface diffusion, membrane permeability (Ps) is RT
governed by both the diffusion coefcient (Ds)
and the solubility coefcient (Ss). The latter is a The term, (1 a)q is positive; therefore, Ps, unlike
thermodynamic factor that correlates the concen- P, decreases with temperature. In the case of dense
tration of molecules in the adsorbed phase to the polymeric membranes, the permeability coef-
external pressure, which typically follow a Lang- cient is also a product of the diffusivity and the
muir isotherm. At the same time, Ss follows solubility coefcients, and the latter, similarly to
Arrhenius relation: the surface diffusion solubility coefcient,
q decreases with temperature. On the other hand,
Ss K o exp (4) the corresponding increase in diffusivity over-
RT compensates a decrease in solubility. Thus, the
permeability coefcient in dense polymeric mem-
where q (>0) is a heat of adsorption, which implies
branes follows Arrhenius-type relationship simi-
that Ss decreases with temperature. Surface diffu-
lar to the permeability coefcient for activated
sion is an activated process, which depends on q:
diffusion in microporous membranes.
aq The third diffusion mechanism, Knudsen dif-
Ds Ds exp (5) fusion, is not an activated process. Although
RT
Knudsen diffusion coefcient increases with tem-
where a is a proportionality constant (0 < a <1). perature, it does not follow Arrhenius-type rela-
The product aq is the activation energy for surface tionship (Thornton et al. 2010).
Diffusive Flow 553
References total ux of component A (N) relative to a xed

frame of reference is
Baker RW (2012) Membrane technology and applications.
Wiley, West Sussex
Burggraaf AJ (1996) Transport and separation properties N A J A yA N J A y A N A N B (2)
of membranes with gases and vapours. In: Burggraaf
AJ, Cot L (eds) Fundamentals of inorganic membrane
science and technology. Elsevier Science, Amsterdam, where yA is a mole fraction of A in the binary
pp 331433 mixture at a given point. If component B is sta-
Koros WJ, Fleming GK (1993) Membrane-based gas sep-
tionary (NB = 0), for example, when A is D
aration. J Membr Sci 83:180
Smit E, Mulder MHV, Smolders CA, Karrenbeld H, van transported across a membrane, then
Eerden J, Feil D (1992) Modeling of the diffusion of
carbon dioxide in polyimide matrices by computer
simulation. J Membr Sci 73:247257 J A N A 1 y A (3)
Thornton AW, Hill JM, Hill AJ (2010) Modelling gas
separation in porous membranes. In: Yuri
Ficks rst law of diffusion is limited to binary
Yampolskii Y, Freeman B (eds) Membrane gas separa-
tion. Wiley, Hoboken, pp 85109 systems. Therefore, if diffusive ow of a binary
mixture takes place across a membrane, counting
the membrane as a stationary component, the sys-
tem is ternary. For ternary and more generally
multicomponent systems, diffusive transport is
Diffusive Flow described by the Stefan-Maxwell equation. Diffu-
sive ow is the main transport mechanism in
Boguslaw Kruczek several membrane separation processes, including
Department of Chemical and Biological gas separation, pervaporation, reverse osmo-
Engineering, University of Ottawa, Ottawa, ON, sis, and diffusive dialysis.
Canada
Driving Force for Diffusive Flow

Introduction
Ficks rst law implies that the driving force for
Diffusive ow ( diffusive transport) refers to the diffusion is the concentration gradient. In
transfer of matter by diffusion. Quantitative study the simplest example of diffusive ow in mem-
of diffusive ow dates back to the nineteenth branes Knudsen diffusion, which occurs in
century to the works of Thomas Graham and porous membranes in which the collisions
Adolf Fick. The latter was formulated to what is between gas molecules and pore walls are more
now known as Ficks rst law of diffusion: frequent than those between the gas molecules
the concentration gradient is expressed in terms of
@c the partial pressure gradient for binary and
J D (1)
@z multicomponent system and as a total pressure
gradient for the single gas transport. However, in
where J is the diffusive ux, D is the diffusion the more rigorous analysis, the true driving force
coefcient, and c is the concentration of species is the gradient of chemical potential. Assuming
transported by diffusive ow and z is a coordinate applicability of the ideal gas law, a more general
along which the net diffusive transport occurs. expression for the ux of component A is given by
The diffusive ow is therefore a concentration Krger et al. (2012):
gradient-driven process, with the net mass transfer
from high concentration to low concentration RT dlnpA dcA
region. Diffusive ow may occur along with a J A uA cA (4)
f dlncA dz
convective ow. For a binary system (A, B), the
554 Diffusive Flow
where uA is the ow velocity of component A, f is is that the later takes into consideration the interac-
a friction coefcient, R is the universal gas con- tions between components i and j, which in dilute
stant, T is the absolute temperature, and pA is the systems are negligible. Consequently, the Fickian
partial pressure of component A. The term d ln and Maxwell-Stefan diffusion coefcients are
pA/d lncA represents the gradient of the equilib- equivalent in dilute ideal systems. On the other
rium isotherm in logarithmic coordinates, which hand, they deviate from each other in concentrated
is a thermodynamic correction factor. This term systems. However, Fickian diffusion coefcient is
may vary substantially with concentration, but in known to be a function of concentration in concen-
general, at low concentrations (i.e., when cA is a trated systems.
linear function of pA), it approaches unity. Com-
parison of Eq. 1 with Eq. 4 reveals that
Solution-Diffusion Model
RT dlnpA
DA (5)
f dlncA Diffusive ow or diffusion is a rate-limiting step
in membrane processes such as gas separation,
This expression for the diffusion coefcient arises pervaporation and reverse osmosis, in which the
from the interpretation of the diffusive ow as transport is governed by the solution-diffusion
being driven by the gradient of chemical potential model (Baker 2012). In this model, it is assumed
and opposed by frictional forces. that penetrant in feed and in feed-side interface of
the membrane and the penetrant in the permeate-
side interface and permeate are in instantaneous
Maxwell-Stefan Equations equilibrium. The corresponding equilibrium
constant, which is referred to as distribution
The Maxwell-Stefan model was originally devel- coefcient or a solubility coefcient (S), is a ther-
oped to describe diffusion in homogeneous gas or modynamic factor. The solubility coefcient is
liquid phase. It considers the diffusion coefcients constant only at low concentrations where the
as inverse drag coefcients representing the inter- concentration of penetrant in the membrane is a
change of momentum between the different types linear function of its concentration in the uid
of molecules. The general form of the Maxwell- phase (e.g., Henrys law is applicable). In this
Stefan equation for diffusion in a multicomponent case, the diffusive ux can be expressed in terms
mixture is (Krger et al. 2012) of the external driving force, for example (Koros
and Fleming 1993),
ci @mi X ci J j cj J i
(6)
RT @z j6i
cDij pf pp pf pp
J SD P (8)
L L
where Ji = uici is the diffusive ux of the i-th
component, Dij represents the Stefan-Maxwell where P = SD is the permeability coefcient, pf and
diffusivities, and c is the total concentration. For pp are the partial pressures of penetrant at the feed
an ideal gas, the left-hand side of Eq. 6 reduces to and permeate sides, respectively, and L is the mem-
@ci =@z , so for equimolar counter-diffusion in a brane thickness. Equation 8 is applicable for gas
binary system permeation; however, similar rate equations can be
written for reverse osmosis and pervaporation.
@ci
J i J j
Dij (7)
@z
Dual-Mode Sorption
which is identical with the usual Fickian formula-
tion given by Eq. 1. The difference between the Most of practical membranes are made from
Fickian and Maxwell-Stefan diffusion coefcients glassy polymers in which sorption does not follow
Diffusivity 555
linear isotherm according to Henrys law, but quantities, thus making this approach more suit-
rather the so-called dual-mode sorption in able for use in the correlation of experimental and
which (Kesting and Fritzsche 1993) predictive models (Krger et al. 2012). This
approach has become quite popular for the analy-
c0H bp sis of diffusive transport in inorganic adsorbent
c cD cH Sp (9)
1 bp membranes, in which adsorption is described by,
but not limited to, Langmuir isotherm, largely
where cD is the concentration in the Henry-type because of the work of Krishna and his associates
sites, cH is the concentration in the Langmuir sites, (Wesselingh and Krishna 2000). Nevertheless, D
b is the hole afnity constant, and cH0 is the hole because of its simplicity compared to the
saturation constant, which is a measure of the Maxwell-Stefan formalism, solution-diffusion
sorption capacity of the unrelaxed volume. Equa- model based on Fickian diffusion is still by far
tion 9 is essentially the summation of Henry and the most commonly used approach to analyze
Langmuir isotherms. The molecules in Langmuir diffusive transport in membranes.
sites are partly or completely immobilized; how-
ever, the two populations i.e., those in Henry sites
and those in Langmuir sites, are in instantaneous References
equilibrium. Assuming partial immobilization of
molecules in Langmuir sites leads to the dual- Baker RW (2012) Membrane technology and applications.
Wiley, West Sussex
mobility model, which accounts for two distinct Krger J, Ruthven DM, Theodorou DN (2012) Diffusion in
molecular environments with different diffusion nanoporous materials. Wiley-VCH, Weinheim
coefcients. Consequently, Ficks rst law for Kesting RE, Fritzsche AK (1993) Polymeric gas separation
dual-mobility model becomes membranes. Wiley, New York
Koros WJ, Fleming GK (1993) Membrane-based gas sep-
aration. J Membr Sci 83:180
@cD @cH Wesselingh JA, Krishna R (2000) Mass transfer in
J D DH (10)
@z @z multicomponent mixtures. Delf University Press, Delf
Zolondz RR, Fleming GK (1992) Gas permeation. In: Ho
where D and DH are the diffusion coefcients in WSW, Sirkar KK (eds) Membrane handbook. Van
Nostrand Reinhold, New York, pp 19101
Henry and Langmuir sites, respectively. Introduc-
ing F = DH/D, which varies from 0 (complete
immobilization) to 1 (no immobilization in Lang-
muir sites), and substituting the expressions for
Henry and Langmuir isotherms, Eq. 10 can be Diffusive Transport
rearranged to Kesting and Fritzsche (1993):
Diffusivity
!
FK @cD
J D 1 (11)
1 acD 2 @z
Equations 9, 10, and 11 are applicable to single gas Diffusivity

permeating through the glassy polymeric mem-
branes. They are easily extendable to binary and Renzo Di Felice
multicomponent systems within the classical Fickian University of Genova, Genova, Italy
diffusion formalism (Zolondz and Fleming 1992),
without resorting to Maxwell-Stefan equations.
However, the Maxwell-Stefan formulation has Synonyms
the important advantage that the rate parameters
are simply related to directly measurable Diffusive transport; Molecular diffusivity
556 Digitized Structure Model
For a species in a homogeneous phase, gas, liquid,

or solid, diffusivity can be dened as any material Digitized Structure Model
transport driven by concentration difference. Its
physical origin lies in the random molecular motion Vasilis Burganos
that leads to an overall phenomenon of smoothing Institute of Chemical Engineering Sciences,
concentration gradients: any batch systems origi- Foundation for Research and Technology
nally showing concentration difference for one or Hellas, FORTH/ICE-HT, Patras, Hellas
more species will eventually be completely mixed if
enough time is elapsed. Diffusivity is easily quanti-
ed by the diffusivity ux, i.e., by how much a Digital Reconstruction
species diffuses per unit area per unit time.
There are various practical situations which for It is a procedure to represent the internal structure
different reasons, both theoretical and practical, of materials in digital form, usually, binary, in
have received particular attention and have been three dimensions using data from microphoto-
investigated in depth. A brief summary is given graphs or other sources of spatial correlation
here (Cussler 1997). data. It has contributed signicantly to the visual
comprehension of the internal conguration of
Molecular diffusivity. In this case, the uid porous materials, human tissues, and composite
where the phenomenon is taking is considered materials. In the case of porous materials, like
unbounded so that there is no physical interac- porous membranes, catalyst supports, rocks, etc.,
tion between the diffusing species and the sys- a binary representation in three dimensions
tem boundaries. Moreover, the species are assigns the value of 0 (or 1) to the voxels that lie
considered to be diluted so that any interaction in the void space and the value of 1 (or 0) to the
between the diffusing molecules can also be voxels in the complementary space (solid)
neglected. This is probably the most studied (Torquato 2002, 2010). In this way, one obtains
cases, and many systems can be considered to a fully three-dimensional array of ags (0 or 1)
fall in these categories. that denes in a discretized form the pore space
Diffusivity of strong electrolytes. As before here inside the material. This array can then be utilized
we considered the system to be unbounded. in simulations of sorption and transport phenom-
However, there is an interaction between the ena (Gelb 2009), typically, diffusion, ow,
diffusing molecules which affects the transport two-phase ow, and dispersion that are of utmost
rate and must be taken into account. importance in the design and modeling of mem-
Diffusivity of associating solutes. The solute brane materials and membrane separation pro-
associates to form aggregates (weak electro- cesses. Raw data for the initiation of the
lytes or micelle), and again this affects the reconstruction procedure can be obtained in vari-
transport rate. ous ways. Physical serial sectioning along planes
Solvation. Solute-solvent interactions cannot that are normal to a xed direction has repeatedly
be ignored in this case. been used using special microtomes or surface
Solute-boundary interactions are important: here grinders following impregnation of the void
we can have diffusion in porous media or com- phase with some suitable substance (resin,
posite, diffusion in cylindrical pore (Knudsen or Woods metal, etc.). Digitization of the photo-
surface or capillary or molecular sieving). graphs of each and every section of the material
provides a set of two-dimensional arrays to be
combined in a deterministic fashion into a three-
References dimensional array that discretizes the working
sample of the medium. Computer tomography
Cussler EL (1997) Diffusion. 2nd edition, Cambridge Uni- provides the same type of information in an auto-
versity Press, Cambridge (UK) mated or semiautomated manner and enjoys
Digitized Structure Model 557
Digitized Structure
Model, Fig. 1 Stochastic
reconstruction of porous
material using the fractional
Brownian motion method,
showing individual volume
elements
tremendous use in medicine and some types of interweaving option that allows three-
porous media, most typically in hydrocarbon res- dimensional reconstruction (Fig. 1) at much
ervoirs. More recently, stochastic techniques are greater scale than the correlation length of the
attracting the increasing interest of investigators structure (Kikkinides and Burganos 2000). For
in this area inasmuch as they require a single materials that have evolved from powders through
section only rather than a whole series of sections. some agglomeration or sintering process, ballistic
Provided that the sample is sufciently isotropic or random placement methods can be used that
and homogeneous, one can extract valuable geo- allow mass transfer between coalescing particles
metric and topological data from the section and at a controlled fashion so as to reproduce the
reconstruct in three dimensions a discretized form actual grain-size distribution and porosity or
of the material that respects the features that char- solid fraction. Reconstruction of ber-type
acterize the two-dimensional image of the section materials can proceed in a similar manner. In
(Quiblier 1984; Adler 1992). Such features the case of porous membranes that contain
include various moments of the correlation func- nanoscale pores, scattering techniques can be
tion, typically porosity and autocorrelation func- used that offer in an implicit manner the autocor-
tion, but also higher-order moments or lineal- relation function, thus sidestepping the stage of
length distribution, chord-length distribution, microscopy and the subsequent phase of image
etc. The latter is succeeded by the simulated analysis for the extraction of two-dimensional
annealing technique (Kirkpatrick et al. 1983) features.
that, theoretically, is capable of reproducing
three-dimensional descriptions of materials from
an arbitrary number of moments of the correlation
References
function at the expense, of course, of heavy
requirements in computational time. A simpler Adler PM (1992) Porous media- geometry and transports.
procedure that has proven efcient for several Butterworth-Heinemann, Stoneham
types of porous membranes including asymmetric Gelb LD (2009) Modeling amorphous porous materials
and conned uids. MRS Bull 34:592601
ones is the so-called fractional Brownian motion
Kikkinides ES, Burganos VN (2000) Permeation proper-
technique that respects the rst two moments of ties of three-dimensional self-afne reconstructions of
the correlation function and, in addition, offers an porous materials. Phys Rev E 62:69066915
558 Diluant
Kirkpatrick S, Gelatt CD, Vecchi MP (1983) Optimization

by simulated annealing. Science 220:671680 Dip-Coating Method for Ceramic
Quiblier JA (1984) A new 3-dimensional modeling tech-
nique for studying porous-media. J Colloid Interface Membrane Preparation
Sci 98:84102
Torquato S (2002) Random heterogeneous Zhaoliang Cui
materials microstructure and macroscopic properties. State Key Laboratory of Materials-Oriented
Springer-Verlag, New York
Torquato S (2010) Optimal design of heterogeneous mate- Chemical Engineering, College of Chemistry and
rials. Annu Rev Mater Res 40:101129 Chemical Engineering, Nanjing Tech University,
Nanjing, China
Dip-coating method for ceramic membrane

Diluant preparation is one of the most important methods
for ceramic membrane preparation. It is usually
Greg Foley employed to form the intermediate layer of a
School of Biotechnology, Dublin City University, ceramic membrane (Lindqvist and Lidn 1997).
Dublin, Ireland The procedure is illustrated in Fig. 1 (Li 2007).
The drying process starts simultaneously when
the substrate is in contact with an atmosphere
with a relative humidity below 100 %. In a multi-
Any dialtration process involves the replace- ple step process, after calcination of the rst layer,
ment of solvent (including undesired impuri- the complete cycle of dipping, drying, and calci-
ties) with a liquid of desired composition. This nation is repeated. The raw material is a more
liquid is known as the diluant. For example, in
constant volume dialtration (CVD), also
known as continuous dialtration, diluant is
added to a tank of unpuried solution while it
undergoes batch ultraltration. The rate of
diluant addition is set equal to the permeate
ow rate, and thus the solution (retentate) vol-
ume remains constant. The effect of this process
is the gradual replacement of the solvent and
any solutes with low rejection coefcients with
the diluant. Mathematically, the reduction in
concentration of any solute, with rejection coef-
cient, s, is given by
c c0 e1sV d =V s
where Vd is the volume of diluant added and Vs is

the solution (retentate) volume. The ratio Vd/Vs is
known as the dialtration factor.
The diluant may be puried water or, more Dip-Coating Method for Ceramic Membrane Prepa-
likely, a buffer. ration, Fig. 1 Dip coating (Li 2007)
Direct Contact Membrane Distillation (DCMD) 559
interactive powder than the materials used for the pore-distillate interface on the other side of the
support and the porosity is governed by the choice membrane at a lower temperature (Fig. 2).
of sintering temperature. The viscosity of the par- Among the possible MD congurations,
ticle suspension and the coating speed or time are DCMD is best suited for applications such as
the critical factors in dip-coating process desalination or the concentration of aqueous solu-
(Lindqvist and Lidn 1997). By this method, tions (orange juice), in which water is the major
pore size of the coating layer is possibly smaller permeating component. Advantages and draw-
than 200 nm, and the process can be conducted backs of DCMD are similar to those of other
continuously (Li 2007). MD congurations (see Membrane Distillation D
for a detailed description of benets and disad-
vantages of DCMD) and summarized briey as
References follows: theoretically 100 % rejection of nonvol-
atile solutes, lower operating temperature with
Li K (2007) Ceramic membranes for separation and reac- respect to thermal evaporation processes, and
tion. Wiley, Chichester
reduced membrane fouling problem with respect
Lindqvist K, Lidn E (1997) Preparation of alumina mem-
branes by tape casting and dip coating. J Eur Ceram Soc to pressure-driven membrane processes, weakly
17:359366 inuenced by feed concentration. Therefore, this
thermal technique is relatively free from ux
reduction by concentrating feed, which is inevita-
ble in the reverse osmosis (RO)-based process.
Consequently, the DCMD process is an appropri-
Direct Contact Membrane ate alternative for achieving high water recovery
Distillation (DCMD) from brackish/seawater feeds, managing the con-
centrate from desalination processes including
Francesca Macedonio
RO and recovering water from industrial brines.
Heat and mass transfer phenomena occurring
in DCMD process are discussed in various recent
Italy
publications (Lawson and Lloyd 1997; Drioli
et al. 2015; Schoeld and Fane 1987; Martnez-
Engineering, University of Calabria, Rende,
Diez and Vzquez-Gonzlez 2000; Gryta
Italy
et al. 1997; Ali et al. 2013 and see Membrane
Distillation (MD)). The ux N can be
Direct contact membrane distillation (DCMD) is represented by the following simple correlation:
one of the four main possible membrane distilla-
tion (MD) congurations. In DCMD process, an N C Pfm Ppm
aqueous solution colder than the feed aqueous
solution is maintained in direct contact with the where C is the membrane characteristic parameter
permeate side of the membrane. Both the feed and and can be calculated by using various models,
permeate aqueous solutions are circulated tangen- depending upon the membrane features and oper-
tially to the membrane surfaces (Fig. 1). In this ating temperature applied. DP Pfm Ppm repre-
MD mode, the transmembrane temperature differ- sents the vapor pressure difference across the
ence induces the required water vapor pressure membrane (function of temperatures and compo-
difference between both membrane sides. Conse- sitions at the membrane surface at feed f and per-
quently, water molecules evaporate at the hot liq- meate p side, respectively). Mass transfer in direct
uid/vapor interface, diffuse in vapor phase contact membrane distillation can be separated
through the pores, and are condensed at the into three steps (e.g., mass transfer in feed
560 Direct Contact Membrane Distillation (DCMD)
Direct Contact
Membrane
Distillation (DCMD),
Fig. 1 Typical DCMD
schema
Direct Contact
Membrane Tf
Distillation (DCMD), vapour
m
Fig. 2 DCMD process Tf
Tpm
vapour
Tp
FEED vapour DISTILLATE

or or
RETENTATE PERMEATE

boundary layer, mass transfer across the mem- Qf Qp ) hf T f T fm hp T pm T p (2)
brane, and mass transfer in permeate boundary
layer, as in membrane distillation). The mass where Qf and Qp represent the heat transferred at
transfer in permeate boundary layer is not taken bulk feed and permeate side, respectively. More-
into account since the mole fraction of the over, each h and each T represent the
transporting species in the permeate stream is corresponding heat transfer coefcients and tem-
approximately equal to one. For what concerns peratures shown in Fig. 2.
the mass transfer in boundary layers, it is general The real challenge is to calculate the tempera-
to neglect surface diffusion and viscous ow and tures at membrane surfaces (Tfm and Tpm). To
to employ a Knudsen-molecular diffusion transi- realize this, various approaches have been used
tion model (Lawson and Lloyd 1997; Srisurichan (Schoeld and Fane 1987; Martnez-Diez and
et al. 2006). Vzquez-Gonzlez 2000; Gryta et al. 1997).
For what concerns heat transfer, the net heat Schoeld (Schoeld and Fane 1987) suggests
transported through conduction Qc and due to the following approach valid when transmem-
evaporation Qv in DCMD can be calculated by brane surface temperature differences are not
using the following relationships (Drioli greater than 10 C. In this case the pure water
et al. 2015): ux can be represented by
Km dP
Q Qc Qv T fm T pm Jl (1) NC jT T fm T pm (3)
d dT m
where dPdT jT m can be expressed according to

at steady state Clausius-Clapeyron equation:
Direct Contact Membrane Distillation (DCMD) Applications 561
dP MlP polarization has been investigated by directly

jT jT (4)
dT m RT 2 m monitoring the bulk and interfacial temperatures.
The authors have concluded that heat transfer
where M is the molar mass of water molecules. coefcient can be predicted by using the relation-
Assuming the same temperature polarization on ship predicted by (Gryta et al. 1997).
up- and downstream,
Tf Tp References
Tm (5) D
2
Ali A, Macedonio F, Drioli E, Aljlil S, Alharbi O a
Thus Eq. 1 can be written in the following form: (2013) Experimental and theoretical evaluation of tem-
perature polarization phenomenon in direct contact
membrane distillation. Chem Eng Res Des
Km dP
Q C jT m l T f T fm (6) 91(10):19661977
d dT Drioli E, Ali A, Macedonio F (2015) Membrane distilla-
tion: recent developments and perspectives. Desalina-
tion 356:5684
Q H T fm T pm : (7) Gryta M, Tomaszewska M, Morawski AW (1997) Mem-
brane distillation with laminar ow. Sep Purif Technol
Combining Eqs. 2, 6, and 7 11:93101
Lawson KW, Lloyd DR (1997) Membrane distillation.
J Membr Sci 124:125
Tf Tp
T fm T pm (8) Martnez-Diez L, Vzquez-Gonzlez MI (2000) A method
H H to evaluate coefcients affecting ux in membrane
1
hf hp distillation. J Membr Sci 173:225234
Schoeld RW, Fane AG (1987) Heat and mass transfer in
membrane distillation. J Membr Sci 33:299313
T fm T pm y T f T p (9) Srisurichan S, Jiraratananon R, Fane AG (2006) Mass
transfer mechanisms and transport resistances in direct
from Eqs. 3 and 8, contact membrane distillation process. J Membr Sci
277:186194

DT 1 1 K m =dm 1
1 (10)
Nl dP=dT lC h h
Direct Contact Membrane Distillation
The permeability
C and overall heat transfer coef- (DCMD) Applications
cient h 1=hf 1=hp can be determined by
plotting DT 1
Nl versus dP=dT as illustrated by Eq. 10 Maria Tomaszewska
(Drioli et al. 2015; Schoeld and Fane 1987). Faculty of Chemical Technology and
Equation 1 was originally developed for Engineering, West Pomeranian University of
Tfm Tpm 10o C and for pure water as feed Technology, Szczecin, Poland
and permeate. A more general approach for the
determination of up- and downstream heat trans-
fer coefcients is based on the use of various DCMD is a membrane distillation (MD) congu-
empirical correlations. Recently some ambitious ration in which both solutions, feed and permeate,
attempts have been observed in measuring the are in a direct contact with a hydrophobic porous
membrane surface temperature directly by using membrane. Thus, the water vapor transferred
different techniques. For example, Ali across the membrane is directly condensed in a
et al. (2013) have used a cell equipped with cold permeate inside the membrane module. This
16 temperature sensors to measure the tempera- conguration has been the most frequently stud-
ture proles on feed and permeate side in DCMD. ied due to a simple operation mode in comparison
The effect of different parameters on thermal with other MD congurations. DCMD is
562 Direct Contact Membrane Distillation (DCMD) Applications
Sea MF/UF NF RO Fresh

water water
Fresh water membrane MD

crystallizer (crystallizer)
NaCl
NaCl, CaCO3,
MgSO4*7H2O
Direct Contact Membrane Distillation (DCMD) Appli- crystallization units operating on the brine from NF and/or
cations, Fig. 1 Scheme of sea water desalination plant RO stages (Based on Di Proo et al. (2010)).
using integrated membrane system containing membrane
considered as the best choice for applications in (Fig. 1). An application of the DCMD/MCr sys-
which water is the major component of the feed- tem allows to recover both water and salts
ing solution. DCMD can be applied in chemical, (commercial products), reducing the environmen-
food, or pharmaceutical industry and in environ- tal problems related to brine disposal (Drioli
mental protection. et al. 2011; Di Proo et al. 2010).
The major area of DCMD application is desali- A DCMD system utilizing solar and wind
nation focused on the production of high purity energy for water supply, designed for small
water and freshwater from sea or brackish water. board and rural or coastal areas with a lack of
The nature of DCMD driving force causes that electricity, is presented in Fig. 2 (Susanto 2011).
nonvolatile solutes such as salts, colloids, and mac- DCMD has also been applied to surface or
romolecules that are almost completely retained by a groundwater treatment. It was found that mem-
membrane. The permeate is pure water with electri- brane scaling by CaCO3, CaSO4, and their mix-
cal conductivity of 0.22.5 mS/cm, which is suitable ture occurring in the system can be signicantly
for medical, pharmaceutical, or semiconductor limited by the feed acidication or antiscalant
industry purposes. Lower temperatures than those addition. Moreover, it was found that scaling in
usually used in the conventional distillation allow to DCMD is less severe than in conventional desali-
achieve the permeate uxes up to 80 kg/m2h with nation processes (Khayet and Matsuura 2011).
signicant reduction in energy costs. Typical feed DCMD can be operated with solutions of
temperatures (3080 C) permit the efcient recycle extremely high concentration of salts, even close
of low-grade or waste heat streams, as well as the to the supersaturated point, and the permeate ux
use of alternative energy sources (solar, wind, or was satisfying. Therefore, DCMD has also been
geothermal). In addition, the possibility of using successfully applied to wastewater treatment,
plastic equipment also reduces corrosion problems either to produce a permeate less hazardous to
(Drioli et al. 2005; Khayet and Matsuura 2011). the environment or to obtain a retentate that con-
Integration of DCMD with pressure-driven tains high concentration of valuable chemicals
membrane processes such as microltration (Table 1) (Khayet and Matsuura 2011).
(MF) or reverse osmosis (RO) creates the oppor- The main disadvantage of DCMD process is a
tunity for enhancement of water productivity from problem with membrane wetting, what limits the
seawater up to 87.6 %. A development of the concentration of organic compounds to diluted
system coupling nanoltration (NF) and RO aqueous solutions. The membranes used in MD
with DCMD/membrane crystallizer (MCr) might are chemically resistant; thus, they can tolerate
lead to an increase in the total water recovery harsh operating conditions. DCMD is a promising
factor of a desalination plant up to 92.8 % method for recovery of HCl from metal pickling
Direct Contact Membrane Distillation (DCMD) Applications 563
Direct Contact Membrane Distillation (DCMD) Applications, Fig. 2 Schematic diagram of DCMD powered by
solar and wind energy (Reprinted from Susanto (2011), copyright (2011), with permission from Elsevier)
Direct Contact Membrane Distillation (DCMD) Applications, Table 1 Examples of DCMD application in various
fields
Application area Examples of application
Water treatment Desalination of seawater and brackish water, removal of various contaminants (heavy metals,
arsenic, boron, humic acid) from surface and groundwater, production of high purity water,
removal of low concentrations of organics (ethanol, acetone, acetonitrile, n-butanol)
Wastewater Treatment of low- and medium-level radioactive wastewater, textile wastewater contaminated
treatment with dyes, oil-water emulsions, HCl removal from etching solutions from electrochemical
industry, separation of photocatalyst in the photocatalysis-DCMD system, reclamation in space
Medicine and Concentration of an extract of traditional Chinese medicine
biology Concentration of bovine plasma and bovine blood, bovine serum albumin, to ameliorate treatment
of uremia
Pharmacy Pharmaceutical wastewaters containing taurine
Food industry Concentration of fruit and juices, whey, sugarcane, etc.
Biotechnology Removal of toxic products from culture broth, production of ethanol in a bioreactor-DCMD
system
Chemical Concentration of acids, separation of a half-nished product (HCl) from the reaction mixture to
industry shift the reaction equilibrium in DCMD-chemical reactor system
Other Ethylene glycol removal from coolant liquids, breaking azeotropic mixtures HCl-water, propionic
acid-water, salt crystallization in DCMD/MCr
564 Direct Contact Membrane Distillation (DCMD) Applications
water
Hydrochloric acid
Metal objects
purified
Etching bath Rinsing objects
Concentration
Concentration of MD II
spent pickling
solution MD I
Crystallisation
HCl solution
salt
HCl concentration water
MD III
Direct Contact Membrane Distillation (DCMD) Appli- solution using DCMD (Reprinted from Tomaszewska
cations, Fig. 3 The ow sheet diagram of integrated et al. 2001, copyright (2001) with permission from
processes of metal pickling with treatment of spent Elsevier)
Direct Contact KCl, H2SO4

Membrane Distillation
(DCMD) Applications,
Fig. 4 Scheme of the MD REACTOR
conversion of potassium KCl+H2SO4=
chloride into potassium
KHSO4+ HCl HCl
sulfate using integrated
Sulfuric acid
system: chemical reactor/ Retentate

DCMD (Based on
Tomaszewska and Mientka
2009) Permeate
Hydrochloric
CRYSTALLIZER acid
KHSO4
Solvent
Solvent DESALTING
recovery 2KHSO4 K2SO4+H2SO4
K2SO4 free of chlorides
solutions (Fig. 3) (Tomaszewska et al. 2001). The industry (fruit juices concentration) (Sotoft
process was also successfully applied in a pilot et al. 2012).
scale to clean low- and medium-level radioactive Novel applications are bio- and chemical reac-
waste from a nuclear center. tors integrated with DCMD in which the volatile
Moreover, DCMD found success in the treat- products are continuously removed from the reac-
ment of solutions containing substances sensitive tion mixture to shift the equilibrium and improve
to high temperature due to the risk of their degra- productivity. It was found that a continuous
dation, especially in biotechnology and food extraction of ethanol from broth using DCMD
Direct Fluorination of Polymer Membranes: Gas Separation Properties 565
resulted in an increase of its productivity by 87 %

(Khayet and Matsuura 2011). The membrane Direct Fluorination of Polymer
reactor based on DCMD was applied in the pro- Membranes: Gas Separation
duction of potassium sulfate for fertilizer industry Properties
(Tomaszewska and Mientka 2009). The concen-
tration of the reaction mixture containing KCl and A. P. Kharitonov
K2SO4 close to the saturated state in the mem- Talrose Branch of the Institute of Energy
brane reactor resulted in periodical precipitation Problems of Chemical Physics of the Russian
of KHSO4, whereas HCl was continuously Academy of Sciences, Moscow Region, Russia D
removed from the reaction mixture to shift the
equilibrium of the conversion (Fig. 4).
MD is generally recommended for special pur- Polymeric membranes can be used for the separa-
poses for which the application of other mem- tion of gas mixtures such as He-CH4, H2-CH4,
brane techniques was considered impossible. CO2-CH4, H2-N2, etc. There is, however, a com-
Until now the DCMD process has been under mon problem in a gas separation when a poly-
development due to high energy consumption meric membrane is used: membranes with high
and problems with the membrane wettability dur- gas permeability often have low gas separation
ing long-time operation. However, it should be factor, and on the contrary, membranes with high
pointed out that there are numerous calculations separation factor have low permeability (Fig. 1)
showing that DCMD can be economically attrac- (Robeson 2008). The direct uorination can be
tive especially when low-grade energy or alterna- effectively used to improve gas separation
tive energy is used.
References
Di Proo G, Curcio E, Drioli E (2010) Membrane

crystallization technology. In: Drioli E, Giorno
L (eds) Comprehensive membrane science and engi-
neering, vol 4. Academic Press/Elsevier, Oxford,
pp 2146
Drioli E, Curcio E, Di Proo G (2005) Review paper: state
of the art and recent progresses in membrane
contactors. Chem Eng Res Des 83(A3):223233
Drioli E, Stankiewicz AI, Macedonio F (2011) Membrane
engineering in process intensication an overview.
J Membr Sci 380:18
Khayet M, Matsuura T (2011) Membrane distillation. Prin-
ciples and applications. Elsevier, Amsterdam
Sotoft LF, Christensen KV, Andrsen R, Norddahl B
(2012) Full scale plant with membrane based
concentration of blackcurrant juice on the basis of
laboratory and pilot scale tests. Chem Eng Process Direct Fluorination of Polymer Membranes: Gas Sep-
54:1221 aration Properties, Fig. 1 Separation selectivity a for
Susanto H (2011) Towards practical implementations of the He/CH4 mixture vs. permeability of He for various
membrane distillation. Chem Eng Process 50:139150 polymer membranes in logarithmic scale. Filled
Tomaszewska M, Mientka A (2009) Conversion of KCl points literature data (Robeson 2008). Empty triangle,
into KHSO4 in membrane reactor: long-term experi- square, and diamond correspond to pristine polyamide
ments. Desalination 245:647656 Matrimid 5218 hollow ber module (points correspond
Tomaszewska M, Gryta M, Morawski AW (2001) Recov- to different treatment conditions) and PVTMS (Kharitonov
ery of hydrochloric acid from metal pickling solutions 2007, 2008) and PTMSP (Langsam et al. 1988) at mem-
by membrane distillation. Sep Purif Technol branes. Filled symbols represent transport properties of
2223:591600 uorine treated membranes
566 Direct Fluorination of Polymer Membranes: Gas Separation Properties
properties of polymer membranes when the gas

mixture consists of gases with markedly different
gas kinetic diameters. In this case substantial
increase of separation selectivity (up to several
tens times for the case of He/CH4 mixture) is
accompanied with a relatively small decrease
(or no change) of permeability of a gas with
smaller gas kinetic diameter (He, H2, etc.). Direct
uorination of polymers is a heterogeneous reac-
tion of gaseous F2 mixtures with a polymer sur-
face. This is a method of the surface modication:
only upper surface layer is modied (~0.01 to
several microns in thickness), but the bulk prop-
erties (e.g., tensile strength) remain unchanged.
The direct uorination proceeds spontaneously
at room temperature with sufcient for industrial
applications rate. Fluorination results in a substi-
tution of H-atoms for F-atoms, saturation of
double (conjugated) bonds with uorine, and Direct Fluorination of Polymer Membranes: Gas Sep-
aration Properties, Fig. 2 Inuence of treatment condi-
disruption of majority of C-N and C-Si bonds tions of PVTMS at membrane on the separation
followed with formation of C-F bonds. The chem- selectivity of CO2/CH4, He/N2, and He/CH4 mixtures.
ical composition of uorinated layer depends on Treatment condition (from left to right in each group at
composition and pressure of uorinating mixture the plot): virgin PVTMS, treatment with 2 % F2 + 98 % He
mixture, treatment with 33 % F2 + 67 %He mixture, treat-
and treatment duration. Treatment at mild uori- ment with 2 % F2 + 98 % He mixture followed by a
nation conditions does not cause disruption of grafting of acrylonitrile, treatment with 60 % F2 + 40 %
C-C bonds in the main polymer chain. Direct O2 mixture
uorination is a dry technology. Polymer hollow
bers, fabricated membrane modules, and com-
posite membranes can be treated. For the case of evidenced by Fig. 1, direct uorination results in a
hollow bers and composite membranes, only the very remarkable increase (by a factor of several
dense separation layer can be uorinated and the tens times or more than hundred times) of separa-
porous support will remain untouched, so the tion selectivity. The permeability of He and H2 is
tensile strength of membrane element will not be not practically changed after uorination. Hence
decreased. There are safe and reliable methods to the direct uorination of PVTMS and polyimide
neutralize (by converting into the solid phase) Matrimid 5218 provides the possibility to
unused F2 and the end product HF. overjump the Robeson boundary (straight line
The direct uorination was used to enhance gas in Fig. 1). The direct uorination can substantially
separation properties of several polymer mem- improve the separation selectivity of CO2/CH4,
branes (both homogeneous and composite) and He/N2 and He/CH4 mixtures (Fig. 2).
hollow ber modules: polyimide (PI), polyvinyl-
trimethylsilane (PVTMS), poly
(1-trimethylsilylpropyne) (PTMSP), poly References
(phenylene oxide), polysulfone, poly(4-methyl-
pentene), polycarbonatesiloxane, etc. (Langsam Amirkhanov DM, Kotenko AA, Tulskii MN (1998) Tech-
et al. 1988; Le Roux et al. 1994; Amirkhanov nological characteristics of the manufacture and use of
uorine-modied graviton hollow gas-separation
et al. 1998; Kharitonov 2007, 2008). Figure 1
bres. Fibre Chem 30:318324
illustrates the inuence of direct uorination on Kharitonov AP (2007) Chapter 2: direct uorination of
separation selectivity for He/CH4 mixture. As it is polymers from fundamental research to industrial
Direct Membrane Emulsification 567
applications. In: Gardiner IV (ed) Fluorine chemistry Membrane Emulsication,) or as a consequence

research advances. Nova Science Publishers, New York, of the shear stress generated by moving the mem-
pp 35103
Kharitonov AP (2008) Direct uorination of polymers. brane or the continuous phase (Dynamic Mem-
Nova Science Publishers, New York brane Emulsication.
Langsam M, Anand M, Karwacki EJ (1988) Chemical sur- Usually, in the DME method, the membrane is
face modication of poly[1-(trimethylsilyl)propyne] for wetted by the continuous phase, and hydrophilic
gas separation membranes. Gas Sep Purif 2:162170
Le Roux JD, Paul DR, Kampa J, Lagow RJ (1994) Modi- and hydrophobic membranes are used in the prep-
cation of asymmetric polysulfone membranes by mild aration of oil-in-water (O/W) and water-in-oil
surface uorination. Part I. Transport properties. (W/O) emulsions, respectively. The mean emul- D
J Membr Sci 94:121141 sion droplet size is controlled by the appropriate
Robeson LM (2008) Upper bound revisited. J Membr Sci
320:390400 selection of the pore size of the membrane, and a
factor of 210 is found between pore size and
droplets size. Additional parameters have to be
considered to tune the production of an emulsion
Direct Membrane Emulsification by DME:
Emma Piacentini1, Alessandra Imbrogno1,2 and Shear stress. Uniform and small droplets are
Richard G. Holdich3 obtained when the shear stress increases as a
1
Institute on Membrane Technology, National consequence of their faster detachment at the
Research Council of Italy, ITM-CNR, Rende, membrane pore surface. In general, the mean
Italy droplet size decreases exponentially until it
2
University of Calabaria, Rende, Italy reaches a size where it becomes, more or less,
3
Department of Chemical Engineering, independent of the applied shear. The shear
Loughborough University, Loughborough, stress can be produced: (i) by owing the con-
Leicestershire, UK tinuous phase tangentially to the membrane
surface (cross-ow ME), (ii) by pulsating the
continuous phase in the lumen side of the
membrane (pulsed ME or pulsed back-and-
In the early 1990s, Nakashima and Shimizu forward ME), (iii) by rotating or vibrating the
(1991) introduced membrane technology for membrane, and (iv) by inducing azimuthal
emulsion preparation by the direct emulsication oscillation of a tubular membrane.
method (DME). In DME, the dispersed phase is Transmembrane pressure. The transmembrane
directly fed through the membrane pores and pressure needs to be higher than a critical pres-
emulsion droplet formation takes place on the sure (capillary pressure) in order to initiate dis-
other side of the membrane, which is in contact persed phase intrusion in the pores. The
with the continuous phase (Fig. 1). Here, droplets dispersed phase ux increases with an increase
can detach after reaching a critical dimension of the transmembrane pressure as more pores are
either for spontaneous deformation (Static gradually activated for droplet production and
the velocity increases in the pores. Larger drop-
Dispersed Phase lets are produced at higher dispersed phase ux.
Emulsier. Droplet size decreases with lower
interfacial tension and, according to the time
Membrane needed by the emulsier, is dissolved in the
dispersed phase to cover the forming droplet
Continuous Phase
surface area. Smaller droplets are produced
because of a faster decrease of retaining forces
Direct Membrane Emulsification, Fig. 1 Direct mem- and a reduction of coalescence between the
brane emulsication method forming droplets at adjacent pores.
568 Direct Methanol Fuel Cell (DMFC)
One of the main limitations of DME is the low micrograph of a DMFC MEA is shown in Fig. 1
dispersed phase ux through conventional matrix- (Aric et al. 2001). The electrodes usually consist
type membranes (like SPG or ceramic mem- of three layers: catalytic layer, diffusion layer, and
branes) which have high thickness and tortuous backing layer. The catalytic layer is composed of a
pore channels giving relatively high ow resis- mixture of catalyst and ionomer, and it is charac-
tance. As an alternative, higher dispersed phase terized by a mixed electronic-ionic conductivity.
ux at low transmembrane pressure can be The catalysts are often based on carbon-supported
achieved using microsieve-type membranes due or carbon-unsupported PtRu and Pt materials at
to their low thickness and well-dened rectilinear the anode and cathode, respectively. The mem-
pores. Another problem encountered when the brane as well as the ionomer consists, in most
amount of the dispersed phase emulsied is cases, of a peruorosulfonic acid polymer. The
increased is the recirculation of the continuous package formed by electrodes and membrane is
phase (including emulsion drops), giving rise to called membrane and electrode assembly
poor control over the droplet size distribution. (MEA). The overall thickness of this package is
This is due to the interactions of the newly generally smaller than 1 mm. Each MEA forms a
forming droplets with droplets in the emulsion, cell. Several cells are usually connected in series
as well as droplet breakup within the circuit. To to form a fuel cell stack that is integrated in a
overcome this problem, advanced methods for system which contains the auxiliaries allowing
dynamic membrane emulsication like rotating, stack operation and delivering of the electrical
vibrating, azimuthally oscillating membrane power to the external load.
emulsication, and pulsed ow have received Typical peruorosulfonic acid membranes
much attention. used in the present DMFCs, such as Naon
The DME method is suitable for the prepara- (DuPont), are permeable to methanol transport
tion of simple and multiple emulsions as well as (Ren et al. 1996), thereby reducing signicantly
polymeric particles. In the latter case, polymeri- the fuel utilization efciency of the device.
zation, gelation, or coacervation processes of the A scheme of the overall reaction process occur-
formed droplets are carried out subsequently to ring in a DMFC equipped with a proton-
droplet generation. conducting electrolyte is outlined below:
CH3 OH H2 O ! CO2 6H
References 6e anode (1)
Nakashima T, Shimizu M, Kukizaki M (1991) Membrane

3=2O2 6 H 6e ! 3H2 O cathode (2)
emulsication by microporous glass, Key Engineering
Materials, 61/62: 513516
CH3 OH 3=2 O2 ! CO2
2H2 Ooverall (3)
In the presence of an alkaline electrolyte, the

Direct Methanol Fuel Cell (DMFC) process is outlined as follows:
Antonino Salvatore Arico CH3 OH 6 OH ! CO2 5 H2 O

CNR-ITAE Institute, Messina, Italy 6e anode (4)
3=2O2 3H2 O 6e ! 6OH cathode (5)

The core of a direct methanol fuel cell is a polymer
electrolyte ion-exchange membrane. The elec- CH3 OH 3=2O2 ! CO2
trodes (anode and cathode) are in intimate contact 2H2 Ooverall (3)
with the membrane faces. A scanning electron
Direct Methanol Fuel Cell (DMFC) 569
Direct Methanol Fuel Cell

(DMFC), Fig. 1 SEM
micrograph of a DMFC
membrane and electrode
assembly equipped with
Naon 112 membrane
Direct Methanol Fuel Cell 1

(DMFC), Fig. 2 Single-cell T = 60C
and in situ half-cell Cell
electrode polarizations for a 0.8
Anode
DMFC operating at 60 C,
ambient pressure, with 1 M Cathode
Potential / V
methanol at the anode and 0.6

air feed at the cathode
0.4
0.2
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Current density / A cm2
The thermodynamic efciency of the process is DG nF DErev (8)

given by the ratio between the Gibbs free energy,
i.e., the maximum value of electrical work (DG ) DErev is the electromotive force. At 25 C and
that can be obtained, and the total available energy 1 atm and with pure oxygen feed, the reversible
for the process, i.e., the enthalpy (DH ). Under potential for methanol oxidation is 1.18 V (Aric
standard conditions: et al. 2001). It does not vary signicantly in the
operating range 20 130 C and 13 bar abs.
rev DG =DH ; reversible energy efficiency pressure.
(6) Usually, the open-circuit voltage of a polymer
electrolyte direct methanol fuel cell is signi-
with
cantly lower than the thermodynamic or reversible
DG DH T DS (7) potential for the overall process (Fig. 2). This is
mainly due to methanol crossover that causes a
and mixed potential at the cathode and to the
570 Discontinuous Diafiltration by Sequential Dilution (Sequential Dilution Diafiltration)
irreversible adsorption of intermediate species at Extending this analysis to n stages (i.e., repeating
electrode potentials close to the reversible poten- steps 1 and 2 above n times) gives
tial. Direct methanol fuel cells nd application in
portable power sources, assisted power units V f =V 0 b1=n
(APU), and electrotraction (Dillon et al. 2004).
For example, if the microsolute concentration is to
References be reduced by a factor of 10 (= b) and then if a
single stage is used, Vf/V0 = 10, while if three
Arico AS, Srinivasan S, Antonucci V (2001) DMFCs: stages are used, Vf/V0 = 101/3 = 2.15.
From fundamental aspects to technology development. The total volume of buffer used in this case is
Fuel Cells 1:133
Dillon R, Srinivasan S, Aric AS, Antonucci V (2004)
given by
International activities in DMFC R&D: Status of tech-
nologies and potential applications. J Power Sources V b =V 0 n V f V 0 =V 0 n V f =V 0 1
127:112
Ren X, Wilson MS, Gottesfeld S (1996) High performance n b1=n 1
direct methanol polymer electrolyte fuel cells. J
Electrochem Soc 143:L12
For a single stage system, this ratio works out as
9, while for a three-stage system, it works out as
3.46. Thus, the more stages that are used, the
Discontinuous Diafiltration by smaller the buffer consumption.
Sequential Dilution (Sequential
Dilution Diafiltration)
Greg Foley
School of Biotechnology, Dublin City University, Discontinuous Diafiltration by
Dublin, Ireland Volume Reduction (Intermittent
Feed Diafiltration)
Discontinuous dialtration by the sequential dilu- Greg Foley
tion method can be summarized in terms of the School of Biotechnology, Dublin City University,
three steps shown below. The apparent rejection Dublin, Ireland
coefcients are assumed to be 1.0 and 0 for
macrosolute and microsolute, respectively.
The steps involved in a discontinuous dialtration
Step 1: Add buffer to increase solution volume by volume reduction ( intermittent feed
from V0 to Vf. Macrosolute (A) and microsolute dialtration) are outlined below. The apparent
(B) concentrations decrease. rejection coefcient of the macrosolute is 1.0,
Step 2: Use batch ultraltration to reduce solution while that of the microsolute is zero.
volume back from Vf back to V0. Macrosolute
concentration returns to its initial value, Step 1: Batch ultraltration from volume V0 to Vf.
microsolute concentration remains Macrosolute concentration increases,
unchanged. (This follows from the rejection microsolute concentration remains the same
coefcients of both solutes.) because it has a rejection coefcient of zero.
Step 3: Repeat steps 1 and 2 as required. Step 2: Replace lost permeate with diluant (pure
water or buffer). Macrosolute concentration
A microsolute balance on the rst step gives returns to initial concentration, microsolute
concentration reduced.
b cB0 =cBf V f =V 0 Step 3: Repeat steps 1 and 2 as required.
Dissolved Oxygen (DO) in Water by Membrane Operations, Removal of 571
Consider a single stage, i.e., steps 1 and (approximately 8 ppm under ambient conditions),
2 above. The concentration of the macrosolute DO leads to oxidation of dissolved components
remains unchanged. A mass balance on the and materials, for example, boiler piping in
microsolute gives the following expression for (nuclear) power plants. In such systems, the
its concentration at the end of the two-step desired DO concentration in the boiler feed
process. water is 710 ppb. Therefore, DO removal from
water presents a challenging problem for various
b cB0 =cBf V f =V 0 industries such as production of semiconductors,
power plants, pharmaceuticals, and biotechnol- D
If this is repeated n times, one gets ogy. In the production of semiconductors, stan-
dard DO level in ultrapure water is the most
V 0 =V f b1=n stringent (below 1 ppb). Removal of DO from
water by membrane technology can be accom-
The buffer consumption is given by plished by using gas-liquid membrane contactors
and membrane reactors.
Gas-liquid membrane contactors generally
V b =V 0 n V 0 V f =V 0 n 1 V f =V 0
incorporate a porous hollow ber membrane for
n 1 b1=n direct contacting of two immiscible phases,
namely, a gas and a liquid, for the purpose of
Thus, for a tenfold reduction in microsolute absorption or stripping without dispersion of one
concentration and a three-stage process, one gets phase into another. The DO removal from water
V0/Vf = 101/3 = 2.15 in each step, while the on a membrane contactor requires generation of a
buffer consumption works out as driving force for oxygen transmembrane transfer,
Vb/V0=3(1101/3) = 1.607. and this process demands reduction in oxygen
If one stage had been used, one would have had concentration in a gas phase. This goal can be
V0/Vf = 10 and Vb/V0 = 1.0(1101) = 0.9. Thus, accomplished in one of three ways: (1) by apply-
increasing the number of stages increases the ing vacuum to the water, (2) by supplying an
amount of buffer required. appropriate sweep gas to the gas side, and (3) by
The single-stage approach is not always possi- a combination of the rst two where a small ow
ble as too large a volume concentration factor of sweep gas is used to improve separation driving
(V0/Vf) might lead to unsustainably high solution force but the gas side is still maintained under
viscosities. vacuum (Sengupta et al. 1998). In industrial
(commercial) Liqui-Cel gas-liquid membrane
contactors, polypropylene porous hollow ber
membranes are used. These membrane contactors
Dissolved Oxygen (DO) in Water by offer a serious alternative to the traditional
Membrane Operations, Removal of two-stage vacuum degassing towers in the high-
purity water treatment systems in the semiconduc-
V. V. Volkov tor industry.
A.V. Topchiev Institute of Petrochemical DO removal from water can be performed on a
Synthesis Russian Academy of Sciences (TIPS membrane-UV reactor because DO can react with
RAS), Moscow, Russian Federation hydrogen to yield water under the action of UV
irradiation with a wavelength of 185 nm. This
reaction requires dissolution of hydrogen in
Dissolved oxygen (DO) is one of the principal water, and this process is provided by gas-liquid
components which should be removed for the membrane contactors based on hollow ber mem-
production of ultrapure water. Even though con- branes which are permeable for hydrogen and UV
centration of DO in water is very low irradiation (Li and Tan 2001).
572 Distributor Reactor
Most efcient existing methods for deep

removal of DO from water are based on the prin- Distributor Reactor
ciple of oxygen reduction by hydrogen on a pal-
ladium catalyst which yields water. The existing Membrane Distributor
catalytic processes for water deoxygenation via
hydrogenation reaction involve two stages:
(1) absorption of hydrogen in water and (2) pas-
sage through a xed-bed catalytic reactor. In con- Donnan Effect
trast to existing processes, the DO removal on a
catalytic membrane reactor can be accomplished Patrick Fievet
in one stage. There are two congurations of Institut UTINAM, University of Franche-Comt,
catalytic membrane reactors. The rst design of Besanon, France
the membrane reactor is based on a polypropylene
porous hollow ber membrane module packed
with a palladium catalyst, namely, spherical beds The Donnan effect (which is named after the
of palladium-deposited granules of anion- British chemist Frederick George Donnan
exchange resin in the void space of the shell side (Donnan 1911, 1995) is related to the presence
(Li et al. 1995). An alternative design of a cata- of impermeant ions (i.e., ions that are unable to
lytic membrane reactor provides water deoxygen- pass through a semipermeable membrane or
ation by catalytic membranes, namely, by boundary) upon one side of a boundary on the
palladium-loaded hollow bers (Volkov distribution of permeant ions across the boundary.
et al. 2009). Water and hydrogen are supplied At equilibrium, the co-ions (i.e., ions with the
into a shell side of the reactor and into a ber same charge polarity as the impermeant ions)
lumen, respectively. Hence, DO removal is will be less concentrated on the impermeant
accomplished by chemical reaction between ions side of the boundary than the other side,
dissolved oxygen and dissolved hydrogen in the whereas the counterions (i.e., ions having the
presence of the palladium-based catalyst. Using a charge polarity opposite to that of the impermeant
catalytic membrane reactor, concentration of DO ions) will be more concentrated there. It results an
in water of ten parts per billion (ppb) and lower is unequal distribution of permeant ions between the
feasible. two sides of the boundary.
The surface charged groups (e.g., sulfonic or
carboxylic acids) of a synthetic membrane are an
References example of impermeant ions. When such a mem-
brane is placed in a salt solution, a Donnan equi-
Li K, Tan X (2001) Development of membrane-UV reactor librium sets up at both membrane-external
for dissolved oxygen removal from water. Chem Eng solution interfaces. Because of the presence of
Sci 56(17):50735083 the membrane xed charge, the ion concentrations
Li K, Chua I, Ng WJ, Teo WK (1995) Removal of
in the membrane are not equal to those in the
dissolved oxygen in ultrapure water production using
a membrane reactor. Chem Eng Sci 50(22):35473556 solution: the concentration of co-ions is lower in
Sengupta A, Peterson PA, Miller BD, Schneider J, Fulk the membrane phase than in the solution, while
CW (1998) Large-scale application of membrane the concentration of counterions is higher there.
contactors for gas transfer from or to ultrapure water.
Sep Purif Technol 14:189200
A potential difference, called the Donnan poten-
Volkov VV, Petrova IV, Lebedeva VI, Plyasova LM, tial, is created at the membrane-solution interface
Rudina NA, van Erkel J, van der Vaart R, Tereshchenko to counteract the transport of the co-ions to the
GF (2009) Catalytic nanoclusters of palladium on the membrane phase and the transport of counterions
surface of polypropylene hollow ber membranes:
to the solution. In other words, the co-ions are
removal of dissolved oxygen from water. In: Starov
VM (ed) Nanoscience: colloidal and interfacial aspects. repelled by the membrane, whereas the counter-
Taylor & Francis Group, New York, pp 11731188 ions are attracted.
Donnan Effect 573
a b
Interface
Solution Membrane Equilibrium Donnan:

(cC)+ (cA) = (c C)+ (cA)
G
Electroneutrality conditions:
C C
z C cC + z A c A = 0 (solution)
A A z C cC + z A c A + X = 0 (membrane)
D
Donnan Effect, Fig. 1 (a) Schematic representation of of cations and anions, respectively. zC and zA are the charge
the distribution of cations C and anions A between the numbers of cations and anions, respectively. The symbol
solution and the membrane having xed ionic groups G; X (positive or negative depending on the charge sign of the
(b) Donnan relation and electroneutrality conditions in the membrane) denotes the volume charge density, which is
solution and membrane. c and c are the molar ion concen- dened as the number of moles of xed charges per unit of
tration at equilibrium in the solution and membrane, pore volume
respectively. n+ and n are the stoichiometric coefcients
To illustrate the inuence of various parame- Donnan Effect, Table 1 Calculation of the co-ion
ters on the Donnan equilibrium, let us consider the partitioning coefficient (kco-ion) at a membrane-solution
interface for a (positively or negatively charged) membrane
equilibrium situation between a membrane with a fixed charge |X| of 102 mol L1 and various salts
containing xed ionic groups G and an ionic solu- and concentrations
tion containing the strong electrolyte CA (Fig. 1). cco-ion (mol L1) |zcoion| |zcounterion| kco-ion
Based on the Donnan equilibrium and taking 104 1 1 0.01
the electroneutrality conditions into account, the 103 1 1 0.10
partitioning coefcient of co-ions at the 102 1 1 0.62
membrane-solution interface (kco-ion), which 103 1 2 0.31
describes the distribution of co-ions between the 103 2 1 0.04
membrane and the solution, can be derived:
ccoion Solving Eq. 1 allows the concentration of

kcoion
ccoion co-ions in the membrane (ccoion ) and, thus, the

z zcoion partitioning coefcient of co-ions (kco-ion) to be
zcoion ccoion counterion calculated. Results presented in Table 1 show that
(1)
zcoion ccoion X exclusion increases (kco-ion decreases) when:
with ccoin and ccoion the concentrations of The salt concentration (or ccoin) decreases
co-ions in the solution and membrane, respec- The charge number of co-ions zcoion increases
tively, and zcoion and zcounterion the charge num- The charge number of counterions zcounterion
bers of co-ions and counterions, respectively. decreases
A co-ion partitioning coefcient that tends to The xed membrane charge (in absolute value)
one means that co-ions (and therefore counterions X increases.
due to electroneutrality requirements) are only
slightly excluded from the membrane (weakly The increase (in absolute value) in the xed
charged membrane, small |X|), whereas a co-ion charge and electrolyte concentration has then
partitioning coefcient that tends to zero means opposite effects on the exclusion phenomenon
that the co-ions are strongly excluded since the former makes the Donnan effect stron-
from the membrane (strongly charged membrane, ger, whereas the latter screens the electrostatic
high |X|). interaction between ions and the membrane xed
574 Donnan Equilibrium
Donnan Effect, Table 2 Calculation of the co-ion References

partitioning coefficient (kco-ion) at a membrane-solution
interface for various membrane volume charges |X|, salts, Donnan FG (1911) Theorie der Membrangleichgewichte
and concentrations und Membranpotentiale bei Vorhandensein von nicht
z+ n+ dialysierenden Elektrolyten. Ein Beitrag zur
csalt csalt = | |X| physikalisch-chemischen Physiologie (theory of mem-
(mol z|n (mol brane equilibria and membrane potentials in the pres-
jXj
Salts L1) csalt L1) z n csalt kco-ion ence of non-dialyzing electrolytes. A contribution to
1:1 103 103 102 10 0.10 physical-chemical physiology). Z Elektrochem Angew
Phys Chem 17(10):572581
1:1 102 102 101 10 0.10 Donnan FG (1995) Theory of membrane equilibria and
1:2 5 103 102 10 0.01 membrane potentials in the presence of non-dialyzing
104 electrolytes. A contribution to physical-chemical phys-
1:2 5 102 101 10 0.01 iology. J Membr Sci 100:4555
103
2:1 5 103 102 10 0.31
104
2:1 5 102 101 10 0.31
2:2 5 103 102 10 0.10
104
Patrick Fievet
2:2 5 102 101 10 0.10
103
Institut UTINAM, University of Franche-Comt,
Besanon, France
Salt 1:2: salt with a monovalent counterion and divalent co-ion
Salt 2:1: salt with a divalent counterion and monovalent
co-ion
Salt 1:1: salt with a monovalent counterion and monovalent The Donnan equilibrium describes the distribu-
co-ion
tion of ions between two ionic solutions separated
Salt 2:2: salt with a divalent counterion and divalent co-ion
by a semipermeable membrane or boundary. The
presence of impermeant ions of one side of the
boundary has profound effects on the nal distri-
bution of permeant ions as they diffuse to an
charge. As a result, as can be seen in Table 2, the equilibrium in the solution, which is named after
co-ion partitioning coefcient is ruled by the ratio the British chemist Frederick George Donnan
of the membrane xed charge (in absolute value) (1911, 1995).
to the charge concentration in solution, z njX cj salt . To illustrate this phenomenon, let us consider a
Table 2 also shows that for salt solutions having boundary that initially separates a solution of
the same charge concentration, the co-ion NaCl (side I) and a solution of NaX (side II),
partitioning coefcients are identical for salts 1:1 NaCl and NaX being completely dissociated and
and 2:2, and kco-ion follows the sequence 2:1 > having the same molar concentrations, and let us
1:1 = 2:2 > 1:2 where the rst gure denotes the assume that the boundary is impermeable to the
charge number of the counterion and the second anion X, but completely permeable to both Na+
one that of the co-ion. and Cl ions (Fig. 1a). X can be, e.g., an anionic
The increase of co-ion exclusion from the protein inside a cell, the walls of which are imper-
membrane leads to a higher salt rejection since meable to the protein or surface charged groups
the latter is determined by rejection of co-ions. (such as, e.g., sulfonic or carboxylic acids) of a
synthetic membrane. Since side II does not con-
tain Cl ions, Cl ions diffuse along the concen-
Cross-References tration gradient from side I to side II (Fig. 1a). The
presence of both Cl anions plus impermeable
Donnan Equilibrium anions on side II makes side II negative relative
Donnan Potential to side I and establishes an electric potential
Donnan Equilibrium 575
a b
I II I II
NaCl NaX
+ -
+ +
+ + X- + + + X- +
+ X- + X-
+ X- + X-
+ X- + + X-
+ + +
+ +
+ + X- + + X- D
+ X- X- + X- X-
+ +
Cl- concentration gradient Cl- electrical gradient
Na + concentration gradient
c d
I II I II
+ -
+ +
+ + X- + + + X- +
+ X- + X-
+ + X- + + X-
+ X- X-
+ + +
+ +
+ + X- + + X-
X- X- X-
+ X- + +
+
Na + concentration gradient
Donnan Equilibrium, Fig. 1 Illustration of the establishment of Donnan equilibrium between two solutions (NaCl and
NaX) separated by a boundary permeable to K+ and Cl ions but impermeable to the anion X
difference at the boundary between sides I and II response to the competing electrochemical gradi-
(Fig. 1b). The development of the negative charge ent so that each side remains electrically neutral
on side II counterbalances the concentration gra- due to equal molar concentrations of both cations
dient of Cl ions and prevents Cl concentration and anions. Between the two sides, there is
from becoming equal on both sides of the mem- unequal distribution of permeant ions with the
brane. At the same time, the negative charge on more total anions on side II. This imbalance in
side II attracts the positive Na+ ions to migrate the distribution of permeant ions results in an
from side I to side II along the electrical gradient. electrical potential difference between sides
The Na+ concentration on side II now exceeds that I and II, which is called the Donnan potential.
of side I, and Na+ ions will diffuse back from side At equilibrium, the electrochemical
II to side I along the concentration gradient potentials of the ions on side I and side II are
(Fig. 1c). At equilibrium, the competing electrical equal, hence:
potential and concentration gradients for each ion
are in balance and there is no net ow of ions m0i , I RTlnaIi zi FcI
through the membrane (Fig. 1d). Na+ and Cl ions
move back and forth between sides I and II in m0, II RTlnaII z FcII
i i i (1)
Initial state
SideI SideII 50 120
Na+ Cl- Na+ X- 40

100
c1 c1 c2 c2
80
100x / c1(%)
30
c1-x / x
Equilibrium state 60
SideI SideII 20
40
Na+ Cl- Na+ X- Cl- 10
c1 - x c1 - x c2 + x c2 x 20
0 0
Eq. (3) leads to the following relations: 0.01 0.1 1 10 100
x c1 c1 x c1 + c2 C2/C1
=
c1 c2 + 2c1 x c1
Donnan Equilibrium, Fig. 2 Effect of the initial con- and the distribution ratio of NaCl between sides I and II at
centration ratio of NaX (c2) to NaCl (c1) on the percentage equilibrium (c1 x/x)
amount of NaCl diffusing from side I to side II (100 x/c1)
where m0i represents the standard chemical poten- where the square brackets indicate molar concen-
tial of the permeant ion i, ai is its activity, zi is its trations at equilibrium.
charge number, C is the electric potential, R is the It should be noted that Donnan equation Eq. 3
ideal gas constant, T is the temperature, and F is was derived without considering the initial con-
the Faraday constant. The superscripts I and II centrations of NaCl and NaX on sides I and II,
refer to the side I and side II, respectively. respectively.
In the derivation of the electrochemical poten- Figure 2 shows the effect of the initial concen-
tial, the difference in pressure between sides I and tration ratio of NaX (c2) to NaCl (c1) on the
II is neglected. In the case of a synthetic mem- percentage amount of NaCl diffusing from side
brane, this assumption can be justied if its swell- I to side II (100 x/c1) and the distribution ratio of
ing is negligible. NaCl between sides I and II at equilibrium
Assuming that the standard chemical potentials (c1 x/x). These results are obtained by making
on sides I and II are equal, the Donnan potential the assumption that the liquid volumes on both
(dened as the potential difference between the sides of the membrane are equal.
side of the boundary with the impermeant ion and As can be seen, if the relative concentration of
the side with the permeant ions) can be calculated NaX on side II is high enough (c2 >> c1), NaCl
for each permeant ion i as: will be expelled to a large extent from side II to
side I and the distribution of NaCl between I and II
RT aIi will be then very unequal, whereas, on the other
DcD cII cI ln (2)
zi F aIIi hand, more NaCl (up to 50 %) will be able to ow
from side I to side II with decreasing c2/c1 and the
The writing of Eq. 2 for permeant ions Na+ and Cl distribution ratio of NaCl between I and II will
(assuming sufciently diluted solutions) and the tend then to 1. At equilibrium, the concentration
combination of the two resulting equations lead to of Cl ions (i.e., ions with the same charge polar-
the so-called Donnan (or Gibbs-Donnan) equation: ity as the impermeant ions) on side II will be lower
than that on side I, whereas the Na+ ions (i.e., ions
Na 1 Cl 1 Na 2 Cl 2 (3) having the charge polarity opposite to that of the
Donnan Potential 577
Donnan Equilibrium,
1 120
Fig. 3 Effect of the initial
concentration ratio of NaX
(c2) to NaCl (c1) on the 100
0.8
concentration ratio of both
Cl ions (x/c1 x) and 80
Na+ ions (c2 x/c1 x)
c2+x / c1-x
0.6
x / c1-x
between sides II and I at 60
equilibrium
0.4
40 D
0.2
20
0 0
0.01 0.1 1 10 100
c2/c1
impermeant ions) will have a higher concentration by a semipermeable membrane or boundary.

on side II than on side I (Fig. 3). Because of the presence of impermeant ion species
on one side of the boundary, the concentrations of
permeant ions are not the same on both sides of
Cross-References the boundary. These concentration differences
between two solutions lead to chemical potential
Donnan Potential differences across the boundary and these differ-
ences will, on their turn, be compensated by an
additional electric potential difference between
References
two solutions. At equilibrium, the electrochemical
Donnan FG (1911) Theorie der Membrangleichgewichte potentials of permeant ions in two solutions are
und Membranpotentiale bei Vorhandensein von nicht identical. Assuming their standard chemical poten-
dialysierenden Elektrolyten. Ein Beitrag zur tials are equal, the electric potential difference
physikalisch-chemischen Physiologie (theory of mem- between two solutions, which is called the Donnan
brane equilibria and membrane potentials in the pres-
ence of non-dialyzing electrolytes. A contribution to potential (DCD), can then be derived:
physical-chemical physiology). Z Elektrochem Angew
Phys Chem 17(10):572581
RT aIi
Donnan FG (1995) Theory of membrane equilibria and DcD cII cI ln (1)
membrane potentials in the presence of non-dialyzing zi F aIIi
electrolytes. A contribution to physical-chemical phys-
iology. J Membr Sci 100:4555
with c the electric potential, R the ideal gas con-
stant, T the absolute temperature, F the Faraday
constant, zi the charge number of ion i, and ai its
activity. The superscript I and II refer to the solu-
Donnan Potential
tion without impermeant ions and the solution
with impermeant ions, respectively.
Patrick Fievet
The impermeant ions may be, e.g., anionic
Institut UTINAM, University of Franche-Comt,
proteins inside a cell whose walls are imperme-
Besanon, France
able or xed charges of a synthetic membrane.
Equation 1 shows that:
The Donnan potential appears as a result of
Donnan equilibrium, which refers to the distribu- The Donnan potential will be negative in case
tion of ions between two ionic solutions separated of anionic impermeant species (membrane
578 Donnan Potential
Donnan Potential, Pore

Fig. 1 Schematic
representation of the
electric potential
0 diff
distribution in a pore under D
a concentration difference.
Dc0D and DcDx D : Donnan Diluted solution
potentials at x
interfaces.Dcdiff: diffusion D
Concentrated m
potential. Dcm: membrane
solution
potential
x
0 x
with a negative xed charge) and positive in (DCdiff) arising through the membrane pores
case of impermeant cations (membrane with a (Fig. 1):
positive xed charge).
The co-ions, i.e., ions with the same charge DCm DCDx
D DCD DCdiff
0
(1)
polarity as the impermeant ions (the latter can
be the xed charges of a synthetic membrane), The diffusion potential nds its origin in a differ-
will be less concentrated on the impermeant ence in transport velocity of cations and anions
ions side than the other side, whereas the through the membrane.
counterions, i.e., ions having the charge polar- Within the scope of the DSP&DE model, ion
ity opposite to that of the impermeant ions, will partitioning at the membrane/solution interfaces is
be more concentrated there. It results in an accounted for by means of modied Donnan
unequal distribution of permeant ions equations including steric and dielectric effects.
(co-ions and counterions) between the two The Donnan potential reads as follows:
sides.
int int
int kB T k c
The Donnan potential is used in the Teorell- D c
DCint cint ln i inti (2)
zi e ci
Meyer-Sievers (TMS) model (Theorell 1935;
Meyer and Sievers 1936) and related models with
such as the DSP&DE model (Lanteri et al. 2008;
Escoda et al. 2010), which notably describe the
0int 0int
i fi exp DWi, Born exp DWi, image
int
membrane potential phenomenon. The membrane kint
potential is dened as the electrical potential dif- (3)
ference arising through a membrane that separates
two solutions of the same electrolyte at the same kB is the Boltzmann constant; T is the absolute
temperature and hydrostatic pressure but different temperature; e is the elementary charge; zi is the
concentrations. Within the scope of the TMS charge number of ion i; fint i represents the steric
model (and related models), the membrane poten- partitioning coefcient for ion i, which is dened
tial (DCm) is split in two components, namely, the as the ratio between the available section (i.e.,
difference between the Donnan potentials at each taking into account the zone inside the pore in
interface (DcDx 0
D and DcD: superscripts Dx and which the ion center cannot penetrate because of
0 refer to the interface between the membrane its nite size) and the pore cross-section; and the
0 int 0 int
and the most diluted solution and the interface terms DWi,Born and DWi,image denote the increase
between the membrane and the most concentrated in the interaction energy due to the Born dielectric
solution, respectively) and the diffusion potential effect (Born 1920) and image forces (Dukhin
Donnan Potential 579
et al. 1988), respectively, where the prime symbol In high concentrations, the Donnan contribu-
indicates that both terms are scaled on kBT. tion tends to zero (the asymptotic value of the
According to Eq. 2, the difference in Donnan membrane potential is then equal to the diffusion
potentials, called the Donnan term, is given by potential) due to the screening of both membrane
(considering c0i 2cDxi ): xed charge density and image force interaction

! DW00 0Dx
i, Born DWi, Born 0 .
0
kB T kDx
i ci
DCDx
D DC0D ln 0 Dx
(4)
zi e 2ki ci For neutral membranes and when the only D
exclusion mechanism is the steric exclusion
For charged membranes, the membrane poten- (i.e., kint fint
i i ) or
the Born dielectric exclu-

tial is bound between limiting values
corresponding to the Nernst potential at low sion ki ; exp DW0int
int
i, Born or both

concentrations and the diffusion potential at 0int
i fi ; exp DWi, Born
int
kint , the Donnan
high concentrations (Lanteri et al. 2008; Escoda
et al. 2010). The Nernst potential corresponds to contribution is null irrespective of the concen-
the asymptotic value of the Donnan term when tration. It means that Donnan potentials are
the co-ion exclusion is total. If the membrane equal at both interfaces although each of them
volume charge density is assumed to be inde- differs from zero. Indeed, since cations and
pendent of the axial position inside pores, the anions have usually different sizes, their steric
concentration of counterions inside pores is then partitioning coefcients are different as well as
0 int
constant too (c0counterion cDx their terms DWi,Born . The smallest ions (e.g., the
counterion ) and Eq. 4
can be rewritten as: anions) will be then less excluded than the
largest ones (e.g., the cations). Consequently,
Dx a slight negative Donnan potential will arise,
kB T kcounterion
D DCD
DCDx 0
ln (5) pulling cation ions into pores and expelling
zcounterion e 2k0counterion
anions to maintain electroneutrality within the
membrane. Since only steric effects or only
Moreover, the steric partitioning coefcients
Born dielectric effect or both are considered,
are identical at both interfaces ( f0i fDx
i 6 0 ) one obtains k0i kDx i . Considering the
since the pore radius is assumed to be
electroneutrality condition within the mem-
constant through the membrane. Born effect is brane, it comes from Eq. 4 that the Donnan
also identical at both interfaces DW00 i, Born
0Dx
potentials are identical at both interfaces, and
DWi, Born 6 0, since it is considered to be inde- thus the Donnan contribution to the membrane
pendent of the salt concentration. Unlike Born potential arising through a neutral membrane is
effect, the magnitude of the interaction via image null whatever the concentration.
charges depends on concentration. However,
there is almost no screening of the image force
interaction at low concentrations, and thus, Cross-References
DW00 0Dx
i, image DWi, image 6 0 . It results that
Donnan Equilibrium
k0counterion kDx
counterion. Equation 5 then reduces to: DSP&DE Model Applications
kB T References
D DCD
DCDx ln2
0
(6)
zcounterion e
Born M (1920) Volumen and hydratationswarme der
ionen. Z Phys 1:4548
which represents the Nernst potential (the value of Dukhin SS, Churaev NV, Shilov VN, Starov VM
which is inversely proportional to the charge num- (1988) Modelling reverse osmosis. Russ Chem Rev
ber of counterions). 57(6):572
580 Donnan Steric Pore (DSP) and Dielectric Exclusion (DE) Model
Escoda A, Lanteri Y, Fievet P, Don S, Szymczyk A (2010) dielectric effects. Its main features can be summa-
Determining the dielectric constant inside pores of rized as follows:
nanoltration membranes from membrane potential
measurements. Langmuir 26(18):1462814635
Lanteri Y, Szymczyk A, Fievet P (2008) Inuence of steric, The active layer of the membrane is considered
electric and dielectric effects on membrane potential. as a charged porous layer and is characterized
Langmuir 24:79557962 by means of three adjustable parameters: the
Meyer KH, Sievers JF (1936) La permabilit des mem-
branes I. Theorie de la permabilit ionique. Helv Chim average pore radius rp, the effective membrane
Acta 19:649664, 665677; 987995 thickness d (including tortuosity or porosity
Theorell T (1935) An attempt to formulate a quantitative and tortuosity of the active layer depending
theory of membrane permeability. Proc Soc Exp Biol on the model version), and the volume charge
Med 33:282283
density X (dened as the mole number of xed
charges per unit of pore volume), which is
assumed to be constant along the membrane
pores. The presence of the coarse-porous sup-
Donnan Steric Pore (DSP) port layer is disregarded because it does not
and Dielectric Exclusion (DE) Model affect the retention properties of the
membrane.
Patrick Fievet The radial variations of both local electric
Institut UTINAM, University of Franche-Comt, potential and ion concentrations inside the
Besanon, France nanopores are neglected. This approximation
(called homogeneous approximation) is valid
provided the electric charge carried out by the
The Donnan steric pore (DSP) and dielectric exclu- nanopore walls is not too high (Szymczyk
sion (DE) model (known as the SEDE model in et al. 1999).
literature) is a model that describes mass transfer of Membrane separation is described as being
neutral and charged solutes through nanoltration the result of the following steps: rst, a distri-
(NF) membranes (Szymczyk and Fievet 2005). It is bution of species at the membrane/feed
based on a macroscopic description of transport solution interface, followed by solute transport
and is an extension of the Donnan steric pore through the pores, and nally a distribution of
model (DSPM) rstly proposed by Bowen and species at the membrane/permeate interface
Mukhtar (1996) and then revised in Bowen (Fig. 1).
et al. (1997) and in Bowen and Mohammad The solute transport is based on the extended
(1998). In addition to the steric/electric exclusion NernstPlanck (ENP) equation, which
mechanisms, the DSP and DE model also includes describes transport in terms of diffusion,
migration (in the case of ions only), and con-
vection. It is modied by hydrodynamic coef-
cients (Ki,c and Ki,d) accounting for the effect
of nite pore size on the various components of
Note for editors: The model developed by Bandini and
Vezzani is known as the DSP and DE model. However, the transport and reads as follows:
their model considers dielectric exclusion in terms of the
only image force contribution. My suggestion, therefore,
would be to replace the topic title Donnan Steric Pore dci zi ci Ki, d Di, 1 F dc
ji Ki, d Di, 1
(DSP) and Dielectric Exclusion (DE) Model by the title dx RT dx
SEDE (Steric, Electric, and Dielectric Exclusion) Model,
Ki, c ci V (1)
which includes dielectric exclusion in terms of both Born
dielectric effect and image force contribution. It would
avoid confusion. Indeed, I wrote this topic by considering
where ji is the molar ux of specie i through
my model, namely, the SEDE model, which was developed
by Szymczyk and me. However, as requested by the edi- the pores; Di,1 its diffusivity in the external
tors, I did use the term DSP and DE in the text below. solution (value at innite dilution); ci its
Donnan Steric Pore (DSP) and Dielectric Exclusion (DE) Model 581
Pore
ci (0)
c+ (0+)
c+ ()
D
X
Feed c (0+) Permeate
+
0
D c () ci ( )
D
0 0+ +
x
Donnan Steric Pore (DSP) and Dielectric Exclusion electric potential proles. Mono-monovalent salt and neg-
(DE) Model, Fig. 1 Schematic representation of a pore atively charged membrane (X < 0). Dcint D : Donnan poten-
separating a feed solution (high-pressure side) from a tial (superscript int stands for 0 or d)
permeate (low-pressure side) and concentration and
concentration inside the pores at the x position cint

i
Gint i
along the pore; zi its charge number; c and cint
V are the local electric potential and the solvent
i

zi e
velocity inside the pores, respectively; R the Fint
i exp DcDint
kB T
ideal gas constant; T the absolute temperature; 0 int
0 int

and F the Faraday constant. exp DW i, Born exp DW i, image
Hydrodynamic coefcients Ki,c and Ki,d
(3)
depend on the solute to pore size ratio, li,
dened as li = ri,Stokes/rp where ri,Stokes is where the superscript int stands for 0 or d depending
the Stokes radius of the solute which is calcu- on the interface that is considered, the symbol D
lated using the well-known StokesEinstein denotes a variation with respect to external solu-
relation: tions, and e is the elementary charge.
The rst term on the r.h.s. represents the con-

kB T tribution of the steric exclusion, fi being the steric
ri, Stokes (2) partitioning coefcient for solute i, which is
6pZDi, 1
dened as the ratio between the available section
(i.e., taking into account the zone inside the pore
with kB the Boltzmann constant and Z the solvent in which the ion center cannot penetrate because
viscosity. of its nite size) and the pore cross section
The meaning and relevance of hindrance (Fig. 2). The steric partitioning coefcient is
factors (Ki,c and Ki,d) have been widely bounded between 0 and 1, Fint i 1 for point
documented by Deen (Deen 1987) as well as solutes, whereas Fint i 0 for solutes larger than
by many other authors (Mason et al. 1980; the pore. It should be noted that for a neutral
Deen et al. 1980; Nakao and Kimura 1982; solute, the partition coefcient is written as:
i Fi (due to its zero charge).
Gint int
Bowen and Sharif 1994).
The partition coefcient of a solute i at the int
DcD is the Donnan potential involved in the
membrane/external solution interfaces electric exclusion mechanism (Fig. 1). By deni-
i =ci ), which is derived by
(dened as cint int
tion, it has the same sign as the membrane charge.
assuming local thermodynamic equilibrium, As shown by the second term on the r.h.s. of Eq. 3,
is given by: the Donnan potential is repulsive for co-ions
582 Donnan Steric Pore (DSP) and Dielectric Exclusion (DE) Model
ri, Stokes
Available pore area
for transport
rp
rp
Pore area
i = (1i)2 i = (1i)
Donnan Steric Pore (DSP) and Dielectric Exclusion (DE) Model, Fig. 2 Schematic representation of available pore
area for transport in a cylindrical (a) and slit-like (b) pore. Fi: steric partitioning coefcient; li = ri,Stokes/rp
(i.e., ions having the same charge polarity as the multivalent ions are more strongly rejected than
membrane) and attractive for counterions (i.e., monovalent ones. Moreover, the Born effect does
ions with the charge polarity opposite to that of not depend on pore size and geometry.
the membrane). The interaction free energy of an ion inside a
Finally, the terms DW0i,Born
int
and DW0i,image
int
de- pore of nanometric dimensions is also affected by
note the increase in the interaction energy (scaled the interaction of the ion with the polarization
on kBT) of ion i due to the Born and image force charge induced by the ion itself at the interface
dielectric effects, respectively. between the membrane matrix and the solution
The Born dielectric effect is connected with the inside the pore (Parsegian 1969; Dresner 1974;
lowering of the dielectric constant of the solution Glueckauf 1976). This phenomenon results from
inside pores ep (with respect to the corresponding the difference in dielectric constants between the
value in the external bulk solution eb) caused by membrane and the solution and is usually
an alteration of the solvent structure when it is described as the production of image forces
conned in nanodimensional pores. The variation (Dukhin et al. 1988) since the interaction between
of the interaction energy resulting from the Born the ion and the polarized interface is formally
effect is given by: equivalent to the interaction with a ctitious
image charge located at the other side of the

z i e 2 1 1 interface at the same distance from it as the ion
DW0i , int
Born (4)
8pe0 kB Tri, cav ep eb (Fig. 3). This topic has been reviewed recently by
Yaroshchuk (2000) who derived approximate
where e0 is the vacuum permittivity and ri,cav is the expressions for the interaction energy in pores of
cavity radius of ion i. cylindrical and slit-like geometries. These equa-
Equation 4 is similar to the relation derived in tions, which are complex, account for the screen-
the original Born model (Born 1920) for ing of the interaction between the ion and the
ionsolvent interactions except that the radius of polarized interface by the membrane xed charge
the cavity formed by the ion i in the solvent (ri,cav) and the ionic atmosphere as well.
is used here (Rashin and Honig 1985; Hyun and Like the Born effect, this dielectric effect is also
Ichiye 1997; Qiu et al. 1997) instead of the ionic controlled by the square of the ion charge. Conse-
radius as stated in the original Born model (which quently, it does not differentiate (qualitatively)
is known to overestimate solvation enthalpies, between co-ions and counterions (unlike the
especially for cations). Donnan exclusion which repels co-ions but attracts
As shown by Eq. 4, the partition due to the counterions), and both co-ions and counterions are
Born DE is unfavorable for any ion, indepen- excluded from the membrane pores provided
dently of the ion polarity (provided ep < eb), and em < ep.
Donnan Steric Pore (DSP) and Dielectric Exclusion (DE) Model 583
Solution inside the pore Membrane matrix

m
p (p > m)
Induced polarisation
charges
Ions Image
charges
+ + +
Donnan Steric Pore (DSP) and Dielectric Exclusion discontinuities. It should be noted that the physical effect
(DE) Model, Fig. 3 Schematic representation of the inter- is the induced polarization charges at the interface whereas
action of ions with their polarization charge induced by the the image charge is just a tool to construct an equivalent
ions themself at the interface between the solution inside system where the net interaction is the same (Jonssn and
the pore and the membrane matrix. Image charges can be Wennerstrm 1983)
used to visualize the force arising from the dielectric
Last, it should be noted that DE (Born and Deen WM (1987) Hindered transport of large molecules in
image forces) plays a relevant role even in the liquid-lled pores. AIChE J 33:14091425
Deen WM, Savat B, Jamieson M (1980) Theoretical model
case of uncharged membranes for which Donnan for glomerular ltration of charged solutes. Am
exclusion is zero. J Physiol 38:10871091
Dresner L (1974) Ion exclusion from neutral and slightly
charged pores. Desalination 15:3957
Dukhin SS, Churaev NV, Shilov VN, Starov VM
Cross-References (1988) Modelling reverse osmosis. Russ Chem Rev
57:572
Donnan Potential Glueckauf E (1976) The distribution of electrolytes
between cellulose acetate membranes and aqueous
Nanoltration solutions. Desalination 18:155172
Hyun JK, Ichiye T (1997) Understanding the Born radius
via computer simulations and theory. J Phys Chem
References B 101:35963604
Jonssn B, Wennerstrm H (1983) Image-charge forces in
Born M (1920) Volumen and hydratationswarme der phospholipid bilayer systems. J Chem Soc Faraday
ionen. Z Phys 1:4548 Trans 2(79):1935
Bowen WR, Mohammad AW (1998) Characterization and Mason EA, Wendt RP, Bresler EH (1980) Similarity rela-
prediction of nanoltration membrane performance a tions (dimensional analysis) for membrane transport.
general assessment. Inst Chem Eng 76A:885893 J Membr Sci 6:283298
Bowen WR, Mukhtar H (1996) Characterization and pre- Nakao S, Kimura S (1982) Models of membrane transport
diction of separation performance of nanoltration phenomena and their applications for ultraltration
membranes. J Membr Sci 112:263274 data. J Chem Eng Jpn 15:200205
Bowen WR, Sharif AO (1994) Transport through Parsegian A (1969) Energy of an ion crossing a low dielec-
microltration membranes particle hydrodynamics tric membrane: solutions to four relevant electrostatic
and ux reduction. J Colloid Interface Sci 168:414421 problems. Nature 221:844846
Bowen WR, Mohammad AW, Hilal N (1997) Characteri- Qiu D, Shenkin PS, Hollinger FP, Still WC (1997) The
zation of nanoltration membranes for predictive GB/SA continuum model for solvation. A fast analyt-
purposes use of salts, uncharged solutes and atomic ical method for the calculation of approximate Born
force microscopy. J Membr Sci 126:91105 radii. J Phys Chem A 101:30053014
584 Dope Solution
Rashin AA, Honig B (1985) Reevaluation of the Born generally has a much greater viscosity and elas-
model of ion hydration. J Phys Chem 89:55885593 ticity than that used for the fabrication of at sheet
Szymczyk A, Fievet P (2005) Investigating transport prop-
erties of nanoltration membranes by means of a steric, membranes by the phase inversion technique.
electric and dielectric exclusion model. J Membr Sci The used polymer limits the solvent(s) and the
252:7788 non-solvent(s) that can be used to prepare the
Szymczyk A, Aoubiza B, Fievet P, Pagetti J (1999) Elec- dope solution for the phase inversion process.
trokinetic phenomena in homogeneous cylindrical
pores. J Colloid Interface Sci 216:285296 Water is frequently used as a non-solvent. For
Yaroshchuk AE (2000) Dielectric exclusion of ions from example, among others, the solvents dimethyl
membranes. Adv Colloid Interf Sci 85:193230 formamide (DMF) and dimethyl acetamide
(DMAC) were used to prepare dope solutions
from polysulfone, poly(ether sulfone), poly(vinyl-
idene uoride), polyacrylonitrile, cellulose ace-
Dope Solution tate, polyimide, or poly(ether imide).
To prepare membranes by thermal precipita-
Khayet Mohamed tion, for example, thermally induced phase inver-
Department of Applied Physics, Universidad sion (TIPS) membranes, the dope solution is a
Complutense de Madrid, Madrid, Spain polymer/diluent system adequate to form a homo-
geneous melt blend, which will be cooled or
quenched to induce phase separation (i.e.,
Dope solution is a homogeneous thermodynami- liquidliquid phase separation, solidliquid
cally stable polymer solution formed by a polymer phase separation, or combined liquidliquid and
or a copolymer, solvent or a solvent mixture, solidliquid phase separation). For TIPS mecha-
and/or additive(s) (i.e., the additive can be a sec- nism the diluent should be miscible with the con-
ond polymer, a non-solvent, organic or inorganic sidered polymer at high temperatures, should have
particles of different sizes, etc.). It is also possible a relatively low molecular weight together with a
to add a non-solvent to the polymer solution to low volatility at high temperatures, and should be
such an extent that all the components are still thermally stable at high temperatures. For
miscible. In this case, care must be taken because example, the used diluent for poly(vinylidene
demixing may occur by the addition of the uoride), PVDF, is dimethyl phthalate, dioctyl
non-solvent to the polymer solution so that it phthalate, dioctyl sebacate, dibutyl phthalate,
becomes thermodynamically unstable. Some- g-butyrolactone, propylene carbonate, or dibutyl
times nanoparticles or nanotubes are dispersed in sebacate.
the polymer solution. The parameters affecting the physical and
To prepare a dope solution, the polymer must chemical characteristics of the dope solution are
be soluble in a solvent or a solvent mixture. The the polymer, the solvent(s), the non-solvent addi-
solvent has a close solubility parameter with the tive(s), and their concentrations as well as the
polymer. The solvent and the non-solvent must be temperature. These characteristics affect the
miscible and the viscosity of the dope solution structural morphology of the formed at sheet
must be high (i.e., more than 100 P). membranes and hollow bers. The polymer con-
After preparing the polymer solution, this centration in the dope solution is one of the rele-
should be ltered to remove impurities and insol- vant factors affecting membrane structure. For
uble contaminants. Subsequently, the resulted example, the increase of the polymer concentra-
dope solution should be degassed before loading tion in the dope solution results in a high-volume
into the dope vessel for fabrication of either at fraction of polymer and consequently a low poros-
sheet or hollow ber membranes. The spinning ity and small pore size. A signicant reduction of
dope suitable for hollow ber fabrication by the porosity was observed with increasing PVDF
dry/wet spinning or wet spinning techniques concentration in the spinning dope from 15 to
Dope Solution 585
20 wt% (Wu et al. 2007). However, over this P

PVDF concentration range, no apparent effect on
the mean pore size of the prepared hollow bers
was observed. Fabrication of membranes by the Miscible
phase inversion technique from dope solutions region
with a very low polymer concentration is practi-
cally impossible. For instance, when the polymer Dope
solution a b c d Membrane
content is less than a threshold value, for example,
10 wt% for PVDF, the resulting membranes D
become inconsistent and holes start to appear Tie Immiscible
line region
within the membrane (Ortiz de Zrate
et al. 1995). In the phase inversion process, S NS
non-solvent additive in the dope solution acts as
Dope Solution, Fig. 1 Schematic phase diagram of an
a pore-forming agent improving the permeability isothermal ternary system (polymer, P; solvent, S and
of the fabricated phase inversion membrane. In non-solvent, NS)
general, when the concentration of the
non-solvent additive is increased in the dope solu- composition will be displaced to the right side in
tion, the pore size and porosity of the prepared the phase diagram toward the P-NS axis. Different
membrane increase (Khayet and Matsuura 2001). composition paths can occur (Khayet
The maximum amount of non-solvent that can and Matsuura 2011). For example, the path (a, b,
be added to a dope solution can be determined c, d) starts from a dope solution with a low initial
from the ternary phase diagram presenting the polymer volume fraction on the P-S axis and the
composition of the polymer, solvent, and the ratio of the rate of solvent outux to that of the
non-solvent. In this case, the dope solution must non-solvent inux is relatively high. When the
be in one phase region where all the components dope solution crosses the binodal, two separate
are completely miscible with each other. Figure 1 phases will begin to form. Those are a polymer-
shows an isothermal ternary system (polymer, P; rich phase represented by the upper end of the tie
solvent, S, non-solvent, NS). The pure compo- line and a polymer-poor phase represented by the
nents (P, S, NS) are localized at the corners of lower end of the tie line. If the amount of the
the triangle, and any point within the triangle polymer-rich phase is large relative to the amount
represents a mixture of all three components. of the polymer-poor phase, droplets of the latter
Within a certain compositionally dened range phase are dispersed in the continuous polymer-
of thermodynamic states, all three components rich phase and form pores in the continuous poly-
are completely miscible, whereas in a different mer matrix. The obtained nal membrane at the
range, the system decomposes into two distinct P-NS axis is asymmetric (i.e., a dense surface
phases. The solid black line is the liquidliquid layer at the lm/gelation media interface followed
boundary, called the binodal which separates the by a porous sub-layer). The lm composition
miscible region at the left from the immiscible starts from the point a, passes through points
region at the right. The thermodynamic parameter b and c (polymer densication point), and
describing the miscibility of two or more compo- nally reaches the point d along the composi-
nents (e.g., polymer and diluent) is the Gibbs free tion path. b is the point where the phase sepa-
energy of mixing, for which an expression was ration starts and c is the point where the polymer
given by the Flory-Huggins lattice theory (Flory solidication starts to occur. Exchange of solvent
1965). and non-solvent will lead to the nal formation of
During membrane preparation, the dope solu- the membrane, the porosity of which is deter-
tion at the left side of the binodal in the P-S axis is mined by the point on the P-NS line (e.g., point
immersed in a non-solvent so that the nal d) (Khayet and Matsuura 2011).
586 Double Emulsion
References can serve as an entrapping reservoir for active

ingredients that can be released from the inner
Flory PJ (1965) Principles of polymer chemistry. Cornell phase to the outer phase by a controlled and a
University Press, Ithaca
sustained mechanism. The major area of applica-
Khayet M, Matsuura T (2001) Preparation and characteri-
zation of polyvinylidene uoride membranes for tions concerns the human pharmaceutical eld:
membrane distillation. Ind Eng Chem Res water-in-oil-in-water (W/O/W) emulsions have
40:57105718 been investigated as potential vehicles for various
Khayet M, Matsuura T (2011) Membrane distillation: prin-
hydrophilic drugs (vaccines, vitamins, enzymes,
Ortiz de Zrate JM, Pea L, Mengual JI (1995) Character- hormones) which would be then slowly released.
ization of membrane distillation membranes prepared Also, double emulsions have promising applica-
by phase inversion. Desalination 100:139148 tions in other elds like the food industry (low
Wu B, Li K, Teo WK (2007) Preparation and characteriza-
calorie products, improved sensoric characteris-
tion of poly(vinylidene uoride) hollow ber mem-
branes for vacuum membrane distillation. J Appl tics, taste masking), cosmetic industry (easily
Polym Sci 106:14821495 spreadable creams with encapsulated ingredients
in both water and oil phases), separation science,
and agriculture.
In practice, double emulsions are thermody-
Double Emulsion namically unstable systems with a strong tendency
for coalescence, occulation, and creaming. Most
Na Li double emulsions consist of relatively large drop-
Department of Chemical Engineering, School of lets, cannot withstand storage regimes, and have a
Chemical Engineering and Technology, Xian strong tendency to release the entrapped matter in
Jiaotong University, Xian City, China an uncontrolled manner. Four possible mecha-
nisms are described for instability of W/O/W emul-
sion: (1) coalescence of the internal aqueous
Synonyms droplets, (2) coalescence of the oil drops, (3) rup-
ture of the oil lm separating the internal and
Multiple emulsions external aqueous phases, and (4) passage of water
(and water-soluble material, e.g., a drug) to and
Double emulsion is a complex system of an emul- from internal droplets through the oil layer (Graaf
sion in an emulsion, where the droplets of the et al. 2005). Much work has been devoted in the
dispersed phase themselves contain even smaller last decade in order to improve the stability of the
dispersed droplets. An emulsion is a dispersion of multiple emulsions and the control of the release
droplets of one liquid in another one with which it rates of the addenda. The most recent achievements
is incompletely miscible. are the use of polymeric emulsiers to improve
Two main types of double emulsion can be interface coverage and to better anchor into the
distinguished: water-in-oil-in-water (W/O/W) dispersed phase and the decrease of droplet size
emulsion, in which a W/O emulsion is dispersed by improving methods of formation (Garti 1997).
as droplets in an aqueous phase, and oil-in-water- Double emulsion is generally prepared in a
in-oil (O/W/O) emulsion, in which an O/W emul- two-step emulsication process using two surfac-
sion is dispersed in an oil phase. W/O/W emulsion tants of opposite solubility. To produce O/W/O
is more common than O/W/O emulsion. Double emulsion, a W/O emulsion is rst prepared with
emulsion contains more interface and is even a large excess of hydrophobic surfactant (HLB
more thermodynamically unstable than single <10), by strong homogenization, to form the
emulsion (Graaf et al. 2005). smallest possible droplets. In the second step a
Double emulsions have signicant potential in W/O/W double emulsion is further formed by
many applications since, at least in theory, they gentle addition of the W/O (termed inner
Double Emulsion Production, Membrane Operations for 587
emulsion) to water in which a hydrophilic one pharmacology industry, and other elds like agri-
(HLB >10) has been dissolved (Bibette culture and separation science.
et al. 1999). Double emulsions are usually prepared in a
The most important conventional emulsica- two-step emulsion process. First, internal emul-
tion devices are stirring apparatuses, rotorstator sion (e.g., water-in-oil (W/O)) is designed under
systems, and high-pressure homogenizers. Mem- high-shear conditions to obtain small droplets.
brane emulsication is a relatively new method Secondary emulsion step for double emulsion
for the production of emulsions that has received (e.g., water-in-oil-in water (W/O/W)) is carried
increasing attention over the last 15 years. The out with mild shear to avoid rupture of the internal D
technique is attractive given the low energy con- droplets. Usually, the primary step can be pro-
sumption, the better control of droplet size and duced by conventional method or also by mem-
droplet size distribution and especially the mild- brane emulsication, while the membrane
ness of the process. Microchannel emulsication emulsication is especially suitable for the second
is a novel method for producing monodisperse emulsication step.
emulsions. It is even more monodisperse than For the conventional emulsication method,
usually obtained with membrane emulsication the most commonly used devices are rotorstator
(Graaf et al. 2005). systems and high-pressure homogenizers. In
rotorstator systems (i.e., colloid mills), a high
shear is generated between a rotor and a stationary
References smooth, roughened, or grooved surface. Turbu-
lence is the primary cause of uid disruption lead-
Bibette J, Calderon FL, Poulin P (1999) Emulsions: basic ing to the formation of droplets (Graaf et al. 2005).
principles. Rep Prog Phys 62:9691033
In a high-pressure homogenizer, the droplets are
Garti N (1997) Double emulsions-scope, imitations and
new achievements. Colloid Surf A 123124:233246 broken up in a nozzle by turbulence and cavita-
Graaf S, Schron CGPH, Boom RM (2005) Preparation of tion. This process can be assisted with the use of
double emulsions by membrane emulsication-a an ultrasonic or electric elds. Due to their high
review. J Membr Sci 251:715
throughput, rotorstator systems and high-
pressure homogenizers are suitable for large-
scale production of emulsions. However, energy
requirements of these systems are high because
Double Emulsion Production, only a small part of the energy input is used for
Membrane Operations for droplet breakup; most of the energy supplied is
lost and converted into heat. Furthermore, shear
Na Li stresses these systems applied may cause loss of
Department of Chemical Engineering, School of functional properties of heat- and shear-sensitive
Chemical Engineering and Technology, Xian components. In addition, they show poor control
Jiaotong University, Xian City, China over droplet size and distribution.
In order to overcome the problems of tradi-
tional emulsication methods, a new technique
Double emulsions are emulsions of emulsions. for making emulsions known as membrane
The droplets of the dispersed phase themselves emulsication arouses peoples interest. This
contain even smaller dispersed droplets. Two method involves using a low pressure to force
main types of double emulsions are water-in-oil- the dispersed phase to permeate into the continu-
in-water (W/O/W) emulsions and oil-in-water-in- ous phase through a membrane which has a uni-
oil (O/W/O) emulsions. The former is more com- form pore-size distribution. Droplets grow at pore
mon. Double emulsions have shown signicant outlets and then detach when reaching a certain
promise in the food industry, cosmetic industry, size (Joscelyne and Tragardh 2000). One of the
588 Drinking Water Treatment
most signicant advantages of membrane emulsi- References

cation is the low energy consumption. Also the
better control of droplet size and distribution Graaf S et al (2005) Preparation of double emulsions by
membrane emulsication-a review. J Membr Sci
and the mildness of the process are highly
251:715
attractive. Compared to conventional methods, Joscelyne SM, Tragardh G (2000) Membrane
a disadvantage of membrane emulsication is emulsication-a literature review. J Membr Sci
that the ux of the dispersed phase through the 169:107117
membrane is fairly low. However, the ux of
dispersed phase could be increased by using a
membrane with a low hydraulic resistance
(e.g., microsieves). Drinking Water Treatment
The commonly used membranes for mem-
brane emulsication are Shirasu porous glass A. Moura Bernardes
(SPG) membranes, ceramic aluminum oxide Universidade Federal do Rio Grande do Sul
(a-Al2O3) membranes, a-alumina- and zirconia- (PPGE3M-UFRGS), RS, Porto Alegre, Brazil
coated membranes, macroporous silica glass
membranes, micro-fabricated metal membranes,
polytetrauoroethylene (PTFE) membranes, and Drinking water treatment includes the pro-
hydrophobized silicon nitride microsieves. cesses and operations applied in order to obtain
For membrane emulsication two methods of an adequate and continuous supply of clear, col-
operation are used: cross-ow membrane emulsi- orless, odorless, palatable, and safe water.
cation and premix membrane emulsication. The drinking water treatment requirements dif-
Generally, the important process controlling fer according to the feed water quality. Many
parameters are membrane pore size, porosity, aquifers and isolated surface waters are high in
thickness, and surface type, the velocity of con- water quality and may be pumped from the supply
tinuous phase, the transmembrane pressure, and and transmission network directly to any number
wall contact angle. It has been found that it is a of end uses, including human consumption.
linear relationship between the droplet size and Source protection, i.e., conservation and protec-
membranes pore size. At high porosities, it is tion of catchment areas, although not strictly in the
more likely for droplets to coalesce at the mem- category of treatment, can play a critical role in
brane surface before they detach. If the porosity supporting the quality of water supplies (Sullivan
is too low, the dispersed phase ux may be not et al. 2005). However, clean water sources are the
sufcient for viable production of emulsions. exception in many parts of the world, particularly
The thinner the membrane, the larger the ux regions where the population is dense or where
through the pores will be, which leads to a there is heavy agricultural use. In these places, the
higher droplet expansion rate. It has been water supply must receive varying degrees of
observed that hydrophilic membranes are t for treatment before distribution (Peirce et al. 1998).
making O/W emulsions and hydrophobic mem- The most common water treatment process,
branes for W/O emulsions. The droplet size known as conventional treatment, consists of
decreases sharply as the cross-ow velocity coagulation, occulation, sedimentation, ltra-
increases from rest and reaches a size where it tion, and disinfection. Taken together, coagula-
becomes more or less independent of the ow tion, occulation, sedimentation, and ltration
velocity. The permeating ux is increased by effectively remove many contaminants. Equally
increasing the transmembrane pressure. Another important, they reduce turbidity, yielding water of
important parameter affecting the droplet size is good clarity and hence enhanced disinfection ef-
the wall contact angle. Higher wall contact ciency. If not removed by such methods, particles
angle results in larger droplets and a more poly- may harbor microorganisms and make nal disin-
disperse emulsion. fection more difcult. Generally, disinfection is
Drinking Water Treatment 589
accomplished through the addition of an oxidant. used for desalinating brackish water or seawater
Chlorine is by far the most common disinfectant (Baker 2012). Competing with RO, electrodialy-
used to treat drinking water, but other oxidants, sis (ED) also plays an important role on the treat-
such as chloramines, chlorine dioxide, and ozone, ment of water in a very large market, a
are also used. Although ultraviolet (UV) light can competition that is continuously intensied by
be used, it does not leave a residual, and usually a the increased shortage of water resources. Both
secondary disinfectant (i.e., chlorine) is added methods have advantages and disadvantages in
(Pepper et al. 2008). drinking water treatment installations, and they
In many cases, an additional treatment is nec- have to be considered in order to choose the D
essary to remove suspended, colloidal, and most optimal concept of membrane desalination.
dissolved constituents remaining after conven- Membrane fouling is one of the most important
tional treatment. Dissolved constituents may factors that limit greater use of desalination mem-
range from relatively simple inorganic ions, such branes. Fouling occurs due to particulate matter,
as calcium, potassium, sulfate, nitrate, and phos- organic matter, microorganisms forming biolms,
phate, to an ever-increasing number of highly and inorganic scaling (Escobar 2010). Bacterial
complex synthetic organic compounds contamination problems have been reported as
(Tchobanoglous et al. 2003). one of the main causes of membrane fouling in
The ability of a water treatment facility to osmosis. This is caused by the characteristic of the
deliver a wholesome and high-quality product to membrane that serves as a barrier between the
its consumers usually occurs when advanced feed water and the product, which not only
treatment methods are used. Activated carbon or removes dissolved solids, but also bacteria,
granulated activated carbon (GAC) has been used viruses, and insoluble substances. ED only
to remove organic chemicals from water since the removes ions. Therefore, any bacteria, colloidal
early 1900s. Ion exchange resins have been material, or silica present in the feed water stream
employed to remove inorganic metals for decades. will remain in the product stream. This fact repre-
Ozonation has been successfully used, instead of sents an advantage in terms of membrane fouling.
chlorine and bromine, to both disinfect water and On the other side, it can be seen as a major disad-
oxidize organic constituents. Advanced oxidation vantage, especially for the production of potable
processes have been used to remove persistent water, since microorganisms or organic contami-
compounds. Ultraviolet light is being evaluated nants are not eliminated.
at a number of municipal treatment facilities for By using ED technology, fouling can be min-
bacterial and viral control. Membrane separation imized, thus decreasing the need for the addition
processes are nowadays increasingly applied. As of chemical products, by reversing the polarity of
far back as 1985, the French have made use of the system with electrodialysis reversal (EDR).
membrane separation technology (i.e., reverse By reversing the polarity (and the solutions ow
osmosis) for the removal of organic and inorganic direction) several times per hour, ions move in the
compounds. Today, a number of reverse osmosis opposite direction through the membranes, mini-
installations are operating worldwide to drinking mizing buildup. This generally minimizes mem-
water treatment. Reverse osmosis membranes can brane fouling, decreasing pretreatment
hold back a wide range of micropollutants includ- requirements in comparison to RO (Seidel
ing pesticides and pathogenic organisms and can et al. 2011).
replace the need for both ozonation and activated In general, electrodialysis reversal (EDR) is
carbon ltration (Sullivan et al. 2005). more attractive for the desalination of brackish
To produce fresh water from seawater and water and in cases where the removal of organic
brackish water, desalination plants in industrial matter and microbial control are not important.
scale are nowadays installed throughout the EDR is also of interest in treating brackish waters
world. Approximately one-half of the reverse where silica is an important limitation (Escobar
osmosis (RO) systems currently installed are 2010). Nevertheless, the operating costs of
590 Drinking Water Treatment
electrodialysis scale almost linearly in proportion Seawater has a salt concentration of 3.24.0 %,
to the salt concentration in the feed. Therefore, depending on the region of the world. Because of
this technology is best suited to low-salt- this high salinity, only RO membranes with salt
concentration feed streams. Reverse osmosis rejections of 99.3 % or more can produce potable
costs also increase signicantly with salt concen- water containing less than 500 ppm dissolved salt
tration but at a lower rate than electrodialysis in a single pass. Application to seawater desalina-
does. The result is that reverse osmosis is the tion of the rst-generation cellulose acetate mem-
lowest-cost process for streams containing branes, with rejections of 9799 %, was limited.
between 3,000 and 10,000 ppm salt. However, With the development of the polyamide hollow
site-specic factors or plant size often make the ne bers and interfacial composites, suitable
technology the best approach for more dilute feed seawater membranes became available, and
water or for streams as concentrated as seawater many plants have been installed. These mem-
(35,000 ppm salt) (Baker 2012). branes can produce permeate water that meets
The salinity of brackish water is usually the WHO standard (<500 ppm salt) in a single
between 2,000 and 10,000 mg/L. The World pass or less than 200 ppm in a two-pass system,
Health Organization (WHO) recommendation typically a single-stage seawater system and a
for potable water is 500 mg/L, so up to 90 % of single-stage brackish water system connected in
the salt must be removed from these feeds. Early series (Baker 2012).
RO cellulose acetate membranes could achieve Recently, membrane separation processes are
this removal easily, so treatment of brackish being studied for the removal of emergent con-
water was one of the rst successful applications taminants (e.g., pharmaceuticals and personal
of reverse osmosis. Several plants were installed care products) from water. Surface and ground-
as early as the 1960s. The osmotic pressure of water are the major renewable resources for sus-
brackish water is approximately 11 psi per tainable drinking water production throughout the
1,000 ppm salt, so osmotic pressure effects do world. However, many chemical constituents that
not generally limit water recovery signicantly. have not been considered historically as contam-
Limitations are generally due to scaling. Typical inants are present not only in wastewater, but also
water recoveries are in the 7090 % range, which in natural waters at global scale. These emerging
means the brine stream leaving the system is up to contaminants (EC) are commonly derived from
ten times more concentrated in calcium, sulfate, municipal, agricultural, and industrial wastewater
and silica ions present in the feed. Disposal of the sources and pathways due to often inadequate
1015 % of the brackish water that remains as treatment in conventional wastewater treatment
concentrated brine represents a signicant prob- plants, being detected in rivers downstream. In
lem. In fact, nowadays, brackish water desalina- fact, the presence of EC can be expected in rivers,
tion is the largest application of electrodialysis. lakes, wells, and even drinking water. Consider-
The competitive technologies are ion exchange ing that conventional water treatment with coag-
for very dilute saline solutions, below 500 ppm, ulation, sedimentation, and ltration achieves low
and reverse osmosis for concentrations above levels of removal for EC, techniques that have
2,000 ppm. In the 5002,000 ppm range, electro- been gaining attention in the past few years are
dialysis is often the low-cost process (Baker pressure-driven membrane processes
2012). An advantage of electrodialysis compared nanoltration (NF) and reverse osmosis (RO).
to reverse osmosis is that a much higher brine These two treatments seem to be able to effec-
concentration can be achieved, since there are no tively remove most organic and inorganic com-
osmotic pressure limitations (Strathmann 2010). pounds and microorganisms from raw water, and
Typically 8095 % of the brackish feed is recov- their application in drinking water treatment has
ered as product water. The degree of water recov- been seen as preferred alternatives for the removal
ery is limited by precipitation of insoluble salts in of emerging organic contaminants (Quintanilla
the brine (Baker 2012). 2010; Radjenovi et al. 2008).
Droplet Size 591
References Mean: average size of a population

Median: size where 50 % of the population is
Baker RW (2012) Membrane technology and applications. below/above
Wiley, Chichester/Hoboken
Mode: size with highest frequency
Escobar IC (2010) Chapter 14 conclusion: a summary of
challenges still facing desalination and water reuse.
P. 389397. In: Escobar IC, Schfer AI (eds) Sustain- There are many different means that can be
able water for the future: water recycling versus desa- dened depending upon how the droplet size dis-
lination, vol 2. Elsevier, Amsterdam/Boston, pp 1416
tribution is analyzed. The three most commonly
Peirce JJ, Weiner RF, Vesilind PA (1998) Environmental pol-
used for particle sizing are: D
lution and control, 4th edn. Elsevier, Oxford, UK. 392p
Pepper I, Gerba C, Gentry T, Maier R (2008) Environmen-
tal microbiology. Academic Press, London. 624p Surface area mean D[3, 2] (Sauter mean
Quintanilla VA (2010) Rejection of emerging organic con-
diameter)
taminants by nanoltration and reverse osmosis mem-
branes: effects of fouling, modelling and water reuse. It is the weighted average surface diameter,
PhD thesis, Delft University of Technology assuming spherical particles of the same sur-
Radjenovi J, Petrovi M, Ventura F, Barcel D (2008) face area as the actual particles:
Rejection of pharmaceuticals in nanoltration and
reverse osmosis membrane drinking water treatment. X
Water Res 42(14):36013610 D3i ni
Seidel C, Gorman C, Darby JL, Jensen VB (2011) An D3, 2 X 2
assessment of the state of nitrate treatment alternatives. D i ni
Final report. The American water works association,
inorganic contaminant research and inorganic water It is most relevant where specic surface
quality joint project committees. 136p
Strathmann H (2010) Electrodialysis, a mature technology
area is important, e.g., bioavailability, reactiv-
with a multitude of new applications. Desalination ity, and dissolution. It is most sensitive to the
264:268288 presence of ne droplets in the size
Sullivan PJ, Agardy FJ, Clark JJJ (2005) Water protection. In: distribution.
The environmental science of drinking water. Butterworth
Heinemann, Amsterdam/Boston, pp 89141
Volume moment mean D[4, 3] (De Brouckere
Tchobanoglous G, Burton FL, Stensel HD, Metcalf & mean diameter)
Eddy, Inc (2003) Wastewater engineering: treatment It is the weighted average volume diameter,
and reuse. McGraw-Hill, Boston, 1819p assuming spherical particles of the same vol-
ume as the actual particles:
X
D4i ni
D4, 3 X 3
Droplet Size D i ni
Emma Piacentini It is relevant for many samples as it reects the

Institute on Membrane Technology, National size of those particles which constitute the bulk of
Research Council of Italy, ITM-CNR, Rende, the sample volume. It is most sensitive to the
Italy presence of large particulates in the size
distribution.
Number length mean D[1,0]
Droplet size is dened by the diameter of an It is often referred to as the arithmetic mean:
equivalent sphere having the same property
(such as volume or mass) as the analyzed droplet. X
Di ni
In order to simplify the interpretation of particle D1, 0 X
size distribution data, a range of statistical param- ni
eters can be opportunely selected for any given
sample. The most common droplet size in a drop- It is most important where the number of par-
let size distribution could be reported as: ticles is of interest, e.g., in particle counting
592 Drug Delivery by Membrane Operations
applications. It can only be calculated if we know body by various anatomical routes. The choice of
the total number of particles in the sample and is the route of administration depends on the disease,
therefore limited to particle counting applications. the effect desired, and the product available. Drugs
For volume-weighted particle size distribu- may be administered directly to the organ affected
tions, such as those measured by laser diffraction, by the disease or given systemically and targeted to
diameter is reported based upon the maximum the organ. Methods of systemic drug delivery are
particle size for a given percentage volume of classied as follows: gastrointestinal system (oral,
the sample. The most common diameters are the rectal), parenteral, transnasal, pulmonary (drug
Dv10, Dv50, and Dv90 that are corresponded to delivery by inhalation), transdermal, and intra-
10, 50, and 90 vol% on a relative cumulative osseous infusion. The oral route of drug administra-
droplet size curve, respectively. tion is the most used for both conventional as well
as novel drug delivery. The reasons for this prefer-
ence are the ease of administration and widespread
acceptance by patients.
Drug Delivery by Membrane Membrane devices used to deliver the drug
Operations consist basically of a drug reservoir (Stamatialis
et al. 2008). Two main types of systems can be
Catherine Charcosset distinguished: (1) osmotic membrane systems and
Universit Lyon 1, Lyon, Villeurbanne, France (2) diffusion-controlled membrane systems. An
osmotic system consists of a reservoir made of a
polymeric membrane permeable to water but not to
Drug delivery is a eld of vital importance to med- the drug (semipermeable membrane) Fig. 1. The
icine and healthcare. A drug delivery system is a reservoir contains a concentrated drug solution. As
formulation or a device that enables the introduction water crosses the membrane due to osmotic pres-
of a therapeutic substance in the body and improves sure, the drug solution is released through the
its efcacy and safety by controlling the rate, time, orice. In diffusion-controlled membrane systems,
and place of its release (Jain 2008). It includes the the drug release is controlled by transport of the
administration of the therapeutic product, the drug across a membrane. The transport is depen-
release of the active ingredients by the product, dent on the drug diffusivity through the membrane
and the subsequent transport of the active ingredi- and the thickness of the membrane. These systems
ents across the biological membranes to the site of nd broad application in pills, implants, and
action. A drug delivery system may be (1) a device patches (see Drug-Delivery Applications).
used to deliver the drug or (2) a formulation of the Membrane-based formulations used to deliver
drug which will be administered for a therapeutic the drug consist of carriers that are biodegradable,
purpose. Drugs may be introduced into the human biocompatible, targeting, and stimulus responsive
Drug Delivery by Drug solution

Membrane Operations, Orifice
Fig. 1 Schematic view of
an osmotic drug delivery
system
Water and
concentrated
solution
Semi permeable
membrane
Drug Delivery, Application of 593
like liposomes and nanocapsules (Zhang transdermal drug delivery (passive diffusion, ion-
et al. 2013). Liposomes are enclosed spherical tophoresis, and skin or device-controlled deliv-
vesicles that are organized in one or several con- ery) (Grassi 2009). These drug delivery
centric phospholipid membrane bilayers with an applications concern mainly transdermal, oral,
aqueous phase inside. Liposomes can entrap and ocular delivery routes of administration.
hydrophilic pharmaceutical agents in their inter- In transdermal drug delivery (Stamatialis
nal aqueous compartment and lipophilic drugs et al. 2008), the drug is incorporated into a patch
within the lipid membrane. Nanoparticles and delivered through the skin due to the concen-
(nanospheres and nanocapsules) are ranging in tration difference or electrical current (Fig. 1). D
size from about 10 to 1000 nm. Nanospheres Transdermal patches are used as a convenient
have a matrix type structure with drugs adsorbed and effective method of drug delivery through
at their surface, entrapped in the particle, or the human skin. A membrane may exclusively
dissolved in it. Nanocapsules have a polymeric control or partially control drug transport together
shell membrane and an inner liquid core, the drugs with other components (such as the adhesive
being dissolved in the inner core or adsorbed at layer). Patches are commercialized, for example,
their surface. Nanoparticles have been investi- for delivery of nicotine for smoking cessation,
gated for the entrapment of a wide variety of estrogen or testosterone for hormonal therapy,
drugs, for applications ranging from ophthalmic and methylphenidate for attention-decit/
delivery to carriers in chemotherapy. hyperactivity disorder. Ocular patches are
membrane-controlled reservoir systems. The
drug, accompanied by carriers, is captured in a
Cross-Reference
thin layer between two transparent, polymer
membranes, which control the rate of drug
Drug Release
release. The device is placed on the eye where it
oats on the tear lm. The drug slowly diffuses to
References the target area.
As a noninvasive transdermal drug delivery
Jain KK (2008) Chapter 1: Drug delivery systems an method, iontophoresis applies electrical current
overview. In: Jain KK (ed) Methods in molecular biol-
to deliver drugs through the skin to the underlying
ogy, vol 437, Drug delivery systems. Humana Press,
Totowa, pp 150 tissue or capillaries and then to the whole circu-
Stamatialis DF, Papenburg BJ, Girones M, Saiful S, lating system (Fan et al. 2008). A voltage applied
Bettahalli SNM, Schmitmeier S, Wessling M (2008) between two electrodes immersed in a drug solu-
Medical applications of membranes: drug delivery,
tion causes the drug to be moved from the donor
articial organs and tissue engineering. J Membr Sci
308:134 part into the skin. The positively charged elec-
Zhang Y, Chan HF, Leong KW (2013) Advanced materials trode (anode) attracts the negatively charged
and processing for drug delivery: the past and the drug ions; the negatively charged electrode
future. Adv Drug Deliv Rev 65:104120
(cathode) attracts the positively charged ions.
These devices are commercialized for example
for delivery of lidocaine and epinephrine in local
Drug Delivery, Application of analgesia.
In oral delivery, the drug is pressed into a tablet
Catherine Charcosset which is coated with a nondigestible hydrophilic
Universit Lyon 1, Lyon, Villeurbanne, France membrane such as cellulose esters like cellulose
acetate, cellulose acetate butyrate, cellulose triac-
etate, and ethyl cellulose (Stamatialis et al. 2008;
Drug delivery membrane systems include osmotic Grassi 2009). Once the membrane gets hydrated, a
membrane systems, diffusion-controlled mem- viscous gel barrier is formed, through which the
brane systems (pills, implants, patches), and drug slowly diffuses. The release rate of the drug
594 Drug Permeability
Drug Delivery, Drug-impermeable

Application of, backing layer
transdermal membrane
delivery systems
Drug reservoir
Adhesive
Skin Membrane
is determined by the type of membrane used. The

key feature of these systems is that drug release is Drug Permeability
independent of pH and hydrodynamics of the
external dissolution medium. The result is a robust Catherine Charcosset
dosage form for which the in vivo rate of drug Universit Lyon 1, Lyon, Villeurbanne, France
release is comparable to the in vitro rate, produc-
ing an excellent in vitro/in vivo correlation.
Another key advantage is that they are applicable An important property of a drug is its bioavail-
to drugs with a broad range of aqueous solubility ability. For an orally administered drug, the bio-
such as diltiazem HCl, carbamazepine, metopro- availability is mainly limited by its solubility and
lol, oxprenolol, nifedipine, and glipizide. permeability through the intestinal epithelia. It is
Another example is implants which consist of a therefore important to have access to drug perme-
membrane reservoir containing a drug in liquid or ability. Today the most often used methods are the
powder form (Grassi 2009). The drug slowly dif- Caco-2 model and the parallel articial membrane
fuses through the nondegradable membrane that permeation assay (PAMPA).
usually consists of silicone and ethylene-co-vinyl Caco-2 cells are human colon epithelial cancer
acetate. By properly choosing membrane thickness cell lines used as a model of human intestinal
and permeability, release kinetics can be controlled permeation of drugs and other compounds (van
and designed for a specic therapeutic target. For Breemen and Li 2005). When cultured as a mono-
example, an implantable 5-year contraceptive per- layer, Caco-2 cells differentiate to form tight junc-
mits the long-term release of levonorgestrel. tions between cells to serve as a model of
paracellular movement of compounds across the
monolayer. Caco-2 cell monolayers are usually
References cultured on semipermeable plastic supports that
may be tted into the wells of multi-well culture
Fan Q, Sirkar KK, Michniak B (2008) Iontophoretic trans- plates. Test compounds are then added to either
dermal drug delivery system using a conducting poly- sides of the monolayer. After incubation, aliquots
meric membrane. J Membr Sci 321:240249
of the solution in opposite chambers are removed
Grassi M (2009) Membranes in drug delivery. In: Pabby
AK, Rizvi SSH, Sastre AM (eds) Handbook of mem- for the determination of the concentration of test
brane separations. CRC Press, Boca Raton, pp 427492 compounds. The apparent permeability coef-
Stamatialis DF, Papenburg BJ, Girones M, Saiful S, cients are then calculated. Radiolabeled com-
Bettahalli SNM, Schmitmeier S, Wessling M (2008)
pounds used in the original Caco-2 cells
Medical applications of membranes: drug delivery,
articial organs and tissue engineering. J Membr Sci monolayer assays have been replaced by the use
308:134 of liquid chromatography mass spectrometry
Drug Release 595
(LC-MS) and LC-tandem mass spectrometry Fischer SM, Flaten GE, Hagesther E, Fricker G, Brandl
(LC-MS-MS). MS eliminates the need for M (2011) In vitro permeability of poorly water soluble
drugs in the phospholipid vesicle-based permeation
radiolabeled compounds and permits the simulta- assay (PVPA): the inuence of non-ionic surfactants.
neous measurement of multiple compounds. J Pharm Pharmacol 63:10221030
The parallel articial membrane permeability Van Breemen RB, Li Y (2005) Caco-2 cell permeability
assay (PAMPA) is a possible low-cost alternative assays to measure drug absorption. Expert Opin Drug
Metab Toxicol 1:175185
to cellular models (Avdeef 2005). Its popularity in
the industry has risen rapidly. On the contrary to
cell-based permeation assays, PAMPA assay con- D
tains articial membrane that is used for evaluat-
ing the permeation properties of the compound. Drug Release
The system consists of phospholipid layer on a
lter plate, half of which is treated with an organic Catherine Charcosset
solvent, mimicking the cell membrane. Various Universit Lyon 1, Lyon, Villeurbanne, France
models with different lipids and/or organic sol-
vent (hexadecane/dodecane) are used. Due to its
articial nature, PAMPA model predicts only pas- Polymeric systems are often used to provide con-
sive permeation. When used in parallel with cell- trolled release and protect the drug from biologi-
based in vitro permeation assays (Caco-2), the cal degradation prior to its release. The
differences in the results can reveal other than development of these devices started with the
passive permeation mechanisms. PAMPA is suit- use of nonbiodegradable polymers, which rely
able for high-throughput evaluation of passive on the diffusion process, and subsequently
permeation characteristics at multi-well culture progressed to the use of biodegradable polymers,
plates, classifying compounds to high or low per- in which swelling and erosion take place. Based
meation. PAMPA assay can also be conducted at on the physical or chemical characteristics of
several physiologically relevant pH conditions polymer, drug release mechanism from a polymer
(28), to see the effect of ionization of the com- matrix can be related to three main mechanisms
pound to its passive permeation. (Arin et al. 2006):
Another recent alternative to access to drug
permeability properties is the phospholipid Drug diffusion from the non-degraded polymer
vesicle-based permeation assay (PVPA) developed (diffusion-controlled system).
for measurement of passive drug permeability Enhanced drug diffusion due to polymer swell-
(Fischer et al. 2011). The system consists of a ing (swelling-controlled system).
tight layer of liposomes on a lter support which Drug release due to polymer degradation and
is obtained by rst depositing smaller liposomes erosion (erosion-controlled system). Diffusion
that will go into the pores and then larger liposomes is involved in the three systems. For a
that will layer on top of the lter, by use of centri- nonbiodegradable polymer matrix, drug
fugation. Freeze-thaw cycling is then used to pro- release is due to the concentration gradient by
mote liposome fusion to generate a tight barrier either diffusion or matrix swelling (enhanced
used for permeability measurement. The character- diffusion). For biodegradable polymer matrix,
ization of the barrier structure showed that the release is controlled by the hydrolytic cleavage
pores of the lter are lled with liposomes and of polymer chains that lead to matrix erosion,
that there is a layer of liposomes on top of the lter. even though diffusion may be still dominant
when the erosion is slow.
References
Empirical and semiempirical mathematical
Avdeef A (2005) The rise of PAMPA. Expert Opin Drug models are often used to describe drug release
Metab Toxicol 1:325342 from matrix-based drug delivery systems
596 Dry Spinning
(Siepmann and Peppas 2012). The most popular case of the power law where n = 0.5. The power
mathematical equation to describe the release rate law describes in many cases dynamic swelling
of drugs from matrix systems is the Higuchis and drug release from glassy polymers.
equation (Higuchi 1961). Initially valid only for
planar systems, it was later modied and extended
to consider different geometries and matrix char- References
acteristics including porous structures. The equa-
tion of the Higuchi model is given as: Arin DY, Lee LY, Wang CH (2006) Mathematical model-
ing and simulation of drug release from microspheres:
implications to drug delivery systems. Adv Drug Deliv
Mt p
D2c0 cs cs t for c0 > cs (1) Rev 58:12741325
A Higuchi T (1961) Rate of release of medicaments from
ointment bases containing drugs in suspensions.
where Mt is the cumulative absolute amount of J Pharm Sci 50:874875
Siepmann J, Peppas NA (2012) Modeling of drug release
drug released at time t, A is the surface area of the
from delivery systems based on hydroxypropyl meth-
controlled release device exposed to the release ylcellulose (HPMC). Adv Drug Deliv Rev 64:163174
medium, D is the drug diffusivity in the polymer
carrier, and c0 and cs are the initial drug concen-
tration and the solubility of the drug in the poly-
mer, respectively. Equation 1 can be expressed as: Dry Spinning
Mt p
K t (2) Khayet Mohamed
M1 Department of Applied Physics, Universidad
where M1 is the absolute cumulative amount of
drug released at innite time (which should be
equal to the absolute amount of drug incorporated Dry spinning is an extrusion technique used for
within the system at time t = 0) and K is a con- fabrication of bers (laments) or hollow bers
stant reecting the design variables of the system.using a viscous dope solution containing polymer
Thus, the fraction of drug released is proportional(s) dissolved in a volatile solvent or in a solvent
to the square root of time. Alternatively, the drugmixture with a high volatility. The dope solution is
release is proportional to the reciprocal of the pumped (extruded) through a lter and then
square root of time. through a spinneret or tiny orices (i.e., jets) in a
Another commonly used semiempirical equa- heated drying chamber. As the dope solution exit
tion to describe drug release from polymeric sys- the spinneret, air or an inert gas is used to evapo-
tems is the power law: rate the solvent(s) so that the ber solidies and
collected on a take-up wheel. Spinning is
1=2 !
Mt Dt X1
nd followed by stretching of the as-spun bers to
1=2
4 2 p 2 1 ierfc p
n
orient the polymer chains along the gas gap
M1 d 2 Dt
n1
(ber axis). In dry spinning the as-spun bers do
(3) not come in contact with any precipitating liquid
(bore liquid, external coagulant). Dry spinning
where Mt and M1 are the absolute cumulative differs from dry/wet spinning and wet spinning
amount of drug released at time t and innite in that the solidication is achieved through evap-
time, respectively; k is a constant incorporating oration of the solvent carried out by a stream of air
structural and geometric characteristics of the or an inert gas. In dry spinning and melt spinning,
device; and n is the release exponent, indicative the dimensions of the spinneret are not so crucial
of the mechanism of drug release. The classical because the ber dimensions are mainly deter-
Higuchi equation (Eq. 1) represents the special mined by the ratio of the extrusion rate and
Dry Spinning 597
Dry Spinning, b Bore fluid

Fig. 1 Schematic
presentation of die swell
(gas)
occurring in the nascent a Dope solution
ber (a) and hollow ber Dope Dope
(b) when extruded from a solution solution
spinneret
D
Spinneret Spinneret
Gas gap Gas gap
Formed fiber Hollow fiber
tearing rate (Mulder 1992). The gas ow used 300 and 1000 m/min. The spinning rate in melt
for solvent evaporation and solidication of the spinning (thousands of meters per minute) is
bers inuences to a signicant extent the quality much higher than that used in the dry/wet spin
of the produced bers. In dry/wet spinning and technique (meters per minute) (Mulder 1992).
wet spinning, precipitating liquids are involved When a polymer solution is extruded through a
(bore liquid, external coagulant). In dry spinning, tube in orice spinneret, shear stress is induced
the ber does not need to be dried like in wet within the spinneret (Fig. 1) (Khayet and
spinning and dry/wet spinning. Examples of Matsuura 2011). There are at least three forces
polymer used for production of bers by dry (stresses) applied upon the dope solution:
spinning are cellulose acetate, polyacylonitrile, (i) shear and elongation stresses within the spin-
polyurethane, and bers based on poly-m- neret, (ii) gravity induced by the weight of the
phenylene isophthalamide, polybenzimidazoles, spun hollow ber, and (iii) stress induced by the
polyamidoimides, and polyimides. In general, take-up system (i.e., wind-up drum or take-up
dry spinning is used for polymers that cannot be wheel). It is known that the stress affects dramat-
used in melt spinning. The concentration of the ically the polymer molecular orientation and
spinning dope solution and the spinning speeds is relaxation at the outer surface of the nascent
higher than those used in dry/wet spinning and ber. The effects of elongation and relaxation
wet spinning. Spinning speeds range between after exiting from the spinneret will play
598 DSP&DE Model Applications
important roles on nal hollow ber structure. volume charge density (X), which is dened
Furthermore, the induced elongation stresses out- as the number of moles of xed charges
side the spinneret from gravity through the air gap per unit of pore volume; the dielectric con-
complicate the kinetics and dynamics of the phase stant of the solution inside pores (ep); and
inversion process. In addition, non-Newtonian the dielectric constant of the membrane
polymer solutions may exhibit die swell and material (em)
relaxation after exiting from spinneret. Characterization using uncharged solutes:
Molecular chains tend to align themselves rp and d are usually determined by the
much better, and this enhanced orientation causes analysis of rejection data of neutral sol-
the polymer molecules to pack closer to each other utes (Bowen et al. 1997; Schaep
leading to a tighter skin ber structure. The molec- et al. 1999; Labbez et al. 2002) since in
ular orientation induced by shear stress within the that case the rejection only depends on
spinneret might relax in the air gap region, before structural parameters. Estimation of rp
reaching the take-up wheel, if the elongation and d with various uncharged solutes
stress along the spin line is small as spinning shows that the variation in rp for best t
dope solution is a viscoelastic uid or might be is relatively small. Hence, a mean value of
enhanced if the spin line stress is high. In other rp can be considered. However, it is
words, the elongation stress caused by gravity reported that the parameter d varies sig-
along the spinning line becomes more pro- nicantly with solute size (Bowen
nounced with increasing the distance between et al. 1997; Schaep et al. 1999; Labbez
the spinneret or jets and the take-up wheel. et al. 2002). For an unknown solute, d can
then be obtained from a correlation
between d and ri/rp, ri being the solute
References radius (Bowen et al. 1997). It was also
found that the values of d obtained from
Khayet M, Matsuura T (2011) Membrane distillation: prin- the tting of the rejection data are differ-
ent from that determined from water per-
ogy. Kluwer, Dordrecht meability measurements by using the
Hagen-Poiseuille equation (Bowen
et al. 1997; Labbez et al. 2002). As the
quality of tting obtained with an inde-
DSP&DE Model Applications pendent t of both rp and d is much less
good than that obtained using d from per-
Patrick Fievet meability data, some authors recommend
Institut UTINAM, University of Franche-Comt, that the value of d is not estimated from
Besanon, France the permeability data for NF membranes.
Characterization using single-salt solutions:
The description of experimental rejec-
DSP&DE Model Applications tion rates is in general satisfactory, and the
tted values of X appear to be physically
The applications of the Donnan/steric pore (DSP) reasonable when calculations are
and dielectric exclusion (DE) model (known as performed by neglecting DE mecha-
the SEDE model in literature) can be summarized nisms, i.e., by using the DSP model
as follows: (based on a Donnan/steric exclusion
mechanism). The analysis of rejection
(i) To characterize a nanoltration (NF) mem- data shows that X is not constant but
brane in terms of the average pore radius (rp); depends very much on the salt nature
the effective membrane thickness (d); the and its concentration. The increase in the
DSP&DE Model Applications 599
volume charge density with the salt con- membranes) (Szymczyk and Fievet
centration is probably caused by interac- 2005; Szymczyk et al. 2007) with physi-
tions between free ions in solutions and cally realistic values of ep (i.e., ranging
the membrane, where each individual ion from 42 to 69). However, literature shows
contributes to the membrane charge by that the approaches used to obtain a good
means of adsorption. However, the DSP agreement between theory and experi-
model fails to describe the high rejection ment differ depending on the couple
rates observed with some NF membranes membrane/salt studied. The description
in the case of solutions containing diva- can be satisfactory considering (a) all D
lent counterions (Vezzani and Bandini exclusion mechanisms (Szymczyk and
2002; Szymczyk and Fievet 2005). Fievet 2005; Szymczyk et al. 2006,
Indeed, the use of the DSP model leads 2007), or (b) steric, electric, and Born
to the prediction of unrealistic high vol- dielectric exclusion mechanisms
ume charges or to a change of the sign of (Cavaco Moro et al. 2008; Bouranene
the membrane charge (which is not et al. 2009), or (c) steric, electric exclu-
supported by streaming potential mea- sion mechanisms and image forces
surements) (Szymczyk and Fievet 2005) (Bandini and Vezzani 2003; Bouranene
or to a ratio of the volume charge density et al. 2009).
to the salt concentration increasing with Characterization using multi-ionic
concentration (which disagrees with com- solutions:
mon adsorption isotherms) (Bowen and In the case of 3-ion solutions, different
Welfoot 2002) or to the increase of the approaches of the DSP&DE model have
rejection rate as the membrane charge also to be used to appropriately describe
density X decreases (which is the oppo- the separation properties (Bouranene
site behavior from the one predicted by et al. 2009; Szymczyk et al. 2006). Unlike
the Donnan/steric exclusion mechanism). the results obtained with 2-ion and 3-ion
This suggests that additional phenomena solutions, the four exclusion mechanisms
have to be accounted for so as to provide a (Donnan, steric, Born, and image effects)
better understanding, characterization, are found to be unable to describe the
and prediction of NF membrane perfor- experimental data in the case of solutions
mances. However, it should be mentioned containing more ions. However, the
that the DSP model was proved to provide description of the experimental rejection
a coherent description of NF behavior for rates is quite satisfactory considering
membranes at the borderline between UF either Born dielectric effect or image
and NF (membranes with larger pores and forces (in addition to steric and electric
very thick active layers) as has been exclusion mechanisms).
shown, e.g., in Lefebvre (2003) and The application of the DSP&DE
Lefebvre et al. (2004) for ion rejection model was also extended to a complex
and ltration potential (i.e., the pressure- solution of ve ions (K+, Cl, NH4+,
induced potential arising across mem- SO42, and clavulanate ion), which
branes containing selective layers). mimics the industrial NF feed in the iso-
The DSP&DE model was shown to lation process of clavulanate, which is a
provide a rather good description of the pharmaceutical product (Cavaco Moro
rejection properties of both symmetric et al. 2008). With a single-tted parameter
(KCl) and asymmetric (MgCl2, CuCl2, ep and considering that the ion selectivity
ZnCl2, NiCl2, CaCl2) single-salt solutions is governed by Donnan, steric, and Born
by negatively charged organic mem- dielectric exclusion mechanisms, a good
branes (AFC 30 and AFC 40 PCI NF agreement was obtained between the
600 DSP&DE Model Applications
calculated rejections rates of the ve ions investigate the signicance of dehydration

and experimental values. and pore swelling phenomena on the NF of
(ii) To investigate the exclusion mechanisms in neutral solutes, which could be responsible
NF for the observed decrease in the neutral sol-
The coupling between the various mech- ute rejection in the presence of salts (Escoda
anisms involved in the ion rejection could be et al. 2013). To this end, rejection properties
investigated. It was shown, e.g., that dielec- of two NF ceramic and polymeric mem-
tric effects (i.e., Born dielectric effect and branes were studied with single polyethylene
image forces) make the Donnan exclusion glycol (PEG) solution and mixed
stronger. Indeed, unlike Donnan exclusion, PEG/inorganic electrolyte solutions. Rejec-
image forces and Born dielectric effect do tion data obtained with the ceramic mem-
not differentiate (qualitatively) between brane (rigid pores) were used to compute
coions and counterions since they are con- the dehydration of the PEG molecules
trolled by the square of the ion charge. As a induced by the surrounding ions.
result, both coions and counterions are Concerning the polymeric membrane, the
excluded from pores by image forces. By decrease in the rejection rate was found to
decreasing the internal concentration, the be systematically higher than for the ceramic
dielectric exclusion phenomenon therefore membrane. The additional decrease was then
reduces the screening of the membrane ascribed to the swelling of the pores. The
xed charge and thus strengthens the electric experimental data of rejection rates were
exclusion. By contrast, the presence of a then used to compute the variation in the
xed charge on the pore walls as well as the mean pore radius in the presence of the var-
ionic atmosphere surrounding an ion screens ious salts. For the two membranes, it was
the interaction of this ion with the pore wall found that the contribution of the dehydra-
via image forces. The coupling between tion effect follows the Hofmeister series
electric and dielectric exclusion mechanisms (Kunz 2004).
may result in a non-monotonous variation of (iii) To predict the rejection properties of a NF
the rejection with membrane charge density membrane
(X). It was also shown that the Born effect It must be stressed that accounting for the
also strengthens the contribution of image dielectric exclusion mechanism inevitably
forces. increases the number of tting parameters,
The analysis of the rejection properties of making the use of the DSP&DE model
a ternary ion mixture by an amphoteric less appropriate for predictive purposes. Nota-
ceramic membrane (Szymczyk et al. 2006) bly, the dielectric constant of the solution con-
has shown that the dielectric constant inside ned inside the membrane pores (ep) is used as
the pores ep (which was tted to experimen- a tting parameter in all current versions of the
tal rejection rates) decreases with increasing DSP&DE model (Szymczyk and Fievet 2005,
normalized volume charge density (dened 2006, 2007; Escoda et al. 2011; Cavaco-
as the ratio of the membrane volume charge Morao et al. 2008). Indeed, although in prin-
density to the total charge concentration in ciple ep could be determined from dielectric
solution), which means that the presence of relaxation measurements, in practice, the mea-
xed charges at the pore walls (in addition to surement of this parameter is extremely dif-
the connement effect) could also have an cult due to (a) the multilayer (thick support
inuence on the orientation of solvent layers) and composite (membrane material
dipoles. and conned solution) structure of NF mem-
The DSP&DE model (more exactly the branes and (b) the pore size distribution. It
DSP model) was also applied to mixtures of must also be pointed out that the multilayer/
uncharged solutes and ions in order to composite structure and nanodimensional
Dual Mode Sorption Model 601
pores make also the determination of the membrane using the DSPM model: effect of pH, salt
membrane charge density extremely difcult. concentration and nature. J Membr Sci 208:315329
Lefebvre X (2003) Etude des modles de transfert en
Finally, unlike the DSP model, the prediction Nanoltration. Application du modle hybride bas
of membrane performances for multi-ionic sur les quations de Nernst-Planck tendues par le
systems by means of the adjustable parameters dveloppement du logiciel de simulation Nanoux.
obtained for single-salt solutions is not possi- Thesis, University of Montpellier II, France (in French)
Lefebvre X, Palmeri J, David P (2004) Nanoltration the-
ble with the DSP&DE model. Indeed, the ory: an analytic approach for single salts. J Phys Chem
choice for X and ep (em being assumed to be B 108:1681116824
known) allowing the description of experi- Schaep J, Vandecasteele C, Mohammad AW, Bowen WR D
mental rejection rates is not unique for (1999) Analysis of the salt retention of nanoltration
membranes using the Donnan-Steric partitioning pore
single-salt solutions because both electric model. Sep Sci Technol 34(15):30093030
exclusion and dielectric exclusion contribute Szymczyk A, Fievet P (2005) Investigating transport prop-
to reject ions. erties of nanoltration membranes by means of a steric,
electric and dielectric exclusion model. J Membr Sci
252:7788
Szymczyk A, Sbai M, Fievet P, Vidonne A (2006) Trans-
Cross-References port properties and electrokinetic characterization of an
amphoteric nanolter. Langmuir 22:39103919
Nanoltration Szymczyk A, Fatin-Rouge N, Fievet P, Ramseyer C,
Vidonne A (2007) Identication of dielectric effects
in nanoltration of metallic salts. J Membr Sci
287:102110
References Vezzani D, Bandini S (2002) Donnan equilibrium and
dielectric exclusion for characterization of
Bandini S, Vezzani D (2003) Nanoltration modeling: the nanoltration membranes. Desalination 149:477483
role of dielectric exclusion in membrane characteriza-
tion. Chem Eng Sci 58:33033326
Bouranene S, Fievet P, Szymczyk A (2009) Investigating
nanoltration of multi-ionic solutions using the steric,
electric and dielectric exclusion model. Chem Eng Sci
64:37893798 Dual Mode Sorption Model
Bowen WR, Welfoot JS (2002) Modelling the
performance of membrane nanoltration critical Donald R. Paul
assessment and model development. Chem Eng Sci
57:11211137
Department of Chemical Engineering, Texas
Bowen WR, Mohammad AW, Hilal N (1997) Characteri- Materials Institute, Austin, TX, USA
zation of nanoltration membranes for predictive
purposes use of salts, uncharged solutes and atomic
force microscopy. J Membr Sci 126:91105
Cavaco Moro AI, Szymczyk A, Fievet P, Brites Alves AM
The dual mode sorption model is a convenient
(2008) Modelling the separation by nanoltration of a and commonly used equation for describing gas
multi-ionic solution relevant to an industrial process. sorption in polymers at temperatures below the
J Membr Sci 322:320330 glass transition temperature, Tg, of the polymer.
Escoda A, Don S, Fievet P (2011) Assessment of dielec-
tric contribution in the modeling of multi-ionic trans- Gas sorption in glassy polymers generally
port through nanoltration membranes. J Membr Sci (a) exhibits nonlinear isotherms that are concave
378:214223 to the pressure axis, unlike the linear ones seen
Escoda A, Bouranene S, Fievet P, Don S, Szymczyk above Tg, (b) shows greater levels than for rub-
A (2013) Dehydration and pore swelling effects on
the transfer of PEG through NF membranes. Membr
bery polymers with the extent being larger the
Water Treat 4(2):127142 higher is the Tg, and (c) depends on the prior
Kunz W, Henle J, Ninham BW (2004) Zur Lehre von der history of the sample and the details of the mea-
Wirkung der Salze (about the science of the effect of surement protocol (Horn and Paul 2012). The rst
salts): Franz Hofmeisters historical papers. Curr Opin
Colloid Interface Sci 9(12):1937
two aspects are well illustrated in Fig. 1 that
Labbez C, Fievet P, Szymczyk A, Vidonne A, Foissy A, shows the concentration of carbon dioxide, C,
Pagetti J (2002) Analysis of the salt retention of a titania dissolved in ve different polymers as a function
602 Dual Mode Sorption Model
70 temperature T, i.e., C0H

Tg T , and is
CO2 at 35C, Bulk Polymers thought to be related to the unrelaxed volume
60 of the glass; see Paul (2010) for a recent
summary of the extensive literature in this area.
50
Matrimid The CH0 term would be zero for butyl rubber, i.e.,
C [cm3(STP)/cm3]
PPO Eq. 1 reduces to Henrys law, since it is not in the

40
glassy state.
PSF
Petropoulos (1970) proposed that the Lang-
30
muir population may have a nite, but different,
Polystyrene mobility than Henrys law population. Paul and
20
Koros (1976) extended this analysis to give
10 expressions for the time lag and the following
Butyl Rubber expression for the permeability coefcient, P, for
0 gases in glassy polymers:
0 5 10 15 20 25
CO2 Pressure (atm)
C0H bDH
P kD DD (2)
Dual Mode Sorption Model, Fig. 1 Sorption isotherms 1 bp2
for CO2 in various polymers at 35 C (Reprinted with
permission from Horn and Paul 2012. Copyright 2012
American Chemical Society) where DD is the diffusion coefcient associated
with Henrys law population and DH is the
diffusion coefcient for the Langmuir popula-
of the pressure, p, in the external carbon dioxide tion and p2 is the upstream pressure of the
gas phase. Butyl rubber shows a linear isotherm, permeating gas. This has become known as
while all four of the glassy polymers show fea- the dual mobility model. It predicts that the
tures (a) and (b) mentioned above at this measure- permeability decreases slightly as the upstream
ment temperature, i.e., T = 35 C. The extent of pressure increases and this is what is commonly
carbon dioxide sorbed into these glassy materials observed experimentally in simple cases where
increases the higher is the Tg; i.e., their glass the polymer is not plasticized signicantly by
transition temperatures descend in the following the permeating gas.
order: 310 C for the polyimide Matrimid 210 C
for poly(phenylene oxide) (PPO) 186 C for
polysulfone (PSF) 100 C for polystyrene. References
The dual mode sorption model consists of the
sum of Henrys law term and a Langmuir term, Barrer RM, Barrie JA, Slater J (1958) Sorption and diffu-
i.e., sion in ethyl cellulose. III. Comparison between ethyl
cellulose and rubber. J Polym Sci 27:177197
Horn NR, Paul DR (2012) Carbon dioxide sorption
C0H bp and plasticization of thin glassy polymer lms
C kD p (1)
1 bp tracked by optical methods. Macromolecules 45:
28202834
where kD is Henrys law coefcient, CH0 is the Paul DR (2010) Fundamentals of transport phenomena in
polymer membranes. In: Drioli E, Giorno L (eds) Com-
Langmuir capacity term, and b is an afnity prehensive membrane science and engineering,
parameter. It appears that Barrer et al. (1958) vol 1. Academic, Oxford, pp 7590
were the rst to propose this now well-known Paul DR, Koros WJ (1976) Effect of partially immobilizing
equation. Equation 1 provides a very good t sorption on permeability and the diffusion time lag.
J Polym Sci Polym Phys Ed 14:675685
of data like that in Fig. 1 where the CH0 obtained Petropoulos JH (1970) Quantitative analysis of gaseous
by the tting would increase as the polymer diffusion in glassy polymers. J Polym Sci Part
Tg increases relative to the measurement A-2(8):17971801
DuraMem Membrane 603
fouling in the rst place. These techniques can

Durability of Membrane (Lifetime of extend a membrane modules lifetime but will
Membrane) not address issues such as chemical degradation,
polymeric aging, and mechanical abrasion.
Colin A. Scholes Hence, all membrane modules separation perfor-
Department of Chemical and Biomolecular mances decrease over time in all industrial pro-
Engineering, University of Melbourne, Parkville, cesses and replacement is necessary.
VIC, Australia Currently, the industry standard is for a mem-
brane module lifetime of
5 years under standard D
process conditions (Li et al. 2008). For the major-
Durability (or lifetime) of membranes is a critical ity of reverse osmosis membrane plants currently
factor in the commercialization potential of any being commissioned or under construction, the
membrane material, as it must be resilient to the RO membranes are guaranteed by the manufac-
process conditions over an extended period of turer for 5 years. Similarly, vapor separation mem-
time. The lifetime of a membrane is also an impor- branes in the petrochemical industry have a
tant factor in the economic evaluation of a mem- lifetime of 35 years, dependent on the conditions
brane process, as replacement of membranes is a of the feed. Seawater ultraltration membranes
signicant cost burden. last up to 7 years because of their resilient material
A range of factors inuence the durability of and the ability to avoid regular chemical cleans.
membrane materials. These include chemical deg- Alternatively, dairy and brine processing mem-
radation of either the active layer or support struc- branes need to be replaced every year because of
ture due to harsh chemical components in the feed biological and particulate buildup, while hemodi-
stream, for example, feeds with low or high pH alysis and hemoltration membranes are only
conditions, as well as high electrolyte concentra- required to function for a few hours, as they are
tions, polymeric aging, and polymer plasticiza- single-use only equipment.
tion, leading to increased probability of the
active layer rupturing. An example of this is cel-
lulose acetate membranes being plasticized by
References
exposure to hydrocarbons in petrochemical gas
separation applications; active layer compaction Li NN, Fane AG, Ho WSW, Matsuura T (2008) Advanced
due to the extended exposure to high-pressure membrane technology and applications. Wiley,
differentials across the membrane, resulting in Hoboken
reduced ux because of loss in diffusivity;
and membrane fouling, which is the buildup of
foreign matter on a membranes surface resulting
in decreased performance. An example is micro- DuraMem Membrane
bial growth in water treatment. Finally, mechani-
cal abrasion of the active and support layers Ludmila Georgieva Peeva
is caused by particulate matter being present in Imperial College London, London, UK
system. Each of these factors inuences mem-
brane materials differently, depending on the
application. DuraMemTM is a range of highly stable organic
There are a number of approaches to extend solvent nanoltration (OSN) membranes
membrane lifetimes, most of which focused on manufactured by Evonik Membrane Extraction
overcoming membrane fouling. These include Technology Ltd. (www.duramem.evonik.com).
feed pretreatment, chemical and mechanical Membranes are of integral asymmetric type and
cleaning, feed stream pulsation, and pressure var- are based on cross-linked polyimide (PI) (Great
iation, as well as chemical additives to prevent Britain 2007; Toh et al. 2007). These membranes
604 Dusty-Gas Model (DGM)
DuraMem Membrane, Fig. 1 A typical SEM micrograph of a cross section (left) and top separating layer (right) of
cross-linked PI membranes (Adapted from Toh et al. (2007))
are available in formats of at sheets and

spiral-wound membrane modules within different Dusty-Gas Model (DGM)
molecular weight cutoff range (150900 g mol1)
and possess excellent stability in most of the Alessio Caravella
organic solvents including polar aprotic solvents National Institute of Advanced Industrial Science
such as dimethylformamide (DMF) and and Technology Research Institute for
N-methylpyrrolidone (NMP). The membranes Innovation in Sustainable Chemistry, ISC-AIST,
have a spongelike structure (Fig. 1) and are usable Tsukuba, Ibaraki, Japan
in acetone, ethanol, methanol, tetrahydrofuran,
dimethylformamide, dimethyl sulfoxide,
dimethylacetamide, isopropanol, acetonitrile, The Dusty-Gas Model is a transport model
methyl ethyl ketone, ethyl acetate, and others describing the motion of uid mixtures through
(www.duramem.evonik.com). The membranes porous media. Its particular name is due to the
have been operated continuously for 120 h in main hypothesis behind its development: in a
DMF and THF and showed stable uxes multicomponent system of n gas species moving
and good separation performances, with DMF within a porous medium, the porous medium itself
permeability in the range of 18 is viewed as an additional (n + 1)-th component of
105Lm2h1Pa1(18 Lm2h1bar1) (Toh the mixture, consisting of giant molecules (or dust
et al. 2007). Possible re-imidization and loss of particles) constrained to remain xed in the space
cross-linking at elevated temperatures limit their by the presence of a virtual external force. How-
range of application to temperatures <50 C. The ever, if this virtual external force is removed, the
membranes can be operated in a pressure range of model can be also used to describe motion of
2060 bar (290870 psi). aerosols.
This model has been independently reinvented
several times in the course of the history, starting
with James Clerk Maxwell in 1860 (Maxwell
References 1980). However, it was only in 1983 that it was
formalized as a mature theory by Mason and
Great Britain 2437519 (2007)
Toh YH, Lim FW, Livingston AG (2007) Polymeric mem-
Malinauskas (1983), who described in detail the
branes for nanoltration in polar aprotic solvents. number of different hypotheses leading to
J Membr Sci 301:310 corresponding different expressions of the model.
Dusty-Gas Model (DGM) 605
The Dusty-Gas Model is particularly important NiDiff NiKn

in Membrane Technology, as it can be used to
describe the transport of a multicomponent uid
NiTot NiVisc NiTot
mixture through porous membranes and/or porous
supports. Its use is not restricted to just ideal
NiSurf
gases, but it can apply to real uids as well,
requiring the use of activities instead of partial
pressures. Dusty-Gas Model (DGM), Fig. 1 Analog electric circuit
The model takes into account the following relative to the transport modes included in the Dusty-Gas D
Model
concurrent and independent transport modes:
Knudsen ow (i.e., free-molecule ow), where Neglecting surface diffusion and thermal dif-
the molecule-wall collisions are dominant over fusion, the mathematical form of the model can be
the molecule-molecule ones, so that the latter expressed in the following manner:
can be neglected.
! !
Continuum diffusion, where the species in Xn
xj N i xi N j
mixture move with respect to each other by Bin e
j1 CT Dij
the effect of chemical potential gradient, t
2 e !3
which can include diffusion driven by concen- !
1 4Ni x i B 0 Xn
!
tration gradient, temperature gradient, and/or t P C xj F j 5
Kn e CT
T
by external forces. In this case, the molecule- Di j1
t
molecule collisions are dominant over the !
molecule-wall collisions. xi lnxi xi lnP xiRTF i
Viscous ow, where the mixture acts as a single (1)
uid whose ow is promoted by the gradient of
!
total pressure. In this case, the molecule- where N i [mol s1 m2] is the total uxes (i.e.,
molecule collisions again dominate over the diffusive + convective) of the i-th species, xi its
molecule-wall collisions. molar fraction, P [Pa] the total pressure, CT [mol
Surface diffusion is where the transport of m3] the total concentration, T [K] the tempera-
the molecules over the surface in adsorbed 2 1
ture, DBin
ij [m s ] is the binary diffusion coef-
state is dominant. This type of transport is cient relative to the generic species i and j, DKn i
usually negligible compared to the others, [m2 s1] is the Knudsen diffusivity, B0 [m2] is a
except for sufciently small pore size, for geometrical parameter depending on the structure
!
which the adsorption forces are strong geometry, [Pa s] is the mixture viscosity, F i is the
enough to constrain the molecule to mainly resultant of all the possible external forces acting
move over the surface. This kind of trans- on the i-th species, and e and t are the medium
port is usually precursors of phenomena like porosity and tortuosity, respectively (all the vecto-
multilayer adsorption and capillary rial entities are denoted with an arrow above).
condensation.
The most important hypothesis that allows References

leading a usable mathematical form of the Mason EA, Malinauskas AP (1983) Gas transport in
model is that all the involved transport modes porous media: the dusty-gas model, Chemical engi-
act independently to each other. Under this neering monographs 17. Elsevier, Amsterdam
hypothesis, the scheme behind the formaliza- Maxwell JC (1962) Illustration of the dynamical theory of
gases. Part II. On the process of diffusion of two or
tion of the Dusty-Gas Model can be
more kinds of moving particles among one another.
represented by the following analog electric Philos Mag 20(1860):21 (Reprinted in Scientic
circuit (Fig. 1): Papers, vol 1, pp 392409, New York)
606 Dynamic Free Volume
with temperature according to the equation:

Dynamic Free Volume vf = vfg[1 + (ar ag)(T Tg)], where ar and ag
are the coefcients of thermal expansion above
Yuri Yampolskii and below Tg, respectively, and vfg is the free
A.V. Topchiev Institute of Petrochemical volume at Tg. Originally, it has been assumed
Synthesis, Russian Academy of Sciences, that the fraction of free volume at Tg, fc = vf/v,
Moscow, Russia where v is the specic volume of the polymer, is
constant for most of the polymers and equal
0.025 0.003 (iso-free volume concept of
Dynamic free volume appears in liquids (Frenkel Simha-Boyer). Further studies showed that fc var-
1946) and in polymers above their glass transition ied in a much wider range and increases with Tg
temperatures Tg, i.e., in rubbers. In liquids, it (Yampolskii et al. 2000).
appears due to mobility of individual molecules, The variation of the dynamic free volume in
while in rubbers due to segmental mobility. rubbers with temperature and as a function of Tg
Dynamic free volume inuences such properties has important implications for membrane science:
of the medium as diffusion and self-diffusion when the difference Texp-Tg increases (at constant
coefcients and viscosity. In glassy state, where Texp), i.e., for the rubbers with lower Tg, e.g.,
the segmental mobility is frozen, still some local polydimethylsiloxane, the transport parameters
mobility of small side groups is possible as well as such as P(Texp) and D(Texp) strongly increase
the corresponding free volume. When a glassy (van Amerongen 1964; Yampolskii et al. 1982).
polymer is plasticized, that is, contains low Here, Texp is the ambient temperature at which the
molecular mass additives, their mobility can permeability and diffusion coefcients are usually
induce some mobility of the adjacent parts of the measured. It means that the polymers with the
main or side chains. So the fraction of dynamic lowest glass transition temperature have the
free volume in the total free volume can increase, greatest dynamic free volume at ambient
the polymer becomes more rubberlike, and its temperatures.
properties (gas permeation parameters) become The size of the elements of dynamic free vol-
more similar to those of rubbers: increase in per- ume in rubbers is relatively large: it should be
meability and decrease in permselectivity. These compared with the size of kinetic segments, so
phenomena are detrimental for separation of gas much larger than the size of the typical gaseous
mixtures that contain great concentration of penetrants. It means that the selectivity of diffu-
highly sorbed components like carbon dioxide sion in rubbers is very weak, and, because of this,
(Sanders 1988). According to molecular dynam- in rubbers solubility-controlled permeation
ics, the average time of settled life of free vol- regime is realized.
ume elements in rubbers is very short about
several picoseconds (Hofmann et al. 2000). The
mobility of any kinetic system is determined by
the relaxation time t: t(T) = to exp(U/kT), where References
U is the activation energy and to = 1012 s. It is
often assumed that U increases at decreased tem- Frenkel YI (1946) Theory of liquids. Clarendon, Oxford
Hofmann D, Fritz L, Ulbricht J, Paul D (2000) Comput
peratures and, hence, the relaxation times at the
Theory Polym Sci 10:419436
glass transition temperature t(Tg) become longer Sanders ES (1988) J Membr Sci 37:6380
than the intrinsic times of cooling of the high van Amerongen GJ (1964) Rubber Chem Technol
elastic system, i.e., vitrication occurs. These 37:10651152
Yampolskii Y, Durgaryan S, Nametkin N (1982)
changes of U are caused by the decreases in the Vysokomol Soed A 24:536541
dynamic free volume in rubbers vf. The dynamic Yampolskii Y, Kamiya Y, Alentiev A (2000) J Appl Polym
(or uctuation) free volume above Tg increases Sci 76:16911705
Dynamic Mechanical Analysis 607
applications. For membrane samples, tension

Dynamic Mechanical Analysis mode (fth mode from the top in Fig. 2b) is
generally adopted.
Chi Hoon Park In the membrane applications, DMA is widely
Gyeongnam National University of Science and used for various elds such as fuel cell mem-
Technology (GNTECH), Jinju-si, branes, bio membranes, gas separation,
Gyeongsangnam-do, Republic Of Korea pervaporation, etc. Page et al. applied DMA to
(South Korea) investigate the structural characteristic of Naon
membranes well known as a fuel cell membrane D
(Page et al. 2005). In the study, two thermally
reversible transitions of Naon over room temper-
Dynamic mechanical analysis (DMA) is a tech- ature were explained as glass transition by the
nique to study viscoelastic properties and modu- short-range segmental motions within a static
lus of elasticity of polymers by measuring the electrostatic network below 150 C and the long-
damping of an oscillatory signal of stress and range chain/side-chain motion by thermally acti-
strain. DMA results generally describe the varia- vated destabilization of the electrostatic network
tion of modulus values as a function of tempera- over 240 C. Also, Xie reported that the peruor-
ture or frequency (e.g., Fig. 1) and are widely used osulfonic acid (PFSA) ionomer membrane
to detect the transitions such as glass transition showed a wide range of glass transition from
temperature (Tg) or melting of polymer materials.
55 C to
130 C (Fig. 1), and based on the
Figure 2a shows general schematic diagram of Tg result, he studied the multi-shape memory
DMA apparatus; after a sinusoidal force is applied effect of PFSA (Xie 2010). Mano focused on the
to the sample as an input signal, the resulting advantage of DMA that the test conditions can be
sinusoidal deection or deformation is detected more closely set to the physiological environ-
as an output signal, which contains information ments which the biomaterials will be applied to
about the storage modulus E0 , the dynamic loss (Mano 2008). Using chitosan membranes, their
modulus E00 , and the mechanical loss factor tan d mechanical properties were measured under
(www.ngb.netzsch.com). According to the feature low/moderate hydration conditions at different
or types of sample materials, there are various relative humidity (RH) levels and in completely
modes of deformation as shown in Fig. 2b wet (immersed) condition. Weng
(www.ngb.netzsch.com), and this variety of sam- et al. synthesized intrinsically dopable polyimide
ple holding method enables DMA to have wide (DPI) membranes containing an amine-capped
Dynamic Mechanical
Analysis, Fig. 1 Dynamic
mechanical analysis (DMA)
curve of peruorosulfonic
acid (PFSA) ionomer. tan d
is the ratio between the loss
modulus E00 and the storage
modulus E0 (Xie 2010)
(Reprinted with permission
from Macmillan Publishers
Ltd: Nature 464:267270,
copyright (2010))
608 Dynamic Mechanical Analysis
aniline trimer (ACAT) having high gas separation sol-gel method for a cation-exchange membrane,
properties, where DMA results reported higher and its mechanical stability was measured by
mechanical and thermal stabilities of the mem- DMA (Nagarale et al. 2004); Gholap
branes than those of polyaniline (PANI) and con- et al. synthesized hydrophobically modied
ventional nondopable polyimide (NDPI) (Weng PVA-g-polyNTBA membranes, where DMA was
et al. 2011). Other examples are as follows: used to detect the crystallinity of the membranes
Budd et al. studied the mechanical stability of as a function of annealing (Gholap et al. 2004);
polymer of intrinsic microporosity (PIM-1) by using DMA, Arjun and Ramesh studied the
DMA in the temperature range from 30 C to mechanical stability of the oxidatively stable
450 C for pervaporation application (Budd polycarbonate urethane (PCU) membranes fabri-
et al. 2004); Nagarale et al. synthesized poly cated by electrospinning method for biomedical
(vinyl alcohol)-SiO2 hybrid membrane by application (Arjun and Ramesh 2012); Ghassemi
Dynamic Mechanical Analysis, Fig. 2 (continued)

Dynamic Mechanical Analysis 609
Dynamic Mechanical Analysis, Fig. 2 (a) General sample holding method) (www.ngb.netzsch.com)
schematic diagram of dynamic mechanical analysis (Reprinted with permission from NETZSCH-Gertebau
(DMA) apparatus and (b) modes of deformation (i.e., GmbH)
et al. synthesized poly(arylene ether)s with pen- References

dant peruoroalkyl sulfonic acid groups for fuel
cell application, where their a-relaxation temper- Arjun GN, Ramesh P (2012) Structural characterization,
mechanical properties, and in vitro cytocompatibility
ature was reported around 196 C (Ghassemi
evaluation of brous polycarbonate urethane mem-
et al. 2011); Love investigated thermal and branes for biomedical applications. J Biomed Mater
mechanical transitions for commercially available Res A Article rst published online. doi:10.1002/jbm.
polymer Li-ion battery separators (Love 2011). a.34255
610 Dynamic Membrane Emulsification
Budd PM, Elabas ES, Ghanem BS, Makhseed S, pores into the continuous phase, and the droplet
McKeown NB, Msayib KJ, Tattershall CE, Wang detachment occurs by a surface shear. The regular
D (2004) Solution-processed, organophilic membrane
derived from a polymer of intrinsic microporosity. Adv droplet detachment from the pore openings is
Mater 16(5):456459 obtained by:
Dynamic Mechanical Analyzer DMA 242 C. Online acces-
sible poster. NETZSCH-Gertebau GmbH. http:// (i) Moving the continuous phase in cross-ow,
www.netzsch-thermal-analysis.com/download/Poster_
DMA_242C_65.pdf. Accessed 25 Aug 2012 stirred, or pulsed mode
Ghassemi H, Schiraldi DA, Zawodzinski TA, Hamrock (ii) Moving the membranes by rotation, vibra-
S (2011) Poly(arylene ether)s with pendant peruoroalkyl tion, or oscillation
sulfonic acid groups as proton-exchange membrane mate-
rials. Macromol Chem Phys 212(7):673678
Gholap SG, Jog JP, Badiger MV (2004) Synthesis and char- In Table 1, techniques, shear stress prole, and
acterization of hydrophobically modied poly(vinyl alco- forces acting on a forming droplet for each mem-
hol) hydrogel membrane. Polymer 45(17):58635873 brane emulsication process are illustrated.
Love CT (2011) Thermomechanical analysis and durabil- In cross-ow mode, the shear stress is gener-
ity of commercial micro-porous polymer Li-ion battery
separators. J Power Sources 196(5):29052912 ated by the recirculation of the continuous phase
Mano JF (2008) Viscoelastic properties of chitosan with tangentially to the membrane surface. This tech-
different hydration degrees as studied by dynamic nique is called cross-ow ME. A typical cross-
mechanical analysis. Macromol Biosci 8(1):6976 ow ME apparatus includes a tubular or at-sheet
Nagarale RK, Gohil GS, Shahi VK, Rangarajan R (2004)
Organicinorganic hybrid membrane: thermally stable microltration membrane, a pump for the
cation-exchange membrane prepared by the Solgel recirculation of the continuous phase along the
method. Macromolecules 37(26):1002310030 lumen side of the membrane, a feed vessel, and a
Page KA, Cable KM, Moore RB (2005) Molecular origins pressured nitrogen gas container for the dispersed
of the thermal transitions and dynamic mechanical
relaxations in peruorosulfonate ionomers. Macromol- phase (Peng and Fellow 1998). Cross-ow ME is
ecules 38(15):64726484 suitable for large-scale production and continuous
Weng C-J, Jhuo Y-S, Huang K-Y, Feng C-F, Yeh J-M, or semicontinuous operation. However, droplet
Wei Y, Tsai M-H (2011) Mechanically and thermally breakup can occur over time when concentrated
enhanced intrinsically dopable polyimide membrane
with advanced gas separation capabilities. Macromol- emulsion is produced due to the high cross-ow
ecules 44(15):60676076 velocities used to obtain narrow droplet size dis-
Xie T (2010) Tunable polymer multi-shape memory effect. tribution at high dispersed phase ux or over long
Nature 464:267270 time of operation.
In pulsed-ow mode, a periodic ow pulsation
is generated in the continuous phase without
Dynamic Membrane Emulsification recirculation. This technique is called pulsed-
ow ME. In a typical pulsed-ow ME apparatus,
Emma Piacentini1, Alessandra Imbrogno1,2 and the dispersed and the continuous phases are
Richard G. Holdich3 injected using a pump, while the pulsed ow in
1
Institute on Membrane Technology, National the continuous phase is generated by a frequency
Research Council of Italy, ITM-CNR, Rende, generator (Holdich et al. 2013) or by inverting the
Italy ow direction back and forward within the mem-
2
University of Calabria, Rende, Italy brane lumen (Piacentini et al. 2014). Pulsed-ow
3
Department of Chemical Engineering, ME is advantageous for the production of emul-
Loughborough University, Loughborough, sions with a high dispersed to continuous phase
Leicestershire, UK ratio, and it is suitable for large-scale production
in continuous or semicontinuous mode.
In an alternative, the shear stress on the mem-
Dynamic membrane emulsication (ME) refers to brane surface can be generated by stirring the con-
the formation of emulsion droplets by extruding tinuous phase. This technique is called stirred
the to-be-dispersed phase through the membrane membrane emulsication. In a typical stirred ME
Dynamic Membrane Emulsification, Table 1 Membrane emulsification techniques, shear stress profile, and forces acting on a forming droplet
Membrane emulsication based on moving the continuous phase
ME plant Forces acting on a drop at the pore level Shear stress prole
Dynamic Membrane Emulsification
(continued)
611
D
Dynamic Membrane Emulsification, Table 1 (continued)
612
Membrane emulsication based on moving the membrane

ME plant Forces acting on a drop at the pore level Shear stress prole
(continued)
Dynamic Membrane Emulsification, Table 1 (continued)
613
D
614 Dynamic Membrane Emulsification
device, the dispersed phase is injected through the production rates, facilitating easy process automa-
membrane pores by using a pump or nitrogen gas tion and reduced operation time.
while stirring the continuous phase with a rotator In dynamic ME, the forces acting on a droplet at
when a tube-shaped submerged membrane is the membrane pore level include detaching forces,
used or with a paddle blade stirrer in the case of driving droplets off the pore, and retaining forces,
a at-sheet membrane (Stillwell et al. 2007). holding droplets on the pore. Detaching forces are
Stirred ME is useful for emulsion production at drag (FD), buoyancy (FBG), inertial (FI), lift (FL),
batch scale, in the laboratory, and it is especially and static pressure (FSP) force, while the main
used to study the effect of different experimental retaining force is the interfacial tension (Fg). The
conditions and chemistry on the formulation drag force (FD) is created either by moving the
properties. continuous phase parallel to the membrane surface
In rotating ME, the dispersed phase is intro- or by moving the membrane; the buoyancy force
duced into the center of a rotating tubular porous (FBG) is due to the density difference between the
membrane, and droplets are detached at the mem- continuous phase and the dispersed phase; the iner-
brane wall in the radial direction into the station- tial force (FI) is caused by the dispersed phase ow
ary continuous phase (Vladisavljevi and moving out from the pore outlet; the dynamic lift
Williams 2006). The pressure is generated in the force (FL) results from the asymmetric velocity
membrane lumen by using nitrogen gas or a prole of the continuous phase near the droplet;
pump, while the membrane is rotated in the con- static pressure (FSP) force is due to the pressure
tinuous phase by an overhead stirrer. Rotating ME difference between the dispersed phase and the
is particularly advantageous for the production of continuous phase at the membrane surface; the
coarse emulsions and fragile structured products interfacial tension force (Fg) represents the effect
in which the droplets and/or particles are subject of dispersed phase adhesion around the edge of the
to breakage during the pump recirculation. How- pore opening. In the case of moving membrane
ever, the use for large-scale production is limited emulsication, additional (or enhanced) forces
due to the complicated and more expensive design have to be considered.
and higher power consumption compared to alter- For rotating membrane emulsication, com-
native designs. pared to a nonrotating system, the drag force created
The vibrating ME principle is similar to the by the membrane rotation (FDR) has to be taken into
rotating ME except for the use of oscillation of account, and the gravitational body force is replaced
the membrane perpendicularly with respect to the by the centrifugal force (FC) resulting from centrif-
dispersed phase injection direction (Holdich ugal acceleration and density difference of the two
et al. 2010). The frequency and amplitude of the phases. High centrifugal force may be responsible
axial oscillation of the membrane was generated for polydisperse emulsion production in the case of
by an electrically driven vertical oscillator and O/W emulsion due to deformation of the emerging
could be controlled separately. oil drops and induced ow of oil drops toward the
In azimuthally oscillating ME, a tubular mem- membrane surface causing coalescence and wetting
brane is periodically rotated back and forward in a of the membrane.
slowly cross-owing continuous phase, while the For an oscillating membrane system where the
dispersed phase is radially injected in the lumen membrane is axially excited, an additional drag
side of the membrane using a pump (Silva force (FDE) and inertial force (FIE) appear in the
et al. 2015). The azimuthal (tangential) oscillation direction parallel to the membrane. This is due to a
of the membrane is controlled by a control panel sudden change in the direction of the membrane
which is connected to the oscillator motor provid- which induces a velocity difference in the
ing separate control over the frequency and column of liquid inside the membrane and local-
membrane displacement. Azimuthally oscillating ized high pressure at the peak positions of the
ME operates in a continuous mode, and it is suit- vertical oscillation (i.e., top and bottom of the
able for large-scale production providing high membrane tube).
Dynamic Membrane Emulsification 615
In dynamic ME, the droplet breakup process is 1

tmax 0:825mc or trans for r < r trans (2)
correlated with a dripping or jetting behavior d
depending on the relative strength of the different
r 0:6 1
forces acting on the process. In the dripping region, t 0:825c or trans
trans
for r < r trans
the inertial force is negligible compared to the drag r d
force or unable to overcome the surface tension (3)
force. Thus, the droplet size is affected mainly by
the drag force imparted by the continuous phase where mc [Pa s] is the continuous phase viscosity,
ow or moving membrane and the surface tension rtrans is the transitional radius, and d is the bound- D
force. The inertial force of the dispersed phase has ary layer thickness given
qfrom the Landau-
negligible effect in drop detachment in the dripping Lifshitz equation d m=or . The shear stress
mode. It is shown that the droplet size becomes is not uniformly distributed over the membrane
smaller as the shear stress increases and the inu- surface, and it can be assumed that the maximal
ence is more signicant for low shear values and shear (tmax) is reached at distance rtrans from the
smaller pore size. At high dispersed phase ow center of the membrane; rtrans is the transitional
rates, the inertial force exceeds the surface tension radius in which the rotation changed from a free
force, leading to a transition to jetting behavior. In vortex to a forced vortex.
this case, the droplet detachment point moves fur- In rotating ME, the shear stress is directly pro-
ther downstream from the pore usually resulting in portional to the membrane rotational speed (n)
the formation of a larger droplet than when in the [rpm] but depends also on the width of the
dripping regime. annular gap between the rotating membrane
The different process conditions can be accu- and the stationary vessel. The shear at the
rately tuned to control the shear stress in dynamic surface of the membrane is calculated using
ME according to the mechanism used to generate Eq. 4:
the drag force (owing of continuous phase or
moving membrane); the following provides some I?R2 n
I? 2 1 2 (4)
simplied equations for shear at the membrane 15 R2 R1
surface according to the method of generation of
that shear: where R1 is the radius of the rotating membrane
and R2 is the radius of the stationary vessel.
In cross-ow ME, the shear stress (t) [Pa] is Pulsed ME and vibrating and azimuthally
constant over the tubular membrane area, and it oscillating ME are based on the generation of
is correlated with the tangential velocity of the the shear stress by oscillation of either the
owing continuous phase (v) [ms1], the den- continuous phase or the membrane. Thus,
sity of the emulsion (r) [kgm3], and the fric- there are two parameters affecting shear stress
tion factor (l) which is calculated depending on the membrane surface: frequency (f)
on the Reynolds number (l = 16/Re for [Hz] and amplitude (a) [m] of the oscillation.
Re < 2000; l = 0.0792 Re0.25 for During the oscillation, the shear is variable and
Re > 2000) Eq. 1: the emerging drop detaches when it experi-
ences the maximum shear stress which is cal-
culated using Eq. 5:
lrn2
t (1)
2 3= 1=
tmax 2apf 2 mr 2
(5)
In stirred ME, the shear stress (t) [Pa] depends
on the angular velocity (o) [s1] of the stirrer. where m is the viscosity and r is the density of the
The shear over the whole membrane area can continuous phase. A peak shear event occurs twice
be calculated with the Eqs. 2 and 3: per oscillation: once in each direction that the wave
616 Dynamic Membrane Microfiltration
is moving, for a regular wave form such as a order to prevent the buildup of a layer of rejected
sine wave. solutes on the membrane while exerting a pressure
on the uid to produce a ltrate. The combination
of high feed pressure and ow rate requires pow-
erful and expensive pumps that consume energy
References and the friction creates a pressure decay along the
membrane which reduces the transmembrane
Holdich RG, Dragosavac MM, Vladisavljevic GT
(2010) Membrane emulsication with oscillating and pressure (TMP). In addition, the rejected mole-
stationary membranes. Ind Eng Chem Res cules or particles can create a cake on a
49(8):38103817 microltration (MF) membrane which reduces
Holdich RG, Dragosavac MM, Vladisavljevic GT,
both ltration rate and solute transmission
Piacentini E (2013) Continuous membrane emulsica-
tion with pulsed (oscillatory) ow. Ind Eng Chem Res through the membrane.
52:507515 Dynamic ltration is a recent alternative to
Peng SJ, Fellow RAW (1998) Controlled production of classical cross-ow ltration (Lee et al. 1995;
emulsions using a crossow membrane: part II: indus-
Jaffrin 2008) consisting in creating a high shear
trial scale manufacture. Chem Eng Res Des
76(8):902910 rate (velocity gradient) at the membrane by a disk
Piacentini E, Drioli E, Giorno L (2014) Pulsed back-and- or a rotor rotating near the membrane or by using
forward cross-ow batch membrane emulsication rotating or vibrating membranes, while using a
with high productivity to obtain highly uniform and
low feed ow rate, only slightly larger than the
concentrate emulsions. J Membr Sci 453:119125
Silva PS, Dragosavac MM, Vladisavljevi GT, permeate ow rate.
Bandulasena HCH, Holdich RG, Stillwell M, Williams
B (2015) Azimuthally oscillating membrane emulsi-
cation for controlled droplet production. AIChE
J. doi:10.1002/aic.14894
Advantages and Drawbacks of Dynamic
Stillwell MT, Holdich RG, Kosvintsev SR, Gasparini G, Filtration
Cumming IW (2007) Stirred cell membrane emulsi-
cation and factors inuencing dispersion drop size and The rst advantage is a substantial increase in
uniformity. Ind Eng Chem Res 46(3):965972
permeate ux at all membranes cut-off due to
Vladisavljevi GT, Williams RA (2006) Manufacture of
large uniform droplets using rotating membrane emul- shear rates which can reach 3 105 s1 when
sication. J Colloid Interface Sci 299(1):396402 using disks with radial vanes rotating at high
speed. Due to moderate uid velocities inside
the module, the transmembrane pressure remains
relatively uniform, which contributes to ux
increase. Clarication of a suspension by MF
Dynamic Membrane Microfiltration requires a good solute transmission which is aug-
mented by cake reduction. In ultraltration (UF),
Michel Jaffrin nanoltration (NF), and reverse osmosis (RO) the
Department of Biological Engineering, permeate ux is mainly limited by the concentra-
Technological University of Compiegne, BP tion polarization layer made of molecules rejected
20529, Compiegne Cedex, France by the membrane. This layer is diminished by
high shear rates which increases back transport
of solutes away from the membrane. The ux
Introduction keeps increasing with rising TMP until pressures
of 40 bars at high shear rates. The combination of
To separate molecules or particles from a solution these two factors, high TMP and high shear rates,
by cross-ow ltration through a membrane, it is yields frequently permeate uxes ve times
necessary to circulate the uid at a speed of higher than in cross-ow ltration with the same
36 ms1 in a tubular or plane membrane in membrane and same uid. In addition, in NF and
Dynamic Membrane Microfiltration 617
Dynamic Membrane
Microfiltration,
Fig. 1 Industrial VSEP
vibrating modules
(Courtesy of New Logic
Research)
Dynamic Membrane
Microfiltration,
Fig. 2 Dyno rotating disk
module (Bokela, Germany)
RO applications, such as wastewater treatment, it disks rotating around several parallel shafts in a
is important to have a small microsolutes concen- closed tank, with a total membrane area exceeding
tration in permeate. This is the case in dynamic 120 m2 which are cheaper to build and to service
ltration as the diffusive mass transfer through the than multicompartment modules with metal disks
membrane is greatly reduced by the small solute rotating between xed organic membranes.
concentration at the membrane. Rotating disks modules give a choice between
The main drawbacks are higher costs of equip- using high rotation speeds to maximize the ux
ment and maintenance due to moving parts and and thus minimize membrane area and investment
more complex design and membrane area limita- cost, or to use moderate speeds to obtain the same
tions for some systems. The recent availability of ux as with tubular membranes but saving more
30 cm diameter ceramic membrane disks has per- than 60 % of the energy spent per m3 of permeate
mitted the development of systems with ceramic with these membranes.
618 Dynamic Membrane Microfiltration
Dynamic Membrane
peripheral
Microfiltration, strip rotor
Fig. 3 Sketch of OptiFilter
CR, Metso paper Co,
Finland
support layer
clamp ring
support plate
membrane
concentrate
water out
feed filtrate
water in water out
Dynamic Membrane Microfiltration, Table 1 List of industrial dynamic filtration modules and their characteristics
(Jaffrin 2008)
Membrane Disk MaxTMP
Manufacturer Models Type area m2 Pore size diam, mm bar
Bokela Dyno Rotating disks 0.01312 10 nm1 mm 137850 4
(Germany) organic
Andritz Kraus-M Membrane rotating on 0.0516.4 7 nm2 mm 2
(Germany) DCF two shafts ceramic
Novoow SSDF Membrane rotating on 125 7 nm2 mm 152500 2,5
(Germany) CRD one shaft ceramic
Metso paper Optilter Rotors between 15-84-140 MF-UF 5501000 34
(Finland) CR membranes Organic
Spintek (US) ST Membr rot on 1 shaft 0.12.3 MF-UF 275
Org/ceram
New Logic VSEP 0.05150 organic Organic MF-UF-NF- 300500 40
Research, (US) RO
Industrial Filtration Modules They consist of a stack of circular organic mem-

branes in separate compartments oscillating azi-
VSEP Vibrating Modules (New Logic Research, muthally around a vertical shaft at a resonant
Ca, USA) frequency of about 60 Hz. The maximum mem-
Close to 400 VSEP systems have been installed brane area is 150 m2 and the power spent in
worldwide since their introduction around 1995. vibrations is only 9 kW (Fig 1). It can sustain
Dynamic Optimization of Membrane Operation 619
pressures of 40 bars and is used with a wide choice

of membranes from MF to RO Al-Akoum Dynamic Optimization of Membrane
et al. (2002). Operation
Single Shaft Rotating Disks Systems Radoslav Paulen

Bokela GmbH (Karlsruhe, Germany) commer- Department of Biochemical and Chemical
cializes the Dyno with metal disks rotating Engineering, Technische Universitt Dortmund,
between xed circular membranes of 12 m2 for Dortmund, Germany
D
MF and UF (Fig. 2).
Metso Paper (Raisio, Finland) manufactures
the Optilter CR with blades rotating between Dynamic optimization is usually referred to open-
xed membranes of 1 m diameter and up to loop optimal control. It encompasses several
140 m2 membrane area with a 132 kW motor methods whose purpose is to determine trajecto-
(Fig. 3). ries of dynamic (time-dependent) variables in
optimal manner. These methods may be used to
Multishaft Rotating Ceramic Membrane nd optimal control of processes or to design
systems optimal experiments for parameter estimation.
KMPT GmbH (Vierkirchen, Germany) commer- The generalized representation of dynamic opti-
cializes a two-shaft module with 15 cm diameter mization problem (DOP) is as follows
overlapping ceramic membranes of 16.4 m2 area tf

(Table 1). min G x tf , tf F xt, u t, tdt (1)
u t t0
:
s:t: x t fxt, u t, t (2)
Conclusion
h xt0 , t0 , xt0 , tf 0 (3)
Dynamic ltration is not intended for initial treat-
ment of huge volumes of uids, but it should be g xt, u t, t 0 (4)
well adapted to end of pipe treatment when the
where t is time, u(t) denotes vector of control
goal is to reach high solid concentration in
variables, u(t) is vector of state variables, and
retentate or to extract valuable components in
G() and F() stand, respectively, for nonintegral
highly charged suspensions.
and integral part of optimization objective. The set
of differential Eq. (2) represents dynamical sys-
tem. This is usually the model of the process
whose optimal control one searches for. Sets of
References
functions h() and g() may express various types
Al-Akoum O, Jaffrin MY, Ding LH, Paullier P, Vanhoutte of equality/inequality constraints which are
C (2002) An hydrodynamic investigation of imposed by the physical reality.
microltration and ultraltration in a vibrating mem- In order to nd solution to DOP, one can
brane module. J Membr Sci 197:3752
exploit various techniques, stochastic (such as
Jaffrin MY (2008) Dynamic shear-enhanced membrane
ltration: a review of rotating disks, rotating genetic algorithms, simulated annealing, etc.) or
membranes and vibrating systems. J Membr Sci deterministic ones. Deterministic methods are
324:725 based on the following three principles: varia-
Lee SA, Russoti BG, Buckland B (1995) Microltration
tional calculus (developed by Euler and
of recombinant yeast cells using a rotating disk
dynamic ltration system. Biotechnol Bioeng 48: Lagrange), dynamic programming (Bellman
386400 1957), and Pontryagins maximum/minimum
620 Dynamic-Volume Diafiltration
principle (Pontryagin et al. 1962). The latest one Pontryagin LS, Boltyanskii VG, Gamkrelidze RV,
gave rise to most popular numerical techniques Mishchenko EF (1962) The mathematical theory of
optimal processes. Wiley, New York
such as control vector parameterization, control Rahimpour MR, Elekaei Behjati H (2009) Dynamic opti-
vector iteration, boundary condition iteration, mization of membrane dual-type methanol reactor in
orthogonal collocation, and multiple shooting the presence of catalyst deactivation using genetic algo-
techniques. rithm. Fuel Proc Technol 90(2):279291. doi:10.1016/
j.fuproc.2008.09.008, ISSN 0378-3820
Various dynamic optimization problems were Van Boxtel AJB, Otten ZEH (1993) New strategies for
solved already in literature on membrane pro- economic optimal membrane fouling control based on
cesses. These applications include optimization dynamic optimization. Desalination 90(13):363377.
of membrane reactors performance (Parvasi doi:10.1016/0011-9164(93)80187-R, ISSN 0011-9164
Zondervan E, Roffel B (2008) Dynamic optimization
et al. 2009; Rahimpour and Elekaei Behjati of chemical cleaning in dead-end ultra ltration.
2009), other membrane-assisted processes (Kuhn J Membr Sci 307(2):309313. doi:10.1016/j.
et al. 2009; Bui et al. 2010), membrane separation memsci.2007.09.067, ISSN 0376-7388
(Fikar et al. 2010; Paulen et al. 2011) as well as
membrane fouling and cleaning (Van Boxtel and
Otten 1993; Blankert et al. 2006; Zondervan and Dynamic-Volume Diafiltration
Roffel 2008).
Zoltn Kovcs
Department of Food Engineering, Institue of
References Bioengineering and Process Engineering, Szent
Istvan University, Budapest, Hungary
Bellman R (1957) Dynamic programming. Princeton
University Press, Princeton, United States
Blankert B, Betlem BHL, Roffel B (2006) Dynamic
Dynamic-volume dialtration is a type of batch
optimization of a dead-end ltration trajectory:
blocking ltration laws. J Membr Sci dialtration process that is designed to achieve the
285(12):9095. doi:10.1016/j.memsci.2006.07.044, twin aims of simultaneous macrosolute concen-
ISSN 0376-7388 tration and microsolute removal by employing
Bui VA, Vu LTT, Nguyen MH (2010) Simulation and
optimisation of direct contact membrane distillation
optimal time-varying control of diluant feeding.
for energy efciency. Desalination 259(13):2937. The technological goal in a general batch pro-
doi:10.1016/j.desal.2010.04.041, ISSN 0011-9164 cess is to increase the macrosolute concentration
Fikar M, Kovcs Z, Czermak P (2010) Dynamic optimiza- from c1,0 to c1,f and to reduce the microsolute con-
tion of batch dialtration processes. J Membr
Sci 355(12):168174. doi:10.1016/j.memsci.2010.
centration from c2,0 to c2,f. The fractionation is
03.019 accomplished by introducing a diluant into the
Kuhn J, Lakerveld R, Kramer HJM, Grievink J, Jansens PJ feed reservoir. Choosing the right strategy of sched-
(2009) Characterization and dynamic optimization of uling the addition of diluant is a critical aspect to
membrane-assisted crystallization of adipic acid. Ind
consider in batch processing (Lutz 1997). Processes
Eng Chem Res 48(11):53605369. doi:10.1021/
ie802010z can be characterized upon their a-strategies (a is the
Parvasi P, Khosravanipour Mostafazadeh A, Rahimpour ratio of diluant ow rate to permeate ow rate).
MR (2009) Dynamic modeling and optimization of a Frequently used batch dialtration techniques,
novel methanol synthesis loop with hydrogen-
permselective membrane reactor. Int J Hydrog Energy
such as traditional dialtration, variable-volume
34(9):37173733. doi:10.1016/j. dialtration, pre-concentration followed by
ijhydene.2009.02.062, ISSN 0360-3199 variable-volume dialtration, intermittent feed
Paulen R, Foley G, Fikar M, Kovcs, Czermak P (2011) dialtration, and sequential dilution dialtration,
Minimizing the process time for ultraltration/
dialtration under gel polarization conditions.
employ arbitrarily predened schemes of diluant
J Membr Sci 380(12):148154. doi:10.1016/j. addition. In contrast, dynamic-volume dialtration
memsci.2011.06.044, ISSN 0376-7388 is designed by the evaluation of the optimal time-
Dynamic-Volume Diafiltration 621
varying prole of diluant ow for the entire course or minimization of complex economic index:
of operation.
The process of dynamic-volume dialtration tf
poses a dynamic optimization problem (Fikar JC k1 tf k2 c1 1 R1 t qt dt
et al. 2010). The goal of optimization is to nd tf 0
the optimal time-dependent function of a that k3 atqt dt
drives the process from its initial state to a pre- 0
scribed terminal state. Optimality is dened as the
minimization of the objective function J without D
where total cost is a sum of three terms that are the
violating given constraints. The mathematical operational cost of the pump, cost of the valuable
problem can be formulated as follows: component loss, and the cost of the utilized dilu-
min J s:t: tion water. k1, k2, and k3, are the respective cost
at factors.
: c1 q This technique can be useful to evaluate opti-
c1 R1 a, c1 0 c1, 0,
V mal operation of membrane ltration plants
: c2 q
c1 tf c1, f , c2 R2 a, including reverse osmosis, ultraltration,
V nanoltration, and microltration systems
c2 0 c2, 0, c2 tf c2, f , (Paulen et al. 2012). In general, optimal a control
:
V a 1q, V 0 V 0 , need not to be any of the conventional dialtration
a 0, amax concepts, and thus, it might need advanced con-
trol systems. In many cases, however, it might be
where c1, c2, and V represent the actual attractive to implement the optimal trajectory
macrosolute concentration, microsolute concen- since it can lead to reduced operation time, diluant
tration, and volume in the feed tank, respectively. consumption, and product losses. The computed
It is assumed that the system is well mixed and the optimal prole may vary case by case depending
diluant is solute-free. R1 and R2 denote the rejec- on the complexity of the membrane response
tion of macrosolute and microsolute, respectively. model, the initial and nal values, and the con-
It should be mentioned that, in a general case, the straints involved in the model.
actual value of both q and R depends on the
concentrations of macrosolute as well as on
microsolute. Moreover, process parameters (e.g., References
pressure, temperature, hydrodynamic conditions,
etc.) also affect q and R. Fikar M, Kovcs Z, Czermak P (2010) Dynamic optimiza-
In practice, several objective functions can be tion of batch dialtration processes. J Membr
Sci 355(12):168174. doi:10.1016/j.memsci.2010.
considered for optimization, for instance,time 03.019
minimization: Lutz H (1997) Membrane ltration with optimized addi-
tion of second liquid to maximize ux, United States
J A tf Patent 5,597,486. Assignee: Millipore Investment
Holdings Limited, Wilmington
diluant minimization: Paulen R, Fikar M, Foley G, Kovcs Z, Czermak P
tf (2012) Optimal feeding strategy of dialtration buffer
JB atqt dt in batch membrane processes. J Membr Sci
0 411412:160172
E
Ease of Expansion in Gas Separation References
Adele Brunetti Brunetti A, Bernardo P, Drioli E, Barbieri G (2010) Mem-

brane engineering progresses and potentialities in gas
separations. In: Yampolskii Y, Freeman B (eds) Mem-
Research Council of Italy, ITM-CNR, Rende, brane gas separation. Wiley, New York, pp 281312
Italy Miller GQ, Stcker J (1989) Selection of a hydrogen sep-
aration process NPRA annual meeting, 1921 Mar,
San Francisco
Quite often during the operation of a system for

the treatment of gases, it is necessary to expand it
for treating greater streams. In some cases, future Effective Diffusivity
expansion is contemplated even during the initial
phase of a project. In other cases, it could be a Renzo Di Felice
necessity not foreseen during system design phase University of Genova, Genova, Italy
(Miller and Stcker 1989; Brunetti et al. 2010).
Membrane system expansion is very easy,
since this only requires the addition of identical Effective diffusivity is a convenient parameter
modules. This is the advantage offered by the which is introduced when diffusion takes place
modularity of membrane units and the reduced in non-homogenous media. Consider, for exam-
equipment and control systems required for oper- ple, the case where a species is diffusing through
ating it. In comparison, considering the other ref- porous particles, such as reactant diffusing inside
erence technologies for gas separation, PSA and a catalytic solid. In this case, the molecules have
absorption systems can also be expanded, but it to travel for a longer distance given that the pores
requires additional design considerations and of the catalytic particles are not straight, and
adds cost in the initial phase of the project. moreover diffusion takes place over a smaller
The cryogenic units cannot be expanded if it is area due to the solid being wall impermeable.
not foreseen during the design phase. Generally These effects are taken into account by dening
they can be over-dimensioned, and a capacity an effective diffusivity as
increase is often obtained without modication
e
to the cold box itself through addition of a tail
Deff D
gas compressor. t

DOI 10.1007/978-3-662-44324-8
624 Effectiveness Factor
where e is the particle void fraction and thereby References

takes into account the reduced ow area available
and t is the tortuosity which considers the longer Di Felice R, Cazzola D, Cobror S, Oriani L (2008) Oxygen
permeation in PET bottles with passive and active
distance traveled by the molecules. It should be
walls. Packag Technol Sci 21:405415
stressed, however, that more often than not, the Maxwell JC (1873) A treatise on electricity and magne-
parameter inside the parenthesis is utilized as a tism, vol 1. Clarendon Press, Oxford
tting factor derived from experimental data Nielsen LE (1967) Models for the permeability of lled
polymer systems. J Macromol Sci A1(5):929942
rather than predicting factor from the solid phys-
ical characteristics. These result in the reported
tortuosity factor, for example, to assume rather
wild values, as high as ten, which are difcult to
justify with geometrical arguments alone.
Effectiveness Factor
The concept of effective diffusivity is also
Renzo Di Felice
applied for the case of species diffusing in hetero-
geneous media, such a polymer lled with a sec-
ond, less permeable, material. Maxwell (1873)
has obtained an exact expression, in the case of The concept of effectiveness factor is utilized
composite media lled with spheres, for the effec- when the effect of diffusion on the overall rate of
tive diffusivity given as a reacting process wants to be taken into account.
Lets consider a catalytic solid particle where a
2 1 1 1
2f reaction takes place. For a species to react, it must,
D D D D
Deff D S S at the same time, diffuse inside the catalytic parti-
2 1 1 1 cle and react. It is intuitive that the diffusion step
f
DS D DS D slows the overall process rate, and an effective-
ness factor is dened as
where D is the diffusion coefcient in the primary
media, Ds the diffusion coefcient through the actual rate of process
spheres, and f the sphere volume fraction. For
rate of process if diffusion were infinitively fast
the limiting case of the spheres being completely
impermeable, Maxwell expression simplify to This denition is equivalent to consider reactant
concentration on any point in the catalytic particle
1f
Deff D equal to that at the external surface.
f A rather straightforward mass balance allows
1
2 the determination of the effectiveness factor for a
Various empirical expressions have been reported single cylindrical pore where a rst-order reaction
in literature in order to take into account the devi- is taking place on the wall:
ation from sphericity of the foreign material. Par-
tanhf
ticularly simple and efcient is the expression
f
proposed by Nielsen (1967):
where f is the Thiele module dened as
1f
Deff D s
af
1 k
2 fL
Deff
where a is the dispersed material aspect ratio.
Nielsen relationship has been shown to do a sat- with L the pore length, k the chemical reaction
isfactory job against experimental evidence kinetic factor, and Deff the reacting species effec-
(Di Felice et al. 2008). tive diffusivity.
Electrical Double Layer 625
As expected the effectiveness factor approa-

ches 1 when Thieles module is very small, i.e.,
for very short pores or very slow reaction or very
high diffusion coefcients. On the other hand, for
very high value of Thiele modules, effectiveness
factor approaches the inverse of the Thiele
module.
The above result can be easily generalized for
the case of at plate by putting L equal to half the
plate width, for the case of long cylindrical pellets
E
by putting L equal to half the pellet radius, and to
the case of spherical pellets by putting L equal to
one third the sphere radius. Extension to kinetic
different from the rst order has been presented in
literature, and a summary can be found, for exam-
ple, in Sattereld (1991).
Electrical Double Layer, Fig. 1 Schematic sketch of
structure of the Electrical Double Layer for (a)
References functionalized and (b) nonfunctionaalized membrane
materials
Sattereld CN (1991) Heterogeneous catalysis in industrial
practice. McGraw-Hill, New York
void volume of the material (ion-selective mem-

branes), and (iv) nonfunctionalized polymeric or
ceramic membranes.
Electrical Double Layer Whereas, in the rst two cases, the traditional
denition is valid, in the last two, the situation is
Karel Bouzek and Tomas Bystron different. This is due to the fact that the bulk
Faculty of Chemical Technology, University of membrane material is neither an electronic nor
Chemistry and Technology Prague, Technick 5, an ionic conductor. Hence the double layer only
Prague 6, Czech Republic forms on one side of the phase interface. The
Fig. 1 provides a schematic sketch of the two
mentioned cases. In the case of functionalized
An electrical double layer is usually considered to materials, the charge-carrying groups covalently
form at the phase interface between two electri- bound to the polymeric backbone are oriented
cally conducting media. In such a case, on one towards the void volume of the membrane inte-
side of the interface, an excess of positive charge rior, thus forming an electrically charged lm
exists, which is counterbalanced by an identical covering the phase interface on the side of the
excess of negative charge localized on the oppo- solution. The charge of this lm is compensated
site side of the interface. The overall system by that of the mobile ions present in the solution
charge is thus equal to zero. In membrane science, and located close to the lm (see Fig. 1a).
this denition has, to a certain degree, been mod- In the case of the non-functionalized materials,
ied, the reason being that membranes can be the surface charge xed to the surface of the solid
produced from a broad spectrum of materials, phase is formed by specic adsorption of ions of
including (i) electronic conductors (e.g., metals), one sign. Both the sign of the adsorbed ions and
(ii) ionic conductors (e.g., O2-conducting the extent of the adsorption are determined by the
ceramics), (iii) functionalized polymers providing properties of the membrane material. In this case,
ionic conductivity through the liquid lling the too, the xed charge is compensated by the charge
626 Electrical Interactions in Membranes
of ions moving in the system and located close to membrane separation processes, such as the exis-
the absorbed surface lm (see Fig. 1b). tence of Donnan potential, the selectivity of mass
An electrical double layer signicantly inu- transport, or mass transport enhancement under
ences the properties of the membrane, mainly with current load by electroosmotic ux, but also for
respect to its transport properties and behavior the existence of life itself. These facts document
under current load. In selected cases the surface the importance of electrical interactions in mem-
of the membrane materials is modied to achieve brane materials with respect to their transport
the preferred adsorption and thus the required properties.
transport and properties, including selectivity.
References
Electrical Interactions in Membranes Stuart McLaughlin (1989) The Electrostatic Properties of

Membranes. Annu. Rev. Biophys. Biophys. Chem.
18:113136
Karel Bouzek
Faculty of Chemical Technology, University of
Chemistry and Technology Prague, Technick 5,
Prague 6, Czech Republic
Electrical Potential
In general electrical (electrostatic) interactions Karel Bouzek and Tomas Bystron

arise as a result of the interaction of electrically Faculty of Chemical Technology, University of
charged particles (charge carriers) with sur- Chemistry and Technology, Prague, Technick 5,
rounding matter (electrically charged and neu- Prague 6, Czech Republic
tral) (Stuart 1989). In the case of membranes,
probably the most important interactions to con-
sider are those between electrically charged bod- Electric potential is a scalar physical quantity
ies, such as ions dissolved in solution and determined by the electric energy (or work in
ion-conducting, functionalized, or even inert general) necessary to transfer a unit positive elec-
membranes. In living systems the electrostatic tric charge from innity to the point of interest in a
interactions between the charged/polar and neu- studied system. The point in innity is formally
tral/nonpolar side of a solute molecules with each chosen as a reference state characterized by zero
other and with a solvent are responsible for the electric potential. From this more or less abstract
formation/assembly of monolayer and bilayer and intuitive denition of the reference state, it is
biological membranes and their functioning. apparent that the absolute value of electric poten-
On the uid side the charge carriers are typi- tial is not available experimentally. However, it is
cally mobile ions or colloidal particles of organic possible to measure electric potential difference,
and inorganic nature. These carriers interact with known as voltage.
the charge carriers xed in the material of the We can distinguish between outer (external)
membrane (see also the electrical double layer). and inner (internal) electric potential. In the case
The charge xed in the membrane material is, in of outer electric potential, the charge is transferred
principle, of dual origin: (i) it comes from the to the closest proximity of the external surface of
component introduced and incorporated during the phase P. However, the distance from this sur-
its synthesis, and (ii) it is the result of the specic face has to be large enough to avoid the appear-
adsorption of ions on the surface of the solid ance of molecular and image forces. At the same
phase. Electrostatic interactions between xed time it should not cause any weakening of the
and freely moving charge carriers are responsible electric interaction with the charge located on the
not only for several phenomena characteristic of internal side of the phase interface. The inner
Electrically Enhanced Processes by Membrane Operations 627
electric potential is determined by the strength of

the electric eld in the interior of the phase P. It Electrically Enhanced Processes by
thus represents the work necessary to transport the Membrane Operations
charge from innity to the phase P. The difference
between the inner electric potential and the outer Karel Bouzek
electric potential corresponds to the surface poten- Faculty of Chemical Technology, University of
tial of phase P (Arthur and Alice 1997). Chemistry and Technology Prague, Technick 5,
Prague 6, Czech Republic
References E
Electrically membrane operations are typically
Arthur W. Adamson, Alice P. Gast (1997) Elecrical limited to systems involving electrically charged
Aspects of Surface Chemistry, in Physical Chemistry
of Surfaces, 6th edn. John Wiley & Sons, Inc.
particles. In such a case, the application of an
ISBN:0-471-14873-3) electric eld can often signicantly enhance the
ongoing process. In selected cases, it is the appli-
cation of an external electric eld that makes the
desired processes possible at all. By applying an
electric eld, electrically charged particles are
Electrical Work in Membranes dragged in the direction determined by the polar-
ity of their charge and the vector of the electric
Karel Bouzek eld. This movement is responsible for electric
Faculty of Chemical Technology, University of current ow through the system. Part of the
Chemistry and Technology Prague, Technick 5, electric charge carried by one type of particles
Prague 6, Czech Republic corresponds to what is known as the transference
number of the given particle. It is proportional to
the particle mobility in the system and the electric
An electrical work is dened as the work charge it carries (Colin 1976).
connected with the charge carrier transport In the case of membrane operations enhanced
between two points in an electric eld, i.e., from by an electric eld, generally functionalized
point of potential 1 to point of potential 2. Thus, ion-selective membranes are considered. Electro-
electrical work in membranes is the product of the dialysis and dialysis are typical representatives of
transported electrical charge and the difference in an electrically enhanced and a nonenhanced pro-
electric potentials between which it has passed cess, respectively. Electrodialysis is used to desa-
(Arthur and Alice 1997). Since the resistivity of linate a treated stream containing dissolved ions
the membrane can be approximated by ohmic by transporting them, due to the action of the
resistivity, the electrical work dissipated in the electric eld, against the concentration gradient
membrane is equal to the ohmic potential loss in to the concentrate stream (Fernando et al. 2011).
the membrane multiplied by the current load. In the case of dialysis the situation is the reverse,
Expressed alternatively, it is equal to the square since the dissolved species move in the direction
of the current load multiplied by the ohmic of the concentration gradient away from the
resistivity. treated stream to the solution free of the removed
components owing along the opposite side of the
membrane. It is thus clear that electrodialysis is
References typically a signicantly more intensive process in
terms of the desalination rate. At the same time, it
Arthur W. Adamson, Alice P. Gast (1997) Elecrical
produces a signicantly lower volume of the
Aspects of Surface Chemistry, in Physical Chemistry
of Surfaces, 6th edn. John Wiley & Sons, Inc. ISBN: stream carrying the removed components. The
0-471-14873-3 electric eld thus enhances the intensity of the
628 Electrochemical Deposition
mass ux during membrane processes involving

charged species and makes it possible to transport Electrochemical Impedance
the charged species against the concentration gra- Spectroscopy (EIS)
dient. On the other hand, dialysis is not limited to
charged particles. Impedance Spectroscopy
Electrically enhanced membrane operations
can also be carried out with nonfunctionalized
membrane materials or with membrane materials
carrying no xed electric charge, electroltration Electrochemical Membrane
being an example. Here, the applied electric eld Bioreactor
inuences the membrane operation indirectly by
preventing blockage of the membrane surface. Bioelectrochemical MBR
This is ensured by the electrophoretic transport
of the charged colloidal particles concentrated in
the solution away from the surface of the ltering
membrane, see also Electroltration. Electrochemical Processing
It is, therefore, evident that the application of
an external electrical eld can signicantly Tanja Vidakovic-Koch
enhance or modify processes, including mem- Max Planck Institute for Dynamics of Complex
brane operations. Technical Systems, Magdeburg, Germany
References In electrochemical processing, electrical energy is

supplied to or obtained from the electrochemical
Colin A (1976) Vincent: The Motion of Ions in Solution system in order for chemical production or energy
under the Inuence of an Electric Field. J. Chem. Educ, conversion to take place (Bard and Stratmann
vol. 53, pg. 490
Fernando Valero, Angel Barcel and Ramon Arbos
2007). The rst group of processes, also called
(2011). Electrodialysis Technology - Theory and electrolytic, is not spontaneous. The second group
Applications, Desalination, Trends and Technologies, of processes, called galvanic, is spontaneous and it
Michael Schorr (Ed.), ISBN: 978-953-307-311-8, delivers electrical energy. Electrolytic processes
InTech, DOI: 10.5772/14297. Available from: http://
www.intechopen.com/books/desalination-trends-and-
can be further divided into two categories: inor-
technologies/electrodialysis-technology-theory-and- ganic and organic. In inorganic electrochemical
applications processing, important commodity chemicals such
as sodium hydroxide, chlorine, and pure metals are
produced. The major inorganic electrochemical
processing technologies are chlor-alkali electroly-
Electrochemical Deposition sis and electrowinning of metals like aluminum or
copper. Nowadays, hydrogen production by water
Track-Etch Membranes as Tools for Template electrolysis gets more on importance in context of
Synthesis of Nano-/Microstructures and Devices chemical storage of renewable electrical energy
(wind and photovoltaic) in hydrogen. The most
signicant commercial electroorganic synthesis is
Monsantos electrohydrodimerization (EHD) of
Electrochemical Impedance acrylonitrile to adiponitrile. Adiponitrile has an
Spectroscopy importance in production of nylon 6-6. Examples
of galvanic systems are fuel cells and batteries. The
Impedance Spectroscopy, Membrane Charac- main product of galvanic systems is electrical
terization by energy.
Electrochemical Regeneration 629
Electrochemical systems have some intrinsic

advantages over other types of chemical systems Electrochemical Regeneration
like better control of a reaction rate, operation at
lower temperatures, and less environmental Tanja Vidakovic-Koch
impact. They take place in an electrochemical Max Planck Institute for Dynamics of Complex
reactor. The design of an electrochemical reactor Technical Systems, Magdeburg, Germany
is inuenced by the state of aggregation of reac-
tants (gas, liquid, or solid), necessity of reactants
and/or products separation, required mass trans- In electrochemical regeneration electrical energy
port conditions, and electrode materials. If prod- is applied to restore some important property, like
E
uct or reactant separation is required, an adsorption capacity or catalyst activity of a tech-
electrochemical reactor must contain a separator, nical system. Electrochemical regeneration relies
which is a membrane. Major requirements on a on principles of electrochemistry and relates to
membrane are good separation efciency, low electrochemical processing. Electrochemical
electrical resistance, no electron conductivity, regeneration is conveniently conducted in situ
low cost, long operating life time, good dimen- with an electron as only reagent requiring simple
sional stability, and resistance to plugging and handling and equipment. A technical setup for
fouling. In general permeable and semipermeable electrochemical regeneration requires in general
membranes have been applied in electrochemical two electrodes, an electrolyte and a power supply.
processing. Permeable membranes are porous In addition a membrane can be added to the setup
materials lled with liquid electrolyte which per- in order to separate anode and cathode depart-
mit the bulk ow of liquid through their structure ment. Some examples of electrochemical regen-
and are thus nonselective regarding transport of eration are electrochemical regeneration of
ions or neutral molecules. In electrochemical pro- activated carbon-based adsorbents in wastewater
cesses, these are also referred to as diaphragms. treatment and regeneration of enzymatic cofactors
Permeable membranes can be made of inorganic in electroenzymatic processes.
and organic materials and composites. Examples Organic pollutants in wastewaters can be
of these materials are asbestos (chlor-alkali elec- removed by adsorption using, e.g., activated car-
trolysis), polymers like polyethylene and polypro- bon as an adsorbent. This process is normally
pylene (batteries), or composites like polymer operated using a batch of adsorbent with sufcient
(polypropylene)-modied asbestos. Semiperme- capacity to operate for many months before
able membranes permit the selective passage reaching saturation. Once loaded, adsorbent
of certain species by virtue of molecular size must be disposed or regenerated. One option for
or charge. In electrochemical processes, adsorbent regeneration is electrochemical regen-
ion-conducting membranes (see solid electrolyte) eration (Brown et al. 2004). The loaded adsorbent
are broadly applied. In general, ion-conducting is located in a form of a packed or uidized bed in
membranes have higher separation efciency the anode (anodic regeneration) or cathode
and lower electrical resistance than diaphragms, (cathodic regeneration) compartment of the reac-
but they are also more costly and impose higher tor. The efciency of the regeneration depends on
requirements on system purity. the processing time, voltage gradient, an electro-
lyte, and a compartment. According to literature
the efciency of cathodic regeneration is higher
than of anodic regeneration. The mechanism of
References electrochemical regeneration is ascribed at the
rst place to local pH changes close to anode or
Bard AJ, Stratmann M (eds) (2007) Encyclopedia of elec-
cathode. At the anode side due to oxygen evolu-
trochemistry. Macdonald DD, Schmuki P (eds) Elec-
trochemical engineering, vol 5. Wiley-VCH Verlag tion reaction a pH decrease can be expected, while
GmbH & Co. KGaA, Weinheim at the cathode side due to hydrogen evolution pH
630 Electrochemistry
value will increase. This pH changes induce References

organic pollutants desorption. In the next step,
dissolved pollutants can be oxidized at the anode. Brown NW, Roberts EPL, Garforth AA, Dryfe RAW
(2004) Electrochemical regeneration of a carbon-
In the case of the cathodic regeneration they have
based adsorbent loaded with crystal violet dye.
rst to mitigate from the cathode to the anode. This Electrochim Acta 49:32693281
might be mass transfer controlled leaving some Wichmann R, Vasic-Racki D (2005) Cofactor regeneration
residues in the cathode, unless very large currents at the lab scale. Adv Biochem Eng Biotechnol
92:225260
or long regeneration times are employed.
Further example of electrochemical regenera-
tion is regeneration of enzymatic cofactors in
electroenzymatic processes (Wichmann and
Vasic-Racki 2005). Redox enzymes are very Electrochemistry
selective and specic catalysts, which can enable
a number of partial oxidation or reduction reac- Tanja Vidakovic-Koch
tions for industrial applications at mild conditions. Max Planck Institute for Dynamics of Complex
Broader industrial application of redox enzymes Technical Systems, Magdeburg, Germany
has been so far hindered by their dependence on
expensive cofactors (e.g., nicotinamide adenine
dinucleotide (NAD)), which are consumed in the Electrochemistry is a branch of chemistry which
reaction (e.g., Eq. 1) and have to be regenerated studies charge transfer processes across an elec-
for a process to be economical: tried interface also called an electrochemical
double layer (Bockris and Reddy 1988, Hamann
CO2 NADH H HCOOH NAD (1) et al. 2007). Applications of electrochemistry are
broad including electrochemical processing, elec-
Electrochemical regeneration offers a possibility troanalysis electrochemical sensors, electrochem-
of cofactor regeneration. In this respect especially ical regeneration, and corrosion. In addition,
regeneration of NAD has been studied since many important processes in biological systems
NAD-dependent oxidoreductases are of great like photosynthesis and cell respiration are inher-
industrial interest. The electrochemical regenera- ently electrochemical processes.
tion can be represented by this reaction: The main feature of an electrochemical system
is a separation of ionic- and electronic ows. Ions
NAD H 2e NADH (2) are owing through the electrolyte which is exclu-
sively an ionic conductor, while electrons ow
This reaction is however not selective enough and through an outer electrical circuit which is exclu-
the kinetics is very sluggish on most known elec- sively an electron conductor. These two ows are
trode materials. Some improvements have been interconverted at the electrode/electrolyte inter-
achieved by using surface-modied electrodes. face across the electrochemical double layer by
Another strategy is to add an additional mediator means of an electrochemical reaction The poten-
according to tial difference in the electrochemical double layer
is related to thermodynamics (Nernst equation)
NAD Medred NADH Medox (3) and kinetics (Butler-Volmer or Tafel equations)
of an electrochemical reaction and it is a driving
Medox H 2e Medred (4) force for the electrochemical reaction to take
place. This unique feature of electrochemistry
Electrochemical cofactor regeneration is still not a makes easy to control the reaction rate by elec-
mature technology, and further improvements in trons at different energies.
electrode materials are needed to make this option Electrochemical processes can be spontaneous
feasible. (Gibbs free energy, DG <0), called galvanic, and
Electrodeionization 631
nonspontaneous DG>0, called electro- This technology is based on an electrodialysis

lytic. Instead in terms of Gibbs free energy, spon- unit with a diluate and also, in selected cases, a
taneity of an electrochemical process can be concentrate chamber lled with ion-exchanger
expressed in terms of cell voltage, where a posi- particles. They can be arranged as monopolar
tive value stands for a galvanic system and a beds (formed by particles of one polarity ion
negative for an electrolytic. The relationship exchanger), as layered beds (cation- and anion-
between the cell voltage and Gibbs energy is exchanger particles lled separately in several
given by equation DG = nFUr, where n stands alternating layers), or as a mixed bed (uniform
for number of exchanged electrons, F for a Fara- mixture of both types of ion-changer particles).
day constant, and Ur for an equilibrium cell The ion-exchange phase takes on the role of
E
voltage. electroconductive media, thus reducing ohmic
Many electrochemical systems require pres- drop in the dilute chamber. At the same time it
ence of separators. This is usually a membrane provides a three-dimensional interface for the
which can be a permeable, termed diaphragm, or removal of traces of ions present in the solution.
semipermeable, termed membrane. The latter type Two regions are typically distinguished in the
usually in addition to separation serves as an electrodeionization operation: (i) ions removal
electrolyte, so-called solid electrolyte in electro- and (ii) solvent splitting. Within the rst region
chemical systems. An example is ceramic yttria- the electrodeionization unit works bellow mass
stabilized zirconia (YSZ) membrane which has transfer limitation. It means, ux of ions to the
found an application in solid oxide fuel cells. solution ion exchanger interface driven by the
The ionic conductivity of this material is provided current load used has a value well below mass
by O2 ions. transfer limitation in a dilute chamber. The func-
tion of the ion-exchange bed consists in providing
a pathway for ions trapped in the dilute channel to
References the ion-selective membranes separating dilute/
concentrate chambers. In a second domain, how-
Bockris JOM, Reddy AKN (1988) Modern electrochem- ever, the current load exceeds limiting current
istry. Plenum Press, New York
density, i.e. limiting ux of ions present from
Hamann CH, Hamnett A, Vielstich W (2007) Electrochem-
istry, 2nd edn. Wiley-VCH Verlag GmBH, Weinheim solution to the solution ion exchanger interface.
In such a case sufcient number of ions to trans-
port corresponding electrical charge is provided
by decomposition (dissociation) of the solvent
(typically water). In contrast to electrodialysis,
Electrodeionization this splitting does not take place only at the
solution-membrane interface but also at the con-
Karel Bouzek tact of the cation- and anion-selective phase
Faculty of Chemical Technology, University of (Alvarado and Chen 2014).
Chemistry and Technology Prague, Technick 5, In the case of a concentrate chamber, the role of
Prague 6, Czech Republic the ion-exchange phase again consists in reducing
ohmic drop in the channel while maintaining the
concentration of the ions in the liquid phase at a
Electrodeionization represents a variant of elec- minimum to reduce back diffusion from the con-
trodialysis, modied in order to allow treatment of centrate to the dilute chamber.
low-salinity and low-conductivity media. This The quality of the stream produced is compa-
technique is typically applied to produce high- rable to that of the ion-exchange process. The
purity water suitable for use, for example, in ener- advantage of electrodeionization is that it is a
getics. It combines the advantages of ion continuous process that does not require a regular
exchange with those of electrodialysis. regeneration phase of operation. This feature of
632 Electrodialysis
electrodeionization has a further important between the electrodes is established, the cations
advantage. It saves a signicant amount of migrate toward the cathode. They pass through
corresponding chemicals and reduces the salinity the cation-exchange membrane, but they are
of the waste streams produced. Such technology is retained by the anion-exchange membrane. Like-
thus a suitable component for closed loop tech- wise the anions migrate toward the anode and pass
nologies which, on ending, discharge removed through the anion-exchange membrane but are
salts in the solid form and avoids production of retained by the cation-exchange membrane. The
contaminated liquid streams. overall result is that an electrolyte is concentrated
in alternate compartments, while its ion content is
depleted in the other compartments. In an indus-
References trial size electrodialysis stack, 100400 cell pairs
are arranged between the electrodes. Various
Alvarado L, Chen A, (2014) Electrodeionization: Princi- spacer and stack constructions such as the
ples, Strategies and Applications, Electrochim. Acta
so-called sheet ow are used in practical applica-
132:583
tions (Schaffer and Mintz 1966).
The concept of a sheet ow stack is illustrated
in Fig. 2 which shows the arrangement of the
spacers and membranes in a stack. The spacers
Electrodialysis not only separate the membranes and provide the
proper mixing of the solutions in the cells, but in
Heiner Strathmann their frames, they also contain the manifolds for
Universitt Stuttgart, Institute of Chemical the two different ow streams in the stack.
Process Engineering Stuttgart, Baden- Practical problems which affect the efciency
Wurttemberg, Germany of electrodialysis are concentration polarization
and membrane fouling. Concentration polariza-
tion which determines the so-called limiting cur-
The principle of electrodialysis is illustrated in rent density is controlled by hydrodynamic
Fig. 1 which shows a series of alternating anion- parameters of the feed ow in the stack. Mem-
and cation-exchange membranes arranged brane fouling is controlled by a technique which is
between two electrodes. The membranes are sep- referred to as clean in place or electrodialysis
arated by a spacer gasket and form individual reversal (Katz 1979).
cells, through which an electrolyte solution is The total costs in electrodialysis are the sum of
pumped. When an electrical potential difference the plant capital and the plant operating costs. The
Electrodialysis, Concentrate
illustrating the principle of Diluate
electrodialysis A C A C C
+ + + +
+ +
+
+ +
Anode + + Cathode
+ + + +
+ +
+ + +
+ +
+ + + + + + +
+ +
Electrode + + Electrode
rinse + + rinse
Feed
Repeating unit
Electrodialysis Reversal 633
Ion-exchange
Electrode membrane
cell
Concentrate
Diluate
Electrode
rinse solution
E
Feed solution
Feed solution
Electrode Diluate cell

Spacer Concentrate cell
Electrodialysis, Fig. 2 Schematic drawing illustrating the construction of a sheet ow stack design
capital related costs are proportional to the A major disadvantage especially for the produc-
required membrane area for a given capacity tion of potable water is the fact that only ions are
plant and a for a given feed and product removed, while uncharged components such as
concentration. The operating costs are propor- microorganisms or organic contaminants will not
tional to the required desalination energy which be eliminated. Another disadvantage of electrodi-
is the function of the transferred ions, i.e., the alysis is the relatively high energy consumption
concentration difference between the feed and when solutions with high salt concentrations have
product ion concentration. The energy EV to be processed. Thus, electrodialysis can only be
required to desalinate 1 m3 of water by electrodi- cost-effectively applied in water desalination in a
alysis is given by certain range of feed water salt concentration and
required product water (Strathmann 2010).
Fr A
CP i
f
EV C Cp
x References
Here rACP is the area resistance of the cell pair, Katz WE (1979) The electrodialysis reversal process.
which can be estimated from measurements of Schaffer LH, Mintz MS (1966) Electrodialysis. In:
membrane resistance and conductivities of solu- Spiegler KS (ed) Principles of desalination. Academic,
tions, F is the Faraday constant, i is the current New York, pp 320
density, Cf and Cp are the feed and product water Strathmann H (2010) Electrodialysis, a mature technology
with a multitude of new applications. Desalination
concentrations, and x is an efcient term which is 264:268288
generally close to 1.
Electrodialysis has advantages and limitations
compared to other deionization processes such as
reverse osmosis. A main advantage of electrodi-
alysis compared to reverse osmosis is that very Electrodialysis Reversal
little feed pretreatment is required and higher
brine concentrations can be achieved. Reverse Electrodialysis (RED)
634 Electrodialysis Reversal in High Supersaturation Mode
stack as illustrated in Fig. 1 which shows a typical

Electrodialysis Reversal in High repeating unit of an electrodialysis stack with
Supersaturation Mode bipolar membranes is composed of three cells,
two monopolar membranes and a bipolar mem-
Reverse Electrodialysis in High Supersatura- brane. The outer cells of the repeating unit are fed
tion Mode with a salt solution, the inner cells with water, or a
diluted acid and base. When an electrical potential
gradient is applied across a repeating unit, protons
and hydroxide ions which are generated in the
Electrodialysis with Bipolar bipolar membrane generate with the cations and
Membranes anions removed from the salt solution, an acid and
a base on either side of the bipolar membrane. The
Heiner Strathmann process design is closely related to that of the
Universitt Stuttgart, Institute of Chemical conventional electrodialysis using the sheet ow
Process Engineering Stuttgart, Baden- stack concept (Liu et al. 1977; Simons 1993).
Wurttemberg, Germany The utilization of electrodialysis with bipolar
membranes to produce acids and bases from the
corresponding salts is economically very attractive
The conventional electrodialysis can be combined and has a multitude of interesting potential appli-
with bipolar membranes and utilized to produce cations in the chemical industry as well as in bio-
acids and bases from the corresponding salts. technology and water treatment processes. Its key
A bipolar membrane is a laminate of an anion on component is the bipolar membrane. The bipolar
a cation-exchange layer. In this process membrane schematically illustrated in Fig. 2 con-
monopolar cation- and anion-exchange mem- sists of a laminate of an anion- and a cation-
branes are installed together with bipolar mem- exchange membrane with a 45 nm thick catalytic
branes in alternating series in an electrodialysis transition layer in between. In Fig. 2 this transition
Base Acid Base Acid

Repeating
cell unit
cm bpm am cm bpm am
+ + + +
+ + + +
+ + + +
M + + + + +

+
+ H + + + M + M+ + H+ +
+
+ +
+ + +
+ +
+ OH + +
+

+ X + X +
OH + + + +
+ + + +
+ + + +
+ + + X +
+ + + +
+ + + +
+ + + +
Salt solution Salt solution Salt solution
Electrodialysis with Bipolar Membranes, cell unit consisting of a cation-exchange membrane (cm), a
Fig. 1 Schematic drawing illustrating the principle of bipolar membrane (bpm), and an anion-exchange mem-
the electrodialytic production of an acid and a base from brane (am)
the corresponding salt with bipolar membranes. Repeating
Electrodialysis with Bipolar Membranes 635
Cation-exchange Anion-exchange required membrane area for a given capacity

layer layer plant. They are a function of the current density
applied in a given stack operation. A unit cell
contains a bipolar membrane, a cation- and an
H2O + H2O
Cathode +
Anode
anion-exchange membrane. The bipolar mem-
+
+ brane is rather expensive, and its useful life time
H+ +
OH
+ as well as that of the anion-exchange membrane is
+
+ rather limited in strong bases. The operating costs
Bipolar membrane
in electrodialysis with bipolar membranes are
strongly determined by the energy requirements
E
Electrodialysis with Bipolar Membranes, which are composed of the energy required for the
Fig. 2 Schematic drawing illustrating the electrodialytic water dissociation in the bipolar membrane and
water dissociation in a bipolar membrane with water dif-
fusing into the reaction region between the cation- and
the energy necessary to transfer the salt ions from
anion-exchange layers of the membrane and protons and the feed solution and protons and hydroxide ions
hydroxide migrate to the corresponding electrode from the transition region of the bipolar mem-
brane into the acid and base solutions. The energy
consumption due to the pumping of the solutions
through the stack can generally be neglected.
layer has been articially magnied. Water is dif- Since bipolar membranes became available as
fusing through both membrane layers into the tran- commercial products, a large number of applica-
sition layer where it gets electro-catalytically tions have been identied and studied on a labo-
dissociated into H+- and OH-ions, which ratory or pilot plant scale. However, in spite of the
migrate toward cathode and anode into the outer obvious technical and economical advantages of
solutions. the technology, large-scale industrial plants are
The energy required for the water dissociation still quite rare (Gineste et al. 1996). The main
can be calculated from the Nernst equation for a reasons for the reluctant use of bipolar membrane
concentration chain between solutions of different electrodialysis are poor membrane stability at
pH-values. It is given by: very high or low pH-values and insufcient
permselectivity at high ion concentrations, which
DG FD 2:3RTDpH results in a substantial product salt contamination,
low current efciency, and short membrane life.
Here DG is the Gibbs free energy and DpH and Nevertheless, there are a number of smaller-scale
D are the pH-value and the potential difference applications in the chemical process industry, in
between the two solutions separated by the bipolar biotechnology, in food processing, and in waste-
membrane. For 1 mol/L acid and base in the two water treatment.
phases separated by the bipolar membrane, DG is
0,022 kWh/mol and D is ca. 0,83 V at 25 C.
Compared to the ohmic potential drop over the
membranes, the required potential drop for water References
splitting in the transition layer is much more pro-
nounced. The determination of the costs for the Gineste JL, Pourecelly G, Lorrain Y, Presin F, Gavach C
(1996) Analysis of factors limiting the use of BPM: a
production of acids and bases from the simplied model to determine trends. J Membr Sci
corresponding salts follows the same general pro- 112:199208
cedure as applied for the cost calculation in elec- Liu KJ, Chlanda FP, Nagasubramanian KJ (1977) Use of
trodialysis desalination. The overall costs are the bipolar membranes for generation of acid and base: an
engineering and economic analysis. J Membr Sci
investment-related costs and the operating costs.
2:109124
The investment-related costs are dominated by the Simons R (1993) Preparation of high performance bipolar
membrane costs and are proportional to the membranes. J Membr Sci 78:1323
636 Electrodriven Membrane Processes
alternating anion and cation exchange membranes

Electrodriven Membrane Processes between an anode and a cathode. Within ion
exchange membranes, charged groups are
Seung-Hyeon Moon attached to the polymer backbone of the mem-
School of Environmental Science and brane material. These xed charged groups par-
Engineering, Gwangju Institute of Science and tially or completely exclude ions of the same
Technology (GIST), Gwangju, South Korea charge (co-ions) from the membrane, so that an
anion exchange membrane (AEM) with xed pos-
itively charged groups excludes positive ions but
Introduction is freely permeable to negatively charged ions,
referred to as counterions. Similarly, a cation
Electrolyte membranes or ion exchange mem- exchange membrane (CEM) with xed negatively
branes (IEMs) are ion-selective membranes for charged groups is freely permeable to positively
separation or charge transfer through the mem- charged ions. When an aqueous salt solution is
branes. Electrodriven membrane processes circulated in the cell under an electrical potential,
employ IEMs under an electric eld to perform the positively charged cations migrate toward the
(i) mass separation for concentration or dilution of cathode and the negatively charged anions toward
ions in aqueous phase; (ii) chemical synthesis the anode. The overall result is a potential drop
with an electrochemical reaction, such as across the cell pairs as well as a change in the ion
chlorine-alkaline electrolysis and production of concentration in alternate compartments. The
hydrogen via water electrolysis; and (iii) energy depleted solution is generally referred to as the
conversion and storage involving the conversion diluate and the concentrated solution as the con-
of chemicals into electrical energy and vice versa. centrate. Electrodialysis is used for desalination of
Non-charged porous membranes are also used in various salt solutions including seawater as well
some energy conversion processes to separate as production of table salt (Fig. 1).
anolyte and catholyte solutions. A bipolar membrane (BPM) in which the
anion- and cation-exchangeable layers
(AEL/CEL) are adjoined together easily splits
Separation Processes water molecules into protons and hydroxyl ions
under a reverse bias condition. Accordingly intro-
Applications of the rst type are mainly found in duction of inorganic substances into the bipolar
separation of ionic species under an electric eld. membrane is known to be an effective technique
Cations move toward a cathode and permeate for generation of acid and base. Two or three
through a cation exchange membrane while compartment stack congurations may be chosen
anions move toward an anode and permeate an depending on the product specication, i.e.,
anion exchange membrane. Due to rejection of CEM-BPM, AEM-BPM, or CEM-BPM-AEM. It
non-permeable ion by the membranes, electro- is expected that water-splitting electrodialysis
lytes are concentrated or diluted in a compart- using bipolar membrane can make zero-emission
ment. Typical processes in this category are possible and be applied as a clean technology by
electrodialysis, bipolar membrane electrodiaysis, recycling salt by-products after splitting to acid
electrodeionization, and membrane capacitive and base. Developments of high-performance
deionization. Without an electric eld, diffusion bipolar membranes further expand use of bipolar
dialysis enables to purify acid or base solutions membrane electrodialysis in chemical, biochemi-
containing metallic contaminants. cal, and environmental industries (Fig. 2).
Electrodialysis (ED) is an electrochemical Electrodeionization (EDI) or continuous
separation process using an electrical potential as electrodeionization (CEDI) is a hybrid separation
a driving force. Electrodialysis systems typically process that removes ionized species from liquids
consist of a cell arrangement with a series of using electrically active media and an electrical
Electrodriven Membrane Processes 637
Electrodriven Membrane
Processes, Fig. 1 An
Electrodialysis Cell
Processes, Fig. 2 A
Bipolar Membrane
Electrodialysis Cell
potential activating ion transport. The electrically and the concentration of dissolved ionic solutes at
active media, such as an ion exchange resin in a relatively low salt concentration (Fig. 3).
CEDI devices, may function to alternately collect Membrane capacitive deionization (MCDI)
and discharge ionized species or to facilitate the is an advanced capacitive deionization (CDI) with
transport of ions continuously by ionic or elec- the help of ion exchange membranes. A CDI sys-
tronic substitution mechanisms. Unlike an ion tem is operated by adsorption and desorption of
exchange process, CEDI does not require ions on carbon electrodes. When an electric poten-
chemicals to regenerate the ion exchange resin tial is applied to CDI cells, charged ions in con-
or concentrate the wastewater. In a CEDI system, taminant water are adsorbed onto the surface of
the ion exchange resin bed plays a major role in charged electrodes and forms an electric double
the reduction of the high electrical resistance in layer due to the charged electrode and adsorbed
the diluate compartment, while the ion exchange ions, producing puried water. Often CDI leads to
membranes lead to depletion and concentration of a higher energy consumption and a lower opera-
the solutions in the diluate and concentrate com- tion efciency due to mobility of unwanted ions.
partments, respectively. The CEDI process is To avoid this phenomenon, ion exchange mem-
capable of achieving high levels of purication branes are introduced to CDI for ion selectivity,
Processes, Fig. 3 An
Electrodeionization Cell
which is called membrane CDI (MCDI). A MCDI compartment is enhanced. The acid feed enters the
employs two types of ion exchange membranes, bottom of every alternate compartment, as water
i.e., cation exchange and anion exchange mem- is fed from the upper part of the stack to contact
branes. The AEM and CEM are positioned in the acid feed counter currently. Permeate acids are
front of the positively and negatively charged collected from the bottom of the permeate com-
electrodes, respectively. While counterions are partment as puried acid product. The depleted
attracted onto the electrode surface, simulta- feed ows out of the upper part as an efuent
neously co-ions expelled from the counter elec- containing high concentrations of metal salts and
trode. The selectivity of ion exchange membranes a low concentration of remaining acid. Similarly a
prevents reverse adsorption and prohibits the cation exchange membrane is used to recover base
mobility of unwanted ions. from metal salts.
Diffusion dialysis is a membrane separation

process in which transport of selective ions is
driven by a concentration difference over an ion Chemical Synthesis Processes
exchange membrane. The diffusion dialyzer for
acid recovery consists of a multitude of compart- Applications of the second types are electrolysis
ments made of gaskets arranged alternately with processes to generate chemicals by electrode reac-
anion exchange membranes, which allow acids to tions and ion substitution through membranes.
permeate but retain metal salts. The gasket sur- A chlor-alkali process involves electrode
rounds a meshlike spacer by which mixing in a reactions for the electrolysis of NaCl solution.
Electrodriven Membrane Processes, Fig. 4 A Membrane Capacitive Deionization Cell
The process produces caustic soda (NaOH) and NaA + HX ! HA + NaX. A four-compartment
chlorine (Cl2) that are used in a wide range of double decomposition cell consists of a series of
industries. The most common chlor-alkali process anion and cation exchange membranes arranged in
is the electrolysis of aqueous sodium chloride in a an alternating pattern between anode and cathode.
membrane cell. The membrane is a cation perme- When the cell is operated with four different
able one which allows positive ions to pass streams, ion substitution or salt metathesis reaction
through it. That is, the only the sodium ions occurs, AX + BY ! AY + BX (Fig. 4).
from the NaCl solution may pass through the
membrane and not the chloride ions. The advan-
tage of the membrane process is that the sodium Energy Conversion Processes
hydroxide solution is formed in a cathode com-
partment which is not contaminated with NaCl Applications of the third types are found in vari-
solution. Also puried NaCl should be supplied ous energy conversion and storage systems where
to obtain pure NaOH solution. Chlorine is gener- charge carriers are selectively permeate through a
ated at anode by oxidation of chloride ion. As a membrane. Charge carriers are generated from
clean energy, hydrogen had been drawing more electrode reactions or initially supplied as
attention. One of hydrogen sources is electrolysis supporting electrolytes. When charge carriers
of water, splitting water to hydrogen and oxygen. move between anolyte and catholyte without
Various catalysts and energy sources are investi- mixing of active materials, a separator type mem-
gated to improve the efciency as a reverse pro- brane may be used. This type of applications
cess of a fuel cell. includes fuel cell, lithium ion battery, redox ow
Electro-ion substitution reaction can be car- battery, reverse electrodialysis, etc.
ried out in an ED stack in which membranes are all Polymer electrolyte fuel cells (PEFCs) are
of the same kind, either anion or cation exchange commonly considered to be one of the most likely
membranes. In a unit cell consisting of two com- potential alternative power sources due to their
partments with CEMs, acid and feed streams ow. convenient operating temperature, high-energy
When an electric current is supplied to the stack, conversion efciency, and ease of integration
protons in the acid stream (HX) are transported to into hybrid systems. Electricity is generated by
the feed stream (NaA) and convert the negatively potential difference between anode where hydro-
charged organics (A) to free acid (HA). As a result gen or fuel is oxidized to proton with the help of
the inorganic cations in the feed stream, in this case catalyst and cathode where oxygen is reduced and
sodium ions, are transported to the acid compart- reacts with proton. A polymer electrolyte mem-
ment to meet the electroneutrality (NaX) in the feed brane is placed between the anode and cathode
stream, completing an ion exchange reaction, catalysts to insulate between electrodes. Polymer
Processes, Fig. 5 A
Diffusion Dialysis Cell
electrolyte membrane (PEM) sandwiched with used for energy storages and electric vehicles. In
two catalyst layers is called a membrane-electrode these applications, safety is a key parameter. The
assembly (MEA). Peruorinated sulfonic acid safety of battery is closely linked to the battery
membranes such as Naon , Flemion , separator absorbing a highly ammable liquid
Aciplex , and Dow series have been widely electrolyte. Although typical separators based on
used due to their high proton conductivity and semicrystalline polyolen materials such as poly-
good thermal and chemical stability. ethylene (PE) and polypropylene (PP) are
Hydrocarbon-based membranes, however, are mechanically strong and relatively inexpensive,
generally less stable and subject to fast degrada- their low melting temperature possibly causes
tion, yet to be improved for practical applications. physical contact between two electrodes over
Recent efforts have focused on the commerciali- 180 C. The physical contact triggers thermal
zation of PEFCs in a range of applications, includ- runaway, which results in explosive burning of
ing use in portable devices, residences, stationary the highly ammable liquid electrolyte (Fig. 6).
applications, and vehicles (Fig. 5). A redox ow battery is an energy storage
A lithium ion battery is a rechargeable battery device in which potential difference of a redox
in which lithium ions move between anode and couple generates (discharge) or stores (charge)
cathode with redox reactions during charge and electric energy. Positive and negative electrolytes
discharge cycles. A separator is required to pre- are stored separately and the electrolyte solutions
vent mixing of anolyte and catholyte while lith- are fed into a cell for energy conversion. A variety
ium ions permeate freely in electrolyte solution. of membranes are used for RFB systems
Often the separator is prepared as a multilayer depending on the charge carrier, electrolyte sol-
porous membrane due to requirement of shut- vent, and redox couple. Membranes must possess
down and meltdown. The lithium ion battery is two main properties: preventing the undesirable
rapidly growing due to the widespread use of cross-mixing of the negative and positive electro-
wireless electronics such as cell phones and laptop lytes to minimize self-discharge, while selectively
computers. Moreover, concerns about global conducting the charge carrier ions through the
warming and oil shortages have led to the devel- membrane to complete the redox circuit during
opment of high energy density Li ion batteries the passage of current. Typically monovalent
Processes, Fig. 6 Electro-
Ion Substitution Cell
Processes, Fig. 7 A Salt
Methathesis Cell
selective cation exchange membrane or a thin researchers. Furthermore, nonaqueous organic

anion exchange membrane is used to prevent electrolytes provide a wider potential window of
crossover of the cationic redox couples when pro- operation due to the high solubility of various
ton is a charge carrier. A RFB is a promising metalligand complexes in an organic solvent.
technology due to several advantages such as its Nonaqueous RFB systems require ion exchange
large capacity, convenient operating temperature, membranes or non-charged porous membranes
long cycle life, and so on. On the other hand, the depending on the electrochemical properties
use of an aqueous electrolyte severely restricts the and molecular size of charge carriers employed
energy density as is determined by the cell (Fig. 7).
potential that is constrained by the electrochemi- Reverse electrodialysis (RED) is an ion
cal stability window of water. Along with com- exchange membrane process for renewable
mercialization of aqueous RFBs, nonaqueous energy generation. Electric potential occurs as a
RFBs which can offer a higher density have result of Donnan equilibrium where two solutions
drawn considerable interest from energy storage containing permeable ions are separated by a
642 Electroendosmosis
Processes, Fig. 8 A
Polymer Electrolyte
Membrane Fuel Cell
corresponding ion exchange membrane. Using a i.e., ionic conductivity, transport number, oxida-
similar cell structure to electrodialysis, electricity tive stability, and power density.
can be generated by diffusion of ions using con-
centrated salt solution and diluted solution, called
reverse electrodialysis (RED). Typically seawater
and river water may be used as feed solution in a Electroendosmosis
modied electrodialysis cell. Other salinity
gradient power generation is pressure-retarded Electroosmosis, Overview of
osmosis using pressure difference due to the
concentration difference across an osmotic mem-
brane including a polymer electrolyte active layer
(Fig. 8).
Electro-Fenton Process
Marc Cretin1,2 and Mehmet A. Oturan3

1
Perspectives Institut Europen des Membranes,
ENSCM-UM-CNRS (UMR 5635), Universit de
Electrodialysis has been studied and industrial- Montpellier, Montpellier Cedex 5, France
2
ized for nearly a century as a leading electrodriven National School of Chemistry of Montpellier,
membrane process. Along with the appearance of University of Montpellier, Montpellier, France
3
advanced ion exchange membranes, various sep- Laboratory of Earth Materials and Environment,
aration and reaction processes are following elec- University Paris-Est, Marne La valle,
trodialysis in many industries as demands for Champs-sur-Marne, France
compact and environmentally friendly processes
are growing continuously. Last several decades
energy-related electrodriven membrane processes The electro-Fenton process is one of the most
are growing rapidly with the help of membranes popular electrochemical advanced oxidation pro-
as a key component. Many membranes originally cesses. As all the advanced oxidation process
developed for separation processes have (AOPs), electro-Fenton (EF) process is based on
expanded their applications for energy conversion the in situ production and reactions of the hydroxyl
and storage. Accordingly the membranes should radicals (OH), the second powerful oxidant (after
be characterized in terms of electrochemical prop- uorine) with E0 (OH/OH) = 2.80 V/SHE and a
erties to improve the performance appropriately, very reactive species (Brillas et al. 2009; Oturan
Electrofiltration 643
and Aaron 2014). OH is able to mineralize any their mineralization, i.e., transformation to CO2,
organic pollutants and therefore used in AOPs in H2O, and inorganic ions.
order for destruction of persistent organic This EF process has been shown to be more
pollutants. efcient and cost-effective than some widely
Contrary to anodic oxidation process (Panizza AOPs such as Fenton oxidation and ozonation
and Cerisola 2009) in which OHs are produced at for the treatment of organic pollutants species
anode surface from water oxidation (Panizza and (Brillas et al. 2009; Oturan and Aaron 2014). It
Cerisola 2009), in the EF process, OHs are gen- was successfully applied to the removal of toxic
erated in the bulk of solution from electrochemi- and persistent organic pollutants from water such
cally assisted Fentons reactions: as synthetic dyes, chlorinated aromatics, chloro-/
E
nitrophenol herbicides, insecticides, fungicides
and other pesticides, and pharmaceutical and per-
H2 O ! OH H e
sonal care products. The results of these studies
have shown that the EF process using a large
H2 O2 Fe2 H ! Fe3 OH H2 O surface area carbon-felt cathode is a very promis-
ing technology due to its simplicity, low cost, and
In the classical Fenton process, the Fentons outstanding performance, which is mainly due to
reagent (H2O2 + Fe2+) is added externally to the the quick and efcient cathodic regeneration of
reaction medium. This leads to the formation of the Fe2+ catalyst.
Fe(OH)3 sludge and promotes wasting reactions
that decrease process efciency. EF process has a
great advantage to generate in situ H2O2 (from References
two-electron reduction of O2) and regenerate
ferrous iron ions (from initially added catalyst) Brillas E, Sirs I, Oturan MA (2009) Electro-Fenton process
and related electrochemical technologies based on
as schematized on the following electrochemical Fentons reaction chemistry. Chem Rev 109:65706631
reactions: Dirany A, Sirs I, Oturan N, Oturan MA (2010) Electro-
chemical abatement of the antibiotic sulfamethoxazole
from water. Chemosphere 81:594602
O2 2H 2 e ! H2 O2 Oturan MA, Aaron JJ (2014) Advanced oxidation pro-
cesses in water/wastewater treatment: principles and
applications. A review. Crit Rev Environ Sci Tech.
Fe3 e Fe2 doi:10.1080/10643389.2013.829765
Panizza M, Cerisola G (2009) Direct and mediated anodic
oxidation of organic pollutants. Chem Rev 109:65416569
There is no accumulation of Fe3+ and other
reagents (H2O2 and Fe2+), so sludge formation
occurs and parasitic reactions are minimized to
enhance mineralization (transformation of
organics to CO2 and water) efciency. Electrofiltration
In the case of using a high O2 evolution anode
such as boron-doped diamond (BDD) lm Karel Bouzek
anode, OH are formed simultaneously at the Faculty of Chemical Technology, University of
anode as well as in the bulk solution from the Chemistry and Technology Prague, Technick 5,
electrogenerated Fentons reagent. The oxidation Prague 6, Czech Republic
power and mineralization efciency are extremely
high (Dirany et al. 2010).
The OH thus formed by the process in a con- Electroltration represents a modication of
tinuous and catalytic way will react with the dead-end membrane micro- or ultraltration. It
organic pollutants present in the medium until targets a signicant reduction of the ltration
644 Electrolyzers
Electrofiltration,
Fig. 1 Schematic sketch of
electroltration principle,
Fw stands for driving force
due to the friction between
the particle and owing
solvent molecules and Fe
stands for driving force
resulting from action of the
applied electric eld E on
the particle carrying
electrical charge
time and focuses especially on the ltration and/or time they are electrically charged. The mild con-
concentration of colloidal substances that other- ditions of electroltration thus enable their prop-
wise rapidly build up a deposit of colloidal parti- erties to be preserved during the process of their
cles on the surface of the membrane, which separation (Gzke and Posten 2010).
strongly hinders permeation of the uid phase.
The basic principle is that colloidal particles
usually carry an electric charge. Hence, by apply- References
ing an appropriate electric eld, colloidal particles
can be moved in the direction opposite to the uid Henry jr. JD, Lawler LF, Kuo CHA (1977) A solid/liquid
separation process based on cross ow and
ow, thus keeping the surface of the ltrating electroltration. AIChE Journal 23:851
membrane free of the deposit (Henry et al. Gzke G, Posten C (2010) Electroltration of Biopoly-
1977). A schematic sketch of this arrangement is mers. Food Eng. Rev. 2:131
shown in the Fig. 1.
As the ltrating membrane remains unimpeded
by a deposit of colloidal particles, the pressure
drop across it remains relatively low. In contrast, Electrolyzers
the separator covered by the layer of colloidal
particles attracted by the electric eld does not Antonino Salvatore Arico
permit a signicant uid ow. This results in a CNR-ITAE Institute, Messina, Italy
reduction of the shear stress forces in the depos-
ited lm of separated colloid. These facts make
electroltration especially promising for the sep- One of the main processes occurring in an
aration of biotechnology-derived products, the electrolyzer device is the water electrolysis. Elec-
reason being that such products are typically sen- trolysis of water is the dissociation of water mol-
sitive to high shear stress forces while at the same ecules into hydrogen and oxygen gases. For this
Electrolyzers 645
process, in the presence of liquid water at 298 K Anode Cathode

and 1 bar, DG is 237 kJ mol1 (corresponding to H2O 2H+ + 2e- + 0.5O2 2H+ + 2e- H2
~1.23 V), DS is 163 J mol1 K1 Erev = 1.23 V vs. RHE Erev = 0.00 V vs. RHE
(TDS ~0.25 V), whereas DH is 286 kJ mol1. Metal
Pt/C
The thermoneutral potential at which this reaction Oxides
occurs in the absence of external heat supply is O2 H2
Eth,DH = 1.48 V (upper heating value
3.54 kWhNm3 H2) (Millet et al. 2011). If
Solid Polymer
Electrolyte C
steam is fed to the device, the reaction enthalpy A A
H+
is reduced by ~40 kJ mol1 corresponding to the N T
O H+
H E
vaporization enthalpy. Water electrolysis is tradi- D H+
O
tionally carried out in alkaline media with several E D
commercial electrolyzers available on the market. E
Water electrolyzers using a solid polymer electro-
Electron Flow
lyte are less common and generally use expensive H2O H2O
materials such as noble metal electrocatalysts and
H2O H2 + 0.5O2 E rev = 1.23 V
Naon membranes (Barbir 2005; Siracusano
et al. 2010). Polymer electrolyte membrane Electrolyzers, Fig. 1 Principle of operation of a PEM
(PEM) electrolyzers represent a viable alternative water electrolysis cell
to alkaline electrolyzer using KOH or NaOH as
electrolytes for hydrogen generation. The advan-
tages of SPE water electrolyzers especially con-
cern with increased safety, high energy density,
pressure operation that may reach 50100 bars.
and low maintenance. In the PEM water
Thus, a proper thickness is necessary for the poly-
electrolyzer, water is usually supplied to the
mer electrolyte separator (around 100 mm). For
anodic compartment where oxygen evolution
high-pressure operation in PEM electrolyzers,
occurs, whereas hydrogen is produced at the cath-
reinforced PFSA membranes provide a proper
ode by protons transported through the protonic
combination of good conductivity and high
membrane (Fig. 1). The electrodes are usually
mechanical strength.
composed of a platinum electrocatalyst for hydro-
gen evolution, whereas metal oxides (e.g., IrO2,
RuO2, etc.) are used for the anode due to their
enhanced activity and stability than Pt for this
References
reaction (Marshall et al. 2007; Siracusano
Barbir F (2005) PEM electrolysis for production of
et al. 2010). The performance of an SPE hydrogen from renewable energy sources. Sol Energy
electrolyzer is strongly related to the characteris- 78:661
tics of the membrane and electrode assembly Marshall A, Brresen B, Hagen G, Tsypkin M, Tunold R
(2007) Hydrogen production by advanced proton
(MEA) where the electrochemical reactions take
exchange mebrane (PEM) water electrolysers
place at triple-phase boundary. Therefore, the Reduced energy consumption by improved
interface between solid polymer electrolyte and electrocatalysis. Energy 32:431
electrocatalyst layers should be characterized by a Millet P, Mbemba N, Grigoriev SA, Fateev VN,
Aukauloo A, Etivant C (2011) Electrochemical per-
suitable extension; furthermore, the contact resis- formances of PEM water electrolysis cells and perspec-
tance between the catalytic layer and the mem- tives. Int J Hydrog Energy 36:4134
brane should be as low as possible. Generally, Siracusano S, Baglio V, Di Blasi A, Briguglio N, Stassi A,
Naon membrane is used as conducting polymer Ornelas R, Trifoni E, Antonucci V, Arico AS
(2010) Electrochemical characterization of single cell
electrolyte in PEM electrolyzer systems. How-
and short stack PEM electrolyzers based on a nanosized
ever, low levels of H2 and O2 crossover are nec- IrO2 anode electrocatalyst. Int J Hydrog Energy
essary for PEMWE application due to the high- 35:5558
646 Electromembrane
In a cation-exchange membrane, the xed nega-

Electromembrane tive charges are in electrical equilibrium with
mobile cations (counterions) in the interstices of
Grarld Pourcelly the polymer as shown in Fig. 1 (Strathmann
Institut Europeen des Membranes, CC 047, 2010). In this case, the mobile anions are referred
Universite Montpellier II, Place Eugene Bataillon, to as coions. They are more or less excluded from
Montpellier, France the polymer matrix because of their electrical
charge which is identical to that of the xed ions
(Donnan exclusion (Donnan 1911)). Thus, the
Electromembrane or charged membrane stands selectivity of an IEM results from the exclusion
for ion-exchange membrane [IEM]. They are used of coions from the membrane phase. The proper-
in a number of processes which are rather different ties of IEM are determined by different parame-
in their basic concept, their practical applications, ters such as the density of the polymer network,
and their technical relevance (Strathmann 2004). the hydrophobic/hydrophilic character of the
All IEM separation processes are based on the polymer matrix, the nature and the ratio of xed
same fundamental principle which is the coupling ion-exchange groups, the cross-linking ratio, etc.
of the transport of electrical charges, i.e., an elec- The most desired properties of IEM are (i) high
trical current with a transport of mass, i.e., cations chemical and thermal stabilities, (ii) high mechan-
or anions, through a permselective membrane due ical and dimension stabilities, (iii) high
to an externally applied or internally generated permselectivity, (iv) low electrical resistance,
potential gradient. There are two types of IEM: (v) and a low cost.
(i) monopolar and (ii) bipolar membranes. Bipolar membranes (BPMs) are composed of
Monopolar membranes are either cation- two layers of ion exchangers joined by a hydro-
exchange membranes which contain negatively philic junction (Pourcelly et al. 2009). The diffu-
charged groups xed to the polymer matrix or sion of water from both sides of the BPM allows
anion-exchange membranes which contain posi- its dissociation under the electrical eld to gener-
tively charged groups xed to the polymer matrix. ate protons and hydroxyl ions, which further
Electromembrane, a
Fig. 1 (a) Cation-
exchange membrane with a Counter-ion pathway
homogeneous structure; (b)
ion-exchange membrane
with a heterogeneous
structure prepared from an
ion-exchange resin powder
in a binder polymer (From Counter-ion Co-ion Fixed ion
Strathmann 2010)
Polymer matrix
b
Counter-ion pathway
Solution filled gaps

Ion-exchange resin Binder polymer
Electromembrane Processes 647
Electromembrane, Fig. 2 Principle of a bipolar membrane. Left hand: water dissociation under electrical eld. Right
hand: the two ion-exchange layers bearing xed anion- or cation- exchange groups
migrate from the junction layer through the

cation- and anion-exchange layers of the BPM as Electromembrane Processes
depicted in Fig. 2. The requirements for suitability
of BPM include that for monopolar membranes Gerald Pourcelly
but also an experimental potential to achieve the Institut Europeen des Membranes, CC 047,
water-splitting capability as close as possible as Universite Montpellier II, Place Eugene Bataillon,
the theoretical value equal to 0.83 V at 25 C. Montpellier, France
Nowadays, superior styrene-divinylbenzene
copolymer membranes can be easily purchased,
peruorinated membranes with great chemical Electromembrane processes involve ion-exchange
stability are on the market, and BPM with an membranes. They can be divided into three
industrial-scale lifetime (>20,000 h) is available. types: (i) Separation of components such as
salts or acids and bases from electrolyte solu-
tions. In this case, the driving force for the ion
transport across the membrane is an electrical
References potential as in electrodialysis or a concentration
gradient as in both diffusion dialysis and Donnan
Donnan FG (1911) Theory of membrane equilibrium and dialysis. (ii) The second type of processes
membrane potential in the presence of non-dialysing
electrolyte. Z Electrokem Angew Phys Chem
involves an electrochemical reaction producing
17:572581 chemicals such as chlorine, acids and bases, or
Pourcelly G, Bazinet L (2009) Developments of organic and inorganic compounds. The most
BPM technology in food and bio-industries. In: known process of this type is the chlorine-alkali
Pabby AK, Rizvi SSH, Sastre AM (eds) Handbook of
membrane separations, CRC Press, Boca Raton,
production where the ion-exchange membrane is
pp 581634 the key component. (iii) The third type of
Strathmann H (ed) (2004) In: Ion exchange membrane ion-exchange membrane processes involves the
separation processes. Membrane technologies series, conversion of chemical into electrical energy and
Elsevier, Amsterdam
Strathmann H (2010) Electrodialysis: a mature technology
vice versa. The solid polymer electrolyte fuel cell
with a multitude of new applications. Desalination is the most signicant application (Strathmann
264:268288 2004).
648 Electromembrane Processes
CONCENTRATE (OUTLET)
DILUATE (OUTLET)
ELECTROLYTE
Na+ Na
+
Na
+
Na
+ +
Na Na
+
+ + + + +
Na Na Na Na Na
ANODE CATHODE
Cr Cr Cr Cr Cr
Cr Cr Cr Cr Cr Cr
D C D C D C D C D C D
DILUATE (INLET)
CONCENTRATE (INLET)
Electromembrane Processes, Fig. 1 Scheme of a conventional two-compartment electrodialysis stack (C): concen-
trate, (D): diluate compartments
(i): In dialysis, the driving force for the com- and used to produce acids and bases from their
ponent transport is a difference of concentration corresponding salts, Fig. 2. This process is eco-
between the two solutions separated by a mem- nomically very attractive and has several interest-
brane. In diffusion dialysis and Donnan dialysis ing potential applications mainly in food industry,
the driving force is also a concentration gradient ne chemicals, or biotechnologies (Pourcelly
but the membrane is carrying anions or cations et al. 2009).
according to its permselectivity. Diffusion dialysis (ii): The most known membrane electrolysis
with anion exchange membranes is used on a large processes are for the production of chlorine and
scale to recover acids from pickling solutions caustic soda (Pletcher and Walsh 1990) or special
(Kobuchi et al. 1987). Donnan dialysis is used organic compounds (Sata 1991). In the chlorine-
for water softening or for recovering organic alkali process, the cells are arranged in a
acids from their salts (Kliber and Wisnieskwi similar way as in an electrodialysis stack but in
2011). two different congurations using monopolar or
Electrodialysis (ED) is used to transport salt bipolar electrodes. Due to the severe operating
ions from one solution through ion-exchange conditions (80 C, 35 wt% NaOH, wet chlorine),
membranes to another solution under the inu- the membrane is a composite structure of
ence of an applied electric potential difference. sulfonated and carboxylated peruorinated
The elementary cell consists of a feed (diluate) reinforced polymers.
compartment and a concentrate (brine) compart- (iii): Ion-exchange membranes are used today
ment formed by an anion exchange membrane also as key components in energy storage and
and a cation exchange membrane placed between conversion systems such as batteries and fuel
two electrodes. In almost all practical ED pro- cells. A fuel cell is an electrochemical reactor in
cesses, multiple ED cells (few hundreds) are which energy is converted into electrical energy.
arranged into a conguration called an ED stack Generally fuel cells are fed with hydrogen which
Fig. 1. is transformed into protons at a catalytic anode
Conventional ED is used mainly for desalina- with electrons. Protons then migrate across a pro-
tion of saline solutions (chemicals) or milk whey ton conducting membrane and combine with oxy-
(food industry). It can be combined with bipolar gen to produce water at the catalytic cathode
membranes (bipolar membrane electrodialysis) (Lucia 2014).
Electroosmosis 649
Electromembrane Processes, Fig. 2 Electrodialysis with bipolar membrane. Scheme of a three-compartmentcell
References
Electroosmosis
Kliber S, Wisnieskwi JA (2011) Removal of bromate and
associated anions from water by Donnan dialysis with Karel Bouzek
anion exchange membranes. Desalin Water Treat Faculty of Chemical Technology, University of
35:158163
Kobuchi Y, Motomura H, Noma Y, Hanada F (1987) Chemistry and Technology Prague, Technick 5,
Application of ion exchange membranes to recover Prague 6, Czech Republic
acids by diffusion dialysis. J Membr Sci 27:173179
Lucia U (2014) Overview on fuel cells. Renew Sustain
Energy Rev 30:164169
Pletcher D, Walsh FC (1990) Industrial electrochemistry. Electroosmosis is the term used for the convective
Chapman & Hall, London ow of a solution containing charge carriers, e.g.,
Pourcelly G, Bazinet L (2009) Developments of BPM in the form of dissociated salts, through the
technology in food and bio-industries. In: Pabby AK,
pores or structure of a material with an
Rizvi SSH, Sastre AM (eds) Handbook of membrane
separations. CRC Press, Boca Raton, pp 581634 electrically charged internal surface. In the eld
Sata T (1991) Ion exchange membranes and separation of membrane processes, this phenomenon is typ-
processes with chemical reactions. J Appl Electrochem ically discussed in connection with the utilization
21:283294
of ion-selective membranes under current load.
Strathmann H (2004) In: Ion exchange membrane separa-
tion processes. Membrane technologies series. However, it can potentially occur in any mem-
Elsevier, Amsterdam brane material. The schematic sketch of the prin-
ciple of electroosmotic ux initiation in an
ion-selective membrane pore (void space) is
shown in Fig. 1.
Electronic and Ionic Conductivity The driving force of electroosmotic ux is the
electric eld. Flux velocity is directly proportional
Mixed Conducting Membranes to the hydraulic permeability of a membrane and
650 Electroosmosis, Overview of
Cross-References
Electroosmosis, Overview of
References
Fila V, Bouzek K (2008) The effect of convection in the

external diffusion layer on the results of a mathematical
model of multiple ion transport across an ion-selective
membrane. J Appl Electrochem. 38:1241
Squires TM, Bazant MZ (2004) Induced-charge electro-
osmosis. J Fluid Mech 509:217
Electroosmosis, Fig. 1 The schematic sketch of a prin-

ciple of electroosmotic ux initiation in the membrane with
an electric charge carriers xed on the pore (void fraction)
wall
Dianne Wiley1 and Gustavo Fimbres Weihs2

1
indirectly proportional to the viscosity of the pore School of Chemical and Biomolecular
uid. The latter parameter represents an excess of Engineering, The University of Sydney, Sydney,
free charge carriers of one sign in the bulk of the NSW, Australia
pore uid. An excess of one-sign charge carriers is 2
Ctedras CONACYT, Instituto Tecnolgico de
possible on account of an intrinsic property of an Sonora, Cd. Obregn, Sonora, Mexico
ion-selective membrane, i.e., the presence of
charge-carrying functional groups in the mem-
brane structure. These ion groups are bonded by Synonyms
a covalent bond to the backbone of the membrane
material. They are thus xed and immobilized in Electroosmosis; Electroendosmosis; Electro-
the structure. In the case of conventional mem- osmosis; Electroosmotic ow; EOF
brane materials, the role of xed charge can be
carried out by ions specically adsorbed on the Electro-osmosis is the movement of liquid in
pore walls or along the backbone of membrane response to an applied electric eld across a con-
materials. The charge carried by functional groups duit such as a membrane, capillary tube,
and/or adsorbed in the internal structure of the microchannel, or porous material.
membrane has to be compensated by an excess The application of an external electric eld
charge of opposite sign present in the pore uid. near a solid/liquid interface may result in the
The electroosmotic convective ow of the pore motion of liquid with respect to an adjacent
uid is induced by friction force between the charged surface. These associated effects are
solvated mobile charge carriers (ions) moving known as electrokinetic phenomena (Hunter
under the action of the electric eld and the sur- 1981), rst observed by Reuss in 1808 (Reuss
rounding liquid phase molecules (Squires and 1809). Electrokinetic phenomena include
Bazant 2004). electro-osmosis, electroacoustics, electrophoresis,
Electroosmotic ux represents one of the diffusiophoresis, streaming potential, zeta poten-
mechanisms of mass transfer through an tial, and sedimentation potential. The earliest
ion-selective membrane. Its signicance increases experiments on electrokinetic phenomena focused
with increasing current load and, thus, the inten- on electro-osmosis because these are the easiest to
sity of the electric eld applied across the mem- conduct (Wall 2010). Early investigators included
brane (Fila and Bouzek 2008). Wollaston, Porrett, Napier, Faraday, and Daniell.
Electroosmosis, Overview of 651
Electroosmosis,
Overview of, Membrane
Fig. 1 Schematic
representation of electro- - - - - - - - - -
osmosis in a membrane EDL + + + + + + + +
pore, showing the effect of
the electric eld on the
electric double layer (EDL) + Fluid
and the resulting drag on the
bulk uid Electric Field
EDL + + + + + + + +
- - - - - - - - - E
Membrane
In electro-osmosis, the bulk uid moves relative electric potential (f) and the permittivity of the
to a charged surface due to an external electric uid (e) by Poissons equation:
eld.
The chemical equilibrium between a solid sur- e2 f re (1)
face (such as a membrane) and a uid containing
charged species typically leads to the interface Under negligible convective or electrophoretic-
acquiring a net xed electrical charge (most induced ow, the charge density follows a
often negative for membranes), characterized by Boltzmann distribution (Hunter 1981), and the
the zeta potential (z). The presence of co-ions charge density can be obtained by solving the
(similarly charged ions) and counterions nonlinear PoissonBoltzmann equation:
(oppositely charged ions) results in attraction
and repulsion of ions in the vicinity of a charged 2 f k2 sinhf 0 (2)
surface. This phenomenon, coupled with the ran-
dom thermal motion of the ions, creates an electric where the constant k depends on the ionic com-
double layer (EDL) or Debye layer in the region position (Russel et al. 1989). Although solutions
near the interface. Electroneutrality is not to Eq. 2 can only be obtained numerically for
maintained within the EDL, because there is an complex geometries, it can be linearized using
excess of counterions compared to co-ions in the DebyeHuckel approximation, which leads
order to neutralize the surface charge (Probstein to the following electric potential distribution:
2005). When an electric eld is applied to the

uid, the net charge in the electrical double layer y
f zexp (3)
is induced to move by the resulting Coulomb lD
force. The movement of ions drags bound solvent,
causing bulk uid motion due to momentum where y is the distance from the surface, and
transfer, as depicted in Fig. 1. The resulting ow lD = 1/k is the Debye length.
is termed electro-osmotic ow. Electro-osmosis In membrane systems, electro-osmosis can be
can only occur if there are charged species in the approximated through the HelmholtzSmo-
uid that can respond to the electric eld luchowski (HS) velocity equation:
(Jagannadh and Muralidhara 1996).
Electro-osmosis can be described mathemati- ezEx
us (4)
cally using the momentum transport equation, as m
the resulting body force due to an electric eld on
a charged uid (Hu and Li 2007). The net charge The HS approximation described by Eq. 4 is an
density of the uid (re) can be related to the articial slip velocity implemented on a charged
652 Electroosmosis, Overview of
surface. It can be used to simulate the effect of Cross-References

electro-osmosis on the velocity in a conduit at the
outer edge of the EDL (Probstein 2005). This Backwashing
equation implies that the electro-osmotic ow Electrical Double Layer
velocity is proportional to the applied electric Electroosmotic Drag in Membranes
eld. Other equations that are relevant to electro- Membrane Fouling
osmosis under different conditions include the Surface Charge Density
HuckelOnsager equation and the Henry Zeta Potential
equation.
Most common applications of electro-osmosis
are in capillary electrophoresis, capillary References
electrochromatography, and microuidic devices
in elds as diverse as biophysics, dewatering, Bowen WR (1993) Electrochemical aspects of
microltration and ultraltration. In: Howell JA,
geomechanics, medicine, microchips, oil and gas
Sanchez V, Field RW (eds) Membranes in
production, and separations including in mem- bioprocessing: theory and applications. Springer, Dor-
brane separation. There are a number of important drecht, pp 265291
applications in membrane technology. At the fun- Bowen WR, Ahmad AL (1997) Pulsed electrophoretic
lter-cake release in dead-end membrane processes.
damental level, electro-osmosis can be used to
AIChE J 43(4):959970
determine the zeta potential of membranes Bowen WR, Clark RA (1984) Electro-osmosis at micropo-
(Bowen and Clark 1984). It can also be benecial rous membranes and the determination of zeta-
for the operation of membrane separation sys- potential. J Colloid Interface Sci 97(2):401409
Bowen WR, Sabuni HAM (1994) Electroosmotic membrane
tems, and has found numerous applications in
backwashing. Ind Eng Chem Res 33(5):12451249
microltration and ultraltration processes Bowen WR, Kingdon RS, Sabuni HAM (1989) Electrically
(Bowen 1993), in particular to do with the miti- enhanced separation processes: the basis of in situ
gation of membrane fouling. This is because sur- intermittent electrolytic membrane cleaning (iiemc)
and in situ electrolytic membrane restoration (iemr).
face activity and membrane charge, two
J Membr Sci 40(2):219229
phenomena closely related to membrane fouling, Hu G, Li D (2007) Multiscale phenomena in microuidics
cannot be inuenced through hydrodynamic and nanouidics. Chem Eng Sci 62(13):34433454
forces such as increased shear. Electro-osmosis Hunter RJ (1981) Zeta potential in colloid science: princi-
ples and applications. Academic, London
can be used for membrane cleaning and restora-
Jagannadh SN, Muralidhara HS (1996) Electrokinetics
tion (Bowen et al. 1989), for backwashing methods to control membrane fouling. Ind Eng Chem
(Bowen and Sabuni 1994), for releasing lter Res 35(4):11331140
cakes in dead-end membrane processes (Bowen Liang YY, Fimbres Weihs GA, Wiley DE (2014) Approx-
imation for modelling electro-osmotic mixing in the
and Ahmad 1997), and for enhancing ux (Sarkar
boundary layer of membrane systems. J Membr Sci
and De 2010). 450:1827
Although the use of DC electric elds in mem- Probstein RF (2005) Physicochemical hydrodynamics: an
brane separation processes has been in practice for introduction, 2nd edn. Wiley, New York
Reuss FF (1809) Sur un novel effet de llectricit
several decades, anode corrosion is a signicant
galvanique. Mmoires de la Socit Impriale des
problem that has discouraged its wider application Naturalistes de Moskou 2:327337
(Jagannadh and Muralidhara 1996). Nevertheless, Russel WB, Saville DA, Schowalter WR (1989) Colloidal
the development and availability of corrosion- dispersions. Cambridge University Press, Cambridge,
UK
resistant materials has enabled recent develop-
Sarkar B, De S (2010) Electric eld enhanced gel con-
ments and commercialization of these techniques. trolled cross-ow ultraltration under turbulent ow
Electro-osmosis has also found applications in conditions. Sep Purif Technol 74(1):7382
RO processes, leading to improved recovery Spiegler KS, Macleish JH (1981) Molecular (osmotic and
electro-osmotic) backwash of cellulose acetate
through membrane backwashing (Spiegler and
hyperltration membranes. J Membr Sci 8(2):173192
Macleish 1981) or by increasing the level of Wall S (2010) The history of electrokinetic phenomena.
mixing in the boundary layer (Liang et al. 2014). Curr Opin Colloid Interface Sci 15:119124
Electroosmotic Drag in Membranes 653
due to momentum transfer (viscous drag). This

Electroosmotic Drag Flux bulk uid motion due to an external electric eld
is termed electroosmosis (De Groot 1966) and can
Electroosmotic Drag in Membranes occur in any uid with a net charge, including in
membrane channels or within membrane pores
(Fig. 1).
Electroosmotic drag (EOD) is the underlying
Electroosmotic Drag in Membranes mechanism responsible for solvent transport in a
number of membrane techniques including:
Dianne Wiley1 and Gustavo Fimbres Weihs2
E
1
School of Chemical and Biomolecular EOD in electrodialysis (ED), often called
Engineering, The University of Sydney, Sydney, water transport, is one of the reasons for
NSW, Australia losing current density (Indusekhar and
2
Ctedras CONACYT, Instituto Tecnolgico de Krishnaswamy 1965). ED has many industrial
Sonora, Cd. Obregn, Sonora, Mexico applications including the production of chlo-
rine, NaOH, and other inorganic acids and
bases (de Groot et al. 2011).
Synonyms EOD of water in proton exchange membranes
(PEMs, also referred to as polymer electrolyte
Electroosmotic drag ux; Electroosmotic water membranes), which leads to water manage-
transport; EOD ment issues (Zawodzinski et al. 1995).
EOD of methanol in direct methanol fuel cells
Electroosmotic drag in membranes refers to the (DMFCs), leading to methanol crossover
movement of water or other electroneutral sol- (Heinzel and Barragn 1999).
vents through a membrane, associated with the In vanadium redox batteries (VRBs), EOD can
movement of ions under the inuence of an elec- cause vanadium crossover and capacity loss
tric eld. (Agar et al. 2013).
The zeta potential at the interface between a EOD in electroosmotic pumps (EOPs) is used
membrane and an electrolyte generally leads to to generate ow or pressure (Wang et al. 2006).
the creation of an electric double layer (EDL). EOPs can be used to aid with the hydration of
When an electric eld is applied to uid within PEMs for fuel cells.
an EDL, the net charge is induced to move by the EOD is a mechanism for producing
resulting Coulomb force. The movement of ions anomalous ux through charged membranes
drags bound solvent, causing bulk uid motion (e.g., NF membranes) due to the selective
Electroosmotic Drag
in Membranes, Membrane
Fig. 1 Schematic
representation of
electroosmotic drag in a
membrane pore. The
Drag
electric eld causes the + Electric Field
dissolved ions to experience +
a net force, and these ions Drag
transfer momentum to the
solvent molecules through
viscous drag
Membrane
654 Electroosmotic Flow
passage of one type of charged ion over the References

other, leading to the axial rearrangement of the
EDLs (Sasidhar and Ruckenstein 1982). Agar E, Knehr KW, Chen D, Hickner MA, Kumbur EC
(2013) Species transport mechanisms governing capac-
ity loss in vanadium ow batteries: comparing Naon
EOD is of particular importance in PEMs such and sulfonated Radel membranes. Electrochim Acta
as Naon and sulfonated polyetherketones, used 98:6674
in fuel cells and VRBs. Water management criti- De Groot SR (1966) Thermodynamics of irreversible pro-
cesses, 4th edn. North-Holland, Amsterdam
cally impacts the performance of PEMs for fuel
de Groot MT, de Rooij RM, Bos AACM, Bargeman G
cells, as their proton conductivity increases with (2011) Bipolar membrane electrodialysis for the alka-
hydration. Too little water can dry the membrane, linization of ethanolamine salts. J Membr Sci
leading to a reduction in conductivity and conse- 378(12):415424
Heinzel A, Barragn VM (1999) A review of the state-of-
quently lower cell performance. Too much water
the-art of the methanol crossover in direct methanol
can ood the membrane and lead to cathode fuel cells. J Power Sources 84(1):7074
ooding problems. In addition to the generation Indusekhar VK, Krishnaswamy N (1965) Diffusion effect
of water at the cathode catalyst layer due to the during electrodialysis with ion-exchange membranes.
J Appl Polym Sci 9(7):26312632
electrochemical reaction, water is also transported
Ise M, Kreuer KD, Maier J (1999) Electroosmotic drag in
through the membrane (Lee et al. 2008). The polymer electrolyte membranes: an electrophoretic
permeate ux of water in a PEM is governed by NMR study. Solid State Ion 125(14):213223
two main mechanisms: (a) the osmotic ux due to Lee PH, Han SS, Hwang SS (2008) Three-dimensional
transport modeling for Proton Exchange Membrane
a chemical potential gradient across the mem-
(PEM) fuel cell with micro parallel ow eld. Sensors
brane, which is inuenced by both the activity of (Basel) 8(3):14751487
the solution (osmotic pressure) and the hydro- Ren X, Gottesfeld S (2001) Electro-osmotic drag of water
static pressure, and (b) the electroosmotic ux in poly(peruorosulfonic acid) membranes.
J Electrochem Soc 148(1):A87A93
due to the drag induced by the proton ux (Ren
Sasidhar V, Ruckenstein E (1982) Anomalous
and Gottesfeld 2001). effects during electrolyte osmosis across charged
In order to optimize water concentration in a porous membranes. J Colloid Interface Sci
fuel cell membrane, it is important to quantify 85(2):332362
Wang P, Chen Z, Chang H-C (2006) A new electro-osmotic
electroosmotic drag. The electroosmotic water
pump based on silica monoliths. Sens Actuators
ux can be related to the proton ux through B 113(1):500509
the membrane by means of the electroosmotic Zawodzinski TA, Davey J, Valerio J, Gottesfeld S (1995)
drag coefcient (psdrag), which is dened as The water content dependence of electro-osmotic
drag in proton-conducting polymer electrolytes.
the number of water molecules transported
Electrochim Acta 40(3):297302
through the membrane per proton as the chemical
potential gradient of water tends to zero (Ise
et al. 1999).
Electroosmotic Flow
Cross-References
Direct Methanol Fuel Cell (DMFC)
Electrical Double Layer
Electrodialysis
Polymer Electrolyte Membrane Fuel Cell
(PEMFC) Electroosmotic Water Transport
Proton-Exchange Membranes for Fuel Cells
Zeta Potential Electroosmotic Drag in Membranes
Electrophoresis 655
The main limitation of electrooxidation is the

Electrooxidation decrease of the electrocatalytic activity of the
electrode because of the fast formation of a poly-
Marc Cretin mer layer from oxidation intermediates on the
Institut Europen des Membranes, ENSCM-UM- anode surface. Film formation depends on the
CNRS (UMR 5635), Universit de Montpellier, adsorption properties of the electrode surface,
Montpellier Cedex 5, France the concentration and the nature of organic com-
National School of Chemistry of Montpellier, pounds, their degradation intermediates, the pres-
University of Montpellier, Montpellier, France ence of oxygen, etc. Then electrodes presenting a
surface with weak adsorption properties are pro-
E
moted to gain in stability (Panizza and Cerisola
Electrooxidation is a direct anodic oxidation tech- 2003). Nevertheless, in anaerobic conditions, the
nique to remove biorefractory pollutants by direct formation of the polymer fouling layer remains
electrolysis in which pollutants exchange elec- the limiting factor of the mineralization.
trons with the anode surface without intermediate To overcome this poisoning effect, anode must
substances. It differs from indirect anodic oxida- be polarized over oxygen evolution to promote
tion. Indeed by indirect anodic oxidation, pollut- water discharge, leading to hydroxyl radical
ants are oxidized by OHs radicals generated by (OH) formation on specic electrode like boron-
water oxidation on a high O2 evolution anode doped diamond. The electrochemical formation of
such as boron-doped diamond (BDD). this powerful oxidant avoids then polymer fouling
Direct anodic oxidation is possible using noble layer formation and then the decrease of the
metals (Pt and Pd), metal oxide (IrO2, PbO2), or electrocatalytic activity of the electrode. More
carbon-based (boron-doped diamond) anodes, details can be found concerning this electrochem-
before oxygen evolution. The rst stage is the ical advanced oxidation process (EAOP) in the
pollutant adsorption on the anode surface. The entry indirect anodic oxidation.
second stage is the oxygen (as a source of elec-
trons) transfer from water to the organic pollutant
References
using electric energy (i.e., anodic polarization)
through the so-called electrochemical oxygen Panizza M, Cerisola G (2003) Inuence of anode material
transfer reaction (EOTR). A typical example of on the electrochemical oxidation of 2-naphthol part 1.
EOTR often cited in the literature (Panizza and Cyclic voltammetry and potential step experiments.
Cerisola 2009) to illustrate direct electrooxidation Electrochim Acta 48:34913497
Panizza M, Cerisola G (2009) Direct and mediated anodic
is anodic incineration of phenol as follows: oxidation of organic pollutants. Chem Rev
109:65416569
C6 H5 OH 11H2 O ! 6CO2 28H 28e
The complete phenol oxidation to CO2 (i.e., min-

eralization) is obtained at the anode, and protons Electrophoresis
liberated in this reaction are reduced at the cathode
to hydrogen. The pollutant mineralization rate Catherine Charcosset
depends mainly on the nature of the anodic Universit Lyon 1, Lyon, Villeurbanne, France
material, the choice of the anodic potential
(xed before oxygen evolution), and chemical
and physical parameters such as electrolyte and The term electrophoresis refers to the motion of
pollutant composition and concentration and suspended particles in an applied electric eld.
temperature. Among separation techniques, electrophoresis is
656 Electrophoresis
widely used in research and development and (SDS-PAGE), which separates proteins according
quality control in disciplines such as biochemis- to their molecular size. The method involves plac-
try, immunology, genetics, and molecular biology ing the sample in gel with a pH gradient and
(Westermeier 2001). Electrophoresis is based on applying a potential difference across it.
the differential migration of charged species in a Cellulose acetate membranes are other current
semiconductive medium under the inuence of an supporting media for electrophoresis separation
electric eld. Separation of many different kinds (Westermeier 2001). They are used for routine clin-
of species including proteins, DNA, nucleotides, ical analysis and related applications, as well as for
drugs, and many other biochemicals is obtained the analysis of molecules in physiological uids.
upon differences in size, charge, and hydropho- These membranes have large pores and therefore
bicity. The technique was rst reported in 1937 by have a low sieving effect on molecules. The elec-
Arne Tiselius who won the Nobel Prize in Chem- trophoretic separation is thus entirely based on
istry in 1948 for the separation of different serum charge density. The matrix exerts little effect on
proteins by a method called moving-boundary diffusion so that the separated zones are relatively
electrophoresis. Since then, a number of wide and the resolution and limit of detection area
improved techniques have been introduced such is low. For these reasons, cellulose acetate mem-
as gel electrophoresis, capillary electrophoresis, branes are often replaced by gel electrophoresis.
and two-dimensional electrophoresis. Other supporting membranes for electrophore-
Gel electrophoresis uses an electric current sis include Naon membranes, a type of peruor-
passed through an agarose or polyacrylamide gel osulfonic acid membrane, and cation-exchange
(SDS-PAGE) to separate the molecules in a sample membranes, which are chemically resistant and
on the basis of their differences in molecular size consist of a pore-structure cluster network (Fang
and charge. As the sample migrates in the gel in et al. 2004). These membranes are widely used in
response to the electric current, the smaller species the eld of chloralkali industry and in fuel cells.
move more quickly than the larger species, which A Naon membrane contains hydrophilic pores
results in a distinct banded pattern in the gel. This (1020 and 5060 in size) acting as very
banded pattern may be visualized via the applica- narrow electrophoresis channels. The xed-
tion of staining agents, such as ethidium bromide, charge sites (SO3) on the hydrophilic pore
which reveals the gel bands under UV light, or surface provide a strong charged background.
silver stain, which is typically used to detect pro- Naon membrane electrophoresis is a potentially
teins. The silver stain is compatible with mass attractive technique for the separation of small
spectrometry techniques for further analysis of the organic molecules like amino acids or ions.
protein composition. Capillary electrophoresis
(CE) involves a combination of both polyacryl-
amide gel electrophoresis (SDS-PAGE) and high- Cross-References
performance liquid chromatography (HPLC)
(Ahuja and Jimidar 2008). High voltages of Ion-Exchange Membranes
500 V/cm or greater are generated within narrow
capillaries (20200 mm). The high voltages cause
electroosmotic and electrophoretic movement of References
buffer solutions and ions, respectively, within the
Ahuja S, Jimidar M (2008) Capillary electrophoresis
capillary. Two-dimensional gel electrophoresis methods for pharmaceutical analysis. Academic,
(2-D electrophoresis) separates species in two Amsterdam
steps, according to two independent properties. In Fang C, Wu B, Zhou X (2004) Naon membrane electro-
a common technique, the rst dimension is isoelec- phoresis with direct and simplied end-column pulse
electrochemical detection of amino acids. Electropho-
tric focusing, which separates proteins according to resis 25:375380
their isoelectric points; the second dimension is Westermeier R (2001) Electrophoresis in practice, 3rd edn.
SDS polyacrylamide gel electrophoresis Wiley-VCH, Weinheim
Electrophoretic Deposition 657
suspension, including metals, polymers, ceramics,

Electrophoretic Deposition and glasses.
The fundamental mechanisms of EPD are
Catherine Charcosset described in the literature mainly in the framework
Universit Lyon 1, Lyon, Villeurbanne, France of the DerjaguinLandauVerweyOverbeek
(DLVO) theory and in relation to the distortion of
the particle double layer under the application of a
Electrophoretic deposition (EPD), also called DC electric eld (Corni et al. 2008). However,
electrocoating, e-coating, cathodic electrodeposi- numerous other theories (occulation by particle
tion, or electrophoretic coating, is a simple and accumulation, particle charge neutralization, elec-
E
effective technique for coating of charged parti- trochemical particle coagulation, electrical double
cles on substrates (Besra and Liu 2007). It has layer distortion, and thinning mechanism) have
several advantages including continuous been proposed to explain the particle interactions
processing, uniform deposition and control of and the kinetics of deposition.
the thickness, and morphology of a deposited Several examples of membrane preparation
lm by adjustment of the deposition time and using an EPD technique are reported. For exam-
applied potential. In EPD, charged powder parti- ple, EPD is used as a seeding method for making
cles, dispersed or suspended in a liquid medium, zeolite membranes (Abdollahi et al. 2007). Zeo-
are attracted and deposited onto a conductive sub- lites are crystalline structures which possess uni-
strate of opposite charge on application of a DC form and molecular-sized pores. Zeolite
electric eld. There are two types of electropho- membranes have a great potential in separation
retic deposition (Fig. 1). When the particles are and catalysis processes owing to their unique
positively charged, the deposition happens on the pore structures and adsorption properties and
negative electrode (cathode) and the process is their high thermal, mechanical, and chemical sta-
termed cathodic electrophoretic deposition. The bility compared with polymeric membranes.
deposition of negatively charged particles on pos- Since zeolites are negatively charged particles,
itive electrode (anode) is called anodic electro- they can be effectively attracted to the substrates
phoretic deposition. By suitable modication of of positive charge. By the aid of EPD, an oriented
the surface charge on the particles, any of the two continuous layer of nano-sized zeolite seeds is
modes of deposition is possible. This technique is formed on the support and acts as nuclei for the
convenient for stable suspensions containing next step which is crystal growth under hydrother-
charged particles free to move when an electric mal situation.
eld is applied. Therefore, EPD can be applied to The EPD technique is also used for the prepa-
any material that is available as a ne powder ration of membrane electrode assembly for fuel
(e.g., <30 mm particle size) or as a colloidal cell (Morikawa et al. 2004). A membrane
Electrophoretic a b
Deposition,
Fig. 1 Schematic of
electrophoretic deposition +
process. (a) Cathodic EPD +
and (b) anodic EPD +

+ +

+

+ +
+ +
658 Electrophoretic Mobility
electrode assembly can be prepared by EPD pro- factors such as the strength of electric eld or
cess onto a Naon membrane. A suspension voltage gradient, the dielectric constant of the
consisting of ethanol, carbon powders with Pt medium, the viscosity of the medium, and the
catalyst, and Naon polymer is used to obtain a zeta potential of the particles.
stable dispersed solution. The thickness of the The electrophoretic mobility, m (m2/Vs), is the
prepared catalyst layer is controlled by EPD dura- observed electrophoretic velocity, v (m/s), divided
tion or concentration of suspension. The cell by electric eld strength, E (V/m):
obtained may present a higher electrochemical
performance compared with ordinary cells. v
m
E
Cross-References The electrophoretic velocity is the distance of

migration divided by time, also called velocity
Electrophoretic Painting of migration. Mobilities are sometimes expressed
with a negative sign, because migration of the
Fuel Cell Membrane
Zeolite Membrane solutes or particles generally occurs in the direc-
tion opposite to the electrophoretic eld (which is
taken as reference).
References The electrophoretic mobility of charged sol-
utes, including proteins, is predicted using the
Abdollahi M, Ashrazadeh SN, Malekpour A (2007) Prep- Debye-Huckel-Henry theory (OConnor
aration of zeolite ZSM-5 membrane by electrophoretic
deposition method. Microporous Mesoporous Mater et al. 1996). This theory is valid only for spherical
106:192200 nonconducting particles at low zeta potentials,
Besra L, Liu M (2007) A review on fundamentals and with ions present in the electrical double layer
applications of electrophoretic deposition (EPD). Prog behaving as point charges. The Debye-Huckel-
Mater Sci 52:161
Corni I, Ryan MP, Boccaccini AR (2008) Electrophoretic Henry theory gives the mobility as
deposition: from traditional ceramics to nanotechnol-
ogy. J Eur Ceram Soc 28:13531367 ze f kr
Morikawa H, Tsuihiji N, Mitsui T, Kanamura K (2004) m
6pr 1 kr
Preparation of membrane electrode assembly for fuel
cell by using electrophoretic deposition process.
J Electrochem Soc 151:A1733A1737 where z is the net charge (dimensionless), e the
elementary charge (C), r the particle radius (m),
the viscosity (kg m1 s1), and f(kr) is the
Henry correction factor, which has values
Electrophoretic Mobility between 1 and 1.5. The Debye-Huckel parameter,
k, is dened by
Catherine Charcosset
Universit Lyon 1, Lyon, Villeurbanne, France 2e2 N A I
k2
ekT
When an electric eld is applied across a given where NA is the Avogadros constant (kmol1), I is
medium, charged solutes or particles suspended in the ionic strength (kmol m3), e is the permittivity
the electrolyte are attracted toward the electrode (J1 C2 m1), k is the Boltzmanns constant
of opposite charge. Viscous forces acting on the (J K1), and is T the temperature. OBrien and
component tend to oppose this movement. When White (1978) obtained a more rigorous solution to
equilibrium is reached between these two oppos- the equations describing the electrophoretic
ing forces, the solutes or particles move with mobility of a rigid spherical particle with a uni-
constant velocity. The velocity is dependent on form charge density. The predictions of this model
Electrophoretic Painting 659
coincide with those of the Debye-Huckel-Henry coat many industrial products such as automobile
theory for systems with zeta potentials less than bodies and parts, tractors and heavy equipment,
25 mV. metal furniture, and beverage containers. The pro-
The electrophoretic mobility of proteins has cess is carried out with the use of anodes and
been determined through several microltration cathodes in the painting tank. There are two
membranes under conditions of zero concentra- types of electrophoretic painting: anaphoretic
tion driving force and with negligible adsorption and cataphoretic painting. In anaphoretic painting
on to the membranes occurring during the exper- the proles will be the anode, and in cataphoretic
iment (OConnor et al. 1996). The initial mobility painting the proles will be the cathode. In the
of both proteins is close to free-solution values at electrophoretic painting process, evolution of
E
the same pH, after correction for the ionic strength oxygen gas will take place at the anode, and
of the buffer. This effect is explained by the effect hydrogen gas will be evolved at the cathode.
of reduced free area in the membrane which is The electrophoretic paint is continuously cir-
compensated by a corresponding increase in the culated to avoid settling of the paint solids and
potential gradient. Data were also obtained for heat resulting from the pumping process as well as
mixtures of two proteins. Interactions occur from the passage of electric current. Dry coating
when the two proteins are oppositely charged, thicknesses of the order of 20 mm are normally
changing the apparent mobilities from the free applied, using a voltage of 150200 V for a time
solution values. of 12 min. After painting, the proles are rinsed
before stowing at temperatures of 180200 C for
2030 min. The electrophoretic paint may contain
Cross-References a water-soluble acrylic-based paint, the resin
being rendered soluble by the addition of suitable
Electrical Potential amines such as melamine. When current is passed
Electrophoresis through the paint, the resin and pigments will
migrate to the anode while the amine will be
discharged at the cathode. During paint deposition
References
the level of amine will increase, and this causes a
OBrien RW, White LR (1978) Electrophoretic mobility of decrease in paint deposition rate if allowed to
a spherical colloidal particle. J Chem Soc Faraday continue. The most usual processes used for
Trans 1:16071626 removal of excess amine are ultraltration often
OConnor AJ, Pratt HRC, Stevens GW (1996) Electropho- in combination with reverse osmosis.
retic mobilities of proteins and protein mixtures in
porous membranes. Chem Eng Sci 51:34593477 After electrophoretic painting, ultraltration
continuously treats the paint bath to produce per-
meate needed for rinsing metal parts (Agana
et al. 2011). While permeate is used for rinsing,
the mixture of paint and rinse water is returned to
Electrophoretic Painting the paint bath from the downstream rinse system.
The volume in the paint bath remains constant,
Catherine Charcosset and the process leads to a closed loop cycle
Universit Lyon 1, Lyon, Villeurbanne, France including a multistage rinse system. No wastewa-
ter is produced and almost no deionized water is
required for the purpose of rinsing. Different types
Electrophoretic painting, also called electropho- of membranes are available for various types of
retic deposition of paint, electrodeposition paint- paint. The type and size of the membrane modules
ing, or E-coating, is an economical and corrosion- are related to the needs of bath volume and char-
resistant technique for applying coatings to elec- acteristic parameters of the paint bath, such as
trically conducting materials. It is widely used to total solids, pH, and temperature.
660 Electrospun Nanofiber Membrane for Biosensors
Cross-References Within this scenario, electrospinning is an eco-

nomic and versatile technology for producing
Electrophoretic Deposition one-dimensional nanomaterials that possess high
Permeate specic surface area and high porosity. In this
technology, an electried jet of polymer solution
is obtained by applying a relatively high electric
References
tension. If the viscoelastic forces on the polymer
Agana BA, Reeve D, Orbell JD (2011) Optimization of the solution are sufcient to contrast jet breaking, a
operational parameters for a 50 nm ZrO2 ceramic mem- nonwoven web of bers ranging from few nano-
brane as applied to the ultraltration of post- meters to several microns is collected on a proper
electrodeposition rinse wastewater. Desalination support. Several reviews and books about
278:325332
electrospinning (Bhardwaj and Kundu 2010;
Ramakrishna et al. 2005; Wendorff et al. 2012)
are available for having a better understanding of
this technique. The majority of electrospun bers
Electrospun Nanofiber Membrane are polymer-based or polymer-inorganic compos-
for Biosensors ites. It has been shown they have a good biocom-
patibility and consequently they represent a good
Michele Modesti, Carlo Boaretti and candidate for biosensor fabrication especially as a
Martina Roso matrix for protein or enzyme immobilization
Department of Industrial Engineering, University (Li et al. 2014).
of Padova, Padova, Italy Among biosensors, monitoring changes in
extracellular glucose concentration is one of the
most important ones because it is related to both
A biosensor is an analytical device comprising brain disease and neurological disorders, and it is
three main elements: a biological recognition ele- also the key analyte for medical diagnostics and
ment able to interact specically with a target, a management of diabetes. A complete and exhaus-
transducer able to convert a change in property of tive review on glucose biosensor and all the basic
the solution or surface into a recordable signal, principles of them can be found in the scientic
and a processing system which is able to amplify literature (Yoo and Lee 2010). The most common
and quantify the signal. The involved biomolecu- glucose biosensors achieve specic recognition of
lar interactions in the bio-receptor/analyte glucose by immobilization of the enzyme glucose
complex can be based on antibody-antigen inter- oxidase (GOD) on the surface of electrodes Wu
actions, enzymatic interactions, nucleic acid inter- and Fan (2013). The basic concept of the glucose
actions, and cellular or analyte/synthetic receptor biosensor is based on the fact that the immobilized
interactions. GOD catalyzes the oxidation of b-D-glucose by
Biosensors are generally classied according molecular oxygen, producing gluconic acid and
to transduction modes and recognition elements hydrogen peroxide. Three general strategies are
(DOrazio 2003; Rodriguez-Mozaz et al. 2004) in used for the electrochemical sensing of glucose:
ve principal transducer classes that are electro- by measuring oxygen consumption, by measuring
chemical, optical, thermometric, piezoelectric, the amount of hydrogen peroxide produced by the
and magnetic. Moreover, electrochemical sensors enzyme reaction, or by using a diffusible or
may be subdivided into potentiometric, ampero- immobilized mediator to transfer the electrons
metric, or conductometric types (Thevenot from the GOD to the electrode (Yoo and Lee
et al. 2001). Biosensors represent nowadays the 2010). At this regard, polyvinyl alcohol (PVA)
answer to a pressing need of advanced solutions nanobers have been shown to be useful in pro-
for sensitive detection of the physiological ducing an amperometric glucose biosensor (Ren
chemicals involved in clinical diagnosis. et al. 2006) which exhibited a rapid response time
Electrospun Nanofiber Membrane for Biosensors 661
(1 s) and a higher output current (mA level) to chemical stability and good electrical
glucose in the normal and diabetic level. The conductivity. Poly(L-lactide) (PLLA) has been
linear concentration response range from 1 to electrospun on Pt microelectrodes followed by
10 mM and the lower detection limit (0.05 mM) the electrochemical deposition of PEDOT-GOD
of the sensor can meet the demand in the detection on the Pt microelectrodes and around PLLA
of the glucose in medical diagnosis. nanobers. Comparing the performance of the
In enzyme-based biosensors, the sensitivity two different nanostructured sensors (PEDOT-F
and longevity of the biorecognition element are vs. PEDOT-NFs), it has been shown that the
functions of the physical design and the enzyme increase of surface area related to the nanobers
stability over time (Wilson and Gifford 2005). affects both the impedance of the biosensor
E
Moreover, temperature, pH, and other chemicals (which is sensitively lower) (Abidian et al. 2010)
can easily affect the activity and stability of the and it allows more GOD to be entrapped on the
immobilized enzyme as well as the immobiliza- biosensor. Both of these results contribute to the
tion chemistry and the microenvironment of the increased sensitivity of the PEDOT-NFs-GOD
enzyme on the electrode (Marx et al. 2011). biosensors.
In order to get an efcient enzyme encapsula- Nevertheless, the scientic research on biosen-
tion and the absence of interference from other sors is moving also toward the design and devel-
coexisting electroactive species, the biocomposite opment of nonenzymatic glucose sensors based
based on Prussian blue, chitosan, and polyvinyl on direct oxidation of glucose at modied elec-
alcohol has been fabricated by electrodeposition trode surface (Kong et al. 2012). According to
and subsequent electrospinning in enzyme- Ding et al. (2010), metals (Au, Pt, Ni, Cu) and
friendly conditions. Prussian blue (PB), known metal oxides (ZnO, CuO, etc.) have been
as articial peroxidase, has been employed in exploited to construct a variety of enzyme-free
fabricating amperometric glucose biosensors glucose sensors. Among the aforementioned
because of its well-known capabilities for enhanc- metal oxides, cobalt oxide nanobers (Ding
ing electron transport and excellent catalytic et al. 2010) have been obtained by electrospinning
activity toward H2O2 reduction at a low and subsequent calcination, and they have been
overpotential. The modied electrode has been used for glucose detection in an alkaline medium.
shown to exhibit such a rapid direct electron trans- The developed sensor has been shown to have a
fer rate that the glucose sensor exhibited excellent fast response time (less than 7 s), a high sensitivity
performance, which includes wider linearity, of 36.25 mA mM1cm2, good reproducibility
lower detection limit, and good stability in opti- and selectivity, and a detection limit of 0.97 mM
mized conditions (Wu and Yin 2013). (S/N = 3). The high sensitivity and low detection
Another way for improving sensitivity and limit has been hypothesized to be related to the
lifetime of the immobilized enzyme is the appli- excellent catalytic property of the as-prepared
cation of conducting polymers to bioelectronic Co3O4 nanobers and to their highly porous 3D
surfaces; the advantage of using such electrospun network. The latter provides high specic surface
polymers is related to their ability to increase the area and numerous active sites, and it allows the
signal-to-noise ratio (S/N), and, at the same time, access of analytes to the active catalytic sites with
they are a suitable matrix for the immobilization minimal diffusion resistance. Another attempt to
and entrapment of enzymes (Yang et al. 2014). get nonenzymatic glucose sensor has been
For instance, Yang et al. compared the reported in the research of Cao et al. (2011),
performance of enzyme-entrapped conducting wherein nickel oxide microbers (NiO-MFs)
polymer nanobers (NFs) with reference to its have been directly immobilized onto the surface
conducting polymer lm (F) counterpart. Among of uorine tin oxide (FTO) electrode by
different conducting polymers, poly electrospinning followed by calcination
(3,4-ethylenedioxythiophene) (PEDOT) has been without using any immobilization matrix. The
chosen as support for GOD enzyme because of its amperometric performance of glucose at
662 Electrospun Nanofiber Membrane for Biosensors
NiO-MFs-modied electrode have shown an activity, thanks to an effective reduction of the

ultrasensitive current (1785.41 mA mM1 cm2) diffusion resistance of the substrates/products.
response and low detection limit of 3.3 108M Further challenge in detecting DNA repair rep-
(S/N = 3). resents a key factor in cancer diagnostic. Due to
Despite the presence of enzyme-based and the subtlety of DNA damage, it is difcult to sense
nonenzyme-based biosensors, there are also other the presence of damaged repair with high selec-
support-enzyme hybrid substrate materials for bio- tivity and sensitivity. In this regard immobilizing
sensors. An example that has been investigated by DNA on gold/polymeric electrospun nanobers
researchers was based on water-stable PVA and and their use in electrochemical sensing of DNA
PVA/poly(ethyleneimine) (PEI) nanobers deco- repair processes have been shown to be very prom-
rated with silver nanoparticles (AgNPs) (Zhu ising, showing the potential of these low-cost
et al. 2013). By combining an in situ reduction devices for sensing applications (McWilliams
approach and electrospinning technique, a H2O2, et al. 2014). Polyacrylonitrile electrospun bers
glutathione (GSH), and glucose detection coated with gold nanoparticles have been tested
biosensor has been obtained, showing a high sen- and thiolated DNA has been assembled on these
sitive detection of H2O2 with a detection limit of bers. It has been demonstrated that the bers
5 mM and a fast response, broad linear range, low themselves become the working electrode of the
detection limit, and excellent stability and sensors, modied with either the lesion-bearing or
reusability. defect-free DNA monolayers. From the reported
Looking at the aforementioned literature, it assays, the device sensitivity has been found to be
appears evident that the ongoing trend in biosen- on the order of femtomoles per electrode, and it
sor fabrication is the use of specic nanomaterials means that sensing can be accomplished with even
to get better sensitivity, lower detection limit, 1 ng of enzyme.
and linear response range. Further evidence of
this is the H2O2-modied electrode based on
hemoglobin (Hb) collagen and carbon nanotubes References
(CNTs) proposed by Li et al. (2014). Hb has been
used because of its similar structure to peroxidase, Abidian MR, Corey JM, Kipke DR, Martin DC
CNTs for the electrical properties, and collagen (2010) Conducting-polymer nanotubes improve electri-
cal properties, mechanical adhesion, neural attachment,
for its excellent biocompatibility. Therefore, and neurite outgrowth of neural electrodes. Small 6:421
the resulting hemoglobin-collagen-carbon nano- Bhardwaj N, Kundu SC (2010) Electrospinning: a fasci-
tube (Hb-collagen-CNT) composite nanober- nating ber fabrication technique. Biotechnol Adv
modied electrode has excellent performance, 28:325347
Cao F, Guo S, Ma H, Shan D, Yang S, Gong J (2011) Nickel
thanks to the great surface area and the oxide microbers immobilized onto electrode by
high porosity of the three-dimensional reticular electrospinning and calcination for nonenzymatic glu-
structure which ensures more channels for both cose sensor and effect of calcination temperature on the
electron transfer and diffusion of H2O2. performance. Biosens Bioelectron 26:27562760
DOrazio P (2003) Biosensors in clinical chemistry. Clin
Among enzymes currently used in biosensor Chim Acta 334:4169
fabrication, urease (Sawicka et al. 2005), fructose Ding Y, Wang Y, Su L, Bellagamba M, Zhang H, Lei Y
dehydrogenase (Marx et al. 2011), (2010) Electrospun Co3O4 nanobers for sensitive and
a-chymotrypsin (a proteolytic enzyme acting in selective glucose detection. Biosens Bioelectron
26:542548
the digestive systems of many organisms) (Jia Jia HF, Zhu GY, Vugrinovich B, Kataphinan W, Reneker
et al. 2002), and lipase (Li et al. 2007; Ye DH, Wang P (2002) Enzyme-carrying polymeric
et al. 2006) represent just a few examples of the nanobers prepared via electrospinning for use as
ongoing research. The large surface area of unique biocatalysts. Biotechnol Prog 18:10271032
Kong FY, Li XR, Zhao WW, Xu JJ, Chen HY
electrospun nanobers has been found to be (2012) Graphene oxidethionineAu nanostructure com-
important for increasing both the amount of posites: preparation and applications in non-enzymatic
immobilized enzymes and their own catalyzing glucose sensing. Electrochem Commun 14:5962
Electrospun Nanofibers for Water and Wastewater Treatment Applications 663
Li SF, Chen JP, Wu WT (2007) Electrospun polyacryloni-

trile nanobrous membranes for lipase immobilization. Electrospun Nanofibers for Water
J Mol Catal B: Enzym 47:117124
Li J, Mei H, Zheng W, Pan P, Sun XJ, Li F, Guo F, Zhou and Wastewater Treatment
HM, Ma JY, Xu XX, Zheng YF (2014) A novel hydro- Applications
gen peroxide biosensor based on hemoglobin-collagen-
CNTs composite nanobers. Coll Surf 284B. 118:7782 Michele Modesti, Carlo Boaretti and
Marx S, Jose MV, Andersen JD, Russel AJ
(2011) Electrospun gold nanober electrodes for bio- Martina Roso
sensors. Biosens Bioelectron 26:29812986 Department of Industrial Engineering, University
McWilliams MA, Anka FH, Balkus JK, Slinker JD of Padova, Padova, Italy
(2014) Sensitive and selective real-time electrochemi-
cal monitoring of DNA repair. Biosens Bioelectron E
54:541546
Ramakrishna S, Fujihara K, Teo WE, Lim TC, Ma Z Water pollution is the contamination of water
(2005) An introduction to electrospinning and bodies (e.g., lakes, rivers, oceans, aquifers, and
nanobers. World Scientic, Singapore groundwater), and it occurs when pollutants are
Ren G, Xu X, Liu Q, Cheng J, Yuan X, Wu L, Wan Y
(2006) Electrospun poly(vinyl alcohol)/glucose oxi- discharged directly or indirectly into water bodies
dase biocomposite membranes for biosensor applica- without adequate treatment to remove harmful
tions. React Funct Polym 66:15591564 compounds. Direct sources include efuent out-
Rodriguez-Mozaz S, Marco MP, Lopez de Alda MJ, falls from factories, reneries, waste treatment
Barcel D (2004) Biosensors for environmental moni-
toring of endocrine disruptors: a review article. Anal plants, etc., that emit uids of varying quality
Bioanal Chem 378:588598 directly into urban water supplies. Indirect
Sawicka K, Gouma P, Simon S (2005) Electrospun sources include contaminants that enter the water
biocomposite nanobers for urea biosensing. Sensor supply from soils/groundwater systems and from
Actuat B-Chem 108:585588
Thevenot DR, Toth K, Durst RA, Wilson GS (2001) Elec- the atmosphere via rainwater. Soils and ground-
trochemical biosensors: recommended denitions and waters contain the residue of human agricultural
classication. Biosens Bioelectron 16:121131 practices (fertilizers, pesticides, etc.) and improp-
Wendorff JH, Agarwal S, Greiner A (2012) erly disposed of industrial wastes. Atmospheric
Electrospinning. Materials, processing and applica-
tions. Wiley-VCH, Weihneim contaminants are also derived from human prac-
Wilson GS, Gifford R (2005) Biosensors for real-time in tices (such as gaseous emissions from automo-
vivo measurements. Biosens Bioelectron 20:2388 biles, factories, and even bakeries).
Wu J, Fan Y (2013) Sensitive enzymatic glucose biosensor Contaminants can be broadly classied into
fabricated by electrospinning composite nanobers and
electrodepositing Prussian blue lm. J Electroanal organic, inorganic, radioactive, and acid/base.
Chem 694:15 Moreover, bacteria and virus represent a danger-
Wu J, Yin F (2013) Sensitive enzymatic glucose biosensor ous class of contaminating agents, which are
fabricated by electrospinning composite nanobers and responsible of diseases in developing countries.
electrodepositing Prussian blue lm. J Electroanal
Chem 694:15
Yang G, Kampstra KL, Abidian MR (2014) High-
performance conducting polymer nanober biosensors
for detection of biomolecules. Adv Mater. doi:10.1002/
Microbial Control of Water
adma.201400753
Ye P, Xu ZK, Wu J, Innocent C, Seta P (2006) Nanobrous Currently the basic water treatments for microbial
membranes containing reactive groups: electrospinning control require chemical disinfectants and
from poly(acrylonitrile-co-maleic acid) for lipase
membrane-based ltration systems. If the forma-
immobilization. Macromolecules 39:10411045
Yoo EH, Lee SY (2010) Glucose biosensors: an overview tion of harmful disinfection by-products from
of use in clinical practice. Sensors 10:45584576 chemical disinfectants is the potential drawback
Zhu H, Du ML, Zhang Z, Wang P, Bao SY, Wang LN, Fu that makes their use not completely reliable
YQ, Yao JM (2013) Facile fabrication of AgNPs/
(Li et al. 2008), water ltration membrane is
(PVA/PEI) nanobers: high electrochemical efciency
and durability for biosensors. Biosens Bioelectron affected by biofouling and virus penetration. Bio-
49:210215 fouling can be considered as a biotic form of
664 Electrospun Nanofibers for Water and Wastewater Treatment Applications
organic fouling, and it represents a major problem Nanoltration is a type of pressure-driven mem-
in nanoltration (NF) and reverse osmosis brane process with properties in between reverse
(RO) membrane ltration. osmosis (RO) and ultraltration (UF), which has
At this regard, electrospun nanobers have been found to be adequate for desalination with
been found to be a good candidate for improving low salt content. According to the established
the performance of ltering media. Antifouling technology, in order to reduce the resistance and
properties can be achieved by plasma treatment enhance the ux, thin lm composite membranes
or surface graft polymerization of nanobers, as (TFC) are commonly used. They are based on
suggested by the literature (Musale and Kumar three layer, that is, a nonwoven fabric as support,
2000). Another advantage related to the use of a middle porous support, and a top layer which is
electrospun nanobers for microbial control is responsible for the selective salt removal. Within
the capability to get a wide range of assemblies, this scenario, electrospun nanobers can be used
for instance, antimicrobial nanobers themselves as a mid-layer membrane in order to achieve
and nanobiocides, which include metal, metal- higher ux and low fouling behavior than the
oxide nanoparticles, engineered nanomaterials commercial NF membranes (Yung et al. 2010).
(fullerenes and carbon nanotubes), and natural Nevertheless, the applicability of electrospun
substances (Botes and Cloete 2010). For instance, membranes as self-supporting NF membranes
nanoparticles of Ag, TiO2, and ZnO represent has been proved (Kaur et al. 2012) even though
low-cost materials with a well-known antimicro- a careful optimization of the whole system needs
bial activity that can be properly attached to further studies in terms of process optimization
nanobers reducing potential toxicity and and cost analysis.
leaching effects. As regards desalination by membrane distilla-
Polyurethane, poly(vinylidene) uoride-co- tion (MD), nanobrous membranes based on
hexauoro-propylene (PVDF-HFP), and polycar- electrospun polymers (polyvinylideneuoride,
bonate with quaternary ammonium salt are just PVDF) and copolymers (PVDF-co-F6PP, PVDF-
few examples of antimicrobial nanolters which co-hexauoropropene) have been proven to
have been shown to have a very good ltration reduce the energy requirements of direct contact
efciency of 0.3 mm size particles (99.93 %) and a membrane distillation process, thanks to the
very strong antimicrobial activities against both higher permeate ows (Khayat et al. 2011; Shih
S. aureus and E. coli (Yao et al. 2008, 2009; Kim 2011) than PTFE commercial membranes. Very
et al. 2007). good performance in a direct contact membrane
distillation setup could be achieved also by self-
supporting carbon nanotube (CNT) membranes
Desalination (Dume et al. 2011), which have to be properly
modied for enhancing their hydrophobicity and
Another challenging task related to the water issue that allow to get a 99 % salt rejection and a ux
regards the need of alternative sources in order to rate of 12 kg/m2.
tackle the higher demand for fresh water and
energy resources. Desalination technology that
converts seawater into clean water has been Heavy Metal Ions Removal
found to be a suitable approach for it. Among
the different membrane separation processes, Water pollution by heavy metal ions is another
membrane distillation (MD), electrodialysis important related issue because long-term expo-
(ED), reverse osmosis (RO), freeze desalination sure to metal ions such as lead (Pb) can result in
(FD), ion exchange (IX), and nanoltration acute or chronic damage to the brain and the
(NF) have been developed as desalination tech- central nervous system on humans. Furthermore,
nologies (Sundarrajan and Ramakrishna 2013). heavy metals are dangerous because they tend to
Electrospun Nanofibers for Water and Wastewater Treatment Applications 665
bioaccumulate, i.e., the concentration of a chem- durability and hydrophobicity of carbon nanotube
ical increases in a biological organism over time, bucky paper membranes in membrane distillation.
compared to the chemicals concentration in the Kaur S, Sundarrajan S, Gopal R, Ramakrishna S
environment (Lee et al. 2013). Among the tech- (2012) Formation and characterization of
nologies employed for the treatment of contami- polyamide composite electrospun nanobrous mem-
nated water (chemical oxidation, precipitation, IE, branes for salt separation. J Appl Polym Sci 124:
E205E215
RO), adsorption and membrane separation could Khayet M, Payo G, Carmen AS, Carmen M (2011) Nano-
be enhanced by electrospun nanobers. In the structured at membranes for direct contact membrane
literature several examples are reported, the cen- distillation. WO/2011/117443
tral ideas of which are: Kim SJ, Nam YS, Rhee DM, Park HS, Park WH
(2007) Preparation and characterization of antimicro- E
bial polycarbonate nanobrous membrane. Eur Polym
i. The functionalization of the bers with such J 43:31463152
molecules able to bind metal ions Lee CH, Chiang CL, Liu SJ (2013) Sep Purif Technol
ii. The surface modication of polymer in order 118:737743
Li Q, Mahendra S, Lyon DY, Brunet L, Liga MV, Li D,
to vary certain properties, such as hydrophi- Alvarez PJJ (2008) Antimicrobial nanomaterials
licity or hydrophobicity, water absorbency, for water disinfection and microbial control:
adsorption capacity, and resistance to micro- potential applications and implications. Water Res
bial attack 42:45914602
Mahapatra A, Mishra BG, Hota G (2013) Electrospun
iii. The production of mixed nanocomposite Fe2O3Al2O3 nanocomposite bers as efcient
bers adsorbent for removal of heavy metal ions
from aqueous solution. J Hazard Mater 258259:
As regards the functionalization of bers (point 116123
Musale DA, Kumar A (2000) Solvent and pH resistance of
i), rhodanine is a heterocyclic molecule that surface crosslinked chitosan/poly(acrylonitrile) com-
belongs to the sulfur-containing N and O organic posite nanoltration membranes. J Appl Polym Sci
compounds which possesses uptake performance 77:17821793
toward Ag (I), Pb (II), and Hg (II) ions (Lee Musyoka SM, Ngil JC, Mamba BB (2013) Remediation
studies of trace metals in natural and treated water using
et al. 2013). On the other hand cellulose and surface modied biopolymer nanobers. Phys Chem
chitosan-/polyacrylamide-modied nanobers Earth 66:4550
have been proven to be suitable for the adsorption Shih JH (2011) A study of composite nanober membrane
of trace of metals (Cd, Pb, Cu, Cr, and Ni) from applied in seawater desalination by membrane distilla-
tion. Masters Thesis, National Taiwan University of
the river water and treated water (Musyoka Science and Technology
et al. 2013). Their adsorption capacity has been Sundarrajan S, Ramakrishna S (2013) New directions in
enhanced (point ii) by surface modication nanoltration applications are nanobers the right
using furan-2,5-dione. The electrospun iron materials as membranes in desalination? Desalination
308:198208
oxidealumina ber systems represent an exam- Yao C, Li X, Neoh KG, Shi Z, Kang ET
ple of nanocomposite bers (point iii) that have (2008) Surface modication and antibacterial activity
been used as adsorbent for the removal of heavy of electrospun polyurethane brous membranes with
metal ions, i.e., Cu2+, Ni2+, Pb2+, and Hg2+ from quaternary ammonium moieties. J Membr Sci
320:259267
aqueous system (Mahapatra et al. 2013). Yao C, Li X, Neoh KG, Shi Z, Kang ET (2009)
Antibacterial activities of surface modied electrospun
poly(vinylidene uoride-co-hexauoropropylene)
References (PVDF-HFP) brous membranes. Appl Surf Sci
255:38543858
Botes M, Cloete TE (2010) The potential of nanobers and Yung L, Ma H, Wang X, Yoon K, Wang R, Hsiao BS,
nanobiocides in water purication. Crit Rev Microbiol Chu B (2010) Fabrication of thin-lm nanobrous
36:6881 composite membranes by interfacial polymerization
Dume L, Germain V, Sears K, Sch utz J, Finn N, Duke M, using ionic liquids as additives. J Membr Sci
Cerneaux S, Cornu D, Gray S (2011) Enhanced 365:5258, new container
666 Electrospun Nanofibrous Membranes
the advantages offered in terms of cost and pro-

Electrospun Nanofibrous ductivity over more complicated bottom-up
Membranes approaches, exibility of the membranes, and
the potential upscaling for industrial production
Martina Roso, Carlo Boaretti, Alessandra (Persano et al. 2013).
Lorenzetti and Michele Modesti
Department of Industrial Engineering, University
of Padova, Padova, Italy Fundamentals
The generation of nanobers by electrospinning

Introduction starts from a solution of a proper solvent in which
the desired material is dissolved at a suitable con-
Nanobers are an interesting and versatile class of centration. This solution is usually poured in a
one-dimensional (1-D) nanomaterials, with diam- syringe or a pipette from which it is let own
eters ranging from tenths to hundreds of nanome- through a nozzle using a micropump, and by
ters, which have been recognized as promising means of a high-voltage direct current power sup-
due to their outstanding properties in terms of ply (550 kV), an electrostatic eld is applied
high porosity, excellent pore interconnectivity, between the nozzle and a metal collector with
small diameters, and high surface-to-volume opposite polarity. By increasing the intensity of
ratio. Among the different nanober manufactur- the electric eld, a charge is induced on the sur-
ing technologies, electrostatic spinning or face of the solution at the tip of the capillary. The
electrospinning represents the easiest, most prom- mutual repulsion of the charges elongates the
ising, and versatile method for the generation of hemispherical drop to form a conical shape
aligned or randomly distributed nanobers of a (Taylor cone). Once that the electric eld has
rich variety of different materials like synthetic reached a threshold value, the electrostatic
and natural polymers (Huang et al. 2003), com- force overcomes the surface tension and a
posites (Sahay et al. 2012), ceramics (Dai charged jet of solution is ejected from the tip of
et al. 2011), and metals (Wu et al. 2007). the Taylor cone, producing a jet that travels
With this technique that is essentially a varia- toward the metal collector. During this last phase
tion of the electrostatic spray (or electrospraying), the jet undergoes a whipping process in which it is
a ber is produced thanks to the solidication of stretched and the solvent of the solution evapo-
an electried jet of a solution that is stretched by rates leaving a ber that is collected on a grounded
the repulsive action of surface charges and evap- collector.
oration of solvent. The main difference with The morphology and diameter of electrospun
respect to electrospraying is the presence of high nanobers depend on a wide range of parameters,
elongational viscous forces generated by the chain which can be classied according to three main
entanglements of the polymer present in the solu- different categories:
tion that, thanks to the application of a strong
electric eld, prevent the formation of liquid drop- Solution parameters: type of solvent, viscosity,
lets in favor of a rapid whipping jet. The molecular weight and molecular weight distri-
electrospun solution is essentially polymer bution, vapor pressure, surface tension, and
based, with the possibility to add metal or ceramic solution conductivity
precursors that after electrospinning treatments Process parameters: solution feeding rate,
can generate the corresponding nanobers. electric eld strength, distance between the
An extensive research activity has been carried tip and the collector, needle diameter, compo-
in the last years toward the evaluation of sition, geometry, and motion of the collector
electrospun nanobrous membranes (ENMs) in Ambient parameters: temperature, humidity,
the context of membrane technology, fueled by and air velocity.
Electrospun Nanofibrous Membranes 667
Between them viscosity has a prominent role in including the most traditional and advanced tech-
relation to the morphology of the nanobers. For nologies: air and water ltration, fuel cell mem-
polymer solutions of low viscosity, surface ten- branes and battery membrane separators,
sion is the dominant aspect and just beads or biomedical applications, and many others. Most
beaded nanobers are obtained. As the viscosity of these applications havent already reached their
of the solution is increased, the beads become industrial level, but the promising results from the
bigger, the average distance between them longer, research eld and their different potential applica-
and the shape of the beads changes from spherical tions are believed to encounter the interest of
to spindle-like with a decrease of the ber diam- industry and governments. Actually the most
eter. At a suitable concentration, smooth industrially developed products based on ENMs
E
nanobers without defects can be obtained, but are lters for air purication. Potential products
with a progressive increment of its value, the ready to market are related to liquid ltration and
bers increase their diameter until they change energy storage, in the short-medium period, and to
their shape into micro-ribbons. Usually the proper the biomedical eld, in the medium-long period
viscosity is dened in relation to the specic (Persano et al. 2013).
polymer-solvent system, and for a given
polymer-solvent system, it can change according Air and Water Filtration
to the molecular weight of the polymer. Complex ENMs high specic surface, porosity and perme-
architectures like core-shell, porous, and hollow ability, tailorable thickness, and ber diameters
structures can be formed by electrospinning are desirable features for ltration applications.
through the incorporation of nanoparticles into These materials have been employed for several
bers, co-electrospinning of different polymers, years for air ltration and recently, at a research
or employing solutions with different boiling level, even for microltration, ultraltration, and
point solvents, while modication of the nanoltration in the eld of water treatment. The
electrode-collector system allows to orientate the advantages of electrospun nanobers compared to
bers and to obtain 3-D structures (Teo conventional air lter media are related to a higher
et al. 2011). ltration efciency, lower pressure drops, and
Electrospun nanobers can also be obtained energy savings. Indeed, nanobers with diameters
from polymer melts at high temperature, from a lower than 0.5 mm have much higher capability to
combination of a magnetic and electric eld, aer- collect ne particles because of the slip ow that
ated polymer solution, and from multi-jet devices increases diffusion, interception, and inertial
in order to improve production rate. impaction efciencies, determining a lower drag
force around the bers and, thus, lower pressure
drops. In addition these media are easy to clean,
Applications enabling them to signicantly extend the life of
lters and reducing the maintenance costs. Their
The interesting properties related to the possibility potentials have shown possible applications in air
of obtaining nanometric bers from ltering media (Barhate and Ramakrishna 2007),
electrospinning allow their use for several poten- especially for high-efciency particulate and
tial applications, with a multidisciplinary perspec- low-penetration air lters, lters for transportation
tive. Broadly speaking it is possible to identify applications, adsorptive catalytic gas lter for res-
four main areas of interest (Ramakrishna pirators, lter media for pulse clean cartridges in
et al. 2005): bioengineering, environmental engi- dust collection, and cigarette lters for smoke
neering and biotechnologies, energy and electron- ltration.
ics, and, nally, defense and security. In this ENMs can be employed also for water treat-
framework the specic elds of applications in ment (Cloete et al. 2013) and generally show high
which electrospun nanobers are directly in con- ux rate and low transmembrane pressure with
tact with membrane technology are several, performance comparable to those commercially
668 Electrospun Nanofibrous Membranes
available. The main interesting application in this nonconductive or less conductive polymer into a
area is related to desalination, VOC gas stripping, porous matrix, which acts as mechanical rein-
oil/water emulsion separation, microbial, organic forcement when the pores are lled with a highly
compounds, and heavy metal removal. However, proton-conductive component. Alternatively, a
for solid-liquid ltration, very high surface-to- highly proton-conducting matrix is electrospun
volume ratios promote membrane fouling. For into a porous ber mat and subsequently
this reason and for improving membrane perfor- reinforced with a secondary polymer to provide
mances, ENMs can be surface modied (Nasreen mechanical stability. The proton-conductive poly-
et al. 2013), in order to add functional groups or mers are usually chosen between peruoro-
functionality, with processes like in situ graft and sulfonic acid and sulfonated polymers that can
interfacial polymerization, plasma treatment, wet be coupled with organic or inorganic particles as
treatment, coating, and blending with surface- conductor enhancers or to improve strength and
modifying agents. A main drawback for the hydrophilicity.
employment of electrospun nanobers for such Electrospun mats are also attractive alterna-
applications is their mechanical strength which is tives to polymer gel electrolytes for lithium-ion
not sufcient to withstand macroscopic impact batteries, since they can be employed as matrix in
during ltration application such as normal liquid which the electrolyte can be encapsulated improv-
or air ows passing through them. Hence, they ing ionic conduction across the membrane and
need to be used as active coating layer on existing mechanical strength while providing good water
melt-blown supportive brous media (composite uptake. In these elds the most widely studied
membrane) or the bers need to be bonded to polymer has been poly(vinylidene uoride)
enhance mechanical properties. (PVDF) thanks to its good electrochemical and
thermal stability. However, PVDF-based gel poly-
mer electrolyte with its high crystallinity limits the
Fuel Cell and Battery Separator Membranes ion migration, lowering the battery performance.
Another area in which ENMs can have promising ENMs with their porous membrane are capable to
applications is that of energy-related applications overcome this problem, and PVDF and other
and devices (Cavaliere et al. 2011), especially for alternative polymers and blend have been
polymer fuel cell and lithium-ion battery electro- explored by electrospinning with encouraging
lyte membranes. results.
Research efforts on polymeric proton
exchange membranes for fuel cells have led the Biomedical Applications
interest of researchers toward the employment of Nanober research applications in the biomedical
ENMs as porous reinforcing mats to minimize eld have a multifaceted perspective spreading
in-plane swelling and shrinking. Nanobers are between tissue engineering, drug delivery, and
also able to increase proton conductivity with wound dressing (Agarwal et al. 2008). Tissue
respect to bulk lm, thanks to their highly oriented engineering is an emerging multidisciplinary
ionic domains (Dong et al. 2010) and provide area in which nanobers represent an important
good mechanical strength, while the remarkable advancing front for the production of suitable
exibility of their production process allows to scaffolds of different materials that can mimic
adequately tailor their nal morphology for com- natural extracellular matrix. To this purpose
posite membranes. Fuel cell membranes are semi- ENMs have been tested as natural, synthetic, and
permeable membranes that have the function to be composite scaffolds for different types of targeted
proton conductive, electron insulating, and dense tissue including blood vessel, cartilage, bone,
in order to avoid fuel crossover. For this reason nerve, and many others. The small diameter of
two types of approaches can be used when the bers and their high surface area are benecial
employing electrospinning for such application. for cell attachment and bioactive factor loading,
The rst is related to the electrospinning of enhancing cell growth.
Electrospun Nanofibrous Membranes 669
Drug delivery membranes incorporate a drug catalysts for the abatement of selected
component that can be patched on wound of sur- compounds.
gery or encapsulated into pharmaceutical capsules The characteristics of ENMs match well the
to deliver the drug through the digestive system of requirements for different types of sensing
the patient. Polymeric electrospun drug delivery devices (Ding et al. 2009) including acoustic
systems are advantageous for such task because wave, resistive, gravimetric, photoelectric, opti-
they can deliver drugs efciently to a localized cal, and amperometric sensors. Indeed, ENMs
area, with the possibility to vary the release rate by high surface area has the potential to provide
simply varying the ber diameter or the loading unusually high sensitivity, fast response time,
dosage. The possibility of choosing different and lower detecting limits. In this case different
E
materials, processes, and processing conditions approaches can be employed to provide a sensing
allows to reach the desired encapsulation ef- capability to nanobers, such as electrospinning
ciency, preserving bioactivity (Zamani of a polymeric sensing material, incorporation of
et al. 2013). Several drugs have been incorporated sensing molecules into nanobers, or application
successfully into electrospun media obtaining bet- of sensing material on nanober surface via coat-
ter performance over normal cast lm and with the ing/grafting technique, employing organic and
possibility to load insoluble drug for enhancing inorganic polymer.
their dissolution. Other potential applications of ENMs include
ENMs have also exhibited potential in wound afnity membranes for protein purication
dressing thanks to the possibility to generate (Ma et al. 2005) and protective clothing against
homogeneous scaffolds, provide uniform nanoparticles (Faccini et al. 2012) on textile
adherence, and wet wound surface without uid support.
accumulation. They can provide high gas perme-
ation and protection from infection and dehydra-
tion, extending their applications on different
types of wounds as compared to traditional Cross-References
wound dressing materials and opening new
doors for the next generation of wound dressing Nanober
materials. Oil-Water Emulsion
Other Applications Ultraltration (UF)
Electrospun materials from stable polymers or
ceramic bers are also ideal candidates as sup- References
ports for homogeneous and heterogeneous catal-
ysis in gas phase, since they can provide high Agarwal S, Wendorff JH, Greiner A (2008) Use of
surface area and high porosity. The catalyst, electrospinning for biomedical applications. Polymer
which is usually a semiconductor, can be the 49:56035621
Barhate RS, Ramakrishna S (2007) Nanobrous ltering
nanobrous mat itself, or it can be subsequently media: ltration problems and solutions from tiny
added by surface modications or deposition. materials. J Membr Sci 296:18
Examples are organic or inorganic nanobers on Basheer C (2013) Nanober-membrane-supported TiO2 as
which metals or semiconductor nanoparticles are a catalyst for oxidation of benzene to phenol. J Chem
2013:17
deposited or embedded for catalytic (Basheer Cavaliere S, Subianto S, Savych I, Jones DJ, Roziere J
2013) and photocatalytic oxidation (Modesti (2011) Electrospinning: designed architectures for
et al. 2014) and synthesis (Formo et al. 2009) of energy conversion and storage devices. Energy Envi-
organic compounds in gas phase. The advantages ron Sci 4:47614785
Cloete TE, de Kwaadsteniet M, Gule NP, Klumperman B
of the high surface areas of the catalytic lters (2013) Electrospun nanobrous membranes for water
obtainable with this technique are related to treatment applications. In: Lens PNL, Virkutyte J,
higher activity with respect to common porous Jegatheesan V, Kim SH, Al-Abed S (eds)
670 Electrospun Polyethersulfone Nanofiber Membranes
Nanotechnology for water and wastewater treatment. Further Readings

IWA Publishing, London, pp 283294 Andrady AL (2008) Science and technology of polymer
Dai Y, Liu W, Formo E, Sun Y, Xia Y (2011) Ceramic nanobers. Wiley, Hoboken
nanobers fabricated by electrospinning and their Ding B, Yu J (2014) Electrospun nanobers for energy and
applications in catalysis, environmental science, environmental applications, Nanostructure science and
and energy technology. Polym Adv Technol technology series. Springer, Berlin
22:326338 Jayakumar R, Shantikumar N (2012) Biomedical applica-
Ding B, Wang M, Yu J, Sun G (2009) Gas sensors based on tions of polymer nanobers, advances in polymer sci-
electrospun nanobers. Sensors 9:16091624 ence. Springer, Berlin
Dong B, Gwee L, Salas-de la Cruz D, Winey KI, Elabd YA Modesti M, Lorenzetti A, Roso M (2011) Nanobers via
(2010) Super proton conductive high-purity Naon electrospinning. Encyclopedia of nanoscience and
nanobers. Nano Lett 10:37853790 nanotechnology, vol 17. America Scientic, Valencia,
Faccini M, Vaquero C, Amantia D (2012) Development of California, pp 231312
protective clothing against nanoparticle based on Wendorff JH, Agarwal S, Greiner A (2012)
electrospun nanobers. J Nano Mater 2012(892894) 9 Electrospinning: materials, processing and applica-
pages. doi:10.1155/2012/892894 tions. Wiley-VCH, Singapore
Formo E, Yavuz MS, Lee EP, Lane L, Xia Y (2009)
Functionalization of electrospun ceramic nanobre
membranes with noble-metal nanostructures for cata-
lytic applications. J Mater Chem 19:38783882
Huang ZM, Zhang YZ, Kotaki M, Ramakrishna S (2003) Electrospun Polyethersulfone
A review on polymer nanobers by electrospinning and Nanofiber Membranes
their applications in nanocomposites. Compos Sci
Technol 63:22232253
Michele Modesti, Carlo Boaretti and
Ma Z, Kotaki M, Ramakrishna S (2005) Electrospun
nanober as afnity membrane. J Membr Sci Martina Roso
265:115123 Department of Industrial Engineering, University
Modesti M, Roso M, Boaretti C, Besco S, Hrelja D, of Padova, Padova, Italy
Sgarbossa P, Lorenzetti A (2014) Preparation of smart
nano-engineered electrospun membranes for methanol
gas-phase photoxidation. Appl Catal B 144:216222
Nasreen SAAN, Sundarrajan S, Nizar SAS, Electrospun polyethersulfone (PES) nanober
Balamurugan R, Ramakrishna S (2013) Advancement membranes are semipermeable membranes
in electrospun nanobrous membranes modication
obtained by electrospinning that have been
and their application in water treatment. Membranes
3:266284 explored for possible applications concerning
Persano L, Camposeo A, Tekmen C, Pisignano D (2013) water ltration (microltration, nanoltration,
Industrial upscaling of electrospinning and applications and engineered osmosis), composite polymer
of polymer nanobers: a review. Macromol Mater Eng
electrolyte membranes, and tissue engineering.
298:504520
Ramakrishna S, Fujihara K, Teo WE, Lim TC, Ma Polyethersulfone (PES) is a high-performance
Z (2005) An introduction to electrospinning and aromatic polymer which belongs to the family of
nanobers. World Scientic, Singapore polysulfones. The basic repeating unit of the polymer
Sahay R, Kumar PS, Sridhar R, Sundaramurthy J,
backbone consists of para-linked aromatic groups
Venugopal J, Mhaisalkar SG, Ramakrishna S
(2012) Electrospun composite nanobers and their connected by ether and sulfone groups (see Fig. 1).
multifaceted applications. J Mater Chem PES is synthesized by polycondensation of
22:1295312971 4,4-sulfonylbisphenol (bisphenol S) with
Teo WE, Inai R, Ramakrishna S (2011) Technological
advances in electrospinning of nanobers. Sci Technol
Adv Mater 12:013002 (19pp) O
Wu H, Zhang R, Liu X, Lin D, Pan W (2007)
Electrospinning of Fe, Co, and Ni nanobers: synthe- O S
sis, assembly and magnetic properties. Chem Mater
19:35063511 O n
Zamani M, Prabhakaran M, Ramakrishna S (2013) Polyethersulfone (PES)
Advances in drug delivery via electrospun and
electrosprayed nanomaterials. Int J Nanomedicine Electrospun Polyethersulfone Nanofiber Mem-
8:29973017 branes, Fig. 1 Repetitive unit of polyethersulfone
Electrospun Polyethersulfone Nanofiber Membranes 671
4,4-dichlorophenyl sulfone via nucleophilic sub- Other potential applications concern polymeric
stitution at high-reaction temperature (up to scaffolds as substrate for stem cell culture made of
285 C) in the presence of a suitable solvent like pristine (Christopherson et al. 2009;
diphenyl sulfone, sulfolane, or N-methyl-2- Ardeshirylajimi et al. 2013) and modied (Chua
pyrrolidone (Parker et al. 2012). et al. 2006; Shabani 2009) PES, for which both the
PES is a widely employed polymer for member diameter and the type of functionalization
brane fabrication due to its interesting properties in can have a prominent role in the inltration, dif-
terms of high glass transition temperatures, good ferentiation, and proliferation of the cells.
mechanical strength and stiffness, and outstanding The high stability of the polymer and its chain
thermal and oxidative resistance. This polymer can exibility due its structure make PES a suitable
E
be electrospunned from solutions of suitable con- candidate as non-uorinated replacement for pro-
centration with polar aprotic solvents (e.g., ton exchange membranes, especially for direct
dimethylacetamide) in order to produce nanobers methanol fuel cells. By sulfonation it is possible
with average diameters of 0.752 mm and average to introduce proton conducting functional groups
specic surface area of 1530 m2/g, by changing along the polymer chain, while by electrospinning
the process parameters (Kwak et al. 2013). a suitable nanobrous web can be produced. This
Electrospun nanobrous membranes (ENMs) nonwoven structure can be lled with Naon to
are a new class of energy-saving membranes that obtain a compact membrane with higher electro-
are under extensive study because of their high chemical performances with respect to Naon
interconnected porosity and tunable pore size that 112 and Naon 117 dense membranes (Shabani
can give high permeability and selectivity. One of et al. 2010; Hasani-Sadrabadi et al. 2011).
the most important elds of interest for ENMs is
water ltration for which these membranes could
overcome some intrinsic limitations of conven-
Cross-References
tional porous polymeric membranes. However
ENMs suffer of poor mechanical strength, while
Electrospun Nanobrous Membranes
PES, even if extensively used for commercial
membranes, is a highly hydrophobic polymer so
Ultraltration (UF)
the synthesis of composite membranes is the usual
approach to solve these issues. PES ENMs can be
employed for dye and heavy metal removal by References
blending with polymers that incorporate func-
tional groups that have binding/chelating capabil- Ardeshirylajimi A, Hosseinkhani S, Parivar K,
Yaghmaie P, Soleimani M (2013) Nanober-based
ity (Min et al. 2012; Wu et al. 2014). The polyethersulfone scaffold and efcient differentiation
mechanical properties of these membranes can of human induced pluripotent stem cells into osteoblas-
be improved by interber adhesion/junction tic lineage. Mol Biol Rep 40:42874294
(Yoon et al. 2009; Homaeigohar et al. 2012), Bui NN, Lind ML, Hoek EMV, McCutcheon JR
(2011) Electrospun nanober supported thin lm com-
while the hydrophilicity can be increased by oxi- posite membranes for engineered osmosis. J Membr
dation treatment (Yoon et al. 2009). Sci 385386:1019
A combination of both improvements can be Christopherson GT, Song H, Mao HQ (2009) The inu-
achieved by incorporation of hydrophilic inor- ence of ber diameter of electrospun substrates on
neural stem cell differentiation and proliferation. Bio-
ganic particles which can also allow a higher materials 30:556564
thermal stability (Homaeigohar et al. 2011). PES Chua KN, Chai C, Lee PC, Tang YN, Ramakrishna S,
ENMs can be employed even as a middle layer in Leong KW, Mao HQ (2006) Surface-animated
thin lm composite polymeric membranes for electrospun nanobers enhance adhesion and expan-
sion of human umbilical cord blood hematopoietic
engineered osmosis with the advantage to provide stem/progenitor cells. Biomaterials 27:60436051
higher osmotic uxes than their commercial coun- Hasani-Sadrabadi MM, Shabani I, Soleimani M,
terparts (Bui et al. 2011). Moaddel H (2011) Novel nanober-based triple-layer
672 Embryonic Stem (ES) Cell
proton exchange membranes for fuel cell applications. fertilization and contains from 64 to several hun-
J Power Sources 196:45994603 dred cells organized in an outer shell, the
Homaeigohar SS, Mahdavi H, Elbahri M (2011) Extraor-
dinarily water permeable sol-gel formed trophectoderm, and the inner cell mass (ICM).
nanocomposite nanobrous membranes. J Colloid The ICM is the locus of pluripotent cells destined
Interface Sci 366:5156 to yield all the tissues of the developed organism.
Homaeigohar S, Koll J, Lilleodden ET, Elbahri M (2012) In the process of obtaining embryonic stem cells,
The solvent induced interber adhesion and its inuence
on the mechanical and ltration properties of the trophectoderm is removed by immunosurgery,
polyethersulfone electrospun nanobrous microltration and the ICM is disaggregated and plated on feeder
membranes. Sep Purif Technol 98:456463 cells. Ethical issues surround the derivation of
Kwak NS, Jung WH, Park HM, Hwang TS human ES cells from in vitro fertilized blastocysts.
(2013) Electrospun polyethersulfone brous mats: sul-
fonation, its characterization and solution-phase ammo- Embryonic stem cells have pluripotency and
nium sorption behavior. Chem Eng J 215216:375382 indenite replication characteristics. These cells
Min M, Shen L, Hong G, Zhu M, Zhang Y, Wang X, have the capacity to give rise to differentiated prog-
Chen Y, Hsiao BS (2012) Micro-nano structure poly eny representative of all three embryonic germ
(ether sulfones)/poly(ethyleneimine) nanobrous afn-
ity membranes for adsorption of anionic dyes and heavy layers (ectoderm, endoderm, and mesoderm). ES
metal ions in aqueous solution. Chem Eng J 197:88100 cells are able to differentiate in all cell types differ-
Parker D, Bussink J, Hendrik T, van De Grampel J, ently from adult stem cells that can produce only a
Wheately GW, Dorf EH, Ostlinning E, Reinking K, limited number of cell types. It is possible to modu-
Schubert F, Junger O, Wagener R (2012) Polymer,
high-temperature. In: Ullmanns encyclopedia of late the differentiation of stem cells in a given phe-
industrial chemistry, 7th edn. Wiley-VCH Verlag notype by using specic growth factors and
GmbH & Co. KGaA, Weinheim molecules, which trigger the differentiation process.
Shabani I, Haddadi-Asl V, Seyedjafari E, Babaeijandaghi, ES cells for their ability of propagating them-
Soleimani M (2009) Improved inltration of stem cells
on electrospun nanobers. Biochem Biophys Res selves indenitely represent a valuable tool for
Commun 382:129133 both research and regenerative medicine. They
Shabani I, Hasani-Sadrabadi MM, Haddadi-Asl V, can serve as an unlimited source of any cell type
Soleimani M (2010) Nanober-based polyelectrolytes in the body; human ES cells could yield highly
as novel membranes for fuel cell applications. J Membr
Sci 368:233240 effective in vitro models for use in drug discovery
Wu JJ, Lee HW, You JH, Kau YC, Liu SJ (2014) Adsorp- programs and provide a renewable source of cells
tion of silver ions on polypyrrole embedded for use in transplantation therapy. Cell therapies
electrospun nanobrous polyethersulfone membranes. based on the use of ES cells have been proposed
J Colloid Interface Sci 420:145151
Yoon K, Hsiao BS, Chu B (2009) Formation of functional for tissue replacement after injury or disease.
polyethersulfone electrospun membrane for water puri- Adult stem cells, isolatable from bone marrow,
cation by mixed solvent and oxidation processes. adipose tissue, tooth pulp, and many other loca-
Polymer 50:28932899 tions of the body, are capable of self-renewal and
can be readily expanded ex vivo for several pas-
sages without losing their self-renewal capacity.
Mesenchymal stem cells can differentiate into
Embryonic Stem (ES) Cell multiple tissue-forming cell lineages such as oste-
oblasts, chondrocytes, adipocytes, tenocytes, and
Loredana De Bartolo myocytes. Recent work on the differentiation of
Institute on Membrane Technology, National bone marrow-derived mesenchymal stem cells
Research Council of Italy, ITM-CNR, Rende, into neuron-like cells is another demonstration of
Italy considerable plasticity of adult mesenchymal
stem cells. Stem cells are often a preferred cell
source for regeneration of multiple cell lineage
Embryonic stem (ES) cell is a totipotent stem cell tissues. The ability to expand stem cells is desired
derived from an early stage embryo, which is called to generate cells for tissue engineering in clinical
blastocyst. This stage is reached 45 days after and pharmaceutical applications (Rahaman and
Emulsification 673
Mao 2005). Conventional methods for expanding Siegmund M, Laera S, Favia P, De Bartolo L,
stem cells or progenitor cells comprised polysty- Bader A (2009) Rat embryonic liver cell expansion
and differentiation on NH3 plasma-grafted PEEK-
rene culture dishes and component of extracellular WC-PU membranes. Biomaterials 30:65146521
matrix such as collagen. Alternatively biocompat- Piscioneri A, Campana C, Salerno S, Morelli S, Bader A,
ible and biodegradable materials have been pro- Giordano F, Drioli E, De Bartolo L (2011) Biodegrad-
posed to support cells and promote their able and synthetic membranes for the expansion and
functional differentiation of rat embryonic liver cells.
differentiation and proliferation toward the forma- Acta Biomater 7:171179
tion of a tissue (Piscioneri et al. 2011). Rahaman MN, Mao JJ (2005) Stem cell-based composite
Advances in cell-based therapies would not tissue constructs for regenerative medicine. Biotechnol
have been possible without the innovative design Bioeng 91(3):261284
E
and fabrication of several generations of biomate-
rials. Novel biomaterials with distinct properties
are necessary to accommodate the growth and
interactions of multiple cell lineages in composite Emulsification
tissue constructs (Grifth 2000). Membrane sys-
tems provide the temporary structural framework Emma Piacentini
for tissue-forming cells to synthesize extracellular Institute on Membrane Technology, National
matrices and other functional components in the Research Council of Italy, ITM-CNR, Rende,
intended shape and dimensions (De Bartolo and Italy
Bader 2013). They can respond on the key
demands for utilizing cell-based therapies to engi-
neer composite tissue constructs with a purposeful Emulsication is a process by which one phase is
orientation toward anatomic structures that the broken up, dispersed, and distributed in a second
tissue-engineered constructs are intended to regen- immiscible or partially miscible phase (Leal-
erate. Development of new intelligent biomaterials Calderon et al. 2007). Many different emulsica-
in synergy with cell biology will advance stem cell- tion methods can be identied, and they can be
based clinical therapeutics. Engineered membranes distinguished in nonmechanical and mechanical
have the potential to mimic and control the physi- methods. The nonmechanical methods include the
cal, chemical, and biological factors necessary for dispersed phase precipitation and the phase inver-
guiding stem cell differentiation (Pavlica sion. Changes in the phase behavior of the sub-
et al. 2009). They are currently being investigated stances to be emulsied, promoted by variation of
to act as scaffolds to guide and improve 3D tissue temperature or composition or by mechanical
formation, substrates to enhance cell culturing stress, are used to achieve the desirable state of
techniques, vehicles for cell delivery, and sources the system. The mechanical methods of producing
of immobilized and/or time-released factors. They emulsions include the use of high-speed mixers,
can be applied to the regeneration of numerous colloid mills, high-pressure valve homogenizers,
tissue types, including the liver, pancreas, bone, ultrasonic homogenizers, microuidization, and
cartilage, skin, and nerves, and are being used for membrane emulsication. Depending on the nature
the in vitro realization of physiological models. of the starting materials, emulsication can be dis-
tinguished into two categories. The creation of an
emulsion directly from two separate liquids is
References
dened as primary emulsication, whereas the
De Bartolo L, Bader A (2013) Biomaterials for stem cell reduction in size of the droplets in preformed emul-
therapy: state of art and vision for the future. CRC sion is dened as secondary emulsication (Fig. 1).
Press, Boca Raton The physical processes that occur during emul-
Grifth LG (2000) Polymeric biomaterials. Acta Mater
sication can be highlighted by considering the
48:263277
Pavlica S, Piscioneri A, Peinemann F, Keller M, behavior of two immiscible liquids in a container
Milosevic J, Staeudte A, Heilmann A, Schulz- such as oil and water. Their thermodynamically
674 Emulsification
PRIMARY SECONDARY
EMULSIFICATION EMULSIFICATION
Oil
Water
Emulsification, Fig. 1 Emulsication
most stable state consists of a layer of oil on top of uid. When the viscous forces generated within a
a layer of water that allows to minimize the con- uid dominate the inertial forces (low Re), the
tact area between the two phases. To create and ow prole is laminar; when the inertial forces
emulsion, it is necessary to supply energy in order dominate (high Re) in the ow prole, it is turbu-
to disrupt and mix the oil and water which is lent. The size of the droplets produced by a
usually achieved by mechanical agitation. The homogenizer depends on a balance between the
droplets formed are constantly moving around two opposing physical processes: droplet disrup-
and frequently collide and coalesce with neigh- tion and droplet coalescence. The interfacial
boring droplets. The presence of an emulsier forces that tend to hold the droplets together and
prevents the merging together of the droplets the disruptive forces generated within the homog-
after they are formed. The emulsier adsorbs to enizer that tend to pull the droplets apart are
the surface of the droplets during emulsication involved in droplet disruption process. To deform
and forms a protective membrane that prevents and disrupt a droplet during homogenization, it is
the droplets from coming close enough together necessary to apply an external force that is signif-
to coalesce. The rates of droplet disruption, drop- icantly larger than the interfacial force. The inter-
let coalescence, and emulsier adsorption within a facial force is described by the Laplace equation:
particular homogenizer depend on the ow prole
that the uids experience: (i) laminar ow which 4g
DP
is a regular, smooth, and well-dened ow with d
relatively low ow rate; (ii) turbulent ow which
is an irregular, chaotic, and ill-dened ow with where g is the interfacial tension between the two
relatively high ow rate characterized by the for- liquids, d is the droplet diameter, and DP is the
mation of eddies within the uid; and (iii) Laplace pressure which acts across the interface
cavitational ow which is an extremely complex toward the center of the droplet. The equation
ow because of the formation of small cavities indicates that the pressure required to disrupt a
that implode and generate shock waves. The ten- droplet increases as the interfacial tension
dency for one ow regime is a consequence of the increases or as the droplet size decreases. For a
balance of viscous and inertial forces acting on the droplet to be broken up during homogenization,
uid expressed by the Reynolds number: the disruptive forces must exceed the interfacial
forces and their duration must be longer than the
inertial forces L vrc time required for droplet deformation. The rela-
Re tive magnitude of disruptive and interfacial forces
viscous forces c
is characterized by the Weber number (We):
where L is some characteristic length of the sys-
tem, v is the average uid ow velocity, rc is the disruptive forces
We
density of the uid, and c is the viscosity of the interfacial forces
Emulsification by Membrane Operations 675
Emulsification, Table 1 The type of homogenizers used for emulsification

Homogenizer Droplet size and size Energy density
type Droplet formation mechanism Productivity distribution (J m3)
High-speed mixer Droplets break up in TI, TV, and Batch or >2 mm, polydisperse Low-high
LV ow regime continuous
Colloid mill Droplets break up in LV and TV Continuous >1 mm, polydisperse Low-high
ow regime 103108
High-pressure Droplets break up in TI, TV, LV, Batch or >0.1 mm, Medium-high
homogenizer and CI ow regime continuous polydisperse 106108
Ultrasonic Droplets break up in CI ow Continuous >0.1 mm, Medium-high
regime polydisperse 106108
E
Microuidization Droplets break up in TI and TV Batch or <0.1 mm, Medium-high
ow regime continuous polydisperse 106 to 2 108
Membrane Droplet detachment by wall shear Continuous >0.3 mm, narrow Low-medium
stress <103108
TI turbulent inertial, TV turbulent viscous, LV laminar viscous, CI cavitational
During homogenization droplet-droplet collisions

are particularly rapid because of the intense Emulsification by Membrane
mechanical agitation of the emulsion. Droplet Operations
coalescence will depend on the time taken for
the emulsier to be adsorbed to the surface of Emma Piacentini and Lidietta Giorno
the droplets relative to the time between droplet- Institute on Membrane Technology, National
droplet collisions. The ow prole and the nature Research Council of Italy, ITM-CNR, Rende,
of the emulsier used inuenced these times. Italy
The characteristics of the different type of
homogenizers are reported in Table 1.
High-speed mixer and colloid mills are suitable
Emulsions can be generated using membrane
for preparing emulsions with relatively large drop-
operations by: (i) direct membrane emulsication,
let sizes (>1 mm), while the other types of homog-
where a microporous membrane is used to dis-
enizers can be used to prepare submicron droplets.
perse one of two immiscible liquids as a form of
High-speed mixers, ultrasonic homogenizers,
droplets into another liquid or by (ii) premix mem-
microuidizer, and membrane homogenizers can
brane emulsication, where a course emulsion is
be used for primary emulsication, whereas high-
extruded through the membrane pores in order to
pressure valve homogenizers and colloid mills are
generate ne droplet emulsion. A pressure is
most suitable for secondary emulsication. Most of
required to cause the dispersed phase to permeate
these homogenizers have high productivity and
through the membrane. If any shear is applied on
they are able to work in a batch or continuous
the membrane surface, droplets can be spontane-
operation mode. In particular, membrane homoge-
ously detached from the pore outlets on the action
nizers have appreciably lower productivity than the
of the interfacial tension when they reach a certain
other major types of homogenizers. The use of
size. The method is referred as static membrane
membrane homogenizers may be particularly use-
emulsication (Kukizaki 2009). The shear on the
ful where narrow droplet size distributions are
membrane surface can be generated by moving
important such as for drug delivery.
the continuous phase or using moving mem-
branes. The method is referred as dynamic mem-
References brane emulsication. The shear stress is generated
by the continuous phase owing tangentially to
Leal-Calderon F, Schmitt V, Bibette J (2007) Emulsion the membrane surface (cross-ow membrane
science basic principles. Springer, New York emulsication) (Williams et al. 1998) or stirred
676 Emulsification by Membrane Operations
Emulsification by Membrane Operations, Table 1 Different membrane operations used for the production emulsion
droplets
Dynamic membrane emulsication
Stirred ME Suitable for low volumes of continuous Emulsion production at batch scale
phase
Useful to study the effect of different The shear stress at the membrane surface is
experimental conditions on the emulsion nonuniform and depends on the cell geometry
preparation
Cross-ow ME Constant shear stress at the membrane Emulsion droplet breakup as a consequence of
surface the recirculation
Suitable for large-scale production and High dispersed phase concentration is obtained
continuous or semicontinuous operation after a long time of operation
Cross-ow ME Constant shear stress at the membrane High dispersed phase concentration is obtained
with static surface after a long time of operation
promoter Suitable for large-scale production and
continuous or semicontinuous operation
Suitable to prevent droplet breakup
Pulsed-ow ME Suitable for large-scale production and

Able to prevent droplet breakup
High dispersed phase concentrations
obtained in a single pass
Rotating ME Suitable for large-scale production and Complicated and more expensive design
continuous or semicontinuous operation Higher power consumption
Able to prevent droplet breakup
High dispersed phase concentrations
obtained in a single pass
(continued)
Emulsification by Membrane Operations 677
Emulsification by Membrane Operations, Table 1 (continued)

Vibrating ME Suitable for large-scale production and Complicated and more expensive design
Able to prevent droplet breakup Higher power consumption
Nonuniform temporal distribution of shear stress
on the membrane surface
E
Static membrane emulsication
Simple experimental setup Emulsion production at batch scale
Low energy input The production of uniform droplets is possible
Able to prevent droplet breakup only at low dispersed phase ux with low
productivity
(stirred membrane emulsication) (Stillwell phases, phase composition, pH, and ionic
et al. 2007). The use of static turbulence pro- strength
moters in cross-ow membrane emulsication Process parameters, including wall shear
allows increasing the shear stress at the membrane stress, transmembrane pressure, membrane
surface while maintaining a low value along the module conguration, and temperature
circuit out of the membrane (Koris et al. 2011).
Alternatively, a periodic ow pulsation is gener- Comparing to the conventional emulsication
ated in the continuous phase without recirculation processes, emulsication by membrane operation
(pulsed-ow membrane emulsication) (Holdich permits to obtain a better control of droplet size
et al. 2013). Moving membranes can be used and droplet size distribution, low energy and
instead of the commonly used stationary mem- material consumption, and modular and easy
branes. The droplet detachment from the mem- scale-up. Nevertheless, productivity (m3/day) is
brane surface is controlled by rotating much lower, and therefore the challenge in the
(Vladisavljevi and Williams 2006) or vibrating future is the development of new membranes
(Holdich et al. 2010) the membrane within an and innovative membrane operations to keep the
otherwise static continuous phase. Membrane known advantages and maximize the
operations used for the production emulsion drop- productivity.
lets are described in Table 1.
Emulsication by membrane operation is con-
trolled by: References

Membrane parameters, including porosity, Kosvintsev SR (2010) Membrane emulsication with
mean pore size, pore geometry, pore distance, oscillating and stationary membranes. Ind Eng Chem
and membrane surface wettability Res 49:38103817
Phase parameters, including interfacial ten-
Piacentini E (2013) Continuous membrane emulsica-
sion, emulsier type and concentration, viscos- tion with pulsed (oscillatory) ow. Ind Eng Chem Res
ity and density of dispersed and continuous 52:507515
678 Emulsifier
Koris A, Piacentini E, Vatai G, Bekassy-Molnar E, contains both a polar or hydrophilic head and a
Drioli E, Giorno L (2011) Investigation on the effects nonpolar or hydrophobic tail (Fig. 1). A measure
of a mechanical shear-stress modication method dur-
ing cross-ow membrane emulsication. J Membr Sci of the degree to which an emulsier is hydrophilic
371:2836 or lipophilic is given by the hydrophilic-lipophilic
Kukizaki M (2009) Shirasu porous glass (SPG) membrane balance (HLB) determined by calculating values
emulsication in the absence of shear ow at the mem- for the different regions of the molecule. Emulsi-
brane surface: inuence of surfactant type and concen-
tration, viscosities of dispersed and continuous phases, ers adsorb at interfaces between two immiscible
and transmembrane pressure. J Membr Sci 327:234243 liquid anchoring its hydrophilic part into water
Stillwell MT, Holdich RG, Kosvintsev SR, Gasparini G, and its lipophilic part into oil decreasing the inter-
Cumming IW (2007) Stirred cell membrane emulsi- facial tension between them (Fig. 1). This facili-
cation and factors inuencing dispersion drop size and
uniformity. Ind Eng Chem Res 6:965972 tates droplets disruption during homogenization
Vladisavljevi GT, Williams RA (2006) Manufacture of and, in the case of membranes, lowering the min-
large uniform droplets using rotating membrane emul- imum emulsication pressure. Emulsiers have
sication. J Colloid Interface Sci 299:396402 an important role in emulsion stabilization against
Williams RA, Peng SJ, Wheeler DA, Morley NC,
Taylor D, Whalley M, Houldsworth DW (1998) Con- droplets coalescence and/or aggregation provid-
trolled production of emulsions using a crossow mem- ing electrostatic repulsion, steric repulsion, and/or
brane, part II: industrial scale manufacture. Chem Eng strength to the interfacial layer of the droplets.
Res Des 76:902910 Emulsiers are normally classied according
to the head group type as ionic (anionic and cat-
ionic), nonionic, and amphoterics (zwitterionics).
Anionic emulsiers contain anionic functional
Emulsifier
groups at their head, such as sulfate (sodium
dodecyl sulfate, SDS), sulfonate, phosphate, and
Emma Piacentini
carboxylates. Cationic emulsiers contain cat-
ionic functional groups at their head, such as
pH-dependent primary, secondary, or tertiary
Italy
amines and permanently charged quaternary
ammonium cation (cetyl trimethylammonium
Emulsiers are a class of molecules with surface bromide, CTAB). Nonionic emulsiers have
active properties (Dickinson 2009). This behavior polar headgroups and they include glycol and
is due to their amphiphilic structure which glycerol esters, polyoxyethylene esters,
Emulsifier,
Fig. 1 Emulsier structure EMULSIFIER
and emulsier distribution
at water-in-oil and oil-in- Hydrophilic Head
water emulsion interface
Hydrophobic Tail
Water-in-oil emulsion Oil-in-water emulsion

Emulsion 679
polyoxyethylene ethers, polyoxyethylenepo- multiple emulsions, for example, water-in-oil-in-

lyoxypropylene copolymers, sorbitan derivatives, water emulsions (W/O/W), in which water droplets
and sucrose esters. Zwitterionic (amphoteric) are dispersed within larger oil droplets which are
emulsiers have both cationic and anionic centers themselves dispersed in an aqueous phase and
attached to the same molecule. The cationic part is oil-in-water-in-oil emulsions (O/W/O) consisting
based on primary, secondary, or tertiary amines or of oil droplets dispersed in larger water droplets
quaternary ammonium cations. The anionic part which are themselves dispersed in an oil phase.
can be more variable and include sulfonates, The preparation of an emulsion is termed emul-
sultaines, betaines, and phosphates (lecithin). sication and the agents used for this purpose are
A variety of emulsiers are natural products termed emulsiers. Other agents, such as emul-
E
derived from plant or animal tissue such as hydro- sion promoters or stabilizers, are often added to an
colloids (high molecular weight polysaccharides) emulsion to promote the emulsifying process, for
and proteins. example, by increasing the viscosity or providing
a protective colloid action. The preparation of
emulsions involves breaking up the internal
References phase by supplying mechanical or chemical
energy. When an emulsion is formed, the interface
Dickinson E (2009) Hydrocolloids as emulsiers and between the phases is considerably increased as a
emulsion stabilizers. Food Hydrocolloids 23:
result of the droplet formation. The liquid always
14731482
tends to reduce its surface or interface to a mini-
mum; therefore, an increase in interface is possi-
ble only if energy is supplied. The work that must
be expended on drop division is:
Emulsifiers
dA g dI
Amphiphilic Molecules
where dA is the work to be expended and dI is the
increase in interface. The proportionality factor is
the interfacial tension g between the phases to be
Emulsion emulsied. Thus, if the interfacial tension
between the two phases is high, considerable
Emma Piacentini mechanical energy is required for emulsication
Institute on Membrane Technology, National unless an emulsier is added; if the interfacial
Research Council of Italy, ITM-CNR, Rende, tension is low, little mechanical energy is
Italy consumed.
According to the droplet size, emulsions are
classied as follows:
An emulsion consists of two immiscible liquids
(usually oil and water) with one of the liquids Macroemulsions: these usually have a size
(dispersed phase or internal) dispersed as a form range of 0.15 mm.
of spherical droplets in the other (continuous phase Nanoemulsions: these usually have a size
or external) (Israelachvili 1994). Depending upon range of 20100 nm.
the nature of the dispersed phase, the emulsions are Micellar emulsions or microemulsions: these
classied as (i) oil-in-water emulsions (O/W) usually have a size range of 550 nm.
consisting of oil droplets dispersed in an aqueous
phase and (ii) water-in-oil emulsions (W/O) If the droplet size exhibits a wide statistical
consisting of aqueous droplets dispersed in an oil distribution, the emulsion is described as polydis-
phase. It is also possible to prepare various types of perse, in contrast to monodisperse systems with a
680 Emulsion
Emulsion, Fig. 1 The

various breakdown
processes in emulsions
No change in droplet size and size distribution
Change in droplet size and size distribution

FLOCCULATION OSTWALD
RIPENING
CREAMING EMULSION COALESCENCE
SEDIMENTATION PHASE
INVERSION
uniform droplet size. The ideal particle size particle size distribution and density difference
depends on the available methods of preparation between the droplets and the medium (Fig. 1).
and industrial application in each case. Another In sedimentation, the uniform dispersion of the
important emulsion property is the ratio of the droplets is disturbed by aggregation, which leads
volume of the dispersed phase (Vi) to the volume to settling or creaming of the internal phase. This
of the continuous phase or (Ve) is called the phase process results from external forces usually grav-
volume ratio (F). If F < 0.43 (Vi = 30 % of total itational or centrifugal. When such forces exceed
volume), the ow properties of the emulsion are the thermal motion of the droplets (Brownian
determined primarily by the continuous phase. If motion), a concentration gradient builds up in
F > 0.43, the viscosity of the emulsion generally the system with the larger droplets moving faster
increases. As F increases, either phase reversal or to the top (if their density is lower than that of the
cream formation occurs. medium) or to the bottom (if their density is larger
Emulsion stability should match its applica- than that of the medium) of the container. To keep
tion. Thus, for a number of applications, the emul- an emulsion stable, such aggregation must be
sion should be stable under very specic prevented since droplet aggregates sediment
conditions, but it should break after its purpose faster than individual small droplets. Sedimenta-
has been achieved according to a specic condition is not always necessarily accompanied by
tion (such as temperature, pH, or salts, or the like). coalescence. Although the distribution has been
An emulsion is stable if fusion of the droplets is altered, the original dispersion can be restored by
prevented by a sufciently high energy barrier shaking or stirring. Flocculation refers to aggre-
(Tadros 2013). In general, the energy barrier is gation of the droplets (without any change in
built up by the lm of emulsier that forms at primary droplet size) into larger units. It is the
the surface of the droplets. Several breakdown result of the van der Waals attraction that is uni-
processes may occur on storage depending on versal with all disperse systems. Flocculation
Emulsion 681
occurs when there is no sufcient repulsion to absorbed by the larger droplets, and then increas-
keep the droplets apart to distances where the ingly larger drops merge together until two con-
van der Waals attraction is weak. Flocculation tinuous phases are nally formed. The driving
may be strong or weak, depending on the force for coalescence is the surface or lm uctu-
magnitude of the attractive energy involved. One ations which results in close approach of the drop-
way to overcome the van der Waals attraction is lets whereby the van der Waals forces is strong
by electrostatic stabilization using ionic surfac- thus preventing their separation. Two droplets can
tants, which results in the formation of electrical only coalesce if the intervening layer of liquid is
double layers that introduce a repulsive energy pierced when they approach each other. There-
that overcomes the attractive energy. The second fore, coalescence is opposed in two ways by the
E
and most effective method of overcoming occu- emulsier lm surrounding the droplets. First, as
lation is by steric stabilization using nonionic in the case of aggregation, the like charges of the
surfactants or polymers. Ostwald ripening electrical double layer prevent them from
(disproportionation) results from the nite solu- approaching each other. Second, the buildup of
bility of the liquid phases. Liquids that are referred an elastic surface lm causes the emulsion drop-
to as being immiscible often have mutual solubil- lets to bounce off each other when they collide.
ities that are not negligible. With emulsions, Coalescence is always followed by accelerated
which are usually polydisperse, the smaller drop- settling or creaming, which destroys the emulsion
lets will have larger solubility when compared completely. The emulsion is then broken and can-
with the larger ones (due to curvature effects). not be reconstituted by shaking or stirring. The
With time, the smaller droplets disappear and driving force for prevention of coalescence is to
their molecules diffuse to the bulk and become produce a stable lm that can be achieved by two
deposited on the larger droplets. With time, the mechanisms and their combination: (i) increased
droplet size distribution shifts to larger values. repulsion both electrostatic and steric and
Several methods may be applied to reduce Ost- (ii) dampening of the uctuation. In general,
wald ripening: (i) Addition of a second dispersed smaller droplets are less susceptible to surface
phase component that is insoluble in the continu- uctuations and hence coalescence is reduced.
ous medium. In this case, partitioning between This explains the high stability of nanoemulsions.
different droplet sizes occurs, with the component The phase inversion refers to the process whereby
having low solubility expected to be concentrated there will be an exchange between the disperse
in the smaller droplets. During Ostwald ripening phase and the medium. For example, an O/W
in a two-component system, equilibrium is emulsion may with time or change of conditions
established when the difference in chemical invert to a W/O emulsion. In many cases, phase
potential between different size droplets (which inversion passes through a transition state
results from curvature effects) is balanced by the whereby multiple emulsions are produced. Phase
difference in chemical potential resulting from inversion of emulsions can be one of two types:
partitioning of the two components. This effect transitional inversion induced by changing the
reduces further growth of droplets. facers that affect the HLB of the system, for
(ii) Modication of the interfacial lm at emulsion example, temperature and/or electrolyte concen-
interface. By using surfactants that are strongly tration, and catastrophic inversion, which is
adsorbed at the emulsion interface (i.e., polymeric induced by increasing the volume fraction of the
surfactants) and that do not desorb during ripening disperse phase.
(by choosing a molecule that is insoluble in the Emulsions have application in several indus-
continuous phase), the rate could be signicantly trial systems such as food emulsion, for example,
reduced. In coalescence, the individual droplets mayonnaise, salad creams, deserts, and bever-
fuse together. First, the smaller droplets are ages; personal care and cosmetics, for example,
682 Emulsion Characterization
hand creams, lotions, hair sprays, and sunscreens; These tests cannot identify if a double emulsion
and pharmaceuticals, paints, and bitumen (e.g., oil in water in oil phase) is present. This must
emulsions. be investigated with microscopic methods.
The two main characteristics of an emulsion,
which inuence the physical stability, the color,
References and the rheological behavior, are the drop size
distribution (DSD) and the disperse phase content
Israelachvili J (1994) The science and applications of (DPC) (Schuchmann 2007).
emulsions - an overview, Colloids and Surfaces A:
The DPC is set by the manufacturer and is
Physicochemical and Engineering Aspects 91: 18.
Tadros TF (2013) Emulsion Formation, Stability, and constant when using premix membrane emulsi-
Rheology, in Emulsion Formation and Stability cation. Using other membrane emulsication pro-
(ed T. F. Tadros), Wiley-VCH Verlag GmbH & Co. cesses like cross ow devices, the DPC of
KGaA, Weinheim, Germany.
emulsions is adjusted by the rate of ow of the
oil and water phase or by recirculation time of the
continuous phase/emulsion and changes continu-
ously. Regarding multiple emulsions, additional
Emulsion Characterization instability mechanisms may lead to changes of
DPC while emulsifying and storing. The DPC
Volker Gaukel, Heike Schuchmann and can be measured with differential scanning calo-
Richard Bernewitz rimetry (DSC) (Schuch et al. 2013; Dalmazzone
Food Process Engineering, Karlsruhe Institute of et al. 2009), NMR (Bernewitz et al. 2011; van
Technology (KIT), Karlsruhe, Germany Duynhoven et al. 2007), or rheological character-
istics of single and double emulsions.
The DSD depends on several factors as the kind
The production of emulsions via membrane pro- and pore size of the membrane, the transmembrane
cesses is investigated since 30 years and started ux, the shear forces at the membrane surface, as
with investigations from Nakashima et al. with well as the composition of the emulsion (kind of
porous glass membranes (Nakashima emulsier, viscosity of the phases, etc.). The DSD
et al. 1991). There are several principles of mem- can be measured with various methods which are
brane emulsication as, e.g., direct emulsication referred here. The DSD can also change during
and premix emulsication. But nevertheless storage and distribution of an emulsion leading to
which process is used, the characterization of the a product deterioration and phase separation.
emulsion is an important issue. A rough indication for the physical stability of an
Besides the emulsions ingredients, which are emulsion can be the zeta potential (z), whereas z
dened by the manufacturer, there are many struc- >30 mV indicates a stable emulsion. Another
ture parameters which inuence the physical sta- method is a Dynamic Mechanical Analysis
bility, the color, the rheological behavior, or the (DMA) of the emulsion (Brummer 2006).
controlled release properties of the emulsion. As the changes of DSD are often very slow and
If the kind of emulsion is not known, the rst cannot be monitored over the whole shelf life of a
characterization method must identify if it is an oil product, there are several test procedures to accel-
in water (o/w) or water in oil (w/o) emulsion. In erate this process. A procedure is to rise the stor-
many cases this can be easily done by a dilution age temperature from room temperature to
test. This means that an o/w emulsion can be 4050 C which accelerates the deterioration pro-
diluted with a hydrophilic phase and w/o emulsions cesses by the factor 2. Other means are the expo-
can be diluted with a hydrophobic phase. Measure- sure of the emulsion to many short temperature
ment of conductivity of the emulsion is another abuses or to a centrifugal eld which can shorten
possibility as o/w emulsions have in general a the monitoring time by the factor 102,000. It
much higher conductivity than w/o emulsion. should be kept in mind that all the acceleration
Emulsion Liquid Membrane (ELM) 683
techniques may change the structure of the emul- aqueous phase and oil-in-water emulsion dis-
sion and may lead to differing results compared to persed in an outer organic phase. The membrane
real storage conditions. phase in the water-in-oil-in-water (W/O/W) type
is the immiscible oil phase separating the aqueous
phases, while in the O/W/O type, the immiscible
References water phase separating the two organic phases
acts as the LM. Hence, the liquid membrane
Bernewitz R, Guthausen G, Schuchmann HP (2011) NMR serves here a dual purpose: (a) permitting selec-
on emulsions: characterisation of liquid dispersed sys-
tive transfer of one or more components through it
tems. Magn Reson Chem 49:93104
from external phase to internal droplets and vice
Brummer R (2006) Most important test methods. In: Rhe- E
ology essentials of cosmetic and food emulsions. versa a`nd (b) preventing mixing of external and
Springer, Berlin, pp 7580 internal phases. The emulsion is dispersed in the
Dalmazzone C, Noik C, Clausse D (2009) Application of
feed solution, and mass transfer from the feed to the
DSC for emulsied system characterization. Oil & gas
science and technology rev. IFP 64(5):543555 internal receiving phase takes place. ELMs were
Nakashima T, Shimizu M, Kukizaki M (1991) Membrane rst used for separation of hydrocarbons from
emulsication by microporous glass. Eng Mater wastewater with high separation efciency. Com-
61&62:513516
pared to conventional processes, emulsion liquid
Schuch A, Khler K, Schuchmann HP (2013) Differential
scanning calorimetry (DSC) in multiple W/O/W emul- membrane (ELM) process has some attractive fea-
sions: a method to characterize the stability of inner tures, for example, simple operation, high ef-
droplets. J Therm Anal Calorim 111(3):18811890 ciency, extraction and stripping in one stage, larger
Schuchmann HP (2007) In: Brckel U, Meier W,
interfacial area, and scope of continuous operation.
Wagner G (eds) Product design and engineering: best
practices, vol 1. Wiley-VCH, Weinheim, p 63 Since an ELM is a thin lm of liquid (oil or
van Duynhoven PM, Maillet B, Schell J, Tronquet M, aqueous) composed of surfactants and their sol-
Goudappel GJW, Trezza E, Bulbarello A, van vents between a feed and a receiving phase, any
Dusschoten D (2007) A rapid benchtop NMR method
immiscible liquid can serve as a membrane
for determination of a droplet size distributions in food
emulsions. Eur J Lipid Sci Technol 109(11):10951103 between two liquid or gas phases containing a
solute at different concentrations. If the solute is
Further Reading soluble in the membrane phase and has a reason-
Mcclements DJ (2007) Critical review of techniques and able diffusivity through the membrane, then its
methodologies for characterization of emulsion stabil- selective transport through the membrane from
ity. Crit Rev Food Sci Nutr 47(7):611649
higher to lower concentration can be achieved.
Schramm LL (1992) Emulsions fundamentals and appli-
cations in the petroleum industry, vol 237. American This type of permeation has simple mechanism
Chemical Society, Washington, DC and not of much technical importance.
At facilitated transport mechanism, liquid
membrane incorporates a reactive component or
carrier, reacting reversibly and selectively with
Emulsion Liquid Membrane (ELM) species of interest to carry the formed complexes
across the LM to the internal phase, and dissoci-
Vladimir S. Kislik ates, discharging the solute to the internal phase.
Campus Givat Ram, Casali Institute of Applied The unchanged carrier then diffuses back to the
Chemistry, The Hebrew University of Jerusalem, membrane-external phase interface (see Fig. 1).
Jerusalem, Israel A small amount of carrier is required in the mem-
brane phase even for achieving a high degree of
separation. At different proton concentrations in the
Emulsion liquid membrane, ELM, is a system in aqueous phase or using another ions, ion exchange
the form of double emulsions (for details, see processes between two LM surfaces occur. This
Chakraborty et al. 2010). It may be of two types: phenomenon is called coupled mass transport. If
water-in-oil emulsion dispersed in an external the transports of these two different species occur
684 Emulsion Liquid Membrane for Wastewater Treatment
has been used as carrier. One more plant with the

Organic 200 m3/h capacity in Germany and the 200 m3/h
capacity plant in the Netherlands.
Weak acids like phenol and cresol and weak
M2+ bases like ammonium and amines have been suc-
MR2 M2+ cessfully removed from wastewater. Among them,
Strip
H+ the separation-concentration of phenol has been
intensively investigated. Phenol removal from
RH wastewater was commercialized in China. Phenol
H+ is removed from about 1,000 ppm to 0.5 ppm with
Feed an extraction efciency of greater than 99.95 %.
ELM technology has been applied to a great
extent for separation of mixtures of saturated and
Feed Phase Membrane Strip Phase
I Phase II III aromatic hydrocarbons, of amino acids, and of
H+
strong acids like nitric acid and thiocyanate. Cya-
Cm nide removal from wastewater in gold processing
Cmc is commercialized in China. Cyanide is reduced
Ci
from about 130 ppm to 0.5 ppm with an extraction
efciency of 99.6 %.
ELM has promise in the elds of biotechnol-
Cmc
H+ ogy and biomedicine and has found application in
Ci the separation of organic acids, fatty acids, puri-
cation of antibiotics, enzyme-catalyzed reactions,
and detoxication of blood.
Emulsion Liquid Membrane (ELM), Fig. 1 Schematic
representation of the liquid membrane globule and concen-
tration prole of solute through ELM
References
in the same direction, it is called cotransport, while Chakraborty M, Bhattacharya C, Datta S (2010) Emulsion
hybrid liquid membranes: denitions and classication,
transport in opposite direction is called
theories, module design, applications, new directions and
countertransport. Evidently the process leads to perspectives. In: Kislik V (ed) Liquid membranes princi-
the transport of targeted ionic species across the ples and applications in chemical separations & waste-
membrane against their concentration gradient. water treatment, 1st edn. Elsevier, Boston, pp 141200
This so-called uphill transport will continue
until one driving factor (difference of chemical
potentials) is balanced by the difference between
chemical potentials of another transported ion. Emulsion Liquid Membrane
The ELM technique has great potential for for Wastewater Treatment
recovery and removal of different metal ions.
Separation of metals like copper, zinc, cadmium, N. Othman
cobalt, nickel, mercury, uranium, chromium, rhe- Faculty of Chemical and Energy Engineering,
nium, and several others, including noble metals Universiti Teknologi Malaysia, Skudai, Johor,
like gold and silver, lanthanides, and rare earths, Malaysia
was studied. To date, there are two industrial
plants installed for zinc recovery from wastewater
in Austria, having a capacity of 75 m3/h and Emulsion liquid membranes are known as double
700 m3/h, removing zinc selectively from 500 to emulsion system. The advantage of this process is
3 ppm. Bis(2-ethylhexyl) dithiophosphoric acid extraction and stripping process occurred
Emulsion Liquid Membrane for Wastewater Treatment 685
simultaneously in one single-step operation, and phase. At the external interface, the carrier will
equilibrium limitation can be removed. It also can form carrier-solute complexes and diffuse to the
reduce the amount of expensive extractant; high internal interface and release the solute into the
uxes and high selectivity are possible. It is also receiving phase by the reaction with the stripping
possible to treat any source of wastewater agent. For example, in metal separation from
containing organics and metals even at low con- wastewater, even in very low concentration, the
centration. The primary emulsions of water in oil carrier will selectively combine with the solutes to
emulsion were prepared by emulsifying two form a metal-carrier complex, and the complex
immiscible phases of the stripping solution and will permeate through the membranes from the
organic liquid membrane phase with a surfactant outer to the inner interface. At the inner interface,
E
to produce an emulsion. This primary emulsion is the complex decomposes by the reversal of the
then dispersed in the solution or phase to be equilibrium reaction, and the metal ion is liberated
treated. Mass transfer takes place between the into the internal phase and the regenerated carrier
feed phase and the internal phase through the goes back into the membrane phase. The mecha-
liquid membrane phase. The illustration and sche- nism of removal and recovery of the solute/metal
matic diagram of the process is shown in Fig. 1. assisted by the carrier is illustrated in Fig. 2.
The organic liquid membrane may contain a car- The major problem associated with ELMs in
rier to facilitate the extraction process. The carrier the wastewater treatment is emulsion stability. If
will act as a shuttle to carry the solute from exter- the emulsion globules break and the inner droplet
nal interface to the internal interface or receiving phase spills into the continuous phase, the sepa-
ration is lost. Interfacial shear between the contin-
uous phase and membrane phase causes the liquid
membrane to thin and, in some cases, rupture.
Another problem is osmotic swelling although
it is rarely mentioned in the literature. This phe-
nomenon occurs when water in external phases
diffuses through the organic membrane phase and
Phase I Phase II Phase III swells the inner aqueous droplet phase. The
increased volume of the internal phase leads to
increased breakage and dilution of the concen-
trated solute in the droplet phase.
For the last three decades, this method has
attracted many studies in the area of hydrometal-
lurgy such as separation of metal ions either from
Emulsion Liquid Membrane for Wastewater Treat-
ment, Fig. 1 A schematic diagram of emulsion liquid wastewater or from ores (Othman et al. 2006; Reis
membranes and Carvalho 1993; Reed et al. 1987; Marr 1984;
Emulsion Liquid
Membrane phase
Membrane Phase III Phase II
for Wastewater
Treatment, Fig. 2 The
A A A A
mechanism of couple
transport in emulsion liquid
membrane
Carrier A solute
molecule
686 Emulsion Pertraction Technology for Zinc Recovery
Draxler and Marr 1986; Babcock et al. 1986). It Reed DL, Bunge AL, Noble RD (1987) Inuence of reac-
has been reported that emulsion liquid membrane tion reversibility on continuous-ow extraction by
emulsion liquid membrane. In: Noble RD, Way JD
system has been successfully used to recover cop- (eds) Liquid membranes: theory and applications.
per selectively from waste stream of mine solu- American Chemical Society, Washington, DC
tions (Wright et al. 1995). Reis MTA, Carvalho JMR (1993) Recovery of zinc from
From the viewpoint of practical application in industrial efuent by emulsion liquid membranes.
wastewater treating processes, recovery processes Volkel W, Halwachs W, Schugerl K (1980) Copper extrac-
for copper (Volkel et al. 1980; Goto et al. 1989a), tion by means of a liquid surfactant membrane process.
uranium (Hayworth et al. 1983), and zinc (Marr J Membr Sci 6:1931
1984) were examined in test plant. The rst appli- Wright JB, Nilsen DN, Hundley G, Galvan GJ (1995) Field
test of liquid emulsion membrane technique for copper
cation of emulsion liquid membrane (ELM) on an recovery from mine solutions. Miner Eng 8:549556
industrial scale was the process to remove zinc
from wastewater at a textile plant in Austria
(Draxler and Marr 1986).
In order to get a better understanding of emul-
sion liquid membrane process and the system Emulsion Pertraction Technology
potentials, the extraction performance must be for Zinc Recovery
studied based on the kinetics and thermodynamic
aspects. The parameters that affect the solute Immaculada Ortiz and Eugenio Bringas
extractability and selectivity should be identied. Department of Chemical and Biomolecular
The parameters such as stripping agent types and Engineering, University of Cantabria, Santander,
acidity that control the mass transfer of solute, Cantabria, Spain
volume ratio of emulsion to external phase that
affects the mass transfer area of extraction pro-
cess, and carrier concentration, type of diluents, The emulsion pertraction technology (EPT) is a
swelling, residence time, and agitation rate that separation process that combines the ability of
control the extraction performance and the liquid membranes to promote the uphill transport
breakup rate of emulsion should be studied. of target species by the coupling between mass
transfer and chemical reaction and the benets of
using membrane contactors, namely, large inter-
References facial area, nondispersive contact, and indepen-
dent ow of the uid phases. Figure 1 shows the
Babcock WC, Friesen DT, Lachapelle ED (1986) Liquid ow diagram of the EPT process that comprises
membranes for separating uranium from vanadium and
uranium from molybdenum. J Membr Sci
two essential process units: a microporous hollow
26(3):303312 ber membrane contactor (see Fig. 2) and the
Draxler J, Marr R (1986) Emulsion liquid membranes. Part emulsion vessel that contains a pseudo-emulsion
I: phenomenon and industrial application. Chem Eng consisting of the organic phase formulated with a
Process 20:319329
selective organic carrier and the dispersed strip-
Goto M, Kondo K, Nakashio F (1989) Acceleration effect
of anionic surfactants on extraction rate of copper with ping solution. The target solute is chemically
liquid surfactant membrane containing LIX65N and transferred from the aqueous feed to the organic
nonionic surfactant. J Chem Eng Jpn 22:7998 phase that is embedded in the pores of the hollow
Hayworth HC, Ho WS, Burns WA Jr, Li NN (1983) Extrac-
tion of uranium from wet process phosphoric acid by
bers due to their hydrophobic character. Next,
liquid membranes. Sep Sci Technol 18(6):493521 the solute-carrier complex diffuses to the interface
Marr R (1984) Pilot plant studies of liquid membrane between the organic and the droplet of stripping
separation. Proceeding of Eng. Found. Conf. On New phase where the back-extraction reaction occurs.
Directions in Separation technology, Davos
Othman N, Mat H, Goto M (2006) Separation of silver
The solute is recovered from the internal aqueous
from photographic wastes by emulsion liquid mem- phase after emulsion settling (San Romn
brane system. J Membr Sci 282(12):171177 et al. 2010; Urtiaga et al. 2010; Carrera
Emulsion Pertraction Technology for Zinc Recovery 687
Hollow Fiber Contactor

Feed Solution Rafinate AC A + C
Mass
AC
transfer
A C
Back-extraction
Extraction

A + C AC
Emulsion
Concentrate AC E
Pore
C
Emulsion Tank
Feed A
Solution Hollowfiber
Stripping Emulsion
droplet
Emulsion Pertraction Technology for Zinc Recovery, Fig. 1 Performance of the EPT process
et al. 2009; Bringas et al. 2006; Urtiaga employed in zinc electroplating operations (see
et al. 2005; Klaassen and Jansen 2001). In the composition in Table 1).
forthcoming section, the potential of the EPT Figures 3 and 4 depict the separation and
technology to develop separation processes with recovery objectives to be achieved by the appli-
zinc recovery is evaluated. cation of EPT to the treatment of SPA and SPB.
Under the usual composition of SPA, zinc
forms anionic chlorocomplexes (ZnCl42 and
Recovery of Zinc from Liquid Wastes ZnCl3) while iron is present in the form of neu-
by EPT tral or cationic compounds (Regel et al. 2001).
Tributyl phosphate (TBP) and water are reported
Zincs electropositive nature makes it well-suited as the most suitable extraction and back extraction
for use as a coating for protecting iron and steel reagents enabling the selective separation and
products from corrosion. For this reason, the sur- concentration of zinc with minimum iron extrac-
face treatment industry accounts for almost half tion (Cierpezewski et al. 2002). On the other hand,
zinc modern-day demand. The processes involved the information provided by the equilibrium iso-
in the surface treatment of components are pre- therms depicted in Fig. 5 conrms the commercial
dominantly water-based, and thus the generation selective carrier bis(2,4,4- trimethylpenthyl)
and management of complex liquid wastes is an phosphinic acid (Cyanex272) as a suitable reagent
issue of concern (Bringas et al. 2012). EPT has to formulate the liquid membrane due to its capac-
been proven to be an efcient technology to per- ity to selectively separate Fe3+ and Zn2+ (tramp
form both the selective removal and recovery of ions) from chromium under the typical pH condi-
zinc from different wastes produced in the context tions (1.82.5) of the passivation baths. Sulphuric
of surface treatment industry: (i) spent pickling acid is selected as stripping agent (Urtiaga
acids (SPA) generated in the hot-dip galvanizing et al. 2010).
process (see composition in Table 1) and (ii) spent Figure 6 shows the kinetic results obtained
chromium-based passivation baths (SPB) with the SPA-TBP-water and the EPT process.
Emulsion Pertraction
Technology for Zinc
Recovery,
Fig. 2 Characteristics of
hollow ber contactors for a
bench scale EPT process
Parameter/Characteristic Value/Description
Fiber material Polypropylene
Shell material Polypropylene
Type of fiber X-50
Internal diameter of the

fiber 240 m
Thickness 30 m
Average pore diameter 0.04 m
Porosity 40%
Effective lenght 0.15
Interfacial area 1.4 m2
Number of fibers 10200
Emulsion Pertraction Technology for Zinc Recovery, It is concluded that the extraction (EX) and back-
Table 1 Physical-chemical properties of spent pickling extraction (BEX) percentages of zinc and iron,
acids (SPA) and spent chromium-based passivation baths
(SPB) when steady state conditions (regarding zinc
kinetics) are reached, are respectively, 79 %
SPA SPB
(EX Zn), 98 % (BEX Zn), 23 % (EX Fe), and
Parameter Value Parameter Value
38 % (BEX Fe). Under these operation condi-
pH 0 pH 1.82.5
tions, the maximum value of selectivity of zinc
Zn2+ (mg/L) 122,000 Zn2+ 250011,780
(mg/L) over iron in the stripping solution is 15 kg of
Fe2+ (mg/L) 100,000 Fe3+ 2090 Zn/kg of iron (Ortiz et al. 2004; Samaniego
(mg/L) et al. 2006; Samaniego et al. 2007; Carrera
Cl (mg/l) 300,000 Cr3+ 45009350 et al. 2009; Bringas et al. 2012). On the other
(mg/L) hand, the efciency of the EPT process to carry
Free acidity, 1 NO 67,520
3
out the regeneration of real passivation baths is
H+ (mol/L) (mg/L)
demonstrated in Fig. 7 which shows reductions
Emulsion Pertraction Technology for Zinc Recovery 689
FEED SOLUTION SEPARATION PROCESS SEPARATION AND RECOVERY OBJECTIVES

TBP Raffinate Coagulant (FeCl3)
(FeCl2, HCl)
Spent Pickling Acids Emulsion Pertraction

(Zn2+, HCL, Fe2+) Technology
Electrolytic
recovery of zinc
Stripping solution
Service (ZnCl2, HCl) Galvanizing
E
water
process
Emulsion Pertraction Technology for Zinc Recovery, Fig. 3 Treatment of spent pickling acids by EPTwith zinc recovery
FEED SOLUTION SEPARATION PROCESS/OBJECTIVES RECOVERY OBJECTIVES
Metallic pleces from zinc Regenerated passivation bath

electroplating: Incoming of (Cr3+)
Zn2+ and Fe3+
Spent Passivation Emulsion Pertraction

Bath Technology
Electrolytic
recovery of zinc
Stripping solution
(ZnSO4)
H2SO4 Cyanex272 Electroplating
process
Emulsion Pertraction Technology for Zinc Recovery, Fig. 4 Regeneration of chromium-based passivation baths by
EPT with zinc recovery
Emulsion Pertraction
Technology for Zinc Selective separation
Recovery,
Fig. 5 Extraction 100
iron
isotherms of iron, zinc, and
chromium with Cyanex272
75
% Extraction
50
25
chromium
zinc
0
0 1 2 3 4 5 6 7
Equilibrium pH
Emulsion Pertraction Zn in SPA

Technology for Zinc 120
Recovery, Zn in stripping
Fig. 6 Evolution with time 100 Fe in SPA
of iron (discontinuous line) Fe in stripping
and zinc (continuous line)
CZn - CFe (g L-1)

concentration in the SPA and 80
stripping solution (0.5 L of
SPA with the composition 60
indicated in Table 1; 1 L of
organic solution containing
50 % v/v TBP in Shellsol 40
D70; 1 L of water)
20
0
0 30 60 90 120 150 180
t (min)
Emulsion Pertraction 100

Technology for Zinc
% Extraction
Recovery,
80
Concentration (kg m3)
Fig. 7 Extraction Stripping composition

percentages and
% Extraction
concentrations in the
stripping solution after 3 h 60
of EPT regeneration of SPB
using Cyanex272 as 40
selective extractant
0.43 kg m3
20 0.003 kg m3
0
[Zn] [Fe] [Cr]
in the zinc and iron contents higher than 80 % Future Directions

after 3 h of experimental running with an
almost negligible variation of the chromium Future development will require research and
concentration. Furthermore, zinc is selectively development activities in the following areas:
recovered in the stripping solution (con-
centration >35 kg/m3) being the concentrations 1. Determination of the optimal operational con-
of iron (<0.45 kg/m3) and chromium ditions to: (i) maximize the extraction and
(<0.003 kg/m3) almost negligible (Urtiaga back-extraction kinetics and (ii) achieve max-
et al. 2010; Bringas et al. 2011; Diban imum values of selectivity which permit the
et al. 2011; Bringas et al. 2012). exploitation of the different process streams
Therefore, these results conrm the EPT pro- generated after the application of the EPT
cess as a suitable alternative to perform the treat- process.
ment of spent pickling acids and spent passivation 2. Analysis of the long-term performance of the
baths allowing at the same time the zinc recovery separation process to evaluate the stability of
for further reuse as was illustrated by Figs. 3 the selective carrier which guarantees the pro-
and 4. cess economic viability.
Emulsion Rupture by Membranes 691
References
Emulsion Rupture by Membranes
Bringas E, San Romn MF, Ortiz I (2006) Separation and
recovery of anionic pollutants by the emulsion
Jose Coca
pertraction technology. Remediation of polluted
groundwaters with Cr(VI). Ind Eng Chem Res Department of Chemical and Environmental
45:42954303 Engineering, University of Oviedo, Oviedo,
Bringas E, Mediavilla R, Urtiaga A, Ortiz I (2011) Devel- Spain
opment and validation of a dynamic model for regen-
eration of passivating baths using membrane
contactors. Comp Chem Eng 35:918927
Bringas E, San Romn MF, Urtiaga AM, Ortiz I (2012)
Integrated use of liquid membranes and membrane Emulsions are homogeneous mixtures that consist E
contactors: enhancing the efciency of L-L reactive of a dispersed phase (oil droplets in O/W emul-
separations. Chem Eng Process. doi:10.1016/j. sions) distributed uniformly in a continuous phase
cep.2012.11.005
(water). Ultraltration (UF) and microltration
Carrera JA, Bringas E, San Romn MF, Ortiz I (2009)
Selective membrane alternative to the recovery of zinc (MF) membranes have been used for the treatment
from hot-dip galvanizing efuents. J Membr Sci of O/W emulsions produced in industries such as
326:672680 steel works, metal nishing, pharmaceuticals,
Cierpezewski R, Miesiac I, Regel-Rosocka M,
cosmetics, food, etc. Oil droplets are recovered
Sastre AM, Szymanowski J (2002) Removal of zinc
(II) from spent hydrochloric acid solutions from as retentate and the aqueous phase permeates the
zinc hot galvanizing plants. Ind Eng Chem Res membrane. The production of 1 t of steel may
41:598603 demand up to 200 t of water, the largest water
Diban N, Mediavilla R, Urtiaga A, Ortiz I (2011)
usage corresponding to the rolling mills. Mem-
Zinc recovery and waste sludge minimization from
chromium passivation baths. J Hazard Mater brane performance diminishes when oil droplets
192:801807 fall below the micron range and problems become
Klaassen R, Jansen AE (2001) The membrane contactor: more pronounced when ionic surfactants are pre-
environmental applications and possibilities. Environ
sent, as they increase repulsive forces between
Prog 20:3743
Ortiz I, Bringas E, San Romn MF, Urtiaga AM droplets.
(2004) Selective separation of zinc and iron from Demulsication is commonly achieved by
spent pickling solutions by membrane-based solvent chemical or electrostatic methods. Chemical
extraction. Sep Sci Technol 39:24412455
demulsication destabilizes the disperse phase
Regel M, Sastre AM, Szymanowski J (2001) Recovery
of zinc(II) from HCl spent pickling solutions by two mechanisms: coagulation (inorganic
by solvent extraction. Environ Sci Technol 35: salts) and occulation (organic polymers).
630635 The disadvantage of chemical demulsication
Samaniego H, San Romn MF, Ortiz I (2006) Modelling of
is that the bulk phase has to be further treated
the extraction and back-extraction equilibria of zinc
from spent pickling solutions. Sep Sci Technol before discharge. Electrostatic demulsication
41:757769 is ineffective for emulsions with high water
Samaniego H, San Romn MF, Ortiz I (2007) Kinetics of content and sparking during treatment may
zinc recovery from spent pickling efuents. Ind Eng
generate new compounds from the surfactant
Chem Res 46:907912
San Romn MF, Bringas E, Ibez R, Ortiz I (2010) and the oil. However, membranes can be used
Liquid membrane technology: fundamentals and as coalescers by forcing the emulsion through
review of its applications. J Chem Technol Biotechnol the pores of the membrane (Kajitvichyanukul
85:210
et al. 2011).
Urtiaga A, Abelln MJ, Irabien JA, Ortiz I (2005) Mem-
brane contactors for the recovery of metallic com- Hydrophilic membranes induce coalescence
pounds. Modelling of copper recovery from WPO of O/W emulsions while hydrophobic membranes
processes. J Membr Sci 257:161170 can be used for the demulsication of W/O emul-
Urtiaga A, Bringas E, Mediavilla R, Ortiz I (2010)
sions (Fig. 1). The main factors affecting mem-
The role of liquid membranes in the selective
separation and recovery of zinc for the regeneration brane demulsication are (Daiminger et al. 1995;
of Cr(III) passivation baths. J Membr Sci 356:8895 Hong et al. 2003; Kocherginsky et al. 2003):
692 Emulsion Separation by Membranes
qO/W > 90 resistant, easy to clean by backushing, and

have a low cost and longer life than the previous
Oil generation membranes.
References
Daiminger U, Nitsch W, Plucinski P, Hoffmann S (1995)

Hydrophilic membrane
Novel techniques for oil/water separation. J Membr Sci
99:197203
Hong A, Fane AG, Burford R (2003) Factors affecting
membrane coalescence of stable oil-in-water emul-
qO/W < 90 sions. J Membr Sci 222:1939
Kajitvichyanukul P, Hung Y-T, Wang LK (2011) Mem-
brane technologies for oilwater separation. In: Wang
Oil
LK, Chen JP, Hung Y-T, Shammas NK (eds) Handbook
of environmental engineering, vol 13, Membrane
and desalination technologies. Springer, Dordrecht,
Hydrophobic membrane pp 639668
Kocherginsky NM, Tan CL, Lu WF (2003) Demulsication
of water-in-oil emulsions via ltration through a
Emulsion Rupture by Membranes, Fig. 1 Inuence of
hydrophilic polymer membrane. J Membr Sci 220:
membrane characteristics in oil droplet rejection
117128
Membrane pore size, transmembrane pressure

(DP), and imposed in-pore shear rates affect Emulsion Separation by Membranes
coalescence. The smaller the membrane pore
size and the lower DP, the better Jose Coca
demulsication efciency. However, small Department of Chemical and Environmental
pores coupled with a low DP would lead to Engineering, University of Oviedo, Oviedo,
low permeation ux. Spain
If the DP is below a critical value (DPc), the
emulsion rejection can be maximized. Above a
DPc the membrane acts as a coalescer and the Separation of oily emulsions from wastewaters is
droplets wet the membrane enabling the oil generally carried out by a combination of chemi-
droplets to coalesce. cal and mechanical methods. Chemicals
Membrane demulsication seems to be inde- (coagulants and occulants) are used to destabi-
pendent of the initial disperse phase concentra- lize the oil-water interface allowing the oil drop-
tion and membrane thickness. lets to coalesce. Mechanical methods include
The demulsication process is determined by gravity-based settlers, skimmers, dissolved air o-
interactions between droplets and membrane tation (DAF), centrifuges, electro-coalescers, etc.
surface. Pressure-driven membrane processes
(microltration MF, ultraltration UF, and reverse
So far, membranes as coalescers do not show osmosis RO) may be used to separate oil and
great advantages with respect to conventional water phases. The water molecules move through
coalescing techniques, e.g., ber-bed coalescence the membrane (permeate) and a concentrated
and electrostatic coalescers. New chemically emulsion (retentate or concentrate) is obtained.
treated ceramic membranes and silicon carbide Some components of the emulsion may also
supports have a great potential for oil-water appli- move through the membrane, depending on their
cations since they are abrasion and solvent characteristics and size and the nature of the
Emulsion Treatment with Membranes 693
Oily wastewater Free oil removal
Concentrate
Centrifugal
separator
Equalization UF
Process
tank
tank
Filter
Final concentrate Oil-free permeate
disposal discharge
Settleable solids
E
Emulsion Separation by Membranes, Fig. 1 Hybrid UF system for oily wastewater treatment
membrane, leading to contamination of the per- emulsion) is then transferred to a process tank
meate. Fouling and scaling of the membrane lead and pumped through the UF unit to remove the
to a poor performance and membrane cleaning emulsied oil. The retentate containing the oil is
procedures must be taken into account. recycled to the process tank, and the permeate is
When MF is used for oil-water separation, continuously withdrawn. This process is com-
particulates, emulsied oil, and microbial contam- monly used in the automobile industry (Cheryan
inants can be removed. Cross-ow velocities may and Rajagopalan 1998).
range between 3 and 6 m/s and transmembrane
pressures (TMPs) between 100 and 770 kPa. The
main limitation of membrane processes is the References
decline of the permeate ux over time because
of concentration polarization, membrane fouling Cheryan M, Rajagopalan N (1998) Membrane processing
of oily streams. Wastewater treatment and waste reduc-
(due to surfactant or oil adsorption on the pore
tion. J Membr Sci 151:1328
walls), gel layer formation, pore blocking by oil Coca J, Gutirrez G, Benito JM (2013) Treatment of oily
droplets, pH, and temperature. wastewater by membrane hybrid processes. In: Coca-
Often it is not possible to use a simple mem- Prados J, Gutirrez-Cervell G (eds) Economic sustain-
ability and environmental protection in mediterranean
brane system to perform an oil-water separation:
countries through clean manufacturing methods.
some efuents may cause severe membrane foul- Springer, Dordrecht, pp 3561
ing and pretreatment is necessary to maintain a
high and steady ux. In these situations, inte-
grated-membrane or membrane-based hybrid
processes may be suitable alternatives to obtain Emulsion Treatment
good process performance and to extend mem- with Membranes
brane life (Coca et al. 2013). A typical membrane
hybrid process for oily wastewaters is shown in Jose Coca
Fig. 1. Department of Chemical and Environmental
Usually the process starts with the removal of Engineering, University of Oviedo, Oviedo,
settleable solids and free-oating oil prior to Spain
membrane treatment, mainly UF. This can be
accomplished in a tank with free-oil removal
equipment, such as a skimmer, or by a rotating An emulsion is a homogeneous mixture of two
brush strainer, a pressure or vacuum lter to immiscible liquids, with one of them (typically oil)
remove solids, and a centrifugal separator or a dispersed as droplets into the other. Emulsion drop-
hydrocyclone to remove oil and solids. The lets usually range between 0.1 and 20 mm in diam-
remaining oily wastewater (mainly stable O/W eter. The two main categories of emulsions are
694 Emulsion Treatment with Membranes
Chemical addition
(coagulants/flocculants)
Oily Free oil

wastewater
Centrifugation
Feed UF
Process
tank
tank
Concentrate Oil-free permeate

disposal (to sewer)
Settleable solids
Emulsion Treatment with Membranes, Fig. 1 UF process for oily wastewater treatment
oil-in-water (O/W, >30 % water) and water-in-oil osmosis (RO), occulation followed by
(W/O, <25 % water): water and highly polar liquids microltration (MF), MF, membrane distillation,
are hydrophilic, while nonpolar liquids are consid- nanoltration (NF), and ultraltration
ered oils. Emulsions consist of two phases: an (UF) (Cheryan 1998; Chakrabarty et al. 2010).
internal or discontinuous phase (nally divided A typical UF-based system for oily wastewa-
droplets) and an external or continuous phase ters, operated in a semi-batch recycle mode, is
(which keeps the droplets in suspension), which shown in Fig. 1. The nal concentrate volume
are bound together at the interphase. A surfactant may be only 35 % of the initial oily wastewater
reduces the interfacial tension between the two volume. The system must be cleaned after a cer-
phases binding them together. Industrial oily waste- tain time to restore the permeate ux. O/W emul-
waters can be classied as: free oil (Dp 150 mm), sions may be reduced by 8590 % by volume and
dispersed oil (Dp = 20150 mm), stable emulsied with an oil concentration in the retentate of
oil (Dp 20 mm), and dissolved oil (Dp 5 mm). 7075 %.
Free oils and dispersed O/W emulsions can be MF or UF processes cannot remove dissolved
removed by mechanical methods such as gravity oil components in water. For that purpose, other
settling, skimming, coalescence, centrifugation, methods, such as RO or NF, are required. The
etc. (Alther 1998; Stewart and Arnold 2008). Stable choice of membrane material is important: inor-
emulsied oils and dissolved oils cannot be ganic membranes are chemically robust and
removed efciently by conventional methods expensive; polymeric membranes have lower
because of the small droplet size and low oil resistance to aggressive feeds and are more sus-
concentration. ceptible to fouling, but are considerably cheaper.
Membrane processes are increasingly being In spite of the fact that efuent oil concentrations
applied for treating O/W emulsions due to their of 5 ppm or less can be achieved with membranes,
advantages: high-quality permeate and removal they have not found wide practical applications so
efciency, lower capital costs than with thermal far. Membrane systems suffer from concentration
processes, and compact design. Some of the most polarization and fouling problems that lead to a
promising membrane O/W treatments are dehy- substantial ux decline with time. Membranes
dration of emulsions by pervaporation, reverse must be replaced every 35 years.
Emulsions Drop Size Distribution, Measurement of 695
O/W emulsions can sometimes be treated by a often that they are fast and may be used online or
combination of two membrane processes (e.g., at least without diluting the sample.
UF/NF, UF/RO or NF/RO) to obtain a high- For the measurement of DSD, a physical char-
quality water efuent. acteristic which is connected to the drop size of a
single drop-like sedimentation velocity, diffrac-
tion of light at the drop surface, projected area in
References a microscopic image, etc., is necessary. This
makes clear that the basis of the calculation of
Alther G (1998) Put the breaks on wastewater emulsions. DSD varies between different methods, and one
Chem Eng 105:8288
should not expect the same DSD results from
Chakrabarty B, Ghoshal AK, Purkait MK (2010) Cross- E
ow ultraltration of stable oil-in-water emulsion using different measuring methods. In addition during
polysulfone membranes. Chem Eng J 185:447456 the calculation procedure, it is partly necessary to
Cheryan M (1998) Ultraltration and microltration hand- make assumptions and simplications, and there-
book, 2nd edn. Technomic, Lancaster
fore DSD results are very sensitive to the calcula-
Stewart M, Arnold A (2008) Emulsions and oil treating
equipment: selection, sizing and troubleshooting. Gulf tion parameters which are set by the manufacturer
Professional Publishing, Burlington or which can be set by the user of the equipment.
The most common techniques for DSD mea-
surement are presented in Table 1 which shows the
measuring principle as well as the analyzable drop
Emulsions Drop Size Distribution, size range and some restrictions of the method.
Measurement of Progress in the eld of emulsions has evolved
complex structures, like multiple emulsions
Volker Gaukel, Richard Bernewitz and (Muschiolik 2007; Jimnez-Colmenero 2013).
Heike Schuchmann As in a multiple emulsion, there are more than
Food Process Engineering, Karlsruhe Institute of one DSD to determine the complex structure that
Technology (KIT), Karlsruhe, Germany challenges the common measuring techniques.
However, progress in the eld of DSD determina-
tion of double emulsions has been made. Espe-
The drop size distribution (DSD) of an emulsion cially PFG-NMR and IA of confocal laser
has inuence on the physical stability, the color, scanning microscopy images offer possibilities
and the rheological behavior of the emulsion and is for the characterization of double emulsions
therefore an important means of characterization. (Schuster et al. 2012).
The DSD can be measured with various Another important issue in the context of DSD
methods. In principle one can distinguish between measurements is the illustration and interpreta-
methods which measure physical characteristics tion. DSDs are statistical distributions and can be
of single drops or physical characteristics of the shown as cumulative or density distributions.
bulk emulsion. The latter are often used for the Especially showing the latter, it is very important
characterization of an emulsion in terms of quality to consider all the rules of their calculation. To the
control where it is only necessary to notice differ- authors knowledge, there are many measurement
ences between a reference and a product or where devices with very weak software concerning this
it is sufcient to attain the DSD results only after a point. For simplication and interpretation of
calibration against a system with known DSD. results, it is very common to show only mean
Examples are rheological behavior, dielectric values of the DSD as, e.g., the Sauter Mean Diam-
spectrometry, dynamic scanning calorimetry, eter or statistical values like the median or modal
focused beam reectance, or dynamic reection value. Some insights on this topic are given in
measurements. An advantage of these methods is (Hess 2004; Sommer 2001).
696 Enantiocatalytic Membrane
Emulsions Drop Size Distribution, Measurement of, Table 1 Overview on the most important measurement
techniques for drop size analytics in emulsions
Physical
characteristic for
Method size measurement Size range Additional information
Sedimentation Sedimentation (50 nm) Range for the use of centrifuges in brackets. For
velocity of a 1 mm1 mm small drops, very high dilution necessary
single drop
(Statistical) image Projected area (0,1 nm) Size range depending on the image source.
analytics (IA) 1 mm20 mm Electron microscopy in brackets. Information on
structure and double emulsion detection possible.
High number of drops necessary for reliable
statistical analysis
Laser diffraction (LD) Diffraction of (50 nm) In brackets: with additional light sources and
light at the drop 1 mm2 mm scattering information, necessity of complex
surface refraction index. High dilution necessary
Dynamic laser light Diffusion rate 1 nm1 mm Dispersed phase content up to 10 % possible.
scattering (DLS) Drop sedimentation leads to measurement error
Pulsed eld gradient Coefcient of 0,2100 mm Measurement without dilution possible.
nuclear magnetic diffusion Characterization of some parameters of double
resonance (PFG-NMR) emulsions
References process of a particular enantiomer from a racemic

mixture. The most popular reaction using the
Hess WF (2004) Representation of particle size distribu- technique is esterication, where Candida rugosa
tions in practice. Chem Eng Technol 27(6):624629
lipase or Candida antarctica lipase is the biocat-
Jimnez-Colmenero F (2013) Potential applications of
multiple emulsions in the development of healthy and alyst for the reaction (Giorno et al. 2007; Lau
functional foods. Food Res Int 52(1):6474 et al. 2010). The application of membranes, espe-
Muschiolik G (2007) Multiple emulsions for food use. cially in the production of drugs and other ne
Curr Opin Colloid Interface Sci 12(45):213220
chemicals, has become a trend in recent years and
Schuster S, Bernewitz R, Guthausen G, Zapp J, Greiner
AM, Khler K, Schuchmann HP (2012) Analysis of it will be one of the best techniques for the bulk
W1/O/W2 double emulsions with CLSM: statistical synthesis of such compounds (Lau et al. 2010; Li
image processing for droplet size distribution. Chem et al. 2003). Hollow ber membrane with a par-
Eng Sci 81:8490
ticular molecular weight cutoff is the best choice
Sommer K (2001) 40 years of presentation particle size
distributions yet still incorrect? Part Part Syst Charact to use as the immobilization matrix.
18(1):2225
References
Giorno L, DAmore E, Drioli E, Cassano R, Picci N (2007)

Inuence of -OR ester group length on the catalytic
Enantiocatalytic Membrane activity and enantioselectivity of the free lipase and
immobilized in membrane used for the kinetic resolu-
Mohamad Hekarl Uzir tion of naproxen esters. J Catal 247:194200
School of Chemical Engineering, Universiti Sains Lau SY, Uzir MH, Kamaruddin AH, Bhatia S (2010)
Lipase-catalysed dynamic resolution of racemic ibu-
profen ester via hollow bre membrane reactor: model-
ling and simulation. J Membr Sci 357(1-2):109121
Li N, Giorno L, Drioli E (2003) Effect of immobilization
Enantiocatalytic membrane is a type of mem- site and membrane materials on multiphasic
enantiocatalytic enzyme membrane reactors. In: Li N,
brane immobilized with a particular catalyst Drioli E, Ho W, Lipscomb G (eds) Advanced mem-
(mainly enzyme) onto the shell side for the pur- brane technology. New York Academy of Sciences,
pose of carrying out an in situ reaction-separation New York, pp 436452
Enantiomers 697
discrimination illustrated by high resolution crystal

Enantiomer Discrimination structure of type 4 phosphodiesterase. J Med Chem
49(6):18671873
(Enantioselectivity) Klaholz B, Mitschler A, Belema M, Zusi C, Moras D (2000)
Enantiomer discrimination illustrated by high-resolution
Mohamad Hekarl Uzir crystal structures of the human nuclear receptor hRARg.
School of Chemical Engineering, Universiti Sains Proc Natl Acad Sci USA 97(12):63226327
McNaught AD, Wilkinson A (Eds.), (1997) IUPAC - Com-
Malaysia, Penang, Malaysia pendium of Chemical Terminology, Blackwell Science,
Oxford, England, UK
Molabaasi F, Talebpour Z (2011) Enantiomeric discrimi-
Enantiomer discrimination (or chiral discrimina- nation and quantication of the chiral organophospho-
rus pesticide fenamiphos in aqueous samples by a novel E
tion) is a method used to distinguish between two and selective 13P nuclear magnetic resonance spectro-
enantiomers for the purpose of separating the scopic method using cyclodextrin as chiral selector.
compounds. The terms is also related to enantios- J Agric Food Chem 59(3):803808
electivity where it refers to the discrimination of a
given reactant A when it reacts with two alterna-
tive reactants B and C or in two different ways
Enantiomers
(probably at two different sites) with a reactant
B (McNaught and Wilkinson 1997). A number of
Mohamad Hekarl Uzir
methods have been established to undergo enan-
School of Chemical Engineering, Universiti Sains
tiomer discrimination; these include mass spec-
troscopy (Czerwenka and Lindner 2004;
Czerwenka et al. 2004), nuclear magnetic reso-
nance (NMR) (Gafni et al. 1998; Molabaasi and
Enantiomers are types of stereoisomer whose
Talebpour 2011), and crystal recognition
structures are mirror image of each other, and
(Ballesteros et al. 1995; Huai et al. 2006; Klaholz
they are not superimposable. The property is spe-
et al. 2000). These methods have been specically
cial to most chiral compounds and exists both
developed based on a particular compound with
naturally or from a particular chemical synthesis
its own characteristics of crystal structure.
(Solomons and Fryhle 2004). It represents the
intrinsic property of the building block of life,
which consists of amino acids, sugars, peptides,
References
proteins, and polysaccharides (Maier et al. 2001).
Ballesteros E, Gallego M, Valcarcel M, Grases F (1995)
The simplest compound that exhibits enantio-
Enantiomer discrimination by continuous precipitation. meric property is 2-butanol with the structural
Anal Chem 67(18):33193323 formulae given in Fig. 1.
Czerwenka C, Lindner W (2004) Enantiomer discrimina- The existence of enantiomers of a particular
tion of peptides by tandem mass spectrometry: inu-
ence of the peptide sequence on chiral recognition.
compound can be determined when there is a
Rapid Commun Mass Spectrom 18:27132718 molecule containing one tetrahedral atom with
Czerwenka C, Maier NM, Lindner W (2004) Enantiomer four different groups attached to it, as shown in
discrimination by mass spectrometry: noncovalent the two structures of 2-butanol above (McMurry
interaction of an N-derivatized dipeptide with various
cinchona alkoloid derivatives and comparison with
enantioselective liquid-phase separations. Anal CH3 CH3
Bioanal Chem 379:10391044 HO H
Gafni A, Cohen Y, Kataky R, Palmer S, Parker D (1998) H OH
C C
Enantiomer discrimination using lipophilic cyclodex-
trins studied by electrode response, Pulsed-
GradientSpin-Echo, (PDSE) NMR and relaxation rate C2H5 C2H5
measurements. J Chem Soc Perkin Trans 2(1):1924
Huai Q, Sun Y, Wang H, MacDonald D, Aspiotis R, Enantiomers, Fig. 1 Structural formula of 2-butanol
Robinson H, Huang Z, Ke H (2006) Enantiomer showing two different enantiomers
698 Enantiomers Production by Membrane Operations
2004). Specialized drugs are mainly made of com- References

pounds that exhibit enantiomeric properties; these
include (S)-ibuprofen, (S)-naproxen, (S)- Giorno L, Li N, Drioli E (2003) Use of stable emulsion to
improve stability, activity and enantioselectivity of
citalopram, and (+)-norcisapride. As can be appar-
lipase immobilized in a membrane reactor. Biotech
ently seen, only one type of enantiomer acts as an Bioeng 84(6):677685
active compound compared to their Giorno L, DAmore E, Drioli E, Cassano R, Picci N (2007)
corresponding structures. This shows that the syn- Inuence of -OR ester group length on the catalytic
activity and enantioselectivity of the free lipase and
theses of such drugs are somewhat important, in
immobilized in membrane used for the kinetic resolu-
particular, for use in pharmaceutical and medical tion of naproxen esters. J Catal 247:194200
sectors. Some of the processes carried out to Lau SY, Uzir MH, Kamaruddin AH, Bhatia S (2010)
obtain the compounds are tedious and require a Lipase-catalysed dynamic resolution of racemic ibu-
profen ester via hollow bre membrane reactor: model-
number of reagents as well as different catalysts.
ling and simulation. J Membr Sci 357(12):109121
This will then lead to high cost of production Lau SY, Fadzil NG, Kamaruddin AH, Bhatia S, Uzir MH
(Maier et al. 2001; Tramper 1996). (2011) Conceptual design and simulation of a plant for
the production of high purity (S)-ibuprofen acid using
innovative enzymatic membrane technology. Chem
Eng J 166(2):726737
References
Maier NM, Franco P, Lindner W (2001) Separation of

enantiomers: needs, challenges, perspectives.
J Chromatogr A 906(1):333
McMurry J (2004) Organic chemistry. Thomson Books/ Enantiomers Separation by
Cole, Belmont Membrane Operations
Solomons TWG, Fryhle CB (2004) Organic chemistry.
Wiley, New Jersey
Mohamad Hekarl Uzir
Tramper J (1996) Chemical versus biochemical conver-
sion: when and how to use biocatalysts. Biotechnol School of Chemical Engineering, Universiti Sains
Bioeng 52(1):290295 Malaysia, Penang, Malaysia
Enantiomers separation is a process of separat-

Enantiomers Production by ing isomers in a racemic mixture into their
Membrane Operations individual enantiomer. The use of membranes in
chiral separation can be divided into two
Mohamad Hekarl Uzir main categories, namely, adsorption-type
School of Chemical Engineering, Universiti Sains enantioselective membranes and membrane-
Malaysia, Penang, Malaysia assisted resolution with non-enantioselective
solid membranes (Pirkle and Bowen 1994; Xie
et al. 2008). The former is entirely based on the
Enantiomer production is a process of forming transport mechanism with an aid of chiral carriers.
enantiomer of interest through membrane separa- The work of Gumi and co-workers reported a
tion. The current method of kinetic resolution and detail account of the transfer of the (S)-
in situ membrane separation is a highly effective propranolol attached to the N-hexadecyl-L-
process which requires less energy and short reac- hydroxyproline through the formation of ionic
tion time (Lau et al. 2011). Most of the works pair, which is then being transported through the
related to product separations involved lipase polysulfone-based membrane (Gumi et al. 2005).
enzyme immobilized onto the membrane matri- The latter is a technique mostly coupled with
ces. These include the production of naproxen enzymes as biocatalysts. The enzyme is initially
ester (Giorno et al. 2003, 2007) and ibuprofen immobilized onto the shell side of the membrane
ester (Lau et al. 2010). matrix before carrying out the required resolution,
Enantioselective Separations, Membrane Operations 699
either kinetic resolution or dynamic kinetic resolu- diffusion, but provides highly selective sorption
tion. For example, the work of Long and (Hadik et al. 2002; Higuchi et al. 2010).
co-workers successfully utilized Candida rugosa
lipase as the source of biocatalyst for the kinetic
resolution of (S)-ibuprofen ester (Long et al. 2005). References
Ceynowa J (1998) Separation of racemic mixtures by

membrane methods. Chem Anal 43(6):917933
References Hadik P, Szabo L, Nagy E (2002) D, L-lactic acid and D,
L-alanine enantio separation by membrane process.
Gumi T, Ferreira Q, Viegas R, Crespo J, Coelhoso I,
Palet C (2005) Enantioselective separation of propran-
Desalination 148(13):193198
Higuchi A, Tamai M, Ko Y-A, Tagawa Y-I, Wu Y-H,
E
olol by chiral activated membranes. Sep Sci Technol Freeman B, Bing J-T, Chang Y, Ling Q-D (2010) Poly-
40(4):773789 meric membranes for chiral separation of pharmaceuti-
Long W, Kow P, Kamaruddin A, Bhatia S (2005) Compar- cals and chemicals. Polym Rev 50(2):113143
ison of kinetic resolution between two racemic ibupro- Xie R, Chu L-Y, Deng J-G (2008) Membranes and mem-
fen enters in an enzymatic membrane reactor. Process brane processes for chiral resolution. Chem Soc Rev
Biochem 40(7):24172425 37:12431263
Pirkle W, Bowen W (1994) Preparative separation of enan-
tiomers using hollow-bre membrane technology. Tet-
rahedron Asymmetry 5(5):773776
Xie R, Chu L-Y, Deng J-G (2008) Membranes and mem-
brane processes for chiral resolution. Chem Soc Rev Enantioselective Separations,
37:12431263
Membrane Operations
Mohamad Hekarl Uzir

Enantioselective Membrane Malaysia, Penang, Malaysia
Mohamad Hekarl Uzir

School of Chemical Engineering, Universiti Sains Enantioselective separations refer to separation
Malaysia, Penang, Malaysia methods used to separate racemic mixture into
individual enantiomer with the aid of membrane
matrix/module. In order to measure the level of
Enantioselective membrane is a type of membrane selectivity, the percent enantiomeric excess (% ee)
used to separate chiral compounds into their indi- of a mixture needs to be determined (see enantios-
vidual enantiomer. The usual type of membrane electivity). Membranes have been successfully
used for separation purposes is of polymeric type used in chiral separation with various types as
(Ceynowa 1998; Higuchi et al. 2010; Xie well as with different abilities: direct separation
et al. 2008). The mechanism of chiral separation type of membrane made from polymer/liquid and
can be categorized into two different types of separation with nonselective membrane assisted
membranes, namely, diffusion-selective mem- by chiral carriers/support (Maier et al. 2001;
branes and sorption-selective membranes. Pirkle and Bowen 1994).
Diffusion-selective type of membranes are basi-
cally made of polymer without any specic chiral References
selectors (i.e., the polymer itself has chiral prop-
erties); however, sorption-selective membranes Maier NM, Franco P, Lindner W (2001) Separation of
can be designed in such a way that a particular enantiomers: needs, challenges, perspectives.
J Chromatogr A 906(1):333
chiral selector can be immobilized can be
immobilized onto the surfaces of the polymer tiomers using hollow-bre membrane technology. Tet-
membranes, which later gives less selective rahedron Asymmetry 5(5):773776
700 Enantioselective Synthesis
both chemical and biological catalyses utilizing

Enantioselective Synthesis Zr for the carboalumination reaction and later
applying lipase originated from Pseudomonas
Mohamad Hekarl Uzir cepacia for the acetylation reaction. However,
School of Chemical Engineering, Universiti Sains by comparing both types of reactions, biocatalysis
Malaysia, Nibong Tebal, Penang, Malaysia has been known to provide better yield as well
as %e.e (Tsai and Dordick 1996). One of the
recent works in the development of pharmaceuti-
cally active compounds has been reported by Wen
Enantioselective synthesis is a chemical/bio- and colleagues where the group managed to syn-
chemical reaction where only one enantiomer of thesize (S)-propranolol through a coupled reac-
a chiral product is preferentially formed. The syn- tion with kinetic resolution as a recycling route
thesis can be divided into two different types, utilizing Candida antarctica lipase B. The result
namely, catalytic synthesis and biocatalytic trans- showed an improve %e.e up to 100 % (Wen
formation. Nowadays, chemical catalysis has et al. 2014).
been continuously improved, especially in terms
of yields and percentage enantiomeric excesses
(%e.e), for instance, the synthesis of spirocyclic References
oxindolo-b-lactam, a combination of indole and
b-lactam, which acts as antibacterial, antiviral, Tsai S-W, Dordick JS (1996) Extraordinary enantiospe-
cicity of lipase catalysis in organic media induced by
and antifungal agents (Zhang et al. 2014). The
purication and catalyst engineering. Biotechnol
structure of spirocyclic oxindolo-b-lactam or Bioeng 52(2):296300
commercially known as chartelines is given in Wen Y, Hertzberg R, Gonzalez I, Moberg C (2014) Minor
Fig. 1. enantiomer recycling: application to enantioselective
synthesis of beta-blockers. Chem-A Eur J 20(13):
Zhang and coworkers reported that the con- 38063812
struction of chartelines started from the bifunc- Xu S, Oda A, Kamada H, Negishi E (2014) Highly
tional N-heterocyclic carbine undergoing enantioselective synthesis of g-, d-, and e-chiral 1-
Staudinger reaction of ketenes with isatin-derived alkanols via Zr-catalyzed asymmetric carboalu-
mination of alkenes (ZACA)-Cu- or Pd-catalyzed
ketimines. For such a highly complex compound,
cross-coupling. Proc Natl Acad Sci U S A 111(23):
the %e.e reached up to 95 % with the yield of 83688373
89 %. Similar goes to the synthesis of g-, d-, and Zhang H, Gao Z, Ye S (2014) Bifunctional N-heterocyclic
e-chiral-1-alcoloids, reported by Xu and carbene-catalyzed highly enantioselective synthesis of
spirocyclic oxindolo-beta-lactam. Org Lett 16(11):
coworkers (2014).The reaction was based on
30793081
Cl
R1
Enantioselective Synthesis by
R2
N
N
Membrane Operations
Br Mohamad Hekarl Uzir
N School of Chemical Engineering, Universiti Sains
N H
Br Malaysia, Penang, Malaysia
Compound A: R1=R2=Br
Compound B: R1=Br, R2=H
Compound C: R1=R2=H
Enantioselective synthesis by membrane opera-
Enantioselective Synthesis, Fig. 1 Chemical structure tions refers to chemical reactions to produce one
of spirocyclic oxindolo-b-lactam enantiomer from a racemic mixture utilizing
Enantioselective Transport of Amino Acids by Membrane Operations 701
membrane as a medium of separation. The process are the structural units of proteins. The most of
occurs in situ, where the product formed from the amino acids found in nature are categorized as
reaction will be directly separated into the a-amino acid in which the amino group is attached
required enantiomeric compound (Giorno to the carbon atom adjacent to the carboxylic acid
et al. 1995; Long et al. 2005). Another enantio- group. An a-amino acid can exist in either of two
meric synthesis, which applied the hollow-ber optical isomers except glycine, and the isomer is
membrane module, is from the work of Hadik and normally described as L- or D-amino acid. The major-
co-workers. The polymeric membrane made ity of amino acids in living organisms are L-isomers.
from polypropylene was used together with a Optical isomers, including amino acids, frequently
chiral selector, N-3,5-dinitrobenzoyl-L-alanine show different bioactivity from each other. So, opti-
E
octylester, initially dissolved in toluene (Hadik cal resolution of amino acids has been an important
et al. 2002). Hollow-ber membrane has also subject in various industries dealing with pharma-
been applied with different types of chiral selec- ceuticals, foods, agrochemicals, and so on.
tors for large-scale separation processes. The Optical resolution with membrane processes has
reported results suggested that N(1-naphthyl)- potential merits over other separation methods like
leucine provided a wide range of enantiomeric crystallization or chromatography. Membrane oper-
separation of amino acid derivatives with %ee of ation can be carried out continuously and it enables
up to 95 % (Pirkle and Bowen 1994). the large-scale separation. Optical resolution of
amino acids can be achieved by the function of chiral
recognition sites of solid membranes and liquid
References membranes. The method using liquid membranes
is described in another site, so solid membranes for
Giorno L, Molinari R, Drioli E, Bianchi D, Cestic P (1995) enantioselective separation are explained here.
Performance of a biphasic organic/aqueous hollow
A main research topic on enantioselective mem-
bre reactor using immobilized lipase. J Chem Technol
Biotechnol 64(4):345352 branes is the design of the membranes that have the
Hadik P, Szabo L, Nagy E (2002) D, L-lactic acid and D, chiral recognition sites. Some enantioselective solid
L-alanine enantio separation by membrane process. membranes have been reported as follows.
Desalination 148(13):193198
Long W, Kow P, Kamaruddin A, Bhatia S (2005) Compar-
ison of kinetic resolution between two racemic ibupro-
fen enters in an enzymatic membrane reactor. Process Membranes Prepared from Chiral
Biochem 40(7):24172425 Polymers
tiomers using hollow-bre membrane technology. Tet-
rahedron Asymmetry 5(5):773776 Polypeptide and polysaccharide, like cellulose,
alginate, and chitosan, and their derivatives are
chiral polymers, and membranes from these
chiral polymers have been applied to the
Enantioselective Transport of Amino enantioselective separation of racemic amino
Acids by Membrane Operations acids. The separation factor of these membranes
toward racemic tryptophan is usually in the range
Keiji Sakaki of 1.15.0 (Higuchi et al. 2010). Membranes pre-
National Institute of Advanced Industrial Science pared from polymers with chiral branch have been
and Technology (AIST), Research Institute for also reported.
Sustainable Chemistry, Tsukuba, Ibaraki, Japan
Membranes Added with Chiral Selectors

Amino acids are molecules that contain an
amino group (NH2), a carboxylic acid group Enantioselective membranes can be prepared by
(COOH), and a side chain in each structure and the addition of chiral selectors that show
702 Enantioselectivity
enantioselective afnity toward amino acids. Yoshikawa M, Izumi J, Kitano T (1995) Molecularly
Cyclodextrins, bovine serum albumin (BSA), imprinted polymeric membranes for optical resolution.
and DNA have been used as chiral selectors. The
membranes added with these chiral selectors
showed the separation factor of 1.12.7 toward
amino acids (Higuchi et al. 2010).
Enantioselectivity
Mohamad Hekarl Uzir

Molecular Imprinting Membranes
(MIMs)
Molecular imprinting membranes are fabricated
by incorporating the template molecules into the
Enantioselectivity is the preferential formation of
membranes and then removing the template mol-
a chemical reaction (mainly with metal complexes
ecules. The formed cavities in the membranes
catalysts) of one enantiomer compound over the
show the specic afnity with the template mole-
other (McNaught and Wilkinson 1997). The
cules and their analogue (Yoshikawa et al. 1995).
extent of enantiomeric reactions is determined
Although MIMs currently show low enantios-
by the percentage enantiomeric excess (% e.e)
electivity, this method has the potential of higher
given by;
enantioselectivity (Xie et al. 2008).
Solid enantioselective membranes can be
SR
divided into two classes: diffusion-selective mem- % e:e 100
SR
branes and sorption-selective membranes (van der
Ent et al. 2001; Xie et al. 2008). A diffusion- References
selective membrane is dened as a membrane
with no specic chiral selectors for the chiral McNaught AD, Wilkinson A (Eds.), (1997) IUPAC - Com-
interaction but consists of a chiral polymer. Dif- pendium of Chemical Terminology, Blackwell Science
fusion selectivity is caused by the chiral discrim- Ltd., Oxford, England, UK
ination during diffusion. On the other hand, in a
sorption-selective membrane, chiral selectors are
embedded in a polymer matrix. To reach perme-
ation selectivity in sorption-selective membranes, Enantiospecificity
the selectively adsorbed population has to be
mobile. An electrical potential is sometimes used Mohamad Hekarl Uzir
for the purpose. School of Chemical Engineering, Universiti Sains
References
Enantiospecicity is the ability of an enzyme
Higuchi A, Tamai M, Ko Y, Tagawa Y, Wu Y, Freeman B, to convert a racemic mixture into a particular
Bing J, Chang Y, Ling Q (2010) Polymeric membranes enantiomer. The degree of specicity is interre-
for chiral separation of pharmaceuticals and chemicals. lated with that of selectivity, which equation
Polym Rev 50:113143
suggested by Lopez and coworkers (Lopez
van der Ent EM, vant Riet K, Keurentjes JTF, van der Padt
A (2001) Design criteria for dense permeation-selective et al. 1990):
membranes for enantiomer separations. J Membr Sci
185:207221 " 1 0:5 #2
Xie R, Chu L, Deng J (2008) Membranes and membrane tanh F=E0:5 E =F
Eo E 2
processes for chiral resolution. Chem Soc Rev tanhF1 F1
37:12431263
Encapsulation 703
where Complex coacervation: Coacervation caused

by the interaction of two oppositely charged

ln 1 c 1 eep macromolecules.
E
Internal phase separation: Phase separation

ln 1 c 1 eep
within emulsion droplets consisted of a mixture
and of polymer, good solvent and poor solvent, usu-
ally triggered by removal of good solvent.
Micelles: Spherical aggregates of surfactant
vobs R 2
F molecules dispersed in an aqueous solution.
Deff So 3
Liposomes: Spherical vesicles whose shells
E
consist of single or multiple concentric bilayers
and c is the conversion of the particular
resulting from the self-assembly of phospholipids
reaction (Tsai and Dordick 1996). If the calculated
in an aqueous solution.
intrinsic enantioselectivity is extremely high,
Colloidosomes: Shperical vesicles whose shells
therefore, the reaction can be considered as
consist of coagulated or fused colloid particles.
enantiospecic.
Polymersomes: Shperical vesicles whose
shells consist of amphiphilic block copolymers.
Amphiphilic molecules: Surface active mole-
References
cules made up of two parts, a polar or electrically
Lopez JL, Wald SA, Matson SL, Quinn JA
charged hydrophilic part and a hydrophobic part,
(1990) Multiphase membrane reactors for separating most often an alkyl chain.
stereoisomers. Ann N Y Acad Sci 613(Enzyme Engi- Polycondensation: A polymerization in which
neering 10):155166 the growth of polymer chains proceeds by con-
Tsai S-W, Dordick JS (1996) Extraordinary enantiospe-
cicity of lipase catalysis in organic media induced by densation reactions between molecules of all
purication and catalyst engineering. Biotechnol degrees of polymerization.
Bioeng 52(2):296300 Layer-by-layer polyelectrolyte deposition: A
thin lm fabrication technique based on deposit-
ing alternating layers of oppositely charged poly-
electrolytes with wash steps in between.
Encapsulation Encapsulation is a process of enclosing or
entrapping a core material (liquid, gas, solid par-
Goran T. Vladisavljevi and Richard G. Holdich ticles, cells, dissolved active ingredients, etc.)
Department of Chemical Engineering, inside a solid shell or within a solid or liquid
Loughborough University, Loughborough, matrix for the purpose of controlled or triggered
Leicestershire, UK release, immobilization, isolation, or protection of
the encapsulated material. The material being
encapsulated is called the core material, and the
Synonyms carrier material used for envelopment or entrap-
ment is called the shell material. Typical examples
Encapsulation: entrapping, enclosing, enveloping; of core materials are ink or dye for carbonless
Amphiphilic substances: Amphiphiles copy papers, liquid crystals for microparticle-
based displays, phase-change material for smart
textiles, high-molecular-weight gases for ultra-
Definitions sound contrast imaging, genetic material for
in vitro compartmentalization, active food ingre-
Encapsulation: Coating or entrapping of active dients for functional food products, enzymes for
ingredient inside a solid shell or within a liquid biocatalytic reactors, drugs, pesticides, fra-
or solid matrix of another material. grances, antimicrobial agents, etc.
704 Encapsulation
a Core/shell b Matrix type deliberately compromising the integrity of the

shell, e.g., by exposing the microcapsule to envi-
ronmental stresses such as mechanical forces,
change of pH, temperature, or ionic strength, or
by chemical or biochemical degradation of the
shell. A special class of core/shell capsules are
micelles and vesicles (liposomes, polymersomes,
and colloidosomes), formed by self-assembly of
amphiphilic molecules (phospholipids and
c Multiple cores d Multiple shells diblock copolymers) or particles (Dinsmore
et al. 2002). Micelles are formed spontaneously
when amphiphiles are dispersed in a polar solvent
at concentrations that exceed a critical level,
known as the critical micelle concentration
(CMC). Micelles containing solubilized materials
are referred to as microemulsions. A difference
between micelles and vesicles is that vesicles are
Encapsulation, Fig. 1 Schematic diagrams of capsules
not a thermodynamically stable state of amphi-
with different morphologies philes and do not form spontaneously, e.g., with-
out input of external energy (Tadros 1993).
In the matrix-type capsule (Fig. 1b) the core
Encapsulation efciency (EE) is dened as the material is distributed uniformly throughout a
percentage of core material incorporated into the matrix of shell material. The most common
microcapsules relative to the total amount of the release pattern from matrix-type capsules is rst
core material added during encapsulation process: order in which the release rate decreases exponen-
EE = (mE/mT) 100 %, where mE is the mass of the tially with time until the active ingredient is
core material incorporated and mT is the total mass exhausted. Typical matrix-type capsules are mul-
of the core material added. Loading capacity tiple emulsions, hydrogel particles, solid lipid
refers to the percentage of core material incorpo- particles, polymeric particles, etc. Hybrid struc-
rated within the microcapsules relative to the total tures consisting of a number of hierarchically
mass of the microcapsules (i.e., core + shell assembled homogenous phases may be
material). engineered that allow for even ner control over
Two main types of capsules can be distin- the functionality of microcapsules (Fig. 1c, d).
guished, core/shell (reservoir) type and matrix
type. In the core/shell capsule (Fig. 1a), the shell
material completely surrounds and contains an
internal core material (the internal phase). The References
strategies used for formation of shell are spray
Dinsmore AD, Hsu MF, Nikolaides MG, Marquez M,
coating, complex coacervation, polymer precipi- Bausch AR, Weitz DA (2002) Colloidosomes: selec-
tation by internal phase separation, interfacial tively permeable capsules composed of colloidal parti-
reaction that may include polycondensation or cles. Science 298:10061009
cross-linking, and layer-by-layer polyelectrolyte Tadros TF (1993) Industrial applications of dispersions.
Adv Colloid Interface Sci 46:147
deposition (Yow and Routh 2006). The core mate-
Yow HN, Routh AF (2006) Formation of liquid
rial can be released from core/shell capsules by corepolymer shell microcapsules. Soft Matter
simple molecular diffusion through the shell or by 2:940949
Encapsulation Application 705
and fragrances, textiles, paper, paints, coatings

Encapsulation Application and adhesives, toner applications, and many
other industries.
Goran T. Vladisavljevic Carbonless copy paper developed by Green
Department of Chemical Engineering, and Schleicher in the 1950s was the rst commer-
Loughborough University, Loughborough, cial product to employ microcapsules. A coating
Leicestershire, UK of microencapsulated colorless ink is applied to
the top sheet of the paper, and a developer is
applied to the subsequent sheet. When pressure
Synonyms is applied by writing, the capsules break and the
E
ink reacts with the developer to produce the dark
Bichromal particles synonym: Two-colored color of the copy. Paper-like displays with low
particles; Microencapsulation synonym: power consumption such as rotating bichromal
Microentrapment microspheres system and microencapsulated elec-
trophoretic system are other examples of commer-
Definitions cial applications of microencapsulated ink
(Yoshizawa 2004). In the rotating bichromal sys-
Carbonless Copy Paper - Paper coated on the back tem (Gyricon), bichromal capsules with oppo-
side with colorless dye or ink capsules, that burst sitely charged hemispheres are free to rotate
under the pressure of writing or typing and react within oil-lled cavities. In the microencapsulated
with colorless clay particles coated on the upper electrophoretic display system (E ink), oppositely
surface of the second sheet.; Phase Change Mate- charged black and white particles move under an
rial - A substance with a high heat of fusion that applied electric eld in a clear liquid encapsulated
absorb or release a high amount of energy during within a transparent capsule.
melting of freezing, thus acting as a thermal Todays textile industry makes use of
energy storage; Electrophoretic display - a display microencapsulated materials to enhance the
that forms images by rearranging charged pigment properties of nished goods. One application
particles with an applied electric eld. increasingly utilized is the incorporation of
microencapsulated phase change materials
Microencapsulation can be done: (PCMs), such as parafn wax. Phase change
materials absorb and release heat in response to
1. To protect the encapsulated material against changes in environmental temperatures. When
oxidation or deactivation due to reactions temperatures rise, the phase change material
with reactive species from the environment. melts, absorbing excess heat, and feels cool. Con-
2. To mask the organoleptic properties like color, versely, as temperatures fall, the PCM releases
taste, and odor of the actives. heat as it solidies, and feels warm. Microencap-
3. To achieve controlled/triggered/targeted sulation is also used in thermochromic and pho-
release. tochromic fabrics, which change color with
4. For safe handling of toxic materials. changes in temperature or light, insect-repellent
5. To achieve in vitro compartmentalization or fabrics, which ward off mosquitoes, and scented
immobilization of biological materials and fabrics, which release fragrance (Nelson 2002).
catalysts. Pesticides are encapsulated to be released over
time, allowing farmers to apply the pesticides less
Microencapsulated materials are utilized in often rather than requiring very highly concen-
agriculture, pharmaceuticals, foods, cosmetics trated and perhaps toxic initial applications
706 Encapsulation Efficiency
followed by repeated applications to combat the

loss of efcacy due to leaching, evaporation, and Encapsulation Efficiency
degradation.
Ingredients in foods are encapsulated for sev- Emma Piacentini
eral reasons (Gouin 2004). Most avorings are Institute on Membrane Technology, National
volatile; therefore, encapsulation of these compo- Research Council of Italy, ITM-CNR, Rende,
nents extends the shelf life of products by Italy
retaining the food avors within that would oth-
erwise evaporate out and be lost. Some ingredi-
ents are encapsulated to mask taste, such as The encapsulation efciency (EE%) is dened by
nutrients added to fortify a product without the concentration of the incorporated material
compromising the products intended taste. Alter- (such as active ingredients, drugs, fragrances, pro-
natively, avors are sometimes encapsulated to teins, pesticides, antimicrobial agents, etc.)
last longer, as in chewing gum. Some food ingre- detected in the formulation over the initial con-
dients must be encapsulated to be protected from centration used to make the formulation.
oxidation or other degradation reactions caused Encapsulation efciency (EE %) was calcu-
by exposure to light, moisture, or oxygen. Micro- lated using below formula:
encapsulation preserves lactic acid bacteria, both
starters and probiotics, in food and during the EE % Wt =Wi 100%
passage through the gastrointestinal tract, and
may contribute to the development of new func- where Wt is the total amount of the incorporated
tional foods. material and Wi is the total quantity of incorpo-
Many drug formulations for oral, intravenous, rated material added initially during the prepara-
ocular, and subcutaneous administration are tion. Wt and Wi can be determined using
microencapsulated to achieve controlled, spectroscopic or chromatographic method. If the
targeted, or triggered release of active ingredients. capsule shell material is a polymer, it can be
Aspirin, for example, can cause peptic ulcers and dissolved in the solvent, and the incorporated
bleeding if doses are introduced all at once. molecule will get soluble and it can be quantied.
Microencapsulation of cells and enzymes is If the incorporated molecule is not soluble in that
used to improve efciency of bioreactors since solvent, it can be extracted by adding the capsules
very high volumetric productivity can be in a liquid in which the target molecule is soluble
achieved; encapsulated biocatalysts typically (also by multiple extraction). If the core material is
have greater thermal and operational stability a liquid (such as emulsions), the amount of the
and downstream processing is simplied, since encapsulated material can be evaluated after
the encapsulated biocatalyst can easily be recov- induced separation of the liquid dispersed phase
ered and reused. In molecular biology, single-cell and liquid continuous phase (i.e., simple emul-
encapsulation is used to achieve high-throughput sions) or in the outer liquid phase (i.e., W2 in
screening in directed evolution experiments. water-in-oil-in-water (W1/O/W2) emulsions).
The amount of water retained within the oil drop-
lets during emulsication is also signicant for
double emulsion. The methods used in this partic-
References
ular case enclose the measure of the outer water
Gouin S (2004) Microencapsulation: industrial appraisal of phase conductivity by differential scanning calo-
existing technologies and trends. Trends Food Sci rimetry (DSC) (Schuch et al. 2013).
Technol 15:330347 The encapsulation efciency can be inuenced
Nelson G (2002) Application of microencapsulation in
by (i) the partition coefcient of the target mole-
textiles. Int J Pharm 242:5562
Yoshizawa H (2004) Trends in microencapsulation cule in the solvents used in the preparation of the
research. KONA 22:2331 formulation, (ii) the method used to carry out the
Encapsulation Techniques 707
encapsulation process (temperature, pH, mechan- Encapsulation Techniques, Table 1 Classification of

ical stress), and (iii) the size distribution of the dispersion techniques used in encapsulation processes
capsules (Jyothi et al. 2010). Liquid/air dispersion Liquid/liquid dispersion
Atomization Emulsication
Pressure nozzle High pressure
References homogenizers
Two-uid nozzle Ultrasound homogenizers
Jyothi NVN, Prasanna PM, Sakarkar SN, Prabha KS, Spinning disc Static mixers
Ramaiah PS, Srawan GY (2010) Microencapsulation Dripping/jet breakup Rotor/stator devices
techniques, factors inuencing encapsulation ef- Simple dripping Microuidic devices
ciency. J Microencapsul 27:187197 Electrostatic Membrane emulsication E
Schuch A, Khler K, Schuchmann HP (2013) Differential extrusion
scanning calorimetry (DSC) in multiple W/O/W emul-
Coaxial air/liquid Microchannel
sions, a method to characterize the stability of inner
ow emulsication
droplets. J Therm Anal Calorim 111:18811890
Jet cutting Inkjet printing
Centrifugal nozzle Micellization
Vibrating nozzle
Encapsulation Techniques
Encapsulation Techniques, Table 2 Common methods
Goran T. Vladisavljevic of solid shell/matrix formation in encapsulation processes
Department of Chemical Engineering, Mechanical/
Loughborough University, Loughborough, thermal Physicochemical Chemical
Leicestershire, UK Cooling Solvent removal Suspension
polymerization
Freezing Evaporation or One stage
drying (direct)
The encapsulation technique of choice depends on
Two stage
the type and physical properties of the core and (droplet swelling)
shell material. The chosen encapsulation tech- Pan coating Liquid Interfacial
nique should give a high encapsulation efciency extraction polycondensation
and loading capacity of actives, capsules should Fluidized Layer-by-layer Sol-gel chemistry
not exhibit aggregation or adherence, capsules bed coating deposition
Top spray Self-assembly
should have a narrow particle size distribution
Bottom Simple/complex
without tails, threads, or dents on the surface, spray coacervation
and the process should be suitable for industrial Tangential Ionotropic
scale production. spray gelation
Depending on the initial physical state of the Wurster Internal phase
core phase, two different fabrication routes can be process separation
distinguished: (i) coating solid particles by shell-
forming material in a uidized bed or pan coater
and (ii) dispersing a capsule-forming material in evaporation, cooling, or cross-linking in a hard-
the form of droplets in another immiscible liquid ening bath. Emulsication routes involve emulsi-
or air, followed by droplet solidication. There is cation of a solution or suspension of actives,
a variety of atomization and dripping processes by followed by shell/matrix formation by internal
which a liquid phase can be dispersed in another gelation, polymerization, layer-by-layer electro-
immiscible uid (Table 1). static deposition, internal phase separation, coac-
Depending on the chemical composition of the ervation, etc. (Table 2).
shell-forming material, solidication of air Hydrogel capsules contain a hydrophilic active
suspended droplets can be achieved by solvent entrapped within a hydrophilic polymer network
708 Encapsulation: Entrapping, Enclosing, Enveloping
that can absorb and hold large amount of water. is emulsied with an aqueous phase at room tem-
A gel network can be formed by chemical gelation perature. The solid particles are then formed by
(polymerization by free radical processes or via evaporation of the organic solvent. Hydrophilic
condensation) or by physical gelation, which can actives can be encapsulated by forming a W/O/W
involve heating (heat-setting gels), cooling (cold- emulsion prior to solvent evaporation or cooling
setting gels), or the addition of multivalent coun- (Jaspart et al. 2005).
terions (ionotropic gelation). In the internal Liposomes are usually prepared using two
ionotropic gelation, the droplets of W/O emulsion approaches: (i) hydration of dry lipid lms with
contain a gel-forming polymer (e.g., alginate) and an aqueous solution of actives resulting in forma-
a cross-linking agent in a nondissociated form tion of large multilamellar vesicles (MLV), which
(e.g., calcium carbonate), whereas the continuous is then followed by size reduction of MLVs by
oil phase contains a species (e.g., hydrogen ions) sonication, microuidization, repetitive freezing
that diffuses into the droplets and triggers the and thawing, or extrusion through track-etch
release of cross-linking agent in its active form polycarbonate membranes, and (ii) mixing a
and subsequent gelation. In the external nonaqueous lipid solution with an aqueous solu-
ionotropic gelation, the droplets initially contain tion (Walde and Ichikawa 2001).
only a gel-forming polymer, and cross-linking
occurs in a hardening bath (Zhang et al. 2007).
Coacervation involves the phase separation of References
one or more polymers from the initial solution and
the subsequent deposition of the newly formed Jaspart S, Piel G, Delattre L, Evrard B (2005) Solid lipid
microparticles: formulation, preparation, characterisa-
coacervate phase around the active ingredient
tion, drug release and applications. Expert Opin Drug
suspended or emulsied in the same reaction Deliv 2:113
media. In simple coacervation, phase separation Walde P, Ichikawa S (2001) Enzymes inside lipid vesicles:
is achieved by addition of desolvating agent preparation, reactivity and applications. Biomol Eng
18:143177
(alcohol or salt) or by change in temperature or
Zhang H, Tumarkin E, Sullan RMA, Walker GC,
pH, whereas complex coacervation involves reac- Kumacheva E (2007) Exploring microuidic routes to
tion between two oppositely charged polymers. microgels of biological polymers. Macromol Rapid
The three basic steps in coacervation are Commun 28:527538
Zuidam NJ, Shimoni E (2010) Overview of microencap-
(i) phase separation in a suspension or emulsion
sulates for use in food products or processes and
of active ingredient which leads to a three-phase methods to make them. In: Zuidam NJ, Nedovi VA
system consisting of a polymer-rich liquid (eds) Encapsulation technologies for active food ingre-
phase (coacervate phase), a polymer-lean liquid, dients and food processing. Springer, New York,
pp 329
and a solid or liquid phase containing active
ingredient; (ii) deposition of the coacervate
phase onto the dispersed particles or droplets;
and (iii) hardening of the coating (Zuidam and
Shimoni 2010).
Solid lipid microparticles contain the active Encapsulation: Entrapping,
ingredient entrapped within a high melting point Enclosing, Enveloping
lipid, such as fatty alcohols, fatty acids, fatty acid
esters of glycerol, hydrogenated fatty acid esters, Encapsulation
waxes, etc. Solid lipid microparticles can be fab-
ricated by a temperature-controlled emulsication
of high melting point lipids followed by cooling
or by emulsication/solvent evaporation method. Engineered Organs
In the latter case, a high melting point lipid is
dissolved in an organic solvent, and the mixture Bioarticial Organs, Membrane Operations of
Enhanced Oil Recovery (EOR) 709
hydrocarbon gaseous phase in order to have it

Engineered Organs and Tissue recycled to the reservoir (after pressurization).
Since the quantity of CO2 dissolved in oil varies
Bioarticial Organs and Tissue over a large range of values during the whole
production of the reservoir (levels from 40 % up
to 90 % can be observed after a few years of
production (Cooley and Dethloff 1985)), it is rather
Enhanced Oil Recovery (EOR) difcult to design a cost-effective CO2 removal
unit that could operate during the entire life of the
Arnaud Baudot reservoir. Therefore, modular CO2 removal units
E
Physics and Analysis Division, IFP Energies such as membrane operations offer much exibility
Nouvelles, Solaize, France for such an operation (Table 1).
The rst large-scale EOR project operating
membranes has been carried out in the Sacroc oil
Enhanced oil recovery (EOR) aims at increasing eld located in Western Texas. Carbon dioxide
the oil elds recovery yield, thanks to high- injection was launched in 1972 (at a ow
pressure injection of water or carbon dioxide. range up to 240,000 Nm3/h). Three CO2 recovery
Injection of pressurized carbon dioxide leads to units were initially installed: two hot potassium
the maintenance of high pressures in the reservoir carbonate-based absorption units (operated by
and to an improvement of oil displacement, Sun Explo with a 190,000 Nm3/h capacity leading
thanks to several physical effects: decrease of oil to a reduction of carbon dioxide from 24 % to
viscosity due to carbon dioxide solubilization, 0.5 % and Chevron, with a 54,000 Nm3/h lower-
stripping effects, and modication of multiphase ing CO2 content down to 1 %) and an amine
equilibria due to CO2 solubilization in the reser- (MEA)-based absorption unit (operated by
voir aqueous phase. When the oil reaches the Monsanto, with a 20,000 Nm3/h capacity). At
surface, carbon dioxide is purged with the associ- the end of the 1970s, realizing that the level of
ated gas. Since the volumes of concerned carbon CO2 was raising at a higher rate than expected,
dioxide are very large from 140 to 280 Nm3 par Chevron contacted Cynara in order to install and
extracted barrel (Mazur and Chan 1982) it is operate two gas permeation units upstream of the
necessary to separate carbon dioxide from the hot potassium carbonate absorption columns. The
Enhanced Oil Recovery (EOR), Table 1 Comparison of relative investment costs (CAPEX) and operating costs
(OPEX) of membrane-based, solvent-based, and hybrid processes for carbon dioxide capture (EOR applications)
CAPEX
rf. Cryo.a Mem.
Operating Study realized OPEX KCO3
conditions by Year rf. + Mem. Mem. +DEA DEA Cryo.a TEA chaud
175.000 m3/h Amoco 1982 47 M$ 1 2.2 1.56 1.48
90 % CO2, (Goddin 20.4 M$/a 1 1.76 1.39 1.12
18 bar 1982)
118.000 m3/h Permea 1983 16.1 M$ 1.50 1 1.31
80 % CO2 (Boustany 6.5 M$/a 1.31 1 1.82
et al. 1983)
190.000 m3/h Fluor 1984 1 0.63 0.63b
40 % CO2 Engineering 1 0.56 0.68b
Schendel and
Seymour
1985)
a
Cryo.: cryogenic distillation (Ryan-Holmes process (Kohl and Nielsen 1997))
b
Hot potassium carbonate + membrane
710 Enhanced Oil Recovery (EOR)
Enhanced oil recovery 1 2 3 4 5 6

(Kvaerner Membrane Systems)
3
Flow rate (103 Nm3/h) 26.5 7.4 18.9 11.2 15.3 3.8
Pressure (bar) 16 25 20 26 5 1
5 6 2
CO2 70 10 93.4 35.4 95.2 86.2
CH4 9.6 26.9 2.9 20 2.0 6.7
C2 6.3 10.0 1.0 14.0 0.7 2.4 1
Composition (% mol.) C3 5.6 18.4 0.6 12.9 0.4 1.5
i-C4 2.5 8.6 - 6.0 - 0.2
n-C4 1.3 4.3 - 3.0 - 0.1
H2S 0.6 0.01 0.8 0.1 0.9 0.4 4
N2 4.1 11.8 1.1 8.6 0.7 2.5
CH4 (C2+) recovery yield = 77,8 % (73 %)
Enhanced Oil Recovery (EOR), Fig. 1 EOR example of permeation application (Spillman 1989)
treatment capacities of the membrane units were, Medal:

respectively, 60,000 Nm3/h at 35 bar feed pressure A Medal membrane unit was operated at the
prior to the Sun unit and 25,000 Nm3/h at 33 bar need of the 1990s to process 14,200 Nm3/h
feed pressure prior to the Chevron unit. Cynara at 56 bar, 48 % CO2, reducing the retentate
chose to operate a single-stage conguration, the CO2 concentration down to 6 %.
membrane module being installed in parallel. The Cynara (Kohl and Nielsen 1995):
membrane operation proved to be highly exible Amoco, Texas (1994): Cynara membranes
and reliable, as the membrane modules remained were used to process 35,000 up to
in operation twice longer than the expected life- 120,000 Nm3/h of gas containing 80 % CO2.
time (more than 5 years instead of 23 years) Mobil, Salt Creek (1992): Cynara membranes
without any selectivity decrease. A slight decrease were used to process 76,000 up to
in membrane permeability has nevertheless been 120,000 Nm3/h of containing 70 % CO2.
observed throughout the operation, and it was Shell, Texas: Cynara membranes were used to
necessary to add more modules in order to main- process 13,000 Nm3/h of gas containing
tain the production level of the membrane opera- 70 % CO2 (Fig. 1).
tion (Marquez and Hamaker 1986).
Since then, other membrane companies have
seen their products involved in other EOR References
projects:
Boustany K, Narayan RS, Stookey DJ (1983) Economics
of removal of carbon dioxide from hydrocarbon gas
Kvaerner membrane systems (Chapel
mixtures. In: Abstracts 62nd Annual Gas Processors
et al. 1999): Association, San Francisco, CA, pp 146149
Dallas Production Inc. (Texas) treated Chapel DG, Mariz CL, Ernest J (1999) Recovery of CO2
200,000 Nm3 of gases containing up to from ue gases: commercial trends. In: Oral Presenta-
tion Canadian Society of Chemical Engineers Annual
25 % of carbon dioxide in 1994 during a
Meeting, Saskatoon, Sakatchewan, Canada.
feasibility study that lasted for 18 months. Cooley TE, Dethloff WL (1985) Field-tests show mem-
After this test, the membrane unit was in brane processing attractive. Chem Eng Prog 81:4550
operation treating 1200 Nm3/h. Goddin CS (1982) Comparison of processes for treating
gases with high CO2 content. In: Abstracts 61st Annual
Hydrocarbon Operating, Inc. (Texas)
Gas Processors Association Convention, Dallas, TX,
processed 590 Nm3/h at 51 bar feed pres- pp 6068
sure in 1994 on a pilot skid composed of Kohl A, Nielsen R (1997) Gas purication, 5th edn. Gulf
two tubular membrane modules. Publishing Co, Houston, pp 12381295
Environmentally Friendly Solvents and Diluents 711
Marquez JJ, Hamaker RJ (1986) Development of mem- many membrane processes are employed with
brane performances during SACROC operations. In: the objective of environmentally improving the
Oral presentation at AIChE Spring National Meeting
and Petro. Expo86. New Orleans, LA process industry, toxic solvents/diluents reduce
Mazur WH, Chan MC (1982) Membranes for natural-gas their contributions to environmental protection.
sweetening and CO2 enrichment. Chem Eng Prog Thus, nding more environmentally friendly
78:3843 solvents/diluents is becoming an important
Schendel RL, Seymour JD (1985) Take care in picking
membranes to combine with other processes for CO2 topic in the membrane preparation eld (Cui
removal. Oil Gas J 83:8486 et al. 2013a).
Spillman RW (1989) Economics of gas separation mem- Recently, efforts have been done to develop
branes. Chem Eng Prog 85:4162 environmentally friendly solvents for NIPS pro-
E
cess and diluents for TIPS process. These sol-
vents/diluents should yield the membranes
promising properties and performances without
discharging toxicity to environment. It is much
Environmentally Friendly Solvents better if the solvents/diluents can be recovered
and Diluents easily. Dimethyl sulfoxide (DMSO) and glycerin
triacetate (GTA), which can be employed for
Zhaoliang Cui NIPS and TIPS process, are relatively less
State Key Laboratory of Materials-Oriented toxic. They have been used for polymeric mem-
Chemical Engineering, College of Chemistry and brane preparation. Acetyl tributyl citrate (ATBC)
Chemical Engineering, Nanjing Tech University, (Cui et al. 2013b), which is more environmentally
Nanjing, China friendly, was introduced to prepare poly(vinyli-
dene uoride) (PVDF) membranes via TIPS pro-
cess. Dimethyl sulfone (DMSO2) (Liang
et al. 2013) was used as a universal crystallizable
Non-solvent-induced phase separation (NIPS) diluent to fabricate polar polymer membranes
and thermally induced phase separation (TIPS) (PVDF, polyacrylonitrile (PAN) and cellulose
are two major methods for polymeric porous acetate (CA)) via TIPS method. This diluent is
membrane preparation. Solvents and diluents easy to recover because of its crystallizable prop-
play important roles in NIPS and TIPS processes, erty. The membranes fabricated by above diluents
respectively, inuencing the nal membrane presented promising mechanical properties and
properties during polymeric membrane formation. water permeability. Ionic liquid is a type of
Applicable solvents for the NIPS process are N, green solvents, which probably can be used for
N-dimethyl acetamide (DMAc), N,N-dimethyl polymeric membrane preparations. CA mem-
formamide (DMF), hexamethyl phosphoramide branes (Xing et al. 2010) have been formed via
(HMPA), N-methyl pyrrolidone (NMP), phase inversion employing 1-butyl-3-
tetramethylurea (TMU), triethyl phosphate methylimidazolium thiocyanate ([BMIM]SCN)
(TEP), trimethyl phosphate (TMP), acetone as the solvent and were recycled. The morphol-
(Ac), and tetrahydrofuran (THF), while applicable ogy, porosity, and pure water permeability of the
diluents for the TIPS process are dimethyl phthal- CA membranes fabricated by reused solvents
ate (DMP), diethyl phthalate (DEP), dibutyl were quite comparable to the CA membranes
phthalate (DBP), dihexyl phthalate (DHP), ethyl produced by new solvents.
acetoacetate (EAA), propylene glycol carbonate Researchers started to pay attention to
(PGC), diphenyl ketone (DPK), diphenyl carbon- develop environmentally friendly solvents/dil-
ate (DPC), cyclohexanone, camphor, and uents, and preliminary progress has been
butyrolactone. Most of the above solvents/dilu- made. However, this is far from the require-
ents are toxic, making the work atmosphere ments, and more efforts are needed to contrib-
detrimental for workers and environment. Since ute to this eld.
712 Environmentally Responsive Membranes
References separation of the enzyme and its reuse in the

enzymatic process.
Cui Z, Drioli E, Lee YM (2013a) Recent progress in This technology is especially suitable in con-
uoropolymers for membranes. Prog Polym Sci
tinuous operation by ensuring the recovery and
39(1):164198. doi:10.1016/j.progpolymsci.2013.
07.008 reusability of the enzymes, allowing the
Cui Z, Hassankiadeh NT, Lee SY, Lee JM, Woo KT, minimization of large consumptions of biocata-
Sanguineti A, Arcella V, Lee YM, Drioli E (2013b) lyst. Various enzymatic membrane bioreactor
Poly (vinylidene uoride) membrane preparation with
congurations are possible, i.e., enzyme
an environmental diluent via thermally induced phase
separation. J Membr Sci 444:223236 suspended in solution in a reactor connected
Liang HQ, Wu QY, Wan LS, Huang XJ, Xu ZK with a membrane unit, immobilized enzyme
(2013) Polar polymer membranes via thermally within the membrane matrix itself, or entrapped
induced phase separation using a universal crystalliz-
enzyme in gel or microcapsules.
able diluent. J Membr Sci 446:482491. doi:10.1016/j.
memsci.2013.07.008 Among these setups, the retention of the
Xing DY, Peng N, Chung TS (2010) Formation of cellulose enzyme by an ultraltration membrane is a very
acetate membranes via phase inversion using ionic interesting alternative to overcome the washing
liquid, [BMIM]SCN, as the solvent. Ind Eng Chem
out of the catalyst with the treated efuent but
Res 49:87618769
also to avoid the enzyme immobilization which
is usually related with high costs, complex pro-
cedures, and loss of enzyme catalytic activity. In
this system the enzyme is added into a reactor tank
Environmentally Responsive which is coupled with an ultraltration membrane
Membranes to enable the retention of the free enzyme and its
recycling back to the reaction vessel. Thereby, by
Magnetically Responsive Membranes using an enzymatic membrane reactor, it is possi-
ble to separate the biocatalyst from products
and/or other substrates by a semipermeable mem-
brane that creates a selective physical/chemical
Enzymatic (Peroxidase) Membrane barrier (Fig. 1).
Bioreactor This type of bioreactor offers important bene-
ts: high enzyme loads, prolonged enzyme activ-
Mara. T. Moreira1, Gemma Eibes2, ity, high ow rates, reduced energy requirements,
Thelmo Lu-Chau2, Roberto Taboada-Puig2, straightforward operation, and scale-up, and also,
Adriana Arca-Ramos2, Gumersindo Feijoo2, fresh enzyme can be easily added to maintain
Juan M. Lema2 and Lucia Lloret3 constant enzymatic levels. Indeed, this system
1
Department of Chemical Engineering, School of has been successfully operated for the continuous
Engineering, University of Santiago de application of peroxidases (Lpez et al. 2004) and
Compostela, Santiago de Compostela, Spain laccases (Lloret et al. 2013). Specically,
2
Department of Chemical Engineering, Institute peroxidase-based membrane bioreactors were
of Technology, University of Santiago de applied for the transformation of various com-
Compostela, Santiago de Compostela, Spain pounds with bioremediation purposes, for exam-
3
Chemical and Environmental Engineering ple, for the removal of dyes (Orange II) and
Department, Federico Santa Maria Technical estrogenic compounds (estrone, estradiol,
University, Santiago, Chile ethinylestradiol, etc.). Orange II was successfully
transformed by manganese peroxidase (MnP) in
an enzymatic membrane bioreactor which
Enzymatic (peroxidase) membrane bioreactor is consisted of a 250-mL stirred tank reactor
an enzymatic reactor coupled to a membrane in a (Lpez et al. 2004 in Fig. 1) connected to a
number of congurations that allows the polyethersulfone synthetic membrane (Prep/
Enzymatic Membrane Reactor (EMR) 713
MnP
[Fe3+]
H2O2
Mn3+
H2O Mn2+
MnP [2]
H2O2 Products
MnP-I MnP-II
[Fe4+=0 P] [Fe4+=0]
H2O2 Mn2+ Mn3+
Organic acid
Mn2+ E
H2O
Target compounds
MnP-III
[1]
Enzymatic (Peroxidase) Membrane Bioreactor, Fig. 1 Scheme of an enzymatic (peroxidase) membrane reactor
Scale-TFF Millipore) with a molecular weight References

cutoff of 10 kDa (Lloret et al. 2013 in Fig. 1).
Three stock solutions: H2O2 (94312 mM); Mn2+ Lloret L, Eibes G, Moreira MT, Feijoo G, Lema JM
(2013) Removal of estrogenic compounds from ltered
(33 mM), Orange II (0.1 g/L), and organic acid
secondary wastewater efuent in a continuous enzy-
(1 mM); and MnP (7,000 U/L, 125225 U/L in the matic membrane reactor. Identication of transforma-
tank), were fed into the bioreactor by independent tion products. Environ Sci Technol 47:45364543
variable speed peristaltic pumps, thus initiating Lpez C, Moreira MT, Feijoo G, Lema JM (2004) Dye
decolourization by manganese peroxidase in an enzy-
the catalytic cycle of the enzyme. Under the best
matic membrane reactor. Biotechnol Prog 20:7481
conditions evaluated, high decolorization yield
can be achieved and minimal enzymatic deactiva-
tion, rendering an efciency of 42.5 mg Orange II
oxidized per unit consumed of MnP.
The ndings obtained with the proposed reac- Enzymatic Membrane Reactor (EMR)
tor conguration allow its application as a novel
treatment method for contaminated wastewaters. Lidietta Giorno
The simplicity of the operation, which mainly Institute on Membrane Technology, National
relies on the control of H2O2 dosage, promotes Research Council of Italy, ITM-CNR, Rende,
the peroxidase membrane bioreactor for further Italy
scale-up applications. The oxidative potential of
the MnP-H2O2-Mn2+ system appears to be gen-
eral enough to be applied for the transformation of Synonyms
a wide range of compounds. The main require-
ment is that the oxidation potential of a particular Enzyme membrane reactor
substrate has to be lower than that provided by the
enzymatic cycle, being the main factors affecting A general denition of an enzyme membrane
the stoichiometry and kinetics of the process the reactor (EMR) is a reactor system where a mem-
MnP activity, dosage rate of H2O2, and concen- brane separation is used to keep larger compo-
tration of the target compound. nents in the reactor vessel (i.e., enzymes and/or
714 Enzymatic Membrane Reactor (EMR)
macromolecular substrates), while low- P = product concentration, ML3

molecular-weight molecules (i.e., products Pf = feed product concentration, ML3
and/or inhibitors) are allowed to pass freely
through the membrane, thus leaving the reactor Assuming that substrate conversion obeys the
as permeate (Drioli and Giorno 1999). In this simple Michaelis-Menten model, substrate steady
setup, several advantages of immobilized prepa- state mass balance reduces to:
rations, together with easy recovery of deactivated
enzymes and replacement with fresh catalysts are XSf X=1 X K0M K E V=Q (3)
achieved; inhibitors are continuously removed
from the reaction vessel. The direct and deep R being expressed as K E S= K0M S .
contact between substrates and biocatalysts limits
diffusional resistances, while no activity losses X Sf S=Sf = conversion degree
due to xation to the support occur, thus maxi- KM0
= enzyme Michaelis-Menten constant, ML3
mizing the activity of the biocatalyst. On the other K = enzyme kinetic constant, T1
hand, enzyme stability is not improved. E = enzyme concentration, ML3
Both dead-end and continuous stirred tank V = reaction volume, L3
reactor (CSTR) ultraltration cells with at mem- Q = ow rate, L3 T1
branes can be used as enzymatic reactors.
Performance of dead-end units is largely The equation in this form is a useful tool to
affected by the ow dynamics of the system; in estimate parameters of reaction kinetics. Instead
fact, mixing of substrates and catalysts is not fully of performing nonlinear parameters estimation
accomplished, and concentration polarization procedures, the functional dependence of XSf on
phenomena strongly limit reactor performance. X=1 X can be plotted. The plotshould look
Continuous stirred tank reactors have been like a stright line whose slope is K0M and
more widely adopted, both due to the possibility whose intersection with the XSf axis is given by
of concentration polarization control and easy the point of coordinate (Vmaxt).
modeling of enzyme kinetic behavior. Enzyme activity is usually not constant with
Assuming complete mixing within the reactor time. Physicochemical changes in enzyme struc-
so that enzyme and substrate concentration in the ture, thermal denaturation, and microbial contami-
reactor vessel are uniform, and the latter is equal nation cause enzyme activity to continuously
to its value in the permeate, substrate mass bal- decrease with time. When enzymes or cells are
ance in molar form can be written as: compartmentalized in UF cells, biocatalyst
losses can even occur due to the wrong choice of
Sf S=t R S, P (1) membrane molecular weight cutoff. It is conven-
tional to measure the enzyme stability in terms
where of its half-life time (t1/2), that is, the time at
which enzyme activity is reduced to half its initial
Sf = feed substrate concentration, ML3 value. It can be calculated from the following
S = substrate concentration, ML3 equations:
t = reactor time constant, T
R = reaction rate, MT1 L3 2:303 AE 0:693
Kd log 0 t1=2 (4)
# AE # Kd
Product steady state mass balance similarly
will be: Kd = enzyme deactivation constant, T1
# = operation time, T
P Pf =t R S, P (2) AE0 = initial enzyme activity, or product mass per
unit time and reaction volume, M T1
where AE# = enzyme activity at time #, M T1
Enzymatic Membrane Reactor (EMR) 715
Since biotransformations by means of enzymes lnfSf S t =t dS t=dtg vs t. A straight line

are continuous processes, as long as the reactor is thus obtained whose slope is given by Kd
working life is longer than the native enzyme half- and the line intersects the vertical axis at the point
life, enzyme activity decay with time must be of coordinate (ln Vmax).
taken into account in order to correctly assess When the reacting solution is fed to the system,
reactor performance. A transient substrate mass low-molecular-weight products leave the reactor
balance on the CST reactor leads to the equation permeating the membrane, whereas enzymes are
partially or totally rejected. Then, enzymes tend
V dS t=dt Q Sf S t Vmax V (5) to accumulate in a thin layer immediately
upstream from the membrane causing polarization
E
where phenomena to occur. The extent to which concen-
tration polarization affects reactor performance
V = reaction volume, L3 depends on the balance of rejected solutes, e.g.,
S = substrate concentration, ML3 enzymes or cells, accumulation due to membrane
Sf = feed substrate concentration, ML3 rejection and back diffusion to the bulk phase, and
Vmax = enzyme reaction rate at saturating sub- eventually on the ow dynamics of the reacting
strate concentration, ML3 T1 vessel.
Q = ow rate, L3 T1 For macromolecules (like most biocatalysts), if
membrane properties are carefully chosen, mem-
Let us assume that substrate inlet concentration brane rejection is usually very good, while biocat-
is much higher than the enzymes apparent alyst back diffusion towards the bulk phase is
0
Michaelis constant, KM so that enzyme kinetics extremely slow. The effect of concentration polar-
is expressed according to zero order kinetics, that ization on such reactor performances can then be
is, R KE Vmax . Enzyme activity decay can signicant.
be expressed in terms of the Arrhenius equation: Applying the thin lm theory to a region
immediately upstream from the membrane results
E Eo exp Kd t (6) in the following steady state mass balance
equation:
where
J E De dE=dx (9)
E = enzyme concentration, ML3
E0 = initial enzyme concentration, ML3 under the boundary condition (BC)
Kd = enzyme deactivation constant, T1 x ! 1 E Es
where
Equation 5 then takes the form
J = volumetric ux, L3L2T1
Sf S t =t dS t=dt E = enzyme concentration, ML3
Vmaxo exp Kd t (7) ES = enzyme concentration in the bulk liquid
phase, ML3
Integration of the differential Equation 7 leads De = enzyme diffusion coefcient, L2T1
to estimate the outlet substrate concentration:
Under the assumption of totally rejected
S t Vmaxo exp t=t exp Kd t = enzyme macromolecules, integration of Equa-
1=t Kd Sf tion 9 allows one to express the ratio of enzyme
(8) concentration in the bulk solution in the presence
of concentration polarization to its value in
Estimation of the deactivation constant can be the absence of concentration polarization phe-
graphically carried out from Equation 7 plotting nomena as:
716 Enzymatic Membrane Reactor (EMR)
ES =Eo 1=f1 a a=b exp b 1 g products and/or inhibitors are continuously

(10) removed from the system. Soluble enzymes can
then act directly on substrate macromolecules
Enzyme concentration at the membranesolution without diffusion limitations and steric hindrance
interface can thus be related to enzyme concentra- imposed by enzyme xation to a solid support.
tion in the bulk phase by the following equation: The stirring features of CST EMRs moreover
assures that substrates and/or inhibitors within
the reactor vessel are maintained at the lowest
Ew ES exp b (11) possible concentration level. Such reactor cong-
uration is then extremely useful when substrate
where the dimensionless parameter b is a measure inhibited reaction patterns are involved or when
of convective mass transfer, J, relative to the over- inhibiting species are assumed to exist in the feed
all mass transfer, D/d, more generally KS. stream.
The change of enzyme bulk concentration can Dialtration, semicontinuous, and continuous
be dramatic, with changes in permeate ow rate, operational modes can be used.
and applied pressure. Changes in enzyme bulk Operating the reactor in a semicontinuous con-
concentration may induce a large reduction in dition and adding substrate so as to keep its bulk
reaction rate within the membrane reactor. Under concentration constant leads to meaningful
given stirring conditions, therefore, a critical changes in reactor performance as compared to
value of ow rate and hence of applied pressure the dialtration mode. Under both operational
does exist. At ow rates lower than the critical mode, permeate ow rate continuously decreases
value, the reactor always attains steady operation with time.
conditions, and correspondingly outlet product When EMRs are operated continuously, feed-
concentration is constant. Beyond the critical ing a slurry of substrate macromolecules to the
value, concentration polarization phenomena pro- reactor, concentration polarization phenomena
mote the localization of a large fraction of play a dominant role. The presence within the
enzymes near the membrane surface, seldom lead- reaction vessel of contaminants or intermediate
ing to an enzymatic gel formation. Correspond- products, which are not fully hydrolyzed by the
ingly, an accelerated deactivation of enzyme enzymatic system under study can lead to mem-
activity is superimposed on reactor performance brane fouling or to the formation of a gel layer at
thus hindering the attainment of steady state con- the membrane surface. Under such conditions, the
ditions. Experimental evidence suggests that the ltration rate continuously decreases with time,
contribution of polarized enzymes to overall con- and it may happen that substrate conversion does
version can be negligible, due to the consistency not attain steady state conditions. The addition of
of product concentration in the bulk phase of the enzymes capable of hydrolyzing such foulants to
reactor and in the permeate. low-molecular-weight compounds usually
When macromolecular substrates are involved improves reactor performance, eventually
in the transformation under study, concentration approaching steady state conditions in terms of
polarization phenomena affect the EMR perfor- both permeate ow rate and substrate conversion.
mance more severely. Diffusion limitations of In addition, antifouling procedures including suit-
macromolecular substrates hamper the use able feed pretreatment or procedures to reduce
of immobilized enzymes in the hydrolysis of concentration polarization can be used.
high-molecular-weight substrates. By selecting Equation 11 demonstrates the dependence of
membranes with an appropriate molecular weight enzyme concentration at the upstream membrane
cutoff, both enzyme and substrate are retained in surface on the ow dynamics of the reaction ves-
an EMR in touch with each other, and hydrolysis sel. Due to the monotonicity of the exponential
Enzymatic Membrane Reactor in Supercritical CO2 717
function, at given applied pressure operational present very specic activities. They are one of
conditions improving mass transfer in the bulk the main tools of life as it developed since the
phase (i.e., increasing axial ow rate improves origin. Enzymes are divided into six classes,
KS or D/d) diminish concentration polarization. based on the type of reaction that is catalyzed.
High stirring speeds usually improve ltering per- Lipases are important enzymes that belong to the
formance of a CST at membrane reactor, albeit at class of hydrolases; they catalyze hydrolytic reac-
the expense of partial enzyme deactivation. tions. Candida antarctica lipase B (CalB) is one
Reactor congurations other than dead end or of the most widely used lipases, because this
CSTR/UF cell can offer improvements. enzyme accepts a broad spectrum of different sub-
The use of hollow ber ultraltration modules strates. Most of the time enzymes are immobilized
E
in cross-ow ltration mode strongly improves for industrial applications. The main reason to
both reactor performance and economics. Capil- immobilize enzymes lies in the reduction of
lary membranes are characterized by a favorable costs by enabling efcient separation, recycling,
surface to volume ratio; advantages from this fea- and reuse of the expensive enzymes. The choice
ture are not only related to lower overall plant size for the best immobilization method depends on
but also to the increase of the surface to price ratio. the enzyme, the type of reaction, and the reaction
Flushing enzyme/substrate solution through UF environment.
module tangentially to the membrane surface at The use of porous membrane reactor for enzy-
a high linear velocity reduces the extent of con- matic catalysis represents from this viewpoint a
centration polarization avoiding the formation of a breakthrough technology which offers many
secondary enzymatic gel membrane. Provided advantages as compared to other more classical
that the volume of the UF module is small relative systems (immobilization on divided solid parti-
to the total volume, and that recirculation ow rate cles in xed, uidized, or moving beds): continu-
is much larger than permeate ux, system kinetic ous mode, reduction of product/substrate
behavior can be modeled in terms of a CST inhibition, free enzyme end-products, and the pos-
reactor. sibility to set up monophasic or biphasic
systems even if there can be some drawback
related to membrane fouling. The full activity may
References be maintained inside the reactor by the rejection of
enzyme molecules (Fig. 1) either freely circulat-
Drioli E, Giorno L (1999) Biocatalytic membrane reactors: ing (a) or immobilized on/in the membrane (b). In
application in biotechnology and the pharmaceutical
that case, the membrane is the catalyst media
(functionalization) and the reaction occurs during
membrane crossing. The convective control due
to the application of a low transmembrane pres-
sure (very small thickness of the porous layer: a
Enzymatic Membrane Reactor few microns) is responsible for the mass transfer
in Supercritical CO2 increase. The membrane is a macrosystem
resulting from the assembly of swarms of
Gilbert M. Rios microsystems: each pore represents a particular
Place Eugene Bataillon, IEM European microreactor.
Membrane Institute, Montpellier, France The main practical use of enzymatic catalysis
in supercritical carbon dioxide (scCO2) has to be
found in producing high-added-value products
Enzymes are catalysts obtained from nature since the process remains expensive. Lipases
(proteins), which work in mild conditions and have proven to be rather stable and active during
718 Enzymatic Membrane Reactor in Supercritical CO2
a S P b S
a b
S E
E S P
E
P P
P
membrane membrane Diffusion convection
Enzymatic Membrane Reactor in Supercritical CO2, Fig. 1 Basic mechanisms
Pressure SCFE + NF process

20 MPa
Retentate
(recycled)
Pure SCF Extraction Membrane
chamber separation
Permeate :
1.0 XCO2 SCF + Dispersed molecules
SCF
Fluidification process Molecules
Retentate Pressure
Viscous
SCF (recycled)
liquid Membrane 20 MPa
injection
separation
chamber
Permeate :
Viscous liquid + Dissolved SCF
SCF
0.2 XCO2
Viscous liquid
Enzymatic Membrane Reactor in Supercritical CO2, Fig. 2 Plant schemas
long-term reactions in scCO2. However, it cannot the possibility of easier separation steps with com-
be stated that performing reactions in scCO2 will plete removal of the solvent (Fig. 2).
always result in a higher activity. This is highly The future perspectives for enzymatic mem-
dependent on the type of reaction, the enzyme, brane reactors in scCO2 can be found in the
and the applied conditions. The main advantage of combination of improving the stability and
CO2, a very stable and nontoxic uid, is therefore activity of the potentially interesting
Enzymatic Ultrafiltration Membranes (Immobilization of Urease and Trypsin) 719
enzymes other types of enzymes are attractive

for use in scCO2 as well, such as oxygenases, Enzymatic Ultrafiltration Membranes
epoxide hydrolases, aldolases, or decarbo- (Immobilization of Urease
xylases and an improved integration of the and Trypsin)
reaction and separation sections. Very interest-
ing studies conducted lately on transesteri- Christophe Innocent
cation reactions with the xed enzyme Institut Europen des Membranes, ENSCM,
conguration (Fig. 1b), either using pure SC CNRS, Universit de Montpellier, Montpellier,
uid solvent containing dissolved substrates or France
for transforming oils after the injection of a
E
small quantity of pressurized CO2 in order to
decrease viscosity, are worth mentioning. The sodium alginate was used as an anionic poly-
electrolyte. The polyethyleneimine was used as a
cationic polyelectrolyte. Trypsin (enzyme which
catalyzed the hydrolysis of peptide bonds in
References
which the carboxyl groups are contributed by the
Dijkstra ZJ (2006) The potential of enzymatic catalysis in lysine and arginine residues) was applied as a
supercritical uids. PhD thesis, Eindhoven University positively charged polyelectrolyte and the urease
of Technology (NL). ISBN-10: 90-386-2698-3/ISBN- (enzyme which catalyzes the hydrolysis of urea to
13: 978-90-386-2698-7
ammonium and carbon dioxide) as a negatively
Pommier E, Delebecque N, Paolucci-Jeanjean D, Pina M,
Sarrade S, Rios GM (2007) Effect of working condi- charged polyelectrolyte. For adsorption of the
tions on vegetable oil transformation in an enzymatic polyelectrolyte layers, the supporting membrane
reactor combining membrane and supercritical CO2. (polyacrylonitrile-modied membrane negatively
J Supercritl Fluids 41(3):380385
charged) was immersed in the solution of the
Rios GM (2009) Enzymatic membrane reactor. Plenary lec-
ture. PERMEA, Prague. 911 June 2009. http://www. cationic polyelectrolyte, rinsed with the buffer
imc.cas.cz/sympo/permea09/programme_plenary.html solution to remove excess of polycations on
Enzymatic Ultrafiltration Membranes (Immobilization of Urease and Trypsin), Fig. 1 Scheme of layer-by-layer
adsorption of polyelectrolytes on membrane
720 Enzyme Compartmentalization
membrane surface, immersed in the solution of used for the production of membrane bioreactors
the anionic polyelectrolyte, and rinsed with the or biocatalytic membrane reactors (Mazzei
buffer solution again. These steps were repeated et al. 2010). In particular, asymmetric hollow
to form multiple polyelectrolyte bilayers based on ber membrane can be used as efcient support
the electrostatic layer-by-layer self-assembly to compartmentalize enzymes by entrapment into
(Fig. 1). the sponge membrane layer. The general mem-
brane requirement is that the pores in the dense
layer are small enough to retain enzyme mole-
References cules and at the same time large enough to freely
pass substrates and products (Mazzei et al. 2013).
Decher G (1997) Fuzzy nanoassemblies: toward layered
polymeric multicomposites. Science 277:12321237
References
Enzyme Compartmentalization Mazzei R, Drioli E, Giorno L (2010) Biocatalytic mem-

branes and membrane bioreactors. In: Drioli E and
Giorno L (eds) Comprehensive membrane science and
Rosalinda Mazzei engineering, vol 3. Elsevier B.V., Oxford, pp 195212
Institute on Membrane Technology, National Mazzei R, Piacentini E, Drioli E, Giorno L (2013) Mem-
Research Council of Italy, ITM-CNR, Rende, brane bioreactors for green processing. In: Boodhoo K,
Harvey A (eds) Process intensication for green chem-
Italy istry: engineering solutions for sustainable chemical
processing. John Wiley & Sons, Ltd., Chichester,
pp 227250
Synonyms
Immobilization of enzymes
Enzyme Membrane Reactor
In biology enzyme compartmentalization is ef-
ciently carried out by biomembrane or biological Enzymatic Membrane Reactor (EMR)
membrane that permits to conne specic func-
tions given by enzymes in a precise space. The
enzymes can be localized inside the delimited
compartments or on the membrane surface. Enzyme Membrane Reactors
In eukaryotic cells the corresponding delimited
spaces or compartments are called organelles, Membrane Bioreactors
from which the main are the endoplasmic reticu-
lum, Golgi apparatus, nucleus, mitochondria,
lysosomes, endosomes, and peroxisomes. Each
membrane-enclosed organelle contains a specic Enzymes Immobilized on Lumen
set of proteins free or on the membrane surface
that regulates many vital biochemical processes. Rosalinda Mazzei
For instance, lipid metabolism is catalyzed mostly Institute on Membrane Technology, National
by membrane-bound enzymes. Research Council of Italy, ITM-CNR, Rende,
Starting from nature simulation, in biotechnol- Italy
ogy process, enzyme compartmentalization can
be obtained by immobilizing proteins on supports
or delimiting them by support compartments. Synonyms
Membranes can be optimal supports for enzyme
compartmentalization; in fact they are widely Immobilization of enzymes
Ethanol Production by Continuous Fermentation-Pervaporation 721
EOF
Ethanol Production by Continuous

Fermentation-Pervaporation
E
Enzymes Immobilized on Lumen, Fig. 1 Scheme of Wanqin Jin
tubular biocatalytic membrane with biocatalyst State Key Laboratory of Materials-Oriented
immobilized on lumen surface Chemical Engineering, Nanjing University of
Technology, Nanjing, China
This particular kind of immobilization is generally

Ethanol production by continuous fermentation-
referred to enzymes immobilized into the mem-
pervaporation is a process that couples ethanol
brane lumen surface of tubular membranes for the
fermentation with in situ pervaporation
production of biocatalytic membrane and biocat-
(PV) recovery of ethanol from fermentation broth.
alytic membrane reactors. A scheme of a cross
With the fast increasing of fossil fuel consump-
section of a biocatalytic asymmetric hollow ber
tion and the deterioration of environment, it is
membrane is reported in Fig. 1, in which the place
commonly accepted that the fossil fuel for
that the biocatalyst was immobilized on the mem-
human energy use has to be replaced generally
brane lumen surface is highlighted.
by renewable resources. Compared with tradi-
The biocatalyst can be also compartmental-
tional resources, bioethanol produced by fermen-
ized/segregated in the lumen of tubular membrane
tation process can be sustainably developed
so conned in a dened region of the membrane
without extra yield of greenhouse gas (Peng
module space. In these reactors, the enzymes or
et al. 2011). Bioethanol is produced from renew-
cells are not linked to the membrane. The mem-
able resources (biomass) by ethanol fermentation
brane in this particular case can retain the enzyme
process. However, usually, the maximum concen-
or the cell which is not lost in the efuent stream,
tration of solvents (ethanol) does not exceed
while low-molecular-weight products and inhibi-
8 wt% owing to the end-product inhibition on
tors can be removed through the membrane
microorganism, resulting in high energy cost to
(Strathmann et al. 2006).
separate ethanol from the dilute fermentation
broth by distillation. Therefore, several in situ
product-removal technologies, such as adsorp-
References tion, gas stripping, liquid-liquid extraction,
perstraction, pervaporation, and reverse osmosis,
to membrane science and technology. Consiglio have been developed. These technologies could
Nazionale delle Ricerche, Roma be coupled with ethanol fermentation process to
decrease the effect of product inhibition and
improve the sugar utilization and solvent produc-
tivity (Vane 2005).
Pervaporation coupled with fermentation,
EOD which is also called fermentation-pervaporation
coupled process, has been regarded as the most
Electroosmotic Drag in Membranes promising separation method because this process
722 Ethanol Production by Continuous Fermentation-Pervaporation
Ethanol Production by Continuous Fermentation-Pervaporation, Fig. 1 Ethanol fermentation-pervaporation

coupled process
can in situ extract organic components which is (PTMSP) membranes, poly(ether block amide)
harmless for microorganisms in mild condition. (PEBA) membranes, liquid membranes, other
Figure 1 shows its typical ow chart (Vane 2005). modied polymer membranes, as well as porous
A pervaporation membrane module is connected polypropylene (PP) membranes and polytetra-
to the fermentor to form a circulation of fermen- uoroethylene (PTFE) membranes. Among
tation broth. Preheated fermentation solution is them, the commonly used PDMS membranes
continuously circulated from fermentor through with good selectivity and stability are regarded
the membrane upstream side with another side as the most potential PV membranes for ethanol
being vacuumed using vacuum pump. Therefore, recovery application (Xiangli et al. 2008). The
ethanol solvent can be selectively and continu- typical inorganic membrane for PV is hydropho-
ously removed from the fermentation broth and bic zeolite membrane, such as ZSM-5 and
then concentrated in the membrane downstream Silicalite-1 membrane.
side. As a result, ethanol concentration in the
broth is kept at low level to facilitate the continu-
ous fermentation process as well as enhance the
solvent productivity. In addition, a microltration/ References
ultraltration unit can be alternatively installed
before the pervaporation module, in order to lter Peng P, Shi B, Lan Y (2011) A review of membrane
microbes and avoid the biofouling of materials for ethanol recovery by pervaporation. Sep
pervaporation membranes.
Vane LM (2005) A review of pervaporation for product
PV membranes can be made from either poly- recovery from biomass fermentation processes. J Chem
meric or inorganic materials, even both of them Technol Biotechnol 80:603629
(Peng et al. 2011). The polymeric pervaporation Xiangli FJ, Wei W, Chen YW, Jin WQ, Xu NP (2008) Opti-
membranes for extracting ethanol from fermenta-
siloxane (PDMS)/ceramic composite pervaporation
tion broth include polydimethylsiloxane (PDMS) membranes using response surface methodology.
membranes, poly[1-(trimethylsilyl)-1-propyne] J Membr Sci 311:2333
EthanolWater Mixtures: Separation by Pervaporation 723
Pervaporation mostly allows a variety of possible

EthanolWater Mixtures: Separation application areas: dewatering of organic uids
by Pervaporation like alcohols, ketones, ethers, etc. (EP 765682A1
1997); separation of mixtures from narrow boiling
Kew-Ho Lee temperatures to constant (azeotrope) boiling tem-
Center for Membranes, Korea Research Institute peratures (EP 811420A1 1997); removal of
of Chemical Technology (KRICT), Yuseong-gu, organic pollutants from water and air streams
Daejeon, South Korea (EP 749351A1 1996); separation of fermentation
products; and separation of organic-organic liquid
mixtures (Kujawski 2000). Pervaporation is also
E
Pervaporation is an important membrane process considered as so-called clean technologies, espe-
in chemical industries in which valuables are iso- cially well suited for the treatment and recycling
lated from the liquid mixture. Liquid and vapor of volatile organic compounds and pollution pre-
separation by thermal processes has always been vention (Anne et al. 2002).
highly energy intensive, and new separation pro- Transport mechanism of PV through poly-
cesses taking advantage of mass transfer through meric membranes was studied by many research
dense membranes have already shown they enable groups, and it was explained by the solution-
very signicant energy savings as compared to diffusion model (Binning et al. 1961; Paul and
more classic technologies (Anne et al. 2002). Paciotti 1975; Lee 1975; Mulder and Smolders
Membranes can be used for the selective removal 1984; Kataoka et al. 1991a, b). According to the
of water from aqueous organic mixtures. solution-diffusion model, each component of the
Pervaporation (PV) is a separation process that permeation molecules dissolves into the mem-
involves separation of liquid mixtures, in contact brane and diffuses through the membrane due to
with a membrane. With feed solution on one side, the concentration gradient (Mikihiro et al. 1998)
permeate is removed as a vapor from the other (Fig. 1).
side (Brian et al. 2011); pervaporation (PV) is a Transport through the membrane is driven by
very well-known membrane process for the sepa- the vapor pressure difference between the feed
ration of liquid and vapor mixtures due to its solution and the permeate vapor. The vapor pres-
energetic aspects (EP 909209A1 1999; EP sure difference can be maintained by applying a
944575A1 1999; EP 880400A1 1998). vacuum on the permeate side or by cooling the
EthanolWater Mixtures:
Separation by
Pervaporation, 1. Sorption
Fig. 1 Solution-diffusion
mechanism (Graham 1866) Micrivoids
Polymer
2. Diffusion
Microchannels
3. Desorption
724 EthanolWater Mixtures: Separation by Pervaporation
EthanolWater Mixtures: Separation by Pervaporation, Table 1 PV dehydration of ethanol through various

polymeric membranes (Brian et al. 2011)
Polymer Feed (wt% water) Temp. ( C) Separation factor Flux (g/m2h)
Regenerated cellulose 50 45 5.0 2060
Cellulose acetate 4 60 5.9 200
Teon-g-polyvinylpyrrolidone 4 25 2.9 2200
Peruorinated polymer on PAN support 1.3 50 387 1650
Naon-H+ 4 70 2.5 5000
Polyacrylonitrile-polyvinylpyrrolidone 4 20 3.2 2200
Poly(maleimide-co-acrylonitrile) 15 15 33 8
Poly(acrylic acid-co-acrylonitrile) 18 15 877 13
Polystyrene 4 40 101 5
Poly(vinyl chloride) 4 40 63 3
Alginic acid 4 40 8.8 48
5 60 13 2800
Chitosan 4 40 2208 4
Chitosan acetate salt 4 40 2556 2
Chitosan/glutaraldehyde 4 40 202 7
PVA/25 % TEOS, annealed at 160 C 15 40 329 5
PVA/25 % TEOS, annealed at 130 C 15 40 893 4
permeate vapor so that it condenses, thus creating necessary to use polymers, which contain specic
a partial vacuum. Commercial systems for the groups/active centers, capable of strong interac-
dehydration of concentrated alcohol and other tions with water.
solutions have been developed since the 1980s, The highest uxes are those for the hydrophilic
much of the push coming from interest in the membranes based on cellulose and Naon and
production of pure ethanol as an alternative liquid grafts of hydrophilic poly(vinylpyrrolidone) on
fuel, where PV can be used to dehydrate (Brian Teon and polyacrylonitrile. The PVA/TEOS
et al. 2011). membranes are exceptions in that they are hydro-
philic but exhibit low uxes (Brian et al. 2011)
(Table 1).
Pervaporation Membranes The emphasis is on selectivity, postulated to be
determined by selective sorption and selective
Polymeric Membranes diffusion. Selective sorption is governed by the
For dehydration, where the small molar volume presence in the membrane of active centers such
favors the preferential sorption of water, materials as charged sites which are capable of specic
have to be selected with a higher afnity for water interaction with water, while selective diffusion
than for the other component. The polymeric is governed by the rigidity and regularity of the
materials can be broadly classied into three cat- polymer structure and the nature of the polymer
egories: glassy polymers, rubbery or elastomeric interspace, exemplied by the degree of swelling
polymers, and ionic polymers. In general, the and the frequency of the cross-links. The results
glassy and ionic polymers are more suited for for a series of membranes made by grafting neu-
making water-selective membranes for tral or charged polymers onto supporting mem-
dehydration. branes are reported in (Table 2).
For water-selective membranes, the most Polysalts, formed from anionic and cationic
important factor responsible for the separation is polyelectrolytes, would be appropriate for
the specic interaction between water and the obtaining both highly permeable and highly selec-
polymer. To obtain high selectivities, it is tive membranes (Semenova et al. 1997) (Table 3).
EthanolWater Mixtures: Separation by Pervaporation 725
EthanolWater Mixtures: Separation by Pervaporation, Table 2 PV dehydration of 20 % aqueous ethanol at 70 C

using graft polymer membrane having different charges (Brian et al. 2011)
Host polymer Grafted polymer Graft site charge Separation factor
Polyvinylidene uoride 4-Vinylpyridine Neutral 9
Polyvinylidene uoride N-Vinyl-imidazole Neutral 10
Polyvinyl uoride N-Vinylmethyl-acetamide Neutral 4
Polyvinyl uoride N-Vinyl-pyrrolidone Neutral 7
Polyacrylonitrile Acrylic acid Neutral 10
Polyacrylonitrile K+ acrylate Anionic 500
Polyvinyl uoride K+ acrylate Anionic 156
Polyvinylidene uoride Quaternized 4-vinylpyridine Cationic 175 E
Polyvinylidene uoride Quaternized N-vinylimidazole Cationic 61
Polyvinylidene uoride 4-Vinylpyridine/BrCH2COOH Zwitterionic 76
Polyvinyl uoride Vinylimidazole/BrCH2COOH Zwitterionic 63
EthanolWater Mixtures: Separation by Pervaporation, Table 3 PV dehydration of aqueous ethanol with mem-
brane based on various polysalts (Brian et al. 2011)
Feed (wt% Temp. Separation Flux
Polyanion Polycation water) ( C) factor (g/m2h)
Poly(acrylonitrile-co- Poly(acrylonitrile-co-vinyl pyridine) 10 5000 400
acrylic acid)
Cellulose-SO3-Na+ Polyethyleneimine 16 50 295 1900
Cellulose-SO3-Na+ PolyDADMAC, linear 16 50 140 3200
Cellulose-SO3-Na+ Same, but branched 16 50 123 4900
Cellulose-SO3-Na+ Poly-N, N-dimethyl-3, 5- 16 50 123 2700
dimethylenepiperidine chloride
Aromatic polyamide Polyethyleneimine 20 60 15 300
sulphonate
Poly(acrylic acid) Chitosan
On polysulphone supporting membrane 5 30 1008 132
No supporting membrane 5 30 2216 33
Na+ polystyrene Polyallylamine HC1 6.2 70 70 230
sulphonate
Na+ CMC Chitosan 10 70 1062 1140
Na+ CMC N-Ethyl-4-vinyl-pyridinium bromide 10 70 782 1320
Anionic PVA Cationic PVA
DS 2.3 % DS 2.9 % 4.6 75 2250 378
DS 5.0 % DS 5.2 % 4.6 75 1910 284
The best performers in terms of ux, which at a Hybrid Membranes

maximum of 5 kg/m2h never achieve high values, Many a times, the polymeric membranes may fail
are charged polymers of one type or another, to meet the desired separation requirements. In
including polysalts. Anionic and polysalt mem- such cases, it becomes necessary to add ller
branes are superior. For anionic polymers, the materials such as ceramics and zeolites to improve
proton form has a signicantly higher ux than the separation properties of the membrane. There
the metal or quaternary ammonium salt versions, are several reports showing good separation per-
owing to the greater free space within the polymer formance for ethanol/water mixture using zeolite
network (Table 4). membranes (Kita et al. 1995; Sano et al. 1994).
726 EthanolWater Mixtures: Separation by Pervaporation
EthanolWater Mixtures: Separation by Pervaporation, Table 4 Highest fluxes for PV dehydration of aqueous
ethanol (Brian et al. 2011)
Separation
Membrane polymer Mem. type Feed (wt% water) Temp. ( C) Flux (g/m2h) factor
Naon-H+ Anionic 4 70 5000 2.5
Cellulose-SO3-Na+ and Polysalt 16 50 4900 125
polyDADMAC, branched
K+ acrylate graft on poly(vinyl Anionic 20 70 4700 156
uoride)
PEI/PAA on RO membrane Polysalt 10+ 70 4050 1075
Cellulose-SO3-Na+ and Polysalt 16 50 3200 140
polyDADMAC, linear
K+ acrylate graft on PAN Anionic 20 70 3000 500
EthanolWater Mixtures: Separation by Pervaporation, Table 5 PV dehydration of ethanol using PVA/inorganic

hybrid membranes (Brian et al. 2011)
Crosslinker Feed (wt% water) Temp. ( C) Separation factor Flux (g/m2h)
TEOS (160 C) 15 40 329 50
TEOS (130 C) 15 40 893 40
PEG blend and TEOS 15 50 300 46
No PEG 15 50 160 500
Poly(acrylic acid) copolymer and TEOS 15 40 250 18
g-Aminopropyl-triethoxysilane 5 50 537 36
Sulphated zirconia 5 50 263 10
10 50 142 105
20 50 86 183
30 50 61 1036
Kita et al. made NaA-type zeolite membrane by Binning RC, Lee RJ, Jennings JF, Martin EC (1961) Sepa-
hydrothermal synthesis. NaA zeolite membrane is ration of liquid mixtures by permeation. Ind Eng Chem
53:45
a water-selective membrane, and the PV separa- Brian B, Manh H, Zongli X (2011) A review of membrane
tion factor of water/ethanol system was over selection for the dehydration of aqueous ethanol by
10,000 at 348 K. For ethanol permselective mem- pervaporation. Chem Eng Process 50:227235
branes, Sano et al. (1994) prepared polycrystalline EP 749351A1 (1996) Device for separating mixtures or for
purifying substances by pervaporation
silicalite membrane by the hydrothermal EP 765682A1 (1997) Apparatus for separating liquid
synthesis. media with two membranes having their primary sides
The silicalite membrane showed high ethanol connected by an intermediate space
permselectivity, and a separation factor of 58 was EP 811420A1 (1997) Composite membrane for selective
separating organic substances by pervaporation
realized at 333 K by PV. Silicalite membranes EP 880400A1 (1998) Composite membrane with a support
seem to have great potential for the ethanol recov- membrane made in particular of a microporous material
ery by PV (Table 5). EP 909209A1 (1999) Pervaporisation and module for car-
rying out said process
EP 944575A1 (1999) Esterication of fermentation-
derived acids via pervaporation
References Graham T (1866) On the absorption and dialytic separation
of gases by colloid septa. Philos Mag J Sci 32:401420
Anne J, Robert C, Pierre LND, Bruno C (2002) Industrial Kataoka T, Tsuru T, Nakao S, Kimura S (1991a) Perme-
state-of-the-art of pervaporation and vapour permeation equations developed for prediction of membrane
ation in the western countries. J Membr Sci 206:87117 performance in pervaporation, vapor permeation and
Ethyl Cellulose (EC) Membrane 727
reverse osmosis based on the solution

diffusion can be substituted with epoxy groups up to about
model. J Chem Eng Jpn 24:326 55 %, where ethyl cellulose exhibits good thin
Kataoka T, Tsuru T, Nakao S, Kimura S (1991b) Mem-
brane transport properties of pervaporation in lm-forming ability, excellent mechanical proper-
ethanol
water system using polyacrylonitrile and cel- ties and durability, outstanding gas separation
lulose acetate membranes. J Chem Eng Jpn 24:334 performance, and cost-effectiveness. Ethyl cellu-
Kita H, Horii K, Ohtoshi Y, Tanaka K, Okamoto K (1995) lose can be dissolved in organic solvents (e.g.,
Synthesis of a zeolite NaA membranefor pervaporation
of water/organic liquid mixtures. J Mater Sci Lett 14:206 chloroform) and ethyl cellulose thin lm can be
Kujawski W (2000) Application of pervaporation and formed via solvent evaporation. The nonporous
vapor permeation in environmental protection. Pol ethyl cellulose membrane has been used to sepa-
J Environ Stud 91:1326 rate light gases, where the solubility coefcient
Lee CH (1975) Theory of reverse osmosis and some other E
membrane permeation operations. J Appl Polym Sci 1983 and the diffusion coefcient are two major param-
Mikihiro N, Takeo Y, Sin-ichi N (1998) Ethanol/water eters to determine the membrane performance for
transport through silicalite membranes. J Membr Sci gas separation. This performance is also
144:161171 inuenced by the content of epoxy groups,
Mulder MHV, Smolders CA (1984) On the mechanism of
separation of ethanol/water mixtures by pervaporation which vary with different manufacturers (Houde
I. Calculations of concentration proles. J Membr Sci et al. 1997; Li et al. 2001; Ito et al. 1989) (Fig. 1).
17:289 Another important application of ethyl cellu-
Paul DR, Paciotti JD (1975) Driving force for hydraulic lose is in lm coating of drug tablets. Although the
and pervaporative transport in homogeneous mem-
branes. J Polym Sci 13:1201 tablet coating does not have a therapeutic value, it
Sano T, Yanagishita H, Kiyozumi Y, Mizukami F, Haraya K does offer the psychological importance in aiding
(1994) Separation of ethanol/water mixture by silicalite patients to take the medication. The incorporation
membrane on pervaporation. J Membr Sci 95:221 of color, avor, and other additives in the ethyl
Semenova SI, Ohya H, Soontarapa K (1997) Hydrophilic
membranes for pervaporation: an analytical review. cellulose-coated tablets can improve esthetic and
Desalination 110:251286 organoleptic values and improve stability and
handling capability and drug release properties.
For example, ethyl cellulose is a water-insoluble
polymer, which can be used to protect the hydro-
Ethyl Cellulose (EC) Membrane lysis of the drug by surface coating. The drug
release rate can be adjusted by varying the thick-
Hongyang Ma, Benjamin S. Hsiao and ness of the lm and/or the composition of hydro-
Benjamin Chu phobic and hydrophilic components in the coating
Department of Chemistry, Stony Brook (Sakellariou et al. 1995). Similarly, ethyl cellulose
University, Stony Brook, NY, USA can be used to formulate microcapsules where the
target drug(s) can be tailored with controlled
Ethyl cellulose (EC) membrane has mainly been release rate, as well as improved efcacy, safety,
used for separation of light gases, such as He, O2, processibility, and stability (Rogers et al. 2012). It
N2, CH4, CO2, propane, and propylene. The is interesting to note that hydrophobic ethyl cel-
hydroxyl groups on the glucose unit of cellulose lulose and hydroxypropyl cellulose, which is
OR
OH
RO OR
HO OH Etherification O
O O O
O O
RO OR O
HO OH O
OR n
OH n
R = H or CH2CH3
Ethyl Cellulose (EC) Membrane, Fig. 1 Etherication of cellulose with maximum content of epoxy of about 55 %
728 Ethylene Off-Gas
more hydrophilic, are often combined together Renery and petrochemical ethylene-
synergistically for varying biomedical applica- containing off-gas streams usually contain vary-
tions (Donbrow et al. 1980). ing amounts of heavy C4C10 hydrocarbons,
including alkanes, olens, and aromatics, together
with small quantities of water, nitrogen oxides,
References carbon monoxide, carbon dioxide, acetylene,
methylacetylene, propadiene, butenes, and higher
Donbrow M, Sanuelov D (1980) Zero order drug delivery hydrocarbons. FCC off-gases are available at
from double-layered porous lms: release rate proles
1017 bar and typically contain 818 vol.% of
from ethyl cellulose, hydroxypropyl cellulose and
polyethylene glycol mixtures. J Pharm Pharmacol ethylene, 39 vol.% of propylene, and 1220
32:463470 vol.% of hydrogen, depending on the FCC feed
Houde AY, Stern SA (1997) Solubility and diffusivity of composition and cracking severity (Netzer 2003).
light gases in ethyl cellulose at elevated pressures:
In general, ethylene-bearing off-gas streams con-
effects of epoxy content. J Membr Sci 127:173183
Li XG, Kresse I, Xu ZK, Springer J (2001) Effect of tain relatively dilute concentrations of olens,
temperature and pressure on gas transport in ethyl cel- such as ethylene, with methane, ethane, and
lulose membrane. Polymer 42:68016810 hydrogen which are potentially recoverable.
Ito A, Hwang ST (1989) Permeation of propane and pro-
Hydrogen and light hydrocarbons (ethylene,
pylene through cellulosic polymer membranes. J Appl
Polym Sci 38:483490 propylene, and LPG) have an increasing world-
Sakellariou P, Rowe RC (1995) Interactions in cellulose wide demand. Being more valuable as chemical
derivative lms for oral drug delivery. Prog Polym Sci and plastic feedstocks than they are as fuels, the
20:889942
recovery from off-gas streams of these valuable
Rogers TL, Wallick D (2012) Reviewing the use of
ethylcellulose, methylcellulose and hypromellose in compounds represents an interesting possibility to
microencapsulation. Part 3: applications for microcap- increase revenue in reneries (Hoffmann and
sules. Drug Dev Ind Pharm 38:521539 Kaufmann 2012). Indeed, renery and petro-
chemical ethylene off-gas streams are typically
burnt in are stacks or as fuel, since the recovery
of the olens by conventional means, such as
Ethylene Off-Gas fractionation, is considered as not economically
viable. Considering the high cost of producing
Paola Bernardo ethylene by thermal cracking of hydrocarbon
Institute on Membrane Technology, National feedstocks, which is the primary production pro-
Research Council of Italy, ITM-CNR, Rende, cess for ethylene, the recovery of olens repre-
Italy sents a substantial conservation of resources
(Eldridge 1993). The available separation
methods are absorption, adsorption, distillation,
Ethylene off-gas is a light gas mixture containing and membrane separation. Membrane technology
various concentrations of ethylene, propylene, represents a cost-effective, exible, and safe pro-
hydrogen, methane, ethane, propane, and minor cess for recovering ethylene and associated
amounts of higher hydrocarbons, nitrogen, and hydrogen from renery and petrochemical
other trace components. off-gas streams (Ohlrogge et al. 2010).
The main ethylene-bearing off-gas streams are The use of a membrane system for vaporgas
produced during the petroleum rening from uid separation allows valuable feedstocks to be recov-
catalytic cracking (FCC) units and delayed ered and recycled to the polymerization section in
cokers. FCC units alone produce signicant polyethylene and polypropylene production
amounts of ethylene and hydrogen, owing to the (Baker et al. 1998). Polyolen plants involve pre-
high feed ow rates processed. Other sources for sent losses of monomers and other hydrocarbon
ethylene off-gas streams are the deep catalytic feedstocks (typically $1$3 million per year in
cracking processes. one plant). Vent streams, containing monomer
Ethylene Production by Membrane Operations 729
contaminated with light gases (e.g., N2 and H2),

are usually ared. Monomer recovery with Ethylene Production by Membrane
hydrocarbon-selective rubbery membranes is the Operations
largest application of vapor separation mem-
branes. The vent stream is compressed and cooled Paola Bernardo
to condense hydrocarbons. The gas not con- Institute on Membrane Technology, National
densed, still containing a signicant amount of Research Council of Italy, ITM-CNR, Rende,
hydrocarbons, is sent to the membrane system Italy
and separated into a hydrocarbon-enriched stream
as permeate and a nitrogen stream. The permeate
E
is recycled to the compressor and then to the Ethylene (C2H4), also called ethene, is a colorless
condenser where the hydrocarbon is recovered; and ammable gas derived from natural gas and
the nitrogen stream is recycled to the degassing petroleum. Being particularly reactive, ethylene rep-
bin (Baker et al. 1998). resents one of the largest volume compounds and a
Analogous membrane processes are applied to key building block in the petrochemical industry,
recover ethylene in ethylene oxide and in vinyl with a worldwide demand that has grown steadily.
acetate production. The purge gas for a typical The conventional ethylene production process
ethylene oxide plant contains approximately is steam cracking. A hydrocarbon feedstock (such
2030 % ethylene, 1012 vol.% argon, 110 as naphtha, liqueed petroleum gas, ethane, pro-
vol.% carbon dioxide, 13 vol.% ethane, pane, or butane) is thermally cracked in the pres-
50 vol.% methane, and 45 vol.% oxygen. ence of steam, producing a mixture of hydrogen,
A similar vent gas mixture is produced in the methane, ethylene, ethane, and heavier compo-
vinyl acetate process. A membrane recovery unit nents. Most of the dilution steam and heavier
fed with these purge streams will produce an products are condensed by a rapid quenching,
ethylene-enriched permeate stream and an while the cooled gases are compressed, sent to a
argon-enriched residue stream (Jacobs and Billig scrubbing for the acid gas removal, and subcooled
2005). These systems achieve an ethylene for the successive separations. A distillation train
recovery greater than 70 %, partly recover meth- performs the separation and purication of the
ane (the diluent gas), and reduce hydrocarbon gaseous products, representing an energy con-
burning. suming step, highly heat integrated and requiring
refrigeration (Ren et al. 2006). This provides a
strong drive for evaluating alternative methods
References for separation and recovery purposes. Coke for-
mation requires discontinuous operation for reac-
Baker RW, Wijmans JG, Kaschemekat JH (1998) The tor cleanup. Different wastes are generated
design of membrane vapour-gas separation systems. including caustic scrubber efuent, dilution
J Membr Sci 151:5562 steam condensate, coke and tar at the condensate
Eldridge RB (1993) Olen/parafn separation technology:
separators, and vent gases generated during start-
a review. Ind Eng Chem Res 32:22082212
Hoffmann K, Kaufmann D (2012) Recovery of valuable ups and shutdowns. Overall, this process is very
olen products from renery off-gas streams. Chem capital intensive and energy consuming (Ren
News 1419 et al. 2006). This type of hydrocarbon feedstock
Jacobs ML, Billig BJ (2005) Achieving ethylene ef-
ciency. Hydrocarb Eng
affects the process capital cost as well as the
Netzer D (2003) A combination process for manufacturing amounts and types of waste streams. Cleaner and
ethylene, ethylbenzene and styrene. EP Patent simplied processes are related to the use of eth-
1,017,651 B1 ane. Multiple companies are considering this
Ohlrogge K, Wind J, Brinkmann T (2010) Membranes for
option, driven by the large volumes of shale gas
recovery of volatile organic compounds. In: Drioli E,
Giorno L (eds) Comprehensive membrane science and being recently recovered which provides sizable
engineering, vol 2. Elsevier, Oxford, pp 213242 volumes of very low-cost ethane (Armor 2013).
730 Ethylene/Ethane Separation by Membranes
The integration of membrane separation systems continuous feeding to control the reaction selec-
and of membrane contactors (such as membrane tivity. Some examples, related to the conversion
distillation, membrane strippers and scrubbers, of ethane, include the dehydrogenation by means
etc.) and membrane reactors might benecially con- of hydrogen permeable metallic membranes
tribute to redesign industrial operations. Indeed, (Galuszka et al. 2008) and the exothermic oxida-
membrane technology has many advantages over tive dehydrogenation using perovskite mem-
other conventional technologies, including lower branes, which operate under a mixed
capital and operating costs, low maintenance, low ion-electron conduction (Lobera et al. 2011).
space requirements, exibility, and ease of installa-
tion and operation (Strathmann 2011).
An ethylene production cycle by steam crack- References
ing was redesigned by integrating different mem-
brane systems ( gas separation modules, Armor JN (2013) Emerging importance of shale gas to both
the energy & chemicals landscape. J Energy Chem
microltration, and membrane contactors) into
22:2126
a reference plant producing 800,000 t/year of eth- Bernardo P, Criscuoli A, Clarizia G, Barbieri G, Drioli E,
ylene and consuming ca. 30 GJ/t ethylene Fleres G, Picciotti M (2004) Applications of membrane
(Bernardo et al. 2004). Membrane gas separation unit operations in ethylene process. Clean Technol
Environ Policy 6(2):7895
to remove hydrogen from compressed gas allows
Engler Y, Dupuis G (2000) Process for recovering olens.
a debottlenecking, reducing signicantly the U.S. Patent 6,141,988
energy required in the cryogenic distillation for Gaeta S (2009) Membrane contactors in industrial applica-
hydrogen/methane separation, permitting less tions. In: Drioli E, Giorno L (eds) Membrane opera-
tions: innovative separations and transformations.
severe operating conditions with respect to the
Wiley-VCH, Weinheim, pp 499512
conventional cycle (Engler and Dupuis 2000). Galuszka J, Giddings T, Clelland I (2008) Catalytic dehydro-
Gas separation was also proposed to produce genation of ethane in hydrogen membrane reactor. In:
oxygen-enriched air to be used instead of air for Bose AC (ed) Inorganic membranes for energy and envi-
ronmental applications. Springer, New York, pp 299311
combustion and/or decoking. Hydrocarbon
Lobera MP, Escolstico S, Serra JM (2011) High ethylene
removal from water streams using membrane production through oxidative dehydrogenation of eth-
contactors, capable to operate more efciently ane membrane reactors based on fast oxygen-ion con-
with respect to conventional scrubbing units ductors. Chem Cat Chem 3:15031508
Ren T, Patel MK, Blok K (2006) Olens form conventional
(Gaeta 2009), could achieve a removal greater
and heavy feedstocks: energy use in steam cracking and
than 90 %. Membrane contactors could be also alternative processes. Energy 31(4):425451
applied for acid gas removal from the furnace Strathmann H (2011) Introduction to membrane science
efuent. Water used for scrubbing of the nest and technology. Wiley-VCH, Weinheim
fraction of coke particles during decoking can be
processed in a microltration membrane unit,
recovering up to 90 % of the water and allowing
its reuse in the plant. The replacement of tradi- Ethylene/Ethane Separation
tional operations with membrane units, such as by Membranes
membrane contactors for washing water purica-
tion in the hot section and the membrane gas Paola Bernardo
separation for the hydrogen recovery, determine a Institute on Membrane Technology, National
smaller exergetic loss with respect to the conven- Research Council of Italy, ITM-CNR, Rende,
tional industrial cycle (Bernardo et al. 2004). Italy
Membrane reactors with inorganic membranes
able to operate at high temperature were investi-
gated for alternative ethylene productions, The separation of ethylene from ethane and pro-
benetting from the product continuous removal pylene from propane is a challenging operation. It
to shift a chemical equilibrium or from a reactant is involved in the production of light olens, such
Ethylene/Ethane Separation by Membranes 731
as ethylene and propylene, which represent the transport membranes were studied more exten-
most important building blocks in the petrochem- sively due to low-cost and relatively weak inter-
ical industry. Cryogenic distillation, convention- actions with olens (possible decomplexation).
ally applied to this separation, is highly energy However, carrier poisoning and short life span of
intensive, requiring expensive tall columns, oper- the polymeric membranes are typically reported
ated under low temperature and high pressure, (Rungta et al. 2013). Chitosan-based membranes,
with very high reux ratios and requiring ethyl- impregnated with silver nitrate as a facilitation
ene/propylene refrigeration systems for agent, showed performance stability over thou-
low-temperature cooling (Eldridge 1993). sands of hours of operation in eld tests (Kamza
Due to the complexity and difculty of the et al. 2013). Ionic liquids, selected as additives for
E
separation process, mixtures including olens facilitated transport membranes, have a negligible
produced in the petroleum and rening process vapor pressure that avoids solvent losses by evap-
are often used as fuel. Membrane separation, rec- oration and offer more afnity for the olens
ognized as a low-energy method, appears the most compared to the parafns, providing stability to
promising among alternative separation tech- the metallic cation dissolved inside and acting as a
niques, such as extractive distillation, adsorption, medium for facilitated transport with mobile car-
and absorption, to recover these valuable com- rier (Fallanza et al. 2013).
pounds (Rungta et al. 2013). Polymeric mem- Carbon molecular sieve membranes showed
branes, cheap and with an easy processability, the potential to surpass the polymeric ethylene/
have low olen/parafn separation factor, espe- ethane upper bound performance (Rungta
cially those based on rubbery polymers (Rungta et al. 2013). These membranes are usually pre-
et al. 2013). The experimentally observed ethyl- pared by the pyrolysis of polymeric precursors
ene/ethane and propylene/propane upper bound, such as polyimide materials. However, major
based on the most reliable data of various poly- issues are represented by their brittleness and
meric membranes, were presented in two critical poor mechanical strength.
reviews (Rungta et al. 2013; Burns and Koros In the eld of inorganic membranes, metal
2003, respectively). Glassy polymers have been organic frameworks were recently considered for
studied intensively for this separation. However, preparing membranes to be applied to the olen/
when applied to mixtures and/or at high gas activ- parafn separation (Bux et al. 2011).
ities, these materials are prone to plasticization The concept of hybrid processes, combining a
which causes swelling of the polymer matrix and membrane system with a conventional one, has
results in a higher permeability coupled with a drawn signicant attention for the olen/parafn
loss of selectivity. Strategies to overcome plasti- separation. Savings in total costs and energy
cization include thermal curing and chemical (up to 30 %) could be obtained with a mem-
cross-linking, which reduce the polymer free vol- brane/distillation hybrid system (Caballero
ume, and the addition of nanollers (Goh et al. 2009).
et al. 2011; Ploegmakers et al. 2013).
Facilitated transport membranes based on the
p-complexation mechanism are broadly investi-
gated for olen/parafn separation (Faiz and Li References
2012). The specic interaction between the hybrid
Burns RL, Koros WJ (2003) Dening the challenges for
molecular orbitals of the olens and the atomic C3H6/C3H8 separation using polymeric membranes.
orbitals of the transition metals results in revers- J Membr Sci 211:299309
ible chemical bonds stronger than those formed by Bux H, Chmelik C, Krishna R, Caro J (2011) Ethene/
van der Waals forces, but can be broken by simply ethane separation by the MOF membrane ZIF-8:
molecular correlation of permeation, adsorption, diffu-
increasing the temperature or decreasing the pres-
sion. J Membr Sci 369:284289
sure. These systems allow high selectivity for Caballero JA, Grossmann IE, Keyvani M, Lenz ES
the bound component. Silver-based facilitated (2009) Design of hybrid distillation vapor membrane
732 Evapomeation (EV)
separation systems. Ind Eng Chem Res Evapomeation (EV),

48(20):91519162 Fig. 1 Principle of Feed liquid
Eldridge RB (1993) Olen/parafn separation technology: pervaporation (PV)
a review. Ind Eng Chem Res 32:2208 Membrane
Faiz R, Li K (2012) Polymeric membranes for light olen/
parafn separation. Desalination 287:8297
Fallanza M, Ortiz A, Gorri D, Ortiz I (2013) Polymerionic
liquid composite membranes for propane/propylene
separation by facilitated transport. J Membr Sci
444:164172
Goh PS, Ismail AF, Sanip SM, Ng BC, Aziz M (2011)
Recent advances of inorganic llers in mixed matrix vacuum
membrane for gas separation. Sep Purif Tech
81:243264
Kamza A, Keyvani M, Towe G (2013) Stable facilitated
transport membrane for olen/parafn separation. In: Evapomeation (EV), vacuum
Fuels and petrochemicals division, AIChE Spring Fig. 2 Principle of
Meeting and Global Congress on Process Safety, San evapomeation (EV)
Antonio. 28 Apr 2 May 2013
Ploegmakers J, Japip S, Nijmeijer K (2013) Mixed matrix
membranes containing MOFs for ethylene/ethane sep-
aration. Part A: membrane preparation and characteri- Membrane
zation. J Membr Sci 428:445453
Rungta M, Zhang C, Koros WJ, Xu L (2013) Membrane-
based ethylene/ethane separation: the upper bound and
beyond. AIChE J 59(9):34753489 Feed liquid
Evapomeation (EV) A novel membrane separation technique known

as evapomeation (EV) (Uragami et al. 1988;
Tadashi Uragami Uragami and Saito 1989; Uragami 1991, 1992,
Organization for Research and Development of 1993a, b; Uragam 1994; Uragami 1998, 2005,
Innovative Science and Technology (ORDIST), 2006a, b, 2008) makes use of the advantages of
Kansai University, Suita, Osaka, Japan PV but reduces the negative effects of swelling on
membrane performance. In this technique, the
feed solution is fed to the membrane without
Presently, pervaporation (PV) is applied as the directly contacting the polymer membrane. This
chosen membrane separation technique for the is accomplished by vaporizing the liquid feed so
separation of water/organic, organic/water, and that only vapor is supplied to the polymer mem-
organic/organic mixtures. However, it seems brane as shown in Fig. 2. Therefore, swelling or
that conventional PV is not the most efcient shrinking of the polymer membranes due to con-
membrane separation process for the treatment tact with the feed solutions is minimized.
of some liquid mixtures as follows. Because the The advantages of EV compared to PV are as
polymer membranes used in PV are directly in follows:
contact with the liquid feed solutions as shown in
Fig. 1, however, specically designed chemical 1. In the EV process, membranes are not in direct
and physical properties of the membrane are often contact with liquid feed mixtures as only
impaired by swelling or shrinking of the mem- vapors are supplied to the membranes. Accord-
brane due to sorption of the feed components. ingly, any swelling or shrinking of the mem-
Swelling or shrinking of the polymer membranes brane due to the feed mixtures is minimized,
is disadvantageous for the membrane perfor- and consequently an improvement in mem-
mance with respect to the separation of mixtures. brane performance may be expected.
Evaporation Casting 733
2. Because the organic liquid mixtures are vapor- Uragami T (2008) Structural design of polymer membranes
ized, interactions between component mole- for concentration of bio-ethanol. Polym J 40:485
Uragami T, Saito M (1989) Analysis of permeation and
cules are signicantly weakened, and separation characteristics: a new technique for separa-
consequently the separation performance is tion of aqueous alcoholic solution through alginic acid
remarkably improved. membranes. Sep Sci Technol 24:54
3. In EV, contaminants in a liquid feed mixture, Uragami T, Saito M, Takigawa K (1988) Comparison of
permeation and separation characteristics for aqueous
such as macromolecular solutes, can lead to alcoholic solutions by pervaporation and new
fouling of the membrane; this problem is evapomeation methods through chitosan membranes.
avoided in EV. Makromol Chem Rapid Commun 9:361
4. During EV, both temperature of the feed solu-
E
tion and the membrane surroundings can be
controlled; hence, an improvement in the per-
meation and separation characteristics of the Evaporation Casting
membrane can be achieved.
Francesco Galiano
References
Italy
Uragam T (1994) Separation method of mixed solutions.
Japanese Patent 1,883,353
Uragami T (1991) Method of separating a particular com- Evaporation casting or dry-casting involves the
ponent from its liquid solution. US Patent 4,983,303, evaporation of a solvent (or a mix of solvents)
European Patent 0,273,267
Uragami T (1992) Separation of organic liquid mixtures from a starting solution and the subsequent for-
through chitosan and chitosan derivative membranes mation of a polymeric membrane by precipitation.
by pervaporation and evapomeation methods. In: Brine In this process a polymer is dissolved in a suitable
CJ, Sandford PA, Zikakis JP (eds) Advance in chitin solvent and the solution obtained is spread out
and chitosan. Elsevier Applied Science, Oxford,
pp 594603 across an appropriate support. Then, the solvent
Uragami T (1993a) Method of separating a particular com- is left to evaporate, in inert atmosphere or con-
ponent from its liquid solution. European Patent trolled environment, inducing the polymer precip-
0,273,267 itation and generating a membrane generally with
Uragami T (1993b) Method of separating a particular com-
ponent from its liquid solution. Brazilian Patent a dense structure (Fig. 1). In evaporation casting,
P-8707041-3 in fact, the precipitation process is much slower
Uragami T (1998) Structures and properties of membranes than the precipitation obtained by immersion cast-
from polysaccharide derivatives. In: Dumitriu S ing. As a consequence, the membranes present,
(ed) Polysaccharide structural diversity and func-
tional versatility. Marcer Dekker, New York/Basel/ usually, an isotropic and less porous structure.
Hong Kong, pp 887924 Some other components, called commonly
Uragami T (2005) Structures and functionalites of mem- nonsolvents, can be also added to the initial poly-
branes from polysaccharides derivatives. In: Dumitriu meric solution. In this case, the solvent, the more
S (ed) Polysaccharides structural diversity and func-
tional versatility, 2nd edn. Marcer Dekker, New York/
Basel/Hong Kong, pp 10871122
Uragami T (2006a) Polymer membranes for separation of
organic liquid mixtures. In: Yampolskii Y, Pinnau I,
Freeman BD (eds) Materials science of membranes
for gas and vapor separation. John Wiley & Sons,
Chichester, pp 355372
Uragami T (2006b) Separation materials derived from
chitin and chitosan. In: Uragami T, Tokura S (eds) Mate-
rial science of chitin and chitosan, KODANSHA.
Springer, Tokyo/Berlin/Heidelberg/New York, Evaporation Casting, Fig. 1 Evaporation casting
pp 113163 process
734 Ex Situ Zeolite Synthesis
volatile element of the system, will evaporate gap (i.e., distance between the spinneret and the
faster leading to a higher polymer/nonsolvent coagulation bath) through which the external sur-
concentration responsible for the polymer precip- face experiences coalescence and orientation of
itation. The evaporation process can proceed until polymer aggregates (wet/dry spinning). In gen-
the membrane has completely formed or it can be eral, the nascent ber is partially coagulated by
stopped by immersing the cast lm in a coagula- the internal coagulant (bore liquid) and/or through
tion bath containing a nonsolvent (Baker 2004). the air gap by solvent evaporation, and nally the
By prolonging or reducing the evaporation time coagulation is completed in the coagulation bath
before immersion in the coagulation bath, it is by an external coagulant.
possible to tune the pore size of the membrane. The type of the external coagulant affects con-
When the starting polymeric solution is cast on siderably the structural morphology of the
different support, as porous thin lms or other resulted ber. Usually, water is the preferred
kinds of membranes, an asymmetric membrane external coagulant because of its low cost and
with a dense skin layer deposited on a porous environmental friendliness. The proper choice of
support can be obtained. the coagulant is very important because the rate of
demixing (i.e., phase separation) and the resultant
inner and outer surface structures of the hollow
References ber strongly depend on the chemistry and the
composition of the coagulant. The molecular
Baker W (2004) Membrane technology and applications, sizes and solubility parameters of the used solvent
2nd ed. Wiley, West Sussex, England, Chapter 3,
and the internal and the external coagulant play
pp 112113
important roles on hollow ber morphology. Sol-
vent of a large size may have difculties to leach
out during the coagulation step. The difference in
solubility parameters between the spinning dope
Ex Situ Zeolite Synthesis solution and the bore liquid (internal coagulant) or
the external coagulant affects the coagulation rate
Seeded Hydrothermal Synthesis for Zeolite and subsequently the porosity of the spun hollow
Preparation ber membrane.
The interactions between solvent and external
coagulant and between solvent and polymer play
important roles affecting the rate of polymer pre-
External Coagulant in Fiber cipitation. The temperature of the external coagu-
Preparation lant affects also the polymer precipitation rate.
The maximum amount of solvent that can be
Khayet Mohamed added to the non-solvent in the coagulation bath
Department of Applied Physics, Universidad can be roughly determined by the position of the
Complutense de Madrid, Madrid, Spain binodal (liquid-liquid boundary separating
the miscible region from the immiscible one) in
the ternary phase diagram presenting the compo-
External coagulant is a non-solvent liquid or a sition of the polymer, solvent, and non-solvent.
mixture of solvent/non-solvent in a collecting When the binodal shifts toward the polymer/sol-
bath (coagulation bath) used for fabrication of vent axis, more solvent can be added in the exter-
bers by the dry-/wet-spinning technique or the nal coagulant. It is also possible to change from
wet-spinning technique. Coagulation of the exter- porous to nonporous external ber surface by
nal surface of the nascent ber can start immedi- adding solvent to the coagulation bath (Mulder
ately after its extrusion from the spinneret (wet 1992). For example, when a strong coagulant
spinning) or after the as-spun ber crosses an air such as water for poly(vinylidene uoride),
External Coagulant in Fiber Preparation 735
External Coagulant in Fiber Preparation, (b) internal coagulant water and external coagulant 50 %
Fig. 1 Cross-sectional scanning electron microscopy ethanol in water (Reprinted from (Khayet et al. 2002).
(SEM) structure of PVDF hollow ber membranes pre- Copyright 2002, with kind permission from Elsevier)
pared with different coagulants: (a) water coagulant;
PVDF, polymer is used for hollow ber fabrica- ber membranes can be explained on the basis of
tion, a dense and smooth surface with no obvious the mutual diffusivity of solvent/non-solvent
pores is formed. However, when a weaker coagu- exchange and solubility parameters of the mate-
lant is employed such as methanol, which may be rials involved in hollow ber spinning. The addi-
mixed with water, the roughness of the formed tion of ethanol in water reduces the diffusion of
hollow ber surface may increase considerably non-solvent into the nascent hollow ber mem-
due partially to the formation of pores. These are brane and consequently decreases the rate of pre-
the reasons for using the coagulants: 80 wt.% cipitation. Khayet et al. (2002) observed an
N-methylpyrrolidone (NMP) aqueous solution increase of the external surface roughness of
by Wang et al. (2008) for preparation of PVDF PVDF hollow bers when ethanol was added to
hollow ber membranes and 80 wt.% methanol the external coagulant and attributed the result to
aqueous solution by Bonyadi and Chung (2007) the increase of the pore size at the outer surface as
for fabrication of dual hydrophilic/hydrophobic ethanol was added in the coagulation bath.
hollow ber membranes for direct contact mem-
brane distillation (DCMD) process. Figure 1
shows the effects of the external coagulant on
the cross-sectional structure of porous hydropho- References
bic hollow ber membranes (Khayet et al. 2002).
Water and 50 % (by volume) ethanol in water Bonyadi S, Chung TS (2007) Flux enhancement in mem-
brane distillation by fabrication of dual layer
were used as internal and external coagulants. In hydrophilic-hydrophobic hollow ber membranes.
Fig. 1a prepared with water as internal and exter- J Membr Sci 306:134146
nal coagulants, long nger-like voids are formed Khayet M, Feng CY, Khulbe KC, Matsuura T (2002) Prep-
near the inner surface, while smaller cavities are aration and characterization of polyvinylidene uoride
hollow ber membranes for ultraltration. Polymer
formed near the outer surface. Between the inner 43:38793890
and outer layers, sponge-like structure appears. In Mulder M (1992) Basic principles of membrane technol-
Fig. 1b the outer layer is eliminated when ethanol ogy. Kluwer Academic Publishers, Dordrecht
is added to the external coagulant. The addition of Wang KY, Chung TS, Gryta M (2008) Hydrophobic PVDF
hollow ber membranes with narrow pore size distri-
ethanol delays the external coagulation process bution and ultra-skin for the fresh water production
and the nger-like structure changes to a sponge- through membrane distillation. Chem Eng Sci
like structure. The slow coagulation of the hollow 63:25872594
736 Extracellular Polymeric Substance (EPS)
power industry, microelectronics, and food indus-

Extracellular Polymeric Substance tries to water purication (Flemming et al. 2009).
(EPS) In water supply systems, such problems are due to
either organic or biological fouling of reservoirs,
Loreen O. Villacorte pipelines, media lters, and separation membranes.
FMC Technologies, Separation Innovation and EPS accumulation in membrane ltration sys-
Research Center, Arnhem, The Netherlands tems can cause increase of operating pressure and
cleaning frequency due to blockage of membrane
pores as well congestion along the feed channel
Extracellular polymeric substances or EPSs are (Flemming et al. 1997). In NF/RO systems, it can
biosynthetic polymers from prokaryotic directly cause decline in permeate water quality
(bacteria, archaea) and eukaryotic (algae, fungi) (e.g., salt rejection) due to hindered back diffusion
microorganisms, which either form (loose or of rejected salts and can indirectly cause mem-
tight) slimes around the microbial cells or are brane damage due to long exposure of cleaning
excreted as discrete gels to the surrounding envi- chemicals and very high feed channel pressure
ronment. Typically, EPSs are heterogeneous mix- drop (Vrouwenvelder et al. 2011). It has been
tures of polysaccharides, proteins, nucleic acids, proposed that the removal of planktonic EPS
lipids, and other polymeric compounds. The from the feedwater by pretreatment with dissolved
highly diverse chemical composition of EPS is a air otation (DAF) and/or UF/MF membranes can
result of the different processes related to their effectively minimize biofouling in NF/RO sys-
production and their immediate environment: tems. However, this is still subject to extensive
active microbial secretion, shedding of cell sur- debate and investigations.
face materials, cell lysis, and adsorption from the During algal blooms, very sticky EPS materials
environment (Wingender et al. 1999). (including TEPs) produced by phytoplankton and
EPSs are often associated with the formation of bacterioplankton can cause physically irreversible
biolms and microbial aggregates. In biolm sys- (or non-backwashable) fouling in dead-end MF/UF
tems, they are mainly responsible for binding cells systems (Villacorte et al. 2010). In this case, chem-
and other particulate materials together and to the ically enhanced backwashing (CEB) can be effec-
solid-liquid interface (Characklis and Wilderer tive in restoring membrane permeability. However,
1989). In surface water sources, suspended EPSs applying an optimal dose of coagulant
are responsible for the formation of large aggre- (conventional or in-line) prior to membrane ltra-
gates of organic and inorganic suspended mate- tion has been found to effectively minimize the need
rials including living and dead microorganisms in for chemical cleaning (e.g., Schurer et al. 2012).
the water (e.g., marine snow or mucilage). These
aggregates are held together by a type of EPS
mainly comprising of surface-active, algae- Cross-References
derived anionic polysaccharides collectively
known as transparent exopolymer particles Transparent Exopolymer Particle
(Alldredge 2002). The latter has been recently
identied as one of the major initiators of aquatic
biolms as it can form conditioning lms on the
References
solid-liquid interface conducive for bacterial Alldredge AL (2002) Marine snow. In: Nybakken JW,
attachment (Bar-Zeev et al. 2012). Broenkow WW, Vallier TL (eds) Interdisciplinary
Over the years, the important role of EPS in the encyclopedia of marine sciences. Grolier Academic
microbial activities in natural aquatic systems is Reference, Danbury
Bar-Zeev E, Berman-Frank I, Girshevitz O, Berman
widely recognized. Moreover, it has also been T (2012) Revised paradigm of aquatic biolm forma-
identied to cause serious problems in technical tion facilitated by microgel transparent exopolymer
systems ranging from the shipping industry, particles. Proc Natl Acad Sci USA 109(23):91199124
Extracorporeal Blood Oxygenation Devices, Membranes for 737
Characklis WG, Wilderer PA (1989) Glossary. In: within tissues and the transport of carbon dioxide
Characklis WG, Wilderer PA (eds) Structure and func- in the opposite direction. The heart and lungs
tion of biolms. Wiley, Chichester, pp 369371
Flemming H-C, Schaule G, Griebe T, Schmitt J, work together to ensure the circulation and the
Tamachkiarowa A (1997) Biofouling the Achilles exchange of the respiratory gases. During circu-
heel of membrane processes. Desalination 113:215225 lation, the heart pumps oxygen-depleted blood to
Flemming H-C, Murthy PS, Venkatesan R, Cooksey KE the lungs and then receives oxygen-enriched
(eds) (2009) Marine and industrial biofouling,
vol 4, Springer Series on Biolms. Springer, Berlin blood from the lungs for distribution to the rest
Lens P, OFlaherty, Moran AP, Stoodley P, Mahony T (eds) of the body. In the lungs, the gas exchange pro-
(2003) Biolms in medicine, industry and environmen- cess takes place in grapelike clusters of air sacs
tal biotechnology: characteristics, analysis and control. known as the alveoli. There are about 300 million
IWA Publishing, London E
Schurer R, Janssen A, Villacorte L, Kennedy MD alveoli in the lungs with a combined surface area
(2012) Performance of ultraltration & coagulation in of approximately 70 m2. This is the area through
an UF-RO seawater desalination demonstration plant. which the bloodgas exchange takes place
Desalination Water Treat 42:5764 (Weibel 2009). Fresh air entering the lung carries
Villacorte LO, Schurer R, Kennedy MD, Amy G,
Schippers JC (2010) The fate of transparent oxygen with a partial pressure of oxygen (PO2) of
exopolymer particles in integrated membrane systems: 160 mmHg; the presence of moisture in the lung
a pilot plant study in Zeeland, The Netherlands. Desa- alveoli results in reduction of the PO2 to
lination Water Treat 13:109119 104 mmHg. The same air that enters the lung
Vrouwenvelder JS, Kruithof J, Van Loosdrecht M (2011)
Biofouling of spiral wound membrane systems. IWA carries carbon dioxide with a partial pressure of
Publishing, London carbon dioxide (PCO2) of 0.3 mmHg. The carbon
Wingender J, Neu TR, Flemming H-C (eds) (1999) Micro- dioxide delivered to the lung from the blood in
bial extracellular polymeric substances: characteriza- the venous ends of the pulmonary capillaries
tion, structure, and function. Springer, Berlin
raises the PCO2 in the alveoli to 40 mmHg
(McArdle et al. 2005). At the arterial ends of
the pulmonary capillaries, oxygen diffuses from
the air in the alveoli into the blood, and carbon
Extracorporeal BALs dioxide diffuses from the blood into the alveoli
because of differences in partial pressures. The
Bioarticial Liver Support System (BLSS) heart then receives the oxygenated blood from
the lungs and distributes it to tissues in the rest of
the body. In most tissues, the PO2 is less than
40 mmHg and the PCO2 is greater than
Extracorporeal Blood Oxygenation 45 mmHg so, because of the differences in partial
Devices, Membranes for pressures, oxygen diffuses out of the arterial ends
of tissue capillaries into the cells, and carbon
Mnica Faria1 and Maria Norberta De Pinho2 dioxide diffuses out of the cells into the blood.
1
Chemical Engineering, Columbia University, The heart pumps the oxygen-depleted blood back
New York, NY, USA to the venous ends of the pulmonary capillaries
2
Chemical Engineering, Instituto Superior entering the alveoli with a PO2 of 40 mmHg and
Tcnico, Universidade de Lisboa, Lisbon, PCO2 of 45 mmHg (Fox 2007).
Portugal In June of 2011, the World Health Organiza-
tion classied heart diseases as the worldwide
number one cause of death (WHO 2011). The
Introduction to Extracorporeal Blood treatment of many cardiac-related diseases such
Oxygenation as coronary artery bypass grafting, valve repair,
and aortic aneurysm repairs involves complex
In physiology, respiration is dened as the trans- open-heart procedures where surgeons need to
port of oxygen from the outside air to the cells operate on a motionless heart in an almost
738 Extracorporeal Blood Oxygenation Devices, Membranes for
Extracorporeal Blood Oxygenation Devices, Membranes for, Fig. 1 Main components and the function of the
extracorporeal membrane blood oxygenation system
bloodless surgical environment. Cardiopulmo- components of the bypass machine include spe-
nary bypass (CPB) is a technique in which a cial lters that capture air bubbles and a heat
heartlung machine temporarily takes over the exchanger that cools the blood during surgery
function of the heart and lungs during surgery, and warms it at the conclusion of surgery to pre-
maintaining the circulation of blood as well as pare for returning the body to its own cardiovas-
the oxygen and carbon dioxide contents of the cular circulation. The blood is cooled to maintain
body. Figure 1 shows the main components and the bodys metabolism at a lower rate, reducing
the function of the heartlung machine. In the the oxygen consumption. At the conclusion of the
operating theater, the heartlung machine pro- operation, the surgeon withdraws the cannulas
vides mechanical circulatory replacement of the and restores the ow of blood in the body and
heart by means of a pump and of the lungs through the heartbeat.
a blood oxygenator (BO). During CPB, the sur- The heartlung machine is also used in inten-
geon inserts large catheters or cannulas into the sive care units and cardiac catheterization labora-
vena cava and the aorta and then administers a tories, as an extracorporeal life support (ECLS)
chemical solution to cause the heart to stop beat- system for maintaining blood ow and respira-
ing (cardioplegia). Oxygen-depleted blood is tion. The diseased heart or lung(s) is replaced by
diverted from the upper chambers of the heart to this technology, providing time for the organ to
a reservoir in the heartlung machine. From there, heal. The heartlung machine can also be used
the blood is then transferred to the BO which with venoarterial extracorporeal membrane oxy-
infuses the blood with oxygen and allows carbon genation (ECMO), which is used primarily in the
dioxide to diffuse in the opposite direction. Other treatment of lung disease and is considered a
standard therapy for the treatment of respiratory body inammatory response (Kirklin et al. 1983;
failure in neonatal patients (Grenvik et al. 1999; Edmunds and Colman 2006). Although the safety
Roy et al. 2000). of CPB has been established with low mortality
In adult and pediatric patients, clinical ECMO rates, it is still associated to risks such as blood
is a treatment of last resort for individuals who clots and air bubbles that can cause embolism
would otherwise die despite maximal therapy (occlusion of a blood vessel), hemolysis or dam-
(Bartlett et al. 2000). Since the 1990s, this treat- age to red blood cells, and immune system acti-
ment has grown continuously in the treatment of vation or systemic inammatory response.
adult patients with respiratory dysfunction, for Studies show that presently patients undergoing
oxygenation and carbon dioxide removal, and as cardiac surgery with CPB can still experience
E
a lung-protective ventilatory strategy in order to systemic inammatory response syndrome
rest the lungs and provide optimal conditions (SIRS) (Segal et al. 1998; Onorati et al. 2010).
for the recovery of lung function (Ghosh This results in severe and irreversible effects such
et al. 2009). Despite the current limited applica- as cerebral injury and cognitive impairment (Lund
tion of ECMO in adults, a randomized controlled et al. 2005), endothelial injury (Verrier and Boyle
trial (RCT) is recruiting patients in the United 1996), organ dysfunction and impaired hemosta-
Kingdom to evaluate the cost-effectiveness and sis (Hsu 1997) also known as the process of blood
clinical benet of modern ECMO technique clotting, blood coagulation or thrombus forma-
(Peek et al. 2006). This RCT will be one of the tion, graft occlusion, perioperative strokes and
rst RCTs performed in adults since the bench- pulmonary thromboembolism (Valley
mark NIH trial during the 1970s (Zapol et al. 2010), and even postoperative morbidity
et al. 1979) and could reveal an increase in the and mortality (Hammon 2008; Vallely
rate of survival of patients with severe, but poten- et al. 2010).
tially reversible, respiratory failure without severe In order to prevent in-circuit thrombosis,
outcomes of disability, if technical and clinical or clotting, patients undergoing CPB or
advancements improve outcomes. ECMO require the administration of systemic
anticoagulation agents. Unfractionated heparin,
Clinical Practice, Risks, and History which suppresses the action of thrombin, the key
of Extracorporeal Blood Oxygenation enzyme in the thrombotic portion of the defense
Blood is a complex tissue designed to sustain the reaction, has been the elected anticoagulant for
body by continuous circulation within a vast net- over half a century. Although the results are gen-
work of blood vessels coated with endothelial erally satisfactory, there are sometimes controver-
cells. These unique cells simultaneously maintain sial effects due to (i) thrombin being, in certain
the uidity of blood and ensure integrity of the patients, only partially suppressed by heparin
vascular system. Contact with the surgical wound (Janvier et al. 1996; Edmunds and Colman
and diversion of blood into the heartlung 2006) or (ii) major hemorrhaging even after the
machine trigger a massive defense reaction that patients are subjected to correction of the
stimulates all types of blood cells and ve plasma coagulopathy with antibrinolytic drugs
protein systems to produce a myriad of vasoactive (Chalwin et al. 2008). About 3070 % of patients
cytotoxic and cell-signaling substances into the undergoing coronary artery bypass grafting under
circulation. Platelets, neutrophils, monocytes, CPB require homologous blood transfusions
and endothelial cells are the major cellular actors, (Scott et al. 1992). The major risks such as
and complement, contact, intrinsic coagulation, SIRS, activation of the blood coagulation cascade,
extrinsic coagulation, and brinolytic protein sys- blood cell trauma, thrombosis, clotting,
tems are the primary plasma participants. When hemorrhaging, etc., which patients submitted to
activated, these cells and proteins initiate complex CPB and EMCO encounter, are all due to the
and overlapping reactions and interactions with a contact of the blood with foreign surfaces of mate-
multitude of target molecules to create a whole rials with inadequate blood compatibility also
known as hemocompatibility. Among these sur- DeWallLillehei bubble oxygenator (Lillehei

faces contacting blood in the extracorporeal blood et al. 1955; DeWall et al. 1966) was a major
oxygenation system, the membranes for the gas breakthrough and well accepted by surgeons. It
exchange are of primordial importance due to was a simple disposable device where oxygen
their extensive surface area. bubbles were dispersed into an oxygenation
The start of the historical development of chamber containing venous blood that was
extracorporeal circulation is strongly related to pumped from the patients body and then was
Csar Julian Jean Le Gallois (17701814). In the defoamed in a defoaming compartment before
monograph published in 1812, Experiences sur le returning to the patient. Although the benets of
principe de la vie, he suggested that a part of the bubble versus lm BOs have been often debated,
body might be preserved by a mechanical heart both have their own limitations. For lm BOs,
replacement and some kind of external perfusion though the blood surface area was large, the gas
device (Le Gallois 1812). After the rst idea of transfer efciency was often compromised by
external perfusion devices by Le Gallois, the rst channeling of the blood ow (Niiya and Noble
attempts of oxygenating blood outside the body 1985; Ho and Sirkar 1992) and a very large prim-
were made by the physiologist Alexander ing volume was usually required to obtain suf-
Schmidt, in 1876, when he perfused an isolated cient gas exchange. On the other hand, bubble
dog kidney with oxygenated blood (Kramme BOs caused signicant damage to the blood due
2011). The original BO, invented by von Frey to the foam formation with devastating conse-
and Gruber (Kay 1992), in 1885 was the lm quences such as denaturation of plasma proteins,
BO; it used rotating cylinders to spread blood in gas emboli, fat emboli, brin emboli, and
a continuously renewed thin lm while oxygen neurocognitive decits (Lee et al. 1961, 1969;
owed over it. Though there was direct contact Hill et al. 1969). Despite the nonphysiological
between the blood and gas phases, the gas conditions of the bloodgas interface present in
exchange efciency was low, leading to devices bubble oxygenators, these were used until
that had very large priming volumes. Using the 15 years ago in around 23 % of open-heart sur-
same principle, John Gibbon, best known for gery worldwide (Wiese 2010).
inventing the heartlung machine, developed his
rst lm BO in 1937 (Gibbon 1937) and, after
years of tests and improvements, performed the Membrane Blood Oxygenators
rst successful open-heart surgery with extracor- and Principle of Membrane Blood
poreal blood oxygenation in 1953 (Gibbon 1954; Oxygenation
Edmunds 2004). After this, several forms of sup-
plying oxygen to the blood were attempted, with Membrane blood oxygenators (MBOs) were rst
more or less success, enabling the development of developed when the scientists introduced a
many models of oxygenators of which only a few nonporous symmetric membrane between the
were clinically employed such as the bubble BOs. blood and gas phases to avoid direct contact
In these devices, the gas exchange efciency was between the two phases in an attempt to improve
increased by dispersing bubbles of oxygen the hemocompatibility of BOs (Kolff and Balzer
directly into the blood, which resulted in signi- 1955; Clowes et al. 1956; Marx et al. 1962;
cantly reduced priming volumes compared to the Dorson et al. 1968; Guidoin et al. 1978). This
lm BOs (Cahn and Li 1974). On the other hand, represented a signicant breakthrough in the
bubbling gas through blood led to foam forma- development of blood oxygenation as the mem-
tion, and therefore defoaming agents such as brane provides a barrier in the bloodgas interface
silicone compounds had to be added to the and that minimizes the risk of air embolism
blood (Kremesec 1981). In 1955, DeWall (Iwahashi et al. 2004). Microporous MBOs are,
et al. performed the rst successful CPB proce- in fact, the state-of-the-art technology for extra-
dure using a bubble BO (DeWall et al. 1966). The corporeal blood oxygenation. Another advantage
Extracorporeal Blood Oxygenated blood Membrane

Oxygenation Devices, 100 mmHg O2 O2
Membranes for, 40 mmHg CO2
Fig. 2 The principle of
membrane blood
oxygenation
200 cm3 (STP) / min CO2
250 cm3 (STP) / min CO2
Oxygen depleted blood E

40 mmHg O2 CO2 / O2
46 mmHg CO2
of MBOs is in the device design and fabrication emphasis was placed on oxygen transfer as the
which, because of the dened blood channels, controlling factor in the design of the bloodgas
allows for precisely controlled blood ow charac- exchanger system. However, in some cases,
teristics and, therefore, possibilities for optimal removal of CO2 from the blood may be a limiting
efciency in gas exchange. factor in the design (Bungay et al. 1986; Stamatialis
As in the natural lungs, transport of gas into et al. 2008).
blood is driven by the partial pressure difference
of the respiratory gases between the blood and
the gas separated by the membrane. According Development of Membrane Blood
to the pressure differences, gas diffuses through Oxygenators
the membrane from the side of high partial pressure
to the side of low partial pressure. Figure 2 sche- The very rst discovery of oxygen transfer across
matically shows the principle of membrane blood an articial membrane was made by Kolff and
oxygenation. In order to be clinically functional for Berk in 1944 when they found that venous blood
an average healthy adult weighing around 70 kg, at was being oxygenated while owing through a
rest, the MBO must deliver, at standard tempera- cellophane dialyzer that was in contact with oxy-
ture and pressure conditions, approximately gen containing dialysate (Kolff et al. 1944). This
250 mL of oxygen per minute and remove the discovery stimulated the development of the use
corresponding carbon dioxide production of about of gas permeable membranes in order to separate
200 mL per minute. The controlling factor in the the blood phase from the gas phase in the BOs.
design of these bloodgas exchange devices is the The rst plate-type membrane oxygenator was
permeability of the membranes toward oxygen and built by Clowes Jr. in 1956 (Clowes et al. 1956).
carbon dioxide. The equivalent oxygen partial Clowes membrane oxygenator was based on the
pressure in venous blood is about 40 mmHg and SkeggsLeonards (Skeggs et al. 1949; Skeggs
about 104 mmHg for arterial blood. Use of air as 2000) plate dialyzer with its dialysis membrane
the supply gas and at ambient pressure gives a replaced with a supported ethylcellulose at sheet
driving force for oxygenation of about 88 mmHg. membrane. In order to have the sufcient oxygen-
In contrast, in venous blood, the CO2 partial pres- ating capacity for perfusion in an adult, it had to
sure is about 45 mmHg which, assuming that the have a multilayered system with a very large
exchanging air has negligible CO2 content, consti- membrane surface area of approximately 25 m2.
tutes the maximum driving force for CO2 removal In 1956, Kolff developed the rst coil-type
from the blood. The initial challenge in membrane membrane oxygenator (Kolff et al. 1956b) by
technology was to produce reliable membranes substituting the cellophane membrane in his own
with high permeabilities for oxygen, and much developed articial kidney, the Kolff rotating
drum (Kolff 1954; Kolff et al. 1956a; Kolff and with an irregular structure. This membrane was
Watschinger 1956), with a nonporous symmetric originally manufactured by Fuji Systems in
dense polyethylene (PE) membrane. The blood Tokyo, and its unique structure permitted
owed in a spiral mode, whereas the oxygen enhanced secondary ow effects that improved
owed parallel to the axis of the cylinder. To use the gas exchange capability. The improved
this membrane oxygenator for an adult patient, it Kolobow oxygenator was successfully used in
was necessary to assemble eight coil units with an the clinical eld, rst as an oxygenator for CPB
enormous membrane surface area for effective gas during cardiac surgery, and later in respiratory
exchange (Kolff et al. 1956a). assist devices was the only oxygenator available
The second generation of nonporous symmet- for long-term application of ECMO for acute
ric membranes appeared in the 1960s with the respiratory failure patients (Kolobow
development of silicone rubber which displayed et al. 1971). In 1972, J. P. Hill reported the rst
increased oxygen and carbon dioxide permeabil- successful treatment of adult respiratory distress
ities. The carbon dioxide permeability of silicone with ECMO (Hill 1977), and after 3 years of this
is about ve times greater than the oxygen perme- success, the rst neonatal ECMO survival case
ability, and this partially compensates for the was described by R. H. Bartlett (Bartlett
smaller available driving force for carbon dioxide et al. 1976). This was the beginning of successful
transfer (Haworth 2003). Since the diffusion coef- ECMO application, and it has been shown to be
cient of oxygen and carbon dioxide in air is most effective in the treatment of newborn and
about four orders of magnitude higher than in pediatric patients (Bartlett et al. 1982).
blood, the gas-side mass transfer resistance was The real breakthrough in MBOs came in the
negligible. The major resistance to respiratory gas 1980s with the development of hydrophobic
transfer was due to the membrane and the blood- microporous membranes. These membranes
side concentration boundary layer (Weissman and were made from polytetrauoroethylene (PTFE)
Mockros 1969). This brought about a major and the membrane pore diameters ranged from
advance in establishing the technical feasibility 0.02 to 0.1 mm. The respiratory gases pass through
of membrane oxygenators in the 1960s and the membrane pores rather than diffuse through
1970s as the required membrane area could be the membrane material and, consequently, offer
optimized to less than 6 m2. Kolobow took advan- a much lower resistance to gas transfer than
tage of the development of silicone rubber and in the nonporous membranes. Polypropylene
1963 developed an oxygenator utilizing the same (PP) microporous membranes were introduced
conguration of Kolffs coil lung but with a dif- soon after and displaced the PTFE membranes
ferent assembly. A silicone rubber envelope because they have lower cost and better mechan-
reinforced with nylon knit was wrapped around ical properties, which offered better control of the
a central core, and pure humidied oxygen was blood channel geometry (Haworth 2003).
passed through it under a negative pressure. The In 1981, Y. Nose (Nos and Malchesky 2000)
blood owed across the at tubing parallel to the started to work on the rst microporous hollow
axis of the cylinder. The average priming volume ber membrane oxygenator the Monsanto oxy-
of these units was of 0.1 L/m2, and they could genator. Although the gas exchange characteris-
oxygenate venous blood at rate of approximately tics were reasonable, the device leaked a large
1 L/m2/min (Kolobow and Bowman 1963). In amount of plasma through the membrane walls.
1965, Bramson et al. commercialized the rst The rst commercially available hollow ber oxy-
disposable nonporous symmetric silicone MBO genators, the Capiox MBO, composed of PP
with an integrated heat exchanger (Bramson microporous membranes was developed in 1981
et al. 1965). In 1971, Kolobow (Kolobow by Terumo Corporation (Terumo Corporation,
et al. 1971) improved his rst disposable coil Tokyo, Japan) and Suma (Suma et al. 1981). In
membrane oxygenator by using membranes fab- order to avoid plasma leakage, the inner surface of
ricated from silicone rubber deposited on a fabric the microporous PP hollow bers was coated with
a silicone layer with a thickness of approximately hollow bers where the blood owed over and
0.2 mm. In these two MBOs, the blood was under the outside of the bers (extraluminal ow)
pumped inside the bundle of hollow bers while oxygen crossed from within the hollow bers
(luminal ow), while the gases circulated on the into the blood (Iatridis et al. 1985). In the 1980s, the
outer side. As the venous blood entered each ber, Kuraray Corporation (Osaka, Japan) began on the
it was surrounded by an oxygen-rich ventilating development of a new device the MENOX
gas mixture which passed through the ber walls oxygenator that contained a silicone-coated
into the bloodstream (Weissman and Mockros microporous polyolen membrane. The reason
1968; Dutton et al. 1971; Suma et al. 1981; behind the silicone coating is to reduce the plasma
Tanishita et al. 1994). It was soon found that, for leakage through the membrane pores and to correct
E
routine short-term perfusion, these types of MBOs any defects at the surface of the membrane. The
presented no real advantage over bubble models developers claim that the membranes are
mainly due to thrombus formation within the non-plasma-leaking and that the MENOX oxygen-
bers (Calaore et al. 1987; Bergdahl and ator could be used in ex vivo experiments for up to
Bergdahl 1989). In the 1980s, microporous 5 months (Nos 2001). Another coating developed
MBOs with intraluminal ow accounted for only by Medtronic Inc. (Minnesota, USA) comprised of
20 % of all blood oxygenators sold in the United an alkoxysilane/alkylsilane copolymer, preferably
States (Voorhees and Brian 1996). At that time, aminoalkylsiloxane, was used to coat microporous
the gas transfer performance of MBOs was not yet hollow ber membrane blood oxygenators,
comparable with that of bubble blood oxygena- increasing the resistance of the bers to passage
tors, and, furthermore, the MBOs were complex of blood plasma through the micropores (Plunkett
to operate. 2002).
The disadvantages presented by the MBOs trig- Nevertheless, in both cases, the coating pro-
gered a series of studies on mass transfer efcien- vides an additional layer or second skin and
cies, pressure drops, and shear stresses through introduces an additional barrier to the gas transfer,
blood ow paths in different module congura- altering therefore a severe limiting factor of the
tions. Designers of at sheet MBO modules oxygen mass transfer to the blood.
focused on incorporating passive mixing of the The technological advances of MBOs devel-
blood to reduce the blood-side resistance to gas oped during the 1980s permitted the increase of
transfer, and in 1984, an oxygenator was intro- the gas transfer efciency and reduction of
duced in the market by Cobe Cardiovascular membrane surface area and priming volume. As
(Cobe Cardiovascular, Arvada, Colorado), with a a consequence of these improvements and
high gas transfer efciency and a membrane sur- with the general migration from blood owing in
face area of only 2.5 m2. This was achieved the luminal to the extraluminal space of hollow
through the introduction of spacers over the at bers, by 1992, MBOs accounted for more than
sheet microporous membranes to induce mixing 98 % of all blood oxygenators sold in the United
on the blood side (Elgas and Gordon 1984). Exten- States.
sive studies of MBO modules with hollow ber
intraluminal blood ow (Alpha et al. 1986; State of the Art of Commercial Membrane
Gassman et al. 1987) led to the design of hollow Blood Oxygenators
ber modules with blood owing on the outside of The technical and medical progress of MBOs has
the hollow bers (extraluminal ow), while oxy- been intrinsically associated to the development
gen was administered through the inside of the of membranes that can assure physiological trans-
membranes. In 1985, the Johnson and Johnson port rates of oxygen and carbon dioxide and of
Extracorporeal Maxima Oxygenator (Johnson and module congurations that are compact with min-
Johnson Cardiovascular, Anaheim, California) was imal membrane surface area per unit volume and
presented as a highly efcient and compact with low resistance to oxygen transfer, particu-
MBO. It contained microporous cross-wound larly on the blood side.
A membrane, in a very general sense, is a to ensure that the blood cells and blood compo-
semipermeable barrier that can be made of inor- nents are not damaged.
ganic or organic materials. In MBOs, organic Early microporous MBOs contained up to
polymers are the membrane materials of selection 5.5 m2 of membrane surface area (Von Segesser
due to their physicochemical and structural versa- 1987), so, in order to overcome slow gas transfer
tility. According to their structure, membranes can rates due to the high blood-side mass transfer
be divided in four main categories: nonporous resistance, modern MBOs have utilized secondary
symmetric membranes, microporous membranes, ow effects to disrupt the blood-side concentra-
integral asymmetric membranes, and composite tion boundary layer and hence reduce the mem-
asymmetric membranes. All of them can be brane surface area and priming volume. In the
processed in the form of at sheets or hollow case of at sheet MBOs, such as the Cobe CML
bers (Bungay et al. 1986; Mulder 1996; Baker 30 (Cobe Cardiovascular, Arvada, Colorado,
2004). Nonporous symmetric membranes consist USA) (Voorhees and Brian 1996), the membrane
of a dense lm where the selective transport of surface area has been reduced to between 2.5 and
gases is governed by a solution/diffusion mecha- 3.0 m2, whereas for hollow ber designs with
nism. Microporous membranes have a rigid, extraluminal blood ow such as the Cobe Optima
highly voided structure with randomly distrib- (Medtronic Maxima, Medtronic, Anaheim, Cali-
uted, interconnected pores of different shapes fornia, USA) (Ulrich 1990), the membrane sur-
and sizes from 0.01 to 10 mm in diameter face has been reduced to between 1.7 and 2.3 m2.
depending on the preparation process. Asymmet- In developed countries, the rate of growth of
ric membranes, rst developed by Loeb and the elderly population is exceeding the growth
Sourirajan in 1963 (Loeb 1981), consist of two rate of the rest of the population, and this means
layers: a dense and nonporous symmetric active that in the near future cardiac surgery will be
skin layer with thickness ranging from 0.1 to performed on older and less healthy patients,
50 mm and a relatively thick porous support layer some of which may even need to undergo more
with thickness between 50 and 500 mm. The over- than one operation (Voorhees and Brian 1996).
all performance of the membrane is governed by This changing patient population as well as clin-
the top dense layer, while the porous sublayer ical and economic factors will demand improved
provides mechanical strength to the membrane. MBOs. Designing improved and novel MBOs is a
When the material of the top layer and porous complex procedure given the interdependence of
sublayer are the same, the membrane is called an the important design variables as described
integral asymmetric membrane and when com- below:
posed of different materials known as asymmetric
composite membranes. Maximizing the gas transfer per unit priming
The packaging and arrangement of the mem- volume of the device is essential. In the past,
branes play a fundamental role in the development the patients body temperature was lowered by
of MBOs. Efcient packaging of large surface cooling the blood during CPB, thus reducing
areas of membrane into small volume devices the oxygen requirement. In current practice,
lead to MBOs of high gas transfer rates and low there is a trend toward normothermic coronary
priming volumes. Furthermore, the blood ow perfusion which increases the oxygen require-
path must be carefully designed. While disrupting ment (Galletti 1993). In addition, the need to
the blood-side mass transfer boundary layer will minimize the transfusion of donated blood due
lead to higher gas transfer efciencies, it also to possible transmission of pathogens will
leads to increased shear stresses on the blood drive module design toward lower priming
cells. Damage to blood cells depends both on the volumes (Voorhees and Brian 1996).
applied shear stress and on the time for which the Minimizing the membrane surface area of the
shear stress is applied (Zydney 1985). Thus, while device has clinical and economic benets. Dur-
mixing on the blood side is desirable, it is essential ing CPB surgery, hematological and immune
responses are observed as a result of contact carried out (Vaslef et al. 1994; Gabelman and
between the blood and oxygenator surfaces. Hwang 1999; Wickramasinghe and Han 2002;
While these side effects are relatively minor Taskin et al. 2010). Nagase et al. (2005a, b) have
in healthy patients, they are of concern for thoroughly investigated the relationship between
older and less healthy patients. An increased the oxygen transfer rate and hollow ber arrange-
gas transfer efciency would allow a reduction ment and found better mass transfer performance
in membrane surface area and hence a reduc- in parallel ow hollow bers. This was due to the
tion in the average contact time between the fact that in the hollow ber module with blood
blood and foreign surfaces. Since the mem- perpendicular ow, the hollow bers in the front
brane is often the most expensive component layer conceal those in the rear layer at the cross-
E
in a MBO, minimizing the membrane surface over point, and this decreases the mass transfer
area will reduce manufacturing costs. rates. Furthermore, the membrane surface areas of
The blood ow path must be carefully these parallel modules were larger than the ones in
designed. While disrupting the blood-side the other modules. Other studies show that high
mass transfer boundary layer will lead to gas transfer rates can be obtained when micropo-
higher gas transfer efciencies, it also leads to rous hollow ber membranes are arranged in an
increased shear stresses on the blood cells. orderly way in cross-laid double layer mats at an
Damage to blood cells depends both on the optimal membrane angle (Catapano et al. 2001).
applied shear stress and the time for which The MBOs used in ECMO for long-term lung
the shear stress is applied (Kolff et al. 1944). support system applications in infants and adults
Thus, while mixing on the blood side is desir- account for a very small part of the MBOs used
able, it is essential to ensure that the cells are worldwide. The at sheet coil-type modules are
not damaged. generally composed of a nonporous, dense at
sheet reinforced silicone rubber membrane enve-
In the last decade, rather than focusing on the lope in a spiral coil around a polycarbonate spool.
research of novel blood oxygenation membranes, The blood ows between turns of the envelope in
most researchers have been concerned with blood a thin lm, and oxygen from the gas compartment
ow congurations for mass transfer enhance- diffuses through the membrane into the blood-
ment in hollow ber MBOs. Extraluminal blood stream. At the same time, carbon dioxide diffuses
ow may be concurrent, crosscurrent, or counter- through the membrane into the gas compartment
current to the ow of gas within the bers, and all and is ushed from the oxygenator by the oxygen
these ow patterns have been studied. Some ow. More recent oxygenators aimed for long-
researchers demonstrated that a hollow ber mod- term ECMO usage are based on hollow ber
ule, in which the hollow bers are perpendicular silicone-based membranes with extraluminal
to the blood ow, is the most efcient as it pro- blood ow (Kawahito et al. 2002a, b; Motomura
motes the disruption of the momentum and con- et al. 2003).
centration boundary layer on the blood side. The Hollow ber and at sheet MBOs have almost
subsequent decrease of the resistance to gas mass equivalent hemocompatibility and gas exchange
transfer leads therefore to the increase of oxygen performance. However, the gross air handling is
transfer rates (Mockros and Leonard 1985; Yang different due to differences in the packing density
and Cussler 1986, 1986; Rajasubramanian dened by the total available membrane surface
et al. 1997; Wickramasinghe and Han 2005). area per unit volume. Gu et al. (2000) compared
However, this blood ow arrangement compli- at sheet and hollow ber MBOs in terms of
cates the design of oxygenators because of pressure drop, shear stress, and activation of leu-
the complex blood ow path and the kocytes or white blood cells. They found that both
non-applicability of conventional mass transfer congurations displayed similar gas transfer per-
correlations. As a consequence, many studies of formance. However, the pressure drop along the
momentum and mass transfer in MBOs have been blood ow path of the at sheet MBOs was higher
than that for hollow ber MBOs. Moreover, acti- In order to be technologically feasible, the MBOs,
vation of leukocytes in at sheet MBOs was with minimal membrane surface area, have to
greater. De Somer et al. (1996) compared at comply with two main properties: (i) be efcient
sheet and hollow ber MBOs in terms of hemoly- gas exchangers and provide physiological oxygen
sis, shear stress, and white blood cell and platelet and carbon dioxide transfer rates of approximately
counts and found that although they have important 250 cm3 per minute and 200 cm3 per minute,
differences in pressure drop and shear stress, no respectively, at standard temperature and pressure
statistical differences were found in hemolysis gen- conditions and (ii) be hemocompatible.
eration or other tormented blood elements. There- The majority of current MBOs are made of
fore, pressure drop as a single element may not be hollow bers of microporous hydrophobic mate-
considered to inuence hemocompatibility. rials with pore sizes below 0.1 mm. The resistance
A tremendous amount of research, as of these membranes toward the respiratory gases
evidenced by the numerous scientic articles and is very low as the gas exchange between the blood
patents, has been devoted to the design of the side and the gas side is performed by diffusion and
MBOs available today. In 1999, the total BO convection of oxygen and carbon dioxide via the
market was worth about $460 million and by open pores. MBOs with these types of membranes
2001 had risen to over $500 million (Hanft have high gas permeabilities and are routinely
2002). In 2008, an estimated 1.4 million MBOs used to provide temporary cardiopulmonary
were used worldwide for acute surgical CPB, bypass (CPB) during open-heart surgery and are
corresponding to a consumption of approximately also occasionally used for extracorporeal life sup-
3 million km2 of oxygenation membranes (Wiese port (ECLS) to substitute short-term cardiopulmo-
2009). nary function in patients with respiratory failure.
Presently, most of the MBOs sold worldwide One of the drawbacks of microporous hydropho-
contain hydrophobic microporous membranes, bic membranes is plasma breakthrough which
and more than 95 % of them are composed of occurs when plasma protein layers are formed at
extraluminal ow microporous hollow ber mem- the membrane surface and grow through the mem-
branes with pore sizes below 0.1 mm, outer diam- brane pores increasing the surface energy of the
eters between 300 and 500 mm, and surface areas pore, allowing plasma to leak through from the
of approximately 2 m2. The hollow ber mem- blood side to the gas side (Montoya et al. 1992).
branes are assembled into carefully spaced cross- Another main concern that has troubled the
laid double layer woven mats which provide uni- membrane industry for the past 30 years and to
form extraluminal ow channels for the blood date remains a serious issue is the blood compat-
(Hanft 2002). The main polymers used in the ibility or hemocompatibility of the oxygenation
production of the blood oxygenation membranes membranes. The denition of biocompatibility
are polyolens such as polypropylene (PP), poly- has been updated for over 20 years and can be
ethylene (PE), and poly-4-methylpentene (PMP) considered as being the bodys acceptance of the
and, at a smaller scale, polyvinylidene uoride material, i.e., the ability of an implant surface to
(PVDF) and silicone rubber (Wiese 2009). interact with the cells and uids of the biological
system and to cause exactly the reactions which
the analogous body tissue would bring about
Current Problems (Ratner 1993; Williams 2008). Presently, the bio-
materials community has been unable to more
Extracorporeal membrane blood oxygenators accurately assign the term blood compatible to
(MBOs) used in the medical eld to substitute a given biomaterial in spite of 50 years of inten-
the respiratory function of the lung have at their sive research on the subject. There is no clear
very core membranes that act as a bloodgas consensus as to which materials are blood com-
barrier and ensure oxygen delivery and carbon patible as blood material interactions are com-
dioxide removal from the patients bloodstream. plex and involve multiple factors including
surface chemical composition, charge, exibility, heparin-based coatings are inherently expensive
wettability, and conditions of blood ow and complex due to the use of sodium heparin, a
(Cazenave 1987). Furthermore, unlike biocom- complex, costly, and fragile biologically derived
patibility assays, there are no standardized substance. The Carmeda Bioactive Surface
methods to assess blood compatibility (Ratner (Medtronic, Inc.) uses a polymer primer coat
2007). Labarres (2010) tentative denition of a followed by covalent attachment of heparin to
blood compatible surface is a surface able to primary amino groups present in the primer coat
keep under control coagulation and inammation (Stenach et al. 1992; Spiess et al. 1998). The
processes at its interface with normal blood, in Durao Bonded Heparin Surface (Baxter, Inc.)
given hemodynamic conditions. According to uses a heparin-polymer blend-based coating that
E
Ratner, we still do not have truly blood compatible is intended to achieve similar results (Toomasian
surfaces after 60 years of intense research of blood et al. 1988; Mottaghy et al. 1989). Other heparin-
compatibility and after numerous approaches based coatings intended to improve blood com-
have been tried (Ratner 2007). The hemo- patibility are available from other medical device
compatibility of biomaterials is known to be manufacturers such as Terumo (Fukasawa 1983),
dependent on various surface properties such as 3M Sarns (Zacour 1988; Fried and Bell-Thomson
surface morphology, surface chemical composi- 1992), etc.
tion, surface charge, surface wettability, and sur- Despite these developments, in order to pre-
face topography. The overall blood-contacting vent in-circuit thrombosis, or clotting, when
surfaces must avoid damage to the blood cells subjected to extracorporeal blood oxygenation
and components but also to avoid stagnation and using any of the MBOs available on the market,
localized clotting. patients still require the administration of antico-
To improve blood compatibility of transport agulant drugs that often lead to major hemorrhag-
surfaces such as the ones in MBOs, researchers ing and blood transfusions.
have focused mainly on surface modication Another approach to improve the hemo-
techniques and producing hemocompatible sur- compatibility of membranes destined for MBOs
face coatings. Minimization of blood/material focuses on the development of novel blood oxy-
interactions can be achieved through chemical genation membranes made from materials
modication; however, these methods are still different to the polyolens mentioned before.
far from perfect. Surface chemical modication, Polyurethanes (PUs) rst found a niche in bio-
besides requiring rather complex experimental medical applications mainly because of their
procedures and involving high costs, still does unique mechanical properties, particularly fatigue
not offer surfaces with long-term stability and resistance, tensile strength, and abrasion (Szycher
leads to potential complications downstream and Poirier 1983; Szycher et al. 1983; Graig 1983;
(Klee and Hcker 2000). Most manufacturers Gogolewski 1989; Hepburn 1992; Furukawa
have focused on the development of 1997), but also because of their good bio- and
anticoagulant-based coatings for the entire bypass hemocompatibility properties (Boretos 1980;
circuit including the oxygenation membranes. Wang and Cooper 1983; Yoon and Ratner 1988;
This approach led to the commercialization of Gogolewski 1989). PU membranes are a good
microporous-coated membranes and bypass- candidate for blood oxygenation devices as the
coated circuits. Various manufacturers claim type, length, ratio, and crystallinity of the different
that the added layer of material provides a monomers as well as the method of membrane
blood-surface interaction with improved preparation determine the bulk properties respon-
hemocompatibility compared to that seen for sible for the nal mass transfer properties (Yilgr
the bulk ber materials. For example, heparin is and Yilgr 1999; Hoshi et al. 2000; Jonquires
known to have good blood compatibility et al. 2002) and the surface properties that affect
properties and has been used as a coating on bio- and hemocompatibility (Yoon and Ratner
several commercialized products. Unfortunately, 1988; Gogolewski 1989).
Blood Compatibility and Membrane/Blood chemical composition, etc. This protein layer con-
Interfaces trols further reactions of the other blood compo-
In 2007, an estimated 250 million polymeric nents and cell system. Additionally, the adsorbed
devices contacting the bloodstream were used in proteins are subject to conformational changes as
humans. All of these blood-contacting devices well as exchange processes with other proteins
exhibited thrombosis problems, and all cardiovas- (Ziats et al. 1988). The competitive adsorption
cular device procedures, both short term and long behavior of proteins at the material surface deter-
term, required signicant anticoagulation, at high mines the pathway and the extent of intrinsic
cost and with considerable risk to the patient coagulation and adhesion of platelets. Platelets
(Geckeler et al. 1997). The number of blood- play a fundamental role in hemostasis. Following
contacting medical devices used worldwide is the protein adhesion to the surface of the bioma-
increasing every day, but their hemocompatibility terial, coagulation is initiated by platelets adher-
properties still remain far from ideal. ing to specic binding sites exposed by the plasma
To this date, there is no consensus denition of proteins. The activated platelets swell, grow spiky
blood compatibility. In 2007, Buddy Ratner extensions known as pseudopodia, and release the
labeled as a catastrophe the fact that the scien- contents of their secretory granules, which contain
tic community has still not been able to accu- a variety of substances which include ADP, sero-
rately assign the term blood compatible to a tonin, and thromboxane A2, which stimulate fur-
biomaterial in spite of 50 years of intensive ther platelet adhesion and activation and enhance
research on the subject (Geckeler et al. 1997; the coagulation process. A large variety of clotting
Ratner 2007). While the denition of biocompat- factors take part in a series of chemical reactions
ibility is generally accepted as the ability of a that eventually create a mesh of brin bers. Each
material to perform with an appropriate host of the clotting proteins has a very specic func-
response in a specic application (Williams tion, and the three main ones are prothrombin,
1987), blood compatibility or hemocompatibility thrombin, and brinogen. Thrombin promotes
has received different denitions by various the conversion of brinogen into long insoluble
authors. From a clinical point of view, a biomate- bers or threads of another protein brin. Fibrin
rial can be considered as blood compatible when threads wind around the platelets forming an
its interaction with blood does not provoke either interlocking network of bers and a framework
any damage of blood cells or any change in the for the clot. This net of bers traps and helps hold
structure of plasma proteins. Only in this case can platelets, blood cells, and other molecules tight to
it be concluded that this material fullls the main the material surface, forming a blood clot also
requests of biocompatibility (Gurland 1987). referred to as a thrombus.
A tentative denition, by Labarre (2010), of a It has long been established, in the eld of
blood compatible surface is a surface able to biomaterials, that bio- and hemocompatibility is
keep under control coagulation and inammation mainly governed by the surface properties of the
processes at its interface with normal blood, in material that contacts the blood (Cooper 1982;
given hemodynamic conditions. Lelah and Cooper 1986; Schamberger and
The contact between the blood and the surface Gardella 1994). Predictions about the interactions
of a biomaterial often leads to different degrees of between the biomaterial surface and the adsorbed
clot formation, as a consequence of the proteins can only be formulated by having an
nonspecic protein adsorption and adhesion and exact knowledge of the structure of the biomate-
activation of blood cells (Horbett 1986; Courtney rials surface and the conformation of the
et al. 1994). Within the rst few seconds after adsorbed proteins. These interactions are deter-
blood contacts, material surface plasma proteins mined both by the hydrophobic/hydrophilic,
are deposited. The extent of protein adsorption charged/uncharged, and polar/nonpolar parts of
depends on the properties of the foreign surface, the proteins and the properties of the polymer
namely, the wettability, charge, roughness, surface such as chemical composition,
morphology, charge, wettability, and topography Hb released Hb negative control

(Dawids and Bantjes 1986; Williams and Wil- HI 100
Hb positive control Hb negative control
liams 1989; Stamm 2008).
(1)
Hemocompatibility properties of surface
materials can be evaluated following the
where [Hb]released is the concentration of Hb in the
ISO10993-4 (Biological evaluation of medical
supernatant solution and [Hb]negative control and
devices, Part 4: Selection of tests for interactions
[Hb]positive control are the concentrations of Hb in
with blood. Geneva, Switzerland: International
the negative and positive controls, respectively.
Organization for Standardization; 2002) guide-
According to ASTM F-756, materials can be
lines in terms of protein adsorption, hemolysis,
labeled nonhemolytic when the HI is between E
platelet adhesion and platelet activation, whole
0 and 2, slightly hemolytic when HI is between
blood clotting time, and clinical coagulation
2 and 5, and hemolytic when HI >5.
times.
Thrombosis can be evaluated through an assay
Hemocompatibility properties of surface mate-
based on the determination of the thrombus mass
rials can be evaluated in vitro according to the ISO
formed on the surface of the membranes after
10993-4:2002 standard (ISO standard) guidelines
contact with static, recalcied blood, based on a
in terms of hemolysis, thrombosis, platelet adhe-
modication of the gravimetric assay proposed by
sion and activation, protein adsorption, and clini-
Imai and Nos (1972) and Allmr et al. (1990).
cal coagulation times. The three categories of
Samples of each membrane are cut into pieces
effects on blood most commonly evaluated are
with a surface area of 4 cm2, and 250 mL of
hemolysis, thrombosis, platelet adhesion, and
blood are carefully placed on top. The clotting
activation (Imai and Nose 1972; Goodman
process is then started by addition of 25 mL of
et al. 1989; Allmr et al. 1990; Haycox and Ratner
0.10 M CaCl2 to the blood drop on the sample.
1993; Tze-Man et al. 1993; Bahulekar et al. 1999;
The clotting process is stopped at different time
ASTM International 2000).
periods by addition of distilled water. The throm-
Damage to the membrane of red cells, as a
bus formed at the surface of the membrane is xed
result of blood exposure to foreign materials, can
with formaldehyde and then washed with water.
be evaluated by the hemolysis test. Hemolysis can
The thrombus mass is determined after drying the
be assessed by quantication of released hemo-
thrombus at 37 C until a constant weight was
globin (Hb) after contact between blood and sam-
reached. The results are expressed as a percentage
ple, according to the ASTM F 756-00 standard.
of the thrombus mass formed on glass (positive
The Hb concentration can be determined by col-
control). A pre-weighted lter paper disk that
orimetry at 540 nm, with a UV/VIS spectropho-
followed the above procedure in parallel, but in
tometer. In the cyanmethemoglobin method
the absence of a membrane sample and of blood,
(Moore et al. 1981; Van Assendelft et al. 1996;
was used as a blank. From the thrombus mass
Zwart et al. 1996), blood samples with a plasma
formed on each sample, a thrombosis degree is
Hb concentration below 2 mg/mL are diluted in
calculated as a percentage of the thrombus mass
PBS in order to obtain blood with a total Hb
formed on the positive control after subtracting
concentration of 10
1 mg/mL. The contact
the blank from each thrombus mass. A thrombosis
between diluted blood and membranes with an
degree of 100 % is assigned to the positive
area of at least 7 cm2 is maintained for at least
control;
4 h of static incubation at 37 C, and after which
the blood is centrifuged, the supernatants are ana- mmembrane mnegative control
Thrombosis % 100
lyzed for their Hb content. Two controls are com- mpositive control mnegative control
monly used: distilled water (positive control) and
PBS solution (negative control). The hemolysis The interaction between platelets and the surface
degree is expressed through the hemolytic index of the membranes can be evaluated by a technique
(HI), calculated as follows: often used by researchers based on the analysis of
Extracorporeal Blood Oxygenation Devices, Mem- II dendritic or early pseudopodial: one or more pseudo-
branes for, Fig. 3 Diagram of the platelet spreading podia with no evident attening; stage III spread den-
process divided into ve shape categories for analysis dritic or intermediate pseudopodial: one or more
and correlation between platelet categories and planimet- pseudopodia attened, hyaloplasm not spread between
rically determined platelet spread area. From left to right, pseudopodia; stage IV spreading: hyaloplasm spread
the stages of spreading are dened as follows: stage I between pseudopodia; and stage V fully spread: hyalo-
round or discoid: no pseudopodia present; stage plasm extensively spread, no distinct pseudopodia
scanning electron microscopy (SEM) images percentage of the membrane surface area covered
(Goodman et al. 1989; Haycox and Ratner 1993; by platelets, respectively. The platelets are either
Tze-Man et al. 1993; Bahulekar et al. 1999). passively attached and remain discoid and singu-
Platelet-rich plasma (PRP) is prepared by centri- lar or may be activated subsequently. In the latter
fugation of the pooled blood and is incubated with case, they progressively release their granule con-
the membrane samples for 30 min at 37 C. After tents, thereby attracting further platelets, showing
incubation, the samples are xed with glutaralde- a typical shape change with attening, formation
hyde post xed with osmium tetroxide (OsO4). of pseudopodia, and increasing area of contact.
After xation the samples are dehydrated through Scanning electron microscopic images of adhered
a graded ethanol series to 100 % and dried by platelets with magnications of 1,000 or
critical point drying. The dry samples are then 2,000 magnication have been used by several
observed by SEM. Quantitative and morphology authors for the quantitative analysis of platelet
analysis of platelets adhered to the surface of the material interaction. The extent of platelet spread-
membranes is carried out by imaging software. ing is examined by categorizing platelet shapes
In order to ensure the presence and activity of the into ve morphological forms describing increas-
platelets, glass is generally used as positive control ing activation stages (IV). These are discoid or
for in vitro platelet response to foreign materials as round (stage I), dendritic or early pseudopodial
it promotes enhanced adhesion and activation of (stage II), spread dendritic or intermediate pseu-
platelets (Park et al. 1991; Amiji 1998). dopodial (stage III), spreading or late pseudopo-
The indicators of platelet adhesion used are dial (stage IV), and fully spread (stage V). The
platelet deposition (PD) and platelet coverage spreading stages and respective platelet areas are
(PC), which are the number of adhered platelets/ shown in Fig. 3 and dened in the legend (Allen
10,000 mm2 of membrane surface area and the et al. 1979; Goodman et al. 1984, 1989).
References (Capiox II) and a bubble oxygenator (Harvey 1500).

J Cardiovasc Surg (Torino) 28:633637
Allen RD, Zacharski LR, Widirstky ST, Rosenstein R, Catapano G, Papenfuss HD, Wodetzki A, Baurmeister U
Zaitlin LM, Burgess DR (1979) Transformation and (2001) Mass and momentum transport in extra-luminal
motility of human platelets: details of the shape change ow (ELF) membrane devices for blood oxygenation.
and release reaction observed by optical and electron J Membr Sci 184:123135
microscopy. J Cell Biol 83:126142 Cazenave JP (1987) Blood surface interactions: biological
Allmr K, Hilborn J, Larsson PH, Hult A, Rnby B (1990) principles underlying haemocompatibility with articial
Surface modication of polymers. V. Biomaterial appli- materials. Elsevier Science. http://www.bn.com/w/blood-
cations. J Polym Sci A Polym Chem 28:173183 surface-interactions-j-p-cazenave/1013930926
Alpha D, King E, Bicknell DA (1986) Clinical evaluation Chalwin RP, Tiruvoipati R, Peek GJ (2008) Fatal throm-
of the extracorporeal Maxima hollow ber membrane. bosis with activated factor VII in a paediatric patient on
Proc Am Acad Clin Perfusion 7:3234 extracorporeal membrane oxygenation. Eur E
American Society for Testing of Materials (ASTM) Prac- J Cardiothorac Surg 34:685686
tice for assessment of hemolytic properties of materials Clowes GH, Hopkins AL, Neville WE (1956) An articial
(2000) ASTM International. http://www.astm.org/ lung dependent upon diffusion of oxygen and carbon
DATABASE.CART/HISTORICAL/F756-00.htm dioxide through plastic membranes. J Thorac Surg
Amiji MM (1998) Platelet adhesion and activation on an 32:630637
amphoteric chitosan derivative bearing sulfonate Cooper SL (1982) Biomaterials: interfacial phenomena
groups. Colloids Surf B: Biointerfaces 10:263271 and applications. American Chemical Society, Wash-
Bahulekar R, Tamura N, Ito S, Kodama M (1999) Platelet ington, DC, USA
adhesion and complement activation studies on Courtney JM, Lamba NM, Sundaram S, Forbes CD
poly(N-alkyl mono and disubstituted) acrylamide (1994) Biomaterials for blood-contacting applications.
derivatives. Biomaterials 20:357362 Biomaterials 15:737744
Baker R (2004) Membrane technology and applications. Dawids S, Bantjes A (1986) Martinus Nijhoff Publishers a
West Sussex, England member of the Kluwer Academic Publishers Group,
Bartlett RH, Gazzaniga AB, Jefferies MR, Huxtable RF, Dordrecht, The Netherlands
Haiduc NJ, Fong SW (1976) Extracorporeal De Somer F, Foubert L, Vanackere M, Dujardin D,
membrane oxygenation (ECMO) cardiopulmonary Delanghe J, Van Nooten G (1996) Impact of oxygena-
support in infancy. Trans Am Soc Artif Intern Organs tor design on hemolysis, shear stress, and white blood
22:8093 cell and platelet counts. J Cardiothorac Vasc Anesth
Bartlett RH, Andrews AF, Toomasian JM, Haiduc NJ, 10:884889
Gazzaniga AB (1982) Extracorporeal membrane oxy- DeWall RA, Naja H, Roden TA (1966) Hard-shell
genation for newborn respiratory failure: forty-ve temperature-controlling disposable blood oxygenator
cases. Surgery 92:425433 use in clinical open-heart surgery. JAMA
Bartlett RH, Roloff DW, Custer JR, Younger JG, Hirschl 197:10651068
RB (2000) Extracorporeal life support: the University Dorson W, Baker E, Hull H (1968) A shell and tube
of Michigan experience. JAMA 283:904908 oxygenator. Trans Am Soc Artif Intern Organs
Bergdahl ME, Bergdahl LA (1989) A comparison of at- 14:242249
sheet and hollow-ber membrane oxygenators: the Dutton RC, Mather FW, Walker SN, Lipps BJ, Rudy LW,
Shiley M-2000 vs. the Bentley BOS-CM 40. Tex Severinghaus JW, Edmunds LH (1971) Development
Heart Inst J 16:2731 and evaluation of a new hollow-ber membrane oxy-
Boretos JW (1980) Past, present and future role of poly- genator. Trans Am Soc Artif Intern Organs 17:331336
urethanes for surgical implants. Pure Appl Chem Edmunds LH (2004) Cardiopulmonary bypass after
52:18511855 50 years. N Engl J Med 351:16031606
Bramson ML, Osborn JJ, Main FB, OBrien MF, Wright Edmunds LH, Colman RW (2006) Thrombin during car-
JS, Gerbode F (1965) A new disposable membrane diopulmonary bypass. Ann Thorac Surg 82:23152322
oxygenator with integral heat exchange. J Thorac Elgas RJ, Gordon TM (1984) Blood oxygenator. http://
Cardiovasc Surg 50:391400 www.google.com/patents/US4451562
Bungay PM, Lonsdale HK, de Pinho MN (1986) Synthetic Fox S (2007) Human physiology. McGraw-Hill Science/
membranes science, engineering, and applications. Engineering/Math, New York, USA
Published in cooperation with NATO Scientic Affairs Fried DW, Bell-Thomson J (1992) Oxygen transfer perfor-
Division by D. Reidel mance of a new Sarns membrane oxygenator. Perfusion
Cahn RP, Li NN (1974) Separation of phenol from waste 7:263271
water by the liquid membrane technique. Sep Sci Fukasawa H (1983) http://www.google.com/patents/
9:505519 US4374802
Calaore AM, Glieca F, Marchesani F, Troise G, Furukawa M (1997) Hydrolytic and thermal stability of
Santarelli G, Possati F (1987) A comparative clinical novel polyurethane elastomers. Die Angewandte
assessment of a hollow-bre membrane oxygenator Makromolekulare Chemie 252:3343
Gabelman A, Hwang S-T (1999) Hollow ber membrane Haycox CL, Ratner BD (1993) In vitro platelet interactions
contactors. J Membr Sci 159:61106 in whole human blood exposed to biomaterial surfaces:
Galletti PM (1993) Cardiopulmonary bypass: a historical insights on blood compatibility. J Biomed Mater Res
perspective. Artif Organs 17:675686 27:11811193
Gassman CJ, Galbraith GD, Smith RG (1987) Evaluation Hepburn C (1992) Polyurethane elastomers. Dordrecht,
of three types of membrane oxygenators and their suit- Netherlands
ability for use with pulsatile ow. J Extracorporeal Hill JD (1977) Acute pulmonary failure: treatment with
Technol 19:297304 extracorporeal oxygenation. Med Instrum 11:198201
Geckeler KE, Rupp F, Geis-Gerstorfer J (1997) Interfaces Hill JD, Aguilar MJ, Baranco A, de Lanerolle P, Gerbode
and interphases of (bio)materials: denitions, struc- F (1969) Neuropathological manifestations of cardiac
tures, and dynamics. Adv Mater 9:513518 surgery. Ann Thorac Surg 7:409
Ghosh S, Falter F, Cook DJ (2009) Extracorporeal mem- Ho WS, Sirkar KK (1992) Membrane handbook. Van
brane oxygenation. Cambridge University Press. Nostrand Reinhold, New York
10.1017/CBO9780511635564 Horbett TA (1986) Interaction of the blood with natural and
Gibbon JH (1937) Articial maintenance of circulation articial surfaces. J Biomed Mater Res 20:411412
during experimental occlusion of pulmonary artery. Hoshi M, Ieshige M, Saitoh T, Nakagawa T (2000) Sepa-
Arch Surg 34:11051131 ration of aqueous phenol through polyurethane mem-
Gibbon JH Jr (1954) Application of a mechanical heart and branes by pervaporation. III. Effect of the methylene
lung apparatus to cardiac surgery. Minn Med group length in poly(alkylene glycols). J Appl Polym
37:171185 Sci 76:654664
Gogolewski S (1989) Selected topics in biomedical poly- Hsu LC (1997) Biocompatibility in cardiopulmonary
urethanes. A review. Colloid Polym Sci 267:757785 bypass. J Cardiothorac Vasc Anesth 11:376382
Goodman SL, Lelah MD, Lambrecht LK, Cooper SL, Iatridis A, Chan T, Thompson R (1985) Experimental and
Albrecht RM (1984) In vitro vs. ex vivo platelet depo- clinical trials with the extracorporeal hollow bre lung.
sition on polymer surfaces. Scan Electron Microsc Proc Am Acad Clin Perform 6:4750
Pt. 1:279290 Imai Y, Nose Y (1972) A new method for evalution of
Goodman SL, Grasel TG, Cooper SL, Albrecht RM antithrombogenicity of materials. J Biomed Mater Res
(1989) Platelet shape change and cytoskeletal reorga- 6:165172
nization on polyurethaneureas. J Biomed Mater Res ISO International Organization for Standardization.
23:105123 ISO10993-4 Biological evaluation of medical
Graig RG (1983) Transactions of the ninth annual meeting devices, Part 4: selection of tests for interactions
of the society for biomaterials. http://www.docstoc. with blood. http://www.iso.org/iso/iso_catalogue/cata
com/docs/88308470/Annual-Meeting-of-the-Society- logue_tc/catalogue_detail.htm?csnumber=32086
for-Biomaterials-Birmingham Iwahashi H, Yuri K, Nos Y (2004) Development of the oxy-
Grenvik A, Ayres SM, Holbrook P, Shoemaker W (1999) genator: past, present, and future. J Artif Organs 7:111120
Textbook of critical care. Elsevier Science Health Sci- Janvier G, Baquey C, Roth C, Benillan N, Blisle S, Hardy JF
ence. http://www.ecampus.com/textbook-critical-care- (1996) Extracorporeal circulation, hemocompatibility,
4th-grenvik-ake/bk/9780721672465 and biomaterials. Ann Thorac Surg 62:19261934
Gu YJ, Boonstra PW, Graaff R, Rijnsburger AA, Jonquires A, Clment R, Lochon P (2002) Permeability of
Mungroop H, van Oeveren W (2000) Pressure drop, block copolymers to vapors and liquids. Prog Polym
shear stress, and activation of leukocytes during cardio- Sci 27:18031877
pulmonary bypass: a comparison between hollow ber Kawahito S, Maeda T, Motomura T, Takano T, Nonaka K,
and at sheet membrane oxygenators. Artif Organs Linneweber J, Ichikawa S, Ishitoya H, Hanazaki K,
24:4348 Glueck J, Sato K, Nos Y (2002a) Feasibility of a new
Guidoin RG, Awad JA, Domurado D, Haggis GH hollow ber silicone membrane oxygenator for long-
(1978) Thrombus formation in a staged capillary mem- term ECMO application. J Med Invest 49:156162
brane oxygenator associated with a microaggregate Kawahito S, Motomura T, Glueck J, Nos Y (2002b)
blood lter: a scanning electron microscope study. Development of a new hollow ber silicone membrane
J Biomed Mater Res 12:255269 oxygenator for ECMO: the recent progress. Ann
Gurland H (1987) Uremia therapy. Springer, New Thorac Cardiovasc Surg 8:268274
York, http://www.bn.com/w/uremia-therapy-h-j-gurland/ Kay PH, Munsch CM (1992) Techniques in extracorporeal
1000920138 circulation. 4th Ed, Arnold, London
Hammon JW (2008) Extracorporeal circulation: the Kirklin JK, Westaby S, Blackstone EH, Kirklin JW,
response of humoral and cellular elements of blood to Chenoweth DE, Pacico AD (1983) Complement and
extracorporeal circulation. Card Surg Adult 3:370 the damaging effects of cardiopulmonary bypass.
Hanft SC (2002) Key medical membrane devices for the J Thorac Cardiovasc Surg 86:845857
new millennium. BCC Research, Wellesley, MA, USA Klee D, Hcker H (2000) Polymers for biomedical appli-
Haworth WS (2003) The development of the modern oxy- cations: improvement of the interface compatibility.
genator. Ann Thorac Surg 76:22162219 Biomed Appl Polym Blends 149:157
Kolff WJ (1954) Dialysis in treatment of uremia: articial artery bypass grafting surgery. Ann Thorac Surg
kidney and peritoneal lavage. AMA Arch Intern Med 80:21262131
94:142160 Marx TI, Baldwin BR, Miller DR (1962) Factors inuenc-
Kolff WJ, Balzer R (1955) The articial coil lung. Trans ing oxygen uptake by blood in membrane oxygenators:
Am Soc Artif Intern Organs 1:3942 report of a study. Ann Surg 156:204213
Kolff WJ, Watschinger B (1956) Further development of a McArdle WD, Katch FI, Katch VL (2005) Essentials of
coil kidney; disposable articial kidney. J Lab Clin exercise physiology. Lippincott Williams & Wilkins,
Med 6:969977 Baltimore
Kolff WJ, Berk HTJ, Welle NM, Ley AJ, Dijk EC, Mockros LF, Leonard R (1985) Compact cross-ow tubu-
Noordwijk J (1944) The articial kidney: a dialyser lar oxygenators. Trans Am Soc Artif Intern Organs
with a great area. Acta Medica Scand 117:121134 31:628633
Kolff WJ, Efer DB, Groves JK, Peereboom G, Moraca PP Montoya JP, Shanley CJ, Merz SI (1992) Plasma leakage
(1956a) Disposable membrane oxygenator (heart-lung through microporous membranes. Role of phospho- E
machine) and its use in experimental surgery. Cleve lipids. ASAIO 38:399
Clin Q 23:6997 Moore GL, Ledford ME, Merydith A (1981)
Kolff WJ, Watschinger B, Vertes V (1956b) Results in A micromodication of the Drabkin hemoglobin
patients treated with the coil kidney (disposable dialyz- assay for measuring plasma hemoglobin in the range
ing unit). J Am Med Assoc 161:14331437 of 5 to 2000 mg/dl. Biochem Med 26:167173
Kolobow T, Bowman RL (1963) Construction and evalu- Motomura T, Maeda T, Kawahito S, Matsui T, Ichikawa S,
ation of an alveolar membrane articial heart-lung. Ishitoya H, Kawamura M, Shinohara T, Sato K,
Trans Am Soc Artif Intern Organs 9:238243 Kawaguchi Y, Taylor D, Oestmann D, Glueck J, Nos
Kolobow T, Spragg RG, Pierce JE, Zapol WM Y (2003) Development of silicone rubber hollow ber
(1971) Extended term (to 16 days) partial extracorpo- membrane oxygenator for ECMO. Artif Organs
real blood gas exchange with the spiral membrane lung 27:10501053
in unanesthetized lambs. Trans Am Soc Artif Intern Mottaghy K, Oedekoven B, Pppel K, Bruchm uller K,
Organs 17:350354 Kovacs B, Spahn A, Geisen C (1989) Heparin free
Kramme R (2011) Medizintechnik: Verfahren- Systeme- long-term extracorporeal circulation using bioactive
Informationsverarbeitung. Gabler Wissenschaftsverlage, surfaces. ASAIO Trans 35:635637
Heidelberg, Germany Mulder J (1996) Basic principles of membrane
Kremesec VJ (1981) Modeling of dispersed-emulsion sep- technology. Kluwer Academic Publishers, Dordrecht,
aration systems. Sep Purif Rev 10:117157 Netherlands
Labarre D (2010) Improving blood-compatibility of poly- Nagase K, Kohori F, Sakai K (2005a) Oxygen transfer perfor-
meric surfaces. Trends Biomater Artif Organs 15(1):13 mance of a membrane oxygenator composed of crossed
Le Gallois JJC (1812) Experiments on the principle of life, and parallel hollow bers. Biochem Eng J 24:105113
and particularly on the principle of the motions of the Nagase K, Kohori F, Sakai K, Nishide H (2005b)
heart, and on the seat of this principle: including the Rearrangement of hollow bers for enhancing oxygen
report made to the rst class of the Institute, upon the transfer in an articial gill using oxygen carrier solu-
experiments relative to the motions of the heart. http:// tion. J Membr Sci 254:207217
www.archive.org/details/2561017R.nlm.nih.gov Niiya KY, Noble RD (1985) Competitive facilitated transport
Lee WH, Krumhaar D, Fonkalsrud EW, Schjeide OA, through liquid membranes. J Membr Sci 23:183198
Maloney JV (1961) Denaturation of plasma proteins Nos Y (2001) The ICMT on articial organs, vol II, The
as a cause of morbidity and death after intracardiac oxygenator. ICMT Press, Houston
operations. Surgery 50:2939 Nos Y, Malchesky PS (2000) Therapeutic membrane
Lee WH, Miller W, Rowe J, Hairston P, Brady MP plasmapheresis. Ther Apher 4:39
(1969) Effects of extracorporeal circulation on person- Onorati F, Rubino AS, Nucera S, Foti D, Sica V, Santini F,
ality and cerebration. Ann Thorac Surg 7:562 Gulletta E, Renzulli A (2010) Off-pump coronary artery
Lelah MD, Cooper SL (1986) Polyurethanes in medicine. bypass surgery versus standard linear or pulsatile cardio-
CRC Press, Boca Raton, Florida, USA pulmonary bypass: endothelial activation and inamma-
Lillehei CW, Cohen M, Warden HE, Read RC, Aust JB, tory response. Eur J Cardiothorac Surg 37:897904
DeWall RA, Varco RL (1955) Direct vision intracardiac Park K, Mao FW, Park H (1991) The minimum surface
surgical correction of the tetralogy of fallot, pentalogy brinogen concentration necessary for platelet activa-
of fallot, and pulmonary atresia defects. Ann Surg tion on dimethyldichlorosilane-coated glass. J Biomed
142:418442 Mater Res 25:407420
Loeb S (1981) The Loeb-Sourirajan membrane: how it Peek GJ, Clemens F, Elbourne D, Firmin R, Hardy P,
came about. Synthetic membranes, vol I, ACS Sympo- Hibbert C, Killer H, Mugford M, Thalanany M,
sium Series 153. ACS, Washington, DC Tiruvoipati R, Truesdale A, Wilson A (2006) CESAR:
Lund C, Sundet K, Tennoe B, Hol PK, Rein KA, Fosse E, conventional ventilatory support vs extracorporeal
Russell D (2005) Cerebral ischemic injury and cogni- membrane oxygenation for severe adult respiratory
tive impairment after off-pump and on-pump coronary failure. BMC Health Serv Res 6:163
PlunkettSD(2002)Alkoxysilane/alkysilanecopolymercoating Szycher M, Poirier VL (1983) Synthetic polymers in arti-

of medical devices. http://www.docstoc.com/docs/ cial hearts: a progress report. Ind Eng Chem Prod Res
51355531/Alkoxysilanealkysilane-Copolymer-Coating- Dev 22:588593
Of-Medical-DevicesPatent-6395226 Szycher M, Poirier VL, Dempsey DJ (1983) Development
Rajasubramanian S, Nelson KD, Shastri P, of an aliphatic biomedical-grade polyurethane elasto-
Constantinescu A, Kulkarni P, Jessen ME, Eberhart mer. J Elastomers Plast 15:8195
RC (1997) Design of an oxygenator with enhanced Tanishita K, Panol G, Richardson PD, Galletti PM
gas transfer efciency. ASAIO J 43:710714 (1994) Gas transport in the intracorporeal oxygenator
Ratner BD (1993) New ideas in biomaterials science a with woven tubes. Artif Organs 18:797800
path to engineered biomaterials. J Biomed Mater Res Taskin ME, Fraser KH, Zhang T, Grifth BP, Wu ZJ
27:837850 (2010) Micro-scale modeling of ow and oxygen trans-
Ratner BD (2007) The catastrophe revisited: blood compat- fer in hollow ber membrane bundle. J Membr Sci
ibility in the 21st century. Biomaterials 28:51445147 362:172183
Roy BJ, Rycus P, Conrad SA, Clark RH (2000) The chang- Toomasian JM, Hsu LC, Hirschl RB, Heiss KF, Hultquist
ing demographics of neonatal extracorporeal mem- KA, Bartlett RH (1988) Evaluation of Durao II hepa-
brane oxygenation patients reported to the rin coating in prolonged extracorporeal membrane oxy-
Extracorporeal Life Support Organization (ELSO) reg- genation. ASAIO Trans 34:410414
istry. Pediatrics 106:13341338 Tze-Man K, Jui-Che L, Cooper SL (1993) Surface charac-
Schamberger PC, Gardella JA (1994) Surface chemical terization and platelet adhesion studies of plasma-
modications of materials which inuence animal cell sulphonated polyethylene. Biomaterials 14:657664
adhesion a review. Colloids Surf B: Biointerfaces Ulrich B (1990) Multilayer hollow ber wound body.
2:209223 http://www.docstoc.com/docs/41159322/Multilayer-
Scott W, Rode R, Castlemain B, Kessler R, Follis F, Hollow-Fiber-Wound-BodyPatent-4940617
Pett S, Wernly J (1992) Efcacy, complications, and cost Vallely MP, Bannon PG, Bayeld MS, Hughes CF,
of a comprehensive blood conservation program for car- Kritharides L (2010) Endothelial activation after coronary
diac operations. J Thorac Cardiovasc Surg 103:10011006 artery bypass surgery: comparison between on-pump and
Segal H, Sheikh S, Kallis P, Cottam S, Beard C, Potter D, off-pump techniques. Heart Lung Circ 19:445452
Townsend E, Bidstrup BP, Yacoub M, Hunt BJ Van Assendelft OW, Buursma A, Zijlstra WG (1996) Sta-
(1998) Complement activation during major surgery: bility of haemiglobincyanide standards. J Clin Pathol
the effect of extracorporeal circuits and high-dose 49:275277
aprotinin. J Cardiothorac Vasc Anesth 12:542547 Vaslef SN, Mockros LF, Anderson RW, Leonard RJ
Skeggs LT (2000) Persistence. . . prayer: from the articial (1994) Use of a mathematical model to predict oxygen
kidney to the AutoAnalyzer. Clin Chem 46:14251436 transfer rates in hollow ber membrane oxygenators.
Skeggs LT, Leonards JR, Heisler CR (1949) Articial kid- ASAIO J 40:990996
ney. II. Construction and operation of an improved Verrier ED, Boyle EM (1996) Endothelial cell injury
continuous dialyzer. Proc Soc Exp Biol Med in cardiovascular surgery. Ann Thorac Surg
72:539543 62:915922
Spiess BD, Vocelka C, Cochran RP, Soltow L, Chandler Von Segesser L (1987) Determination of signicant differ-
WL (1998) Heparin-coated bypass circuits (Carmeda) ences in performance of the Bentley BOS-CM 40 hollow
suppress the release of tissue plasminogen activator bre membrane oxygenator and the polystan VT5000
during normothermic coronary artery bypass graft sur- venotherm bubble oxygenator. Perfusion 2:289295
gery. J Cardiothorac Vasc Anesth 12:299304 Voorhees ME, Brian BF (1996) Blood-gas exchange
Stamatialis DF, Papenburg BJ, Girons M, Saiful S, devices. Int Anesthesiol Clin 34:2945
Bettahalli SN, Schmitmeier S, Wessling M (2008) Med- Wang CB, Cooper SL (1983) Morphology and properties
ical applications of membranes: drug delivery, articial of segmented polyether polyurethaneureas. Macromol-
organs and tissue engineering. J Membr Sci 308:134 ecules 16:775786
Stamm M (2008) Polymer surfaces and interfaces. Char- Weibel ER (2009) What makes a good lung? Swiss Med
acterization, modication and application. Springer, Wkly 139:375386
Heidelberg, Germany http://www.springer.com/ Weissman MH, Mockros LF (1968) Gas transfer to blood
chemistry/polymer+science/book/978-3-540-73864-0 owing in coiled circular tubes. J Eng Mech Div
Stenach N, Korn RL, Fisher CA, Jeevanandam V, 94(EM3):857872, paper 6007, American Society of
Addonizio VP (1992) The effects of heparin bound Civil Engineers. American Society of Civil Engineers
surface modication (Carmeda Bioactive Surface) on Weissman MH, Mockros LF (1969) Oxygen and carbon
human platelet alterations during simulated extracor- dioxide transfer in membrane oxygenators. J Med Biol
poreal circulation. J Extra Corp Technol 24:97102 Eng 2:169184
Suma K, Tsuji T, Takeuchi Y, Inoue K, Shiroma K, WHO World Health Organization (2011). The top
Yoshikawa T, Narumi J (1981) Clinical performance 10 causes of death. WHO World Health Organization.
of microporous polypropylene hollow-ber oxygena- http://www.who.int/mediacentre/factsheets/fs310/en/
tor. Ann Thorac Surg 32:558562 index.html
Extraction Index 755
Wickramasinghe SR, Han B (2002) Mass and momentum Zwart A, van Assendelft OW, Bull BS, England JM, Lewis
transfer in commercial blood oxygenators. Desalina- SM, Zijlstra WG (1996) Recommendations for refer-
tion 148:227233 ence method for haemoglobinometry in human blood
Wickramasinghe SR, Han B (2005) Designing micropo- (ICSH standard 1995) and specications for interna-
rous hollow bre blood oxygenators. Chem Eng Res tional haemiglobinocyanide standard (4th edition).
Des 83:256267 J Clin Pathol 49:271274
Wiese F (2009) Membranes for articial lung and gas Zydney AL (1985) Cross-ow membrane plasmapheresis:
separation, membranes in medicine workshop. Lisbon an analysis of ux and hemolysis. Massachusetts Insti-
Wiese F (2010) Membranes for articial lungs. Membr Life tute of Technology, Cambridge, MA
Sci 1:4968. Wiley-VCH. http://onlinelibrary.wiley.
com/. doi:10.1002/9783527631360.ch2/summary
Williams DF (1987) Denitions in biomaterials: proceed-
ings of a consensus conference of the European Society E
for Biomaterials, Chester, England, March 35, 1986. Extraction Index
Elsevier, Amsterdam
Williams DF (2008) On the mechanisms of biocompatibil-
Giuseppe Barbieri
ity. Biomaterials 29:29412953
Williams RL, Williams DF (1989) The spatial resolution of Institute on Membrane Technology, National
protein adsorption on surfaces of heterogeneous metal- Research Council of Italy, ITM-CNR, Rende,
lic biomaterials. J Biomed Mater Res 23:339350 Italy
Yang M, Cussler EL (1986) Designing hollow-ber
contactors. AIChE J 32:19101916
Yilgr I, Yilgr E (1999) Hydrophilic polyurethaneurea
membranes: inuence of soft block composition on the The extraction index is a measure of membrane
water vapor permeation rates. Polymer 40:55755581 operation performance of extracting a desired spe-
Yoon SC, Ratner BD (1988) Surface and bulk structure of
cies from the feed/retentate side in order to have it
segmented poly(ether urethanes) with peruoro chain
extenders. FTIR, DSC, and x-ray photoelectron spec- as permeate. It can be dened/evaluated for any
troscopic studies. Macromolecules 21:23922400 involved species. It gives an indication about the
Zacour RK (1988) Clinical evaluation of the sarns mem- real advantages offered by using a membrane unit
brane oxygenator. Perfusion 3:219224
instead of conventional ones. The higher the var-
Zapol WM, Snider MT, Hill JD, Fallat RJ, Bartlett RH,
Edmunds LH, Morris AH, Peirce EC, Thomas AN, iable value, the higher the performance of the
Proctor HJ, Drinker PA, Pratt PC, Bagniewski A, Miller membrane unit. It is dened as the ratio (Eq. 1)
RG (1979) Extracorporeal membrane oxygenation in of the permeate ow rate of the species permeated
severe acute respiratory failure. A randomized prospec-
through the membrane to that totally fed to the
tive study. JAMA 242:21932196
Ziats NP, Miller KM, Anderson JM (1988) In vitro and membrane unit. Equation 1 shows different ways
in vivo interactions of cells with biomaterials. Bioma- for evaluating the extraction index:
terials 9:513
Permeate flow ratei

Extraction Indexi
Feed flow ratei
Permeate flow rate Permeate molar fractioni

Feed flow rate Feed molar fractioni
Permeate flow rate Permeate Partial Pressurei

(1)
Feed flow rate Feed Partial Pressurei
Permeate flow rate Permeate concentrantioni

Feed flow rate Feed concentrantioni
Permeate flow rate Permeate mass fractioni

Feed flow rate Feed mass fractioni
756 Extremophiles
The extraction index assumes a specic and more be fed in it but also produced by reaction. In this
interesting form when a chemical reaction takes case, Eq. 1 has to take into account also the pro-
place inside the membrane unit such as in the duction term:
membrane reactors since a desired chemical can
Permeate flow ratei

Extraction Indexi (2)
Feed flow ratei Production by Reaction for i th
If the water-gas shift reaction is used as an example, totally available in the feed stream both as H2
the extraction index can be dened as the ratio of molecules and also obtainable from other chemicals
the H2 permeated through the membrane to that (e.g., CO) by reaction, to the membrane unit (Eq. 3):
Flow ratePermeate
H2
H2 Extraction Index
Flow rateAvailable
H2
in the feed
Flow ratePermeate
H2
in the MR (3)
Flow rateFeed
H2 a Flow rateFeed
CO

H2 O=CO feed molar ratio per H2 O=CO feed molar ratio < 1
a
1 per H2 O=CO feed molar ratio 1
The extraction index (Barbieri et al. 2008) takes into

account the hydrogen fed as H2 molecules and that Extremophiles
is contained in the feed stream in other chemicals
(e.g., H2O). The term (a ow rate CO feed) of the Chiara Schiraldi and Mario De Rosa
Eq. 3 considers the maximum H2 extractable from Department of Experimental Medicine, Section of
the chemicals (other than hydrogen) present in the Biotechnology, Medical Histology and Molecular
system. The coefcient (a) takes into account the Biology, Second University of Naples, Napoli,
defecting reactant (CO or H2O) by means of the Italy
feed molar ratio H2O/CO. It (a) is equal to the feed
molar ratio if the latter is lower than 1 (CO in defect
with respect to H2O). It (a) will be equal to 1 when The discovery of life in demanding environments
the CO exceeds the H2O. As dened, the extraction continues to challenge conventional concepts of
index is determined by the membrane properties, the growth-limiting conditions of many cellular
feed molar ratio, and CO conversion achieved in the organisms. Extremophiles may have diverse fea-
membrane reactor, at set operating conditions. tures; they may live at temperatures higher than
60 C (thermophiles) or prefer colder sites, being
able to grow between 5 C and 20 C; they may
References need pH <4 or >9 (acidophiles or alkaliphiles,
respectively); or they can survive to high-salinity
Barbieri G, Brunetti A, Tricoli G, Drioli E (2008) An environments such as salty lakes (halophiles).
innovative conguration of a Pd-based membrane reac-
They have been isolated in challenging biotopes
tor for the production of pure hydrogen. Experimental
analysis of water gas shift. J Power Sources from terrestrial solfataric elds to marine volcanic
182(1):160167. doi:10.1016/j.jpowsour.2008.03.086 areas (Rothschild and Mancinelli 2001). The
Extremophiles 757
specically founded company (bitop).

Alkaliphilic enzymes were found as natural deter-
gents, thus suggesting specic industrial interest.
Psychrophilic enzymes may also be of applicative
interest when biotransformations need to be car-
ried out at low temperature. Other innovative
products from extremophiles are cyclodextrins
and polyunsaturated fatty acids, for which
industrial applications are foreseen. Most of
these kind of microorganisms isolated to date
E
show specic needs to increase cell density.
Recent alternatives to technological applications
Extremophiles, Fig. 1 Electron micrograph of most commonly employed when the production
Sulfolobus solfataricus MT4
of extremocompounds is approached at a bioreac-
tor scale describe batch, fed batch, and continuous
or in situ product removal fermentations in
phylogenetic assessment of the isolated species specically developed bioprocesses. However,
widens between Archaea, Bacteria, and Eukarya. when large-scale production is needed for
There has been a steady increase in the isolation of commercialization of these novel compounds,
these microorganisms documenting the enormous there is the need to bring together genetic and
scientic effort in the last 30 years. Although bioreaction engineering with separation tech-
major advances have been made in the last niques. The former prompted the development
decade, our knowledge of the physiology, metab- of specic plasmids and vectors to obtain a
olism, enzymology, and genetics of this fascinat- sound expression of extremophilic gene sequence
ing group of organisms is still limited. However and thus enzymes into mesophilic hosts; these
there is little doubt that extremophiles will supply recombinant strains are easier to cultivate, and
novel catalysts and will be a source of biomole- the whole strategy is in fact recommended
cules with unique and biotechnologically relevant to solve the typical problems faced in
properties. It has been argued that membranes of extremophile-based bioprocesses and will
extremophiles contain surfactants bearing unique serve to open up new opportunities for the devel-
stability that can be used in pharmaceutical opment of unexplored elds such as renewable
and cosmeceutical formulations. Amylases, energies.
pullulanases, and glycosidases from hyperther-
mophiles were studied and proved efcient in
starch biotransformation obtaining process
throughput enhancement (Burg 2003).
References
Trehalose-forming enzymes were found in
Burg BVD (2003) Extremophiles as a source for novel
Sulfolobus shibatae and S. solfataricus isolated enzymes. Curr Opin Microbiol 6:213218
in diverse solfatara elds worldwide (Fig. 1). Kristjansson JK, Hreggvidsson GO (1995) Ecology and
Also lipases and esterases from thermophilic habitats of extremophiles. World J Microbiol
Biotechnol 11:1725
microorganisms and Archaea proved unique fea-
Rothschild LJ, Mancinelli RL (2001) Life in extreme envi-
tures (Schiraldi et al. 2002). Compatible solutes, ronments. Nature 409:10921101
of key importance for halophile survival, were Schiraldi C, Giuliano M, De Rosa M (2002) Perspectives
isolated from Halomonas and Marinococcus spe- on biotechnological applications of archaea. Archaea
1(2):7586
cies and thoroughfully characterized proving their Sellek GA, Chaudhuri JB (1999) Biocatalysis in organic
outstanding DNA/enzyme stabilization capacity; media using enzymes from extremophiles. Enzyme
their commercialization was pursued by a Microb Technol 25:471482
758 Extremozymes
stability and wide operational range suggest appli-

Extremozymes cation in hydrogen peroxide removal in pulp and
paper bleaching, textile bleaching, food pasteuri-
Chiara Schiraldi and Mario De Rosa zation, and surface decontamination of food pack-
Department of Experimental Medicine, Section of aging (Thompson et al. 2003). Enzymes isolated
Biotechnology, Medical Histology and Molecular from psychrophiles, such as lipases, proteases,
Biology, Second University of Naples, Napoli, and cellulases, have been used as additives for
Italy the preparation of detergents working at low tem-
peratures or in frozen food preparations. Further-
more, thermophilic enzymes have been used for
The biocatalysts produced by extremophilic the construction of optical nanosensors, stable and
microorganisms, so-called extremozymes, are nonconsuming analytes. These innovative devices
proteins with outstanding stability to temperature, are based on the ability of thermophilic enzymes
pH, and organic solvents, thus becoming excellent to bind the substrate at room temperature, without
candidates to improve industrial biotransforma- transforming it (de Champdor et al. 2007). The
tions (Schiraldi and De Rosa 2002). Polymer- binding of substrate to thermophilic enzyme is
degrading enzymes from hyperthermophiles, monitored as uorescence variations of the
psychrophiles, and acidophiles may play an enzyme. Many of these isolated enzymes have
important role in food, detergent, and pulp and been cloned and expressed in mesophilic hosts
paper industry (e.g., amylases, pullulanases, to overcome the issues of extremophilic microor-
xylanases, proteases). Extremozymes also include ganisms cultivations, such as the unconventional
cellulases, proteases, pectinases, keratinases, fermentation parameters, special construction
lipases, esterases, catalases, peroxidases, and materials need, the low growth rates that are typ-
phytases. Owing to the unusual properties of ical of most of these species, and the low biomass
these classes of enzymes, they are expected to yield (despite the good enhancement proved in
ll the gap between biological and chemical membrane bioreactors) (Schiraldi et al. 2001).
industrial processes (Taylor et al. 2011). Despite The recombinant enzymes (Cimini et al. 2008),
this, actually few are the current processes based easily produced at high yield, may be commer-
on these biocatalysts, the most known of which is cialized, and furthermore, modern techniques like
the polymerase chain reaction (PCR) technology mutagenesis and gene shufing will lead to
that in fact is based on DNA-modifying enzymes in vitro tailored enzymes that are highly specic
Taq polymerase isolated from Thermus aquaticus. for countless industrial applications.
The use of this enzyme allowed the automation of
PCR with a great advantage for research labora-
tories and industries. Beside this, few biocatalysts, References
used in diagnostics and starch liquefaction, are
commercially produced by several biotechnology Cimini D, De Rosa M, Panariello A, Morelli V,
companies. Recently enzymes of interest in bio- Schiraldi C (2008) Production of a thermophilic
maltooligosyl-trehalose synthase in lactococcus lactis.
remediation have also been isolated from J Ind Microbiol Biotechnol 35(10):10791083
extremophiles: a thermoalkaliphilic catalase, de Champdor M, Staiano M, Rossi M, DAuria S (2007)
which initiates the breakdown of hydrogen perox- Proteins from extremophiles as stable tools for
ide into oxygen and water, was isolated from advanced biotechnological applications of high social
interest. J R Soc Interface 4(13):183191
Thermus brockianus, found in Yellowstone Schiraldi C, De Rosa M (2002) The production of
National Park, and operates over between 30 C biocatalysts and biomolecules from extremophiles.
and 94 C (pH range 610). Its outstanding Trends Biotechnol 20(12):515521
Extremozymes 759
Schiraldi C, Acone M, Giuliano M, Carten M, De Rosa M advances in thermophilic whole-cell biocatalysts.

(2001) Innovative fermentation strategies for the pro- Microb Biotechnol 4(4):438448
duction of extremophilic enzymes. Extremophiles Thompson VS, Schaller KD, Apel WA (2003) Purication
5(3):193198 and characterization of a novel thermo-alkali-stable
Taylor MP, van Zyl L, Tufn IM, Leak DJ, Cowan DA catalase from Thermus brockianus. Biotechnol Prog
(2011) Genetic tool development underpins recent 19:12921299
E
F
Facilitated Pervaporation pervaporation assisted reactor for further

explanation.
Bart Van der Bruggen However, facilitated pervaporation may rather
Department of Chemical Engineering, Process be related to a hybrid membrane extraction sys-
Engineering for Sustainable Systems (ProcESS), tem, as described by Bessarabov et al. (1999) for
KU Leuven, Leuven, Belgium the separation of liquid olen/parafn mixtures in
the range up to C6. Nonporous polymeric mem-
branes are used in a membrane contactor and a
A facilitated pervaporation system usually refers high-boiling-point selective liquid extractant
to the combination of a pervaporation as a sepa- owing along polymeric membranes between
ration unit combined with a chemical reactor for the extractor and the membrane stripper. The ole-
esterication reactions or other equilibrium reac- n components permeated through the polymeric
tions. The facilitation then refers in fact to the lms at elevated temperatures in a membrane
reaction, and not the separation. The stripper, making it possible to obtain
pervaporation unit removes the side product of pervaporation transport. This is schematically
the reaction (in case of esterication, this is shown in Fig. 1 (Bessarabov et al. 1999).
water) so that the reaction equilibrium shifts to a Another approach to obtain a pervaporation
higher product yield (Van der Bruggen 2010). effect is by the use of xed recognition sites in
Other reactions than esterications can also be membrane contactors (Touil et al. 2006), so that
facilitated, on condition that they concern an equi- the selectivity needed for pervaporation is
librium reaction and that the side product can be obtained in a pertraction process (in which a
easily removed by a pervaporation membrane porous membrane without any selectivity is
(as is the case for water in an organic reaction used). The combination of selectivity characteris-
mixture). For bioconversions, the reaction prod- tic for pervaporation and the pertraction effect
uct(s) are to be removed instead of the results in an enhancement, which can be denoted
by-products, in order to enhance the activity of as facilitated pervaporation. Cyclodextrins, for
the microbial population. This is the case, for example, can be used to form inclusion complexes
example, in the production of bioethanol and in of different stability with organic molecules in the
ABE (acetone, butanol, ethanol) fermentations, separation of p-xylene, m-xylene, and o-xylene
where ethanol (or butanol) is removed by a hydro- (Touil et al. 2006). Kusumosahyo et al. (2004)
phobic pervaporation membrane. See also made a similar study on polyacrylic acid (PAA)
pervaporation membrane reactor and membranes and concluded that the native PAA

DOI 10.1007/978-3-662-44324-8
762 Facilitated Pervaporation
Inlet for extractant obtained by the incorporation of carbon nanotubes

(CNTs) in the pores of a polyvinylidene uoride
A (PVDF) membrane (Sae-Khow and Mitra 2009),
and for polyvinyl alcohol (PVA) membranes
(Bryant et al. 1997), and many other publications
can be found on similar effects in mixed matrix
Contacting
Hydrocarbon membranes, nanoparticles-enhanced membranes,
spacers
feed layer and hybrid organicinorganic membranes. Chem-
ical modications have also been studied for
enhancement of pervaporation. Wu et al. (2002)
Extractant
layer studied cross-linked polydimethylsiloxane
(PDMS) membranes modied by the sequential
introduction of two different side-arm functional
groups, -(CH2)(3)OC2H5 and -(CH2)(3)NMe2,
for the pervaporative recovery of p-cresol from
aqueous streams, which was successful.
Pumping
Cooling B
Heating
STRIPPER
References
Bessarabov DG, Theron JP, Sanderson RD, Schwarz HH,

Non-porous Schossig-Tiedemann M, Paul D (1999) Separation
membrane of 1-hexene/n-hexane mixtures using a hybrid mem-
brane/extraction system. Sep Purif Technol 16(2):
167174
Bryant DL, Noble RD, Koval CA (1997) Facilitated-
Degassing transport separation of benzene and cyclohexane with
C=C poly(vinyl alcohol)-AgNO3 membranes. J Membr Sci
127(2):161170
Facilitated Pervaporation, Fig. 1 Facilitated Kusumosahyo SP, Kanamori T, Sumaru K, Iwatsubo T,
pervaporation using a hybrid membrane extraction system Shinbo T (2004) Pervaporationof xylene isomer mix-
(for olen separation). A is an extraction module, B is a ture through cyclodextrins containing polyacrylic acid
membrane pervaporation cell (Reprinted with permission membranes. J Membr Sci 231:127132
of Elsevier from Bessarabov et al. 1999) Liu WP, Li B, Cao RJ, Jiang ZY, Yu SN, Liu GH,
Wu H (2011) Enhanced pervaporation performance of
poly (dimethyl siloxane) membrane by incorporating
titania microspheres with high silver ion loading.
membrane was almost impermeable for the xylene
isomers, and the incorporation of cyclodextrins in Sae-Khow O, Mitra S (2009) Fabrication and characteriza-
the PAA membranes resulted in membranes hav- tion of carbon nanotubes immobilized in porous
ing a molecular recognition function, which selec- polymeric membranes. J Mater Chem 19(22):
37133718
tively facilitated the transport of the xylene
Touil S, Tingry S, Bouchtalla S, Deratani A (2006) Selec-
isomers. tive pertraction of isomers using membranes having
A further use of the term facilitated xed cyclodextrin as molecular recognition sites. Desa-
pervaporation refers to the membrane material lination 193:291298
Van der Bruggen B (2010) Pervaporation membrane reac-
itself. Liu et al. (2011) incorporated Ag(+) loaded tors. In: Enrico D, Lidietta G (eds) Comprehensive
titania (TiO(2)) microspheres into a poly(dimethyl membrane science and egineering, vol 3. Academic,
siloxane) (PDMS) matrix for pervaporative desul- Oxford, pp 135163
furization of model gasoline and demonstrated Wu P, Brisdon BJ, England R, Field RW (2002) Preparation
of modied difunctional PDMS membranes and a com-
facilitated transport through these membranes, as
parative evaluation of their performance for the
could be seen from increased uxes and separa- pervaporative recovery of p-cresol from aqueous solu-
tion factors. A similar enhancement effect was tion. J Membr Sci 206:265275
Facilitated Transport 763
at both sides of that membrane phase will be

Facilitated Transport transported through it (Fig. 1a). The component
does not react chemically in the liquid membrane,
Gerardo Len and then it remains in the same chemical form in
Departamento de Ingeniera Qumica y the feed, in the membrane, and in the product
Ambiental, Universidad Politcnica de Cartagena, phases. Transport will end when an equality of
Cartagena, Spain concentration of the component prevails on both
sides of the membrane phase. In order to improve
the effectiveness of the separation process, facili-
Synonyms tated transport mechanisms have been described
(Vitt et al. 2010), which maximize both the extrac-
Assisted transport tion velocity (ux through the membrane phase) F
and the reception capacity of the diffusing specie
Transport mechanism used to improve ux, per- in the product phase. Moreover, facilitated trans-
meability, and selectivity in separations by liquid port mechanisms combine the process of extrac-
membranes. Mass transport selectivity and trans- tion and stripping in a single unit operation, and
fer velocity in a liquid membrane are mainly they offer the possibility of transporting a compo-
determined by the solubility and the diffusion nent against its own concentration gradient
coefcient in the membrane phase of the compo- (Ho and Li 1992; Kislik 2010). Two types of
nent to be separated. The driving force of the facilitated transport are usually considered: type
transport process is the gradient of chemical 1 and type 2. In type 1 facilitation, a stripping
potential, normally due to a concentration gradi- agent is added to the product phase, which quan-
ent between the feed and the product phases. titatively reacts with the diffusing specie to yield a
Then, a component which is soluble in the mem- membrane-insoluble product (Fig. 1b). Thus, con-
brane phase and whose concentration is different centration of that specie is reduced to zero in the
a b c
F M P F M P F M P
S
AC
A A A A A A A A
C
AS
d e f
F M P F M P F M P
S A A
AC A
A ABC
AC
A A
C B C B
B B
BC
AS
Facilitated Transport, Fig. 1 Schematic representation of the different transport mechanisms in liquid membranes
764 Farmin 08
product phase, resulting in a high concentration Vitt ME, Ho WSW, Li NN (2010) Liquid membranes. In:
gradient across the membrane phase. In type Drioli E, Giorno L (eds) Comprehensive membrane
science and engineering, vol 4, 1st edn. Elsevier,
2 facilitation (carrier-mediated transport), a Amsterdam, pp 79106
reagent is incorporated in the membrane phase to
carry the diffusing specie across the membrane to
the product phase. The carrier forms, reversibly
with the specie to be separated, a compound Farmin 08
which is soluble in the membrane. This compound
is formed in the interface feed phase/membrane Alamine 336
phase, it is transported through the membrane
phase due to its concentration gradient, and it is
broken in the membrane phase/product phase
interface. The specie to be separated is then liber- Fat Hydrolysis, Membrane
ated in the product phase, and the carrier diffuses Operations of
in the contrary way through the membrane phase,
as a consequence of its concentration gradient, to Angela C. Macedo1 and F. Xavier Malcata2
the opposite interface, to initiate a new separation 1
LEPABE Laboratory of Engineering of
cycle (Fig. 1c). As the process ends when the Processes, Environment, Biotechnology and
concentration of the specie to be separated equal- Energy, University of Porto, Porto, Portugal
izes on both sides of the membrane phase, carrier- 2
DEQ Department of Chemical Engineering and
mediated transport can be improved by incorpo- LEPABE Laboratory of Engineering of
rating a chemical reaction on the receiving side, as Processes, Environment, Biotechnology and
illustrated in Fig. 1d. The concentration of the Energy, University of Porto, Porto, Portugal
specie to be separated is reduced to practically
zero on the product side through an auxiliary
reaction in the product phase, to form an insoluble Lipids that are solid at room temperature are
compound in the membrane. The transportation of termed fats. In nature, most fats are triglycerides
two components is possible, both from the feed consisting of a glycerol backbone, to which three
phase (Fig. 1e) or one from the feed phase (the individual fatty acid moieties are attached. The vari-
component to be separated) and the other from the ety of fats in nature and their diversity in terms of
product phase (Fig. 1f). Both cases are known as physicochemical properties come from the number
coupled transport, the rst being denominated of combinations of the hundreds of existing fatty
cotransport and the second counter-transport. acid residues which, due to their underlying met-
abolic synthesis pathway, are characterized by an
even number of carbon atoms.
Cross-References In view of their insolubility and the relevance to
water-based life, fats often appear stabilized via
Carrier-Mediated Transport emulsication in aqueous media; this is notable in
the case of milk fat, with globules surrounded by a
membrane containing proteins and phospholipids.
References
Said native globules range in size from less than
Ho WSW, Li NN (1992) Emulsion liquid membranes. 1 mm to over 10 mm. The uneven size distribution
Denitions. In: Ho WSW, Sirkar KK (eds) Membrane provides larger globules with a tendency to oat:
handbook, 1st edn. Van Nostrand Reinhold, New York, under static conditions, that buoyancy effect leads
pp 597610 to creaming at the top of the container, but the
Kislik VS (2010) Introduction, general description, deni-
tions and classication. Overview. In: Kislik VS process may be speeded up by centrifugation or
(ed) Liquid membranes, 1st edn. Elsevier, Amsterdam, ltration. The major disadvantage of the latter is the
pp 115 difculty in removing the fat layer once
Fat Processing, Membrane Operations of 765
formed coupled with clogging of the membrane et al. 1998); these are in higher and higher demand
pores which is particularly detrimental in the case for nutritional purposes.
of hydrophobic matrices. Milk is often homoge- Depending on the level of water in their vicinity,
nized by forcing it into a tiny nozzle under high membrane-immobilized lipases operating on a fat
pressure, so as to reduce the average globule size to substrate may also perform ester synthesis, as well
less than 1 mm; this creates a more uniform distri- as interesterication of the acidolysis, alcoholysis,
bution of globules that improves kinetic stability of and transesterication types depending on whether
the emulsion (Fox and McSweeney 1998). the aqueous substrate is replaced by a fatty acid, an
Hydrolysis of fats can be effected by lipases alcohol, or another fat, respectively. This process is
(and esterases, to a lesser extent), which require a relatively easy to bring about, but suffers from two
fat/water interface be present; this type of interfa- major shortcomings: enzymes tend to deactivate as
cial reaction is thus promoted by high specic time online elapses, thus calling for enzyme F
areas, which may be attained at the expense of makeup; and product separation is often difcult,
addition of emulsiers and vigorous stirring. How- especially when only lipid phases are involved and
ever, lipases are typically much more expensive resorts to distillation (which should be carried out at
than their fat substrates, so some form of immobi- low pressure in the case of heat-labile fats).
lization thereof is in order for economic feasibility;
this approach is particularly simple because such
enzymes do not require cofactors for catalytic References
activity, while possessing a relatively high hydro-
phobic nature that facilitates spontaneous adhesion Fox PF, McSweeney PLH (1998) Milk lipids. In: Fox PF,
McSweeney PLH (eds) Dairy chemistry and biochem-
to hydrophobic materials via adsorption.
istry. Blackie Academic and Professional Publishers,
Several immobilized microbial lipases are avail- London, pp 67146
able from industrial suppliers, using, e.g., ion Willis W, Lencki E, Marangoni A (1998) Lipid modica-
exchange resins and polymeric materials as sup- tion strategies in the production of nutritionally func-
tional fats and oils. Crit Rev Food Sci Nutr 38:639674
ports; they have been applied to hydrolyze fats in
emulsied form, but show the disadvantage that the
fat and enzyme particles must collide for catalysis to
occur. Porous membranes constitute an interesting
alternative, because lipase can be immobilized Fat Processing, Membrane
inside their pores with water pumped along one Operations of
side and (melted) fat along the other. Besides low
rates of fouling, said tangential membranes may Angela C. Macedo1 and F. Xavier Malcata2
offer high volumetric efciencies especially if 1
LEPABE Laboratory of Engineering of
they are manufactured in hollow-ber form and Processes, Environment, Biotechnology and
permit the concentration of fat to remain at its max- Energy, University of Porto, Porto, Portugal
imum level ever by circumventing phase mixing. 2
DEQ Department of Chemical Engineering and
Furthermore, immobilized lipase processes LEPABE Laboratory of Engineering of
offer fatty acid- and stereo-specicity advantages Processes, Environment, Biotechnology and
over plain acid- or base-catalyzed hydrolysis of fats Energy, University of Porto, Porto, Portugal
which are intrinsically nonspecic. Considering
the difference in metabolism of fatty acids
depending on their original regiodistribution and Fat processing is an active business, designed to
chain length, lipases have been largely applied to create added value from slaughter by-products
produce tailor-made derivatives of triacylglycerols downstream the animal chain. It encompasses
with medium- or short-chain fatty acid residues in activities dealing with production of animal fats
external positions (sn-1 and sn-3) and long-chain (and proteins) suitable for human consumption in
fatty acids in the internal position (sn-2) (Willis agreement with directive 77/99/EEC (EU 1997),
766 Fat Processing, Membrane Operations of
as well as processing of category 3 (formerly enhances thermostability, alleviates rate depen-

labeled low risk) materials from animals approved dence on pH, and is more exible in terms of
for human consumption in agreement with regu- reactor conguration.
lation ABPR 1774/2002/EC (EU 2002). Fat Unlike hydrolases at large that operate in plain
processing has historically been associated with aqueous phases, lipases are activated only in the
manufacture of high-grade animal fats for specic presence of water/lipid interfaces provided by
markets, e.g., bakery industry, calf milk replacers, the contacting surface between the immiscible
and pet food as well as commodities for formu- aqueous and fat phases, which can be made to
lation in the oleo-chemical industry (e.g., manu- coincide with a porous membrane. Techniques
facture of soaps) and cosmetics; it uses bovine that met with industrial success for immobiliza-
tallow, porcine lard, and poultry fat as major tion of lipases include adsorption, covalent
feedstocks. binding, and entrapment/microencapsulation
The basic operations in fat processing are wet (Knezevic et al. 2004). Note that the economic
melting and dry rendering or hybrid systems feasibility is directly related to the volumetric
thereof. Fat tissues are subjected to relatively productivity on the long run, which depends in
mild processing conditions to minimize their turn not only on the operational stability of the
decay but high-quality nal products hinge crit- enzyme but also on its actual activity which is
ically on the quality and source of the raw mate- normally lower than that of free enzyme, in addi-
rials. They are usually collected from tion to mass transfer constraints (Vulfson 1994).
slaughterhouses from animals slaughtered on the Physical adsorption is the simplest method for
same day and undergo strict food safety steps lipase immobilization; binding occurs via hydro-
according to HACCP methodology including gen bonds, salt linkages, and van der Waals
monitoring of environmental contaminants, forces. The process is carried out at mild condi-
growth hormones, and veterinary drugs. Mem- tions, without (or with a marginal) support activa-
brane processing is possible, but only on clean tion and in the absence of extraneous reagents; it is
melted fat and chiey to deliver support for thus prone to preserving enzyme activity and
enzymes aimed at controlled chemical modica- specicity. The chemical nature of the carrier
tions of such fat, e.g., interesterication. (including the ratio of hydrophilic to hydrophobic
Fat interesterication can be carried out via groups), the particle size and porosity, and the
chemical or enzymatic processes; the former surface area available determine the amount of
require metal alkalis as catalysts, but produce enzyme bound, as well as the enzyme behavior
random interchange of acyl groups, lead to con- following immobilization (Cao 2005).
tamination of the nal product with residual cata- Chemical reaction between activated amino
lyst, and allow formation of soaps all of which acid residues far from the catalytic and binding
are viewed as drawbacks relative to their enzy- site of the enzyme and active functionalities on the
matic counterparts (Erickson 1995). Lipase- carrier is the basis of (effective) covalent attach-
mediated interesterication with lipases is charac- ment. This is clearly the most efcient mode of
terized by great versatility, reasonable substrate immobilization despite its being much more
selectivity (including regio-and enantios- detrimental for enzyme activity, complex in
electivity), possibility of operation at room tem- terms of implementation, and heavily dependent
perature and pressure, and the absence of side on the properties of the carrier (Cao 2005).
reactions. To fully exploit the technical advan- Finally, entrapment involves capture of
tages of such enzymes, economic feasibility rec- enzyme within a polymeric matrix during the
ommends the use in immobilized form to extend polymerization process whereas (the related)
their useful life, avoid product contamination and microencapsulation entails connement by a
consequent unstability, and circumvent the need membrane-like physical barrier around the
for emulsication (James et al. 2009). Lipase enzyme preparation; entrapped enzymes may be
immobilization also facilitates product separation, further microencapsulated. Both processes require
Faujasite 767
simple equipment and relatively inexpensive

reagents and yield a better stability and an activ-
ity similar to adsorbed enzymes (Cao 2005).
References
Cao L (2005) Carrier-bound immobilized enzymes. Prin-

ciples, application and design. Wiley-VCH, Weinheim
Erickson DR (1995) Practical handbook of soybean
processing and utilization. American Oil Chemists
Society, Champaign
EU (1997) Council Directive 77/99/EEC of 21 December
1976 on health problems affecting intra-community
F
trade in meat products. Off J Eur Commun L26
EU (2002) Regulation (EC) No. 1774/2002 of the Euro-
pean Parliament and of the Council laying down health
rules concerning animal by-products not intended for
human consumption. Off J Eur Commun L273:195
James J, Simpson BK, Marshall MR (2009) Application of Faujasite, Fig. 1 The FAU-type zeolite supercage view
enzymes in food processing. Crit Rev Food Sci Nutr along [111] direction (International Zeolite Association (IZA))
36:437463
Knezevic ZD, Siler-Marinkovic SS, Mojovic LV
(2004) Immobilized lipases as practical catalysts.
window openings with the aperture of 8 and
APTEFF 35:1280
Vulfson EN (1994) Industrial applications of lipases. In: supercages of approximately 12 (McCusker and
Woolley P, Peterson SB (eds) Lipases: their structure, Baerlocher 2001). The FAU-type zeolite
biochemistry and application. Cambridge University supercage viewed along [111] is shown in Fig. 1
Press, Cambridge, UK, pp 271288
(International Zeolite Association (IZA)).
The combination of large void volume and
large pore openings in a three-dimensional chan-
nel system makes FAU thermally stable molecular
Faujasite sieve ideal for many catalytic, ion-exchange, and
adsorption applications. FAU zeolites are gener-
Anne Julbe and Martin Drobek ally synthesized by hydrothermal crystallization
Institut Europen des Membranes, Universit de of reactive alkali metal aluminosilicate gels or
Montpellier, Montpellier, France clear solutions at low temperatures (70300 C,
usually 100 C) and pressure (autogenous) under
alkaline conditions. Starting typically from
FAU (faujasite) belongs to the family of alumino- sodium aluminate and sodium silicate, the zeolite
silicate molecular sieves dened by the formula
is obtained in its Na+ form (Blatter and
Ca, Mg, Na2 H2 O Al58 Si134 O384 FAU Schumacher 1991). Replacement of the Na+ cat-
29 240
(www.iza-online.org). Two types of synthetic ions by protons (leading to the H-zeolite form) is
FAU zeolites referenced as Linde X and Linde required for applications in acid catalysis. In addi-
Y differ in Si/Al atomic ratio, which is typically tion, for specic catalytic applications, incorpora-
in the range from 1 to 1.5. one to two for the X and tion of various metal cations into the structure is
higher for the Y zeolite. The 24-tetrahedra carried out by impregnation or ion exchange and
cuboctahedral units (sodalite cages) in the FAU results in modication of the number and nature of
framework type are connected via hexagonal acid sites inuencing the diffusion of reactants
prisms (double six-rings) forming a three- and products. X-type zeolite has a wide range of
dimensional porous channel structure along industrial applications for gas or vapor adsorption,
[110]. These are characterized by 12-oxygen ring separation, and as catalyst (isomerization of
768 Feed Pressure
1-butene, alkylation of toluene with ethylene or crystallizers, reverse osmosis, nanoltration,

methanol, cycloaddition of carbon dioxide to ultraltration, microltration, etc. as well the
ethylene oxide (Ribeiro et al. 1984)). However, more conventional unit operations, like reactors,
in many large-scale industrial applications, its distillation columns, adsorbers, absorbers, liquid-
chemical analogue with Si/Al ratio higher than liquid extractors, lters, pipelines, etc., are fed by
1.5 (zeolite Y) has superseded zeolite X because a stream entering the unit. The pressure of this
of its higher chemical and thermal stability. USY stream (to be treated by the unit operation) at the
(ultra-stable Y zeolite prepared by dealumination entering section of each unit operation is dened
of zeolite Y) is one of the most widely employed as feed pressure.
zeolitic materials in the petrochemical industry, A further specication is required for mem-
essentially as a uid cracking catalyst (FCC) of brane systems. Often, an auxiliary stream named
heavy petroleum distillates, for increasing the sweep is fed to the membrane unit. This stream
yield of gasoline and diesel fuel from crude oil. is characterized by a different pressure named
Considering membrane applications, the produc- sweep pressure; this pressure cannot be referred
tion of tubular FAU zeolite membranes has been to as feed pressure.
mainly considered for alcohol dehydration by
vapor permeation (Sato et al. 2008).
References Fenton Test
Blatter F, Schumacher E (1991) The preparation of pure P. Knauth

zeolite NaY and its conversion to high-silica faujasite, Madirel UMR 7246, Aix Marseille University-
an experiment for laboratory courses in inorganic CNRS, Marseille Cedex, France
chemistry. J Chem Educ 68:519521
International Zeolite Association (IZA). http://www.iza-
online.org. Accessed July 2013
McCusker LB, Baerlocher C (2001) Zeolite structures. In:
van Bekkum H, Flanigen EM, Jacobs PA, Jansen JC Fentons reagent was developed in the 1890s by
(eds) Introduction to zeolite science and practice,
vol 137, Studies in Surface Science and Catalysis.
Henry John Horstman Fenton as an analytical
Elsevier, London, pp 3767 reagent (Fenton and Jones 1900). It is a solution
Ribeiro FR et al (1984) Zeolites: science and technology. of hydrogen peroxide and an iron catalyst. Iron
Martinus Nijhoff Publishers, The Hague (II) sulfate is a typical iron compound in Fentons
Sato K, Sugimoto K, Nakane T (2008) Mass-production of
tubular NaY zeolite membranes for industrial purpose
reagent.
and their application to ethanol dehydration by vapor The mechanism mostly accepted today,
permeation. J Membr Sci 319:244255 suggested by Haber and Weiss in the 1930s
(Haber and Weiss 1932), is that iron(II) is oxi-
dized by hydrogen peroxide to iron(III), a
Feed Pressure hydroxyl radical and a hydroxide anion
(reaction 1):
Giuseppe Barbieri
Institute on Membrane Technology, National Fe2 H2 O2 ! Fe3 OH OH (1)
Italy Iron(III) is then reduced back to iron(II), a perox-
ide radical and a proton by the same hydrogen
peroxide (dismutation reaction 2):
Membrane operations, for instance, membrane
reactors, membrane gas separators, membrane
contactors, membrane emulsiers, membrane Fe3 H2 O2 ! Fe2 OOH H (2)
Fiber Models 769
In the overall reaction, two molecules of

hydrogen peroxide are converted into two rad- Fiber Models
icals and water by the presence of catalytic
iron ions. Vasilis Burganos
The free radicals generated by Fentons reagent Institute of Chemical Engineering Sciences,
induce then secondary reactions: the free Foundation for Research and Technology
hydroxyl radical is a powerful, non-selective oxi- Hellas, FORTH/ICE-HT, Patras, Hellas
dant used to oxidize organic contaminants in
waste waters to primarily carbon dioxide and
water. A ber model is a model that uses straight or
Nowadays, Fentons test is applied for the curved bers, either nite in length or innitely
accelerated testing of the oxidation stability of long, to represent the solid phase in a porous F
proton exchange membranes (PEM) for fuel cell material or a reinforcement component in a com-
applications (Healy et al. 2005). Generally, the posite material. Fiber models are the appropriate
weight loss after a dened treatment time is used choice for modeling brous media, woven or non-
to monitor membrane degradation. Typical woven, typically synthetic but occasionally also
Fentons test conditions for PEM membranes are natural ones. Typical examples include the repre-
an aqueous solution with 30 % H2O2 and 20 ppm sentation of membranes and porous media for fuel
Fe2+ at 8090 C with three treatment cycles with cells (Mathias et al. 2003), lters for the separa-
fresh reagent and 24 h per cycle. Given that the tion or sieving of particulate matter, or lters for
initial H2O2 concentration is much higher than in the exclusion of bubbles in diverse applications of
an operating fuel cell, softer alternative condi- microuidics. There is a recent rapid growth of
tions are also used with an aqueous 3 % H2O2 interest in the utilization of ber models for the
solution mixed with 4 ppm Fe2+ at 68 C description of the structure of gas diffusion layers
(Ramaswamy et al. 2008). in fuel cells but also of modern textiles and fabrics
Although Fentons test is straightforward and for specialized applications (Thiedmann
considered as a benchmark for PEM evaluation, it et al. 2009; Gaiselmann et al. 2012). The typical
has inherent limitations, given that the membrane features that characterize a ber model include the
degradation in this test is not related to an elec- diameter and length of the ber, the number or
trode process and does not involve variations in length density per unit volume, the solid fraction,
fuel cell operating conditions, e.g., operating the shape of the bers, and the elastic properties of
potential, relative humidity, and the presence of the bers that determine the macroscopic mechan-
fuel and oxygen. ical properties of the material. Carbon cloth,
electrospun polyacrylonitrile, and polyester are
common examples of brous materials that can
be represented by ber models. To convert to a
References pore model, various attempts have been made to
infer some effective pore sizes from the ber
Fenton HJH, Jones HO (1900) The oxidation of organic
acids in presence of ferrous iron part I. J Chem Soc model, usually with the help of inscribed
77:6976 spheres among neighboring ber segments.
Haber F, Weiss J (1932) On the catalysis of hydroperoxide. Skeletonization of a ber model is often part of
Naturwissenschaften 20:948950
Healy J, Hayden C, Xie T, Olson K, Waldo R, Brundage A,
the analysis routine to facilitate the comprehen-
Gasteiger H, Abbott J (2005) Aspects of the chemical sion of the ber cluster articulation and the even-
degradation of PFSA ionomers used in PEM fuel cells. tual determination of the topology of the system.
Fuel Cells 5(2):302308 Useful concepts from straight-line path statistics
Ramaswamy N, Hakim N, Mukerjee S (2008) Degradation
or randomness of secant distribution through con-
mechanism study of peruorinated proton exchange
membrane under fuel cell operating conditions. vex bodies (Coleman 1969) are incorporated in
Electrochim Acta 53(8):32793295 this type of models. The bers can be hollow or
770 FIB-SEM Tomography
solid, randomly oriented or ordered to arbitrary were developed in the early 1970s as a tool to either
degree, and either charged or neutral depending etch or deposit material on solids with (sub-)
on the application. Fiber models of membrane micrometer precision, via either volatilization or
materials and, more generally, brous media deposition processes, respectively, assisted by a
lend themselves to the numerical simulation of beam of ions. They soon became an essential
transport phenomena through their structure, usu- micromachining tool in the semiconductor indus-
ally diffusion, single or, two-phase ow, disper- try. Later, the FIB technology was integrated into
sion, particle attraction and deposition, combined SEM devices. Such combination enables direct
phase transition and ow, heat conduction, elec- imaging of those solid surfaces previously
trical conduction, and light transmission processed with the FIB via detection of either
(Torquato 2002; Tomadakis and Sotirchos 1993). backscattered or secondary electrons. A FIB-SEM
device comprises a built-in FIB gun and a SEM
detector, typically oriented forming an angle of
References 5255 . Advances in FIB resolution and automa-
tion have paved the way to FIB-SEM tomography.
Coleman R (1969) Random paths through convex bodies. Tomography, from Ancient Greek terms tomos
J Appl Probab 6:430441
(slice) and grapho (to write, to record), refers to
Gaiselmann G, Thiedmann R, Manke I, Lehnert W,
Schmidt V (2012) Stochastic 3D modeling of ber- the possibility of reconstructing the three-
based materials. Comput Mater Sci 59:7586 dimensional structure of an object, or a selected
Mathias MF, Roth J, Fleming J, Lehnert W (2003) internal plane, by gathering information on discrete
Diffusion media materials and characterisation.
thin slices or two-dimensional projections of said
In: Vielstich W, Lamm A (eds) Handbook of fuel
cells, volume III, chapter 42, 517537, J. Wiley & object.
Sons, London In a typical FIB-SEM tomography experiment,
Thiedmann R, Hartnig C, Manke I, Schmidt V, Lehnert the specimen, often embedded in a resin, is situated
W (2009) Local structural characteristics of pore space
at the coincident point of the FIB and SEM beams.
in GDLs of PEM fuel cells based on geometric 3D
graphs. J Electrochem Soc 156:B1339B1347 Generally, solid specimens are previously sputtered
Tomadakis MM, Sotirchos SV (1993) Ordinary and tran- with metal overlays to protect the sample surface
sition regime diffusion in random ber structures. during FIB milling and to increase its conductivity,
AIChE J 39:397412
in the case of specimens with insufcient electrical
Torquato S (2002) Random heterogeneous materials
microstructure and macroscopic properties. Springer- conductivity, thereby preventing charging effects in
Verlag, New York the course of the experiment. A trench is rst carved
on the upper surface of the sample with the FIB.
Then several cross-sections of the internal structure
of the specimen, exposed upon sequential erosion of
FIB-SEM Tomography thin layers with the FIB, are consecutively imaged
with the SEM. Figure 1a depicts schematically the
Gonzalo Prieto working principle of the serial milling and imaging
Max-Planck-Institut fur Kohlenforschung, procedure in FIB-SEM tomography.
Mulheim an der Ruhr, Germany The outcome is a stack of SEM micrographs of
the consecutive cross-sections, which after align-
ment and further processing serves to reconstruct
FIB-SEM tomography (also known as tomo- the investigated volume in three dimensions
graphic FIB-SEM) is a micrometer-to-nanometer (tomogram). Typically, sample volumes in the
scale imaging technique which combines a range of 150 mm3 are studied in a single
focused ion beam (FIB) (Yao 2007) with a scan- FIB-SEM experiment. 3D imaging resolutions
ning electron microscope (SEM) (Goldstein down to 1030 nm can be achieved with modern
et al. 2003) to obtain information on the internal equipments, being the limiting factor the resolu-
structure of solid specimens. Focused ion beams tion in the z-direction, i.e., the FIB-milling
FIB-SEM Tomography 771
F
FIB-SEM Tomography, Fig. 1 (a) Schematic represen- membrane incorporating metal-organic-framework
tation of the alternating FIB-milling and SEM-imaging (MOF) crystals within a polymer matrix. MOF particles
during a FIB-SEM tomography experiment on a mem- are depicted in green, membrane void defects are shown in
brane specimen. (b) 3D, surface-rendered view of a red, while the polymer matrix is depicted transparent
reconstructed FIB-SEM tomogram of a mixed-matrix (Rodenas n.d.)
direction. Such resolution closes the gap between image the most common membrane defects such
other three-dimensional imaging methods such as as pinholes, cracks, or voids. In this respect, FIB
electron tomography, which relies on transmitted milling represents an advanced solution to expose
electrons, and techniques based on X-rays such as membrane cross-sections for imaging, as compared
X-ray tomography. As a result of the detailed and to more conventional approaches such as fracturing
three-dimensional information provided on the at cryogenic temperatures, which are prone to
internal structure of a wide array of solid speci- introduce a number of artifacts in the freshly gen-
mens, FIB-SEM tomography has become an out- erated surface to be imaged. The aforementioned
standing tool in a number of scientic and structural features are known to be of signicance
technological disciplines, including life-, earth-, for key performance parameters such as selectivity
and materials science. and permeability. Hence, detailed visualization of
Due to the typical sizes of membrane pores, internal structures with FIB-SEM tomography con-
ranging from nanometers to hundreds of micro- tributes to the design and optimization of the mem-
meters, the resolution of this technique is suitable brane production and processing procedures. FIB
to study the internal porosity of a broad array of erosion can also be applied to produce thin tips or
technologically relevant membranes, including slices from thicker membrane specimens, as
analysis of different layer components in asym- required to be studied with complementary three-
metric membranes. In addition, the information dimensional imaging methods such as atom-probe
gathered in FIB-SEM tomograms can be tomography, X-ray tomography, or electron
interpreted on a quantitative basis by combination tomography.
with suitable image analysis methods (Haidekker
2011). In multicomponent membranes, FIB-SEM
tomography enables assessing the proportion, Cross-References
degree of integration, and relative spatial location
of different internal elements in three dimensions. Asymmetric Membrane
As an example, Fig. 1b depicts a reconstructed Membrane Micrograph
FIB-SEM tomogram of a mixed-matrix mem- Metal-Organic Framework
brane incorporating metal-organic-framework Mixed Matrix Membranes (MMMs)
crystals within a polymer matrix. Three- Permeability
dimensional visualization can also be applied to 3D-Imaging Methods for Membranes
772 Ficks Laws
References and molar fraction of the species A, respectively;

and DA [m2 s1] is the multicomponent diffusion
Goldstein J, Newbury DE, Joy DC, Lyman CE, Echlin P, coefcient, which generally depends on tempera-
Lifshin E, Sawyer L, Michael JR (2003) Scanning elec-
ture, pressure, and species compositions. Equa-
tron microscopy and x-ray microanalysis, 3rd edn.
Springer, New York tions 1 and 2 are vectorial equations, as ux is a
Haidekker M (2011) Image processing and analysis: vector. The scalar versions of Eqs. 1 and 2 along a
advanced biomedical image analysis. Wiley, New specic direction (x) are provided by Eqs. 3 and 4,
Jersey
respectively:
Rodenas T et al. unpublished results
Yao N (2007) Focused ion beam systems: basics
and applications. Cambridge University Press, @wA
jA jx rDA (3)
Cambridge, UK @x
@xA
J A jx cDA (4)
@x
Ficks Laws
For anisotropic media, the diffusion coefcient
Alessio Caravella depends on the diffusion direction, and thus, it is a
National Institute of Advanced Industrial Science second-order tensor (Eq. 5) rather than a scalar:
and Technology Research Institute for 2 3
Innovation in Sustainable Chemistry, ISC-AIST, DA, xx DA, xy DA, xz
Tsukuba, Ibaraki, Japan DA 4 DA, yx DA, yy DA, yz 5 (5)
DA, zx DA, zy DA, zz
Ficks laws (rst and second Ficks laws). These In Eq. 5, the diffusivity matrix is expressed with
two laws of diffusion were originally postulated respect to a general reference system. In general,
by the medical doctor Adolf Eugen Fick the multicomponent diffusion coefcient is not
(18291901) (from whom they take their name) directly known, it being obtained, on the contrary,
by analogy with energy transfer in conduction and from the binary diffusion coefcients, which
not by experiment (Fick 1855; Bird et al. 2002). can be measured experimentally and for which
First Ficks law. For a generic species A in a semiempirical relationships are available in the
mixture, rst Ficks law is referred to as the con- literature. The relationship between the
stitutive equation for mass ux of A stating the multicomponent diffusion coefcient and binary
proportionality between ux and mass gradient in diffusion coefcient is extensively reported, for
a certain domain, which can be used in solid, example, in the book of Krishna and Taylor (1993).
liquid, and gas phase (Bird et al. 2002; Krishna Second Ficks law. For a generic species A
and Taylor 1993). This relationship can be in the mixture, the mass conservation law
expressed both in term of mass (Eq. 1) and (scalar) within a tiny control volume can be
molar quantities (Eq. 2) as expressed both in terms of mass quantities
(Eqs. 6, 7, and 8),
jA rDA wA (1)
@ rA
nA jA r A (6)
J A cDA xA (2) @t
nA rA v rA rwA (7)
where jA and JA are the mass and molar ux of the
species A, respectively; r [kg m3] is the density
X
n
of mixture; c [mol m3] is its total concentration; v wi vi (8)
wA [kgA kg1] and xA [molA mol1] are the mass i1
Filler in Membranes 773
and molar quantities (Eqs. 9, 10, and 11), analo-

gously to what is reported for rst Ficks law. Filler in Membranes
@cA Antonino Salvatore Arico

N A J A RA (9)
@t CNR-ITAE Institute, Messina, Italy
N A cA v cA cxA (10)
Synonyms
X
n
v xi vi (11)
Inorganic llers
i1
where nA and NA are the mass and molar convec- Composite peruorosulfonic acid membranes F
tive ux, rA and RA are the mass and molar containing different types of inorganic llers
generation terms which can be positive such as hygroscopic oxides, surface-modied
(source) and negative (sink) , and vi is the arith- oxides, zeolites, inorganic proton conductors,
metic average of the velocities of all the molecules etc. have shown an increased conductivity with
of the ith species in mixture within the control respect to the bare peruorosulfonic membranes
volume. at high temperature (Aric et al. 1998, 2003). The
If there is no either net convective transport or presence of hygroscopic inorganic oxides inside
sinks/source of mass within the control volume, the composite membrane besides extending the
Eqs. 6 and 9 are simplied to give Eqs. 12 and 13, operation of peruorosulfonic membranes (e.g.,
which represent the common form of second Naon) in the high-temperature range reduces
Ficks law: the crossover effects by increasing the tortuosity
factor in the permeation path. Such effects are
particularly serious at high temperature in DMFC
@ rA
jA (12) systems. An appropriate tailoring of the surface
@t chemistry in these nanoparticles is a key step to
enhance water retention at high temperature.
@ c A Composite recast Naon membranes containing
JA (13)
@t inorganic llers have been employed in high-
temperature (~150 C) direct alcohol (Aric
In membrane technology, rst Ficks law is et al. 1998) and H2-air fuel cells (Watanabe
coupled with the second one to calculate ux et al.1996). These composite membranes were
and concentration proles of the species along originally developed for reduced humidication
and through the membrane in reaction and sepa- operation in polymer electrolyte fuel cells
ration processes. These are of practical interest to (Watanabe et al.1996) due to the enhanced water
model the membrane behavior, to eventually opti- retention inside the membrane by the effect of the
mize the operating conditions, and to maximize inorganic ller (Aric et al. 1998). A further
the process performance. advantage of composite membranes relies in the
barrier effect given by the inorganic ller for
methanol crossover (Ren et al. 1996).
References
Bird RB, Stewart WE, Lightfoot EN (2002) Transport

phenomena, 2nd edn. Wiley, New York References
Fick A (1855) Ueber diffusion. Ann Phys 94:5986
Krishna R, Taylor R (1993) Multicomponent mass transfer. Aric AS, Cret P, Antonucci PL, Antonucci V (1998)
Wiley, New York Comparison of ethanol and methanol oxidation in a
774 Finger-Like Structure
liquid-feed solid polymer electrolyte fuel cell at high developed. The nuclei of the polymer-lean phases
temperature. Electrochem Solid-State Lett 1:6668 initiate the nger-like structure. The nuclei are
Aric AS, Baglio V, Di Blasi A, Creti P, Antonucci PL,
Antonucci V (2003) Inuence of the acidbase charac- mostly formed just beneath the top layer, and
teristics of inorganic llers on the high temperature growth of the nuclei depends on the state of the
performance of composite membranes in direct metha- solution in the frontier of the phase separation
nol fuel cells. Solid State Ionics 161:251265 (Marcel 1996). The nuclei may grow if the poly-
Ren X, Wilson MS, Gottesfeld S (1996) High performance
direct methanol polymer electrolyte fuel cells. mer solution in the front is stable and could supply
J Electrochem Soc 143:L12 enough solvent (or solvent/nonsolvent) to sustain
Watanabe M, Uchida H, Seki Y, Emori M, Stonehart P the growth. The growth ends when new nuclei are
(1996) Self-humidifying polymer electrolyte mem- initiated in the frontier of phase separation or the
branes for fuel cells. J Electrochem Soc 143:3847
concentration of the polymer-lean phases enters
the glass transition point. At an extreme condition,
the nger-like structure in the PSf membrane pen-
Finger-Like Structure etrates through the whole cross section because
bottom PEI solution provides solvent to sustain
Tao He the growth of the polymer-lean phases in the PSf
Lab for Membrane Materials and Separation layer (Xiao et al. 2015).
Technology, Shanghai Advanced Research Finger-like structure is not strong enough for
Institute CAS, Shanghai, China many applications due to the weak mechanical
property, e.g., gas separation, nanoltation, and
reverse osmosis. Suppress of the nger-like struc-
Finger-like structure is one of the typical mem- ture is quite a scientic and artistic task for mem-
brane structures, as shown in Fig. 1. It is also brane scientists and engineers. Upon formation of
called macrovoid structure. Figure 1 shows asym- nuclei of polymer-lean phases, if there is no sol-
metric membranes with a thin top skin layer vent to sustain the growth, no nger-like structure
supported by a nger-like sublayer. Formation of would result. By adding a signicant amount of
nger-like structure follows two processes: pore nonsolvent into the polymer solution, it may gen-
initiation and growth. As a polymer solution erate simultaneously a lot of polymer-lean phase
undergoes a liquid-liquid phase separation, both nuclei, thus preventing the polymer-lean phases
polymer-rich and polymer-lean phases are from growing to full-sized nger-like macrovoid.
Finger-Like Structure, Fig. 1 Finger-like structure of a solution (17/83 wt.%) onto a PEI/NMP (20/80 wt.%)
polyethersulfone at sheet membrane. (a) Cross section solution and then immersed in water for precipitation.
of a PES ultraltration membrane; (b) cross section The ngers penetrate through the membrane
of a PSf membrane prepared by casting a PES/NMP
Fixed Bed Catalytic Reactor 775
Increasing the polymer viscosity may reduce the rst studies on membrane reactors focused on the
growth rate of the nuclei and allow more time for effect of the gas permeation through membranes
the formation of more nuclei in the front, leading on the reaction system (which is often a conven-
to less nger-like macrovoids. tional packed bed reactor).
As a rule of the thumb, instantaneous demixing In a packed bed membrane reactor, the catalyst
is generally related to the formation of a nger-like is conned in xed bed conguration, and it is in
macrovoid structure. Therefore, by adding a solvent contact with a permselective membrane. The most
into the coagulation bath, choosing a solvent/ used packed bed conguration is the tubular one
nonsolvent (as the coagulant) system with low afn- where the catalyst may be packed either in the
ity is common practice to suppress the nger-like membrane tube or in the shell side, while the
structure. For a certain solvent/nonsolvent pair, the permeation stream is collected in the other side
hydrophobicity of the polymer is also crucial. For of the membrane (in the case of hydrogen- F
polar solvent/nonsolvent systems, e.g., NMP/H2O, selective membranes) or one reactant is feed on
DMAc/H2O, and DMF/H2O, a PVDF/solvent solu- the other side of the membrane (in the case of
tion will end up with a sponge-like structure due to oxygen-selective membrane).
the strong hydrophobicity of the polymer in For multitubular membrane reactor congura-
nonsolvent-induced phase separation using water tions, the catalyst in tube conguration can be
as the coagulant. However, by adding hydrophilic preferred especially for construction reason and
polymer additive into the PVDF solution, nger- for the extent of bed-to-wall mass and heat trans-
like macrovoids are generated. Encouraging fer limitations which can be very detrimental in
delayed demixing is key to prevent the nger-like the catalyst in shell conguration.
structure, because onset of the nuclei is delayed and Often, a sweep gas can be used in the perme-
the top skin layer is densied. However, this is often ation side of the membrane in order to keep the
unfavorable for membrane performance. permeation hydrogen partial pressure as low as
possible for minimizing the membrane area
required for the hydrogen separation. This prac-
References tice is, for example, very useful if hydrogen for
ammonia plant is being produced, in which case
Marcel MHV (1996) Basic principles of membrane tech- an amount of nitrogen can be used for sweeping
nology, 2nd edn. Kluwer, Norwel
the permeation side producing a synthesis stream
Xiao P, Nghiem LD, Yin Y, Li X-M, Zhang M, Chen G,
Song J, He T (2015) A sacricial-layer approach to (N2/H2 = 1/3) ready for the nal reaction step. If
fabricate polysulfone support for forward osmosis a sweep gas is used in the permeation side, then a
thin-lm composite membranes with reduced internal packed bed membrane reactor can be used in both
concentration polarisation. J Membr Sci 481:106114
cocurrent or countercurrent modes.
An interesting feature of packed bed mem-
brane reactors is the possibility to operate them
Fixed Bed Catalytic Reactor in a reverse ow mode, integrating the reaction
and separation with the recuperative heat
Fausto Gallucci exchange inside the reactor. This operational
Chemical Process Intensication, Department of mode is quite interesting for partial oxidation of
Chemical Engineering and Chemistry, Eindhoven methane (POM) (Smit et al. 2005). In normal
University of Technology, Eindhoven, POM systems air and CH4 feed streams have to
The Netherlands be preheated to the reaction temperature, while
being POM reaction only slightly exothermic,
the external heat transfer between feed and
The xed bed (or packed bed) membrane reactor exhaust is very expensive. Therefore, recuperative
conguration is the rst and most studied cong- heat exchange is preferably carried out inside the
uration for membrane reactors. This is because the reactor.
776 Fixed Carrier Membrane
Although the tube in tube conguration is quite the production of syngas: an experimental study. Int
useful to work in lab scale and for proof of prin- J Chem React Eng 3
Tosti S, Basile A, Bettinali L, Borgognoni F, Gallucci F,
ciple of membrane reactors, for industrial scale Rizzello C (2008) Design and process study of Pd
some other congurations need to be used in membrane reactors. Int J Hydrog Energy 33(19):
order to increase the membrane area per volume 50985105
of vessel used. In fact, the amount of hydrogen
produced is directly related to the amount of mem-
brane area installed in the reactor. Starting for the
tube in tube conguration, a straightforward way Fixed Carrier Membrane
to increase the membrane area in packed bed is the
tube in shell conguration (Buxbaum 2002; Tosti Marius Sandru
et al. 2008). Polymer Particles and Surface Chemistry,
Although the packed bed conguration is very Materials and Chemistry, SINTEF, Trondheim,
advantageous and easy to operate, the production Norway
of thin membranes brought under the spotlight
one of the disadvantages of xed bed membrane
reactors: the inuence of bed-to-wall mass trans- Fixed (site) carrier membranes (FSC) combine the
fer limitations on the membrane area required. durability of a dense membrane with the selectiv-
Briey, as long as the hydrogen ux through the ity of a supported liquid membrane (SLM) over-
membrane is a limiting step, the effect of external coming the limitation of SLM, degradation due to
mass transfer limitations such as the limitations to the wash out of the carrier solution over time. The
hydrogen transport between the bulk of the cata- carrier is either covalently bonded to the polymer
lytic bed (where hydrogen is produced) and the chain (chained carrier) or immobilized in the poly-
membrane wall (where hydrogen is recovered) mer matrix by physical constraints or weaker
can be neglected. However, by increasing the ionic bonding. A xed carrier membrane will
membrane ux, the external mass transfer limita- separate mainly via facilitated transport mecha-
tions became limiting and determine the extent of nism, implying that the carrier will react speci-
membrane area. This has been demonstrated both cally with one of the components to be separated.
experimentally (Peters et al. 2008) and numeri- Noble (1990, 1991) and Cussler (Cussler
cally (Gallucci et al. 2010). This important draw- et al. 1989) developed two different models to
back has driven the research toward new reactor describe the transport across xed carrier mem-
concepts such as micro-membrane reactors or u- branes. E. L. Cussler explained the transport by
idized bed membrane reactors. chained carrier theory, when carriers are cova-
lently bonded, requiring a certain mobility of
polymeric chains and that a percolation threshold
appears when two chained carriers are too far
References apart.
R. Noble described the transport of neutral mol-
Buxbaum RE (2002) Patent US20026461408
Gallucci F, Van Sint Annaland M, Kuipers JAM ecules such as O2 (Noble 1990) or ions (Noble
(2010) Theoretical comparison of packed bed and u- 1991) across membranes in the terms dual-mode
idized bed membrane reactors for methane reforming. sorption model, assuming the presence of two dis-
Int J Hydrog Energy 35:71427150
Peters TA, Stange M, Klette H, Bredesen R (2008) High
tinct regions exhibiting gas solubility based on
pressure performance of thin Pd-23%Ag/stainless Henrys law and Langmuir sorption isotherm. The
steel composite membranes in water gas shift gas mix- model assumes also that permeation is diffusion-
tures; inuence of dilution, mass transfer and surface limited. This model shows the same behavior as
effects on the hydrogen ux. J Membr Sci
316(12):119127
facilitated transport in liquid membranes and does
Smit J, Bekink GJ, Van Sint Annaland M, Kuipers JAM not predict a percolation threshold as chained
(2005) A reverse ow catalytic membrane reactor for carrier model does. The model implies that the
Fixed-Site (Chained-Carrier) Membranes 777
effective diffusion coefcient of the solute to be Further Reading

transported depends on the morphology of mem- Baker RW (2004) Membrane technology and applications.
McGraw- Hill, New York
brane between two reactive sites (terrain).
Mulder M (2003) Basic principle of membrane technology.
Changes in polymer morphology cause a change Kluwer, London
in this effective diffusion coefcient.
The model presented by R. Noble is based on
the model presented by Barrer (1984) assuming
that a hole (microvoid) of the polymer free vol- Fixed-Site (Chained-Carrier)
ume is equivalent to a complexing site (a carrier). Membranes
Barrers dual-mode transport for glassy polymers
contains four diffusion coefcients: diffusion of a Marius Sandru
compound (A) through the polymer matrix, diffu- Polymer Particles and Surface Chemistry, F
sion of (A) from the polymer matrix to a hole in Materials and Chemistry, SINTEF, Trondheim,
the Langmuir region (excess of free volume), Norway
diffusion of (A) between two holes, and diffusion
of (A) from a hole to the polymer matrix. The
signicance of this analysis is that it predicts Carrier-facilitated transport processes in liquid
facilitated transport even when there is no diffu- membranes often achieve spectacular separations
sion or mobility of the complexing agent. between closely related species because of carrier
Both Nobles and Cusslers models have their selectivity. The instability of supported liquid
limitations. Dual sorption model presented by membranes (SLM) or emulsion type of mem-
Noble assumes that the concentration of the carrier branes (ELM) is a major technical challenge for
is in large excess and consequently constant. This their implementation at large industrial scale
implies that the model will start to deviate from (Mulder 2003; Baker 2004; Koros and Mahajan
experimental results at high loadings of the carrier 2000). One approach to stabilize the membranes
with the permeant (A) (carrier saturation). Also the is to covalently link the carrier complex to the
model does not take into account the reaction matrix polymer resulting in xed-site carrier
kinetic between the carrier and permeant (A). (chained) membranes.
The hopping mechanism proposed by Cussler The term chained carrier refers to transport
will strongly depend on the system permeant- mechanism theory rather than to position of the
membrane material characteristics even though it carrier on the membrane. E. L. Cussler (Cussler
is taking into account the chemical reaction et al. 1989) explained the facilitated transport by
between the carrier and component (A): high dif- a chained carrier theory that implies a certain
fusion coefcients of a compound (A) will allow mobility of polymeric chains and that a percola-
facilitation effect even at low concentration of the tion threshold appears when two chained carriers
carrier inrming a strict percolation threshold are too far apart. Chained carriers, by comparison
predicted by the model. to mobile carriers diffusing freely within a liquid
membrane, have limited diffusion because they
are covalently bound in the polymer membrane.
The solute is transported via a hopping (Tarzan)
References mechanism from one carrier to the next one
between the two interfaces, feed and permeate
Barrer RM (1984) Diffusivities in glassy polymers for the
dual mode sorption model. J Membr Sci 18:25 sides of the membrane.
Cussler EL, Aris R, Bhown A (1989) On the limits of In order to transport components via facilitated
facilitated diffusion. J Membr Sci 43:149164 transport, the polymeric chains must have a cer-
Noble RD (1990) Analysis of facilitated transport with
xed site carrier membranes. J Membr Sci 50:207214
tain mobility degree, and the concentration of the
Noble RD (1991) Analysis of ion transport with xed site carrier must be higher than the percolation thresh-
carrier membranes. J Membr Sci 56:229234 old (Cussler et al. 1989).
778 Flat Sheet Membrane Module
The hopping mechanism proposed by Cussler A recirculation cell containing various types of
will strongly depend on the system permeant- at-sheet membranes which were able to retain
membrane material characteristics even though it the suspended catalyst and partially selective to
is taking into account the chemical reaction the pollutant was studied by Molinari et al. (2001).
between the carrier and a component (A): high The membranes were NTR7410 and NTR7450
diffusion coefcients of a compound (A) will (Nitto Denko), N30F and NF-PES-010
allow facilitation effect even at low concentration (Hoechst), and MPCB0000R98 (SEPAREM).
of the carrier inrming a strict percolation thresh- The measured permeate ux was in the range
old predicted by the model. 530 l/h m2 at 4 bar, and all membranes showed
A particular case of chained carrier for CO2 both a rejection and a capacity to adsorb the pol-
transport is presented by KT.J. Kim et al. (2013). lutant with a transitory phase varying from 80 to
The carrier is represented by the amine groups 400 min at 4 bar. This behavior could be a benet
covalently bonded to the polymeric chain of for the process because oscillations in the pollut-
polyvinylamine. ant concentration were not transmitted in the per-
meate. Three factors, rejection, photocatalytic
degradation, and adsorption, were able to main-
References
tain the 4-nitrophenol (4NP) concentration in the
Baker RW (2004) Membrane technology and applications. permeate at very low values. For the continuous
McGraw- Hill, New York system, the lowest 4NP concentration in the per-
Cussler EL, Aris R, Bhown A (1989) On the limits of meate was 67 % (w/w) of the initial 4NP con-
facilitated diffusion. J Membr Sci 43:149164 centration (40 mg/l) after a transient period of
Kim TJ et al. (2013) Separation performance of PVAm
composite membrane for CO2 capture at different pH 300 min.
levels, J. of Mem. Sci 428:218224 In another method, the catalyst, TiO2 P25
Koros WJ, Mahajan R (2000) Pushing the limits on possi- Degussa, was immobilized by physical deposition
bilities for large scale gas separation: which strategies? on a at-sheet polymeric membrane, and 4NP was
Mulder M (2003) Basic principle of membrane technology. used as a model molecule to evaluate the reactor
Kluwer, Dordrecht performance (Molinari et al. 2000). A preliminary
investigation of the stability, under UV irradiation,
of some eligible polymeric membranes was car-
ried out by using scanning electron microscopy
Flat Sheet Membrane Module (SEM), optical microscopy (OM), determinations
of water permeation ux (WPF), and total
Cross-Flow Membrane Module organic carbon (TOC). These tests showed that
commercial membranes made of uoride + PP
(FS 50 PP-Dow), polysulfone + PP (GR 51
PP-TechSep), and polyacrylonitrile (PAN-
Flat Sheet Membrane Photocatalytic TechSep) seemed to be quite stable to UV light
Reactor over a 24 h period of irradiation. Immobilization
of TiO2 onto membranes by ultraltrating TiO2
Raffaele Molinari suspensions showed an optimal layer density
Dipartimento di Ingegneria per lAmbiente e il slightly >2.04 mg TiO2 per cm2 of membrane
Territorio e Ingegneria Chimica (DIATIC), surface area. Results obtained from membrane
Universit della Calabria, Rende, Italy reactor studies indicated that the observed initial
rate constants for the degradation of 4NP were
almost independent on the amount of TiO2
When membrane conguration is at type and it is employed over the range 0.764.08 mg/cm2.
coupled to a photocatalytic reaction, the reactor can A 50 % weight degradation of 4NP after 5 h of
be named at-sheet membrane photocatalytic reactor. irradiation in the presence of air was obtained. An
Flat Sheet Membrane Photocatalytic Reactor 779
almost complete degradation of 4NP, instead, was air rates, a single orice generates a higher aver-
observed in the presence of TiO2 suspended in the age enhancement than two orices, but the reverse
solution and pure oxygen. The permeate deriving was found at a relatively higher air rate. The
from the membrane photoreactor was clear, and enhancement in the center area of the module
4NP concentration was approximately equal to was relatively higher than that of the edge regions
that found in the retentate. The possibility of the when using a single orice, but more uniformity
continuous reuse of the photocatalyst and the con- was achieved with two.
tinuous separation of products from the reaction Exploration of two major commercialized at-
medium gave some advantages over traditional sheet and hollow-ber membranes in a submerged
approaches. membrane fungi reactor fed with a synthetic textile
A at-sheet membrane photocatalytic reactor wastewater revealed striking differences in the
operates, generally, with suspended solids (the extent and mechanism of fouling between the two F
catalyst) and air bubbling (oxygen feed), so, in types, indicating a case-specic scope of choice
the following, some cases of interest for operating between them for industrial wastewater treatment
such reactors are reported. (Hai et al. 2005). The hollow-ber membrane
Selection of a membrane conguration is a exhibited fouling with a cake layer composed of
crucial step in the design process and has a high fungi and starch, intensity being proportional to the
impact on further plant operations. Despite operating ux (0.050.3 m/d). Conversely, the at-
increasing experience with full-scale applications, sheet membrane suffered from immediate internal
practical knowledge concerning the impact of dif- pore blocking beyond a critical ux of 0.2 m3 m2
ferent membrane congurations on process per- d1. During the experiment with major constitu-
formance and operational costs is still lacking. ents of the synthetic wastewater separately, while
Full-scale performance data of a membrane bio- media containing only starch and only dye induced
reactor (MBR) comparing the use of at-sheet and negligible fouling, ux-dependent pore blocking
hollow-ber membranes and analyses of the con- was evident for both the hollow-ber (0.288 m3
sequences on operation, performance, and treat- m2 d1) and at-sheet membranes (1.3 m3 m2
ment efciency have been reported by d1) for the mixture of starch and dye. Despite that
Krzeminski et al. (2012). Hollow-ber congura- a remarkable 99 % color and 97 % TOC removal
tions, compared to the at sheet, are designed for were achieved by both membranes, fouling with
higher uxes, operated at lower concentrations, different modes and intensity for the two types
cleaned more often, and protected by stricter under similar conditions and for the same type of
pretreatment. Filterability of activated sludge membrane under different exposure conditions
from municipal MBRs was better than from warrants development of suitable modules for
industrial MBRs and did not depend on mem- such recalcitrant wastewater.
brane conguration. The energy consumption The oxygen transfer across a microporous, at-
depends more on the inuent type than on the sheet membrane with and without a nitrifying
membrane conguration. membrane-aerated biolm (MAB) was studied
An aerated at-sheet module focusing on the by Shanahan and Semmens (2006). Clean mem-
effect of bubble distribution on the spatial varia- brane oxygen transfer was quantied via average
tion of surface mass transfer with intermittent slug mass transfer coefcients and local uxes calcu-
bubbling was studied by Field et al. (2011). This lated from proles of dissolved oxygen
mode of operation will delay the onset of the (DO) along the membrane length. Stoichiometric
transmembrane pressure jump if biomass removal calculations and DO proles were used to
is more dependent upon maximum shear stress characterize oxygen transfer with a nitrifying
than mean shear stress. Two basic setups were MAB. Comparison of the local uxes with clean
considered: an orice in the middle of the aeration and biolm-coated membranes revealed that
tube, at the base, and two symmetrically placed oxygen transfer was decreased in upstream sec-
orices in the aeration tube. With relatively low tions of the membrane during cultivation of an
780 Flavanoids Separation
MAB due to reduced advection and/or turbulence malonyl-CoA groups (Bloor 2001). This conver-
near the membrane surface. By contrast, oxygen sion can lead to the formation of avones,
transfer was increased in downstream sections of isoavones, avonols, and anthocyanins. Flavo-
the membrane via bacterial respiration. Oxygen noid chemical structures are based on a C6-C3-C6
uxes generally decreased in the downstream skeleton that differs among each other by the
direction with biolm-coated membranes; how- saturation of heteroatomic ring C (Stobiecki and
ever, variability in biolm structure and membrane Kachilicki 2006). Further reactions as
coverage altered this trend on several occasions. glycosilation, acylation, and alkylation cause the
Furthermore, biolm structures were observed to formation of a huge variety of avonoids in the
form via sloughing and settling of biomass at the plant kingdom. The avonoids consist of six
membrane surface. Thus membrane conguration major subgroups: chalcone, avone, avonol, a-
may have a signicant inuence on oxygen transfer vanone, anthocyanins, and isoavonoids. Exam-
rates in membrane reactors. ples of some of these subgroups are shown in
Fig. 1. Together with carotenes, avanoids are
also responsible for the coloring of fruits, vegeta-
References bles, and herbs. For analysis purposes, there is
basically a avonoids division into glycosides,
Field RW, Zhang KS, Cui ZF, Hwang BK (2011) Flat sheet aglycones, and anthocyanins. Each of these
MBRs: analysis of TMP rise and surface mass transfer
types is quantied by different analytical
coefcient. Desalination Water Treat 35:8291
Hai FI, Yamamoto K, Fukushi K (2005) Different fouling methods. The most important dietary sources are
modes of submerged hollow-ber and at-sheet mem- fruits, tea, and soybean. Green and black tea con-
branes induced by high strength wastewater with contains about 25 % avonoids. Other important
current biofouling. Desalination 180:8997
sources of avonoids are apple (quercetin) and
Krzeminski P, Gil JA, van Nieuwenhuijzen AF, van der
Graaf JHJM, van Lier JB (2012) Flat sheet or hollow citrus fruits (rutin and hesperidin). Flavonoids
bre comparison of full-scale membrane bio-reactor are becoming very popular because they have
congurations. Desalination Water Treat 42:100106 many health-promoting effects. Some of the activ-
Molinari R, Mungari M, Drioli E, Di Paola A, Loddo V,
ities attributed to avonoids include antiallergic,
Palmisano L, Schiavello M (2000) Study on a
photocatalytic membrane reactor for water purication. anticancer, antioxidant, anti-inammatory, and
Catal Today 55:7178 antiviral activities.
Molinari R, Grande C, Drioli E, Palmisano L, Schiavello Essential to the study of avonoids is hav-
M (2001) Photocatalytic membrane reactors for degra-
ing the means available for their separation
dation of organic pollutants in water. Catal Today
67:273279 (analytical and preparative) and isolation.
Shanahan JW, Semmens MJ (2006) Inuence of a nitrify- Thin-layer chromatography is a rapid, simple,
ing biolm on local oxygen uxes across a micro- and versatile method for fractionation avo-
porous at sheet membrane. J Membr Sci 277:6574
noids. However, this method has been
substituted to qualitative and quantitative
applications of high-performance liquid chro-
matography (HPLC). In HPLC analysis, avo-
Flavanoids Separation noids can be separated, quantied, and
identied in only one operation by coupling
Beatriz Mello HPLC to ultraviolet (UV), mass, or nuclear
UFSCar, Federal University of Sao Carlos, Buri, magnetic resonance (NMR) detectors. Another
SP, Brazil technique that has been gaining attention is
capillary electrophoresis (CE). All the methods
use the effects of the presence of the avonoid
Flavonoids are a water-soluble class of polyphe- phenyl ring which is an excellent chromophore
nols derived from plant secondary metabolites and is responsible for the easy detection of
through the condensation of cinnamic acid with avonoids. Their UV spectra provide enough
Flexibility in Gas Separation 781
Flavanoids Separation,
Fig. 1 Examples of some
avonoids. (a)
quercetin avonol, (b)
cyanidin anthocyanin, (c)
hesperidin avanone
information to distinguish the type of phenol References

(Marston and Hostettmann 2006).
Flavonoids can be also separated from each Bloor SJ (2001) Overview of methods for analysis and
identication of avonoids. In: Packer L (ed) Methods
other by the use of a polymeric membrane.
in enzymology avonoids and other polyphenols.
A relatively high concentration of polypheno- Academic, San Diego, pp 314
lics and avonoids present in the juices were Kumar A (2006) Membrane separation technology in
fractionated by various methods including processing bioactive components. In: Functional food
ingredients and nutraceuticals. CRC Press, Boca Raton,
ultraltration to get a retentate that showed
pp 193208
very high oxygen radical absorbance capacity Marston A, Hostettmann K (2006) Separation and quanti-
(ORAC) even at 50 % recovery of thin juice cation of avonoids. In: Andersen OM, Markham KR
(Kumar 2006). Membrane-based processes are (eds) Flavonoids: chemistry, biochemistry and applica-
tions. CRC Press, Boca Raton, pp 132
usually more energy efcient than distillation,
Stobiecki M, Kachilicki P (2006) Isolation and identica-
adsorption, and chromatography. Furthermore, tion of avonoids. In: Grotewold E (ed) The science of
membranes have the advantage of compatibil- avonoids. Springer, New York, pp 4770
ity with a wide range of solvents and chemical
products, ability to process thermally sensitive
compounds, easy amenability to automation,
smaller footprint, and seamless scale-up. Flexibility in Gas Separation
These advantages open up several possibilities
of membrane application in the production of Adele Brunetti
bioactive compounds through preconcentration Institute on Membrane Technology, National
of dilute solutions, fractionation of a complex Research Council of Italy, ITM-CNR, Rende,
mixture, recovery of intermediates, and Italy
recycling of solvents. However, it is also
important to be aware of the membrane-based
separation drawbacks, such as limited selectiv- The exibility (also called easy of expansion) is
ity, fouling leading to performance decline, the ability to operate under variable feed quality
cleaning, and higher capital costs in certain conditions, either on a short-term or long-term
installations. basis (Miller and Stcker 1989; Brunetti
782 Fluidized Bed Membrane Reactors
et al. 2010). The changes in feed composition The purity can be maintained constant by a simple
occur very often in renery applications, particu- cycle time adjustment. The response time to var-
larly when the source of the feed is a catalytic iations is rapid but not abrupt, generally requiring
process or when the feedstock to the upstream 515 min for responding to a step change in feed
unit changes as well as in ue gas streams. quality. The new steady-state upon restart follow-
In membrane processes, the increase in feed ing shutdown is reached in about 1 h.
impurity concentrations tend to cause a decrease The cryogenic process has very low exibility
in product purity, which, however, can be because changes in the concentration of the lower
maintained for small feed composition changes boiling components of the feed affect the product
by adjusting the feed-to-permeate pressure ratio. purity directly. Recovery is not strongly affected.
In most renery membrane applications, however, Response time is not as rapid as for PSA or mem-
the major product impurity is methane and this brane systems. Start-up is 824 h depending on
can be allowed to increase slightly in the product the procedure used.
without major downstream impact. The ue gas
stream contains an amount of CO2 ranging from
5 % to 35 % depending on the process, the rest References
being mainly nitrogen and few percentage of oxy-
gen. Whether the CO2 concentration is always Brunetti A, Bernardo P, Drioli E, Barbieri G (2010) Mem-
brane engineering progresses and potentialities in gas
higher than 20 %, the membrane systems can be
separations. In: Yampolskii Y, Freeman B (eds) Mem-
considered highly exible. On the contrary, for brane gas separation. Wiley, New York, pp 281312
CO2 concentration lower than 20 % the membrane Miller GQ, Stcker J (1989) Selection of a hydrogen sep-
operation offers a very low exibility. A further aration process. In: NPRA annual meeting, 1921 Mar,
San Francisco
reduction in CO2 concentration, in fact, produces
a reduction of the driving force with a consequent
decrease in CO2 recovery and in product purity.
The response time of membrane systems is
essentially instantaneous and corrective action Fluidized Bed Membrane Reactors
has immediate results. The start-up time required
by the process is extremely short. Fausto Gallucci
The exibility of the absorption process is Chemical Process Intensication, Department of
moderate. This system, in fact, responds quite Chemical Engineering and Chemistry,
well to the changes in the feed composition; how- Eindhoven University of Technology, Eindhoven,
ever, signicant differences in the fraction of the The Netherlands
desired specie contained in the feed imply higher
solvent ow rate that is strictly related to the size
of equipment apart from the absorber. Further- A uidized bed membrane reactor, also indicated
more, the fact that the feed stream contains oxi- as membrane-assisted uidized bed (MAFBR), is
dizing compounds (oxygen or sulfur dioxide) an integrated reactor where membranes (in form
induces major problems of amine deactivation of tubes, hollow bers, or planar modules) are
and promotes corrosion. immersed in a uidized bed of catalyst particles.
The PSA process shows a great ability to main- Typical example of MAFBR for hydrogen pro-
tain purity and recovery under changing condi- duction is reported in (Gallucci et al. 2010a) and
tions. The process is self-compensating and even consists in a bundle of hydrogen-selective mem-
relatively large changes in feed impurity concen- branes immersed in a catalytic bed operated in
trations have little impact on performance. As the bubbling regime. The use of uidized bed mem-
concentration of a feed impurity increases, its brane reactors makes possible the reduction of
partial pressure increases, increasing also the bed-to-wall mass transfer limitations, but also
amount of the impurity which will be adsorbed. allows operating the reactor at a virtually
Fluorescence 783
isothermal condition (due to the movement of production through Pd-based membranes (see
catalyst). This possibility can be used for operat- among others Adris et al. 1991; Prasad and
ing the autothermal reforming of hydrocarbons Elnashaie 2002) and on autothermal reforming
inside the membrane reactor. In fact, as indicated reactions (see a.o. Grace et al. 2001; Gallucci
by Tiemersma et al. (2006), the autothermal et al. 2010b).
reforming of methane in a packed bed membrane
reactor is quite difcult due to the hot spot at the
reactor inlet which can melt down the membrane. References
This problem is completely circumvented in u-
idized bed membrane reactors. In this case both Adris AM, Elnashaie SSEH, Hughes R (1991) A uidized
bed reactor for steam reforming of methane. Can
autothermal reforming and hydrogen recovery
J Chem Eng 69:1061
can be performed in a single reactor. Deshmukh SARK, Heinrich S, Mrl L, van Sint F
According to Deshmukh et al. (2007) the main Annaland M, Kuipers JAM (2007) Membrane assisted
advantages of the MAFBR can be listed as uidized bed reactors: potentials and hurdles. Chem
Eng Sci 62:416436
follows:
Gallucci F, Van Sint Annaland M, Kuipers JAM (2010a)
Theoretical comparison of packed bed and uidized
Negligible pressure drop, which allows using bed membrane reactors for methane reforming. Int
small particle sizes resulting in limited internal J Hydrog Energy 35:71427150
Gallucci F, Van Sint Annaland M, Kuipers JAM (2010b)
mass and heat transfer limitations (higher
Pure hydrogen production via autothermal reforming of
effectiveness factors) ethanol in a uidized bed membrane reactor. Int
(Virtual) Isothermal conditions J Hydrog Energy 35:16591668
Flexibility in membrane and heat transfer sur- Grace JR, Li X, Lim CJ (2001) Equilibrium modelling of
catalytic steam reforming of methane in membrane
face area and arrangement of the membrane
reactors with oxygen addition. Catal Today
bundles 64:141149
Improved uidization behavior as a result of: Prasad P, Elnashaie SSEH (2002) Novel circulating
Compartmentalization, i.e., reduced axial uidized-bed membrane reformer for the efcient pro-
duction of ultraclean fuels from hydrocarbons. Ind Eng
gas back-mixing.
Chem Res 41:65186527
Rahimpour MR, Lotnejad M (2008) Co-current and
Reduced average bubble size due to enhanced countercurrent congurations for a membrane dual
bubble breakage, resulting in improved bubble to type methanol reactor. Chem Eng Technol 31(1):3857
Tiemersma TP, Patil CS, Sint Annaland MV, Kuipers JAM
emulsion mass transfer.
(2006) Modelling of packed bed membrane reactors for
MAFBR for hydrogen production also circum- autothermal production of ultrapure hydrogen. Chem
vent the bed-to-wall mass transfer limitations Eng Sci 61(5):16021616
affecting packed bed membrane reactors. It has
been demonstrated that higher membrane uxes
(possible with nowadays membranes) will cause a
signicant increase of the concentration polariza- Fluorescence
tion in packed bed membrane reactors, even in
membrane tubes with a diameter as small as 1 cm. Carla A. M. Portugal
By using the MAFBR, the membrane area REQUIMTE-LAQV, Chemistry Department,
required for a xed hydrogen ux can be FCT-Universidade Nova de Lisboa, Campus da
decreased by 50 % if compared with packed bed Caparica, Caparica, Portugal
reactors (Gallucci et al. 2010a).
Even though Rahimpour and coworkers often
used uidized bed membrane reactor for distribu- Fluorescence is a photoluminescence process
tive hydrogen feeding in methanol reactors (see which consists of the emission of photons along
Rahimpour and Lotnejad 2008), most of the the de-excitation of molecular species from the
literature has focused on pure hydrogen excited state S1 to the ground state S0. The loss
784 Fluorescence
Singlet state
S2 IC
Vibrational
relaxation
Triplet state
Intersystem crossing
S1
T1
Conversion (IC)
Phosphorescence
Absorbance
Fluorescence
fluorescence
Delayed
Internal
S0
Fluorescence, Fig. 1 Perrin-Jablonski diagram schematizing the radiative (bold lines) and non-radiative (dashed lines)
de-excitation processes
of energy by molecular species in the excited state collision, or electron, proton, and energy transfer.
may occur through radiative processes (i.e., with Also, it may involve excited state chemical reac-
photon emission) or via nonradiative (i.e., without tions, leading to the formation of different molec-
photon emission) de-excitation mechanisms, as ular species (excimers or exciplexes). These
schematized in Fig. 1. intermolecular de-excitation mechanisms are
Radiative de-excitation may occur via uores- responsible for uorescence quenching, which is
cence or phosphorescence emission, whereas characterized by a fast electronic decay to the
nonradiative de-excitation may occur through ground state and therefore it may compete
intersystem crossing or internal conversion mech- with uorescence, leading to a decrease of the
anisms, which may be coupled to additional uorescence quantum yield, fF. Hence, fF
energy losses through vibrational relaxation to expresses the ability of an excited molecule to
the lowest energy level. The former corresponds decay to the ground state through uorescence
to electronic transitions between states with the radiative process and can be determined by the
same multiplicity whereas intersystem conversion equation
implies an electronic transition between
isoenergetic vibrational levels of different multi- kr
fF
plicity (singlet to triplet states) (Valeur 2002). kr knr
Electronic transitions from triplet state to the
ground state are possible through radiative pro- where kr and knr are the rate constants for radiative
cesses, such as phosphorescence or delayed uo- and nonradiative deactivation processes in the
rescence (upon returning to the excited singlet singlet state (Valeur 2002), respectively.
state). Fluorescence emission is highly sensitive to
Nonradiative de-excitation results from inter- the intrinsic physicochemical characteristics of
actions between the excited species and other uorophores and also to changes of its involving
molecules in their close proximity media. Indeed, steady-state uorescence emission
(intermolecular processes) where energy loss is spectra acquired upon excitation with randomly
consequence of heat release due to molecular oriented light allows at eliciting information about
Fluorescence 785
the excited state molecular events occurring in protein chain may report protein folding/
equilibrium, based on changes in uorescence unfolding processes or changes of protein molec-
intensity or maximum emission shifts (Portugal ular exibility, e.g., caused by protein adsorption/
et al. 2006, 2007; Martinho et al. 2003). The desorption mechanisms (Portugal et al. 2006,
former may be related with the variation of con- 2007).
centration of uorophores present in the sample The emission uorescence spectra obtained in
but also it may be ascribed to other molecular steady state correspond to the contribution of all
events such as uorescence quenching, inner lter the uorescing species present in the system under
effects (e.g., light scattering or reabsorption of analysis that are able to absorb energy at the
emitted light), and chemical reactions with the selected excitation wavelength. The ability to dis-
formation of different species. Instead, the pres- criminate the contribution of each uorescing spe-
ence of maximum emission shifts to lower (blue- cies depends largely on the system complexity, F
shift) or higher (red-shift) wavelengths may evi- but it may be possible through determination of
dence changes in the polarity in the vicinity of the their individual uorescence decay times
uorophores. Applied to intrinsic uorophores of (Beechem and Ludwig 1985). Fluorescence
proteins, e.g., tryptophans, these spectral changes decay times may be elicited by tting of
may reveal protein folding/unfolding processes adequate functions, e.g., sum of exponentials
(Portugal et al. 2006, 2007). (where each parcel corresponds to the
Fluorescence anisotropy corresponds to the contribution of the single uorescing species)
fraction of polarized light emitted by a uorophore described below (Beechem and Ludwig 1985;
upon excitation with polarized light, as expressed Valeur 2002).
by the following equations (Valeur 2002):
X t =
I l, t ai e t i
Polarized light Ijj I IVV IVH G i
r
Total light Ijj 2I IVV 2 IVH G
where I (l, t) is the variation of uorescence
IHV emission along the time, ai corresponds to the
G amplitude of the decay, t and ti are the measuring
IHH
time and the uorescence decay time,
where I represents the fraction of emitted parallel respectively.
polarized light (IVV and IHH) and I corresponds Due to the high sensitivity of uorescence
to the fraction of emitted perpendicular light (IVH emission and uorescence decay times to changes
and IHV). of temperature, pH, ionic strength, and solvent
Fluorescence anisotropy relates with the depo- properties (e.g., polarity, dielectric constant) and
larization of light in the excited state due to the the presence of uorescence quenchers in the
rotational freedom of the uorophores. Therefore, surrounding media, some uorophores have
uorescence anisotropy may provide information been efciently used as uorescence probes for
about constraints imposed on the uorophore by monitoring of membrane-based processes. Fluo-
the system, molecule, or matrix where it is rescence spectroscopy has been used toward the
included in, through detection of changes of the development of membrane materials less prone to
uorophore rotational ability (Lakowicz fouling, through inspection of uorophores
et al. 1983). Hence, the increase of uorescence adsorbed at the material surface (Girons
anisotropy may be ascribed to a decrease of et al. 2006). Also, it can be used for monitoring
uorophore mobility. Recalling the example of the performance of membrane permeation pro-
tryptophan, changes of its mobility within the cesses (Crespo et al. 1999; Galinha et al. 2012).
786 Fluorescence Excitation-Emission Matrix (EEM)
Furthermore, it has been used to obtain a better

comprehension about the impact of membrane Fluorescence Excitation-Emission
processes on the structural and functional proper- Matrix (EEM)
ties of biomolecules, e.g., proteins, using trypto-
phan as the structural probe. Using a uorescence Carla A. M. Portugal1 and Claudia F. Galinha2
1
monitoring approach which combined the REQUIMTE-LAQV, Chemistry Department,
complementary information provided by steady- FCT-Universidade Nova de Lisboa, Campus da
state uorescence, uorescence anisotropy, and Caparica, Caparica, Portugal
time-resolved uorescence it was demonstrated 2
REQUIMTE-CQFB, Chemistry Dept. Faculdade
that under specic processing conditions (e.g., de Cincias e Tecnologia, FCT-Universidade
transmembrane pressure, membrane molecular Nova de Lisboa, Campus da Caparica, Caparica,
weight cutoff, membrane chemistry) membrane Portugal
permeation may induce irreversible changes of
protein structure and function (Portugal
et al. 2006). The uorescence emission of molecular species
depends on their intrinsic physicochemical charac-
teristics and on their interactions with the surround-
ing media. Therefore, uorescence emission is
References extremely sensitive to changes of the environmen-
tal conditions in the uorophores vicinity (e.g.,
Beechem JM, Ludwig B (1985) Time resolved uores- pH, temperature, presence of uorescence
cence of proteins. Annu Rev Biochem 54:4371
Crespo JPSG, Trotin M, Hough D, Howell JA (1999) Use
quenchers, solvent properties). Together, these fac-
of uorescence labelling to monitor protein fraction- tors dene the period during which a uorophore
ation by ultraltration under controlled permeate ux. remains in the excited state, i.e., uorescence life-
J Membr Sci 155:209230 time, and the magnitude of each radiative and
Galinha CF, Carvalho G, Portugal CAM, Guglielmi G,
non-radiative de-excitation process component,
Reis MAM, Crespo JG (2012) Multivariate
statistically- based modelling of a membrane bioreactor through which the uorophore decays to the
for wastewater treatment using 2D uorescence moni- ground state (the lowest energy level).
toring data. Water Res 46:36233636 This dependency relation limits the use of uo-
Girons M, Bolhuis-Versteeg LAM, Lammertink RGH,
rescence spectroscopy for quantitative purposes,
Wessling M (2006) Flux stabilization of silicon
nitride microsieves by backpulsing and surface but simultaneously it explains the potential of this
modication with PEG moieties. J Membr Sci technique to provide information about the system
299:831840 status using single or multiple uorescence probes
Lakowicz JR, Maliwal BP, Cherek H, Balter A (1983)
(uorophores). The information about the molecu-
Rotational freedom of tryptophan residues in proteins
and peptides. Biochemistry 22:17411752 lar events occurring in the system may be obtained
Martinho JMG, Santos AM, Fedorov A, Baptista RP, Taipa based on changes of the uorescence emission of a
MA, Cabral JMS (2003) Fluorescence of a single tryp- single probe, through acquisition of the uores-
tophan cutinase: temperature and pH effect on the pro-
cence signal at only one pair of excitation-emission
tein conformation and dynamics. Photochem Photobiol
78:1522 wavelengths, generally corresponding to the spe-
Portugal CAM, Lima JC, Crespo JG (2006) Probing the cic maximum emission of the uorophore. Alter-
change of enzymatic activity of horseradish peroxidase natively, it may be elicited by selecting one of the
induced by membrane permeation using tryptophan
uorescence. J Membr Sci 284:180192
uorescence signal dimensions (excitation or emis-
Portugal CAM, Lima JC, Crespo JG (2007) Monitoring the sion wavelength) and scanning the other within the
structural alterations induced in b-lactoglobulin range of interest. Despite sensitive, this approach
during ultraltration: learning from chemical and may hamper a complete system description since it
thermal denaturation phenomena. J Membr Sci
reports only the media interactions with a single
300:211223
Valeur B (2002) Molecular uorescence principles and probe, discarding possible additional information
applications. Wiley-VCH, Weinheim captured by other uorescence probes also present
Fluorescence Excitation-Emission Matrix (EEM) 787
a 600
b 600
0 0
550 20 550 20
40 40
60 60
500 80 500 80
100 100
lEm (nm)
450 450
400 400
350 350
300 300
F
250 250
250 300 350 400 450 500 550 600 250 300 350 400 450 500 550 600
lExc (nm) lExc (nm)
c 600
0
550 20
40
60
500 80
100
lEm (nm)
450
400
350
300
250
250 300 350 400 450 500 550 600
lExc (nm)
Fluorescence Excitation-Emission Matrix (EEM), used for wastewater treatment: (a) feed wastewater, (b)
Fig. 1 Fluorescence excitation-emission matrices activated sludge inside the bioreactor and (c) claried
(EEMs) obtained at different locations of a MBR system permeate upon treatment
in the system, emitting in different spectral regions. concerning the presence of the uorescence
Thus, a more thorough system characterization can probes in the system (Marose et al. 1998) as well
be achieved by scanning simultaneously excitation as their interactions with their involving media
and emission wavelengths in order to cover a larger (Galinha et al. 2011). EEMs are thus a sum of
spectral region and capture the information pro- constructive and destructive interferences of the
vided by each uorophore present in the system. media conditions with the probe intrinsic uores-
This 2D uorescence monitoring approach results cence. Constructive interferences are those that
in tridimensional uorescence maps having as contribute to the increase of the uorescence
coordinates the uorescence emission intensity, emission of a specic probe, thus facilitating its
excitation (lExc) and emission (lEm) wavelengths, detection, such as those resulting from decrease of
so-called excitation-emission matrices (EEM), as temperature or solvent polarity. Oppositely,
those illustrated in Fig. 1. destructive interferences are those that limit the
Therefore, EEMs may be regarded as system detection of specic uorescence response.
ngerprints encoding valuable information Destructive interferences are quite common in
788 Fluorinated Polymers
complex systems being caused by a large diversity anticipation of process problems without the need
of phenomena, such as (1) superimposition of of the traditional time-consuming analytical
emission spectra from different uorescent sol- methodologies.
utes; (2) quenching effects, prompted by molecu-
lar species in close proximity to the probe, which
interfere in the uorophore de-excitation process References
reducing its uorescence emission component;
and (3) inner lter effects, including the Galinha CF, Carvalho G, Portugal CAM, Guglielmi G,
Reis MAM, Crespo JG (2011) Two-dimensional uo-
re-absorbance of the emitted light by other solutes
rescence as a ngerprinting tool for monitoring waste-
(e.g., when the absorption spectra of solutes water treatment systems. J Chem Technol Biotechnol
superimpose, at least partially, the uorophore 86:985992
emission spectra) and light scattering promoted, Galinha CF, Carvalho G, Portugal CAM, Guglielmi G,
Reis MAM, Crespo JG (2012) Multivariate
for instance, by the presence of suspended solids
statistically-based modelling of a membrane bioreactor
or turbidity. for wastewater treatment using 2D uorescence moni-
Although it perturbs the direct measurement of toring data. Water Res 46:36233636
the uorophores (Galinha et al. 2011; Kobbero Kobbero C, Keiding K, Larsen KL, Nielsen PH (2008)
Quenching effects in the application of multi-channel
et al. 2008) and the access to convoluted data
uorescence in activated sludge suspended solids.
embedded in EEMs, interferences should be Water Res 42:24492456
regarded as an additional rich source of information Marose S, Lindemann C, Scheper T (1998) Two-Dimen-
that can be disclosed by the use of adequate math- sional Fluorescence Spectroscopy: A New Tool for On-
Line Bioprocess Monitoring, Biotechnol Prog
ematic tools. Such approach has been applied for
14:6374
the characterization of different water streams Osburn CL, Handsel LT, Mikan MP, Paerl HW,
(Galinha et al. 2011; Osburn et al. 2012) and for Montgomery MT (2012) Fluorescence tracking of
the monitoring of biological and wastewater treat- dissolved and particulate organic matter quality in a
river-dominated estuary. Environ Sci Technol
ment processes, namely, membrane bioreactors,
2012(46):86288636
MBRs (Fig. 1). Fluorescence monitoring has been Peiris RH, Budman H, Moresoli C, Legge RL
used aiming to get a better understanding of mem- (2010) Understanding fouling behaviour of
brane fouling mechanisms, in order to dene ade- ultraltration membrane processes and natural
water using principal component analysis of uores-
quate strategies, which may allow for the mitigation
cence excitation-emission matrices. J Membr Sci
of fouling problems (Peiris et al. 2010). Mathemat- 357:6272
ical approaches based on the use of multivariate Wolf G, Almeida JS, Pinheiro C, Correia V,
statistical tools, combining principal component Rodrigues C, Reis MAM, Crespo JG (2001) -
Two-dimensional uorometry coupled with articial
analysis (PCA or PARAFAC) and projection to
neural networks: a novel method for on-line moni-
latent structure (PLS) modeling (Galinha toring of complex biological processes. Biotechnol
et al. 2012) or articial neural networks (ANNs) Bioeng 72:297306
(Wolf et al. 2001), allow for establishing relation-
ships between EEMs and relevant process
output variables, leading to an accurate prediction
of some process performance descriptors (e.g.,
transmembrane pressure, chemical oxygen Fluorinated Polymers
demand, total nitrogen, nitrites and nitrates, reaction
rates). Amorphous Fluoropolymers
2D uorescence spectroscopy offers the possi-
bility of simple, real-time, and non-invasive mon-
itoring of biological or non-biological processes
at different points of the system by using optical Fluoroplastics
bers. Moreover, it allows accurate prediction of
process performance parameters and the potential Amorphous Fluoropolymers
Fluxes and Driving Forces in Membrane Separation Processes 789
The driving forces in membrane processes are

Fluxes and Driving Forces gradients in the chemical potential, the electrical
in Membrane Separation Processes potential, and the hydrostatic pressure, which
could result in a diffusion of individual molecules,
Xianhui Li and Jianxin Li a migration of ions, and a convection of mass,
State Key Laboratory of Separation Membranes respectively. The function of the membrane is
and Membrane Processes, School of Materials illustrated in Fig. 1. The transport of the compo-
Science and Engineering, Tianjin Polytechnic nent A from phase 1 through a membrane into the
University, Tianjin, China phase 2 is due to the driving force as shown in
Fig. 1.
According to the driving forces, the membrane
The transport rate of a component through a processes can be classied into the pressure- F
membrane is determined by the permeability driven processes (reverse osmosis, ultra- and
of the membrane and by the driving force. Gener- microltration, and gas separation), the concen-
ally, the ux through a membrane can be tration gradient-driven processes (dialysis), par-
described by tial pressure-driven processes (pervaporation),
and electrical potential-driven processes
dX (electrolysis and electrodialysis). The main driv-
J P (1) ing forces in the different membrane processes are
dz
illustrated in Table 1. It can be seen in Table 1 that
where J is a ux, P is a phenomenological coef- the hydrostatic pressure differences are used in
cient expressing the permeability of the mem- micro- and ultraltration, as well as in reverse
brane, and dX dz is the driving force. The ux osmosis and gas separation as driving force for
through the membrane can be dened as volume the mass transport through the membrane. The
ux (Jv) expressed in volume per time (m.s1), mode of transport in micro- and ultraltration is
mass ux (Jm) expressed in mass per time (kg. convection. However, in reverse osmosis and gas
m2.s1), molar ux (Jn) expressed in mole per separation, the mode of transport is diffusion, and
time (mol.m2.s1), electrical ux (Je) expressed the actual driving forces are chemical potential
in Faraday per time (A.m2), and heat ux (Jq) and fugacity gradients. The mode of transport in
expressed in heat per time (J.m2.s1). dialysis is diffusion with concentration or activity
Fluxes and Driving membrane

Forces in Membrane
Separation Processes,
Fig. 1 The schematic phase 1 phase 2
diagram of mass transport XA ' XA "
though a membrane
driving forces
X A' > XA"
mass transport
790 Food Packaging
Fluxes and Driving Forces in Membrane Separation coefcient, Lp is the hydrodynamic permeability,
Processes, Table 1 The driving forces applied in the and k is the electric conductivity.
various membrane separation processes (Strathmann 2001)
Fouriers law describes the relation between
Transport heat transport and a temperature gradient. Ficks
Process Driving force mode
law describes the relation between the ux of
Microltration Hydrostatic pressure (Dp) Convection
Ultraltration Hydrostatic pressure (Dp) Convection individual components and a concentration gradi-
Reverse Hydrostatic pressure (Dp) Diffusion ent. Ohms law describes the relation between an
osmosis Chemical potential (Dmi) electrical current and an electrical potential gradi-
Dialysis Concentration (DC) Diffusion ent, and Darcys law describes the relation
Activity (Da) between a volume ux and a hydrostatic pressure
Gas separation Hydrostatic pressure (Dp) Diffusion difference. These two factors should be
Fugacity (Dfi) interlinked: a high driving force yields the high
Pervaporation Partial pressure (Dpi) Diffusion
ux and a high rate of foulants deposition on the
Fugacity (Dfi)
membrane surface. Therefore, driving forces and
Electrodialysis Electrical potential (Dj) Migration
uxes are interdependent in membrane processes.
gradients as driving forces. And an electrical References

potential gradient in electrodialysis is used to
achieve a migration of charged ions across the Strathmann H (2001) Membrane separation processes:
membrane. current relevance and future opportunities. AICHE
J 47:10771087
In order to describe the transport processes
through membranes, a number of equations
which are related to the uxes of various compo-
nents to the corresponding driving forces in the Food Packaging
form of linear relations are developed as follows:
Fouriers law Food Packaging, Membranes for
dT
Jq a (2)
dz
Food Packaging, Membranes for
Ficks law
Alberto Figoli
dC
Jn D (3) Institute on Membrane Technology, National
dz Research Council of Italy, ITM-CNR, Rende,
Darcys law Italy
dp
Jv Lp (4) Synonyms
dz
Ohms law Food packaging; Membrane operations
d Food packaging is often dened as a heteroge-

Je k (5) neous group of boxes, envelops, papers, and coat-
dz
ings that are employed in order to increase
Herein T is the temperature, C is the concentra- foodstuffs shelf life (European Council 1994).
tion, p is the pressure, j is the electrical potential, This means that any food packaging system
a is the heat transfer coefcient, D is the diffusion aims to avoid/slow down bacterial/fungal
Food Packaging, Membranes for 791
Food Packaging,
Membranes for,
Fig. 1 Application of
membrane in traditional
food packaging, revised
from ref. (Figoli et al. 2010)
contamination and deterioration and, at the same oxygen and carbon dioxide. By varying the ratio
time, to ensure the maintenance of properties as between polymer and additive concentrations,
rmness, color, and avor as long as possible. dense membranes in modied polyether ether F
Nowadays, beside the traditional concept of ketone (PEEK-WC) and poly-alpha-pinene
packaging as impenetrable barrier, novel systems (PaP) were reported to match the values required
are required, such as modied atmosphere pack- by different products, as apple, Brussels sprouts,
aging (MAP) and active packaging (AP). Poly- carrot, green pepper, and turnip (Torchia
meric membranes, thanks to their unique et al. 2004).
properties of permselectivity and ability to sustain A wide range of CO2/O2 permselectivity
controlled delivery, can be successfully used for values can be obtained by using packaging struc-
food packaging; indeed, their properties can be tures based on biopolymers. These are dened as
exploited for such innovative packaging systems biodegradable materials that can be decomposed
(Figoli et al. 2010), as also shown in Fig. 1. Fur- by microorganisms and enzymatic degradation.
thermore, the expertise of membranologists could Starch, cellulose, chitosan, polycaprolactone
help to produce food packaging systems from (PCL), and polylactic acid (PLA) are some exam-
biodegradable nonpolluting sources. In particular, ples of biopolymers suitable for food packaging
the properties of lms produced from biopolymers (Mensitieri et al. 2011). In literature, several stud-
can be improved by well-developed techniques as ies already demonstrated that biopolymers can
blending, use of plasticizers, and loading with compete with oil-derived plastics for packaging
nanoparticles. of different fruit and vegetable species (see for
MAP is frequently used in the case of fresh, or instance ref. Conte et al. 2009; Kantola and
minimally processed, respiring, products, like Helen 2001; Muratore et al. 2005; Srinivasa
vegetables and fruits. MAP aims at maintaining et al. 2002). However, hydrophilicity and poor
an optimum gas balance in the package head- mechanical properties are the main drawbacks
space. In particular, a selected ratio between car- for use of biopolymers as food packaging sys-
bon dioxide and oxygen is often required in order tems. Several strategies, already developed in the
to allow living vegetal tissues respiration and eld of membrane science and technology, can be
survival, at the same time preventing aging useful to improve properties and processability of
and deterioration. The respiratory quotient biopolymers for food packaging, such as use of
(RQ) (Cameron et al. 1995) is dened as the chemical additives (plasticizers, antioxidants, sta-
optimum balance between O2 and CO2 that should bilizers) (Torchia et al. 2004; Abdorreza
be maintained within the package; this value is et al. 2011), blending (Torchia et al. 2004; John
peculiar of each fruit or vegetable specimen. MAP et al. 1998; Sarazin et al. 2008), chemical cross-
requires lms having specic permselectivity to linking (Gonzlez et al. 2011; Yu et al. 2011), or
gases, such as carbon dioxide and oxygen, rather use of nanollers (Belbekhouche et al. 2011;
than impenetrable materials. In this case, gas sep- Chivrac et al. 2009; Fam et al. 2011; Hu
aration membranes can be particularly useful, et al. 2011; Li et al. 2011).
while membrane materials can be selected, in Active releasing packaging systems are used to
each case, on the basis of their permselectivity to control the delivery of specic substances to the
792 Food Packaging, Membranes for
packed product. Membranes loaded with specic Conte A, Scrocco C, Brescia I, Del Nobile MA (2009) Dif-
substances, coated on the surface, or entrapped in ferent packaging strategies for fresh-cut barattiere
melon cultivar (cucumis melo L.). Int J Food Sci
the polymeric matrix can be useful in this case, Technol 44:14221428
similar to what already developed in the medical European Council (1994) Directive 94/62/EC on packag-
eld (Baker 2002). Membranes can be used to ing and packaging waste. Off J Eur Union L365:1023
sustain the controlled delivery of antimicrobial Fam LM, Pettarin V, Goyanes SN, Bernal CR (2011) -
Starch/multi-walled carbon nanotubes composites with
substances, such as allyl isothiocyanate (AITC) improved mechanical properties. Carbohydr Polym
(Lee et al. 2008) or oxalic acid (Figoli et al. 2004). 83:12261231
Active release can be also combined with use of Figoli A, Drioli E, Jansen JJM, Wessling M (2004) Film
biopolymers. Films based on starch, corn zein, antimicrobico per prolungare la shelf life, ISSN0033-
9687, Rassegna dellImballagio, Nov. 2004, N 16,
and PLA can be used for delivery of lysozyme, Year 25th
nisin, and natural antimicrobial compounds Figoli A, Mascheroni E, Limbo S, Drioli E (2010) Mem-
(Torchia et al. 2004); literature reports several branes for food packaging. In: Peinemann K-V, Pereira
other interesting examples of biopolymer-based Nunes S, Giorno L (eds) Membrane technology,
vol 3. WILEY-VCH Verlag GmbH & Co. KGaA,
lms as active packaging systems (Atars Weinheim
et al. 2011; Chana-Thaworn et al. 2011; Gemili Gemili S, Yemenicioglu A, Altinkaya SA (2009) Develop-
et al. 2009; Snchez-Gonzlez et al. 2011). Micro- ment of cellulose acetate based antimicrobial food
encapsulation is another strategy for controlled packaging materials for controlled release of lysozyme.
J Food Eng 90:453462
delivery. It has been shown that it can be used Gonzlez A, Strumia MC, Alvarez Igarzabal CI
coupled to traditional packaging materials or bio- (2011) Cross-linked soy protein as material for biode-
polymers for active packaging systems releasing gradable lms: synthesis, characterization and biodeg-
natural antimicrobials (Guarda et al. 2011) or radation. J Food Eng 106:331338
Guarda A, Rubilar JF, Miltz J, Galotto MJ (2011) The
aroma compounds (Marcuzzo et al. 2010). antimicrobial activity of microencapsulated thymol
and carvacrol. Int J Food Microbiol 146:144150
Hu Q, Fang Y, Yang Y, Ma N, Zhao L (2011) Effect of
References nanocomposite-based packaging on postharvest
quality of ethylene-treated kiwifruit (Actinidia
Abdorreza MN, Cheng LH, Karim AA (2011) Effects of deliciosa) during cold storage. Food Res Int
plasticizers on thermal properties and heat sealability of 44:15891596
sago starch lms. Food Hydrocoll 25:5660 John J, Tang J, Bhattacharya M (1998) Processing of
Atars L, Prez-Masi R, Chiralt A (2011) The role of biodegradable blends of wheat gluten and modied
some antioxidants in the HPMC lm properties and polycaprolactone. Polymer 39:28832895
lipid protection in coated toasted almonds. J Food Kantola M, Helen H (2001) Quality changes in organic
Eng 104:649656 tomatoes packaged in biodegradable plastic lms.
Baker RW (2002) Future directions of membrane gas sep- J Food Qual 24:167176
aration technology. Ind Eng Chem Res 41:13931411 Lee DS, Yam K, Piergiovanni L (2008) Overview of food
Belbekhouche S, Bras J, Siqueira G, Chappey C, Lebrun L, packaging system, Ch. 1, In: Food Pack. Sci. Tech.,
Kheli B, Marais S, Dufresne A (2011) Water sorption CRC Press, p 57
behavior and gas barrier properties of cellulose whis- Li X, Li X, Ke B, Shi X, Du Y (2011) Cooperative perfor-
kers and microbrils lms. Carbohydr Polym mance of chitin whisker and rectorite llers on chitosan
83:17401748 lms. Carbohydr Polym 85:747752
Cameron AC, Talasila PC, Joles DW (1995) Predicting Marcuzzo E, Sensidoni A, Debeaufort F, Voilley A (2010)
lm permeability needs for modied-atmosphere pack- Encapsulation of aroma compounds in biopolymeric
aging of lightly processed fruits and vegetables. Hort emulsion based edible lms to control avour release.
Sci 30:25 Carbohydr Polym 80:984988
Chana-Thaworn J, Chanthachum S, Wittaya T (2011) Mensitieri G, Di Maio E, Buonocore GG, Nedi I,
Properties and antimicrobial activity of edible lms Oliviero M, Sansone L, Iannace S (2011) Processing
incorporated with kiam wood (Cotyleobium and shelf life issues of food packaging materials and
lanceotatum) extract. LWT Food Sci Technol structures from renewable resources. Trends Food Sci
44:284292 Technol 22:7280
Chivrac F, Pollet E, Averous L (2009) Progress in nano- Muratore G, Del Nobile MA, Buonocore GG, Lanza CM,
biocomposites based on polysaccharides and Nicolosi AC (2005) The inuence of using biodegrad-
nanoclays. Mater Sci Eng R 67:117 able packaging lms on the quality decay kinetic of
Food Processing by Membrane Operations 793
plum tomato (PomodorinoDatterino ). J Food Eng from a uid based on their size and shape and resort
67:393399 to high pressure and tailor-made membranes with
Snchez-Gonzlez L, Chfer M, Gonzlez-Martnez C,
Chiralt A, Desobry S (2011) Study of the release of specic pore sizes. The pore size distribution
limonene present in chitosan lms enriched with ber- allows indeed different possibilities in terms of
gamot oil in food simulants. J Food Eng 105:138143 separation to be attained via microltration, ultra-
Sarazin P, Li G, Orts WJ, Favis BD (2008) Binary and ltration, nanoltration, and reverse osmosis
ternary blends of polylactide, polycaprolactone and
thermoplastic starch. Polymer 49:599609 (sorted by decreasing pore size) (Drioli et al. 2011).
Srinivasa PC, Baskaran R, Ramesh MN, Harish Prashanth Membrane processing has met with industrial
KV, Tharanathan RN (2002) Storage studies of mango success in clarication and concentration of liquid
packed using biodegradable chitosan lm. Eur Food foods because it operates at room temperature,
Res Technol 215:504508
Torchia AM, Clarizia G, Figoli A, Drioli E (2004) Poten- entails low energy consumption, and is highly
tiality of PEEKWC as new material in food packaging. preformant with regard to a wide range of F
Ital J Food Sci XV:173184 contaminants besides scaling up easily
Yu Q, Song Y, Shi X, Xu C, Bin Y (2011) Preparation and (SantAnna et al. 2012); it may also increase yield,
properties of chitosan derivative/poly(vinyl alcohol)
blend lm crosslinked with glutaraldehyde. Carbohydr and its high selectivity can be taken advantage to
Polym 84:465470 improve control (Drioli et al. 2011). Membranes are
able to remove sediments, microorganisms, and
large-sized compounds; they offer indeed an ef-
cient way to gain superior quality and safety with-
out disturbing the basic sensory prole of a food
Food Processing by Membrane product. As major disadvantages, one should out-
Operations line high hydraulic pressure drops, limited maxi-
mum attainable concentration, and concentration
Angela C. Macedo1 and F. Xavier Malcata2 polarization especially in media with high loads
1
LEPABE Laboratory of Engineering of of organic compounds (Petrotos et al. 2010).
Processes, Environment, Biotechnology and Microltration separates suspended particles,
Energy, University of Porto, Porto, Portugal typically from 0.1 to 10 mm (thus including bac-
2
DEQ Department of Chemical Engineering and teria); for sterilization purposes, 0.40.6 mm
LEPABE Laboratory of Engineering of membranes in dead-end and cross-ow ltration
Processes, Environment, Biotechnology and modes are recommended. Permeate uxes of the
Energy, University of Porto, Porto, Portugal order of 100,000 L.m2.h1 are typically
observed when processing (clean) water; liquid
foods usually reduce such a permeate ux to
Food processing consists of transformation of raw 100200 L.m2.h1, owing primarily to viscosity,
ingredients into food, or food itself into other fouling, and concentration polarization
forms. Food processing typically starts with clean, (Accomazzo et al. 1988).
harvested crops or butchered animal products and Ultraltration resorts to membranes with pore
uses them to produce safe and attractive sizes in the range of 0.10.001 mm; it is typically
products along the food chain, down to the con- intended for removal of high molecular weight
sumer at home (Marsh and Angold 2004). Various substances, colloidal materials, and polymeric
unit operations are part of the current practice, molecules of organic or inorganic nature. Ultral-
which rely on either thermal processing or mechan- tration is a cross-ow separation process i.e., the
ical work: the former are less appropriate to heat- liquid feed ows tangentially along the membrane
labile items and are also characterized by poor surface with a pressure drop across it, causing part
thermodynamic efciency. Mechanical work is, to of that stream to percolate the membrane (thus
advantage, applied in membrane-based processes: producing a permeate). The type and amount of
they take advantage of a physical barrier (i.e., a species left in the remaining stream (retentate)
porous membrane or lter) to separate particles depend on the characteristics of the membrane
794 Formulation by Membrane Emulsification
and the operating conditions, as well as the com- following this conguration. This type of
position of the original feed. Optimum design and contactor technology has found applications in
operation should take into account the ow veloc- purication of water, wine fermentation, protein
ity, the pressure drop, the power consumption, the extraction, and beverage carbonation (Stanojevc
rate of membrane fouling, and the module cost. et al. 2003).
Since only high molecular weight species are
removed, the differential in osmotic pressure
across an ultraltration membrane is usually neg- References
ligible (Accomazzo et al. 1988).
Nanoltration is a cross-ow ltration tech- Accomazzo M, Ganzi G, Kaiser R (1988) Deionized (di)
water ltration technology. In: Tolliver D -
nology, usually listed between ultraltration and
(ed) Handbook of contamination control in
reverse osmosis; the nominal pore size of the microelectronic. Noyes Publications, Park Ridge,
membrane is typically ca. 1 nm, but the molec- pp 210254
ular weight cutoff which is typically less than Drioli E, Stankiewicz A, Mecdonio F (2011) Membrane
engineering in process intensication an overview.
1,000 Da, rather than the nominal pore size is to
J Membr Sci 380:18
be used for a more meaningful characterization. Marsh R, Angold R (2004) Sources of contamination in the
Transmembrane pressures can go up to 3 MPa, food chain. In: Edwards M (ed) Detecting foreign bod-
which require powerful pumping equipment; fur- ies in food. CRC Press, Boca Raton, pp 311
Petrotos K, Tsiadi A, Poirazis A, Papadopoulus D,
thermore, nanolter membranes are quite suscep-
Petropakis H, Gkoutsidis P (2010) A description of a
tible to scaling and fouling, so such modiers as at geometry direct osmotic concentrator to concen-
antiscalants are often required. trate tomato juice at ambient temperature and low pres-
Finally, reverse osmosis goes one step further sure. J Food Eng 97:235242
SantAnna V, Marczacl L, Tessaro I (2012) Membrane
in ltration and allows separation of small
concentration of liquids foods by forward osmosis:
dissolved molecules in the Angstrom range; this process and quality view. J Food Eng 11:483489
includes salts, sugars, and organic acids. The only Stanojevc M, Lazarevc B, Radc D (2003) Review of
driving force for this process is the osmotic pres- membrane contactor designs and applications of differ-
ent modules in industry. FME Trans 31:9198
sure difference between the solutions lying on the
two sides of the permeable membrane. The cross-
ow conguration allows the membranes to self-
clean continually; further advantages encompass
relatively low hydraulic pressure and easy scale- Formulation by Membrane
up (SantAnna et al. 2012). Emulsification
Membrane contactors represent another
approach to membrane operations in food Goran T. Vladisavljevi and Richard G. Holdich
processing where separation does not rely spe- Department of Chemical Engineering,
cically on the membrane acting as selective bar- Loughborough University, Loughborough,
rier, but is instead based on phase equilibria. Leicestershire, UK
Hence, mass transfer between phases is promoted
without dispersing one phase into the other, and
such disadvantages as ooding at high ow rates, Membrane emulsication (ME) is a process of
unloading at low ow rates, requirement for emul- forming emulsion by passing a pure dispersed
sied state and density differences, and high interphase or pre-emulsion through a microporous
facial area are dramatically alleviated. In membrane (Fig. 1). The most commonly used
principle, all traditional stripping, scrubbing, membranes for ME are Shirasu porous glass
absorption, evaporation, distillation, crystalliza- (SPG) membrane and microsieve membranes. In
tion, emulsication and liquid-liquid extraction, direct ME, ne droplets are produced at the mem-
as well as enzymatic catalysis can be carried out brane/continuous phase interface by injecting a
Formulation by Membrane Emulsification 795
Formulation by Membrane Emulsification, (ME): a bottom-up approach (direct ME); b Top-down

Fig. 1 Two different approaches used to obtain ne drop- approach (premix ME)
lets of oil-in-water emulsion in membrane emulsication
pure liquid (the dispersed phase) through the rotating (Vladisavljevi and Williams 2006) or
membrane into a second immiscible liquid (the vibrating (Holdich et al. 2010) the membrane
continuous phase) (Nakashima et al. 1991). The within the otherwise static continuous phase or
dispersed phase should not wet the membrane providing periodic ow pulsations in the continu-
wall, i.e., hydrophobic and hydrophilic mem- ous phase (Holdich et al. 2013). In the dynamic
branes are used to produce water-in-oil and and pulsed ow membrane systems, the shear
oil-in-water emulsions, respectively. At low trans- stress on the membrane surface is decoupled
membrane uxes, uniform droplets can be formed from the cross-ow velocity, and consequently,
without applying any shear on the membrane sur- sufciently high shear stress can be achieved at
face, solely by the action of interfacial tension any cross-ow velocity, and emulsions with high
(Kukizaki 2009). In order to achieve commer- dispersed phase concentrations can be produced
cially signicant throughputs in ME, the shear in a single pass through the module.
stress is applied at the continuous phase/mem- In the dripping regime, the droplet size is deter-
brane interface, usually by cross ow or stirring mined by a balance between the shear force and
(Fig. 2ad). Cross-ow systems are easy to scale the capillary force. The size of the droplets pro-
up and can offer a continuous operation and con- duced is mainly affected by the microstructure of
stant shear stress along the membrane surface. the membrane (the pore size distribution, pore
Stirring systems are operated batchwise and do shape, and pore spacing), transmembrane ux,
not provide constant shear stress on the membrane surface shear stress, physical properties of the
surface, but are easier to operate because a closed continuous and dispersed phase, and emulsion
loop recirculation of the continuous phase stream formulation. Surfactant molecules must not
is not needed and a batch size can be very low, just adsorb to the membrane surface, and thus, cat-
10 mL or even less. ionic surfactants must not be used in association
The shear stress can also be generated by a with negatively charged membrane (such as SPG
dynamic membrane (Fig. 1e) or pulsed ow membrane and oxidized silicon microsieves)
(Fig. 1f), in which case the droplet detachment (Nakashima et al. 1993). The transmembrane
from the membrane surface is facilitated by pressure should not exceed ten times the capillary
796 Formulation by Membrane Emulsification
Formulation by Membrane Emulsification, Fig. 2 Different ME systems used to control shear stress on membrane
surface (Vladisavljevi and Williams 2005)
pressure and the shear stress on the membrane membrane surface: inuence of surfactant type and
surface is usually in the range from 2 to 10 Pa concentration, viscosities of dispersed and continuous
phases, and transmembrane pressure. J Membr Sci
(Vladisavljevi et al. 2004). The droplet size is 327:234243
typically two to ten times higher than the pore size Nakashima T, Shimizu M, Kukizaki M (1991) Membrane
and decreases with increasing the shear stress on emulsication by microporous glass. Key Eng Mater
the membrane surface. 6162:513516
Nakashima T, Shimizu M, Kukizaki M (1993) Effect of
surfactant on production of monodispersed O/W emul-
sion in membrane emulsication. Kag Kog Ronbunshu
19:991997
References Vladisavljevi GT, Williams RA (2005) Recent develop-
ments in manufacturing emulsions and particulate
Holdich RG, Dragosavac MM, Vladisavljevi GT, products using membranes. Adv Colloid Interf Sci
Kosvintsev SR (2010) Membrane emulsication with 113:120
oscillating and stationary membranes. Ind Eng Chem Vladisavljevi GT, Williams RA (2006) Manufacture of
Res 49:38103817 large uniform droplets using rotating membrane emul-
Holdich RG, Dragosavac M, Vladisavljevi GT, sication. J Colloid Interface Sci 299:396402
Piacentini, E (2013) Continuous membrane emulsica- Vladisavljevi GT, Lambrich U, Nakajima M, Schubert H
tion with pulsed (oscillatory) ow. Ind Eng Chem Res (2004) Production of O/W emulsions using SPG mem-
52:507515 branes, ceramic a-Al2O3 membranes, microuidizer
Kukizaki M (2009) Shirasu porous glass (SPG) membrane and a microchannel plate: a comparative study. Colloid
emulsication in the absence of shear ow at the Surf A 232:199207
Forward Osmosis (FO) 797
(Klaysom et al. 2013), seawater desalination

Forward Osmosis (FO) (McCutcheon et al. 2005), wastewater treatment
and concentration of diluted streams (Holloway
Abaynesh Yihdego Gebreyohannes and et al. 2007), food processing (Petrotos et al. 1998),
Lidietta Giorno removal of trace organic matter (Cartinella
Institute on Membrane Technology, National et al. 2006), and for use in membrane bioreactor
Research Council of Italy, ITM-CNR, Rende, (Zhang et al. 2012). As shown in Fig. 1, there is an
Italy exponential rise in the number of FO-related publi-
cations in the last 5 years. Similarly, the number of
patents released from 2009 to 2014 is signicantly
In literature, osmotic pressure-driven membrane high (Fig. 2), providing a pertinent prospect to its
processes are known as direct osmosis (DO) or application in various elds. F
forward osmosis (FO). FO dating back to the early The following are summaries of the range of
1960s is a membrane contactor process that uti- potential advantages of FO process:
lizes osmotic pressure difference (Dp) across the
membrane for water transport through a semiper- Requires low/no hydraulic pressure since it is
meable membrane (Cartinella et al. 2006; Cath an osmotically driven process.
et al. 2006) than a hydraulic pressure as in Low energy consumption, thereby lowing
RO. The osmotic pressure difference across the costs, if suitable draw solutes and their regen-
membrane arises from the use of concentrated eration methods can be economically and tech-
solution called draw solution (DS) on the perme- nically developed.
ate side of the membrane (Fig. 3a). Apparently FO Energy can be harvested from the mixing of
may be a viable and sustainable alternative to freshwater and saline water by pressure-
thermal-driven (membrane distillation) and retarded FO (PRO).
pressure-driven (reverse osmosis) concentrating Membrane fouling in FO is relatively low and
methods, which are highly energy intensive, more reversible and can be minimized by opti-
hence very expensive. In the presence of natural mizing the hydrodynamics.
sources of concentrated DS, e.g., seawater, FO Numerous contaminants can be effectively
can be very attractive due to its signicantly rejected by the FO process.
lower energy demand for pumping (Tang Has the potential to help achieve high water
et al. 2010). The equation below is generally recovery, which reduces the volume of desali-
used to describe water transport in FO: nation brine, a major environmental concern
for current desalination.
J w AsDp DP (1) In the food and pharmaceutical processing, FO
can maintain color, taste, aroma, and nutrition
where Jw is the water ux, A the water permeabil- of the feed.
ity constant of the membrane, s the reection In the medical applications, FO can help in the
coefcient, and DP is the applied pressure. For controlled release of drugs with low oral bio-
FO, DP is zero. availability in a controlled manner by osmotic
Currently many researchers are focusing on the pumps.
use of FO for different applications, and remark-
able progresses are observed in optimizing the
process through testing different DS, congura- FO Membranes, Membrane
tions, and development of model equations that Configurations, and Modeling
could best explain the phenomena. Membrane Transport
Hence, FO is reemerging as low-energy demand-
ing membrane operation (<30 kwh/m3) (Beaudry Except for a few authors such as McCutcheon
et al. 1999) for dehydration of aqueous solution et al. (2005) which used a thin-lm composite
798 Forward Osmosis (FO)
Forward Osmosis (FO), Fig. 1 The number of publica- do?product=UA&search_mode=CitationReport&SID=

tions excluding patents retrieved from web of science T1GuVqD4QDADmHXOVbu&page=1&cr_pqid=1&
between the year 1996 and 2015 using the keyword forward viewType=summary
osmosis. http://apps.webofknowledge.com/CitationReport.
Forward Osmosis (FO), 30

Fig. 2 Number of patents
No of patenets per annum
related to FO from 2009 to

2014 retrieved from http://
worldwide.espacenet.com/ 20
using the key word
forward osmosis
10
0
2009 2010 2011
2012
2013
2014
Year
polyamide membrane, most publications are based packing congurations in the membrane industry.
on cellulose or cellulose derivate (cellulose acetate) However, it cannot be used in its current design
membranes mainly from Hydration Technology for FO because a liquid stream cannot be forced to
Inc. In addition, in most cases reverse osmosis ow on the support side (inside the envelope)
(RO) was used to reconcentrate the DS which (Cath et al. 2006). Hence, for a continuous FO
gets diluted in the course of FO. operation, at-sheet membrane and specially
Of the different membrane modules, spiral designed spiral wound membrane are used.
wound modules are one of the most common Although applied in limited number, tubular
membranes are also very suitable for continuous Concentration Polarization

FO processing since they are self-supported, can and Membrane Fouling
be fabricated easily, and allow free ow of liquid
on both sides of the membrane. Unlike pressure-driven membrane process, in FO
There are a number of papers discussing the there are both dilutive and concentrative polariza-
transport phenomena in FO process in particular tion effects that result in a lower-than-expected
related to solution physicochemical properties, con- water ux. Furthermore, the concentration polar-
centration polarization, and membrane fouling. For ization could be either external which occurs near
detailed reading, the reader is referred to Tang the surface of the active layer and the porous
et al. (2010), McCutcheon and Elimelech (2006), support layer or internal which occurs inside the
Tan and Ng (2008), and Zhao and Zou (2011). substructure of porous support layer. As a result,
the process is always accompanied by great reduc- F
tion from the bulk osmotic pressure (Dpbulk) to the
Draw Solutions effective osmotic pressure (Dpeffective) as shown in
Fig. 3b.
In FO process, a concentrated solution on the
permeate side of the membrane is the source of External Concentration Polarization
the osmotic driving force. In literature, various When the process is in FO mode of operation, i.e.,
names are given to this solution including: draw feed solution facing the active side of the mem-
solution, osmotic agent, osmotic media, driving brane, there is accumulation of solute that is
solution, osmotic engine, sample solution, or rejected by the membrane. This will result in the
brine (Cath et al. 2006). The main criterion for phenomena of concentrative external concentra-
selecting the type of draw solution (DS) is its tion polarization (ECP). At the same time, the DS
potential to create higher chemical potential than in contact with the permeate side of the membrane
the feed solution. gets diluted by the permeating water at the
The DS and related osmotic pressures in the permeatemembrane interface resulting in dilu-
FO process are important factors inuencing over- tive external concentration polarization (ECP).
all process performance. Many researchers used Due to these phenomena, the effective osmotic
salt solution as DS (Cartinella et al. 2006; driving force near the membrane is signicantly
Cornelissen et al. 2008; Yong et al. 2012), while lower than the bulk (apparent) osmotic pressure
in some cases sugars or ammoniacarbon dioxide (Cath et al. 2006).
(McCutcheon et al. 2005) solutions were used. Interestingly, several studies have shown that
According to Lutchmiah et al. (2014), based on fouling induced by ECP in FO in most part is
publication from 2005 to 2013, 40 % of the cases reversible due to low foulant compaction as a
utilized NaCl as DS. This is mainly attributed by result of the negligible hydraulic pressure gradi-
its high solubility (Cath et al. 2006), low cost ent. Therefore, FO holds a great potential to treat
(Achilli et al. 2010), and relatively high osmotic wastewater (Lutchmiah et al. 2014), including
potential. The same author indicated that 12 % wastewater with high fouling propensity (Cicci
used MgCl2, 8 % sugars, 10 % sulfates, and other et al. 2013; Karaouzas et al. 2011). The main
7 % used magnetic nanoparticles, real wastewater, strategy utilized to reduce the effect of ECP is
carbonates, etc. In addition to these, seawater has increased turbulence or cross ow velocity. More-
frequently been used as a DS (Cath et al. 2010). over, recovering FO membrane performance was
However, its use was mostly affected by abundant possible using simple osmotic back ush, after it
presence of particles and microorganisms which is fouled. For example, an intermittent osmotic
may foul the subsequent system used to back ush has given the possibility to use a single
reconcentrate the DS (in closed-loop systems) or FO membrane over several cycles bringing the
cause fouling/biofouling in the FO unit, hamper- total membrane use to 172 h, during the treatment
ing performance. of vegetation wastewater (Gebreyohannes
800 Forward Osmosis (FO)
FEED J
solute
a b
Jw
ACTIVE LAYER

membrane
Feed
DS
bulk
POROUS SUPPORT
FO
effective
DRAW SOLUTION
Forward Osmosis (FO), Fig. 3 Water ows in FO sys- of bulk osmotic driving force reduction to effective
tem where (a) water diffuses to the more saline side of osmotic driving force due to concentration polarization
the membrane due to osmotic driving force under phenomena
approximately zero hydraulic pressure and (b) the course
et al. 2015). This is an important savings in life- porosity, and tortuosity of the porous substructure
time, compared to the complete performance loss (Cornelissen et al. 2008). It is worth noting that
observed in various pressure-driven membrane compared to ECP, ICP has a more prominent
operations. The FO fouling is generally more pro- effect in signicantly reducing the effective driv-
nounced when the porous support layer is facing ing force across the dense layer from the bulk
feed solution, since this orientation is prone to osmotic difference.
sever internal pore clogging.
Internal Concentration Polarization Promising Application Areas

The asymmetric FO membrane consists of a dense
active layer and a porous support layer, which FO has been studied for a number of applications
adds a more complex situation of concentration including commercial applications, in the water and
polarization in FO. In FO applications, where the wastewater treatment (e.g., extraction bags) and in
active layer of the membrane faces the feed solu- the pharmaceutical industry (e.g., osmotic pumps),
tion and the porous support layer faces the DS, seawater/brackish water desalination, food
there is also dilutive internal concentration polar- processing, drug delivery, and electric power gener-
ization (ICP) within the porous substructure. This ation. For instance, FO was used to concentrate an
phenomenon happens when water diffuses oily wastewater from olive millings after removing
through the active layer; the DS within the porous courser particles in two different congurations. In
substructure becomes diluted, hence dramatically the rst case the pre-ltered wastewater directly fed
reduced effective osmotic driving force. Unlike to the FO process, while in another case permeate of
the external concentration polarization, playing biocatalytic membrane reactor was used as feed. In
with the hydrodynamic condition will not allevi- both cases, FO helped to reduce up to 70 % of the
ate the problem of ICP. Thus, an engineered solu- initial volume while completely rejecting all pollut-
tion for reduced internal concentration ants in the course of reclaiming puried water and
polarization is optimization of the thickness, pharmacologically important compounds. As a
Forward Osmosis (FO), Table 1 Examples of industrial plants that employed FO process
Water Pollutant rejection
Stream Plant name/location recovery DS (%)
Landll leachate (Osmotek Cofn Butte RO: DS NaCl 99
Inc. 2003) Landll, Corvallis, and water
Oregon 92 %
Direct potable reuse for advanced DOC NASA long NaCl Direct potable reuse
life-support systems (Beaudry mission life-support
et al. 1999) system
Concentration of digested sludge Truckee Meadows RO NaCl 8799
liquids (Holloway et al. 2007) Water Reclamation
Facility, Reno,
Nevada
Hydration bags (Hydration Hydration Edibles, e.g., Free of F
Technologies Inc. n.d.) Technology Inc. sugar, beverage microorganism,
powder macromolecules, and
ions
result it reduced the size of downstream process and high-density packing methods for well-
required to further fractionation and concentration performing at-sheet FO membranes.
of the recovered pharmaceutically interesting Development of DS: Most desirable DS needs
compounds. to induce high osmotic pressure, requires low
Table 1 gives summary of industrial practices energy for reconcentration, and must be easily
on the use of FO process in various elds. It is separable from the product freshwater and
worth noting that in most wastewater treatment nonreactive with low or no toxicity.
applications, FO is used as a high-level
pretreatment step before an ultimate purication
References
step. An exception for this is a hydration bag,
which serves as a nal treatment step to produce Achilli A, Cath TY, Childress AE (2010) Selection of
potable water from wastewater. inorganic-based draw solutions for forward osmosis
applications. J Membr Sci 364:233241
Beaudry EG, Herron JR, Peterson SW (1999) Direct osmo-
sis concentration of waste water: nal report. Osmotek,
Concluding Remark Corvallis
Cartinella JL, Cath TY, Flynn MT, Miller GC, Hunter KW,
Regardless of the limited presence of robust mem- Childress AE (2006) Removal of natural steroid hor-
branes and membrane modules for FO, funda- mones from wastewater using membrane contactor
processes. Environ Sci Technol 40:73817386
mental research on FO and the development of Cath TY, Childress AE, Elimelech M (2006) Forward
new applications of FO are growing fast. How- osmosis: principles, applications, and recent develop-
ever, the following points need to be addressed to ments. J Membr Sci 281:7087
have a reliable commercial success: Cath TY, Hancock NT, Lundin CD, Hoppe-Jones C,
Drewes JE (2010) A multi-barrier osmotic dilution
process for simultaneous desalination and purication
Development of new membranes in both at- of impaired water. J Membr Sci 362:417426
sheet and hollow ber congurations: The Cicci A, Stoller M, Bravi M (2013) Microalgal biomass
development needs to orient in the production production by using ultra- and nanoltration membrane
fractions of olive mill wastewater. Water Res
of membrane and modules with high water 47:47104718
permeability, high solute rejection, limited Cornelissen ER, Harmsen D, de Korte KF, Ruiken CJ, Qin
ICP, high chemical and mechanical stability, J-J, Oo H, Wessels LP (2008) Membrane fouling and
802 Forward Osmosis Membrane Reactor
process performance of forward osmosis membranes

on activated sludge. J Membr Sci 319:158168 Forward Osmosis Membrane Reactor
Gebreyohannes AY, Curcio E, Poerio T, Mazzei R, Di
Proo G, Drioli E, Giorno L (2015) Treatment of
olive mill wastewater by forward osmosis. Sep Purif Rong Wang1,2, Laurentia Setiawan2 and
Technol 147:292302 Anthony Gordon Fane2
Holloway RW, Childress AE, Dennett KE, Cath TY 1
School of Civil and Environmental Engineering;
(2007) Forward osmosis for concentration of anaerobic
digester centrate. Water Res 41:40054014 Singapore Membrane Technology Center,
Hydration Technologies Inc. (n.d.) Hydration bags- Nanyang Environment and Water Research
technology overview: http://www.hydrationtech.com/ Institute; Nanyang Technological University,
merchant.mv?Screen=CTGY&StoreCode=MHTI& Singapore, Singapore
CategoryCode=TECH-FO 2
Karaouzas I, Skoulikidis NT, Giannakou U, Albanis TA Singapore Membrane Technology Centre,
(2011) Spatial and temporal effects of olive mill waste- Nanyang Environment and Water Research
waters to stream macroinvertebrates and aquatic eco- Institute, Nanyang Technological University,
systems status. Water Res 45:63346346 Singapore, Singapore
Klaysom C, Cath TY, Depuydt T, Vankelecom IFJ
(2013) Forward and pressure retarded osmosis: poten-
tial solutions for global challenges in energy and water Introduction
supply. Chem Soc Rev 42:69596989
Lutchmiah K, Verliefde ARD, Roest K, Rietveld LC,
Cornelissen ER (2014) Forward osmosis for applica- The forward osmosis membrane bioreactor
tion in wastewater treatment: a review. Water Res (FOMBR) is a new type of membrane bioreactor
58:179197 (MBR) technology, which combines two basic
McCutcheon JR, Elimelech M (2006) Inuence of concen-
processes biological degradation of organic
trative and dilutive internal concentration polarization
on ux behavior in forward osmosis. J Membr Sci wastes and membrane separation using forward
284:237247 osmosis (FO) into a single process for various
McCutcheon JR, McGinnis RL, Elimelech M (2005) wastewater treatments (Zhang et al. 2012). Spe-
A novel ammoniacarbon dioxide forward (direct)
cically, in a FOMBR, suspended solids and
osmosis desalination process. Desalination 174:111
Osmotek Inc. (2003) Landll leachate treatment http:// microorganisms responsible for biodegradation
www.rimnetics.com/osmotek.htm are separated from the treated water using a FO
Petrotos KB, Quantick P, Petropakis H (1998) A study of membrane module, where water is extracted from
the direct osmotic concentration of tomato juice in
the mixed liquor through a semipermeable FO
tubular membrane module conguration. I. The effect
of certain basic process parameters on the process per- membrane to a high osmotic pressure draw solu-
formance. J Membr Sci 150:99110 tion (Achilli et al. 2009). The output from the
Tan CH, Ng HY (2008) Modied models to predict ux FOMBR is a mixture of treated water permeate
behavior in forward osmosis in consideration of exter-
and the draw solution. A posttreatment step, such
nal and internal concentration polarizations. J Membr
Sci 324:209219 as reverse osmosis (RO) or membrane distillation
Tang CY, She Q, Lay WCL, Wang R, Fane AG (MD), can be used to produce puried water and
(2010) Coupled effects of internal concentration polar- to regenerate the draw solution.
ization and fouling on ux behavior of forward osmosis
Compared with conventional MBRs, where
membranes during humic acid ltration. J Membr Sci
354:123133 pressure-driven membrane ltration units are
Yong JS, Phillip WA, Elimelech M (2012) Coupled reverse used, the FOMBR enjoys important advantages
draw solute permeation and water ux in forward including potentially lower energy requirement
osmosis with neutral draw solutes. J Membr Sci
and better water quality, because the separation
392393:917
Zhang J, Loong WLC, Chou S, Tang C, Wang R, Fane AG does not entail the use of hydraulic pressure, and
(2012) Membrane biofouling and scaling in forward the salt-rejecting dense membrane is an absolute
osmosis membrane bioreactor. J Membr Sci barrier to suspended solids and also retains
403404:814
organic species. In addition, the use of the dense
Zhao S, Zou L (2011) Relating solution physicochemical
properties to internal concentration polarization in for- membrane allows retention of recalcitrant
ward osmosis. J Membr Sci 379:459467 organics which are eventually degraded, as
Forward Osmosis Membrane Reactor 803
Forward Osmosis Purified water

Membrane Reactor,
Fig. 1 Submerged
FOMBR
Draw solution
Feed Draw regeneration
solution process
FO membrane
module
Diluted draw
solution
Air F
Sludge
Retentate
Purified water
Wastewater
in
Draw
solution
Feed Draw solution
Air
regeneration process
Bioreactor FO membrane
Sludge out Diluted draw solution
Forward Osmosis Membrane Reactor, Fig. 2 Side-stream FOMBR
organic retention time (ORT) can be signicantly congurations, the driving force for water transfer
longer than the hydraulic retention time is the osmotic pressure gradient across a salt-
(HRT). A greatly enhanced removal of rejecting dense membrane. When water is
micropollutants is also expected, which is a dis- extracted into the draw solution, the diluted draw
tinct advantage if the efuent is considered for solution is, subsequently, sent to a regeneration
reclamation and reuse. device (e.g., RO or MD) to recover the puried
water. The commonly used FO membrane has an
RO-like or nanoltration (NF)-like selective layer
Configurations of FOMBR that has high rejection toward organic/inorganic
compounds in the wastewater including ions.
Similar to conventional MBRs, the FOMBR can Normally two-phase ow around ber bundles
be operated under two congurations. One is the (for a hollow ber module) or between vertical
submerged FOMBR as shown schematically in at plates (for a at sheet membrane module) by
Fig. 1 and the other is side-stream FOMBR, as air scouring is used to control membrane fouling
illustrated in Fig. 2 (Yap et al. 2012). In both and module blocking.
804 Fossil Fuels Processing by Membrane Operations
Major Challenges for Large-Scale the feed tank, the driving force will be reduced,
Applications leading to ux decline and toxic effects on the
microbial community. In addition, the draw solu-
There are three major technical challenges that tion must be able to be regenerated easily and
impede large-scale applications of FOMBRs. economically for reuse (Cornelissen et al. 2008).
Firstly, an appropriate membrane is essential to
improve the performance of the FOMBR. Com-
monly used FO membranes consist of a dense- References
selective layer supported by a porous substrate.
The membrane should exhibit high water ux and Achilli A, Cath TY, Marchand EA, Childress AE
(2009) The forward osmosis membrane bioreactor: a
high salt rejection and be able to limit feed-side
low fouling alternative to MBR processes. Desalination
fouling and internal concentration polarization 239:1021
(ICP), which is a unique phenomenon in the forward Cath TY, Childress AE, Elimelech M (2006) Forward
osmosis process, causing the reduction of the effec- osmosis: principles, applications, and recent develop-
ments. J Membr Sci 281:7087
tive osmotic driving force. The strategy to overcome
Cornelissen ER, Harmsen D, de Korte KF, Ruiken CJ, Qin
the problem is to design the support layer of the FO J-J, Oo H, Wessels LP (2008) Membrane fouling and
membrane as thin and porous as possible to reduce process performance of forward osmosis membranes
mass transfer resistance, and thus, less severe ICP is on activated sludge. J Membr Sci 319:158168
Nawaz MS, Gadelha G, Khan SJ, Hankins N (2013) Micro-
expected (Wang et al. 2010). Fouling is a common bial toxicity effects of reverse transported draw solute
issue in a conventional MBR as the wastewater is a in the forward osmosis membrane bioreactor (FO-
complex system containing microorganisms, extra- MBR). J Membr Sci 429:323329
cellular polymeric substances (EPS), and soluble Wang R, Shi L, Tang CY, Chou S, Qiu C, Fane AG
(2010) Characterization of novel forward osmosis hol-
macro-/microorganics and inorganics. FOMBR, on
low ber membranes. J Membr Sci 355:158167
the other hand, has an additional potential scaling Yap WJ, Zhang J, Lay WCL, Cao B, Fane AG, Liu
issue caused by inorganic salts due to solute accu- Y (2012) State of the art of osmotic membrane bio-
mulation (Zhang et al. 2012). reactors for water reclamation. Bioresour Technol
122:217222
Secondly, the bioprocess must be optimized to
Zhang J, Loong WLC, Chou S, Tang C, Wang R, Fane AG
handle accumulated recalcitrant organics, nutri- (2012) Membrane biofouling and scaling in forward
ents, and elevated salt concentrations, as the FO osmosis membrane bioreactor. J Membr Sci
membrane possesses a dense-selective layer with 403404:814
high retentions to the organic species and salts.
The accumulation of organic matters improves the
biological degradation which may enhance the
removal of micropollutants. The salt accumula- Fossil Fuels Processing by Membrane
tion, however, (i) could have adverse effects on Operations
biological treatment and cause microbial inhibi-
tion (Zhang et al. 2012; Nawaz et al. 2013); May-Britt Hgg
(ii) reduces the osmotic driving force available, Department of Chemical Engineering, Norwegian
resulting in the reduction of water ux; and (iii) University of Science and Technology,
increases the concentration of sparingly soluble Trondheim, Norway
salts, and spontaneous chemical precipitation may
occur when the saturation limits are exceeded,
leading to scaling on the FO membranes. Control of anthropogenic emissions of green-
The third challenge relates to draw solute house gases (GHG) is one of the most challenging
selection and regeneration. A proper draw solu- environmental issues related to global climate
tion is necessary to provide high osmotic pressure change. The larger part of emissions are coming
and low salt back diffusion (Cath et al. 2006). from fossil fuels. The Energy Information Admin-
When the salts from the draw solution diffuse to istration (EIA) 2010 predicted a 49 % increase of
Fossil Fuels Processing by Membrane Operations 805
Fossil Fuels Processing by Membrane Operations, adsorption, cryogenics, and membrane separation
Table 1 Overview of global stationary CO2 sources larger may be used for CO2 capture. Membranes are
than 0.1 million tons (Mt) of CO2/year
already an alternative and competitive technology
Number Emissions for other selected gas separation processes such as
of (Mt CO2/
Process sources year) natural gas sweetening, biogas upgrading, and
Fossil Power 4,492 10,539 hydrogen production during the last two or three
fuels decades. Alternative ways of using membranes
Iron and steel 269 646 for fossil fuel processing at power plants are illus-
industry trated in Fig. 1. Additionally, membranes are also
Cement 1,175 932 being used for CO2 removal from natural gas and
production
upgrading of biogas.
Reneries 638 798
Petrochemical 470 379
Postcombustion: The membrane is then typi- F
industry cally retrotted to the plant as an end-of-pipe
Oil and gas N/A 50 solution, and the main gas components are CO2
processing and N2. For a coal-red power plant the content of
Other sources 90 33 CO2 will be 1015 %, while for a natural gas plant
Biomass Bioethanol 303 91 the CO2 content may be as low as ~5 % unless
and bioenergy there is an integrated solution. The ue gas will at
Total 7,887 13,468 this point have a temperature of around 70 C and
(IPCC special report on CCS, working group III) is at atmospheric pressure. These conditions are in
general not very favorable for membrane separa-
energy demands from 2007 to 2035 following tion as there is hardly any driving force, and much
more and more countries becoming industrialized. energy must be used in order to create sufcient
The International Energy Outlook 2010 reference partial pressure difference over the membrane.
case reported that world energy-related carbon The recommended membrane solution in this
dioxide (CO2) emissions would increase from case will be to use a membrane with very high
29.7 billion metric tons in 2007 to 33.8 billion ux and selectivity which can be achieved with a
metric tons in 2020 and 42.4 billion metric tons in facilitated transport membrane having a xed or
2035 (IEO 2010). The emissions of the green- mobile carrier to enhance the transport (Kim
house gas CO2 is produced in a variety of ways et al. 2004), or to use an innovative two-stage
as shown in Table 1 (Hay and Firman 2002). process solution where air is being used as
Among them, fossil fuel power plants are respon- sweep on the second stage and fed to the burner
sible for roughly 40 % of total CO2 emissions, and (Merkel et al. 2010). In both cases, energy
coal-red power plants being the main contribu- demand is minimized by using only a fan on the
tor. Changing from fossil energy to alternative feed side and pulling vacuum on the permeate
renewable energy would be ideal; however, we side. The membrane material will be
will still be dependent on fossil fuels for many polymeric. Impurities in the ue gas stream
years ahead both due to limitations of technology (SO2, NOx, particles) may be a challenge for the
development and also increased growth. Hence, membrane material.
carbon capture and sequestration (CCS) is seen as Precombustion: When used in precombustion
an efcient way to mitigate the emission of CO2 in the membrane will be an integrated solution with
order to ght against global warming. The key the gas turbine and CO2 may thus be removed at
motivation for CCS is that we may continue to higher pressure. After the shift reaction H2 and
use fossil fuels without causing signicant CO2 CO2 will be in a mixture, with close to 40 mol%
emissions. Clearly coal-red power plants need to CO2. The pressure will be in the range of
be rstly tackled, especially by retrotting capture 2530 bar, and temperature around 300350 C
to existing plants. Different techniques such as unless there has been a sulfur removal unit up
chemical and physical absorption, physical front of the membrane. The high temperature,
806 Fossil Fuels Processing by Membrane Operations
Fossil Fuels Processing

by Membrane
Operations,
Fig. 1 Alternative
membrane processing at a
power plant
Postcombustion N2
(PC) O2
Coal
Power & Heat CO2 Separation
Air
CO2
Precombustion
(IGCC)
Steam CO2
Air/O2
CO2
Coal Shift, Gas Cleanup H2 Power Compression
Gasification and
+ CO2 Separation & Heat
Dehydration
Air
Oxyfuel Combustion
Air N2
Air Separation
O2
Coal CO2
Power & Heat
pressure, and concentration are very favorable for will be needed for sequestration. The main chal-
Palladium membranes which can achieve lenge for this membrane solution is the module
extremely efcient separation of the two main construction at this high temperature and also
components, allowing hydrogen to permeate potential poisoning of the membrane.
(Jordal et al. 2004). An advantage is also that Oxyfuel combustion: This solution is very ele-
CO2 is captured at high pressure, and less energy gant, using a membrane to separate out pure
Fouling 807
oxygen. The air separating membrane is, how- for the performance of a process is membrane
ever, still not fully developed. These membranes fouling. Membrane fouling is a general term. It
are suitable for advanced power generation requir- may be the result of concentration polarization; it
ing pure oxygen for combustion or gasication may also be the consequence of adsorption of feed
and are based on zirconia and perovskite where solution constituents at the membrane surface and
oxygen is transported through the material as O2. especially in microltration also within the mem-
The materials are stable at very high temperatures brane structure. The control of concentration
(>500 C) but again module design is very chal- polarization and membrane fouling is a major
lenging at these high operating temperatures. engineering aspect in the design of membrane
separation processes and equipment (Brian 1966;
Blatt et al. 1970; Jonsson and Boesen 1984; Bian
Cross-References et al. 2000; Hoek and Elimelech 2003; F
Cornelissen et al. 2004).
Oxygen Transport Ceramic Membranes: Perov- Concentration polarization effects will occur in
skite and Nonperovskite all membrane separation processes. Its conse-
quences, however, are especially severe in
pressure-driven membrane processes. When in a
References
mass separation procedure a molecular mixture is
Hay BP, Firman TK (2002) Inorg Chem 41:55025512 brought to a membrane surface, some components
International Energy Outlook 2010 highlights. http:// will permeate the membrane under a given driving
www.eia.doe.gov/oiaf/ieo/highlights.html. Accessed force, while others are retained. This leads to an
28 Sept 2010 accumulation of retained material and to a deple-
Jordal K, Bredesen R, Kvamsdal HM, Bolland O (2004)
Energy 29:12691278 tion of the permeating components in the bound-
Kim TJ, Li B, Hgg MB (2004) J Polym Sci ary layers adjacent to the membrane surface. This
B 42:43264336 phenomenon is referred to as concentration polar-
Merkel TC, Lin H, Wei X, Baker R (2010) J Membr Sci ization. The causes and consequences of concen-
359:126139
tration polarization may be rather different in
different membrane processes. Often the adverse
effects of concentration polarization are intensi-
ed by an adsorption of certain feed mixture con-
Fouling stituents at the membrane surface. This
phenomenon is referred to as membrane fouling.
Lidietta Giorno and Napoleone DAgostino Concentration polarization can be minimized
Institute on Membrane Technology, National by hydrodynamic means such as the feed ow
Research Council of Italy, ITM-CNR, Rende, velocity and the membrane module design. The
Italy control of membrane fouling, however, is more
difcult. The transition between concentration
polarization and fouling can be expressed by the
Synonyms concept of critical ux. This concept was intro-
duced in the 1990s (Field et al. 1995; Howell
Fouling in membranes; Membrane fouling 1995; Bacchin 2004), and it is the ux value
below which a decline of ux with time does not
An intrinsic phenomenon of all membrane sepa- occur, above it fouling is observed. Critical ux
ration processes is the decline of the ux through values depend on numerous parameters, such as
the membrane as a function of time, due to con- properties of solutions to be treated (particle size,
centration polarization effects and the formation concentration, ionic strength, pH), surface inter-
of cake or gel layers by feed solution constituents actions (zeta potential, surface tension), and
retained by the membrane. Equally devastating hydrodynamic conditions (axial velocity)
808 Fouling
(Bromley et al. 2002; Chen 1998; Ingmar inorganic precipitants or highly impermeable as
et al. 1999; Belfort et al. 1994). some lms of mineral oils or hydrophobic surfac-
When operating below the critical ux, it is tants. The fouling mechanism depends also on the
possible to observe a linear correlation between membrane process. In electrodialysis fouling is
ux and transmembrane pressure. Above it, fur- caused mainly by the precipitation of polyelectro-
ther increase in transmembrane pressures lead to lytes or sparingly soluble salts such as CaSO4 or
additional layer deposit on the membrane surface, CaCO3. Membrane fouling in electrodialysis
until a point where the deposit fully compensates affects mainly the anion-exchange membranes
the increase in pressure drop. At this stage the because most of the colloidal and macromolecular
liming ux is reached, which represents the max- polyelectrolytes present in natural waters such
imum stationary permeation ux. After operating as humic acids or proteins are negatively charged.
above the critical ux value, decreasing the trans- In ultra- and microltration of biological solu-
membrane pressure will not lead to the previous tions, but also in reverse osmosis of sea water,
ux behavior. biological fouling is a severe problem affecting
However, experimental evidences show that, the economics of the processes. In biomedical
although operating at subcritical ux, gradual applications protein adsorption and protein
fouling develops in membrane materials, and it denaturation at the membrane surface is often
proves to be hydraulically irreversible after a long impairing the performance of the membranes. In
period experimentation (Ognier et al. 2004). membrane distillation, adsorption of molecules
The formation of a gel or cake layer is one can change surface energy and hydrophobic
cause for membrane fouling. Gel or cake layer properties.
formation may be caused by a variety of materials The difference between concentration polari-
including inorganic precipitates such as CaSO4, zation and membrane fouling or scaling is illus-
Fe(OH), and other metal hydroxides; organic trated schematically in Fig. 1.
materials such as proteins, humic acids, and Concentration polarization is a reversible pro-
other macromolecular materials; and biological cess based on diffusion and takes place over a few
components such as microorganisms and products seconds; it can be described adequately by a sim-
of their metabolism (biofouling). Membrane foul- ple mathematical model and easily be controlled
ing may also occur without concentration polari- by the proper process design. Fouling is generally
zation, i.e., a direct transport to the membrane irreversible and the ux decline takes place over
surface. The attachment of the substances to the many minutes, hours, or even days. A constant
membrane surface may be caused by adsorption ux is generally not reached at all. Membrane
due to hydrophobic interactions, van der Waals fouling is more difcult to describe and to control
force attractions, or electrostatic forces. The foul- by experimental means. Membrane fouling is
ing layer itself may be rather porous and thus determined by a variety of different parameters
permeable for aqueous solutions as some including the feed solution constituents and their
Fouling, Fig. 1 Schematic pure water flux

diagram illustrating the
flux
difference between the ux

decline due to concentration
polarization and due to concentration polarization
membrane fouling
membrane fouling
time
Fouling 809
Methods to reduce concentration polarization
Reduce pressure Reduce concentration Reduce solids

at membrane surface in feed
Mixing perpendicular Low concentration Remove concentration Increase mass

to membrane factor at membrane surface transfer of solids
back
Paddle Static Boundary layer Mechanical

mixers mixers skimming scouring
F
High velocity gradient Increase diffusivity Thin Short flow

channels length
Move Move Increase
membrane liquid temperature
Fouling, Fig. 2 Methods to reduce concentration polarization
concentration, membrane material, and the uid elimination of all particulate materials is of great
dynamic system design. Membrane fouling can be importance for the proper function of the
caused by simple precipitation of insoluble mate- membrane.
rials or reversible and irreversible adsorption of Membrane surface modications include the
components at the membrane surface and within introduction of hydrophilic moieties or charged
the membrane pores. groups in the membrane surface by chemical
The means of preventing or at least controlling means or plasma deposition.
membrane fouling effects are as heterogeneous as Increasing the shear rate imposed by the feed
the different material and mechanisms causing the solution on the membrane surface will in many
fouling (Ridgway et al. 1983; Nilson 1998). The cases reduce the membrane fouling. High feed
main procedures to avoid or control concentration ow velocities and the proper module design are
polarization and fouling involve: efcient tools in controlling membrane fouling
(Fig. 3).
Pretreatment of the feed solution When in spite of an adequate membrane and
Membrane surface modications module design the membrane ux is decreasing
Hydrodynamic optimization of the membrane with operation time to an unacceptable low value,
module it is necessary to clean the membrane to restore the
Membrane cleaning with the proper chemical ux in part or completely. Typical cleaning agents
agents are acids and bases, such as HNO3 and NaOH,
complexing agents, enzymes, and detergents.
A list of various methods to reduce concentra- Another very effective method to minimize the
tion polarization is reported in Fig. 2. effects of membrane fouling in microltration is
A pretreatment of the feed solution may backushing (Fig. 4). In backushing, the applied
include chemical precipitation, preltration, pH pressure is reversed and the permeate is pushed
adjustment, chlorination, or carbon adsorption. through the membrane, lifting off fouling material
In some membrane module, design concepts as, that had been precipitated on the feed side mem-
for instance, in hollow ber modules, the brane surface and washing it out of the ltration
810 Fouling
Fouling, Fig. 3 Effect of axial velocity on concentration polarization
Fouling,
Fig. 4 Backushing
method to reduce fouling Suspension Suspension
Backflushing
Jv
With
backflushing
Without
backflushing
t
device. Backushing is done in certain time inter- ultraltration: cause, consequences and control tech-
vals for a couple of seconds. niques. In: Flinn JE (ed) Membrane science and tech-
nology. Plenum Press, New York, pp 4797
Brian PTL (1966) Mass transport in reverse osmosis. In:
Merten U (ed) Desalination by reverse osmosis. MIT
References Press, Cambridge, MA, p 161
Bromley AJ, Holdich RG, Cumming IW (2002) Particulate
Bacchin P (2004) A possible link between critical and fouling of surface microlters with slotted and circular
limiting ux for colloidal systems: consideration of pore geometry. J Membr Sci 196:27
critical deposit formation along a membrane. J Membr Chen V (1998) Performance of partially permeable
Sci 235:111122 microltration membranes under low fouling condi-
Belfort G, Davis RH, Zydney AL (1994) The behaviour of tions. J Membr Sci 147:265
suspensions and macromolecular solution in crossow Cornelissen ER, Harmsen D, de Korte KF, Ruiken CJ, Qin
microltration. J Membr Sci 98:158 J-J, Oo H, Wessels LP (2004) Membrane fouling and
Bian R, Yamamoto K, Watanabe Y (2000) The effect of process performance of forward osmosis membranes
shear rate on controlling the concentration polarization on activated sludge. J Membr Sci 319:158168
and membrane fouling. Desalination 131:225236 Field RW, Wu D, Howell JA, Gupta BB (1995) Critical
Blatt WF, Dravid A, Michaels AS, Nelsen LM (1970) Sol- ux concept for microltration fouling. J Membr Sci
ute polarization and cake formation in membrane 100:259
Fouling due to Calcium Phosphate 811
Hoek EMV, Elimelech M (2003) Cake-enhanced concen- lifetime of metastable amorphous phase is
tration polarization: a new fouling mechanism for salt- strongly governed by the exact solution environ-
rejecting membranes. Environ Sci Technol
24:55815588 ment, with the presence of other additive mole-
Howell J (1995) Subcritical ux operation of cules and ions, pH, ionic strength, and
microltration. J Membr Sci 107:165 temperature affecting the lifetimes. The transfor-
Ingmar HH, Eert V, Gun T, Christian T (1999) The inu- mation of amorphous to crystalline hydroxyapa-
ence of the membrane zeta potential on the critical ux
for crossow microltration of particle suspensions. tite is described by a rst-order rate law and is
J Membr Sci 156:153 only a function of pH of the mediating solution at
Jonsson G, Boesen CE (1984) Polarization phenomena in constant temperature. Crystalline calcium phos-
membrane process. In: Belfort G (ed) Synthetic mem- phate apatite is less soluble in neutral and alkaline
brane processes. Academic, New York
Nilson JL (1998) Fouling of an ultraltration membrane by conditions and preferentially dissolves in
dissolved whey protein concentrate. J Membr Sci acidic pH. F
36:147 A calcium phosphate stability index based on
Ognier S, Wisniewski C, Grasmick A (2004) Membrane the phosphate concentration, temperature, and pH
bioreactor fouling in sub-critical ltration conditions: a
local critical ux concept. J Membr Sci 229:171 (Kubo et al. 1979) helps to predict the threshold
Ridgway HF, Justice C, Kelly A, Olson BH (1983) Micro- level of scaling (Filmtec 1995).
bial fouling of reverse osmosis membranes used in
advanced wastewater treatment technology: chemical, SI pHa pHc
bacteriological and ultrastructural analyses. Appl Envi-
ron Microbiol 45:1066
where pHa is the actual pH of a feed water and pHc
is the critical pH calculated as below:
11:755 logCaH logPO4 2log t

pHc
Fouling due to Calcium Phosphate 0:65
Alice Antony and Pierre Le-Clech where CaH is the calcium hardness as ppm
School of Chemical Engineering, The University CaCO3, PO4 is the phosphate concentration as
of New South Wales, UNESCO Centre for ppm PO4, and t is the temperature in C.
Membrane Science and Technology, Sydney, Phosphate can form a variety of polymeric ions
NSW, Australia and therefore many forms of phosphate salts are
possible to be formed as a scale; however, the
majority of calcium phosphate deposits are
Calcium phosphate scaling occurrence is largely reported to be amorphous in nature (Chesters
reported in reverse osmosis (RO) applications for 2009). During membrane autopsy studies, cal-
the reclamation of wastewater (Chesters 2009; cium phosphate scaling was mostly observed as
Greenberg et al. 2005; Zach-Maor et al. 2008). coprecipitated along with other organic and/or
In municipal wastewater, phosphate levels typi- inorganic constituents (Fig. 1).
cally range between 3 and 20 ppm, presenting a The state at which the calcium phosphate
major operational challenge for systems designed enters the RO system is expected to govern the
to operate at high product recoveries. The crystal- nature of the deposit forming on the membrane
line forms of calcium phosphate are hydroxyapa- surface (Ning and Troyer 2007). After
tite and uorapatite, which are calcium pretreatment, if calcium phosphate is in its
orthophosphates with varying amounts of OH-, dissolved form, i.e., calcium and orthophosphate
Cl-, and F-groups. Formation of hydroxyapatite ions, then crystallization of calcium phosphate is
occurs through autocatalytic solution-mediated more likely to occur on the membrane surface. On
transformation of highly unstable amorphous cal- the other hand, if calcium phosphate is in its
cium phosphate, hydrolyzing rapidly to more sta- colloidal form, colloidal deposits are more likely
ble hydroxyapatite microcrystals in water. The to form on the membrane surface, instead of
812 Fouling in Membranes
1500
Ca
1000
P
500 CO
PCr
Si Cr
Ca
S Cr
0
0 2 4 6 8
keV
Fouling due to Calcium Phosphate, Fig. 1 SEM-EDS images showing calcium phosphate scaling on RO membrane
operated with paper mill efuent (Picture provided by Dr Alice Antony, The University of New South Wales, Australia)
crystallization. In these cases, antiscalants will be Ning RY, Troyer TL (2007) Colloidal fouling of RO mem-
ineffective at their typical dosage levels, and branes following MF/UF in the reclamation of munic-
ipal wastewater. Desalination 208:232237
therefore anticoagulants or dispersants should be Zach-Maor A, Semiat R, Rahardianto A, Cohen Y,
added. Wilson S, Gray SR (2008) Diagnostic analysis of RO
Unlike other scales, mitigating calcium phos- desalting treated wastewater. Desalination
phate scales through antiscalant addition is 230:239247
reported to be difcult (Greenberg et al. 2005;
Ning and Troyer 2007). However, the risk of Further Reading
Antony A, Low JH, Gray S, Childress AE, Le-Clech P,
calcium phosphate precipitation can be minimized Leslie G (2011) Scale formation and control in high
by maintaining the feed water below pH at 6.4 pressure membrane water treatment systems: a review.
(Zach-Maor et al. 2008). Use of appropriate dis- J Membr Sci 383:116
persants is an effective control measure when
calcium phosphate is in the form of nanoparticles
in the feed. The risk of calcium phosphate scaling
can also be managed to a certain extent by reduc- Fouling in Membranes
ing the concentration of orthophosphate, calcium,
aluminum, iron, and uoride at the pretreatment Fouling
stage.
References
Fouling Index
Chesters SP (2009) Innovations in the inhibition and Napoleone DAgostino and Lidietta Giorno
cleaning of reverse osmosis membrane scaling and Institute on Membrane Technology, National
fouling. Desalination 238:2229 Research Council of Italy, ITM-CNR, Rende,
FILMTEC (1995) FilmTec membranes technical manual. Italy
The Dow Chemical, England
Greenberg G, Hasson D, Semiat R (2005) Limits of RO
recovery imposed by calcium phosphate precipitation.
Desalination 183:273288 Synonyms
Kubo S, Takahashi T, Morinaga H, Ueki H (1979) Inhibi-
tion of calcium phosphate scale on heat exchanger: the
relation between laboratory test results and tests on heat Combined Fouling Index (CFI); Membrane Foul-
transfer surfaces. Corrosion 79. Atlanta, ing Index; Modied Fouling Index (MFI); Silt
Houston, Texas Fouling Index (SFI)
Fouling Index 813
Membrane fouling is an inevitable phenomenon heavily dependent on the test membrane perme-
in membrane processes and reduces the efciency ability; not applicable for testing high fouling feed
and the economic viability of the membrane pro- water, e.g., raw water(ASTM recommends that
cesses. All the fouling stages such as pore turbidity should be <1 NTU); not applicable for
blocking and cake layer not only decrease the testing UF permeate, which is increasingly being
permeate ux during a membrane process but used in desalination pretreatment; no linear rela-
they could inuence negatively the permeate tion with colloidal/suspended matter; fouling
quality. Understanding the membrane fouling potential of particles smaller than 0.45 mm is not
means that it is possible to develop a method to measured; and it is not based on any ltration
measure and predict the fouling potential with mechanism.
respect to a membrane. There is a pressing need To overcome these deciencies, the Modied
for a reliable method able to do that. Fouling Index using the same 0.45 mm membrane F
Simple membrane experiments such as Silt lters (MFI0.45) was developed. The MFI0.45 can
Density Index (SDI) and Modied Fouling Index achieve the linear relationship between concentra-
(MFI) are fouling index methods used to evaluate tion and ux decline, but still cannot accurately
the particulate fouling potential of a feed water on predict ux decline. Boerlage et al. (2002) con-
reverse osmosis (RO) and nanoltration (NF) rmed that this is due to the fact that in RO,
membrane process (Salinas-Rodriguez SG et al. fouling is caused by smaller colloids that are not
2015). The utilization of fouling indices enables to retained by the microltration membranes used in
determine, e.g., the pretreatment requirements the MFI. Seeing the above complexities of fouling
without conducting a pilot study which needs mechanisms, this is not unexpected. The MFI0.45
considerable time and expenses. Although these test is similar to the SDI test (Schippers and
indices are the standardized parameters and Verdouw 1980), but the volume is recorded
widely used in engineering practices, they are every 20 s for 20 min. It takes into account that
considered to be unsatisfactory indicators that initially pore blocking occurs, followed by cake/
often fail to reect the true fouling strength of gel ltration, and nally cake/gel blocking and/or
the feed water. enhanced compression occurs. The cake ltration
SDI and MFI are performed using a membrane can be identied by plotting t/V versus V, where t
with 0.45 mm pores. The SDI test, dened by a is the time and V is the volume. The slope in this
standard method instructed in ASTM-D4189-07 region is the fouling index. The obtained MFI0.45
(2014), is a dead-end ltration conducted on the value is corrected for temperature and pressure
feed at 207 psi through a 47 mm diameter and shows a linear relation with colloidal/
microltration membrane with 0.45 mm pores. suspended matter concentration. MFI is derived
The rate of plugging is measured and expressed in Eq. 2, and it is dened as the slope of an inverse
as % ux decline per minute. Fouling index by ow rate (1/Q) versus cumulative volume (V)
SDI is obtained by Eq. 1: curve:

1 ti =tf 1
SDI 100 (1) a MFI V (2)
t Q
In Eq. 1, ti is the initial time interval to collect Predicting the rate of fouling in RO systems based
500 ml of permeate, and tf is the nal time interval on the MFI0.45 is possible, assuming that cake/gel
to collect the same volume of permeate after a ltration is the dominant mechanism. However,
time interval t of 5, 10, and/or 15 min. the predicted rate of fouling turns out to be very
The limitations of the SDI test are well low for MFI of 1 s/L2 (equivalent to SDI15 13).
documented (Alhadidi et al. 2011; Rachman A pressure increase of 1 bar is predicted to occur
et al. 2013; Nahrstedt et al. 2008), and include: no in more than 100 years with RO feed water with
correction for test water temperature; the result is an MFI = 1 s/L2.
814 Fouling Index
Both MFI0.45 and SDI underestimate the foul- due to various fouling mechanisms is attempted
ing observed in practice. by combining MFI values measured using multi-
As the SDI and the MFI0.45 do not include ple test membranes. This method, which is desig-
smaller colloid sizes, a new index using an ultra- nated as the Combined Fouling Index (CFI),
ltration (UF) membrane has been developed. allows considering the contribution of particles,
Boerlage et al. tested this MFI-UF as a function hydrophobic matters, colloids, and organics to
of molecular weight cutoff (MWCO, 1100 kDa) RO/NF fouling at the same time. J.-S. Choi
of the UF membranes and obtained values ranging et al. (2009) used three different membranes: a
from 2000 to 13,000 s/L2 (as compared to MFI0.45 hydrophilic microltration membrane, a hydro-
values of 15 s/L2). Higher values were linked to phobic microltration membrane, and a hydro-
the retention of smaller colloids as well as phobic ultraltration membrane in order to study
cake ltration of the retained particles, although the efciency of different pretreatments. The
a correlation with the MWCO was not apparent. Combined Fouling Index value was obtained
While other membrane characteristics may using Eq. 3:
be partly responsible for those varied results,
it is also important to note that fouling at CFI o1 M1 o2 M2 o3 M3 o4 (3)
such MWCOs is complex and cannot be solely
attributed to particulates. A 13 kDa membrane where M1, M2, and M3 are the MFI values of the
was established to be the best membrane for three membranes and o1, o2, o3, and o4 are
such tests. weighting factors depending on the characteristics
Boerlage et al. used a 13 kDa UF membrane of RO/NF membrane because they are closely
(estimated pore dimension 9 nm) to measure foul- related to fouling mechanisms. The weighting
ing potential and effectiveness of pretreatment factors were calculated from the ux decline rate
and compare the results with the SDI and the of each membrane.
MFI0.45. MFI-UF can be operated in constant CFI values are in good agreement with
ow or pressure mode. The MFI-UF values were nanoltration ux decline because with this test
in fact 4001400 times higher than the MFI0.45 is possible considering the impact of different
due to the smaller particles captured. The MFI-UF foulants.
can also be used to determine the effectiveness of In addition, it is also possible to use more than
pretreatment with regard to reduction of fouling three test membranes to obtain more information
potential. Roorda and van der Graaf (2001) used on the characteristics of feed water.
the MFI-UF to determine the fouling potential of
UF membranes and conrmed the dependence on
membrane type. As a general evaluation, Reiss References
and Taylor compared three parameters used to
investigate fouling: the Silt Density Index (SDI), Alhadidi A, Kemperman AJB, Schippers JC, Wessling M,
the Modied Fouling Index (MFI), and the linear Van Der Meer WGJ (2011) The inuence of membrane
properties on the Silt Density Index. J Membr Sci
correlation of the water mass transfer coefcient
384:205218
(MTC). Three different NF pilot systems were ASTM, D4189-07 (2014) Standard test method for Slit
used with different pretreatments including acti- Density Index (SDI) of water, ASTM International,
vated carbon and MF. No correlation between the West Conshohocken, PA, 2014
Boerlage SFE, Kennedy MD, Dickson M, El-Hodali DEY,
different parameters was obtained, indicating that Schippers JC (2002) The modied fouling index using
the ltration laws on which the models are based ultraltration membranes (MFI-UF): characterisation,
might not be valid for NF. Hence, these parame- ltration mechanisms and proposed reference mem-
ters need to be used with caution. brane. J Membr Sci 197:121
Choi J-S, Hwang T-M, Lee H, Hong S (2009) A systematic
SDI and MFI seem to be difcult if there is
approach to determine the fouling index for
more than one type of fouling mechanisms. a RO/NF membrane process. Desalination 238:
Another approach to estimate fouling potential 117127
Fouling Release 815
Nahrstedt A, Camargo Schmale J (2008) New insights into adhering foulants are more easily removed from
SDI and MFI measurements. Water Sci Technol Water the depositing surfaces at low hydrodynamic
Supply 8:401412
Rachman RM, Ghaffour N, Waly F, Amy GL shear forces due to the weakened interfacial
(2013) Assessment of Silt Density Index (SDI) as foul- bond between the synthetic surface and different
ing propensity parameter in reverse osmosis foultants, promoting adhesive failure. The release
(RO) desalination systems. Desalination Water Treat ability of foulants is proportional to the surface
51:10911103
Roorda JH, van derGraaf JHJM (2001) New parameter for energy and modulus (gE)1/2 (Brady and Singer
monitoring fouling during ultraltration of WWTP 2000). The lower the surface energy and modulus,
efuent. Water Sci Technol 43(10):241248 the stronger reduction of the enthalpy of adsorp-
Salinas-Rodriguez SG, Amy GL, Schippers JC, Kennedy tion and the higher fouling release potential of the
MD (2015) The Modied Fouling Index Ultraltration
constant ux for assessing particulate/colloidal fouling surface. Surface with compositional heteroge-
of RO systems. Desalination 365:7991 neous hydrophobic/hydrophilic domains can also F
Schippers JC, Verdouw J (1980) The Modied Fouling reduce interactions with the bioadhesives. In the
Index, a method of determining the fouling character- design of surfaces to eliminate the adhesion of
istic of water. Desalination 32:137148
various foulants, polysiloxane- and uorinated
polymer-based surfaces are usually hydrophobic
and most effective in reducing adhesion. The
major advantage of polysiloxane-based surfaces
is the synergy between low surface energy
Fouling Release (to minimize the work of adhesion) and low
elastic modulus. The major advantage of
Zhongyi Jiang1, Xueting Zhao1, Jinming Peng2, uorinated polymer-based surfaces lies in hydro-
Yanlei Su1 and Hong Wu1 phobicity and oleophobicity with low surface
1
School of Chemical Engineering and energy.
Technology, Tianjin University, Nankai, Tianjin, Marine non-fouling coatings provide a source
China of inspiration for fouling release membrane.
2
Key Laboratory for Green Chemical Technology Membranes with fouling release property are
of Ministry of Education, School of Chemical essential for preventing fouling growth on the
Engineering and Technology, Tianjin University, surfaces whose antifouling performance relies on
Tianjin, China the low adhesion strength and easy release of
foulants on such a membrane. In recent years,
surface modication of membranes using low sur-
Fouling release is one of the general strategies face energy materials to enhance fouling release
to achieve desired non-fouling surfaces. Any ability is under active studying microltration,
surface immersed in a specic aqueous system ultraltration, nanoltration, reverse osmosis,
(wastewater, seawater, tap water, etc.) is subjected etc. The surface energy and heterogeneities are
to the undesirable settlement and nonspecic introduced in the elucidation of membrane fouling
adsorption of biomolecules (e.g., protein and release mechanism during membrane ltration.
polysaccharide), natural organic matter, hydrocar- The lower surface energy and larger chemical
bon, and microorganism (e.g., bacteria, algae, and heterogeneity correspond to a lower driving
mollusks), known as fouling. Non-fouling sur- force for the adsorption of the foulants, thereby
faces to resist different foultants are vital in a resulting in effective fouling release ability. Sur-
wide range of applications including chemical faces which fulll these requirements are often
separation and purication, drug delivery sys- derived from polysiloxane- and uorine-based
tems, biomaterials, biosensors, and marine coat- polymers. In situ forced surface segregation of
ings. The concept of fouling release initially amphiphilic copolymers and grafting low surface
comes from marine non-biocidal coating. Fouling energy molecules or monomers to already formed
release refers to the feature that the attached and membranes are popular methods to fabricated
816 Fouling Release Membranes
fouling release membrane. Membranes with low

surface energy chains (such as PDMS, poly(eth- Fouling Release Membranes
ylene glycol)-uoropolymers, peruoroalkyl
polymers, peruoropolyether polymers, and Zhongyi Jiang1, Xueting Zhao1, Jinming Peng2,
uorine-containing block)showed a high release Yanlei Su1 and Hong Wu1
1
rate. The release rates of PDMS and School of Chemical Engineering and
peruoroalkane polymer-functionalized mem- Technology, Tianjin University, Nankai, Tianjin,
branes are about ve times higher than those of China
2
hydrophilic surface as a result of the low interac- Key Laboratory for Green Chemical Technology
tion between the low surface energy surfaces and of Ministry of Education, School of Chemical
protein (Thrien-Aubin et al. 2011). There is a Engineering and Technology, Tianjin University,
growing body of evidence that amphiphilic or Tianjin, China
chemically heterogeneity membrane surfaces can
present both low surface energy and high hydra-
tion energy microdomains and show promising Fouling release membrane is a novel kind of
performance against a wide range of foulants. membrane that has the chemical structure and
Membrane prepared from amphiphilic copoly- main feature of nonpolar low surface energy
mers with hydrophilic segments and uorine- microdomains and hydrophilic domains,
containing segments via nonsolvent-induced inhibiting the nonspecic adsorption of a variety
phase separation (NIPS) process through forced of foulants. The low surface energy microdomains
surface segregation showed enhanced oil, protein, on the membrane surface, which can be
and bacteria release ability (Chen et al. 2011). The constructed with silicone- and uorine-based
low surface energy microdomains on the mem- polymers, are used to minimize the intermolecular
brane surface minimize the intermolecular forces forces of interactions (polar and/or hydrogen-
of interactions (polar or and hydrogen-bonding bonding interactions) between the foulants and
interactions) between the foulants and the mem- the membrane surface. The hydrophilic domains
brane surface, remarkably reducing the ux are used to bind water molecules and generate
decline and reducing energy consumption during hydration layer similar with widely popular
the dynamic ltration process. hydrophilic modication methods. Considering
In the application of antifouling membranes for that foulants tend to adsorb on traditional hydro-
water treatment and biomedical application, foul- phobic membrane surfaces driving by the
ing release membranes offer an interesting alter- decrease of the Gibbs free energy of the system,
native to environmentally friendly strategies to the surface energy of membranes is regarded as
facilitate foulants removal, such as hydrocarbon one of the most relevant physicochemical param-
substances, biomolecules, and microorganisms. eters inuencing free energy of the adsorption
process. The reduction of the surface energy
causes a strong reduction of the enthalpy of
adsorption which describes the strength of inter-
References action between the protein and the surface and
thus strongly decreases the driving force for the
Brady RF, Singer IL (2000) Mechanical factors favoring foulants to adsorb to the surface (Worz
release from fouling release coatings. Biofouling
15:7381
et al. 2012). Due to the presence of nonpolar low
Chen WJ, Su YL, Peng JM, Dong Y, Zhao XT, Jiang ZY surface energy microdomains on the membrane
(2011) Engineering a robust, versatile amphiphilic surface, foulants in the feed solution can easily
membrane surface through forced surface segregation release from the membrane surface and diffuse
for ultralow ux-decline. Adv Funct Mater 21:191198
back to the bulk feed solution during the dynamic
Thrien-Aubin H, Chen L, Ober CK (2011) Fouling-
resistant polymer brush coatings. Polymer membrane ltration process. Inspired by marine
52:54195425 fouling release materials, this kind of membrane is
Fouling Release Membranes 817
HFBM segments surface coverage

F HFBM =0 F HFBM =15.4%
agitating agitating agitating
a b c
agitating agitating
agitating
agitating agitating agitating
feed solution oil PHFBM microdomain membrane
Fouling Release Membranes, Fig. 1 Tentative illustration of the ux decline resistant mechanism of membranes
(Chen et al. 2011b)
initially designed to detach the foulants from the are not able to spontaneously segregate onto the
surface at low hydrodynamic shear force, and now polymer-water interface due to the unfavorable
it is applied to achieve low permeation ux thermodynamics, the forced surface segregation
decline and high permeation ux recovery coupled with nonsolvent-induced phase separa-
simultaneously during the membrane ltration tion process is proposed for achieving membrane
process. surfaces with uorine-containing microdomains.
To construct fouling release membrane sur- In the immersion-precipitation step, hydrophilic
face, majority of the existing researches have segments of the copolymers segregate to the
been carried out to reduce the surface free energy polymer-water interface freely and drag the
through uorine-based polymers. In situ forced uorine-containing hydrophobic segments to
surface segregation and post-surface grafting are enrich on membrane surfaces. Both the composi-
the two general surface modication methods tion of hydrophilic and hydrophobic segments
being explored to introduce low surface energy and the volatility of casting solvent can affect the
microdomains onto the membrane surface. In the surface segregation behavior and the fouling
process of the forced surface segregation, amphi- release ability of membrane surfaces (Chen
philic copolymers with hydrophilic segments and et al. 2011a, b). The membrane materials used in
uorine-containing segments are used as addi- the post-surface grafting method must be rich of
tives or polymer membrane materials as that usu- reactive groups. Surface grafting by chemical
ally applied in surface segregation approach. reaction is a facile method for introducing low
Considering that uorine-containing segments surface energy microdomains onto hydrophilic
818 Fractional Free Volume (FFV)
membrane surfaces. Low surface energy

peruoroalkyl or even polysiloxyl molecules Fractional Free Volume (FFV)
with active groups are used as modiers
to improve the fouling release ability of Yuri Yampolskii
membrane (Thrien-Aubin et al. 2011; Zhao A.V. Topchiev Institute of Petrochemical
et al. 2012). Synthesis, Russian Academy of Sciences,
Although hydrophilic surface modication can Moscow, Russia
signicantly enhance the antifouling property of
the membrane, it, unfortunately, often results in a
drastic ux decline during the separation and puri-
cation process. Fouling release membrane is an Fractional free volume (FFV) is an empirical
alternative to traditional hydrophilic membrane dimensionless parameter that characterizes free
which leads to lower permeation ux decline. volume in polymers. FFV can be dened as the
Figure 1 reports a tentative ux decline resistant ratio Vf/Vsp, where Vf (cm3/g) is the free volume
mechanism of fouling release membrane. The andVsp (cm3/g) is the specic volume of the poly-
random presence of nonpolar low surface energy mer, the latter parameter being the reciprocal den-
(typically uorinated) microdomains on the sity r. The value of Vf can be calculated using
hydrophilic membrane surface creates a surface Bondi (1968) formula Vf = Vsp Voc, where Voc
with chemical heterogeneity that is vital to is the occupied volume, and it can be estimated as
constructing a truly fouling release membrane Voc = 1.3 Vw, where Vw is van der Waals volume
surface. At moderate shear rate along membrane of the repeat unit of the polymer. Such an estimate
surface, nonpolar low surface energy gives the FFV values of most polymers in the
microdomains prevent the coalescence, migra- range of 1025 %. However there are examples
tion, and spreading of the hydrophobic oil drop- when FFV is as large as 35 %
lets or limit the accumulation/coalescence of polar (polytrimethylsilyl propyne). In spite of its
biofoulants, thus endowing the membrane surface wide usefulness, such denition of FFV has
with outstanding fouling release potential without many disadvantages and is being often criti-
inducing a ux decline. cized. In the case of glassy polymers, the den-
sity may depend on prior history of the sample
and, to some extent, on the method of determi-
nation. The coefcient 1.3 in calculating Voc is
References estimated from packing density of molecular
crystals at 0 K and cannot adequately describe
Chen WJ, Su YL, Peng JM, Dong YN, Zhao XT, Jiang ZY
polymers at room temperature. In addition, cal-
(2011a) Engineering a robust, versatile amphiphilic
membrane surface through forced surface culation of Vw of a repeat unit as the sum of the
segregation for ultralow ux-decline. Adv Funct increments of van der Waals volumes of atoms
Mater 21:191198 that form this repeat unit can lead to some
Chen WJ, Su YL, Peng JM, Zhao XT, Jiang ZY, Dong YN,
inaccuracies. Since this approach poorly takes
Zhang Y, Liang YP, Liu JZ (2011b) Efcient wastewa-
ter treatment by membranes through constructing tun- into account the dead volume which depends on
able antifouling membrane surfaces. Environ Sci conformations of the chain and is different for
Technol 45:65456552 different gases, quantum chemical corrections
Thrien-Aubin H, Chen L, Ober C (2011) Fouling-resistant
polymer brush coatings. Polymer 52:54195425
are desirable. More reliable results were
Worz A, Berchtold B, Moosmann K, Prucker O, Ruhe reported by Ronova et al. (2003). So FFV
J (2012) Protein-resistant polymer surfaces. J Mater found in this way can be considered as a semi-
Chem 22:1954719561 quantitative parameter suitable for comparison
Zhao XT, Su YL, Chen WJ, Peng JM, Jiang ZY
of series of polymers. Park and Paul (1997)
(2012) Grafting peruoroalkyl groups onto polyacry-
lonitrile membrane surface for improved fouling assumed that FFV can be considered not as a
release property. J Membr Sci 415416:824834 universal property of polymers. Gases with
Fractionation of Fish Protein Hydrolysates 819
different molecular dimensions can feel dif-

ferent parts of the size distribution of free vol- Fractionation of Fish Protein
ume. So an empirical method for calculation of Hydrolysates
occupied volume specic for each gas was pro-
posed. This group contribution method allows Marie-Pierre Belleville
better predictions of the permeability coef- Institut Europen des Membranes (IEM): UMR
cients based on modied FFV. CNRS 5635, Universit de Montpellier II,
Positron annihilation lifetime spectroscopy Montpellier, France
(PALS) also allows an estimation of fractional
free volume as FFV = Nvf, where N is hole
number density (concentration of holes) and vf is Each year the processing industry of marine spe-
the mean volume of the spherical hole with the cies generates large amounts of solid by-products F
radius R in the polymer or vf = 4/3 (pR3). The such as lleting wastes, heads, etc. which are
FFV values found in this manner are usually usually converted to sh meal and oil for feed.
smaller than FFV estimated via Bondi by a factor However, these by-products often contain high
34. It is explained by the fact that a big fraction of concentrations of proteins that can be transformed
size distribution of free volume in polymers con- into peptides through enzymatic hydrolysis. Fish
sists of holes with the radii inaccessible to protein hydrolysates (FPH) possess good nutri-
o-positronium having the size of 1.06 . This is tional properties and biological activities such as
conrmed by the calculations using molecular antioxidative, anticarcinogenic, immunological
dynamics method (Hofmann et al. 2002): extrap- activities, etc. (Chalamaiah et al. 2012). They
olation of FFV to the zero radius of probe results can thus represent marketable and value-added
in FFV values much greater than those found products to be used as food and feed. Although
via PALS. biological properties of peptides are related to
various characteristics, molecular weight (MW),
amino acid sequence, hydrophobicity, charge, and
Cross-References acidobasic character, the MW seems to be one of
the most important ones. In particular peptides
Dynamic Free Volume with MW between 1 and 4 kDa have been identi-
Free Volume Distribution ed as the most bioactive ones (Hsu 2010). The
Free Volume Element pressure-driven membrane processes thus appear
to be good candidates to fractionate and concen-
trate FPHs in view to produce peptide fractions
References with high specic activity. Firstly peptides can be
separated from nonhydrolysed proteins owing to
Bondi A (1968) Physical properties of molecular crystals, UF with high molecular weight cutoff (MWCO)
liquids, and gases. Wiley, New York membranes (~20 kDa). Then peptide hydrolysates
Hofmann D, Heuchel M, Yampolskii Y, Khotimskii V,
Shantarovich V (2002) Free volume distribution
can be fractionated according to their size with UF
in ultrahigh and lower free volume polymers: membranes of intermediate MWCO (48 kDa);
comparison between molecular modeling and the recovery yields are improved if UF is com-
positron lifetime studies. Macromolecules 35: bined with dialtration. Finally the concentration
21292140
Park JY, Paul DR (1997) Correlation and prediction of gas
of selected fractions can be carried out with
permeability of glassy polymer membrane materials via nanoltration (NF) membranes of low MWCO
a modied free volume based group contribution (200300 Da) (Chabeaud et al. 2009; Vandanjon
method. J Membr Sci 125:2339 et al. 2009). Figure 1 shows the molecular weight
Ronova I, Rozhkov E, Alentiev A, Yampolskii Y (2003)
Occupied and accessible volumes in glassy polymers
distribution of different peptide fractions obtained
and their relation with gas permeation parameters. after UF/NF fractionation of a tuna dark muscle
Macromol Theory Simul 12:425439 hydrolysate.
820 Fractionation of Fish Protein Hydrolysates
a
0.3 kDa FPH
0.40 1 kDa
UF permeate
0.35 UF Retentate
0.30
0.25
AU
0.20
0.15 4 kDa
0.10 7 kDa
0.05
0.00
10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00 50.00
Minutes
b
1.80 0.3 kDa
UF permeate
1.60 1 kDa
NF permeate
NF Retentate
1.40
1.20
1.00
AU
0.80
0.60
4 kDa
0.40
0.20
0.00
10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00 50.00
Minutes
Fractionation of Fish Protein Hydrolysates, hydrolysate (FPH) (feed of the UF step) (b) Comparison
Fig. 1 Molecular weight distribution of peptides obtained of permeate and retentate obtained by NF (NP010,
by enzymatic hydrolysis of tuna dark muscle (a) Com- Microdyn Nadir) with the UF permeate (feed of the NF
parison of permeate and retentate obtained by UF com- step) (Superdex-peptide HR 10/300 Eluant: Water/TFA/
bined with dialtration (8 kDa ceramic membrane, (Inside ACN: 70/0.1/30)
CRAMTM, Tami Industries)) with the raw sh protein
References proximate composition, amino acid composition, anti-

oxidant activities and applications: a review. Food
Chabeaud A, Vandanjon L, Bourseau P, Jaouen P, Chem 135:30203038
Chaplain-Derouiniot M, Guerard F (2009) Performances Hsu K-C (2010) Purication of antioxidative peptides pre-
of ultraltration membranes for fractionating a pared from enzymatic hydrolysates of tuna dark muscle
sh protein hydrolysate: application to the rening of by-product. Food Chem 122:4248
bioactive peptidic fractions. Sep Purif Technol Vandanjon L, Grignon M, Courois E, Bourseau P,
66:463471 Jaouen P (2009) Fractionating white sh llet hydroly-
Chalamaiah M, Dinesh kumar B, Hemalatha R, sates by ultraltration and nanoltration. J Food Eng
Jyothirmayi T (2012) Fish protein hydrolysates: 95:3644
Fractionation of Milk by Membrane Operations 821
volumes of feed and retentate in discontinuous

Fractionation of Milk by Membrane mode or ow rates of feed and retentate in contin-
Operations uous mode)) leads to the protein standardization
of both drinking and cheese milks. The use of
Marie-Pierre Belleville such milks permits a better process control of
Institut Europen des Membranes (IEM): UMR continuous cheese-making processes and leads
CNRS 5635, Universit de Montpellier II, to a constant quality of products throughout the
Montpellier, France year. The use of intermediate-concentrated
retentates (1.7 < VRF <5) and high-concentrated
retentates (VRF 57) permits the continuous pro-
duction of numerous cheese varieties (fresh
Milk is a complex colloidal suspension containing unripened cheeses, soft and semihard cheeses) F
many classes of different size components (salts, with high yield and very little whey drainage.
soluble proteins, casein micelles, fat globules, All these cheese-making processes are derived
microorganisms, etc. (see Table 1)) which can be from the famous worldwide process known as
well separated from others according to their MMV process after the investigators Maubois,
size or charge, thanks to membrane techniques Mocquot, and Vassal (Gsan-Guiziou 2010).
(Brans et al. 2004; Pouliot 2008; Gsan-Guiziou Microltration (MF) is an alternative to centri-
2010). fugation for the removal of bacteria and spores
Since the 1980s, ultraltration (UF) has been from skimmed milk in order to extend its shelf life
widely used in dairy industries in order to concen- without applying a time-temperature treatment.
trate the protein fraction of milk without any MF is thus carried out with multichannel ceramic
denaturation. This operation is generally carried membranes with a pore diameter of 1.4 mm at high
out on spiral wound polymer membranes or to a tangential velocity (69 m s1) and low trans-
lesser extent on tubular ceramic membranes with membrane pressure (TMP) (50 kPa). In order to
molecular weight cutoff (MWCO) 1050 kDa and limit the fouling phenomena, a uniform TMP is
at a transmembrane pressure of 200400 kPa. applied all along the membrane, thanks to the
According to the concentration level, different circulation of the permeate cocurrent of the
types of concentrate can be distinguished. UF of retentate (Bactocatch system , AlphaLaval;
skimmed milk with low volume reduction factors Invesys APV) or thanks to the use of ceramic
(VRF up to 1.7 (the ratio between either the membranes with linear hydraulic-resistance gra-
dient (GP Membralox membrane Pall-Exekia;
Isoux membranes Tami-Industries). MF of
Fractionation of Milk by Membrane Operations, skimmed milk on a 0.2 mm ceramic membrane
Table 1 Approximate composition of milk permits the concentration of casein micelles. One
Size (mm) or hundred percent of the caseins are recovered in the
Concentration molecular weight retentate without any denaturation, while the per-
Components (g L1) (Da)
meate which is sterile and free of phage particles
Water 870875
presents a composition close to that of whey. It
Fat 3444 0.1515 mm
can be used to prepare whey protein concentrate
Lactose 4850 342 Da
Proteins 3235
(WPC) and isolate (WPI). Finally MF can be also
Caseins 2528 50500 mm used for the separation of milk fat into small
(micelles) (<3 mm) and large (>5 mm) globules. This oper-
Soluble 57 14.2150 kDa ation is achieved with ceramic membranes with
proteins nominal pore size higher than 2 mm. The permeate
Ashes 89 which is enriched in small globule fraction leads
(mineral and
to the production of cheeses with smoother and
salts)
ner texture.
822 Fractionation of Whey by Membrane Operations
References Fractionation of Whey by Membrane Operations,

Table 1 Approximate composition of sweet and
Brans G, Schron CGPH, van der Sman RGM, Boom RM acid whey
(2004) Membrane fractionation of milk: state of the art Sweet whey Acid whey
and challenges. J Membr Sci 243:263272 Components (%) (%)
Gsan-Guiziou G (2010) Separation technologies in dairy Water 9394 9495
and egg processing. In: Rizvi SSH (ed) Separation,
extraction and concentration processes in the food, Lactose 4.55 3.84.3
beverage and nutraceutical industries. Woodhead Lactic acid Traces Up to 0.8
Food Series, Oxford, pp 341380 Fat Traces Traces
Pouliot Y (2008) Membrane processes in dairy Total proteins 0.81.0 0.81.0
technology from a simple idea to worldwide panacea. Soluble proteins 0.60.65 0.60.65
Int Dairy J 18:735740
Ashes (mineral and 0.50.7 0.50.7
salts)
foaming, water binding, etc.), and physiological

Fractionation of Whey by Membrane functionalities.
Operations Whey protein concentrates (WPC) containing
3585 % protein are prepared by ultraltration
Marie-Pierre Belleville (UF) with 1020 kDa molecular weight cutoff
Institut Europen des Membranes (IEM): UMR (MWCO) membranes. According to the concen-
CNRS 5635, Universit de Montpellier II, tration level, different types of concentrate can be
Montpellier, France distinguished. For 35 % WPC, the volume reduc-
tion factor (VRF) the ratio between either the
volumes of feed and retentate in discontinuous
Whey is a coproduct of cheese-making and casein mode, or ow rates of feed and retentate in con-
manufacture in the dairy industry; it is the liquid tinuous mode is about 47, whereas VRF of
drained from the curd obtained after addition of 1350 are required to prepare 5060 % WPC.
rennet (chymosin) or mineral/organic, or the per- 7585 % WPC are obtained owing UF (VRF
meate obtained during skimmed milk 3035) combined with dialtration in order to
microltration (MF) in view of casein concentra- remove minerals and lactose. Whey protein iso-
tion. Whey contains many classes of different-size lates (WPI) containing more than 90 % of proteins
components (salts, lactose, soluble proteins are generally produced from the microltration
(mainly b-lactoglobulin and a-lactalbumin), permeate of skimmed milk using ion-exchange
residual casein, microorganisms, etc. (see chromatography (IEC).
Table 1)), but its composition depends on the Reverse osmosis (RO) permits the concentra-
cheese-making process. The addition of enzyme tion of whey and its derivatives before dehydra-
or rennet milk leads to sweet whey (pH 6.26.4), tion. Compared to vacuum evaporation, the
whereas acid whey (pH 4.65) is obtained after energy consumption is lower, however, the con-
acidic coagulation of milk. Considered as waste centration level is limited to about 2528 % dry
for years, whey is now widely used as food ingre- matter (VRF 4) due to the increase of osmotic
dients in many goods and infant and dietetic pressure and viscosity and in order to avoid salt
foods. Owing to chromatographic, electrodialytic, precipitation and lactose crystallization.
and membrane techniques which permit whey As the high salt content of whey decreases its
fractionation, it is possible to prepare valuable nutritional value and generates processing dif-
proteins and peptides with high nutritional quali- culties, demineralization of whey is carried out
ties (high biological value (BV), amino acid pro- before drying. Electrodialysis (ED) and/or
le, etc.), interesting physical properties (gelling, ion-exchange chromatography (IEC) are currently
Fractionation of White Wine Proteins by Membrane Operations 823
used to achieve demineralization of whey in the ethanol, polyphenols, and polysaccharides. How-
range 5090 %. However, as the whey to be ever, they are also involved in some aspects that
demineralized should be previously concentrated can impair the acceptance of the nal product by
by RO, these techniques involve high investment consumers, such as haze formation during bottle
and running costs and generate large volumes of storage and wine allergy (Ferreira et al. 2002).
polluting efuents. On the contrary, nanoltration Therefore, several fractionation methods have
(NF) appears to be more attractive even if this pro- been developed to achieve a selective separation
cess leads to a lower level of demineralization. NF of wine proteins.
permits to concentrate and demineralize the whey First some preparative separation methods are
simultaneously at low transmembrane pressure. The often used such as dialysis (3.5 kDa cutoff mem-
process is thus more cost-effective and environmen- branes) to eliminate low molecular weight com-
tally friendly compared to ED and IEC. In addition, pounds, ultraltration (10 kDa cutoff membranes) F
as nutrition value ions such as calcium are more to isolate and concentrate proteins, or protein pre-
retained by the membrane than monovalent ions cipitation by adding high concentration of salts
(such as Cl), the nal products show higher nutri- (ammonium sulfate) or organic solvents (ethanol,
tional and technological properties. methanol, or acetone). Afterwards, several chro-
matographic, electrophoretic, and membrane sepa-
rations are usually combined that results into
enriched protein fractions. A number of different
Fractionation of White Wine Proteins chromatographic techniques, such as ion exchange,
by Membrane Operations gel ltration/size exclusion, hydrophobic interac-
tion, afnity, and chromatofocusing, have been
Johannes De Bruijn developed for the fractionation and purication of
Department of Agroindustry, Food Engineering wine proteins. In addition to chromatographic tech-
Group, Universidad de Concepcion, Chillan, niques, electrophoresis such as capillary electro-
Chile phoresis and capillary gel electrophoresis can be
used to achieve protein fractionation on a prepara-
tive scale (Le Bourse et al. 2010).
Fractionation is a separation process in which a Sequential fractionation of white wines (Fig. 1)
certain amount of a mixture is divided up into a by using a series of ultraltration membranes with
smaller number of quantities or fractions with decreasing molecular weight cutoff ranging from
different molecular composition. Fractionation is 300, 100, 30 to 10 kDa allows to study foam
based on the difference in specic properties stability of champagne wines (Abdallah
between the individual components. Commonly et al. 2010) and protein haze stability of white
equilibrium has to be aimed between the recovery wines (De Bruijn et al. 2011a, b). Ultracel PLC
of a specic compound and the desired purity of membranes of composite regenerated cellulose
each fraction. within a stirred frontal cell or a cross-ow unit
Proteins from white wines differ in molecular should be preferred due to their ultralow protein
weight, ranging from 9 to 72 kDa, isoelectric point binding capacity. Although the purity of protein
with pI between 3 and 9, hydrophobicity, degree fractions improves during membrane fraction-
of glycosylation, and thermal stability. Proteins ation, a decrease of protein stability has been
can exhibit positive effects such as the stabiliza- observed due to the retention of high macromo-
tion of foam in sparkling wines, the interaction lecular species (glycoproteins superior of
with aroma compounds, the protection of wine 100 kDa) by the ultraltration membrane. This
against precipitation of tartaric salts, and the inter- changing composition of the wine matrix makes
actions with other wine compounds such as membrane fractionation a complex process.
824 Free Diffusion
Fractionation of White
Wine Proteins by MWCO Colloids
Membrane Operations,
300 kDa
Fig. 1 Protein
fractionation by High MW macromolecules
ultraltration membranes
100 kDa
Intermediate MW macromolecules
30 kDa
Low MW macromolecules
10 kDa
Low MW compounds
References as a result of the random walk of the diffusing

particles, self-propelled by thermal energy
Abdallah Z, Agui-Bghin V, Abou-Saleh K, Douillard R, (Brownian motion); from the phenomenological
Bliard C (2010) Isolation and analysis of macromolec-
(macroscopic) point of view, such (diffusion) ux,
ular fractions responsible for the surface properties in
native champagne wines. Food Res Int 43:982987 which can be described by Ficks laws, is propor-
De Bruijn J, Martnez-Oyanedel J, Loyola C, Seiter J, tional to the minus gradient of chemical potential
Lobos F, Prez-Arias R (2011a) Fractionation of and goes from regions of higher concentration to
Sauvignon wine macromolecules by ultraltration and
regions of lower concentration of diffusant in
dialtration: impact of protein composition on white
wine haze stability. Int J Food Sci Technol 46:16911698 appropriate (motionless) environment or solvent
De Bruijn J, Martnez-Oyanedel J, Loyola C, Lobos F, till concentration equalizing. From thermodynamic
Seiter J, Prez R (2011b) Impact of white wine macro- point of view, diffusion (mixing) in chemical sys-
molecules on fractionation performance of ultraltra-
tem increases its entropy (a rate of disorder) and
tion membranes. J Int Sci Vigne Vin 45:1814188
Ferreira RB, Piarra-Pereira MA, Monteiro S, Loureiro tends to state with the lowest inner energy.
VB, Teixeira AR (2002) The wine proteins. Trends An example of free diffusion is as follows (see
Food Sci Technol 12:230239 Fig. 1). If we consider innite cylinder of unit cross
Le Bourse D, Jgou S, Conreux A, Villaume S, Jeandet P
section, not with an instantaneous plane source at
(2010) Review of preparative and analytical procedures
for the study of proteins in grape juice and wine. Anal x = 0, but with an initial distribution of concentra-
Chim Acta 667:3342 tion given by c = c0 for x < 0 and c = 0 for
x > 0, at time t = 0, solution of the Ficks second
law in the one-dimensional case (coordinate x) after
several treatment steps and substitutions gives a
Free Diffusion solution in the form (Jost 1960; Cranck 1975)

Karel Friess c0 x
c 1 erf p (1)
Department of Physical Chemistry, University of 2 2 Dt
Chemistry and Technology Prague, Prague,
Czech Republic where erf x is Gauss error function
erf x erf x; erf 1 1 and D is a
diffusion coefcient. Thus, with if we imagine at
Free diffusion is a random nonrestricted walk or t = 0 in cylinder (column) one solution with dif-
shift of atoms, ions, or particles on distances fusing species with its concentration c = c0
higher than those of the interatomic ones. From (at x < 0) and a second solution (pure solvent)
the atomistic point of view, diffusion is considered with c = 0 (at x > 0) given in contact (with sharp
Free NaA Stabilized Sodalite (SOD) 825
Free Diffusion,
Fig. 1 Schematic draw of
free diffusion with two
solutions: a change of
concentration prole in time
given by Eq. 1 from time
t = 0 (left) to t = 1 (right)
interface at x = 0). If such system is kept at stable (2.65 ), and H2 (2.95 ) from gas or liquid
conditions (constant temperature and no vibra- mixtures (Khajavi et al. 2007). In 1969 alumino-
tions from outside which could cause convection), silicate sodalite was rst synthesized using
at times t > 0 a former sharp interface tends to tetramethylammonium as structure directing
vanish due to diffusion between both solutions, agent (SDA) (Baerlocher and Meier 1969). For
and at long times (t = 1) concentration in the the rst time, Bibby and Dale successfully
whole system became identical and is equal to the reported the synthesis of silica sodalite (Bibby
half of the initial concentration c = c0/2. and Dale 1985). They prepared silica sodalite
using a nonaqueous solvent. In particular, the
ethylene glycol was used as solvent and SDA
References and the unit cell composition obtained was
Si12O24 2C2H4(OH)2. Subsequently, it demon-
Cranck J (1975) The mathematics of diffusion. Clarendon strated the possibility to prepare silica sodalite by
Press, Oxford
using SDA 1,3,5-trioxane (Keijsper et al. 1989).
Jost W (1960) Diffusion in solids, liquids, gases.
Academic, New York Recently, Yang and coworkers studied the effect
of water in a solvothermal system composed of
SiO2, NaOH, and ethylene glycol. They discov-
ered a little amount of water did not inuence the
Free NaA Stabilized Sodalite (SOD)
sodalite crystallization (Yang et al. 2007). How-
ever, when the water molar exceeds the silica one,
Catia Algieri
the a-quartz phase was synthesized. The authors
also demonstrated the addition of germanium into
the system accelerated the crystallization of SOD.
Italy
These and other SDAs (Braunbarth et al. 1997;
Werthmann et al. 2000) can be used to synthetize
Sodalite (SOD) is a small-pore (2.8 ) zeolite and sodalite crystals. However, for each SDA the opti-
so good candidate in the separation of small mol- mization of the synthesis conditions is required
ecules such as NH3 (2.55 ), He (2.6 ), H2O above all the temperature that can be varied from
826 Free Protons in Solid Heteropoly Compounds
some days to different weeks (Munzer

et al. 2008). Considering the pore size of this Free Protons in Solid Heteropoly
zeolite, different research groups tried to prepare Compounds
sodalite membranes for the separation of small
molecules. For example, low silica SOD mem- Ubavka B. Mio
branes were prepared by microwave-assisted Faculty of Physical Chemistry, University of
hydrothermal synthesis (Julbe et al. 2003). The Belgrade, Belgrade, Republic of Serbia
prepared membranes exhibited a low He/N2
permselectivity (6.2) owing to the presence of
intercrystalline defects into the zeolite layer.
Recently, defect-free supported hydroxy sodalite The basic structure of hetropoly acid hydrates
membranes were hydrothermally synthesized on (HPAsnH2O) has already been determined by
alumina disk and used in pervaporation (Khajavi Keggin (1934), but the water and hydrated proton
et al. 2010). The membranes exhibited excep- entities ling the channels (secondary structure)
tional acid and base stability within a pH range have not been studied much until the early 1990s.
of 2.713.7 as conrmed by the very high water As heteropoly acids are superionic proton conduc-
selectivity values (1,000,000) and stable uxes. tors at room temperature and good catalysts, it is
clear that their activities will be the function of
hydration degree and dynamics of proton species.
References Therefore, it was necessary to focus on the studies
that identify protonic entities and follow their
Baerlocher C, Meier WM (1969) Synthesis and crystal
dynamics. The studies started on WPA-6, the
structure of tetramethylammonium sodalite. Helv
Chim Acta 52:18531860 most stable hydrate of WPA with well-known
Bibby M, Dale MP (1985) Synthesis of silica-sodalite from structure. Crystallographic results for WPA-6
non-aqueous systems. Nature 317:157158 show that Keggins anions in WPA-6 are stabi-
Braunbarth CM, Behrens P, Felsche J, van de Goor G
lized by dioxonium ion H5O2+, which was dened
(1997) Phase transitions and thermal behaviour of silica
sodalites. Solid State Ion 101103:12731277 as planar and rigid (Brown et al. 1970). On the
Julbe A, Motuzas J, Cazevielle F, Volle G, Guizard C other hand, IR and Raman spectra taken at room
(2003) Synthesis of sodalite/alpha-Al2O3 composite temperature have shown that different protonic
membranes by microwave heating. Sep Purif Technol
species (H5O2+, H3O+, H2O) exist in dynamic
32:139149
Keijsper J, den Ouden CJ, Post MFM (1989) Synthesis of equilibrium (Mio et al. 1991, 2005; Colomban
high-silica sodalite from aqueous systems; a combined and Tomkinson 1997):
experimental and model-based approach. Stud Surf Sci
Catal 49B:237247
Khajavi S, Kapteijn F, Jansen JC (2007) Synthesis of thin
H5 O2 $ H3 O H2 O $ H 2H2 O "
defect-free hydroxy sodalite membranes: new candidate
for activated water permeation. J Membr Sci 299:6372 Spectroscopic results have also shown that
Khajavi S, Jansen JC, Kapteijn F (2010) Performance of dioxonium ion is not rigid, and even more, four
hydroxy sodalite membranes as absolute water selec-
conformation forms of the H5O2+ have been
tive materials under acidic and basic conditions.
J Membr Sci 356:16 found (Kremenovi et al. 2002). Theoretical ab
Munzer S, Caro J, Behrens P (2008) Preparation and char- initio calculations have proved that too (Mio
acterization of sodium-free nanocrystalline sodalite. et al. 2008).
Microporous Mesoporous Mater 110:310
Finally, to conrm the proposed equilibrium,
Werthmann U, Marler B, Gies H (2000) Pyrrolidine silica
sodalite and ethylamine silica sodalite- two new silica incoherent inelastic neutron spectra (IINS) were
sodalite materials synthesized from different solid silica measured (Mio et al. 1994; Fig. 1). Spectra of
sources. Micropor Mesopor Mat 39:549562 WPA-6 show that all bands characteristic of the
Yang X, Albrecht D, Caro J (2007) Solvothermal synthesis
abovementioned protonic species are present in
of germanosilicate-sodalite and silica sodalite: effect of
water, germanium and uoride. Microporous IINS spectrum (Fig. 1a), and results are in accor-
Mesoporous Mater 100:95102 dance with literature data (Jones et al.1989;
Free Protons in Solid Heteropoly Compounds 827
Free Protons in Solid

Heteropoly Compounds,
Fig. 1 IINS spectra of
WPA-6 (a), WPA-1 (b),
WPA-0.2 (c), and WPA-D
(d) (Reprinted with
permission of Elsevier)
F
828 Free Protons in Solid Heteropoly Compounds
Free Protons in Solid 0.1

Heteropoly Compounds,
Fig. 2 Compton prole of
proton. Momentum
distribution of single 0.05
J (Y)
proton. Solid line forWPA-6
and dotted for WPA-0
(Reprinted with Permission
of Elsevier) 0
20 0 20
MOMENTUM (Y)
Kearley et al. 1990; Tomkinson 1992). Changes in Kozhevnikov obtained by 17O NMR spectros-
IINS spectra during the process of calcinations of copy (Ratajczak et al. 2001, pp. 100117).
WPA-6 are also evident (Fig. 1). When dehydra- Kozhevnikov (1998) speaks also about over-
tion is completed, only protons stabilize the coordinated protons in case of HPAs but in tetra-
Keggins structure, as was postulated. The IINS hedral surrounding of oxygen atoms. He has
spectrum of WPA-0.2, calcinated at about 300 C, emphasized the important role of free protons
was characterized by three main bands at 1149 in acid catalysis. A short time later, Essayem also
(0!1), 2287 (0!2), and 3341 cm1 (0!3). The discusses the role of free protons in acid cata-
single strong band in the spectrum of the lytic processes and activity of Cs-WPA salt
dehydrated phase at 1,149 cm1 was assigned to (Cs1.5H1.1PW12O40) (Essayem et al. 2000).
a proton centered within a symmetric (octahedral) Some other authors, as Janik et al. (2004), also
environment of oxygen atoms. The obtained spec- speak about free protons and deprotonated
trum is very similar to that observed for manga- water formed from protons and few oxygen
nese oxide (g-MnO2) where Fillaux identied atoms from Keggins anions. Recently,
free protons/proton gas for the rst time Ph. Colomban continues to emphasize the impor-
(Fillaux et al. 1991). tance of protons in solids (Ph. Colomban 2013). It
Some additional neutron measurements were could be said that nowadays the idea of free
done at Rutherford Appleton Laboratory on the protons in solids is accepted in the scientic
VESUVIO installation with high-energy (few eV) community.
neutrons, with the aim to get more information Ratio of H+/H2O and H3O+/H2O is also very
about proton dynamics. From Fig. 2, it is evident important for conductive characteristics of these
that Compton scattering of protons takes place in the compounds as charge carriers and species forming
attosecond time scale. The free protons present in networks of hydrogen bonds responsible for fast
WPA-6 and those remaining in WPA-0.2 have a proton transport, i.e., high proton conductivity
different single-proton momentum distribution (Table 1). Many things could be analyzed from
(Fig. 2). data concerning proton conductivity: inuence of
Using Born-Oppenheimer approximation degree of hydration, inuence of protonic ratio,
(Chatzidimitriou-Dreismann et al. 2002), kinetic and inuence of big cations, in the case of salt on
energy of chemical covalent bond formations water dissociation. It is evident that cation param-
was calculated in WPA-6 and WPA-0.2: Ekin eters inuence both the host lattice and
(WPA-6) = 174 meV and Ekin rearrangement of protonic entities and their
(WPA-0.2) = 231 meV. Results show that protons dynamics. In the last row are some data for
are more spatially localized in WPA-0 than in KH2PW12O407H2O (KH2WPA) which have
WPA-6 phase on the time scale of been used as PEM in H2/O2 fuel cell with very
1014s. Ratajczak et al. discussed the obtained promising results (Holclajtner-Antunovi
neutron results and compared them with those of et al. 2010).
Free Protons in Solid Heteropoly Compounds 829
Free Protons in Solid Heteropoly Compounds, References

Table 1 Proton conductivity of some HPCs
Compound Conductivity (S/cm1) Brown GN, Noe-Spirlet MR, Busing WR, Levy HA
H3PW12O4029H2O 8.0102 (1970) Dodecatungstophosphoric acid hexahydrate,
(H5O2+)3(PW12O403). The true structure of Keggins
H3PW12O4021H2O 3.4103 pentahydrate from singlecrystal xray and
H3PW12O406H2O 1.25105 neutron diffraction data. Acta Cristalogr B33:
H3PMo12O4029H2O 0.13 10381045
H3PMo12O4014H2O 1.0104 Chatzidimitriou-Dreismann CA, Abdul-Redah T,
H4 SiW12O4030H2O 2.5103 Streffer RMF, Mayers J (2002) Sub-femtosecond
dynamics and dissociation of C-H bonds in solid poly-
Li3PW12O4027H2O 1.25102
styrene and liquid benzene. J Chem Phys 116:
Na2HPW12O4015H2O 1.7104 15111518
Na3PW12O4016H2O 6.6105 Colomban P (2013) Proton and protonic species: the hid-
K3PW12O4010H2O 1.7102 den face of solid state chemistry. How to measure F
Rb3PW12O406H2O 1.7102 H-content in materials? Fuel Cells 13:618
Colomban P, Tomkinson J (1997) Novel forms of hydro-
Cs3PW12O408H2O 2.5102
gen in solids: the ionic proton and the quasi-free
MgHPW12O4027H2O 1.26102 proton. Solid State Ionics 97:123134
MgHPW12O4014H2O 1.0102 Essayem N, Tong YY, Jobic H, Vedrine JC (2000) Charac-
CaHPW12O4021H2O 5.0103 terization of protonic sites in H3PW12O40 and
BaHPW12O407H2O 2.0103 Cs1.9H1.1PW12O40: a solid-state 1H, 2H, 31P
BaHPW12O4013H2O 1.6102 MAS-NMR and inelastic neutron scattering study on
samples prepared under standard reaction conditions.
K3WPA7.6H2O 8.7105
Appl Catal A Gen 194195:109122
K2HWPA7.4H2O 7.2105 Fillaux F, Ouboumour H, Tomkinson J, Yu LT (1991) An
KH2WPA7.3H2O 1.4104 inelastic neutron scattering study of the proton dynam-
KH2WPA7H2O 1.0102 ics in g-MnO2. Chem Phys 149:459469
(t = 80 C; RH = 70 %) Holclajtner-Antunovi I, Mio UB, Todorovi M,
Jovanovi Z, Davidovi M, Bajuk-Bogdanovi D,
Laievi Z (2010) Characterization of potassium salts
of 12-tungstophosphoric acid. Mater Res Bull
45:16791684
The existence of different protonic species in Janik MJ, Davis RJ, Neurock M (2004) A rst principles
HPCsnH2O was conrmed by dielectric mea- analysis of the location and afnity of protons in the
secondary structure of phosphotungstic acid. J Phys
surements from Cole-Cole diagrams (e00 = f( e0)) Chem B 108:1229212300
too. Debyes type relaxation times were deter- Jones DJ, Penfold J, Tomkinson J, Roziere J (1989) Inco-
mined in two frequency regions: from 5 to herent inelastic neutron scattering studies of proton-
500 kHz and in microwave region from 8 to conducting materials: Sn(HPO4)2H2O and HM
(SO4)2H2O, M=Fe. In: part II. The vibrational spec-
12 GHz. The relaxation time of 1.0 1010 was trum of H3O+. J Mol Struct 197:113121
assigned to the reorientation H3O+ ion and/or H2O Kearley GJ, White RP, Forano C, Slade RCT (1990) An
molecules; and relaxation time of 1.6 1 08 s analysis of the vibrational frequencies and amplitudes
was ascribed to the H+ and/or H3O+ reorientation of the H5O2+ion in H4SiW12O40 6H2O (TSA6H2O)
and H4PW12O40 6H2O (TPA6H2O). Spectrochim
and that 1.0 1011 to proton jumps Acta 46A:419424
(Tjapkin et al. 1993). With temperature increase, Keggin JF (1934) Structure and formula of
the relaxation times are changing in function of 12-phosphotungstic acid. Proc R Soc A144:75100
proton entities participated in dehydration Kozhevnikov IV (1998) Catalysis by heteropoly acids and
multicomponent polyoxometalates in liquid-phase
process. reactions. Chem Rev 98:171198
Whatever we call protons in solids, free pro- Kremenovi A, Spasojevide Bir A, Bourree F,
tons, proton gas, or over-coordinated protons, Colomban P, Davidovi M, Mio UB (2002)
it is a special, new state of protons in solids, not Structural modications of dodecatungstophosphoric
acids hexahydrate induced by temperature in the
only in HPCs, forming covalent bonding on a time 10358 K range. In situ highresolution neutron pow-
scale of 1013 to 1015 s, which determines their der diffraction investigation. Solid State Ionics
catalytic and conductive characteristics. 150:431442
830 Free Radical Polymerization
Mio U, Davidovi M, Tjapkin N, Colomban P, Novak A

(1991) Equilibrium of the protonic species in hydrates Free Volume Distribution
of some heteropolyacids at elevated temperatures.
Solid State Ionics 46:103109
Mio UB, Colomban P, Davidovi M, Tomkinson J (1994) Yuri Yampolskii
Inelastic neutron scattering study of protonic species A.V. Topchiev Institute of Petrochemical
during the thermal dehydration of Synthesis, Russian Academy of Sciences,
12-tungstophosphoric hexahydrate. J Mol Struct
326:99107 Moscow, Russia
Mio UB, Todorovi MR, Davidovi M, Colomban P,
Holclajtner-Antunovi I (2005) Heteropoly
compounds from proton conductors to biomedical
agents. Solid State Ionics 176:30053017
Mio UB, Petkovi M, Davidovi M, Peri M, Abdul- Size distribution of free volume in glassy mem-
Redah T (2008) Proton and protonic entities in solid brane materials is accessible to estimation by
heteropoly compounds: an ab initio calculation of the computer simulations (molecular dynamics,
environmental effect on the H5O2+ ion. J Mol Struct MD) and to experimental studies by the probe
885:131138
Ratajczak HR, Barnes AJ, Bielanski A, Lutz HD, methods. In low permeable and highly selective
Muller A, Pope MT (2001) Polyoxometalate Chemistry conventional glassy polymers like polyimides,
from Topology via Self-Assembly to Applications. In: size distribution of free volume elements
Pope MT, Muller A (eds) Kluwer, Dodrecht, (microcavities), according to MD, is located
pp 100117
Tjapkin N, Davidovi M, Colomban P, Mio UB within the range of microcavity radius 15
(1993) Complex dielectric permittivity, bulk and sur- and can be represented by the Gauss function
face conductivity of 12-tungstophosphoric acid hexa- (Heuchel et al. 2004) (see Fig. 1). For polymers
hydrate and its dehydrated forms. Solid State Ionics with greater gas permeability such as poly
61:179185
Tomkinson J (1992) The vibrations of hydrogen bonds. (trimethylsilyl propyne) (PTMSP) or amorphous
Spectrochim Acta 48A:329348 Teon AF, according to MD, it is much wider
and extends up to microcavity radius 1012
(Hofmann et al. 2003) (see Fig. 2). Positron
annihilation lifetime spectroscopy (PALS) also
indicates widely distributed sizes of free vol-
Free Radical Polymerization ume in highly permeable polymers. However,
there are two interpretations of wide size distri-
Surface Modication: Wet Chemical Oxidation bution. According to Shantarovich et al. (1993)
Free Volume
Distribution, Fig. 1 Size
distribution (MD) of free 0.03
volume in 6FDA-3MPD
polyimide accessible for
o-positronium probe
(Adopted from Heuchel 0.02
FFV
et al. (2004))
0.01
0
0 1 2 3 4 5 6 7 8
R,
Free Volume Distribution 831
Free Volume
Distribution, Fig. 2 Size 0.04
distribution (MD) of free
volume in PTMSP
accessible for
o-positronium probe 0.03
(Adopted from Hofmann
et al. (2003))
FFV
0.02
0.01
F
0
0 2 4 6 8 10 12
R,
Free Volume 0.8

Distribution, Fig. 3 Size
distribution (PALS) of free
volume in PTMSP 0.6
(Adopted from Hofmann
et al. (2002))
0.4
f (R)
0.2
2 3 4 5 6 7 8
R,
and Consolati et al. (1996), lifetime distribution References

can be represented by bimodal size distribution
(Fig. 3), while Dlubek (2008) proposed to con- Consolati TG, Genco I, Pegoraro M, Zandeighi L (1996)
J Polym Sci: Part B: Polym Phys 34:357367
sider monomodal size distribution with variable
Dlubek G (2008) Positron annihilation spectroscopy. In:
dispersion. Both approaches are consistent with Seidel A (ed) Encyclopedia of polymer science and
the results of MD simulation. technology. Wiley, Hoboken
Heuchel M, Hofmann D, Pullumbi P (2004) Macromole-
cules 37:201214
Hofmann D, Heuchel M, Yampolskii Y, Khotimskii V,
Cross-References Shantarovich V (2002) Macromolecules 35:21292140
Hofmann D, Entrialgo-Castano M, Lerbret A, Heuchel M,
Yampolskii Y (2003) Macromolecules 36:85288538
Dynamic Free Volume Yampolskii Y, Shantaovich V, Chernyakovskii F,
Fractional Free Volume (FFV) Kornilov A, Plate N (1993) J Appl Polym Sci 47:8592
832 Free Volume Element
differently plasticized poly(vinyl chloride)s: a positron

Free Volume Element lifetime and PVT study. Polymer 44:19211926
Eldrup M, Lightbody D, Sherwood JN (1981) The temper-
ature dependence of positron lifetimes in solid pivalic
Yuri Yampolskii acid. Chem Phys 63:5158
A.V. Topchiev Institute of Petrochemical Goworek T, Ciesielski K, Jasinska B, Wawzyszuk J (2000)
Synthesis, Russian Academy of Sciences, Temperature variations of average o-Ps lifetime in
porous media. Radiat Phys Chem 58:719722
Moscow, Russia Hofmann D, Entrialgo-Castano M, Lerbret A, Heuchel M,
Yampolskii Yu (2003) Molecular modeling investiga-
tion of free volume distributions in stiff chain polymers
Free volume element (FVE) (synonyms: with conventional and ultrahigh free volume: compar-
ison between molecular modeling and positron lifetime
microcavity, hole) is a conditional notion used studies. Macromolecules 36:85288538
for description of free volume in polymers Shrader DM, Jean YC (eds) (1988) Positron and positro-
(mainly in glassy polymers). Thus, in the positron nium chemistry. Elsevier, Amsterdam
annihilation lifetime spectroscopy (PALS) Tao J (1972) Positron annihilation in molecular substances.
J Chem Phys 56:54995510
(Shrader and Jean 1988; Bartos 2000), it is dened Yampolskii Yu (2010) On estimation of concentration of
as a sphere with the diameter R estimated using free volume elements in polymers. Macromolecules
the Tao-Eldrup equation (Tao 1972; Eldrup 43:1018510187
et al. 1981) via the o-positronium lifetimes ti. In
this approach it is conceived that polymer can be
considered as a continuous matrix with uniformly
distributed in it spheres with the size 4/3(pR3). It Free Volume Estimation
was also assumed that cylindrical shape of FVE is
a more realistic representation, so corresponding Yuri Yampolskii
equation for calculation of the volume of FVE was A.V. Topchiev Institute of Petrochemical
proposed (Goworek et al. 2000). The PALS Synthesis, Russian Academy of Sciences,
method permits an estimation of the mean con- Moscow, Russia
centration of FVE or hole number density N -
(Dlubek et al. 2003). In all the polymers, the
N values are limited within relatively narrow Originally, free volume was interpreted (H. Fujita,
range of (28) 1020 cm3 at the glass transition M.H. Cohen and D. Turnbull, J.S. Vrentas and J.L.
temperature Tg (Dlubek et al. 2003; Yampolsakii Duda) as an abstract parameter that ts the equation
2010). Computer modeling of polymers indicates of free volume model. The methods for quantitative
that the use of FVEs dispersed in dense continu- determination of free volume were proposed by
ous matrix is only a convenient oversimplica- S. Sugden and later by A. Bondi (1968). According
tion. In reality the nanostructure of polymers is to the latter, free volume Vf (cc/g) can be estimated
better approximated as a more or less continuous as the difference: Vf = Vsp Vocc, where the specic
form of free volume, though the sizes of FVEs get volume is dened as Vsp = 1/r (r is the polymer
within the size distribution of free volume density), while occupied volume Vocc can be found
according to molecular dynamics (Hofmann as 1.3 Vw, where Vw is the van der Waals volume of
et al. 2003). the repeat unit of the polymer. The Vw value can be
estimated from the increments tabulated for various
groups (Van Krevelen and te Nijenhuis 2009). The
References factor 1.3 is taken from the packing density of
molecular crystal at 0 K (accounting for inaccessible
Bartos J (2000) Positron annihilation spectroscopy of poly- volume). Bondis formula was criticized: the factor
mers and rubbers. In: Meyers RA (ed) Encyclopedia of 1.3 might not be a universal constant, applicable in
analytical chemistry. Wiley, Chichester, p 7968
Dlubek G, Bondarenko V, Pionteck J, Supej M, Wutzler A, all cases; Vw of repeat units cannot be considered as
Krause-Rehberg R (2003) Free volume in two the sum of increments of the groups that form
Free Volume Model 833
it. However, Vf value found using this formula is a Van Krevelen DW, te Nijenhuis K (2009) Properties of
simple and convenient measure of free volume in polymers. Their correlation with chemical structure,
their numerical estimation and prediction from additive
polymers (especially glassy) and it found a wide use. group contributions. Elsevier, Amsterdam
Often, it is more convenient to use a dimensionless Yampolskii Y (2007) Methods for investigation of free
value, fractional free volume FFV dened as Vf/Vsp. volume in polymers. Russ Chem Rev 76:5978
It is widely used in many correlations. In most poly-
mers, FFV is in the range 0.10.25; however, there
are examples where it reaches the value of 0.35.
Since the 1990s, several experimental methods Free Volume Model
(so-called probe methods) appeared and started to
be used for the direct estimation of free volume in Yuri Yampolskii
polymers (Yampolskii 2007). Among these A.V. Topchiev Institute of Petrochemical F
methods (inverse gas chromatography, 129Xe- Synthesis, Russian Academy of Sciences,
NMR, and others), the most accepted and reliable Moscow, Russia
one is the positron annihilation lifetime spectros-
copy (PALS) (Dlubek 2008). It is based on the
measurement of annihilation lifetimes of o-Ps The concept of free volume is extensively used in
(hydrogen-like atom formed by electron and pos- respect of various properties of condensed phases.
itron): the longer its lifetime, the larger is the Regarding nonporous polymeric membranes, it is
microcavity in the polymer where it resides and quite useful in the description of diffusion phe-
annihilates. The method permits estimation of nomena. The model proposed by Cohen and
discreet lifetimes or its distribution (hence, size Turnbull (1959) for transport in liquids and amor-
distribution of microcavities). In conventional phous glasses considers a matrix consisting
glassy and rubbery polymers, the radius of from hard molecules. It gives the following
microcavity R (it is assumed that it has spherical equation for the diffusion coefcient as a
geometry) is in the range 0.20.3 nm; however, in function of average free volume per molecule
highly permeable polymers, it can be as large as vf : D A expgv =vf , where A is the con-
0.7 nm. In conventional polymers, o-Ps lifetime stant, g is the constant that takes into account
distribution and free volume size distribution is possible overlapping of microcavities, and v is
monomodal (Gauss-type distribution). In highly the minimum required volume of microcavity and
permeable polymers, it is, as a rule, bimodal with often is considered as the size of the diffusing
prevailing larger microcavities. By combining molecule. A modied equation was proposed
PALS and PVT data, it is possible to estimate later by Fujita (1961) where instead of vf, the
the concentration of microcavities or hole number fractional free volume (FFV) was used:
density N (Dlubek 2008). In all polymers, the D Ad RTexpBd =FFV. It is accepted that
N values are in the range 10201021 cm3. It is the constants Ad and Bd depend on the nature
possible also to calculate PALS-based FFV = N (size) of the diffusants. These equations are valid
(4/3)pR3. Different probe methods give the radii only when the concentration of penetrants is small
of microcavities in rather good agreement and they do not exert perturbation on polymer
(Yampolskii 2007). matrix (e.g., swelling). In transport of organic
vapors, much more complicated equations were
proposed and used (Vrentas et al. 1977;
References Petropoulos 1994). It is important that these equa-
tions include an activation term exp(-Ej/RT),
Bondi A (1968) Physical properties of molecular crystals, where Ej is the energy needed for the diffusing
liquids and gases. Wiley, New York molecule to jump into the next hole.
Dlubek G (2008) Positron annihilation spectroscopy. In:
Seidel A (ed) Encyclopedia of polymer science and Free volume found using either of
technology. Wiley, Hoboken abovementioned models usually shows good
834 Freundlich Isotherm
correlations with the diffusion coefcients of In this equation, x is the adsorbed quantity, m is
gases in polymers and often with the gas perme- the mass of the adsorbent, P is the partial pressure
ability coefcients. The concept of free volume of the adsorbate in the gas, and k and n are empir-
was strongly enriched by the use of contemporary ical constants for each adsorbent-adsorbate pair at
simulation methods (molecular dynamics, molec- a given temperature.
ular mechanics, and Monte Carlo) and experimen- For adsorption of a liquid, the relation between
tal probe techniques. However, in many cases the adsorbed amount per gram of the solid at
(especially for highly permeable polymers), it is equilibrium Q (mg/g) and the concentration C in
important to supplement the consideration with solution at the equilibrium (mg/l or mg/l) is given
information on connectivity (openness) of free by the following equation:
volume (Willmore et al. 2006).
Q K:Cn (2)
References in which K and n are constants at a given

temperature.
Cohen MH, Turnbull D (1959) J Phys Chem When the Freundlich equation is written in
31:11641169
logarithmic form, a linear relation between log
Fujita H (1961) Fortshcr Hochpolym-Forsch 3:147
Petropoulos JH (1994) In: Paul DR, Yampolskii Yu P (eds) Q and log C is obtained:
Polymeric gas separation membranes. CRC Press,
Boca Raton, p 17 log Q log K n log C (3)
Vrentas JS, Duda JL (1977) J Polym Sci Polym Phys Ed
15:403417; J Appl Polym Sci 21(1977):17151728
Willmore FT, Wang X, Sanchez IC (2006) J Polym Sci Freundlich isotherms are often used to describe
B Polym Phys 44:13851393 adsorption equilibria between a membrane and a
feed solution. This is essential for the description
of phenomena such as membrane fouling (Van der
Bruggen et al. 2002) and breakthrough effects due
Freundlich Isotherm to desorption (McCallum et al. 2008) in aqueous
solutions. Desorption of adsorbed organic com-
Bart Van der Bruggen pounds to the permeate may result in negative
Department of Chemical Engineering, Process apparent rejections when the concentration of the
Engineering for Sustainable Systems (ProcESS), organic compound in the membrane phase is
KU Leuven, Leuven, Belgium larger than the equilibrium permeate concentra-
tion given by the Freundlich isotherm. The per-
meate concentration during this unsteady-state
A Freundlich isotherm is a mathematical expres- ltration period is similar to breakthrough curves
sion for the adsorption equilibrium between a encountered in the description of adsorption col-
uid (liquid or gas) and a solid material. The umns. Kimura et al. (2003) demonstrated that
Freundlich equation is an empirical expression adsorption of hydrophobic compounds, which
representing the isothermal variation of adsorp- can be described by the Freundlich equation,
tion of a liquid or gas onto the surface of a solid might lead to underestimation of the permeate
material, derived by Freundlich (1909) as an concentration (or overestimation of rejection)
empirical relation. due to saturation of the membrane and subsequent
For adsorption of a gas as the adsorbate to a desorption. This is very important in the transla-
solid adsorbent, the following equation represents tion of experimental results to real scale applica-
a Freundlich isotherm: tions. Thus, a fundamental understanding of
adsorption is a prerequisite.
x 1 Adsorption equilibria using the Freundlich iso-
kPn (1)
m therm can also be used for membrane
Fruit Juice Processing by Integrated Membrane Operations 835
characterization. Ge et al. (2006) demonstrated Van der Bruggen B, Braeken L, Vandecasteele C (2002)
that the variation in membrane thickness and Flux decline in nanoltration due to adsorption of
organic compounds. Sep Purif Technol 29(1):2331
pore size can be predicted by studying Freundlich
isotherms. They found that when the variations
were larger, the time of total saturation is delayed,
the loading capacity at the point of breakthrough
decreases, the solute recovery efciency and
ligand utilization efciency decrease, and the Fruit Juice Processing by Integrated
thickness of unused membrane increases. Membrane Operations
The Freundlich isotherm is often used because
of its simplicity and wide applicability. However, Alfredo Cassano
other isotherms can also be used to describe Institute on Membrane Technology, National F
adsorption on membrane surfaces. The most com- Research Council of Italy, ITM-CNR, Rende,
monly applied are the Langmuir isotherm and the Italy
BET equation. The Langmuir model was derived
theoretically by Irving Langmuir in 1916 as a
mathematical expression for monolayer adsorp- Fruit juice clarication, stabilization, depect-
tion onto a solid surface. Although this model inization, and concentration are typical steps
has a theoretical basis, its application is limited where membrane processes such as microltration
to monolayers; in addition, the Langmuir model (MF), ultraltration (UF), reverse osmosis (RO),
does not work well for rough surfaces, which is osmotic distillation (OD), and membrane distilla-
often the case for membranes. The Langmuir tion (MD) can be utilized as alternative technolo-
model was extended in 1938 by Stephen gies to the conventional transformation
Brunauer, Paul Emmett, and Edward Teller, who technologies. Particularly, MF and UF are valid
developed a new theoretical model for multilayer approaches for the clarication of fruit juices as
adsorption that became known as the BET model alternative to the use of ning agents, such as gel-
(as an acronym for Brunauer-Emmett-Teller). atin, diatomaceous earth, bentonite, and silica sol,
Although more complex than the empirical which cause problems of environmental impact due
Freundlich equation, it has the advantage of a to their disposal (Echavarria et al. 2011).
strong fundamental basis, which goes beyond Claried juices coming from MF or UF pro-
the simple approximation in the Freundlich cesses can be commercialized or submitted to a
model. concentration process in order to obtain a product
suitable for the preparation of juices and beverages.
RO, MD, and OD can be used as concentration
References techniques as alternative systems to thermal evap-
oration or cryoconcentration (Jiao et al. 2004).
Freundlich H (1909) Kapillarchemie: eine Darstellung der Integrated membrane operations have been
Chemie der Kolloide und verwandter Gebiete.
suggested for replacement of conventional juice
Akademische Verlagsgesellschaf, Leipzig
Ge D, Shi W, Ren L, Zhang F, Zhang G, Zhang X, Zhang Q processing unit operations for the clarication and
(2006) Variation analysis of afnity-membrane model concentration of different fruit juices as well as for
based on Freundlich adsorption. J Chromatogr the recovery of aroma compounds.
A 1114:4044
A typical example of integrated membrane
Kimura K, Amy G, Drewes J, Watanabe Y (2003) Adsorp-
tion of hydrophobic compounds onto NF/RO mem- system for the clarication and concentration of
branes: an artifact leading to overestimation of fruit juices is depicted in Fig. 1. The fresh juice is
rejection. J Membr Sci 221:89101 claried by UF; the claried juice is
McCallum EA, Hyung H, Do TA, Huang CH, Kim JH
pre-concentrated by RO and then concentrated
(2008) Adsorption, desorption, and steady-state
removal of 17b-estradiol by nanoltration membranes. by OD. Pervaporation (PV) is used to recover
J Membr Sci 319:3843 aroma compounds from the fresh juice.
836 Fuel Cell Applications
pasteurization
pulp
Fresh juice
permeate
(10-11 Brix)
UF RO
preconcentrated concentrated
juice diluted stripping
stripping solution
(25-26 Brix) solution
OD
PV concentrated
aromatic compounds juice
(64-65 Brix)
Fruit Juice Processing by Integrated Membrane Operations, Fig. 1 Integrated membrane process for the produc-
tion of concentrated fruit juice
References generation and transport applications. The main

types of fuel cells are as follows (Table 1):
Echavarria AP, Torras C, Pagan J, Ibarz A (2011) Fruit Different types of fuel cell have been devel-
juice processing and membrane technology applica-
oped for different applications. For example, solid
tion. Food Eng Rev 3:136158
Jiao B, Cassano A, Drioli E (2004) Recent advances on oxide fuel cell (SOFC) and molten carbonate fuel
membrane processes for the concentration of fruit cell (MCFC) systems are used in stationary power
juices: a review. J Food Eng 63:303324 generation or marine applications, whereas poly-
mer electrolyte fuel cell (PEMFC) systems are
chosen for automotive applications.
In the transport sector, the fuel cell has poten-
Fuel Cell Applications tial to replace internal combustion engine, and in
the recent years, the use of fuel cell to power
Rajnish Kaur Calay automobiles such as buses, motorcycles, train,
Institute for Building, Energy and Material boats, and airplanes has been demonstrated by
Technology, University of Troms, Navrik, various companies.
Norway Fuel cell systems can generate power for dis-
Energy Technology Research Group, Narvik tributed electricity (connected to grid) and grid-
University College, Lodve Langes gt 2, Navrik, independent portable and small-scale generators
Norway for emergency and critical areas applications. Par-
ticularly for disaster areas and defense applica-
tions, fuel cell application is very attractive as
Fuel cell is an energy conversion device which fuel cell generates not only electricity, but it pro-
converts chemical energy of a fuel directly into duces water also. Fuel cell plants prove to be very
electricity which can be used for light, heat, and efcient for cogeneration, i.e., combined heat and
shaft power. Therefore in principle fuel cell can power (CHP) for residential buildings or small
replace heat engines for stationary power communities as the waste heat can be utilized for
Fuel Cell Components 837
Fuel Cell Applications, Table 1 Fuel Cell Applications

Type Electrolyte phase Operating temperature Electrolyte Responsible ion
Molten carbonate fuel Liquid electrolyte 500650 C Molten carbonate CO2
3
cell
Phosphoric acid fuel 160200 C Phosphoric acid H+
cell
Alkaline fuel cell 60220 C Liquid (KOH) OH
immobilized
Solid oxide fuel cell Solid electrolyte 600850 C Ceramic YSZ O2
Polymer electrolyte 4080 C Acid-doped solid H+
fuel cell (120180 C HT) polymer
F
heating. The fuel cell system for CHP varies type is dened based on the nature of the electro-
between 2kWe for an individual house and lyte. The six types are alkaline fuel cell (AFC),
100kWe for small communities. molten carbonate fuel cell (MCFC), phosphoric
Portable fuel cell systems supply power in the acid fuel cell (PAFC), proton exchange membrane
same way as batteries and can also be used to or polymer electrolyte membrane (PEM) fuel cell,
charge batteries for small electrical devices. and solid oxide fuel cell (SOFC). The PEM and
Biofuel cells are at research stage, and several SOFC have solid electrolyte, while other three
researchers are investigating their potential for fuel cell types have liquid electrolyte. In a solid
generating electricity combined with applications electrolyte fuel cell, for example, a PEMFC, the
such as waste treatment and metal recovery from catalyst layer is spread either on the electrolyte
waste streams. (as in PEM or SOFC) or on the electrode. In
addition to these basic components, other compo-
nents are gas diffusion layer, electric connections,
current collectors, separator plates, and seals. In
Fuel Cell Components PEM fuel cell bipolar plates made up of metal or
conductive polymer (or carbon composites) have
Rajnish Kaur Calay more than one function. Bipolar plates allow the
Institute for Building, Energy and Material transfer of electrons from the anode side of one
Technology, University of Troms, Navrik, cell to the cathode side of the adjacent cell, pro-
Norway vide structural rigidity of the stack, distribute
Energy Technology Research Group, Narvik reactant gas to each cell within a stack, and dis-
University College, Lodve Langes gt 2, Navrik, tribute reactant gas within the cell in the stack
Norway through proprietary ow eld topology (shape
designs). In some cases, heat management can
be integrated in BPP design which typically
All fuel cells essentially consist of two involves a thermally conductive medium being
electrodes an anode (negative side) and a cath- supplied through the stack to remove heat from
ode (positive side) and an electrolyte to allow the stack. Thermally conductive mediums typi-
charges to move between the two sides of the fuel cally can be air, de-ionized water, or a coolant
cell. Electrons are drawn from the anode to the with high specic heat capacity. Liquid electrolyte
cathode through an external circuit, producing fuel cells will have electrodes immersed in the
direct current electricity. On both sides of the liquid electrolyte and often benet from simplied
electrolyte and two electrodes (cathode, anode) design and also less expensive catalyst materials.
are the catalyst layers. A catalyst lowers the acti- As a result, liquid fuel cells, such as PAFC and
vation energy to undergo a reaction and helps the MCFC, are well-established technologies and
reaction to take place at a faster rate. The fuel cell have been widely used for medium-scale
838 Fuel Cell Membrane
Fuel Cell Components, Table 1 Basic components of The standard electrolyte membrane is usually a
different fuel cells peruorosulfonic acid membrane such as Naon.
Fuel cell Electrolyte Most of the electrolytes alternative to Naon both
type Electrolyte type Catalyst proton conducting and alkaline type, e.g., hydro-
Alkaline Solution of Liquid Platinum carbon type, are signicantly cheaper, and, in
potassium
hydroxide some cases, they are also characterized by smaller
KOH hydrogen and methanol crossover. However, life-
Molten Sodium or Liquid Nickel time characteristics similar to those shown by
carbonate magnesium Naon-type membranes in fuel cells have not yet
carbonates been demonstrated for the alternative membranes.
(CO3)
Concerning with the conductivity, only recently,
Phosphoric Phosphoric Liquid Platinum
acid acid membranes alternative to Naon type have shown
Polymer Polymer Solid Platinum similar levels of performance. One critical aspect
electrolyte electrolyte is related to the fact that the presence of water is a
membrane requirement of low-temperature PEMFCs and
Solid oxide Ceramic Solid Nickel DMFCs for the occurrence of the electrochemical
compound of
metal (like reactions and for the promotion of ionic conduc-
calcium or tivity (water-assisted conductivity mechanism).
zirconium) High ionic conductivities are often associated to
oxides the presence of large water uptake by the mem-
brane, but this property often causes poor
(<250 kW) stationary applications. Table 1 mechanical characteristics such as large swelling
shown below gives an overview of each type of and relevant crossover (especially methanol).
fuel cell, its electrolyte, and the primary catalyst Phosphoric acid-doped polybenzimidazole mem-
used. branes use a Grotthus mechanism of proton trans-
port and do not require water (Wang et al. 1995).
These membranes operate at about 180200 C,
whereas composite peruorosulfonic acid mem-
Fuel Cell Membrane branes or sulfonated hydrocarbon membranes
including inorganic llers such as silica rely on
Antonino Salvatore Arico the water-assisted mechanism, but they can oper-
CNR-ITAE Institute, Messina, Italy ate up to 145 C, under particular conditions
(3 bar abs pressure), due to the enhanced water
retention of the ller (Aric et al. 2003).
Low-temperature fuel cells can be equipped with Composite recast Naon membranes
a proton or anion exchange membrane in alterna- containing inorganic llers have been employed
tive to liquid electrolytes (Aric et al. 2001). The in high-temperature (~150 C) direct alcohol
core of a polymer electrolyte fuel cell (PEMFC) is (Aric et al. 1998) and H2 air fuel cells
the ion exchange membrane. The electrodes (Watanabe et al. 1996). These composite mem-
(anode and cathode) are in intimate contact with branes were originally developed for reduced
the membrane faces. The membrane determines humidication operation in polymer electrolyte
the fuel cell resistance and the fuel permeation fuel cells (Watanabe et al. 1996) due to the
rate, and it inuences the reaction rate. It is well enhanced water retention inside the membrane
known that the use of non-noble metal catalysts is by the effect of the inorganic ller (Aric
possible in the presence of alkaline electrolytes. et al. 1998). A further advantage of composite
Proton-conducting electrolytes have been pre- membranes relies in the barrier effect given by
ferred to alkaline electrolytes for several decades the inorganic ller for methanol crossover (Ren
for practical reasons, e.g., to avoid carbonation. et al. 1996).
Fuel production (pervaporation application) 839
References However, this hydrotreating process results in a

signicant reduction of octane number due to the
Aric AS, Cret P, Antonucci PL, Antonucci V (1998) saturation of olens (Lin et al. 2009).
Comparison of ethanol and methanol oxidation in a
Pervaporation has been given much more atten-
liquid-feed solid polymer electrolyte fuel cell at high
temperature. Electrochem Solid-State Lett 1:6668 tion as a promising and feasible technology for
Arico AS, Srinivasan S, Antonucci V (2001) Fuel Cells desulfurization of gasoline since Grace Davison
1:133 Company reported the pervaporation membrane-
Aric AS, Baglio V, Di Blasi A, Creti P, Antonucci PL,
based S-brane process in 2002 (White et al. 2005).
Antonucci V (2003) Inuence of the acidbase charac-
teristics of inorganic llers on the high temperature The sulfur components can be preferentially
performance of composite membranes in direct metha- removed from the gasoline feed due to its higher
nol fuel cells. Solid State Ionics 161:251265 afnity with and/or quicker diffusivity in the
Ren X, Wilson MS, Gottesfeld S (1996) High performance
direct methanol polymer electrolyte fuel cells.
membrane. Pervaporation desulfurization demon- F
J Electrochem Soc 143:L12 strates distinct advantages over conventional tech-
Wang J, Wasmus S, Savinell RF (1995) Evaluation of nologies such as low energy consumption, simple
Ethanol, 1-Propanol, and 2-Propanol in a Direct Oxi- operation, and little reduction of octane number.
dation Polymer-Electrolyte Fuel Cell. J Electrochem
The pervaporation membrane materials involved
Soc 142:4218
Watanabe M, Uchida H, Seki Y, Emori M, Stonehart P in desulfurization are mainly hydrophobic poly-
(1996) Self-humidifying polymer electrolyte mem- mers including polyurethane (PU), polyurea/ure-
branes for fuel cells. J Electrochem Soc 143:3847 thane (PUU), polyimide (PI), natural rubber and
poly(styrene-co-butadiene) rubber, and
polydimethylsiloxane (PDMS), as well as some
hydrophilic polymers such as polyvinylpyr-
Fuel production (pervaporation rolidone (PVP), cellulose triacetate (CTA), and
application) polyethylene glycol (PEG) (Lin et al. 2009).
Among the variety of possible biofuels from
Wanqin Jin the biorenery, liquid transportation fuels in the
State Key Laboratory of Materials-Oriented form of bioethanol currently account for the vast
Chemical Engineering, Nanjing University of majority of biofuels produced all over the world.
Technology, Nanjing, China A typical corn-to-ethanol production process con-
sists of pretreatment of the starch to sugars with
enzymes, fermentation of the sugars to ethanol
Fuel is any material that stores potential energy in with yeast, followed by distillation and dehydra-
a form that can be practicably released and used as tion processes of ethanol. Downstream from the
heat energy. The conventional fuels refer to chem- fermentor is usually a dilute aqueous solution
ical fuels including fossil fuels and biofuels. Fos- containing about 512 wt.% ethanol. Separation
sil fuels are hydrocarbons, primarily coal and technologies such as pervaporation are applied for
petroleum (liquid petroleum or natural gas), the subsequent ethanol/water separation. On the
formed from the fossilized remains. Biofuel can one hand, hydrophobic pervaporation membranes
be broadly dened as solid, liquid, or gas fuel such as PDMS (Xiangli et al. 2008) and silicalite-
consisting of or derived from biomass. 1 zeolite membranes could be used for selective
Since the heightened concerns for cleaner air recovery of ethanol from fermentation broth. In
and more stringent environmental quality require- addition, the fermentation process can be coupled
ments for sulfur concentration in oil products, with hydrophobic pervaporation membranes for
ultra-deep desulfurization from fuels, particularly in situ removal of the inhibitors (e.g., ethanol
from gasoline, has become a very important sub- product and phenols) from the fermentation
ject in petroleum rening industry. It is well broth to realize the continuous fermentation. On
known that sulfur removal from FCC gasoline the other hand, hydrophilic pervaporation mem-
can be achieved by catalytic hydrodesulfurization. branes including PVA (Zhu et al. 2010) and NaA
840 Fully Fluorinated Polymers
zeolite membranes (Yang et al. 2012) could be The usual surface modications in the litera-
used for dehydration of concentrated ethanol aque- ture are the following: coating (Dickson
ous solution (usually containing 9095 wt.% etha- et al. 1998), self-assembly (Luo et al. 2005),
nol), obtaining the nal bioethanol product. chemical treatment (Van der Bruggen 2009),
plasma treatment (De Geyter et al. 2008), and
surface graft polymerization (Belfer et al. 2004).
References The coating is a physical surface modication
that can be obtained via one of the following
Lin LG, Zhang YZ, Kong Y (2009) Recent advances in mechanisms:
sulfur removal from gasoline by pervaporation. Fuel
88:17991809
White LS, Wormsbecher RF, Lesemann M (2005) Mem- (a) Adsorption/adhesion the functional layer is
brane separation for sulphur reduction. US Patent: only physically xed on the base material;
6,896,796 the binding strength can be increased via mul-
Xiangli FJ, Wei W, Chen YW, Jin WQ, Xu NP (2008) Opti-
tiple interactions between functional groups in
siloxane (PDMS)/ceramic composite pervaporation the macromolecular layer and on the solid
membranes using response surface methodology. surface.
J Membr Sci 311:2333 (b) Interpenetration via mixing between the
Yang ZZ, Liu YM, Yu CL, Gu XH, Xu NP (2012) Ball-
added functional polymer and the base poly-
milled NaA zeolite seeds with submicron size for
growth of NaA zeolite membranes. J Membr Sci mer in an interphase.
392393:1828 (c) Mechanical interpenetration (macroscopic
Zhu YX, Xia SS, Liu GP, Jin WQ (2010) Preparation of entanglement) of an added polymer layer and
ceramic-supported poly(vinyl alcohol)-chitosan com-
the pore structure of a membrane.
posite membranes and their applications in
pervaporation dehydration of organic/water mixtures.
J Membr Sci 349:341348 The main limitation of this method is that it
does not allow to prepare membranes with stable
surfaces because the new materials on the
membrane surface run away easily.
Self-assembly technique permits to obtain
Fully Fluorinated Polymers supramolecular systems. It entails self-assembly
monolayers (SAMs), in which it has the adsorp-
Amorphous Peruoropolymers tion of an active surfactant on a solid surface and
layer-by-layer (LBL) assembly in which the
assembly is obtained by means of alternative
adsorption of linear polycations and polyanions.
Functionalized Membrane The only condition for the use of the self-
assembly method is that the base membrane
Teresa Poerio must be able to adsorb the rst polyelectrolyte
Institute on Membrane Technology, National layer by means of ionic bonds.
Research Council of Italy, ITM-CNR, Rende, The chemical treatment permits to obtain
Italy membrane surface modication with functional
groups rather stable. The modied membrane
retains the original mechanical properties, while
The functionalized membranes are membranes the interfacial properties are changed. The chem-
modied with appropriate functional groups that ical modication is obtained using a chemical
introduce new membrane properties, afnity reaction such as oxidation, substitution, addition,
or catalytic, for improving the selectivity or and hydrolysis.
creating an entirely novel separation function The plasma treatment permits to obtain mem-
(Ulbricht 2006). brane surface with different properties changing
Functionalized Membrane 841
parameter. Plasma gases are CF4, Ar, H2, He, Ne, active species on the membrane surface, permits
N2, O2, and CO2, and the ionized gas can attack the polymerization of monomer from the surface.
the CC, CH, and CS bonds, excluding the According to the different methods used for the
aromatic CC and CH bonds. Surface modica- surface activation, this technique is classied in
tion with CO2 plasma, generally, allows to obtain chemical, radiation, photochemical, and plasma-
surface oxidation and the formation of hydro- induced grafting.
philic surface; N2 plasma makes amine, amide,
imine, and nitrile functional groups on the mem-
References
brane surface; and O2 plasma leads to incorpora-
tion of oxygen containing functional groups such Belfer S, Fainshtain R, Purinson Y, Gilron J, Nystrm M,
as hydroxyl, carbonyl, and carboxyl groups. Mnttri M (2004) Modication of NF membrane
Instead, to improve hydrophobic property of the properties by in situ redox initiated graft polymeriza- F
membrane surface, high degree of uorinated tion with hydrophilic monomers. J Membr Sci
239:564
compounds such as SF6, CF4, and C2F6 are used De Geyter N, Morent R, Gengembre L, Leys C, Payen E,
as plasma gases. Van Vlierberghe S, Schacht E (2008) Increasing the
Surface grafting is a chemical modication hydrophobicity of a PP lm using a helium/CF4 DBD
method. Compared with the physical methods treatment at atmospheric pressure. Plasma Chem
Plasma Process 28:289298
such as coating, the covalent bond between the Dickson JM, Childs RF, Mc Carry BE, Gagnon DR
polymer chain and the membrane surface avoids (1998) Development of a coating technique for the
the desorption phenomena maintaining the chem- internal structure of polypropylene microltration
ical stability of modied surface for long term. membranes. J Membr Sci 16:2536
Luo ML, Zhao JQ, Tang W, Pu CS (2005) Hydrophilic
The grafting method is generally divided in two modication of poly(ethersulfone) ultraltration mem-
classes grafting to and grafting from based on brane surface by self-assembly of TiO2 nanoparticles.
different strategies of membrane functiona- Appl Surf Sci 249:7684
lization. In the grafting to strategy, the polymer Ulbricht M (2006) Advanced functional polymer mem-
branes. Polymer 47:22172262
containing the reactive groups at the end or side Van der Bruggen B (2009) Chemical modication of
chains is covalently coupled to surfaces, while in polyether sulfone nanoltration membranes: a review.
the grafting from strategy, an initiation, due to J Appl Polym Sci 114:630642
G
Galacto-oligosaccharide Membrane product will be formed subsequently. The concept

Reactor of such immobilization was taking place in 1978
(Chibata 1978). The process of immobilization
Chiranjib Bhattacharjee1 and Dwaipayan Sen2 suggests several techniques attributing to adsorp-
1
Department of Chemical Engineering, Jadavpur tion/covalent linking of the biocatalyst on some
University, Kolkata, West Bengal, India inert solid support, encapsulation within some
2
Department of Chemical Engineering, Heritage semipermeable matrix, or insertion of the biocat-
Institute of Technology, Kolkata, West Bengal, alyst within the matrix network (Rios et al. 2004).
India Snamprogetti S.p.A. (Italy) and Sumitomo Chem-
ical (Japan) are some of the companies that
are using enzyme immobilization technique
One of the preliminary issues with galacto- to produce GOS from lactose (Grosov
oligosaccharide (GOS) synthesis in batch reactor et al. 2008). Snamprogetti S.p.A. used K. lactis
is the retention of monosaccharides, glucose, and b-galactosidase entrapped in cellulose triacetate
galactose, within the reaction medium that hinders bers by ber wet spinning, while Sumitomo
the transgalactosylation (Chen et al. 2003) reac- Chemical has developed high-purity immobilized
tion pathway through competitive inhibition of b-galactosidase from A. oryzae covalently
enzyme b-galactosidase. Hence, the removal of bounded to macroporous amphoteric ion-
these monosaccharides from the reaction medium exchange resin of phenol formaldehyde polymer.
reduces the possibilities of enzyme inhibition and, The simplest form of enzymatic membrane reac-
thereby, increases the yield of GOS (Sen tor as described by Czermaka et al. (2004) is
et al. 2011). In accordance to do so, presently where the reaction was carried out in a CSTR
continuous membrane separation process follow- followed by the separation of enzyme using
ing the reaction is one of the upcoming strategies 50 kDa molecular weight cut-off (MWCO) mem-
that are being introduced in the process (Sen brane. This ensures the separation and, thus,
et al. 2011; Das et al. 2011). As a whole the reutilization of enzyme during enzymatic hydro-
technology has been referred as enzymatic mem- lysis of lactose. However, the criticality here is, as
brane reactor (Das et al. 2011). One of typical said before, the retention of monosaccharides
approaches in enzymatic membrane reactor is within the reaction mixture might hinder the pro-
immobilization of the enzyme over the mem- cess by product inhibition even though with the
brane, as shown in Fig. 1, where the substrate subsequent membrane separation. Therefore, the
comes in contact with the membrane and the researchers became concern to develop a reactor,

DOI 10.1007/978-3-662-44324-8
844 Galacto-oligosaccharide Membrane Reactor
E E E E E
E Physical Adsorption
E E E E E E
E
Covalent binding
Entrapment E E
Cross-linking
Galacto-oligosaccharide Membrane Reactor, Fig. 1 Different forms of enzyme immobilization
PEI GA
E GA: Glutaraldehyde;
GA
PEI: Polyethyelenimine; E: Enzyme
Galacto-oligosaccharide Membrane Reactor, Fig. 2 Enzyme immobilization on membrane
where simultaneous production and separation inert support, like membrane, possibility of such
will be carried out. In view of this, Sen changes could be restricted with enhanced
et al. (2012) had suggested a module that will enzyme stabilization (Mateo et al. 2007). Further-
simultaneously behaves as a production and puri- more, to reduce the possible sugar fouling of
cation unit for monosaccharides. According to membrane reactor, Sen et al. (2012) suggested
the membrane reactor, enzyme b-galactosidase one indigenously fabricated module called rotat-
(from B. circulans) was immobilized over a ing disk membrane bioreactor.
0.3 kDa nanoltration thin-lm composite poly-
amide membrane with an immobilization yield of
83 %. According to the protocol elaborated, the
References
membrane surface was primarily hydrolyzed to
Chen CW, Ou-Yang CC, Yeh CW (2003) Synthesis of
generate free amine groups that are required to galactooligosaccharides and transgalactosylation
form a covalent bond with the glutaraldehyde modeling in reverse micelles. Enzym Microbiol
with subsequent covalent linking with the Technol 33:497507
Chibata I (1978) Immobilized enzymes, research and
polyethylenimine and enzyme (Fig. 2). According
development. Wiley, New York
to them, it was found that with the membrane Czermaka P, Ebrahimia M, Graua K, Netza K, Sawatzki G,
reactor, the GOS yield becomes almost double Pfromm PH (2004) Membrane-assisted enzymatic pro-
compared to GOS yield in batch reactor. More- duction of galactosyl-oligosaccharides from lactose in a
continuous process. J Membr Sci 232:8591
over, it has been reported elsewhere that water-
Das R, Sen D, Sarkar A, Bhattacharyya S, Bhattacharjee
soluble enzyme on bind with membrane poten- C (2011) A comparative study on the production of
tially increases the enzyme activity depending on galacto-oligosaccharide from whey permeate in recycle
the membrane organization and topology of the membrane reactor and in enzymatic batch reactor. Ind
Eng Chem Res 50:806816
binding (Snchez et al. 2013). However, the per-
Grosov Z, Rosenberg M, Rebro M (2008) Perspectives
formance of such enzymatic membrane reactor and applications of immobilised b-galactosidase in
depends on the stability of the enzyme protein food industry a review. Czech J Food Sci 26:114
that might undergo some conformational changes Iyer PV, Ananthanarayan L (2008) Enzyme stability and
stabilization aqueous and non-aqueous environment.
because of unfolding of the protein or in other
Process Biochem 43:10191032
word denaturation of the protein (Iyer and Mateo C, Palomo JM, Fernandez-Lorente G, Guisan JM,
Ananthanarayan 2008). With immobilization on Fernandez-Lafuente R (2007) Improvement of enzyme
Galacto-oligosaccharide Production by Membrane Operations 845
activity, stability and selectivity via immobilization

techniques. Enzym Microbiol Technol 40:14511463
Rios GM, Belleville MP, Paolucci D, Sanchez J (2004)
Progress in enzymatic membrane reactors a review.
Snchez JM, Nolan V, Perillo MA (2013) Galactosidase at Galacto-oligosaccharide Production by Membrane
the membranewater interface: a case of an active Operations, Fig. 1 Fouling of membrane
enzyme with non-native conformation. Colloids Surf
B 108:17
Sen D, Sarkar A, Gosling A, Gras SL, Stevens GW, Kent-
ish SE, Bhattacharya PK, Barber AR, Bhattacharjee to the foulant genre that might reduce the perfor-
C (2011) Feasibility study of enzyme immobilization mance of the separation process (Metzger et al.
on polymeric membrane: a case study with enzymati-
cally galacto-oligosaccharides production from lactose.
2007). Separation of galacto-oligosaccharides
J Membr Sci 378:471478 (GOS) from the reaction medium comprising of
Sen D, Sarkar A, Das S, Chowdhury R, Bhattacharjee sugars and other saccharides using membrane
C (2012) Batch hydrolysis and rotating disk membrane separation technique could be one of the tempting G
bioreactor for the production of galacto-
oligosaccharides: a comparative study. Ind Eng Chem
issues in recent pasts that has been explored by
Res 51:1067110681 food scientists. During membrane separation pro-
cess, the primary objective that has to be decided
by the user is the proper selection of membrane
molecular weight cutoff (MWCO) and the mem-
Galacto-oligosaccharide Production brane material itself that has low vulnerability
by Membrane Operations toward the carbohydrate adsorption on the mem-
brane. The molecular weight of GOS varies
Chiranjib Bhattacharjee1 and Dwaipayan Sen2 from di to penta units of monosaccharides
1
Department of Chemical Engineering, Jadavpur (340828 Da). However, formation of more than
University, Kolkata, West Bengal, India tri could be sometimes trivial because of the steric
2
Department of Chemical Engineering, Heritage hindrance that resists attaching the mono-sugar
Institute of Technology, Kolkata, West Bengal, unit with the galactosyl moiety (Chen et al.
India 2002). Hence, the nanoltration with 300 Da
membrane could be a better cutoff in separating
GOS from other sugars present in the reaction
Membrane separation process is one of the mostly mixture (Sen et al. 2011). Moreover, the mem-
coveted applications in different segments of brane material is one of the key issues that control
industries, especially, in food engineering and the yield of GOS. Sen et al. (2011) studied that
efuent treatment sectors. One of the substantial polyethersulfone (PES) and cellulose triacetate
advantages that attributes to the membrane sepa- (CTA) membrane could interact with the mono-
ration process is the separation of the macromol- saccharides produced during enzymatic reaction
ecules (with respect to the membrane) without and results an inhibition of the enzyme restricting
much operational complexity. This ensures the the production of GOS. Aforesaid fouling of the
reusability and high yield of the separated com- membrane because of sugars and polysaccharides
ponents as retentate. However, membrane separa- leads to an implementation of dialtration, where
tion process has some primal intricacies that the reaction mixture is continuously diluted with a
sometimes constraint the application of the pro- volume of water (either in batch (Fikar et al. 2010)
cess, especially, in food processing industries as or continuous process) depending on the volume
prominent technology. One of such complexities concentration factor (VCF) (given by the ratio of
attribute to primarily membrane fouling (Fig. 1) the feed volume and retentate volume after batch)
that reduces the performance as well as the reus- maintained during the dialtration. Grandison
ability of the membrane. Polysaccharides fouling et al. (2002) studied the purication of produced
of the membrane is such a signicant contributor GOS with two different nanoltration membrane
846 Galacto-oligosaccharides
modules, dead-end and cross-ow, in dialtration Sen D, Sarkar A, Gosling A, Gras SL, Stevens GW, Kent-
mode and concluded cross-ow nanoltration ish SE, Bhattacharya PK, Barber AR, Bhattacharjee
C (2011) Feasibility study of enzyme immobilization
yields more puried GOS. In addition, they had on polymeric membrane: a case study with enzymati-
compared their results with different sets of mem- cally galacto-oligosaccharides production from lactose.
brane, where they concluded thin-lm composite J Membr Sci 378:471478
membrane performs better than cellulose triace-
tate membrane in the purication of GOS. One of
the reasons that attributes to an enhanced puri-
cation in cross-ow system is the sweeping action Galacto-oligosaccharides
across the membrane, which removes the fouling
layer of the deposited sugars on the membrane. In Chiranjib Bhattacharjee1 and Dwaipayan Sen2
1
2007, Mellal et al. had used multi-shaft rotating Department of Chemical Engineering, Jadavpur
disk system in order to purify oligosaccharides to University, Kolkata, West Bengal, India
2
depict the applicability of high-sheared membrane Department of Chemical Engineering, Heritage
modules. Das et al. (2011) had shown that Institute of Technology, Kolkata, West Bengal,
nanoltration with rotating disk membrane mod- India
ule, another high-sheared membrane module, will
create enough turbulence over the membrane
enough to disrupt the concentration polarization Galacto-oligosaccharides (GOS), one of the
and yield puried GOS. Therefore, purication of potential prebiotic carbohydrates that are com-
GOS is one of the prime concerns that can be done monly produced from lactose using enzymatic
with a convenient membrane separation scheme hydrolysis, is basically composed of oligo-
attributing to some mechanically modied low galactose with some lactose and glucose
fouling devices. (Niittyen et al. 2007). It is produced by a hydro-
lytic action of b-galactosidase (EC 3.2.1.23) on
lactose through transgalactosylation (Chen
et al. 2003) reaction, and depending on the source
References of the enzyme, it comprises of different synthetic
by-products mixture. A typical application in the
Chen C-S, Hsu C-K, Chiang B-H (2002) Optimization of utilization of GOS is its medicinal usage as prebi-
the enzymic process for manufacturing low-lactose
milk containing oligosaccharides. Process Biochem otic food-based ingredient that aims to enhance
38:801808 the growth as well as metabolic activity of poten-
Das R, Sen D, Sarkar A, Bhattacharyya S, Bhattacharjee C tially health-promoting indigenous bacteria such
(2011) A comparative study on the production of as bidobacteria and lactic acid bacteria (Depient
galacto-oligosaccharide from whey permeate in recycle
membrane reactor and in enzymatic batch reactor. Ind et al. 2008). GOS cant be hydrolyzed in the small
Eng Chem Res 50:806816 intestine attributed as an indigestible component
Fikar M, Kovcs Z, Czermak P (2010) Dynamic optimiza- within the intestine as lactose for lactose-
tion of batch dialtration processes. J Membr Sci intolerant person. However, intestinal bacteria
355:168174
Grandison AS, Goulas AK, Rastall RA (2002) The use of such as Bidobacterium or Lactobacillus species
dead-end and cross-ow nanoltration to purify prebi- can ferment them to a production of short-chain
otic oligosaccharides from reaction mixtures, fatty acids salt acetate (metabolized by muscle),
Songklanakarin. J Sci Technol 24:915928 propionate (metabolized by the liver), and buty-
Mellal M, Ding LH, Jaffrin MY, Delattre C, Michaud P,
Courtois J (2007) Separation and fractionation of rate (metabolized by the colonic epithelium)
oligouronides by shear-enhanced ltration. Sep Sci (Rodriguez-Colinas et al. 2013). GOS is primarily
Technol 42:349361 composed of di-, tri-, tetra-, and pentasaccharides
Metzger U, Le-Clech P, Stuetz RM, Frimmel FH, Chen V with galactose as the monomeric unit (Chen
(2007) Characterisation of polymeric fouling in mem-
brane bioreactors and the effect of different ltration et al. 2003). Figure 1ac is showing the GOS
modes. J Membr Sci 301:180189 structures of different polysaccharides units.
Galacto-oligosaccharides 847
Galacto-oligosaccharides, Fig. 1 Different forms of GOS: (a) di-GOS with galactose-galactose (1!6) linkage, (b) G
galactose-galactose (1!4), (c) galactose-galactose-galactose (1!4;1!4) linkage
Production of GOS primarily depends on pH, Enzyme + Lactose Enzyme Galactosyl + Glucose
substrate (lactose), and temperature of the enzy- +Water + Saccharides
matic medium. However, as aforesaid, source of Galactose + Enzyme GOS + Enzyme
enzyme will be an obvious reason in the produc-
tion of GOS yield as well as the condition of other Galacto-oligosaccharides, Fig. 2 Reaction schematics
parameters. As, for example, b-galactosidase for GOS production from lactose with enzymatic route
derived from Aspergillus oryzae has GOS yield
of 52 % with lactose concentration of 500 gpL,
pH 4.5, and temperature 40 C (Gosling carbohydrate polymers having different molecu-
et al. 2010), while in the enzyme obtained from lar weight with differential solubility. It had been
Kluyveromyces lactis with 400 gpL lactose con- seen that after producing GOS from lactose
centration, 7 pH, and 40 C temperature, the yield through enzymatic route, it can be separated out
was only 24.2 % (w/w) (Gosling et al. 2010). from the mixture of unreacted lactose, glucose,
Other reason that has an effect on the productivity and galactose with 90 % (v/v) ethanol addition at
of GOS is the enzyme inhibition because of the 40 C temperature (Sen et al. 2011).
produced monosaccharides, glucose, and galac-
tose, also termed as product inhibition. Figure 2
References
shows the enzymatic reaction to produce GOS
from lactose. However, the inhibition differs on Chen CW, Ou-Yang CC, Yeh CW (2003) Synthesis of
the source of enzyme. It had been seen that galactooligosaccharides and transgalactosylation
b-galactosidase extracted from Kluyveromyces modeling in reverse micelles. Enzyme Microb Technol
33:497507
lactis is more prone to galactose inhibition com-
Chockchaisawasdee S, Athanasopoulos VI, Niranjan K,
pared to glucose (Chockchaisawasdee et al. Rastall RA (2005) Synthesis of galacto-
2005). Galactose inhibition takes place in a com- oligosaccharide from lactose using beta-galactosidase
petitive way (Gosling et al. 2010), while glucose from Kluyveromyces lactis: studies on batch and con-
tinuous UF membrane-tted bioreactors. Biotechnol
inhibits in both competitive and uncompetitive
Bioeng 89:434443
way (Gosling et al. 2010). One of the critically Depient F, Tzortzis G, Vulevic J, IAnson K, Gibson GR
acclaimed GOS products that is obtained in the (2008) Prebiotic evaluation of a novel galactooligo-
market with substantial demand is Vivinal GOS saccharide mixture produced by the enzymatic activity
of Bidobacterium bidum NCIMB 41171, in healthy
syrup. The reason behind its syrupy form is its low
humans: a randomized, double-blind, crossover,
solubility in water and is practically insoluble in placebo-controlled intervention study. Am J Clin Nutr
ethanol as GOS is a mixture of different 87:785791
848 Galvanic Replication
Gosling A, Stevens GW, Barber AR, Kentish SE, electrodes with polytetrauoroethylene (PTFE)
Gras SL (2010) Recent advances rening galactooligo- materials. Electrodes made up of PTFEs are stable
saccharide production from lactose. Food Chem
121:307318 and have hydrophilic and hydrophobic properties
Niittyen L, Kajander K, Korpela R (2007) Galacto- which are important for providing required con-
oligosaccharides and bowel function. Scand J Food ductivity for electron transport and for removal of
Nutr 51:6266 produced water. For acidic electrolytes such as the
Rodriguez-Colinas B, Kolida S, Baran M, Ballesteros AO,
Rastall RA, Plou FJ (2013) Analysis of fermentation PEM, the principle catalyst used is platinum. An
selectivity of puried galactooligosaccharides by important feature for MEA preparation and man-
in vitro human faecal fermentation. Appl Microbiol ufacture is to ensure that there is good integration
Biotechnol 97:57435752 between the three components, the electrolyte,
Sen D, Gosling A, Stevens GW, Bhattacharya PK, Barber
AR, Kentish SE, Bhattacharjee C, Gras SL catalyst layer, and gas diffusion layer. These com-
(2011) Galactosyl oligosaccharide purication by eth- ponents must be interconnected to form a triphase
anol precipitation. Food Chem 128:773777 point (TPP) which is critical for electrochemical
reactions to take place and subsequently the per-
formance of the fuel cell. The electrolyte and the
Pt catalyst must be in contact to allow the transfer
Galvanic Replication of the mobile ion to/from the reaction surface. The
Pt catalyst must also have good contact with the
Track-Etch Membranes as Tools for Template GDL which will allow the transfer of electrons to
Synthesis of Nano-/Microstructures and Devices and from the reaction site. Finally, the porous
GDL must allow excellent diffusion of the reac-
tant species to the reaction site while it facilitates
product species away from the site. Careful opti-
Gas Diffusion Electrode (GDE) mization of the GDE is therefore essential as it is
directly linked to the performance of the fuel cell
Rajnish Kaur Calay and also the utilization of the Pt catalyst as the Pt
Institute for Building, Energy and Material catalyst that does not fulll the triphase point
Technology, University of Troms, Navrik, criteria will not participate in electrochemical
Norway reactions.
Energy Technology Research Group, Narvik
University College, Lodve Langes gt 2, Navrik,
Norway
Gas Diffusion Layer (GDL)
The membrane electrode assembly is an assem- Rajnish Kaur Calay

bly of an electrolyte and a catalyst layer either Institute for Building, Energy and Material
side of the electrolyte which is then hot pressed Technology, University of Troms, Navrik,
between two gas diffusion layers to form a Norway
single component. Each side of the membrane Energy Technology Research Group, Narvik
then forms an electrode, often called a gas University College, Lodve Langes gt 2, Navrik,
diffusion electrode (GDE) which is a combina- Norway
tion of the catalyst layer and the gas diffusion
layer.
In the early days, sintered electrodes were used The GDL is important part of polymer electro-
which were thick and quite heavy, but since 1970, lyte membrane PEM and direct methanol DM
bonding methods are employed to produce fuel cell. There are two gas diffusion layers
Gas Permeation Unit (GPU) 849
(cathode and anode sides) which as the name

suggests diffuse reactant gases to the catalyst Gas Permeance
layer (of the electrolyte membrane). Other
main functions include removing excess water Permeance
and heat away from the catalyst layer for the
smooth operation and preventing ooding. GDL
is capable of retaining some water to ensure Gas Permeation Unit (GPU)
good conductivity throughout the membrane. It
provides mechanical strength and support to the May-Britt Hgg
catalyst and membrane. GDL is electrically con- Department of Chemical Engineering, Norwegian
ductive; it also provides pathway for current University of Science and Technology,
collection. Thus, the main requirements of Trondheim, Norway
GDL are permeability of gases and water, con-
ductivity, and mechanical strength. Due to these G
requirements, GDL is most commonly made up The intrinsic permeation property of a gas through
of carbon ber-based products, i.e., cloth and a membrane is characterized by the permeability
carbon paper. As carbon cloth is thicker than coefcient P of a gas A, PA. Barrer was early
carbon paper, it has certain advantages com- introduced as a practical unit for this property
pared to carbon paper. Carbon cloth is made and made it easy to compare the suitability of
by weaving carbon bers and therefore more materials to be used for membrane gas separation.
exible and stronger than carbon paper-based Barrer is also dened below.
GDL layers. New generation GDL material Permeance is directly related to the thickness,
combining the benets of carbon cloth and l, of the membrane and as such dened as
paper is also being investigated by some pressure-normalized steady-state ux, P/l. The
researchers. permeance will characterize the gas transport
Due to different considerations for a GDL, through the membrane. Permeance is an important
many factors such as type of carbon used, parameter when comparing the separation suit-
pore formation, and thickness play an important ability of membranes for mixed gases.
role in the fabrication of GDL for a particular A practical unit often used is GPU (gas perme-
application. In most cases GDL consists of ation unit) (see below).
two layers: macroporous diffusion medium Permeability:
(DM) and microporous layer (MPL) typically
200400 mm thick. DM has pore size ranging 1 Barrer 1010 cm3 STP cm cm2 s1 cm Hg1
from 10 to 30 mm, and MPL has pore size of 1, 33 1014 m3 STP m m2 s1 kPa1
0.055 mm and is purposely made hydrophobic 2:99 1015 kmol m m2 s1 kPa1
by treating with small amount of Teon or PTFE
(510 wt.%). The small pore size creates a cap- Permeance:
illary pressure gradient across the membrane for
effective removal of water away from the reac- 1 GPU 106 cm3 STPcm2 s1 cm Hg1
tion sites.
7:6 109 m3 STPm2 s1 kPa1
3:35 103 kmol m2 s1 kPa1
Gas Permeability (For conversions on membrane key parameters,

please refer to: Table according to Journal of
Permeability Membrane Science.)
850 Gas Permeation: Permeability, Permeance, and Separation Factor
(polymers) and the inorganic materials (carbon,

Gas Permeation: Permeability, ceramic, metal (palladium especially)). The gas
Permeance, and Separation Factor separation will take place according to different
transport mechanisms depending on the intrinsic
May-Britt Hgg membrane material properties and the physical
Department of Chemical Engineering, Norwegian properties of the gases.
University of Science and Technology, Dense polymeric materials may be in their
Trondheim, Norway glassy or rubbery state depending on their glass
temperature (Tg) and the operating conditions.
Separation is basically according to solution-
General diffusion. In its glassy state, the diffusion of the
gas component will be governing the selectivity
Detailed information on the topic of gas between the gases, while in its rubbery state, the
separation with membranes can be found in solution of the gas in the membrane material will
numerous textbooks and chapters in membrane be of major importance. As a rule of thumb, the
handbooks only four book references are men- selectivity will be high and permeation low in a
tioned here: Basic Principles of Membrane Tech- glassy, dense membrane, while selectivity will be
nology (Mulder 1996), Materials Science of low and permeation high in a rubbery material.
Membranes for Gas and Vapor Separation The properties of the gas will also be important for
(Yampolskii et al. 2006), Handbook of Membrane the gas separation: an inert gas (like H2, N2, also
Separations (Pabby et al. 2015), and Membrane CH4) has very low critical temperature and will
Technology and Applications (Baker 2012). not easily dissolve in the material. Hydrocarbons,
The use of membranes for gas applications however, and nonideal gases like CO2 have rela-
dates back to the 1980s when the rst commer- tively high critical temperatures and will easily
cial application was made economically feasible dissolve in the material and hence also swell the
by the work of Henis and Tripodi (1980), for membrane. This will naturally affect the separa-
hydrogen recovery from a gas stream at a petro- tion, and the selectivity will go down. It is thus of
chemical plant. The rst membranes developed major importance to evaluate the suitability of the
were usually polysulfone or cellulose acetate, material for the separation of the gas pair in ques-
while todays commercial gas membranes tion and also with respect to process conditions for
include various types of polymers, especially the potential application. The basic equations for
should be mentioned polyimides. The research gas separation are
internationally on membrane materials for gas
applications is comprehensive; however, rela- dci
tively few new materials have made it to com- J i Dij (1)
dx
mercial applications. The potential for the
industrial use of membranes in gas separation
where Ji is the ux of component i [mol/(m2s)],
and purication is huge. The technology is envi-
Dij is the diffusion coefcient [m2/s], and dci/dx is
ronmentally friendly and a great technology for
the concentration gradient for component i over
process intensication. The commercial pro-
the length x [mol/(m3m)] (Ficks law). This con-
cesses where membranes for gas separation are
verts to
applied today are mainly H2 from gas streams,
O2/N2, CO2/H2, CH4/CO2 (natural gas and bio-
gas), vapor/gas separations, dehydration of air, Pi
Ji Dpi when considering gases (2)
and vapor/vapor separations. The potential for l
other gas separations is many.
There are basically two main categories of where Pi D S is the permeability of i (D is
membrane gas separation: the synthetic materials diffusion S solution) and Pi =l is the permeance
Gas Permeation: Permeability, Permeance, and Separation Factor 851
of i through the membrane, while Dpi is the partial promote the permeation of one of the components,
pressure driving force of component i. or the membrane may be a mixed matrix, i.e., with
The ideal separation factor, a* (Eq. 3), may (nano)particles dispersed in the polymeric phase.
be expressed by the ratio of the pure gas perme- The functionalization will then add an extra driv-
abilities for the individual components i and j: ing force to the permeation, and an extra term
will be added to Eq. 2, as shown in Eq. 7 below.
Pi The characteristic of facilitated or carrier-
aij (3) mediated transport is the occurrence of a revers-
Pj
ible chemical reaction or complexation process in
combination with a diffusion process. The total
The separation factor for gases in mixture aij
ux of a permeant A will thus be the sum of both
(Eq. 4) is expressed by the mole fractions of the
the Fickian diffusion (Eq. 1) and the carrier-
components in the feed (x) and the permeate (y),
mediated diffusion as illustrated below. The driv-
respectively: G
ing force in both terms is here expressed as con-
centrations, c:
aij yi =yj
(4)
xi
= xj DA
JA cA, 0 cA, l
l
As a general rule, the driving force of membrane DAC
cAC, 0 cAC, l (7)
gas separation is the difference in partial pressures l
(concentrations) between feed and permeate side.
It is however more correct to say that the driving where the rst term on the right-hand side of the
force is the difference in chemical potential equation is the Fickian diffusion (DA) and the
thereby including the effect of temperature. second term represents the carrier-mediated diffu-
(An additional driving force may be an electric sion (DAC). l is the thickness of the membrane.
potential or a carrier effect for certain types of The other main gas separation mechanism is
membranes.) New membrane materials may com- the molecular sieving. In this group, one typically
bine different transport mechanisms and thereby has the carbon membranes and other inorganic
increase the ux and selectivities. membranes. Molecular sieving depends on the
Transport through a dense polymer may be pore diameter compared to the molecular diameter
considered as an activated process which can be and will usually take place when diameter,
represented by an Arrhenius type of equation. dp < 5 . If the pore is larger, there is the possi-
This implies that temperature may have a large bility of selective surface ow (5 < dp < 12 )
inuence on the transport rate. Equations 5 and 6 where the critical temperature of the gases will
express the temperature dependence of the diffu- inuence the separation. Too large pores may give
sion coefcient and solubility coefcient in the Knudsen diffusion (dp > 20 ) which is not
Eq. P D S: wanted.
The need for optimized membrane separation
D D0 expEd =RT (5) properties for specic gas mixtures has kicked off
an explosive development with respect to tailor-
S S0 expDH s =RT (6) made polymeric membrane. The rst goal for
most membrane material development is to docu-
where Ed is the activation energy for diffusion, ment performance in the economically interesting
DHs is the heat of solution, and D0 and S0 are region as shown in now so well-known Robeson
temperature-independent constants. plot for polymeric membranes, having an upper-
Several ways are investigated in trying to bound trade-off between gas permeability and
increase the separation factor. New polymeric selectivity (Robeson 2008) shown here in
membrane materials may be functionalized to Fig. 1. The gas separation will, however, depend
852 Gas Separation
Gas Permeation:
Permeability,
Permeance,
and Separation Factor,
Fig. 1 Selectivity for the
gas pair CO2/CH4 as
function of CO2
permeability and the
Robeson plot (squares for
CMS (carbon) membranes,
solid circles for TR
(thermally rearranged)
polymers, triangles for FSC
(xed-site-carrier)
membranes, and solid
diamonds for PIMs)
(Adapted from Robeson
2008)
both on efcient separation of the gases in mixture

and the relevant process conditions and, last but Gas Separation
not least, also durability of the material over time.
Membrane gas separation is complex and chal- Mariolino Carta
lenging, and for a full understanding of the topic, School of Chemistry, The University of
competence on material technology and processes Edinburgh, Edinburgh, Scotland, UK
is needed. The four books listed here are excellent
references on the topic.
Gas separation is a widely used technique in
which the objective is the separation of one or
References more gases from a mixture. It is becoming crucial
for several industrial processes such as the treat-
Baker R (2012) Membrane technology and applications, ment of fumes from coal-red plants, in particu-
3rd edn. Wiley, West Sussex lar, aiming for the removal of CO2 to reduce the
Henis JMS, Tripodi MK (1980) A novel approach to gas
separation using composite hollow ber membranes. greenhouse effect. Growing interest is also given
Sep Sci Technol 15:10591068 to other applications such as the separation and
Mulder M (1996) Principles of membrane technology, purication of commercially important gases
2nd edn. Kluwer, Dordrecht such as H2, CH4, and O2 from natural gas.
Pabby A, Rizvi S, Sastre AM (eds) (2015) Handbook of
membrane separations, 2nd edn. CRC Press, Boca The most common methods to perform gas sep-
Raton aration are:
Robeson LM (2008) The upper bound revisited. J Membr
Sci 320(12):390400 1. Separation with solvent/sorbents
Yampolskii Y, Pinnau I, Freeman B (eds) (2006) Material
science of membranes for gas and vapor separation. 2. Separation by cryogenic distillation
Wiley, Chichester 3. Separation with membranes
Gas Separation 853
Separation with Solvent/Sorbents for running the system, especially for the regen-
eration of the vessel (Fig. 1).
The separation with solvent/sorbents is based on
the afnity of the gas toward a specic sorbent
such as zeolites, alumina, or activated carbon or a Separation by Cryogenic Distillation
solvent, for instance, MEA (methanolamine). The
most illustrative example of this technology is Cryogenic distillation is based on the fact that in a
represented by the pressure swing adsorption mixture of gases, they all have different boiling
(PSA) method. The separation occurs when the points and they could be separated by increasing/
gas mixture comes in contact with the sorbent/ decreasing the temperature and pressure of the
solvent in a vessel which is then pressurized. system in which they are stored, so that they can
The gas with the highest afnity for the adsorbent be divided into their single components. The gas
is trapped, whereas the others pass through the mixture is cooled down to a low temperature
system. The regeneration of the vessel is carried (typically < 50 C). Once in the liquid form, G
out by returning it to atmospheric pressure or by the components of the gas can be directed in a
increasing the temperature, with release of the distillation column, and through a series of com-
trapped gas. The main advantage of this tech- pression, cooling, and expansion steps, they can
nique is the high purity of the separated gas; the be distributed to different channels, depending on
disadvantage consists in the high energy required their boiling points (Fig. 2).
Gas Separation,
Fig. 1 Schematic
representation of pressure
swing adsorption method
(PSA)
Gas Separation,
Fig. 2 Schematic
representation of cryogenic
distillation method
854 Gas Separation
It is a widely used technique for streams that Separation with Membranes

already have a high concentration of desired gas
(typically >90 %), but it is not very appropriate Separation of gases with membranes relies on the
for more dilute gas streams. different afnities of one or more gases toward the
The main advantage of the cryogenic gas sep- membrane material, causing one gas to permeate
aration is that it enables direct production of liquid faster (or slower) than others. It is one of the
gas, which is often very useful for certain trans- fastest growing eld for gas separation tech-
port options, such as transport by ship. niques, especially due to the high variety of mate-
A major disadvantage is connected with the rials which the membrane could be composed of,
high amount of energy required for the refrigera- including microporous organic polymers, zeo-
tion especially for dilute gas streams. lites, ceramic, and metal-containing materials
(for a more in-depth reading, see Yampolskii and
Freeman (Yampolskii et al. 2010)).
The gas mixture is directed into a vessel and put
in contact to the membrane material which is at the
interface with another vessel (Fig. 3). The mixture is
allowed to diffuse into the second vessel under a
pressure gradient which promotes the mass trans-
port through the membrane separating the retentate
(slower gas) from the permeate (faster gas).
The use of membranes for gas separation offers
several benets, probably the most valuable is the
high cost efciency (both for the mechanical sim-
plicity of the system and for low-energy regener-
ation). In fact, they do not require thermal
Gas Separation, Fig. 3 Schematic representation of regeneration, a phase change, or active moving
membranes for gas separation parts in their operation.
Gas Separation,
Fig. 4 An example of
Robeson plot, in this case
O2/N2 (Carta et al. 2013).
The black line represents
the 1991 (Robeson 1991)
upper bound, whereas the
red line is the current (2008)
upper bound (Robeson
2008)
Gas Separation by Membrane Operations 855
Probably the greatest limitation of mem-

branes for gas separation is derived from their Gas Separation by Membrane
trade-off relationship between permeability and Operations
selectivity for a required gas component. This
means that high permeable membranes have low Mariolino Carta1 and Paola Bernardo2
1
selectivity, requiring several run for a good sep- School of Chemistry, The University of
aration, and highly selective membranes have Edinburgh, Edinburgh, Scotland, UK
2
low permeability, meaning long operational Institute on Membrane Technology, National
times. This trade-off was well addressed by Research Council of Italy, ITM-CNR, Rende,
Robeson in two well-known articles (Robeson Italy
1991, 2008) in which he studied the gas separa-
tion performance of several membrane-forming
materials in terms of permeability of a particular The separation of mixtures of gases and vapors is
species (PA) and selectivity toward one compo- required in manufacturing processes across vari- G
nent of a gas pair (aA/B = PA/PB), organizing the ous industries. In the last years, membrane sys-
data in double logarithmic plots for a series of tems are gaining a larger acceptance in industry
commercially selected important gas pairs such for gas separation and are recognized as a cost-
as H2/CH4, H2/CO2, and O2/N2. He conrmed efcient separation able to compete with consoli-
that highly selective membranes generally dated processes such as pressure swing absorption
exhibit low permeability and vice versa. The and cryogenic distillation (Bernardo et al. 2009;
most important outcome of this study is Sanders et al. 2013). Membrane processes have
represented by the so-called Robeson upper several advantages over conventional separation
bound, an empirical line which is drawn for techniques (e.g., distillation, extraction, absorp-
every gas pair plot that is meant to dene how tion, and adsorption), including modularity and
good a material for gas separation is. In Fig. 4, compactness, operational exibility, and no need
there is a typical example (Carta et al. 2013) in for energy-intensive phase changes or potentially
which the red line represents the 2008 upper expensive adsorbents and/or difcult to handle
bound for the gas pair O2/N2. Supposedly, if solvents. The features of membrane operations
we plot the selectivity aA/B versus permeability allow implementing the process intensication
PA for a new membrane and the data point fall strategy in different production cycles. Their ver-
close or go over the upper bound, it is widely satility represents a decisive factor to impose
accepted that the material has an excellent com- membrane processes in most gas separation elds.
promise between P (rate of separation) and a The rst membrane units were installed in
(goodness of separation). ammonia plants for the separation of hydrogen
from nitrogen more than 30 years ago. Today,
the production of nitrogen from air is the largest
application of membrane systems, owing to the
References
demand for nitrogen to inert fuel tanks, also
Carta M, Malpass-Evans R, Croad M, Rogan Y, aboard aircrafts, and for blanketing chemical and
Jansen JC, Bernardo P, Bazzarelli F, McKeown NB liqueed gas shipments. Membrane systems are
(2013) An efcient polymer molecular sieve for also applied to enrich oxygen for medical uses, for
membrane gas separations. Science 339(6117):
303307
hydrogen recovery and purication in reneries,
Robeson LM (1991) Correlation of separation factor ver- for air and gas dehydration, and for ratio adjust-
sus permeability for polymeric membranes. J Membr ment of gas mixtures. Natural gas processing rep-
Sci 62(2):165185 resents an important emerging application eld
Yampolskii Y, Freeman B (eds) (2010) Membrane gas
(Baker and Lokhandwala 2008). The relatively
separation. Wiley, Chichester, UK, 370 pp
Robeson LM (2008) The upper bound revisited. J Membr low volume ow and the relatively high inlet
Sci 320(1+2):390400 carbon dioxide content are strong drivers for the
856 Gas Separation by Membrane Operations
implementation of the membrane technology in glassy polymers, increases with the critical vol-
the biogas upgrading that it is at a developing ume of the penetrant (Matteucci et al. 2006).
stage (Makaruk et al. 2010). The challenging ole- These materials are applied to the separation of
n/parafn separation, not yet commercial, is organic vapors from non-condensable gases,
attracting a lot of interest from the scientic com- treating petrochemical vent and process streams
munity (Rungta et al. 2013). to recover valuable feedstocks (Baker 1999).
Membrane separation allows recovering and An interesting new concept is the use of water-
recycling valuable compounds, such as hydrogen swollen thin lm composite membranes for bio-
and light hydrocarbons (ethylene, propylene, and gas purication, taking advantage of the large
LPG), present in different off-gas streams (Baker difference in solubility in water to become selec-
et al. 1998). Polymeric membranes, cheap and tive for CO2 (Krszov et al. 2012).
with an easy processability, are typically used in Facilitated transport membranes contain carrier
the commercially available membrane system for agents that can react reversibly with the target gas
gas separation (Yampolskii 2012). Commercial component. Therefore, the reaction in the mem-
modules employ composite membranes (Pinnau brane creates another transport mechanism, in
et al. 1988), mainly in the form of compact hollow addition to the simple solutiondiffusion mecha-
bers. These membranes typically operate the nism (Huang et al. 2008). However, carrier poison-
separation based on a solution-diffusion transport ing and short life span of the polymeric membranes
mechanism: sorption of the permeant into the are typically reported (Rungta et al. 2013). Ionic
membrane, permeation by diffusion through the liquids were considered as additives for facilitated
membrane, and desorption at the low-pressure transport membranes. Indeed, their negligible
side of the membrane. vapor pressure avoids solvent losses by evapora-
The experimentally observed upper bound, tion, providing stability to the metallic cation
based on various polymeric membranes, was dissolved inside, and acting as a medium for facil-
reported by Robeson in 1991 and then updated itated transport with mobile carrier (Fallanza
in 2008 (Robeson 1991, 2008), thanks to the et al. 2013). Ionic liquid gel membranes based on
efforts to improve the gas separation performance conventional polymers (Jansen et al. 2011) or on
of ultrahigh free volume and peruoropolymers. polymer ionic liquids (Bara et al. 2008) were pro-
Glassy polymers are chosen for their size- posed to increase the stability compared to
selective behavior (e.g., in O2/N2 or H2 separa- supported liquid membranes.
tions). However, when applied to mixtures and/or The key for new applications of membranes in
at high gas activities, these materials are prone to challenging and harsh environments (e.g., petro-
plasticization, which causes swelling of the poly- chemistry) is the development of new tough, high-
mer matrix and results in a higher permeability performance materials. In the eld of inorganic
coupled with a loss of selectivity. Strategies to membranes, metal organic frameworks were
overcome plasticization include thermal curing recently considered for preparing membranes to
and chemical cross-linking, which reduce the poly- be applied to the olen/parafn separation (Bux
mer free volume (Wind et al. 2002). The addition of et al. 2011) or as additive to a polymer matrix
nanollers to a polymer matrix represents an inter- (Bushell et al. 2013).
esting solution to overcome the trade-off of the High free volume polymers have been investi-
polymeric membranes and the inherent brittleness gated as gas separation membranes, combining
issues of inorganic membranes (Goh et al. 2011). their ease of processing and mechanical stability
Rubbery polymers, instead, present a with the potential to surpass the polymeric upper
solubility-controlled permeation and preferen- bound for different gas pairs (Budd and
tially allow the permeation of large gas or vapor McKeown 2010). Novel PIMs, characterized by
molecules in a gaseous mixture containing also a signicant shape persistence, were developed,
smaller molecules (Grinevich et al. 2011). Their showing interesting performance for the O2/N2
permeability, much higher than in conventional separation (Carta et al. 2013).
Gas Separation Membranes: High-Temperature Glass Joining and Sealing 857
Properly designed hybrid processes, combin- Huang J, Zou J, Ho WSW (2008) Carbon dioxide capture
ing a membrane system with a conventional one using a CO2-selective facilitated transport membrane.
Ind Eng Chem Res 47(4):12611267
(e.g., PSA or absorption), represent technically Jansen JC, Friess K, Clarizia G, Schauer J, Izk P (2011)
and economically viable solutions, able to reduce High ionic liquid content polymeric gel membranes:
energy consumption and total costs (Esteves and preparation and performance. Macromolecules 44:3945
Mota 2007). Krszov M, Vejraka J, Vesel V, Friess K, Randov A,
Hejtmnek V, Brabec L, Izk P (2012) A water-swollen
thin lm composite membrane for effective upgrading of
raw biogas by methane. Sep Purif Technol 89:212216
References Makaruk A, Miltner M, Harasek M (2010) Membrane
biogas upgrading processes for the production of natu-
Baker R (1999) Recent developments in membrane vapour ral gas substitute. Sep Purif Technol 74:8392
separation systems. Membr Technol 1999(114):912 Matteucci S, Yampolskii Y, Freeman B, Pinnau I (2006)
Baker RW, Lokhandwala K (2008) Natural gas processing Transport of gases and vapors in glassy and rubbery
with membranes: an overview. Ind Eng Chem Res polymers. In: Yampolskii Y, Pinnau I, Freeman B (eds)
47:21092121
Baker RW, Wijmans JG, Kaschemekat JH (1998) The
Material science of membranes for gas and vapor sep- G
aration. Wiley, Chichester, pp 148
design of membrane vapour-gas separation systems. Pinnau I, Wijmans JG, Blume I, Kuroda T, Peinemann KV
J Membr Sci 151:5562 (1988) Gas permeation through composite membranes.
Bara JE, Hatakeyama SE, Gin DL, Noble RD (2008) Improv- J Membr Sci 37(1):8188
ing CO2 permeability in polymerized room-temperature Robeson LM (1991) Correlation of separation factor ver-
ionic liquid gas separation membranes through the for- sus permeability for polymeric membranes. J Membr
mation of a solid composite with a room-temperature Sci 62(2):165185
ionic liquid. Polym Adv Technol 19:14151420 Robeson LM (2008) The upper bound revisited. J Membr
Bernardo P, Drioli E, Golemme G (2009) Membrane gas Sci 320(12):390400
separation. A review/state of the art. Ind Eng Chem Res Rungta M, Zhang C, Koros WJ, Xu L (2013) Membrane-
48(10):46384663 based ethylene/ethane separation: the upper bound and
Budd PM, McKeown NB (2010) Highly permeable poly- beyond. AIChE J 59(9):34753489
mers for gas separation membranes. Polym Chem Sanders D, Smith ZP, Guo R, Robeson LM, McGrath JE,
1(1):6368 Paul DR, Freeman BD (2013) Energy-efcient poly-
Bushell AF, Atteld MP, Mason CR, Budd PM, meric gas separation membranes for a sustainable
Yampolskii YP, Starannikova L, Rebrov A, Bazzarelli F, future: a review. Polymer 54(4):7294761
Bernardo P, Jansen JC, Lan M, Friess K, Shantarovic V, Wind JD, Staudt-Bickel C, Paul DR, Koros WJ (2002) The
Gustov V, Isaeva V (2013) Gas permeation parameters of effects of crosslinking chemistry on CO2 plasticization
mixed matrix membranes based on the polymer of intrin- of polyimide gas separation membranes. Ind Eng Chem
sic microporosity PIM-1 and the zeolitic imidazolate Res 41(24):61396148
framework ZIF-8. J Membr Sci 427:4862 Yampolskii Y (2012) Polymeric gas separation mem-
Bux H, Chmelik C, Krishna R, Caro J (2011) Ethene/ branes. Macromolecules 45(8):32983311
ethane separation by the MOF membrane ZIF-8:
molecular correlation of permeation, adsorption, diffu-
sion. J Membr Sci 369:284289
Carta M, Malpass-Evans R, Croad M, Rogan Y, Jansen JC,
Bernardo P, Bazzarelli F, McKeown NB (2013) An Gas Separation Membranes:
efcient polymer-based molecular sieve membranes High-Temperature Glass Joining
for membrane gas separations. Science 339:303307 and Sealing
Esteves IAAC, Mota JPB (2007) Gas separation by a novel
hybrid membrane/pressure swing adsorption process.
Ind Eng Chem Res 46(17):57235733 Marijke Jacobs
Fallanza M, Ortiz A, Gorri D, Ortiz I (2013) Polymerionic Sustainable Materials Management, Flemish
liquid composite membranes for propane/propylene Institute for Technological Research VITO,
separation by facilitated transport. J Membr Sci
Mol, Belgium
444:164172
Goh PS, Ismail AF, Sanip SM, Ng BC, Aziz M (2011) Recent
advances of inorganic llers in mixed matrix membrane
for gas separation. Sep Purif Technol 81:243264 Gas separation membranes such as ceramic mixed
Grinevich Yu, Starannikova L, Yampolskii Yu,
ionic-electronic conducting (MIEC) membranes
Gringolts M, Finkelshtein E (2011) Solubility con-
trolled permeation of hydrocarbons in novel highly are able to transport oxygen or protons through
permeable polymers. J Membr Sci 378:250256 their crystal structure at high temperatures
858 Gas Separation Membranes: High-Temperature Glass Joining and Sealing
(>700 C) in the presence of a driving force difcult to apply on the ceramic membranes, and
(Sunarso et al. 2008). This can be accomplished the seals are generally of lower quality than other
by sealing the membrane to a ceramic (alumina, types of seals. The last group of seals, the glass-
zirconia, or mullite) or metal tube. Since the mem- based ones, form rigidly bonded seals by forming
branes can have different geometries, such as disk chemical bonds with the other materials and pro-
or tubular, diverse seal positions are possible, viding a hermetic joint. Furthermore, the properties
which are illustrated in Fig. 1. of the glass or glass-ceramic seal can be tailored by
This seal must be gastight at the operating tem- changing the composition of the glass, and they are
peratures (7001000 C) of the membranes. How- less expensive than noble metals (Qi et al. 2001;
ever, MIEC membranes have a very high thermal Mahapatra and Lu 2010; Vivet et al. 2011). As
expansion coefcient (CTE > 20 106/K) com- such, glass-based seals are very promising for
pared to the materials to which they are bonded high-temperature sealing of membranes.
(CTE of 815 106/K). As such, at the operating A glass seal can consist of different constitutes
temperatures of the membranes, there is a big ther- such as a network former (SiO2 or B2O3), a network
mal expansion mismatch between the materials, modier (Li2O, Na2O, K2O, BaO, SrO, CaO, MgO),
which can cause failure of the joint. Therefore, to an intermediate (mostly Al2O3), and an additive
achieve a gastight joint, there must be a good (Fig. 2). The rst two constitutes will have the
wettability of the seal on the membrane and the most inuence on the CTE, the glass transition
support, and the seal must have a suitable viscosity temperature Tg, and the softening temperature Ts of
and coefcient of thermal expansion (CTE) in the glass. These are important characteristics since
the required temperature range. Once a seal is the CTE must match as much as possible with the
obtained, it must be thermal, chemical, and joining materials, Tg must be below the operating
mechanical stable during the operation of the temperature of the seal otherwise the seal becomes
membrane to maintain the gastight joint. brittle, and Ts must be higher than the operating
There are different types of sealing methods for temperature to avoid excessive ow of the seal.
joining ceramic gas separation membranes at high An example of a glass-ceramic composite seal
temperature: metal-based seals, compressive seals, which can be used for high-temperature sealing of
and glass-based seals. Metal-based seals can be ceramic membranes for gas separation is based on
applied by active metal brazing or reactive (air) Pyrex 7740 (SiO2 80.6 wt%, B2O3 13 wt%, Na2O
brazing, and they will form a compliant seal. How- 4 wt%, Al2O3 2.3 wt%, and K2O 0.1 wt%), mem-
ever, noble metals (Au, Pt) are expensive, while brane material powder, and NaAlO2 or B2O3
seals obtained with braze alloys, mostly Ag or Cu (Qi et al. 2001). This glass-ceramic composite is
based, can suffer from cracks and delamination due able to form a gastight seal between ceramic
to the CTE mismatch. Compressive seals are membranes such as La0.8Sr0.2Co0.6Fe0.4O3d and
formed by using a deformable material, typically SrCe0.95Tb0.05O3 and alumina or mullite in the
mica based, under a compressive load. This load is temperature range of 600950 C.
a Membrane b c
Membrane
Seal
Seal Membrane
Alumina tube
Alumina tube
Seal
Mullite tube
Gas Separation Membranes: High-Temperature tubular membrane to an alumina or a mullite tube for
Glass Joining and Sealing, Fig. 1 Schematic illustra- oxygen separation (Qi et al. 2001)
tion of high-temperature seal positions to x a disk or
Gas/Vapor Transport 859
Gas Separation
Membranes:
High-Temperature Glass
Joining and Sealing,
Fig. 2 Schematic
illustration of a glass
network structure
(Mahapatra and Lu 2010)
G
References itself, and porous membranes, where the transport
takes place through the open space of the pores in
Mahapatra MK, Lu K (2010) Glass-based seals for solid the membrane. In the latter, different transport
oxide fuel and electrolyzer cells a review. Mater Sci
modes exist, depending on the size of the pores
Eng R 67:6585
Qi X, Akin FT, Lin YS (2001) Ceramicglass composite and on the interaction of the gases and vapors with
high temperature seals for dense ionic-conducting the membrane material and with themselves. The
ceramic membranes. J Membr Sci 193:185193 most representative examples of transport mech-
Sunarso J, Baumann S, Serra JM, Meulenberg WA, Liu S, Lin
anisms are shown in Fig. 1.
YS, Diniz da Costa JC (2008) Mixed ionic-electronic
conducting (MIEC) ceramic-based membranes for oxy-
gen separation (Review). J Membr Sci 320:1341 (A) In large pores, convective or viscous ow
Vivet A, Geffroy PM, Coudert V, Fouletier J, Richet N, occurs. Such membranes nd application as
Chartier T (2011) Inuence of glass and gold sealants
lters to remove particulate matter from gas
materials on oxygen permeation performances in
La0.8Sr0.2Fe0.7Ga0.3O3 d perovskite membranes. and liquid streams, but they are not able to
J Membr Sci 366:132138 separate gases, which move as a homoge-
neous mixture through the pores.
(B) If the average pore diameter is smaller than the
mean free path between the molecules in the
Gas/Vapor Transport gas mixture at the given pressure, so-called
Knudsen diffusion takes place (Knudsen
Johannes Carolus Jansen1 and Marek Lan2 1909; Datta et al. 1992). In this case, the
1
Institute on Membrane Technology, National transport rate is inversely proportional to the
Research Council of Italy, ITM-CNR, Rende, square root of the molecular weight of the gas
Italy species and the selectivity is only a function of
2
University of Chemistry and Technology Prague, their molecular weight ratios.
Prague 6, Czech Republic (C) For even smaller pores, where the size of the
pores is in the range of the size of the gas
molecules themselves, molecular sieving can
General Introduction occur. Molecules that are larger than the pore
size are completely excluded, and only smaller
Transport of gases and vapors in membranes molecules may diffuse through the pores of the
depends rst of all on their micro- and macro- membrane. Such membranes can have very
scopic structure. There is a fundamental differ- high selectivities, in the case of very narrow
ence between dense membranes, where transport pore size distributions. Typical examples are
takes place through the material of the membrane carbon membranes (Vu et al. 2001) or zeolite
860 Gas/Vapor Transport
than larger molecules. However, if the larger

molecules condense inside the pore, the con-
densed phase obstructs the permeation of
smaller molecules. Thus, the mixed gas selec-
tivity may be opposite to the ideal selectivity.
(F) In dense membranes, the molecules move
through the bulk of the membrane material
itself, or more precisely through its free vol-
ume. The rest of this chapter will focus mostly
on the transport in dense membranes because
of their relevance for gas separation
applications.
Gas/Vapor Transport in Dense

Membranes
Ficks Law of Diffusion

The penetrant ow, J, through a nonporous
Gas/Vapor Transport, Fig. 1 Schematic representation membrane can be described by Ficks rst
of different transport mechanisms in porous and dense
membranes
law of diffusion (Fick 1855), derived analogi-
cally to Fouriers law for the description of heat
transfer.
membranes (Caro and Noack 2008; Rangnekar dci

et al. 2015). Also dense polymer membranes, J i Di (1)
dx
when they have a combination of very high
free volume and high stiffness of the polymer
where D is the diffusion coefcient and dc/dx the
chains, may exhibit behavior that comes close
concentration gradient. The concentration in
to molecular sieving (Carta et al. 2013).
nonstationary conditions depends not only on
(D) In the case of strong interaction of the mem-
the position in the continuous phase but also on
brane material with the permeating species or
time. Assuming one-dimensional diffusion, the
in case of readily condensable species, the
transport behavior can be described by Ficks
latter condenses on the pore wall. In this
second law of diffusion:
case, what is the most permeable species
depends on various factors, including the
molecular dimensions of the permeating spe- @ci @ 2 ci
Di 2 (2)
cies of the mixture and on the remaining @t @x
aperture of the pores, as well as the mobility
of the condensed species on the pore surface. Both equations assume that the diffusion coef-
(E) In the extreme situation of the previous, cap- cient is independent on concentration, which is
illary condensation takes place and the whole true at low penetrant activity. Figure 2 describes
pore is lled with the condensed liquid. The the gas concentration prole at different times,
selectivity of these membranes for gas mix- for a previously evacuated polymer membrane
tures may be completely different from the after single-sided and double-sided exposure to
ideal selectivity for single gases. Based on the a gas, based on Eq. 2 (Crank 1975). These pro-
Kundsen diffusion mechanism, small mole- les correspond to those in a typical permeation
cules would always be much more permeable and sorption experiment, respectively.
Gas/Vapor Transport, Fig. 2 Concentration proles sorption experiment (Relative feed concentration = 1)
in a at sheet membrane (100 mm) calculated by Eq. 2, Arrows represent the increasing time during the experi-
for D = 1012 m2 s-1 (Crank 1975) in the case of: (a) ideal ment. Dashed lines represent the concentration prole in
time lag permeation experiment (Relative feed time equal to permeation time lag (see below)
concentration = 1, permeate concentration = 0); (b) ideal
The Solution-Diffusion Mechanism PA DA SA

aA=B (5)
The transport of gases and vapors in dense poly- PB DB SB
mer membranes is governed by the solution-
diffusion model (Wijmans and Baker 1995,
2006). The driving force of this process in dense In ideal systems, such as for the transport of light
membranes is the partial pressure gradient of the gases in rubbers, P, D, and S are constants, but this
permeating species across the membrane, or more is rather an exception and in many cases D and
correctly its chemical potential gradient. The pro- S depend on the concentration of the permeating
cess consists of three steps: rst, the gas is species in the membrane and thus on the gas
absorbed by the membrane at the polymer/gas pressure. Numerous correlations have been pro-
interface on the feed side, then it diffuses across posed, relating the gas and vapor transport prop-
the polymer bulk, and nally it desorbs from the erties of polymeric membranes to the molecular
membrane at the permeate side. The permeability properties of the penetrating species and to the
coefcient, P, is the product of the diffusion coef- chemical and physical properties of the polymer
cient, D, and the solubility coefcient, S: matrix, and they can be found in an excellent
review (Matteucci et al. 2006).
PDS (3) In phenomenological terms, the membrane
productivity is expressed by means of its
where S is dened as the ratio of the equilibrium permeance, dened as an amount of permeate
gas concentration, C, and the gas pressure, p: per unit membrane area, time, and driving force.
A commonly used unit is the gas permeation
S C=p (4) unit, GPU:
The selectivity aA/B of two species A and B is

cm3STP
dened as the ratio of the two individual perme- 1 GPU 106
ability coefcients and this can be decomposed cm2 s cm Hg
(6)
into a diffusion selectivity term and a solubility m3
2:70 109 2 STP
selectivity term: m h bar
The permeance is a property of the membrane and Temperature dependency of diffusivity is usually
depends on its effective thickness. The permeabil- stronger than that of solubility, and therefore the
ity coefcient is an intrinsic property of the mate- permeability usually increases with increasing
rial. The most commonly used unit to describe the temperature (Ghosal and Freeman 1994).
permeability coefcient is the Barrer:
cm3STP cm Gas/Gas, Gas/Vapor, and Vapor/Vapor

1 Barrer 1010 (7) Separation
cm2 s cm Hg
The most important industrial applications of gas

Diffusion is an activated process and solubility is
and vapor separations vary from (A) gas/gas sep-
a thermodynamic property. The temperature
arations to (B) gas/vapor separations, where the
dependence of diffusivity and solubility can there-
membrane is in contact with highly condensable
fore be described by the following Arrhenius and
species. In the extreme case of pervaporation (C),
vant Hoff relationships, respectively (Van
the membrane is at one side in contact with a
Amerongen 1946; Costello and Koros 1992).
liquid phase and at the downstream side it is in
contact with a gas phase. The type of separation
D D0 eED =RT (8) process dictates the possible operation conditions
and the choice of the membrane materials.
S S0 eDHs =RT (9)
(A) For simple gas/gas separations, for instance,
where DHs is the enthalpy of sorption of the O2/N2 separation from air for pure nitrogen
penetrant in the polymer, ED activation energy of production or for O2 enrichment, in principle
diffusion, and D0 and S0 are the preexponential many membrane materials can be safely used
factors. (Baker and Low 2014). The choice depends
Assuming solution-diffusion model (Eq. 3), mainly on the need to achieve a high separa-
temperature dependency of permeability is: tion factor at relatively low ux or if a high
ux is needed and the separation factor is less
P P0 eEP =RT (10) important.
(B) For gas/vapor separations, the transport prop-
where P0 is preexponential factor and EP activa- erties of the different species vary widely.
tion energy of permeation which is equal to: Often the vapor consists of readily condens-
able large molecules, which have a high sol-
EP DH S ED (11) ubility in combination with a low diffusion
coefcient, in contrast to the light gas, with a
As a temperature-activated process, diffusion usu- lower solubility and a high diffusion coef-
ally accelerates with temperature. Dissolution of cient. In this case, the species in the mixture
the gas can be considered as a two-step process are likely to inuence each other, directly via
of condensation of the gas phase, followed competitive sorption in the limited free vol-
by mixing with the polymer matrix. For light ume available and indirectly via plasticiza-
gases, the solubility therefore increases with tion of the polymer matrix by the
increasing temperature, because the negative condensable species. The same situation
enthalpy of condensation is negligible with occurs in gas/gas separations, where one of
respect to the positive enthalpy of mixing. On the two gases readily condenses at higher
the other hand, enthalpy of sorption of more con- pressure, for instance, CO2. Typical examples
densable gases and vapors is negative due to the are volatile organic compounds (VOC)
high negative enthalpy of condensation and the removal from air (Leemann et al. 1996), or
solubility decreases with increasing temperature. CO2 removal from natural gas (Adewole
et al. 2013) or biogas, air dehydration, 4

etc. For such separations, it may be conve-
nient to use rubbery membranes, which are
Permeate pressure (mbar)

less prone to plasticization and which are 3
solubility selective rather than diffusivity
selective.
(C) Pervaporation is the extreme case of vapor/ 2
vapor separation, with condensed vapors

(=liquid mixture) at the feed side of the mem-
1
brane and gaseous species at the permeate
side, either by application of a vacuum or by
the use of a sweeping gas (Mulder and Smol-
0
ders 1991). Pervaporation is particularly 0 25 50 75 100 125
advantageous in the case of azeotropic liq- Time (s) G
uid mixtures, where the membrane can
break the azeotrope. Typical examples of Gas/Vapor Transport, Fig. 3 Typical permeation curve
for analysis of pure gas permeability by the time lag
practical separations are the dehydration of method, with indication of time lag Y determined via the
alcohol with hydrophilic membranes, etha- tangent method
nol recovery from hydroalcoholic solu-
tions, or VOC removal from wastewater
with hydrophobic membranes. In terms of fully evacuated inside a closed permeation cell,
the transport properties, these membranes and after exposure of the membrane to the gas at
are often strongly affected by the swelling the feed side, the pressure at the permeate side is
of the polymer by the permeating species. recorded as a function of time. For ideal systems,
An exception is the alcohol dehydration the concentration prole in the early stage of the
with glassy peruoropolymer membranes experiment takes the form of Fig. 2a and the
(Scholes et al. 2015a, b). A further compli- resulting permeation curve takes the form of
cation in pervaporation with respect to Fig. 3, which is described by the following equa-
gas/vapor separation is the existence of the tion (Jansen et al. 2011):
Schroders paradox, according to which the
membrane material often behaves differ- pt p0 dp=dt0 t

ently in contact with a liquid phase or in RT A l Dt 1
contact with the saturated vapor phase pf S ser
VP Vm l2 6
(Vallieres et al. 2006). A curiosity is that (12)
with
pervaporation membranes can be porous
and nonselective in the dry state and 2X1
1n D n2 p2 t
ser 2 exp
become dense and selective in contact p n1 n2 l2
with the feed mixture (Van Der Bruggen
et al. 2004, 2006). in which pt is the permeate pressure at time t and
p0 is the starting pressure, typically as close to
zero as the vacuum pump allows. In a leak-proof
Methods for Analysis of the Transport instrument, the baseline slope (dp/dt)0 is usually
Parameters in Membranes negligible for a defect-free and well-evacuated
membrane. R is the universal gas constant, T the
Time Lag Method for Pure Gases absolute temperature, A the exposed membrane
The most common way to determine the basic gas area, VP the permeate volume, Vm the molar vol-
transport parameters is the so-called time lag ume of a gas at standard conditions (0 C and
method (Crank 1975), in which the membrane is 1 atm), pf the feed pressure, and l the membrane
thickness. The permeability coefcient, P, is cal- J Permeate

culated in the regime of quasi steady state perme- QPermeate (15)
A
ation, which is dened by the simplied Eq. 13,
describing the tangent to the linear part of the When using a sweeping gas, the permeate ux can
permeation curve: also be calculated from the known sweeping gas
ow rate and the gas concentration in the perme-
pt p0 dp=dt0 t ate/sweeping gas mixture. The individual gas
permeance, P, of the ith species in a gas mixture
RT A pf P (13)
t Y is obtained as the ratio of its volumetric permeate
VP Vm l ux, QPermeate, to the partial pressure difference
between the feed and permeate sides:
The last term is the so-called permeation time lag,
Y, which is usually determined from the intersec- xi Permeate QPermeate
Pi (16)
tion of the tangents before the onset of permeation xi Feed pFeed xi Permeate pPermeate
and after reaching the quasi steady state:
in which xi is the volume fraction or mole fraction
l2 of the ith species, pFeed and pPermeate are the total
Y (14)
6D feed and permeate pressures, respectively. The
mixed gas selectivity of species A and B, aA/B, is
For a membrane of known thickness, it allows the then calculated as the ratio of their individual
determination of the diffusion coefcient of the permeances:
gas. The gas solubility coefcient, S, can then be
obtained indirectly as the ratio of the permeability PA
aA=B (17)
to the diffusion coefcient, using Eq. 3. PB
While the transport parameters P and D are
usually obtained by using the tangents to the Sorption Analysis
permeation curve, the permeation curve can Direct sorption analysis is the most reliable way to
also be tted directly with Eq. 12, after expansion determine the solubility of the gas in the polymer
into a sufcient number of terms (Scheichl matrix. Sorption can be determined volumetri-
et al. 2005; Jansen et al. 2011). This yields the cally, gravimetrically, and with the pressure
values of P, D, and S directly. An inaccurate t in decay method (Keller and Staudt 2005), or by
the case of a deviating curve shape is a direct inverse gas chromatography (IGC) (Danner
indication of nonideal behavior. This happens, et al. 1998). Equilibrium sorption in polymers is
for instance, in the case of clustering or in the one of the basic characteristics describing the
case of strong dual mode sorption (DMS) interaction between a penetrant and a polymer.
behavior. In all these methods, the polymer sample is placed
inside the test cell and exposed to a given pene-
Constant Pressure: Variable Volume Method trant pressure. In gravimetric measurements, the
In this method, the pure feed gas or the feed gas mass uptake can be measured electromagnetically
mixture ows through the membrane cell in cross- (Mamaliga et al. 2004), by a quartz crystal
ow mode and the permeate is either collected as microbalance (QCM) (Mikkilineni et al. 1995)
such or it is transported by a sweeping gas to the or McBains quartz spiral balance (Friess
gas analyzer. The total permeation rate can be et al. 2011; Vopika et al. 2013).
determined directly, for instance, by a bubble Since the sorption coefcient (solubility) is
ow meter or electronic ow meters, measuring usually a function of pressure or activity, knowl-
the volumetric permeate ow rate, JPermeate. The edge of the sorption isotherm shape is important.
permeate ux, QPermeate, is the volumetric ow There are many different types of sorption iso-
rate per unit area: therms (Rouquerolt et al. 1994) depending on
the polymer structure and the relative difference

between penetrant/polymer and penetrant/pene-
trant interaction.
Sorption of light gases in rubbery polymers
increases linearly with pressure according to
Henrys law. On the other hand, vapors sorption
can be described by numerous equations such as
Flory-Huggins theory (Flory 1953), Flory-Rehner
(Flory and Rehner 1943a; Flory and Rehner
Jr. 1943b; Izk et al. 2003), Koningsveld-
Kleintjens equations (Koningsveld and Kleintjens
1971), or the ENSIC model (Favre et al. 1996).
In the case of dense glassy polymeric mem-
branes, three models are the most often used. G
Permanent gases behave almost linearly, at low
pressures following Henrys law. Alternatively, Gas/Vapor Transport, Fig. 4 Gravimetric sorption iso-
the dual-mode sorption model, Eq. 18, gives usu- therm of CO2 in Amine-PIM-1 (Mason et al. 2014), extrap-
ally a satisfactory description of the behavior olated with GAB and DMS models
(Barrer et al. 1958). This model is a combination
of Henrys law and the Langmuir sorption iso- pressure from a chosen reference pressure. For
therm, assuming monolayer sorption at existing gases, the following form of the GAB model
sorption sites. was proposed (Vopika and Friess 2014).
ch p b v m h p p
C CD C H k d p v (20)
1bp
(18) p p h p p p
where C is the gas concentration in the polymer, where p* is a pressure independent constant
CD and CH are the Henry and Langmuir concen- which has the meaning of a reference pressure.
tration contributions, respectively, kd is the linear When p* is equal to saturated vapor pressure, the
sorption parameter, ch is the monolayer capacity, model is equivalent to the BET model (Brunauer
and b is the afnity parameter, reecting the inter- et al. 1938) (Fig. 4).
action strength between polymer and penetrant. For samples with a well-dened geometry,
In the case of vapor sorption in glassy poly- sorption kinetics measurements allow the deter-
mers, the sorption curves often have a typical mination of the diffusion coefcient by equations
S-shape and the DMS model cannot be used. In based on Ficks second law (Crank 1975). The
such cases, the Guggenheim, Anderson, and de relative sorbed amount Qt/Q1 in a at lm with
Boer (GAB) model (Guggenheim 1966) gives a thickness l is given as a function of time by the
better description: following equation:
vm h a f Qt 8X1
1
D2n12 p2 t
v (19) 1 2 e l2 (21)
1 f a 1 f a h f a Q1 p n0 2n 12
where v is the mass of adsorbed vapor per mass of The corresponding concentration-time prole in a
polymer adsorbent, vm is the capacity of the rst at membrane during a sorption experiment is
adsorption monolayer, h denes the ratio of the shown schematically in Fig. 2b. Under real con-
adsorption strength in the rst and the subsequent ditions, a nite time is needed to charge the gas in
layers, a is the vapor activity, and f is a constant the sorption apparatus and a correction for the
dening the deviation of the saturated vapor assumed step-pressure-increase is necessary
reproduced well in the simulations (Macchione

et al. 2007; Jansen et al. 2010). Whereas molecu-
lar dynamics simulations work fairly well for
small molecules at low concentration, like in gas
separation, models to analyze and predict the mass
transport in pervaporation require different
approaches (Lipnizki and Trgrdh 2001), such
as the UNIQUAC model (Heintz and Stephan
1994).
Transport in Heterogeneous
and Homogeneous Mixtures
The description of the transport in polymers with

homogeneously or heterogeneously dispersed
additives or in polymer blends is much more
complex than that in neat polymers. Mixed matrix
Gas/Vapor Transport, Fig. 5 Typical gravimetric sorp-
membranes are currently receiving much attention
tion kinetics curve. CO2 sorption tted with the model
proposed by Vopika (Vopika et al. 2009) in microporous (Aroon and Ismail 2010; Rezakazemi et al. 2014)
Trgers base polymer EA-TB-PIM (Carta et al. 2013) after because they have the potential to combine the
MeOH treatment high permeability and selectivity of inorganic
(e.g., zeolites (Miller et al. 2007)), carbonaceous
(Vopika et al. 2009). An example of a typical (Vu et al. 2003), and organometallic ller particles
sorption kinetics curve obtained with this model is (metal organic frameworks, MOFs (Bushell
shown in Fig. 5. et al. 2013; Zornoza et al. 2013; Adatoz
et al. 2015)) with the good mechanical properties
Modeling of Transport of polymers. There is a large number of predictive
With the increasing computational power of mod- models to describe the performance of MMMs
ern computers, modeling of structural (Heuchel (Vinh-Thang and Kaliaguine 2013). One of the
et al. 2008) and transport properties of gases simplest and most commonly used models to
(Hofmann et al. 2000; Frentrup et al. 2015) and describe the transport in MMMs is the Maxwell
vapors (Giacinti Baschetti and De Angelis 2015) model (Shimekit et al. 2011), valid for a low
in polymeric membranes has gained a prominent concentration of spherical particles dispersed in
position in membrane research. Especially in the the continuous phase:
description of the free volume distribution of
membrane materials, computational methods Pd 2Pc 2fd Pd Pc
Peff (22)
offer a level of insight that no single experimental Pd 2Pc 2fd Pd Pc
method can give. The transport can be studied at
the atomistic level (Hofmann et al. 2000; where the Peff is the effective permeability of the
Theodorou 2006), showing, for instance, the mixed matrix membrane, Pc and Pd represent the
hopping mechanism of a gas molecule from gas permeabilities in the continuous and dispersed
one free volume element to the next, conrming phase, respectively, and fd is the volume fraction
the activated mechanism seen experimentally. of the dispersed phase. Generally, the permeabil-
Although there is often a large discrepancy ity of the gases through the dispersed phase
between the calculated sorption and diffusion depends on the overall void volume, its distribu-
coefcients and the experimental values, the tion in the ller particles, and on the channel size,
trends between different gas species are usually which affect the free volume of the overall system.
More sophisticated models take into account also

the particle shape of the dispersed phase (Cussler
1990). Such particles may have a pronounced
effect in the case of high aspect ratios (Rodenas
et al. 2014) due to the strong effect on the diffu-
sion path length (Falla et al. 1996). Interestingly,
also impermeable fumed silica (Merkel
et al. 2002) or graphene (Althumayri et al. 2016)
ller particles, with intrinsic barrier properties,
can have a positive effect on the permeability of
the membranes, when additional free volume is
created at the polymer-particle interface.
When the dispersed phase is another polymer,
in the case of immiscible polymer blends, the G
transport can be described by fundamentally the
same equations as the mixed matrix materials
(e.g., Eq. 22). Instead, for miscible polymer
blends, the permeability, Pb, is reported to obey Gas/Vapor Transport, Fig. 6 O2/N2 1991 and 2008
upper bound curves for the selectivity versus permeability
the following equation (Robeson 2010): trade-off relation (Robeson 1991, 2008) with the latest
upper bound suggested by Swaidan et al. (2015). The
lnPb f1 lnP1 f2 lnP2 (23) oval represents the approximate cloud of experimental data
in which f1 and f2 are the volume fractions of the

two polymers, respectively, and P1 and P2 are upper bound (Freeman 1999) and concluded that a
their permeabilities. It shows a linear trend when combination of high free volume and extreme
the permeabilities are plotted on a logarithmic rigidity of the polymer chains is needed to exceed
scale and deviates from linearity in the case of the current upper bound (Robeson et al. 2009).
(partial) immiscibility of the two polymers This was conrmed by McKeown et al. with a
(Jansen et al. 2013). novel polymer of intrinsic microporosity based on
Trogers base and ethanoanthracene (Carta
et al. 2013) or benzotriptycene units (Rose
Overall Performance et al. 2015). Alentiev presented a similar approach
for the trade-off in diffusion coefcient and diffu-
Robeson Trade-Off Behavior sion selectivity (Alentiev and Yampolskii 2013).
Although membrane separations may have many
advantages compared to traditional separation Effect of Physical Aging on the Transport
processes such as distillation or pressure swing Properties
adsorption, a limitation is the trade-off behavior The global nonequilibrium state of glassy poly-
between selectivity and permeability. This trend meric membranes tends to relax over time. This
was rstly discussed by Robeson in 1991, who process, where no chemical changes occur, is
suggested a linear so-called upper bound for many called physical aging (Struik 1978) and affects
relevant gas pairs (Fig. 6) (Robeson 1991), which different properties of a polymer. One of these is
were subsequently updated and extended the free volume distribution, which, in turn, is
(Robeson et al. 1994; Robeson 2008). In 2015, a reected in the gas transport properties of the
new upper bound was set for O2/N2, H2/N2, and membrane. Physical aging therefore has a strong
H2/CH4, based on mostly the development of impact on the performance of amorphous glassy
polymers of intrinsic microporosity (Swaidan gas separation membranes (Pfromm 2006), and
et al. 2015). Freeman discussed the basis of the gas diffusion is a very sensitive method to probe
changes in the free volume of a polymer mem- References

brane (Jansen et al. 2009). Different aging mech-
anisms have been proposed. Harms claims that Adatoz E, Avci AK, Keskin S (2015) Opportunities and
challenges of MOF-based membranes in gas separa-
free volume elements diffuse towards the surface
tions. Sep Purif Technol 152:207237. doi:10.1016/j.
of the polymer (Harms et al. 2012). However, due seppur.2015.08.020
to the very low expected diffusion coefcient of Adewole JK, Ahmad AL, Ismail S, Leo CP (2013)
this process, it is signicant only in the case of thin Current challenges in membrane separation of
CO2 from natural gas: a review. Int J Greenh Gas
layers. McCaig et al. proposed that the aging
Control 17:4665. doi:10.1016/j.ijggc.2013.04.
consists of two distinct processes (McCaig and 012Review
Paul 2000): (i) thickness independent lattice con- Alentiev A, Yampolskii Y (2013) Correlation of gas per-
traction and (ii) thickness dependent diffusion of meability and diffusivity with selectivity: orientations
of the clouds of the data points and the effects of
free volume. Usually, the polymer chains pack
temperature. Ind Eng Chem Res 52:88648874.
more efciently during physical aging and the doi:10.1021/ie302680r
polymer becomes denser, resulting in a decrease Althumayri K, Harrison WJ, Shin Y, et al (2016) The
in permeability and an increase in selectivity. inuence of few-layer graphene on the gas permeability
of the high-free-volume polymer PIM-1. Philos Trans
Many approaches have been used to overcome
R Soc London A Math Phys Eng Sci 374:20150031:in
physical aging and to prepare time-stable mate- press. doi:10.1098/rsta.2015.0031
rial, such as crosslinking and addition of llers. Aroon MA, Ismail AF (2010) Performance studies of
Lau et al. showed that addition of an ultraporous mixed matrix membranes for gas separation: a review.
Sep Purif Technol 75:229242. doi:10.1016/j.
additive can prevent the effect of aging on the
seppur.2010.08.023
transport properties of super glassy polymer mem- Baker RW, Low BT (2014) Gas separation membrane mate-
branes (Lau et al. 2014). rials: a perspective. Macromolecules 47:69997013.
doi:10.1021/ma501488s
Barrer RM, Barrie JA, Slater J (1958) Sorption and diffu-
Anomalous Transport
sion in ethyl cellulose. Part III. Comparison between
As already anticipated, Ficks rst and second ethyl cellulose and rubber. J Polym Sci 27:177197.
laws give a rather simplied representation of doi:10.1016/0010-4361(90)90120-L
diffusion in polymers, which usually applies Brunauer S, Emmett PH, Teller E (1938) Adsorption of
gases in multimolecular layers. J Am Chem Soc
only to permanent gases at low pressures. In the
60:309319. doi:10.1021/ja01269a023
majority of cases, the diffusion coefcient is not a Bushell AF, Atteld MP, Mason CR et al (2013) Gas
constant, because of mutual interactions between permeation parameters of mixed matrix membranes
the permeating species, giving rise to clustering based on the polymer of intrinsic microporosity
PIM-1 and the zeolitic imidazolate framework ZIF-8.
(Jansen et al. 2011) and interaction between the
J Membr Sci 427:4862. doi:10.1016/j.memsci.2012.
polymer and the permeating species, 09.035
immobilizing the latter (Mason et al. 2014) or Caro J, Noack M (2008) Zeolite membranes recent devel-
plasticizing the polymer (Lo et al. 2010) and opments and progress. Microporous Mesoporous
Mater 115:215233
favoring diffusion. The solubility of gases and
Carta M, Malpass-Evans R, Croad M et al (2013) An
vapors in the polymer matrix is usually only con- efcient polymer molecular sieve for membrane gas
stant at very low gas pressure or vapor activity. As separations. Science 339:303307. doi:10.1126/
a result, the simple expression for the solution science.1228032
Costello LM, Koros WJ (1992) Temperature dependence
diffusion model in Eq. 3 becomes concentration
of gas sorption and transport properties in polymers:
dependent: measurement and applications. Ind Eng Chem Res
31:27082714. doi:10.1021/ie00012a012
P c D c S c (24) Crank J (1975) The mathematics of diffusion, 2nd edn.
Clarendon, Oxford
This anomalous behavior becomes even more Cussler EL (1990) Membranes containing selective akes.
J Membr Sci 52:275288. doi:10.1016/S0376-
complex in the case of mixed gas or vapor perme- 7388(00)85132-7
ation, where there may also exist a coupling effect Danner RP, Tihminlioglu F, Surana RK, Duda JL
between the different species in the mixture. (1998) Inverse gas chromatography applications in
polymer-solvent systems. Fluid Phase Equilib polymer with intrinsic microporosity (PIM-1).
148:171188. doi:10.1016/S0378-3812(98)00203-9 J Membr Sci 318:8499. doi:10.1016/j.
Datta R, Dechapanichkul S, Kim JS et al (1992) memsci.2008.02.038
A generalized model for the transport of gases in Hofmann D, Fritz L, Ulbrich J et al (2000) Detailed-atomistic
porous, non-porous, and leaky membranes. molecular modeling of small molecule diffusion and
I. Application to single gases. J Membr Sci solution processes in polymeric membrane materials.
75:245263. doi:10.1016/0376-7388(92)85067-S Macromol Theory Simul 9:293327. doi:10.1002/1521-
Falla WR, Mulski M, Cussler EL (1996) Estimating 3919(20000701)9:6<293::AID-MATS293>3.0.CO;2-1
diffusion through ake-lled membranes. J Membr Izk P, Bartovsk L, Friess K et al (2003) Comparison of
Sci 119:129138. doi:10.1016/0376-7388(96)00106-8 various models for transport of binary mixtures through
Favre E, Nguyen QT, Clment R, Nel J (1996) The dense polymer membrane. Polymer 44:26792687.
engaged species induced clustering (ENSIC) model: a doi:10.1016/S0032-3861(03)00137-X
unied mechanistic approach of sorption phenomena in Jansen JC, Macchione M, Tocci E et al (2009) Comparative
polymers. J Membr Sci 117:227236. doi:10.1016/ study of different probing techniques for the analysis of
0376-7388(96)00060-9 the free volume distribution in amorphous glassy
Fick A (1855) Ueber diffusion. Ann Phys 170:5986. peruoropolymers. Macromolecules 42:75897604.
doi:10.1002/andp.18551700105 doi:10.1021/ma901244d G
Flory PJ (1953) Principles of polymer chemistry. Cornell Jansen JC, Friess K, Tocci E, Macchione M (2010) Amor-
University Press, Ithaca phous glassy peruoropolymer membranes of hyon
Flory PJ, Rehner JJ (1943a) Statistical mechanics of cross- AD : free volume distribution by photochromic prob-
linked polymer networks. J Chem Phys 11:512520. ing and vapour transport properties. In: Yampolskii Y,
doi:10.1063/1.1723792 Freeman B (eds) Membrane gas separation. Wiley,
Flory PJ, Rehner J Jr (1943b) Statistical mechanics of Chichester, pp 5983
cross-linked polymer networks II. Swelling. J Chem Jansen JC, Friess K, Drioli E (2011) Organic vapour trans-
Phys 11:521526. doi:10.1063/1.1723792 port in glassy peruoropolymer membranes: a simple
Freeman BD (1999) Basis of permeability/selectivity semi-quantitative approach to analyze clustering phe-
tradeoff relations in polymeric gas separation mem- nomena by time lag measurements. J Membr Sci
branes. Macromolecules 32:375380. doi:10.1021/ 367:141151. doi:10.1016/j.memsci.2010.10.063
ma9814548 Jansen JC, Darvishmanesh S, Tasselli F et al (2013) Inu-
Frentrup H, Hart K, Colina C, M uller E (2015) In silico ence of the blend composition on the properties and
determination of gas permeabilities by non-equilibrium separation performance of novel solvent resistant
molecular dynamics: CO2 and He through PIM-1. Mem- polyphenylsulfone/polyimide nanoltration mem-
branes 5:99119. doi:10.3390/membranes5010099 branes. J Membr Sci 447:107118. doi:10.1016/j.
Friess K, Hynek V, pek M et al (2011) Permeation and memsci.2013.07.009
sorption properties of poly(ether-block-amide) mem- Keller JU, Staudt R (2005) Gas adsorption equilibria.
branes lled by two types of zeolites. Sep Purif Technol Springer, Boston
80:418427. doi:10.1016/j.seppur.2011.04.012 Knudsen M (1909) Die Gesetze der Molekularstrmung
Ghosal K, Freeman BD (1994) Gas separation using poly- und der inneren Reibungsstrmung der Gase durch
mer membranes: an overview. Polym Adv Technol Rhren. Ann Phys 333:75130. doi:10.1002/
5:673697. doi:10.1002/pat.1994.220051102 andp.19093330106
Giacinti Baschetti M, De Angelis MG (2015) Vapour per- Koningsveld R, Kleintjens LA (1971) Liquid-liquid phase
meation modelling. In: Basile A, Khayet AF (eds) separation in multicomponent polymer systems.
Pervaporation, vapour permeation and membrane dis- X. Concentration dependence of the pair-interaction
tillation: principles and applications. Woodhead Pub- parameter in the system cyclohexane-polystyrene. Mac-
lishing, Oxford, pp 203246 romolecules 4:637641. doi:10.1021/ma60023a026
Guggenheim EA (1966) Applications of statistical Lau CH, Nguyen PT, Hill MR et al (2014) Ending aging in
mechanics. Clarendon, Oxford, pp 186207 super glassy polymer membranes. Angew Chem Int Ed
Harms S, Rtzke K, Faupel F et al (2012) Aging and free 53:53225326. doi:10.1002/anie.201402234
volume in a polymer of intrinsic microporosity Leemann M, Eigenberger G, Strathmann H (1996) Vapour
(PIM-1). J Adhes 88:608619. doi:10.1080/ permeation for the recovery of organic solvents from
00218464.2012.682902 waste air streams: separation capacities and process
Heintz A, Stephan W (1994) A generalized optimization. J Membr Sci 113:313322. doi:10.1016/
solution diffusion model of the pervaporation process 0376-7388(95)00130-1
through composite membranes Part I. Prediction of Lipnizki F, Trgrdh G (2001) Modelling of pervaporation:
mixture solubilities in the dense active layer using the models to analyze and predict the mass transport in
UNIQUAC model. J Membr Sci 89:143151. pervaporation. Sep Purif Rev 30:49125.
doi:10.1016/0376-7388(93)E0222-6 doi:10.1081/SPM-100102985
Heuchel M, Fritsch D, Budd PM et al (2008) Atomistic Lo C-H, Hung W-S, De Guzman M et al (2010) Investiga-
packing model and free volume distribution of a tion on CO2-induced plasticization in polycarbonate
membrane using positron annihilation lifetime spec- Robeson LM (1991) Correlation of separation factor
troscopy. J Membr Sci 363:302308. doi:10.1016/j. versus permeability for polymeric membranes.
memsci.2010.07.049 J Membr Sci 62:165185. doi:10.1016/0376-7388(91)
Macchione M, Jansen JC, De Luca G et al (2007) Exper- 80060-J
imental analysis and simulation of the gas transport in Robeson LM (2008) The upper bound revisited.
dense Hyon AD60X membranes: inuence of resid- J Membr Sci 320:390400. doi:10.1016/j.memsci.
ual solvent. Polymer 48:26192635. doi:10.1016/j. 2008.04.030
polymer.2007.02.068 Robeson LM (2010) Polymer blends in membrane trans-
Mamaliga I, Schabel W, Kind M (2004) Measurements of port processes. Ind Eng Chem Res 49:1185911865.
sorption isotherms and diffusion coefcients by means doi:10.1021/ie100153q
of a magnetic suspension balance. Chem Eng Process Robeson LM, Burgoyne WF, Langsam M et al (1994) High
Process Intensif 43:753763. doi:10.1016/S0255- performance polymers for membrane separation. Poly-
2701(03)00077-1 mer 35:49704978. doi:10.1016/0032-3861(94)
Mason CR, Maynard-Atem L, Heard KWJ et al (2014) 90651-3
Enhancement of CO2 afnity in a polymer of intrinsic Robeson LM, Freeman BD, Paul DR, Rowe BW
microporosity by amine modication. Macromolecules (2009) An empirical correlation of gas permeability
47:10211029. doi:10.1021/ma401869p and permselectivity in polymers and its theoretical
Matteucci S, Yampolskii Y, Freeman BD, Pinnau I (2006) basis. J Membr Sci 341:178185. doi:10.1016/j.
Transport of gases and vapors in glassy and rubbery memsci.2009.06.005
polymers. In: Yampolskii Y, Pinnau I, Freeman BD Rodenas T, Luz I, Prieto G et al (2014) Metalorganic
(eds) Materials science of membranes for gas and framework nanosheets in polymer composite materials
vapor separation. John Wiley & Sons, Ltd, Chichester, for gas separation. Nat Mater 14:4855. doi:10.1038/
UK, doi: 10.1002/047002903X.ch1 nmat4113
McCaig MS, Paul DR (2000) Effect of lm thickness on Rose I, Carta M, Malpass-Evans R et al (2015) Highly
the changes in gas permeability of a glassy polyarylate permeable benzotriptycene-based polymer of intrinsic
due to physical aging: part I. Experimental observa- microporosity. ACS Macro Lett 4:912915.
tions. Polymer 41:629637. doi:10.1016/S0032- doi:10.1021/acsmacrolett.5b00439
3861(99)00172-X Rouquerolt J, Avnir D, Fairbridge CW et al (1994) Recom-
Merkel TC, Freeman BD, Spontak RJ et al (2002) mendations for the characterization of porous solids.
Ultrapermeable, reverse-selective nanocomposite Pure Appl Chem 66:17391758. doi:10.1351/
membranes. Science 296:519522. doi:10.1126/ pac199466081739
science.1069580 Scheichl R, Klopffer M, Benjelloundabaghi Z,
Mikkilineni SPVN, Tree DA, High MS Flaconneche B (2005) Permeation of gases in poly-
(1995) Thermophysical properties of penetrants in mers: parameter identication and nonlinear regression
polymers via a piezoelectric quartz crystal microbal- analysis. J Membr Sci 254:275293. doi:10.1016/j.
ance. J Chem Eng Data 40:750755 memsci.2005.01.019
Miller SJ, Koros WJ, Vu DQ (2007) Mixed matrix mem- Scholes CA, Kanehashi S, Stevens GW, Kentish SE
brane technology: enhancing gas separations with poly- (2015a) Water permeability and competitive perme-
mer/molecular sieve composites. Elsevier, Amsterdam/ ation with CO2 and CH4 in peruorinated polymeric
New York, 170, pp 15901596 membranes. Sep Purif Technol 147:203209.
Mulder MHV, Smolders CA (1991) Mass transport phe- doi:10.1016/j.seppur.2015.04.023
nomena in pervaporation processes. Sep Sci Technol Scholes CA, Kentish SE, Stevens GW et al (2015b) Thin-
26:8595. doi:10.1080/01496399108050458 lm composite membrane contactors for desorption of
Pfromm PH (2006) The impact of physical aging of amor- CO2 from monoethanolamine at elevated temperatures.
phous glassy polymers on gas separation membranes. Sep Purif Technol 156:841847. doi:10.1016/j.
In: Yampolskii Y, Pinnau I, Freeman BD (eds.) Mate- seppur.2015.11.010
rials science of membranes for gas and vapor separa- Shimekit B, Mukhtar H, Murugesan T (2011) Prediction of
tion. John Wiley & Sons, Ltd, Chichester, UK the relative permeability of gases in mixed matrix
Rangnekar N, Mittal N, Elyassi B et al (2015) Zeolite membranes. J Membr Sci 373:152159. doi:10.1016/
membranes a review and comparison with MOFs. j.memsci.2011.02.038
Chem Soc Rev 44:71287154. doi:10.1039/ Struik LCE (1978) Physical aging in amorphous
C5CS00292C polymers and other materials. Elsevier Scientic,
Rezakazemi M, Ebadi Amooghin A, Montazer-Rahmati New York
MM et al (2014) State-of-the-art membrane based Swaidan R, Ghanem B, Pinnau I (2015) Fine-tuned intrin-
CO2 separation using mixed matrix membranes sically ultramicroporous polymers redene the perme-
(MMMs): an overview on current status and future ability/selectivity upper bounds of membrane-based air
directions. Prog Polym Sci 39:817861. doi:10.1016/ and hydrogen separations. ACS Macro Lett 4:947951.
j.progpolymsci.2014.01.003 doi:10.1021/acsmacrolett.5b00512
GasLiquid Membrane Contactor 871
Theodorou DN (2006) Principles of molecular simulation

of gas transport in polymers. In: Yampolskii Y, GasLiquid Exchangers Using
Pinnau I, Freeman B (eds) Materials science of mem-
branes for gas and vapor separation. Wiley, Chichester, Membrane as Interface
pp 4994
Vallieres C, Winkelmann D, Roizard D et al (2006) On GasLiquid Membrane Contactor
schroeders paradox. J Membr Sci 278:357364.
doi:10.1016/j.memsci.2005.11.020
Van Amerongen GJ (1946) The permeability of different
rubbers to gases and its relation to diffusivity and
solubility. J Appl Phys 17:972. doi:10.1063/1.1707667
GasLiquid Membrane Contactor
Van Der Bruggen B, Jansen JC, Figoli A et al (2004) Deter-
mination of parameters affecting transport in polymeric Denis Roizard
membranes: parallels between pervaporation and Laboratoire Ractions et Gnie des Procds,
nanoltration. J Phys Chem B 108:1327313279.
CNRS- Universit de Lorraine, Nancy, France
doi:10.1021/jp048249g
Van Der Bruggen B, Jansen JC, Figoli A et al (2006) Char- G
acteristics and performance of a universal membrane
suitable for gas separation, pervaporation, and Synonyms
nanoltration applications. J Phys Chem
B 110:1379913803. doi:10.1021/jp0608933
Vinh-Thang H, Kaliaguine S (2013) Predictive models for Gasliquid exchangers using membrane as inter-
mixed-matrix membrane performance: a review. Chem face; Nondispersive gasliquid contactor
Rev 113:49805028. doi:10.1021/cr3003888
Vopika O, Friess K (2014) Analysis of gas sorption in
glassy polymers with the GAB model: an alternative to
the dual mode sorption model. J Polym Sci History
B 52:14901495. doi:10.1002/polb.23588
Vopika O, Hynek V, Zgaar M et al (2009) A new sorption Gasliquid contactors are devices which are
model with a dynamic correction for the determination
designed to promote mass transfer between a gas
of diffusion coefcients. J Membr Sci 330:5156.
doi:10.1016/j.memsci.2008.12.037 phase and a liquid phase, thanks to gasliquid
Vopika O, Friess K, Hynek Vet al (2013) Equilibrium and contact. Among the various types of existing
transient sorption of vapours and gases in the polymer contactors, e.g., valve trays, random or structured
of intrinsic microporosity PIM-1. J Membr Sci
packing, demister, vacuum towers, etc., and mem-
434:148160. doi:10.1016/j.memsci.2013.01.040
Vu DQ, Koros WJ, Miller SJ (2001) High pressure CO2/ branes, one can clearly distinct two categories:
CH4 separation using carbon molecular sieve hollow rstly, contactors requiring a straight mixing
ber membranes. Ind Eng Chem Res 41:367380. between the gas and the liquid phases and,
doi:10.1021/ie010119w
secondly, contactors where the direct physical
Vu DQ, Koros WJ, Miller SJ (2003) Mixed matrix mem-
branes using carbon molecular sieves II. Modeling per- contact between the two phases does not exist,
meation behavior. J Membr Sci 211:311334. i.e., a contactor which does not need dispersion
doi:10.1016/S0376-7388(02)00429-5 of one phase into the other one to be efcient
Wijmans JG, Baker RW (1995) The solution-diffusion
(Fig. 1).
model: a review. J Membr Sci 107:121. doi:10.1016/
0376-7388(95)00102-I Up to now, membrane contactors are the only
Wijmans JG, Baker RW (2006) The solution-diffusion example of the second category, i.e., systems
model: a unied approach to membrane permeation. which are simultaneously able to avoid phase
In: Yampolskii Y, Pinnau I, Freeman BD (eds.) Mate-
mixing while promoting mass transfer. Obviously
rials science of membranes for gas and vapor separa-
tion, John Wiley & Sons, Ltd, Chichester, UK. doi: the higher the membrane permeability, the better
10.1002/047002903X.ch5 is mass transfer efcacy. Currently, most of the
Zornoza B, Tellez C, Coronas J et al (2013) Metal organic industrial applications are using microporous
framework based mixed matrix membranes: an increas-
membrane contactors (Liqui-Cell 2014).
ingly important eld of research with a large applica-
tion potential. Microporous Mesoporous Mater Like with other gas/liquid contactors, the mass
166:6778. doi:10.1016/j.micromeso.2012.03.012 transfer obtained by means of contactors can be
872 GasLiquid Membrane Contactor
GasLiquid Membrane Contactor, Fig. 1 Schematic

representation of a G/L membrane contactor
directed to carry out either separation and puri-

cation operations of a gas mixture by the selective
removal of a given gas component or, conversely, GasLiquid Membrane Contactor, Fig. 2 Micropo-
rous polypropylene
absorption of gas (e.g., N2, O2, O3, etc.) in the
liquid phase. The driving force is the gas partial
pressure according to the Henry law: pi = Hxi microporous membrane contactor is to avoid
with p the partial pressure, x the concentration of pore wetting by the liquid phase, because pore
gas at equilibrium, and H the Henry law coef- wetting would induce a detrimental decrease of
cient. Note also that the driving force of the mass the overall mass transfer coefcient (Fig. 3a, b). In
transfer can also be due to a chemical reaction case of pore ooding, the mass transfer coefcient
between the gas species and the liquid phase. can drop by a factor 1,00010,000.
Hence there is a breakthrough pressure to be
respected; this limiting pressure (Dp) difference
Phenomenon between the two phases can be calculated by the
YoungLaplace equation (Kim and Harriott 1987):
In the 1980s, Qi and Cussler (1985) achieved
Dp 4gL : cos y=dmax (1)
pioneering work devoted to the understanding of
mass transfer in gasliquid membrane contactors. with gL the surface tension of water, y the contact
Within this specic type of contactors, the mem- angle, and dmax the diameter of the biggest
brane is primarily a physical barrier between a gas micropores.
phase and a liquid phase. So to get an efcient As a guide, using pure water (gL = 72.8 mN/m
device, a proper choice of membrane and operat- at 20 C) and PP (with pore dmax = 1 mm and
ing conditions must be done to ensure a high level y = 115 ), the breakthrough pressure is 1.23 bar.
of mass transfer. For the membrane selection, it However in reality, the measured breakthrough
turned out that a microporous structure looked to pressure is often much lower due to the presence
be the best one to gather appropriate mechanical of impurities or organic solutes in the water; any
properties and high mass transfer coefcient. alteration of the polymer surface properties can
Most of the time, the liquid phase used is an also lower this value. On the other hand, to avoid
aqueous phase; hence, it logically orientated the any gas bubbling in the liquid phase, a slight
selection of membranes to hydrophobic rigid overpressure is usually applied at the membrane
polymers, i.e., glassy or semicrystalline struc- liquid interface.
tures, like polypropylene (PP) (Fig. 2), polyvinyl
diuoride (PVDF), or Teon. Thus the aqueous
phase cannot enter spontaneously the small pores Mass Transfer Theory
(0.11.106 m) of the membrane structure, as
predicted by the YoungLaplace equation given As one can expect, the mass transfer extent is
below. Nevertheless, the main concern with a closely related to the operating parameters, i.e.,
GasLiquid Membrane Contactor 873
a Gas phase Liquid phase b Gas phase Liquid phase
Cgas Cgas-Mb CMb-Liquid
Cliquid
1/kov = 1/kgas + 1/kmembrane + H/kliquid
GasLiquid Membrane Contactor, Fig. 3 (a) In a nonwetted porous membrane, the membrane transport occurs in
gas-lled pore. (b) In a wetted porous membrane, the membrane transport in liquid-lled pore is much slower
G
mainly gas partial pressure in each phase, temper- gas/liquid contact area is well known and remains
ature, gas and liquid ow rates, hydrodynamic constant whatever the gas or liquid ow rates.
conditions, and the specic transfer area. To As shown in Fig. 3, the overall mass transfer
achieve a given purication target, the mechanism resistance is due to the successive resistances of the
of the gas/liquid transfer must be known and used gas phase, of the membrane, and of the liquid
to model the mass transfer and thus to predict phase. Clearly the analogy can be made with elec-
adequate operating parameters. trical resistances, and the reciprocal of the overall
The performance of the mass transfer can be mass transfer coefcient (kov) can be written as the
related to the variation of gas composition either sum of the individual mass transfer coefcients:
between the inlet and the outlet of the gas phase or
between the inlet-dissolved gas and the outlet- 1=kov 1=kgas 1=kmembrane H=kliquid (4)
dissolved gas of the liquid phase. For a fast reac-
H being the Henry law constant between the gas
tion with a steady gas phase velocity, the same
and liquid phases.
basic equations, which are used to predict trans-
ferring gas in conventional columns, also apply
for membrane contactors as recalled below:
Pros and Cons
CGout
ek:a:L=ug (2) Compared to conventional gas/liquid contactors
CGin
(tower, packed columns, mixer settler), membrane
where CGout and CGin are respectively the outlet contactors are known to avoid a number of draw-
and inlet gas concentrations, k is the overall mass backs due to phase dispersion such as solvent loss,
transfer coefcient, L is length of the contactor, a foaming, unknown specied area, and column
is the contact area between the two phases, and ug ooding. It is worth also to underline that as
the gas phase velocity. polymer is the core of membrane contactors, one
Noting one can dene the gas transfer efciency can get a strong reduction in weight of the equip-
Z = (CGinCGout)/CGin, thus one can write: ment compared to conventional ones. At last one
can note as well that a membrane contactor can be
CGout installed horizontally or vertically without prob-
1 Z ek:a:L=ug (3)
CGin lems and that its efciency is not dependent of roll
and pitch marine.
It is worth noting that modeling the mass transfer Note that liquidliquid extraction can also be
properties of a membrane contactor is even easier achieved using membrane contactors, hence
than with conventional contactors because the avoiding any hazard of emulsion formation.
874 GasLiquid Membrane Contactor
All these advantages are linked to the mixing removed by contacting the gas feed ow with
nondispersive feature of a membrane contactor. a liquid properly chosen to trap the acidic spe-
On the other side, drawbacks can come cies. The gas removal can be due to a physical
from the facts that the pores of a membrane do chemical dissolution of the acid gas into the
not have the same size, that pore fouling can liquid or to a chemical reaction with the
occur and block the transfer, and that the hydro- liquid; in this case, the liquid must be endowed
phobic properties of the surface can be altered of basic properties or contains a solute which
inducing a dramatic decrease of the breakthrough is itself a base. When the principle of gas removal
pressure. is linked to physical chemical afnities, it is wise
to operate under pressure to promote higher sol-
ubility of the gas into the liquid. On the other
Intensification Potential of Mass case, if a chemical reaction is involved between
Transfer the gas and the liquid, pressure is not a key
parameter.
As with any membrane modules, the membrane As example of gas dissolution, one can cite
contactors can be prepared with different geome- nitrogenation in the beverage industry or blood
tries, going from plate and frame modules, spiral oxygenation. It is worth to note that blood oxy-
modules, or hollow ber modules. The last type genation has been one of the very rst examples of
has received the most attention because it allows using membrane contactors in 1975 (Esato and
the creation of very large interfacial areas, up to Eiseman 1975); currently the total annual market
10,000 m2/m3 that is up to 20-fold the interfacial is above 500 million.
area of a structured packed column (Gabelman Some examples of use of membrane contactors
and Hwang 1999; http://docnum.univ-lorraine.fr/ are listed here:
public/INPL/2011_NGUYEN_P_T.pdf6). Thus
the factor k.a., which is one of the key parameters Liquid degassing: O2, CO2, and N2 removal from
predicting the mass transfer efciency (Eq. 3), liquids, used for carbonation (food and beverage
indicates clearly that for a constant value of k, a industry), nitrogenation (microelectronics),
strong intensication of the transfer can be deoxygenation, etc.
reached with membrane contactors. Bubble-free gas/liquid mass transfer primarily
This high value of area will be obtained with for ozonation of semiconductor cleaning water
bers having very low inner diameter, typically Dehydration
in the 50100 mm range. Hence, a limit of the Blood oxygenator (health sector)
transfer with membrane contactor can now be
foreseen: the increase of the specic area shall
correspond to an increase of the pressure drop in
the bers of smaller diameter. This drawback will References
be amplied by using uid of high viscosity.
Liqui-Cell web site (visited Apr 2014). http://www.
Thus the potential of intensication is also liquicel.com/
strongly linked to the hydrodynamic conditions Esato K, Eiseman B (1975) Experimental evaluation of
prevailing in the contactor. This shall depend Gore-Tex, membrane oxygenator. J Thorac Cardiovas-
merely on the nature of the uid circulating in cular Surg 69(5):690697
Gabelman A, Hwang S-T (1999) Hollow ber membrane
the lumen of the bers. contactors. JMS 159:61106
Kim B, Harriott P (1987) Critical entry pressure for liquids
in hydrophobic membranes. J Colloid Interface Sci
Applications 115(1):18
Qi Z, Cussler EL (1985) Microporous hollow bers
for gas absorption. I, II. Mass transfer in the
As example of separations, acid impurities of ue liquid. J Membr Sci 23(3):321332, and ibid 23(3):
gases like SOx, NOx, or even CO2 can be 333345
Genotoxin Removal 875
polymeric and ceramic membranes stable in

Genotoxin Removal organic solvents include Koch SelRO, Evonik
DuraMem series, SolSep NF series, Borsig
Gyorgy Szekely GMT-oNF series, as well as Inopor and Pervap
School of Chemical Engineering and Analytical ceramic membranes. The feasibility of API
Science, The University of Manchester, degenotoxication by organic solvent
Manchester, UK nanoltration (OSN) has been successfully
applied in dialtration mode where fresh solvent
is continuously added to the retentate in order to
Synonyms push the relatively small genotoxic molecules
through the membrane, while the relatively
Degenotoxication large API molecules are retained as depicted in
Fig. 1 (Szkely 2011). The aim is achieving high
genotoxin removal and low API loss which G
Definition mainly depends on the molecular size gap
between the impurity and the API as well as the
Genotoxin removal by membranes is an emerg- polarity of the impurity. Hybrid processes featur-
ing purication technology aiming to reach ing OSN and adsorbents have been developed for
ultralow genotoxic impurity levels in order to the removal of genotoxic palladium, impurity
address health and environmental regulations. retained (Pink 2008), and 1,3-diisopropylurea,
API retained (Szkely 2012), from post-reaction
solutions. In the former case, the Suzuki post-
Characteristics reaction solution was subjected to OSN, and the
palladium was retained by the membrane, while
The main challenge of genotoxin removal by the API was collected in the permeate. On the
membranes lies in the fact that most of them are contrary, in the latter case, the genotoxin was
highly reactive species and thus can attack poly- removed with the permeate, while the API was
meric membrane framework at molecular level. retained. In both cases the genotoxin levels were
Degenotoxication of active pharmaceutical reduced further by using commercially available
ingredients (API) is of utmost importance to adsorbents or tailor-made molecularly imprinted
drug manufacturers. Genotoxicity of pharmaceu- polymers. Not only has the combination of
tical impurities is being continuously evaluated molecular imprinting and membrane technology
increasing the list of such highly toxic impurities. provided an efcient genotoxin removal process
Pharmaceutical regulatory authorities draw the but their fusion as well. Molecularly imprinted
attention to API purication and strict the thresh- membranes (MIM) can also be used for
old limits of genotoxins via annually issued API degenotoxication. Potentially genotoxic
guidelines. Besides the conventional API down- 4,40 -methylenedianiline has been applied as a
stream processes such as recrystallization, frac- template for MIM adsorbents (De Luca 2011).
tional distillation, chromatography, extraction, Environmental and economic analysis proved the
and adsorbents, new technologies are being sustainability of API degenotoxication by
implemented in the pharmaceutical industry. membrane processes in comparison with conven-
Application of nanoltration alone and com- tional API purication methodologies such as
bined with adsorbents are proposed as alternative chromatography and recrystallization (Szkely
purication technologies assuring ultralow 2013). Besides pharmaceutical purication, con-
genotoxin contaminations in drug products. API trol of genotoxins is also an environmental con-
post-reaction streams usually feature organic sol- cern. Nanoltration featuring DESAL 51 HL,
vents, limiting the selection of membranes to be DESAL 5 DL, UTC-20, and UTC-60 membranes
used in the process. Commercially available was applied for the removal of genotoxic
876 Germanium Recovery by Ion Exchange Membranes
Genotoxin Removal,
Fig. 1 Scheme of
genotoxin removal by OSN
with potential reverse
osmosis for solvent
recycling
atrazine banned as a total herbicide in Europe degenotoxication processes. Green Chem.

since 1991 from different water matrices doi:10.1039/c2gc36239b
Zhang Y, Van der Bruggen B, Chena GX, Braeken L,
including distilled water, tap water, and river Vandecasteele C (2004) Removal of pesticides by
water (Zhang 2004). nanoltration: effect of the water matrix. Sep Purif
Technol 38:163172
References
De Luca G, Donato L, Del Blanco SG, Tasselli F,

Drioli E (2011) On the cause of controlling afnity to
Germanium Recovery by Ion
small molecules of imprinted polymeric membranes Exchange Membranes
prepared by noncovalent approach: a computational
and experimental investigation. J Phys Chem H. Kawakita
B 115:93459351
Department of Applied Chemistry,
Pink CJ, Wong H, Ferreira FC, Livingston AG
(2008) Organic solvent nanoltration and adsorbents; Saga University, Saga, Japan
a hybrid approach to achieve ultra low palladium con-
tamination of post coupling reaction products. Org
Process Res Dev 12(4):589595
Szkely GY, Bandarra J, Heggie W, Sellergren B, Ferreira
Germanium recovery using catechol complex-
FC (2011) Organic solvent nanoltration: a platform ation and membrane permeation (also called
for removal of genotoxins from active pharmaceutical germanium-recovery-membrane system) is the
ingredients. J Membr Sci 381(12):2133 unique way to recovery germanium. Now, germa-
Szkely GY, Bandarra J, Heggie W, Sellergren B, Ferreira
FC (2012) A hybrid approach to reach stringent low
nium is a rare semimetal and used for semicon-
genotoxic impurity contents in active pharmaceutical ductor, catalysis, and optical apparatus. Due to
ingredients: combining molecularly imprinted poly- heightened interest in renewable energy sources,
mers and organic solvent nanoltration for removal of the production of solar panels has increased. In
1,3-diisopropylurea. Sep Purif Technol 86:7987
production of solar panel, germanium is doped to
Szkely GY, Gil M, Sellergren B, Heggie W, Ferreira FC
(2013) Environmental and economic analysis silicon compound to change the energy gap.
for selection and engineering sustainable API Processing of solar panel production generates
Germanium Recovery by Ion Exchange Membranes 877
Germanium Recovery by Ion Exchange Membranes, Fig. 1 Germanium recovery system using catechol complex-
ation and membrane permeation
the mixed waste of germanium and silicon. For

recycling, germanium as a rare metal should be
recovered from germanium-containing waste.
Using the polymerization technique, the func-
tional group, polyol group, was introduced to the
polymerized resin to recovery germanium along
with the diffusion. The polyol-containing polymer
brush was introduced to the membrane and the
germanium solution was own through the mem-
brane, achieving the high-speed recovery via con-
vection. Catechol forms the complex with
germanium as an anion. This complexation occurs
selectively compared with silicate ion. We pro-
pose the germanium recovery system using mem-
brane and resin, as shown in Fig. 1 (Nozoe
et al. 2012): (1) germanium ion selectively forms
the complex with catechol in solution, forming
anion complex, and (2) germanium complex was
captured to the quaternary-amino-group-
introduced porous membrane and resin. This
study addresses the selective complexation of ger- Germanium Recovery by Ion Exchange Membranes,
manium in the solution and the recovery of anion Fig. 2 pH dependence of germanium recovery against
complex with membrane and resin continuously. silicate ion in a batch mode
In the presence of silicate ion, pH dependence of
germanium recovery was checked in a batch mode, curves at the initial pH of 2.0 and 11.0, as shown in
as shown in Fig. 2. In lower pH, the selective Fig. 3. At pH 11.0, germanium ion was adsorbed
recovery against silicate ion was achieved in more compared with that at pH 2.0, while the
changing the concentration ratio, because the com- selectivity against silicate ion was higher. This
plex formation easily forms in lower pH region. In process is applicable for the wastewater of the
continuous mode, we determined the breakthrough solar panel containing germanium.
878 Glass Transition Temperature (Tg)
Germanium Recovery by Ion Exchange Membranes, Fig. 3 Breakthrough curves of germanium recovery at pH of
11.0 and 2.0
References At the molecular scale, the Tg of a polymer is

the temperature above which large segmental
Nozoe A, Ohto K, Kawakita H (2012) Germanium recov- motions of the polymer chains become possible
ery using catechol complexation and permeation
within the time scale of the experiment. The glass
through an anion-exchange membrane. Sep Sci
Technol 47:6265 transition temperature depends on the molecular
architecture. Substituents that restrict the back-
bone rotation of a very simple polymer, such as
polyethylene, will increase its Tg, and the pres-
Glass Transition Temperature (Tg) ence of polar groups will have an even stronger
effect (Table 1). Besides the substituents on a
Johannes Carolus Jansen exible polymer backbone, the chemical struc-
Institute on Membrane Technology, National ture of the backbone itself obviously has a
Research Council of Italy, ITM-CNR, Rende, major impact on the glass transition temperature.
Italy The presence of sterically hindered groups,
conjugated bonds, fused rings, etc., increases
the glass transition temperature signicantly. In
General Introduction the extreme case, the mobility is so much
restricted that a Tg is no longer observed. For
At the macroscopic scale, the glass transition tem- instance, polymers of intrinsic microporosity
perature, Tg, represents the temperature above (PIMs), a novel class of polymers with growing
which a material changes from a stiff glass into a interest for their high gas permeability
viscous uid or a rubbery material. Besides poly- (McKeown and Budd 2010), consist of ladder
mers, which are the most common materials with structures with such a high rigidity that large-
a glass transition temperature, also various amor- scale motions are impossible and they do not
phous solids, organic liquids, alloys, or inorganic exhibit a glass transition below their degradation
glasses may exhibit a glass transition. temperature.
Glass Transition Temperature (Tg) 879
Glass Transition Temperature (Tg), Table 1 Glass

transition temperature of polymers -[-CH2CH(R)-]n- with
the same backbone and different substituents R (Brandrup
et al. 1999)
Polymer R Tg
Polyethylene H 155
Polypropylene CH3 258270
Poly(vinyl uoride) F 314
Poly(vinyl chloride) Cl 354
Poly(vinyl alcohol) OH 358
Polystyrene C6H5 373
Poly(vinyl acetate) CH3COO 305
The glass transition temperature of polymer G

blends depends on their miscibility. In the rare
Glass Transition Temperature (Tg), Fig. 1 Schematic
case of miscibility at the molecular level, the representation of a DSC trace (bottom) and of the specic
glass transition of the blend has a value between volume (top) of an amorphous polymer
those of the neat polymers. Different equations
describe the behavior of polymer blends more or
less satisfactorily.
In the Fox equation, the Tg of the blend in the specic heat of the sample (Fig. 1). Various
depends exclusively on its composition: other properties undergo more or less pronounced
changes at the Tg, such as the density, the specic
1 w1 w2 heat, the elasticity coefcient or Youngs modu-
(1)
T g T g ,1 T g ,2 lus, the rate of diffusion of gases or liquids
through the polymer, the thermal expansion coef-
where w1 and w2 are the weight fractions and Tg,1
cient, etc.
and Tg,2 are the glass transition temperatures of
There is a clear correlation between the glass
pure polymer 1 and polymer 2, respectively.
transition temperature and the transport properties
The Gordon-Taylor equation is able to describe
in dense membranes (Matteucci et al. 2006). The
slightly asymmetric dependencies of the Tg on the
gas and vapor permeability is much higher in the
blend composition, by means of an adjustable
rubbery state than in the glassy state, and the
parameter, KGT:
selectivity is usually lower. Therefore, the value
w1 T g,1 K GT w2 T g,2 of the Tg is one of the main parameters inuencing
Tg (2) the membrane performance. For most other mem-
w1 K GT w2
brane applications, also those using porous mem-
The Kwei equation introduces a binary parameter, branes, the glass transition temperature is of large
q, which represents the interaction between the interest too. Since porous membranes may col-
two polymers, and can be used to describe even lapse upon softening, it is of fundamental impor-
more asymmetric Tg versus composition proles tance that they are operated at a temperature
(ElMiloudi et al. 2009): sufciently far below the Tg.
w1 T g,1 K Kwei w2 T g,2
Tg qw1 w2 (3)
w1 K Kwei w2
References
Glass Transition Temperature Analysis
Andrews RJ, Grulke EA (1999) Glass Transition Temper-
atures of Polymers. In: Brandrup J, Immergut EH,
Besides a change in the mechanical properties, the Grulke EA (eds) Polymer handbook, 4th edn. Wiley,
Tg is also accompanied by a fairly abrupt change Hoboken
880 Glass Transition Temperature Depression
ElMiloudi K, Djadoun S, Sbirrazzuoli N, Geribaldi S

(2009) Miscibility and phase behaviour of binary and Glassy Membranes
ternary homoblends of poly(styrene-Co-Acrylic acid),
poly(styrene-Co-N, N-Dimethylacrylamide) and Poly
(styrene-co-4-Vinylpyridine). Thermochim Acta Johannes Carolus Jansen
483(12):4954 Institute on Membrane Technology, National
Matteucci S, Yampolskii Y, Freeman BD, Pinnau I (2006), Research Council of Italy, ITM-CNR, Rende,
Transport of gases and vapors in glassy and rubbery
polymers, In: Yampolskii Y, Pinnau I, Freeman BD Italy
(eds.) Materials science of membranes for gas and
vapor separation, John Wiley & Sons, Ltd, Chichester,
UK. doi: 10.1002/047002903X.ch1 Glassy membranes are membranes consisting of
McKeown NB, Budd PM (2010) Exploitation of intrinsic
microporosity in polymer-based materials. Macromol- amorphous polymers which are in their glassy
ecules 43(12):51635176 state at room temperature or under the normal
operation conditions, i.e., their glass transition
temperature is above room temperature or the
operating temperature. The high stiffness of the
glassy polymer provides sufcient mechanical
Glass Transition Temperature strength to the membranes to exist also as porous
Depression or dense integrally skinned at lms or hollow
bers.
Johannes Carolus Jansen The polymer stiffness is an essential aspect in
Institute on Membrane Technology, National the membrane formation process by non-solvent
Research Council of Italy, ITM-CNR, Rende, induced phase inversion, where at a certain point
Italy of the process solidication of the polymer is
required to consolidate the morphology of the
membrane. In glassy polymers, this is typically
Glass transition temperature depression is the by vitrication of the polymer when diffusion of
phenomenon which describes the reduction of the solvent from the polymer rich phase into the
the glass transition temperature by external fac- coagulation bath leads to a gradual increase of
tors, usually by the presence of solvent molecules the glass transition temperature until this exceeds
or of other additives in the polymer matrix. the coagulation bath temperature.
Such additives have the capacity to enhance
the mobility of the polymer chains, thus enabling
long range motions at lower temperature than
in the neat polymer. Although this often goes Glassy Polymer
hand in hand with the phenomenon of plasticiza-
tion, a reduction of the elastic modulus at room Johannes Carolus Jansen
temperature, the two concepts should not be Institute on Membrane Technology, National
confused. Plasticization strictly refers to the Research Council of Italy, ITM-CNR, Rende,
mechanical properties of the polymer and Italy
the molecules, which reduce the elastic modulus
of a polymer by an increase of the chain
mobility generally also reduce their glass transi- Strictly, any amorphous polymer is glassy at tem-
tion temperature. The opposite is not necessarily peratures below its Tg, but in practice a glassy
the case and it may happen that an additive, polymer refers to those polymers which are in
which lowers the Tg, at low concentration their glassy state at room temperature, thus to
increases the elastic or Youngs modulus at amorphous polymers with a glass transition tem-
room temperature. This is then referred to as perature above room temperature, in contrast to
anti-plasticization. rubbery polymers, which have a Tg below room
Global Optimization of Membrane Processes 881
temperature, and (semi-)crystalline polymers with where x represents vector of optimization

a melting point above room temperature. Under (decision) variables, f(x) is optimization criterion,
particular circumstances also semi-crystalline h(x) and g(x) stand for vectors of equality and
polymers can become glassy after quenching, inequality constraints, respectively.
if the crystallization kinetics and the crystal To nd the global optimum, many methods
nucleation rate are sufciently slow to prevent have been developed. These can be basically clas-
crystallization upon rapid cooling. A typical sied into two branches, stochastic and determin-
example of such polymer is poly(ethylene istic ones. Stochastic methods (Weise 2007) such
terephthalate), PET. as simulated annealing, genetic algorithms, and
Glassy polymers are characterized by their particle swarm optimization got inspired by natu-
high stiffness and amorphous non-crystalline ral phenomena. In general, they randomly search
structure. These properties make them size selec- through the solution space (space where con-
tive in their dense form and also mechanically straints are satised) and due to that usually do
sufciently resistant to exist as porous at lm not require gradient information. It is often G
or hollow bre membranes. Examples of the most reported that weakness of these algorithms is
commonly used polymers in commercial mem- that they cannot guarantee convergence to global
branes are cellulose acetate, polysulfone, poly optimum in nite time.
(ether sulfone), polyimide, and polycarbonate. On the other hand, deterministic methods
search through the whole solution space system-
atically and thus can provide so-called
e-convergence (e is arbitrarily small positive num-
ber) guarantee to the global optimum. Most pop-
Global Optimization of Membrane
ular are methods based on branch-and-bound
Processes
algorithms and interval analysis (Horst and Tuy
1990).
Radoslav Paulen
The latter type of methods was successfully
Department of Biochemical and Chemical
used in Chen et al. (2006) for globally optimal
Engineering, Technische Universtt Dortmund,
design of membrane fuel cell. Other applications
Dortmund, Germany
of methods of GO involve parameter estimation
(Papamichail and Adjiman 2004), optimal control
problems (Esposito and Floudas 2000), and prob-
For a number of years, researchers and engineers
lems of nonlinear and mixed integer nonlinear
in many elds have reported the problem of
programming (Chachuat et al. 2006). Other refer-
obtaining multiple solutions while solving
ences on GO problems in chemical engineering
problems of nonlinear programming (NLP).
can be found in Rangaiah (2010).
Multiplicity of solutions is caused by the
nonconvex nature being present typically in
many engineering problems. Global optimization
References
(GO), sometimes called nonconvex optimization,
encompasses several techniques which can be Chachuat B, Singer AB, Barton PI (2006) Global methods
used to handle the problem of nding the best of for dynamic optimization and mixed-integer dynamic
these solutions, the global optimum. optimization. Ind Eng Chem Res 45(25):83738392
General form of GO problem is same as for Chen KI, Winnick J, Manousiouthakis VI (2006) Global
optimization of a simple mathematical model for a
NLP problem and it may be stated as proton exchange membrane fuel cell. Comput Chem
Eng 30(8):12261234. doi:10.1016/j.compch-
min f x emeng.2006.02.009, ISSN 0098-1354
x Esposito W, Floudas C (2000) Deterministic global opti-
s:t: hx 0 mization in nonlinear optimal control problems. J Glob
g x 0 Optim 17:97126
882 Gluconic Acid Production in Hybrid System Membrane Bioreactor with ED
Horst R, Tuy H (1990) Global optimization deterministic methods, i.e., fermentation and chemical synthe-
approaches, 3rd edn. Springer, Berlin sis. In either fermentation or chemical synthesis,
Papamichail I, Adjiman CS (2004) Global optimiza-
tion of dynamic systems. Comput Chem Eng separation, concentration, and purication are
28:403415 necessary for the acid production. Due to the
Rangaiah GP (2010) Stochastic global optimization: tech- excellent performance of membrane separation,
niques and applications in chemical engineering, membrane technology such as electrodialysis
Advances in process systems engineering. World Sci-
entic, Hackensack. ISBN 9789814299206 (ED) can provide a novel alternative to conven-
Weise T (2007) Global optimization algorithms theory and tional procedures for organic acid production
application. http://www.it-weise.de/ (Huang et al. 2007).
ED is a kind of technology which arranges
ion-exchange membranes alternately in a direct
current eld (see Fig. 1). The ion-exchange mem-
Gluconic Acid Production in Hybrid branes include cation-selective, anion-selective,
System Membrane Bioreactor and bipolar membranes. ED technology includes
with ED conventional electrodialysis (CED), electrome-
tathesis (EMT), electro-ion substitution (EIS),
Zhiwei Wang electro-electrodialysis (EED), electrodialysis
Department of Environmental Engineering, with bipolar membranes (EDBM), electrodeio-
College of Environmental Science and nization (EDI), and two-phase electrodialysis
Engineering, Tongji University, Shanghai, China (TPED), which can be applied for organic acid
production (Huang et al. 2007).
Electrodialysis with bipolar membranes
Synonyms (EDBM) has been proved to provide a feasible
solution for gluconic acid production (Novalic
Organic acid production by membrane biotech- et al. 2000). EDBM is a kind of technology that
nology with electrodialysis (ED) integrates solvent and salt dissociation. Com-
pared to conventional organic acid production
(precipitation and acidication), it can realize
Introduction salt conversion without second salt pollution or
provide H+ and OH/alkoxide ions in situ with-
Organic acids are a group of organic compounds out salt introduction (Huang and Xu 2006). In lab
with acidic properties. The most common scale EDBM system for gluconic acid produc-
organic acids are the carboxylic acids (e.g., tion, the conversion rate of sodium gluconate
gluconic acids) containing carboxyl group could reach no less than 98 %, with current
COOH and sulfonic acids comprising the efciency around 70 % and energy consumption
group SO2OH. In addition, organic compounds, about 1 kWh/kg acid (Yu et al. 2000; Huang
which have hydroxyl OH, thiol group SH, et al. 2008).
the enol group, and the phenol group that When a reactor or bioreactor is integrated with
present weak acidity, are also generally referred ED, the combined process can realize in situ sep-
to as organic acids. Organic acids have been aration, purication, and organic acid production
widely used in foods, beverages, pharmaceuti- adjustment without apparent retrogression in its
cals, cosmetics, detergents, plastics, resins, and function. For instance, the use of fermentation to
other biochemical or chemical products. As a produce acids can be integrated into a membrane
result, they have a close relationship with bioreactor system coupled with ED. However,
humans daily life. As one type of organic acid, some improvements are needed for modifying
gluconic acid has been used as food additive, the technology, which include reducing mem-
acidity regulator, and chelating agent. Organic brane cost (especially bipolar membrane),
acids can be produced by two prevailing improving the antifouling properties of
Glucosidase in Membrane Bioreactor 883
Gluconic Acid Anion-selective Cation-selective Bipolar

Production in Hybrid membrane membrane membrane
System Membrane + + +
Bioreactor with ED, + X +
Fig. 1 Schematic of + +
ED. M+: cation; X: anion X
+ +
M+ X
+ +
+ + +
M+
+ +
+ M+
+
+
+ +
M X
+ + +
Cathode Anode
membranes, and increasing the selectivity for

co-ions and developing ion-exchange membranes Glucosidase in Membrane Bioreactor
with a specic selectivity for organic anions or
hydrogen ions (Huang et al. 2007). To sum up, Rosalinda Mazzei
membrane biotechnology with ED is a promising Institute on Membrane Technology, National
solution for organic acid (e.g., gluconic acid) pro- Research Council of Italy, ITM-CNR, Rende,
duction (See also organic acid production by Italy
membrane technology with ED).
b-Glucosidase is the enzyme responsible for the

Cross-References oleuropein hydrolysis; it belongs to the group of
hydrolases widely existing in various sources,
Membrane bioreactors for treatment of tannery such as fungi, bacteria, plant, and animal tissue.
efuents The broad optimum pH and temperature range
Membrane biotechnology makes this enzyme particularly useful for
Ultraltration membrane bioreactor application in biotechnology processes. The
b-glucosidase from different sources has been
studied in various systems, as free or attached
References
to different supports (Yeoh 1991; Esen 1993;
Huang C, Xu T (2006) Electrodialysis with bipolar mem- Montero and Romeu 1993).
branes for sustainable development. Environ Sci A lot of interest was demonstrated in the enzy-
Technol 40:52335243 matic/substrate system b-glucosidase/oleuropein
Huang C, Xu T, Zhang Y, Xue Y, Chen G (2007) Applica-
due to the fact that the oleuropein hydrolysis
tion of electrodialysis to the production of organic
acids: state-of-the-art and recent developments. products are phytochemicals that can be poten-
J Membr Sci 288:112 tially used, if obtained in pure form, in pharma-
Huang C, Li Y, Xu T, Li H, Yang W (2008) Pilot-scale ceutical and food eld. In addition oleuropein
production of gluconic acid using bipolar membranes.
hydrolysis is very important because it is present
J Univ Sci Technol China 38:656659
Novalic S, Kongbangkerd T, Kulbe KD (2000) Recovery in Olive Mill Waste (OMW), following the strat-
of organic acids with high molecular weight using a egy of waste valorization.
combined electrodialytic process. J Membr Sci In in vivo system, olives and leaves, oleuropein
166:99104
is hydrolyzed by b-glucosidases action. In the rst
Yu LX, Lin AG, Zhang LP, Chen CX, Jiang WJ
(2000) Application of electrodialysis to the production reaction step, the products are oleuropein aglycon
of Vitamin C. Chem Eng J 78:153157 with a glucose molecule. Oleuropein aglycon, due
884 Gold Recovery by Supported Liquid Membranes
to its poor water solubility, is not yet puried and reactor containing immobilized -glucosidase for
not yet commercially available, because after its the hydrolysis of oleuropein. J Membr Sci 339:
215223
production by enzymatic hydrolysis, it is fastly Mazzei R, Drioli E, Giorno L (2010) Biocatalytic mem-
rearranged into water-soluble compound. brane reactor and membrane emulsication concepts
Few examples are reported in the literature in combined in a single unit to assist production and
which b-glucosidases are used immobilized on separation of water unstable reaction. J Membr Sci
352:166172
membranes (Mazzei et al. 2009). Mazzei R, Drioli E, Giorno L (2012) Enzyme membrane
A hybrid membrane operation system is devel- reactor with heterogenized b-glucosidase to obtain
oped to (i) compartmentalize the oleuropein phytotherapic compound: optimization study.
hydrolysis, occurring in aqueous phase within J Membr Sci 390391:121129
Montero MA, Romeu A (1993) Properties of almond
the porous membrane, and (ii) separate and stabi- b-glucosidase immobilized on concanavalin
lize the aglycon, occurring in organic phase at the Asepharose. Biochem Mol Biol Int 30:685689
membrane lumen side interface, based on the Yeoh H (1991) Kinetic properties of Cassava
membrane emulsication concept. b-glucosidase immobilized in photo-crosslinkable
resins. Biotechnol Lett 13:697700
The biotechnology used to conduct the hydro-
lysis of oleuropein is a multiphase biocatalytic
membrane reactor.
Polymeric membranes were heterogenized
with a b-glucosidase. The OMW, containing Gold Recovery by Supported Liquid
oleuropein, passed through the enzyme-loaded Membranes
membrane where the reaction occurred. The
water phase containing reaction products passed Pietro Argurio
through the membrane and met an organic Department of Environmental and Chemical
phase recirculated into the lumen side where Engineering, University of Calabria, Arcavacata
water-in-oil emulsions were formed drop by di Rende, Italy
drop. Droplet size and stability were adapted
so that enough exchange area interface was
achieved, and once aglycon extraction was Gold is the chemical element which symbol is
completed, droplets broke and phases separated Au (from Latin aurum) and the atomic number
spontaneously. The integrated system biocata- is 79. The oxidation states in its compounds
lytic membrane reactors and membrane emulsi- range from 1 to +5, and Au(III) is the most
cation process showed that it is possible to common. Typical Au(I) complex is Au(CN)2
produce and simultaneously isolate the isomer (in cyanide media) which is the soluble form
of oleuropein aglycon in one step, starting from of gold encountered in mining, while in chloride
OMW (Mazzei et al. 2009, 2010, 2012; Conidi media Au(III) complexes (Au2Cl6) are the
et al. 2014). typical ones.
Gold production by means of its extraction
from mining can represent an important contri-
bution to environmental pollution. Metal ores,
References which generally contain less than 1 ppm of
gold, are ground and mixed with sodium or
Conidi C, Mazzei R, Cassano A, Giorno L (2014) Inte-
potassium cyanide for gold chemical extraction.
grated membrane system for the production of
phytotherapics from olive mill wastewaters. J Membr Precious metal and heavy metal impurities such
Sci 454:322329 as cadmium, lead, zinc, copper, nickel, and arse-
Esen A (1993) b-Glucosidases: biochemistry and molecu- nic are usually present in an anionic form (i.e.,
lar biology. American Chemical Society, Washington,
cyanide salts) after their extraction from metal
DC
Mazzei R, Giorno L, Piacentini E, Mazzuca S, Drioli ores. These salts are toxic to the liver and kidneys
E (2009) Kinetic study of a biocatalytic membrane because of both cyanide and metal content.
Gold Recovery by Supported Liquid Membranes 885
Thus, it would be desirable to be able to selec- Referring to gold transport, SLM systems have
tively remove these complexes for the recovery been tested in the separation of this precious metal
of precious metals. mainly from cyanide (Au(I)) or chloride (Au(III))
Gold, over its use in jewelry, also has a wide media. Highly acidic conditions are required for
use in various industries, thanks to its physical and the extraction and transport of Au(III) because of
chemical properties. Thus, the recovery of this its easily reducible nature.
metal from the different wastewater generated by Various carriers for Au transport across a SLM
these industries is also of growing interest were reported in literature. Among them, the com-
(Alguacil 2004). mercially available extractant Cyanex 921 has
Gold can be recovered from different been applied in the carrier-facilitated transport of
aqueous solutions by various physicochemical both gold(I) and gold(III) recovery from cyanide
separation techniques as precipitation, ion and chloride media, respectively (Alguacil
exchange, carbon adsorption, cementation, sol- et al. 2005). Gold(I) is transported from alkaline
vent extraction, etc. In the case of convectional pH values (611). Gold(I) extraction takes place G
solvent extraction, metal ion-containing solution in that pH range by the following equilibrium
is placed in contact with a large amount of an reaction:
appropriate organic phase (Kargari et al. 2004).
This extraction step is followed by a stripping
one. The main drawback of solvent extraction M AuCN2aq 3 Lorg
is the large amount of solvent required when , M AuCN2 3Lorg
dilute solutions were processed, making this pro-
cess not very cheap and safe, since the used
solvents are often chlorinated and sometimes where the subscripts aq and org denote the species
carcinogenic. contained in the aqueous and organic phase,
Liquid membrane (LM)-based processes have respectively, and L represents the extractant.
become an attractive alternative to conventional In the case of gold(III), the extraction is
techniques for selective separation and con- governed by the following pH-dependent equilib-
centration of both organic and inorganic com- rium reaction:
pounds from dilute aqueous solutions since
they combine extraction and stripping into a sin- H aq AuCl4 nLorg , H AuCl4 nLorg
gle process, thus reducing the solvent inventory
requirement and then cost signicantly (Molinari
and Argurio 2011). They also allow the use of where n = 1, 2.
expensive and highly selective extractants,
which otherwise would be uneconomic in sol-
vent extractions. References
LM systems include nonsupported liquid
Alguacil FJ (2004) Carrier-mediated gold transport in the
membranes (bulk liquid membrane (BLM) system Cyanex 921HClAu(III). Hydrometallurgy
and emulsion liquid membrane (ELM)) and 71:363369
supported liquid membrane (SLM). SLMs con- Alguacil FJ, Alonso M, Sastre AM (2005) Facilitated
sist of an organic LM phase impregnated in a thin supported liquid membrane transport of gold (I) and
gold (III) using Cyanex 921. J Membr Sci
hydrophobic microltration membrane. This LM 252:237244
phase generally contains an extractant (carrier) Kargari A, Kaghazchi T, Soleimani M (2004) Role of
which binds very selectively the target compo- emulsier in the extraction of gold (III) ions from
nent in the donor phase (feed), transporting aqueous solutions using the emulsion liquid membrane
technique. Desalination 162:237247
it into the acceptor phase (strip), resulting in Molinari R, Argurio P (2011) Recent progress in supported
the so-called facilitated transport (Molinari liquid membrane technology: stabilization and feasible
et al. 2009a, b). applications. Membr Water Treat 2(4):207223
886 Graft Polymerization
Molinari R, Argurio P, Poerio T (2009a) Studies of homo-polymers or copolymers as illustrated in

various solid membrane supports to prepare stable Fig. 1 (A, B, and C are representing different
sandwich liquid membranes and testing copper
(II) removal from aqueous media. Sep Purif Technol types of monomers).
70:166172 Grafting can be accomplished by either
Molinari R, Argurio P, Poerio T (2009b) Flux enhancement grafting to or grafting from approaches. In
of stagnant sandwich compared to supported liquid grafting to approaches, functionalized mono-
membrane systems in the removal of Gembrozil
from waters. J Membr Sci 340:2634 mers react with the backbone polymer to form
the grafted one. On the other hand, grafting
from is achieved by treating a substrate with
some method to generate immobilized initiators
Graft Polymerization followed by polymerization. High grafting den-
sity polymer also can be accomplished using this
Tauqir A. Sherazi technique (Bhattacharya et al. 2009).
Department of Chemistry, COMSATS Institute of Graft copolymerization can be initiated by var-
Information Technology, Abbottabad, Pakistan ious methods including chemical treatment, pho-
tochemical treatment, ionizing radiation (such as
gamma radiation, electron beam radiation, etc.),
Graft polymerization is a process in which mono- photo-irradiation, plasma-induced techniques and
mers are covalently bonded and polymerized as side enzymatic grafting, etc. Grafted polymers can be
chains onto the main polymer chain (the backbone).
very useful as they can be tailored to the require-
Grafting is an attractive approach to impart a ments of particular applications by appropriate
variety of functional groups to a polymer. Graft selection of backbone and monomers to be
polymers are also known as graft copolymer since grafted.
it contains at least two different kinds of monomer Grafted polymers have wide range of applica-
units such as the grafted side chains that are struction such as in the eld of biomedical, textiles,
turally distinct from the main chain. The monomer automobiles, cable technology, separation and
to be grafted may be of one or more than one type; purication, electrolyte membranes, coatings,
thus, the graft chains in grafted copolymer may be adhesives, laminates, commodity plastics, etc.
graft homo-polymer graft copolymer

B
B B
B C C C
B B C
B C
base polmer C C
C
A A A B C
A A A A A A A A
A A A A A A A C
A A A A A A A A A
B A A A A
C C B
B C
B C B
B B
B C C
B B
B C graft homo-polymer B
B
graft copolymer B
B
Graft Polymerization, Fig. 1 Structural representation of graft copolymer

Grain Models 887
Cross-References alternatively, toward some xed center of attraction

(Finney 1970). Ballistic random sphere packs are
Surface Modication: Chemical Treatment usually driven by gravity and are considered to be
more representative of the compaction process. Bal-
listic deposition and central attraction methods pro-
References
vide packings with similar structural properties in
Bhattacharya A, Rawlins JW, Ray P (2009) Polymer terms of the autocorrelation function. Depending on
grafting and crosslinking. John Wiley & Sons, Inc., the type of fabrication process as well as on the
Hoboken, New Jersey characteristic dimensions of the grains, other forces,
either local/short ranged (van der Waals, double
layer, hydrodynamic) or long ranged (electrostatic,
electromagnetic), can also be taken into account
Grain Models during simulations. One, two, or three points of
contact can be assumed to discontinue the downhill G
Vasilis Burganos1 and Eugene Skouras2 motion of the particle. In energy-based models,
1
Institute of Chemical Engineering Sciences, downhill motion of a particle continues until a posi-
Foundation for Research and Technology tion of local energy minimum is reached that is
Hellas, FORTH/ICE-HT, Patras, Hellas considered stable. Such a procedure can be system-
2
Institute of Chemical Engineering Sciences, atically followed through the use of a steepest
Foundation for Research and Technology, Hellas, descent method followed by a conjugate gradient
Patras, Hellas algorithm (Conway and Sloane 1993). Procedures
based on Monte Carlo or similar statistical methods
have also been implemented, where each time a
Grain models involve a set of grain-shaped objects to number of test grains are inserted, but only selected
represent the solid phase in the interior of porous displacements are allowed that lead to minima of
materials. Model representation of the structure of position or energy. Motion and interaction of granu-
porous membranes is very important for visualizing lar particles have also been studied mechanistically
and understanding the structure of amorphous, with computational methods, such as the discrete
porous, and random membrane materials and can element method (DEM), also called a distinct ele-
be implemented, among other ways, by determinis- ment method (Zhu et al. 2007, 2008), for computing
tic or stochastic modeling of the outcome of the the motion and effect of a large number of small
actual fabrication process (e.g., random assembly particles. Although DEM is closely related to molec-
or sequential deposition of particles) with or without ular dynamics, the method is generally differentiated
reference to the detailed physics of the process by its inclusion of rotational degrees of freedom, as
(Preparata and Shamos 1985). Typically, such recon- well as compacted and other complex granular
structions involve some random packing, optionally geometries. Grain models have also been used for
combined with ballistic deposition of hard or soft the description of micro- or nanoparticle additives,
spheres, disks, or ellipsoids of prolate or oblate also called nanollers (at least one of the particles
geometry (Rogers 1964). This stage is usually dimensions is in the nanometer range), into a poly-
followed by simulation of grain sintering as the mer membrane to form a polymeric composite mem-
result of thermal or viscous sintering. In the majority brane or a hybrid, mixed matrix membrane (Sanchez
of packing procedures, random sequential deposi- et al. 2005). This is a promising idea to improve the
tion of overlapping or non-overlapping particles separation properties of the neat polymer and pro-
takes place (Jaeger and Nagel 1992). The particles duce the so-called nanobarriers with numerous
position themselves either randomly or under the applications in gas separation, fuel cells, preserva-
inuence of some unidirectional force (ballistic) or, tion of sensitive products, etc. The enhancement of
888 Granular Media Pretreatment Filtration
Grain Models,
Fig. 1 Ballistic
reconstruction followed by
partial sintering
barrier properties through the addition of inorganic

nanoparticles in polymer matrices can be studied Granular Media Pretreatment
with the help of grain models and elucidate their Filtration
complex dependence on interfacial interactions,
ller shape and orientation with respect to the trans- Graeme K. Pearce
port direction, ller size, distribution of inter-ller Membrane Consultancy Associates Limited,
distances, and degree of agglomeration (Fig. 1). Reading, RG8 7EJ, UK
References Definitions
Conway JH, Sloane NJA (1993) Sphere packings, lattices, The term granular media ltration refers to the
and groups, 2nd edn. Springer, New York use of packed bed of granular media to provide a
Finney JL (1970) Random packings and the structure of
simple liquids. I. The geometry of random close pack-
ltration stage between a raw water intake and a
ing, Proc R Soc Lond A-Math Phys Sci 319:479493 nal advanced treatment stage.
Jaeger HM, Nagel SR (1992) Physics of granular states.
Science 255(5051):15231531
Preparata FP, Shamos MI (1985) Computational geometry:
an introduction. Springer, New York
Objectives
Rogers CA (1964) Packing and covering. Cambridge Uni-
versity Press, Cambridge The objectives of pretreatment using a granular
Sanchez C, Julian B, Belleville P, Popall M (2005) Appli- media lter are to remove a proportion of partic-
cations of hybrid organic-inorganic nanocomposites.
J Mater Chem 15:35593592
ulate constituents that would otherwise foul a
Zhu HP, Zhou ZY, Yang RY, Yu AB (2007) Discrete par- downstream treatment stage such as reverse
ticle simulation of particulate systems: theoretical osmosis or ion exchange. If combined with the
developments. Chem Eng Sci 62:33783392 use of chemicals, or if the media acts as a biolter,
Zhu HP, Zhou ZY, Yang RY, Yu AB (2008) Discrete par-
ticle simulation of particulate systems: a review of
the granular media ltration stage can be used to
major applications and ndings. Chem Eng Sci reduce the concentration of dissolved organics
63(23):57285770 as well.
Granular Media Pretreatment Filtration 889
Description feedstream. Surface lters tend to be used for

feeds with a low solids content and need to be
In a granular media lter, the feed passes through cleaned regularly in order to have a commercially
a packed bed of granules, and contaminants are useful performance. The surface removal mecha-
progressively removed. Due to the ability of the nism is at its most useful in two situations, namely
lter to continue removing particles as the feed for particles which are quite large, e.g., >100 mm,
progresses through the bed, these lters are some- as for a screen product and for quite small parti-
times referred to as depth lters, though this cles, e.g., <10 mm, as for a membrane.
generic description also covers other types of l- Between these ranges, ie for particle sizes of 10
tration device. to 100 mm, the depth or gravity or depth ltration
The simplest granular media lter is a sand mechanism becomes very useful, since products
lter. The original slow sand lters used a very which use this mechanism have a high dirt-
low ltration velocity and relied on capturing holding capacity combined with an acceptable
particles on the surface where a biologically active removal efciency. The granular media lter is G
layer formed (the schmutzdecke layer). Rapid the most common example of a device which
gravity lters (RGF) using sand media were operates using this type of mechanism to remove
developed more than 100 years ago to increase particulates from a feedstream (Voutchkov 2010).
ltration capacity and reduce the footprint These lters have a nominal pore size consider-
required. These rapid sand lters used a much ably greater than the particles they are capturing.
higher linear velocity and a deep bed of granules Thus, for a sand lter with a grain size of normally
to increase dirt-holding capacity and convert the 500 mm (i.e., 0.5 mm) or greater, the nominal pore
device from a surface lter to a depth lter (Rich- size of the media is more than 100 mm, and there is
ardson et al. 2002). a wide distribution of channel sizes. In fact, the
The granular media lter is a development of absolute rating of such a lter will be of the same
these early sand lters. Different types of media magnitude as the particles used to form the media
are combined in current commercial products to bed. However, the bed can routinely capture par-
optimize performance. Generally, the lter ticles down to 10 mm.
operates with a downow loading cycle in which The key attribute of the granular media lter is
particles are removed from the feedstream. that the pore size at the surface of the bed is much
A periodic cleaning cycle is then undertaken greater than most of the particles in the feed chal-
referred to as backwash in which the ow direc- lenge. Particles can therefore enter the media and
tion is reversed, i.e., normally in upow. The ow start to travel with the feed ow through the bed as
rate during backwash is sufcient to cause the bed illustrated in Fig. 1. Potential capture sites are
to expand in volume, though it is not fully uid- created in the bed wherever particles pass close
ized. The efciency of the backwash ow may be to a grain of the media. Gravity forces attracting
well enhanced by the introduction of air known as the particle to stick to the surface of a grain over-
air scour. come the viscous drag forces which would tend to
displace the particle from the surface and continue
in suspension in the uid ow.
Particle Removal Mechanisms For small particles, the forces of gravity attrac-
tion are relatively weak; hence this type of ltra-
Filtration devices such as membranes or screens tion mechanism is only effective down to a certain
use a surface removal mechanism to capture par- particle size, typically around 5 or 10 mm. Below
ticles. The smallest particle that can pass through that, the gravity attraction forces diminish sharply,
such a lter is equivalent to the largest pore or reducing by a factor of 1,000 for every order of
aperture in the surface of the lter. The dirt- magnitude (ie factor of 10) reduction in particle
holding capacity of surface lters is limited by size, since the mass is related to the diameter by a
the area of the lter surface in contact with the cube law.
Flow streamline provides

Large particle approaching pore size of
multiple capture opportunities
bed likely to be sieved out, or may form
as it passes through the bed
a cake at the surface
Medium sized particle likely to be caught

by gravity attraction mechanism
Small particles may be captured by gravity attraction

mechanism if streamline passes very close to
the media surface or may evade capture
Granular Media Pretreatment Filtration, Fig. 1 Granular media bed cross section illustrating particle capture
mechanism
Due to the depth of the media and the tortuous below that of the nominal pore size, the removal
path created for the feed as it moves through the efciency is much lower during the initial ripen-
bed, relatively high removal rates can be ing period in which the media is conditioned by
achieved, even for particles signicantly below particles in the challenge. The conditioning parti-
the nominal pore size. Media beds of 500 mm cles form a ne surface coating on the media
sand particles can achieve removal efciencies which enhances removal. In addition, the lter
of 9099 % for particles down to 1020 mm. rating declines toward the end of the cycle as the
The nominal rating of the depth lter means differential pressure increases across the bed
that the removal efciency is variable and is assisting the viscous drag from the uid ow to
dependent on a host of environmental and oper- dislodge any loosely attached particulates.
ating parameters.
A limitation in this type of device is that large
particles can bridge the pores at the surface, cre- Media Characteristics
ating a cake layer. This effect is known as surface
blinding, and if it happens, it prevents the efcient Media Type
utilization of the pores in the depth of the media, Granular media products can utilize several dif-
bringing the ltration cycle to a premature end. ferent media types of different grain sizes and bed
The surface of the media therefore requires pro- depths to form the deep bed lter. The basic media
tection with a coarse pretreatment. Typically, a used in the devices include anthracite (a type of
screen with a 300 mm rating is used to prevent coal) and sand, with the additional option of a
blinding. layer of garnet. It is also quite common to use
Although high removal ratings can be achieved granular activated carbon (GAC), potentially in
for a granular media lter removing particles well addition to anthracite. Anthracite can allow a
bioltration mechanism to develop in which bac- Polyelectrolyte addition to the feed can also be
terial colonies supported on the surface of the used to increase velocities, and this is standard
media consume dissolved organic carbon, thereby practice where footprint is of critical importance,
improving organic removals. GAC enhances e.g., offshore.
bioltration characteristics, but note that it can
take 46 weeks for the bacterial colonies to Grain Size Diameter: Coarse Versus Fine
become fully effective. Generally, the holdup A typical grain diameter can vary from 0.5 mm for
time in the media bed needs to be at least 15 min a ne bed lter to 1.5 mm for a coarse bed. In a
for bioltration to be effective. DMF of anthracite and sand, it is common to use
A contrasting approach is to apply a chlorine an initial layer of coarse anthracite followed by a
residual to the feed to enhance ltration perfor- ne layer of sand to provide a graded bed with
mance and maintain disinfection in the bed. Cl2 ner removal performance nearer the outlet.
oxidizes a wide range of the contaminants in the Higher velocities are used for coarse beds.
feed improving removal efciencies. G
Bed Depth: Shallow Versus Deep
Number of Layers: Single or Multimedia The total bed depth of a shallow lter could be
The simplest granular media lters just use a around 1 m, whereas a deep bed lter could be
single type of media, usually sand, but also pos- from 2.5 to 3.0 m in depth (Voutchkov 2010). An
sibly GAC. A typical rating for such a lter is allowance of about 5 % needs to be provided to
20 mm. Standard pretreatment lters use a layer allow for settlement. For a shallow bed lter
of coarse anthracite followed by a layer of ne treating a good quality feed, a typical DMF bed
sand, the so-called dual media lter (DMF) to depth would be 0.5 m of anthracite followed by
provide a 10 mm rating. The media is chosen so 0.5 m of sand. A poorer feed quality would require
that the specic gravity of the upper layer is less greater depth of up to 1.5 m for each layer. For a
than that of the lower layer in order to maintain lter providing a bioltration action, a deep bed
stratication in long-term use (Wilf 2007). would be required.
For ner ltration duties of about 5 mm, a triple
media lter can be used in which a nal layer of
high density garnet provides a polishing step. Product Characteristics
There are also commercial products using four
layers, which allow higher velocities and there- Pressure Versus Gravity
fore smaller footprint. Four-layer lters claim to A pressure lter encapsulates the media bed in a
achieve removals down to 1 mm or even lower. large pressure vessel, of 2.4 m or up to 3 m in
Multilayer lters require grains with good sphe- diameter. Although ideally suited for small or
ricity so that the stratication can be maintained medium-sized plants, they are sometimes used
after backwashing. for larger plants in which there is an advantage
to maintain the pressure head of the feed. Pressure
Velocity Range systems also have the advantage of smaller foot-
A single layer lter would tend to use a ow rate print and shorter construction times.
of 512 m3/h per m2 of surface area of the bed, The other type of conguration is the gravity
i.e., a linear velocity of 512 m/h (25 gpm/ft2). lter in which the media bed is contained in a large
The velocity can be increased with the number of tank, potentially made from concrete. Gravity l-
layers and depth of the layers, depending on the ters are often used where there is no disadvantage
degree of removal required. DMF pretreatment in breaking the pressure head of the feed and
plants normally have a velocity of 812 m/h. where the plant layout lends itself to the use of
Four-layer lters can be operated at 2560 m/h, gravity to drive the feed through the bed. Gravity
a factor of ve times as high as a single lters are also normally preferred for plants with a
media, which reduces footprint signicantly. high capacity.
Vertical Versus Horizontal Backwash is mainly conducted with ltered

Vertical lters tend to be used for smaller ows, water. In seawater RO plants, RO concentrate is
whereas it is more economical to use a horizontal sometimes used for lter backwash to improve
arrangement for higher ow rate. The horizontal recovery. The backwash operation is sometimes
conguration can have ow distribution problems combined with air scouring to improve the ef-
during backwash, but this can be overcome by ciency of bed expansion and the removal of deposits.
compartmentalizing the vessel, with independent Air mixed with backwash water passes through the
backwashing of each part. ltration bed at rate of 5590 m/h (35 scf/ft2-min).
Air scouring increases bed expansion and is there-
Downflow Versus Upflow fore followed by a high rate water backwash.
Granular media lters can operate in upow or The typical backwash sequence takes
downow congurations, but in a pretreatment 310 min using an air and water mixture, followed
context, it is more common for them to be oper- by high rate backwash for 510 min. The total
ated with downow. Downow has benets in backwash volume is typically 23 % of the l-
dealing with algae, since the potential for algal trate. Backwash is triggered by a timer with a high
cell rupture is reduced. differential pressure override.
Number of Filtration Stages

For a good quality feed, a single-stage DMF Process Design
would probably be selected, but for a poor quality
feed of say >20 NTU, it is common practice to Figure 2 shows a typical process design for a granular
utilize two-stage ltration in which a high rate media ltration stage in a seawater reverse osmosis
rapid lter is followed by a slow rate ne bed. (SWRO) desalination duty (Voutchkov 2010).
Two-stage ltration tends to reduce coagulant use The pressure drop in a packed bed is related to
for a given removal. It is quite common to adjust the velocity of ow through the bed and the char-
the coagulant dose and/or pH between stages to acteristics of the media and the feed as shown by
optimize the overall performance. the Ergun equation (Richardson et al. 2002). The
rst term in this equation is dominant for the
laminar ow regime and the second becomes
Process Sequence dominant for turbulent ow.
Ergun equation:
A typical ltration cycle for a granular media
lter will be at least 8 h and more probably up to DP 150 Vs m:1 e2 1:75 V2s r1 e

24 h or even more. Treated water quality will be L D2p e3 Dp e3
less good during the ripening period which could
last up to 30 min, and some or all of the ow will be where
wasted or more likely recycled during this period.
After the ltration cycle, a backwash step is car- Dp is the pressure drop across the bed.
ried out. During the backwash step, water passes L is the length of the bed.
through the ltration bed in the reverse direction Dp is the equivalent spherical diameter of the
causing bed expansion and removal of deposits. packing.
Backwash operation is usually conducted at ow r is the density of uid.
rate of twice the ltration cycle ow, e.g., a typical m is the dynamic viscosity of the uid.
range of 1624 m/h for a DMF. Bed expansion during Vs is the supercial velocity (i.e., the velocity that
backwash is in the range of 2050 %. With increased the uid would have through the empty tube at
water temperature, the backwash ow rate has to be the same volumetric ow rate).
increased by about 3 % per c to maintain bed expan- e is the void fraction of the bed (typically 0.6 for
sion due to lower viscosity of the backwash uid. sand grains).
Granular Media Pretreatment Filtration, Fig. 2 Conventional RO pretreatment process
The equation applies for a clean bed. As solids a randomly packed bed formed after expansion
are captured from the feed, the characteristics of during backwash, the bed is prone to this effect.
the bed such as the voidage will change and the Channeling can also occur at the vessel walls, G
pressure drop will rise. Granular media lters and for this reason, the vessel diameter should
have relatively large channels, which means that ideally be 1500 times the mean particle diameter,
both lamina and turbulent terms in the equation a criterion which is readily achieved in large-scale
make important contributions. For media with commercial units.
much ner channels, such as membranes, the A key problem for granular media lters can be
ow regime is almost exclusively laminar. biofouling since the extensive surface of the bed can
The equation shows that physical characteristics support bacterial colonies which are provided with
for the bed such as the sphericity of the particles nutrients from the dissolved organic carbon content
and the voidage between them are important fac- (Wilf 2007). Indeed, this action can be intentional
tors in determining pressure drop. In the rst term and is the basis of bioltration. The behavior is
(characteristic for a laminar regime), pressure drop referred to as biofouling if growth increases to the
is inversely related to the channel diameter, point that pressure drop becomes excessive. The
whereas in the second term (which contributes for phenomenon can be controlled by dosing an oxi-
a turbulent regime), the pressure drop is inversely dizing agent such as chlorine, though this might just
related to the channel diameter squared, thus transfer the biofouling issue to downstream units
becomes much more important for coarser beds. due to the formation of readily assimilable organic
by-products. A strategy to control the problem is to
use shock chlorination in which a high chlorine
Operational Issues concentration of 5 ppm free residual or more is
dosed intermittently, for example, on a daily basis,
At the start of a ltration cycle, granular media for a short period of say 30 min.
lters normally operate with a low pressure drop.
For a shallow bed DMF operated at a moderate
rate, the pressure drop will be less than 0.1 bar. Filtered Water Quality
During a ltration cycle, the pressure drop will rise
to around 0.25 bar. It is important that the pressure Filtrate quality is determined through measure-
drop does not rise too much or channeling may ment of turbidity and SDI. There is no dened
occur in the bed leading to breakthrough. When and consistent correlation between turbidity and
this happens, the excessive pressure drop causes SDI. However, eld experience indicates that to
particle movement in the bed such that low resis- achieve an SDI below three, the ltrate turbidity
tance channels open up in low voidage zones of the has to be below 0.1 NTU, preferably below 0.05
bed resulting in by-passing. If the particles have NTU. By choosing an appropriate ltration sys-
good sphericity and are closely packed, it would tem design in terms of media type, bed depth, and
take a high pressure drop for this to happen, but for number of layers, a turbidity of 0.3 NTU can
894 Graphene Membranes
easily be achieved, and operators target 0.1 NTU formed by a single layer of sp2-bonded carbon
or less with a two-stage system. atoms arranged in honeycomb lattice with the
Dissolved organic carbon (DOC) removal is an atomic thickness of 0.4 nm and a light weight of
important requirement for granular media ltration 0.77 mg/m2.
since this material can cause fouling of downstream In graphene, the carbon atoms are bonded with
RO by direct adsorption and by supporting biofoul- three neighboring carbon atoms by planar s
ing. A high level of DOC may increase turbidity orbitals, which are separated by 120 C with
and SDI, but there is no simple direct correlation. each other, while pi (p) orbitals are delocalized
Chemical dosing holds the key to achieving above and below the graphene sheet (Fig. 2a).
low turbidity ltrate, normally using a clarica- Taking into account the electronic conguration
tion or otation stage. A single-stage system usu- of the carbon at the ground state (1s22s22p2), the
ally achieves a modest reduction in DOC of carbon-carbon s orbitals are covalent bonds
2030 %, whereas a two-stage system may deriving from hybridization of the single 2s
achieve 6070 %. Due to the depth removal mech- orbital with two 2p orbitals; these bonds confer
anism, there will be some degree of variability in robustness and elasticity to the graphene structure.
the treated water quality achieved as feed quality Indeed, a stiffness of 300400 N/m and a breaking
varies and other seasonal changes occur. strength of 42 N/m have been estimated for
defect-free graphene sheet (Lee et al. 2008).
Also, graphene sheets show a spring constant of
References 15 N/m for thickness of 28 nm, resulting in a
great ability to recover the initial size after strain.
Richardson JF, Harker JH, Backhurst JR (2002) Flow of uids
through granular beds and packed columns. In: Chemical These outstanding mechanical properties make
engineering, vol 2, 5th edn. Elsevie, Amsterdam, Ch 4 undoubtedly the graphene a desirable material
Voutchkov N (2010) Seawater pre-treatment. Publisher for reinforcing objects and devices as well as for
water treatment academy, TechnoBiz, Thailand,
the construction of graphene-based nano-
ISBN: 978-974-401-795-6
Wilf M (ed) (2007) The guidebook to membrane desalina- electromechanical systems (NEMS) (Bunch
tion technology. Balaban Desalination Publications, 2007).
L'Aquila. ISBN 0-86689-065-3 The pi (p) orbitals are instead formed from
overlapping of neighboring 2pz perpendicular to
the graphene plane (Fig. 2b); the bands p
Graphene Membranes (bonding) and p* (antibonding) represent the
valence band (VB) and the conduction band
Annarosa Gugliuzza1, Lidietta Giorno1 and (CB), respectively, through which the electrons
Enrico Drioli1,2,3,4 can be delocalized providing excellent electronic
1
Institute on Membrane Technology, National properties to the graphene. The mobility of pi (p)
Research Council of Italy, ITM-CNR, Rende, electrons is indeed estimated to be
Italy 15,000 cm2 V1 s1 with hypothetical limits of
2
Department of Environmental and Chemical 200,000 cm2 V1 s1 for free-impurity sheets.
Engineering, University of Calabria, Rende, Also, the resistivity of the graphene is assessed
Italy to be 1.0 108Om at room temperature with a
3
WCU Energy Engineering Department, capability of the material to transfer electrical
Hanyang University, Seoul, S. Korea current over 108 Acm2. The bands through
4
Center of Excellence in Desalination which the electrons move are represented as two
Technology, King Abdulaziz University, Jeddah, cones touching at the Dirac crossing energy point
Saudi Arabia (Fig. 2c). These two bands, however, degenerate
at the corner, the so-called Dirac point or K of
Among the allotropes of carbon (2D) (Fig. 1), the the hexagonal Brillouin zone, thus resulting in a
graphene is one of the most attractive materials semimetal with a zero-bandgap due to the
Graphene Membranes 895
Graphene
Fullerene Nanotube Graphite
Graphene Membranes, Fig. 1 Carbon allotropes where graphene sheet is the key component: graphene (2), fullerene
(0D), nanotubes (1D) and graphite (3D)
identical environment of two carbon atoms Currently, the graphene can be produced
located in the unit cell of the graphene lattice according to some techniques, including
(Fig. 2a) (Lau 2012). This could represent a lim- mechanical exfoliation, epitaxial growth, and
itation for the use of graphene in electronics, graphene oxide (Soldano 2010). As always, the
optoelectronics, photonics, bio-sensing, and choice of each single-manufacturing approach
solar energy harvesting. Nevertheless, the appli- implies advantages and disadvantages, respec-
cation of an electrical eld can induce changes in tively. The mechanical exfoliation takes the
the Fermi level, thereby causing doping or holes advantage to be handled and competitive from
events (Lu 2013). When the Fermi level is above the economics point of view, but the technique is
the Dirac point, there are holes in the VB. On the not reproducible and reliable enough to be scaled
contrary, a predominance of electrons is in the CB up. The growth of graphene on metallic sub-
when the Fermi level is over the Dirac point. In strates by chemical vapor deposition (CVD) of
both the cases, electrons and holes enable the hydrocarbons is another possible route for syn-
graphene to be transformed from semimetal to thesizing single layers of graphene on larger
semiconductor with much higher conducting scale, in spite of the high-processing tempera-
properties. This approach, along with the excep- tures. However, structural quality and preserved
tional capability of the charge carrier to travel electronic properties represent two important
sub-micrometer distances without scattering, challenges to reach yet. The chemical exfoliation
opens new horizons and promising perspectives is regarded as a promising scalable and handling
in the development of new approaches for mass approach for mass production of graphene
production of high-quality semiconducting sheets. It relies on the use of reactants in
graphene at low costs. the interspace between graphene layers, thus
b
a
pz orbital pz orbital
!120!
bond
bond
Plane of the sp2 orbitals bond
Fermi level
(n-type doping)
Dirac point
ky
Fermi level
kx (p-type doping)
Graphene Membranes, Fig. 2 (a) Honeycomb lattice of graphene with hybridized sp2-bonded carbon; (b) delocalized
electrons in pi (p) orbitals above and below the graphene sheet; (c) the Dirac point or K of the hexagonal Brillouin zone
promoting weak interactions such as van der water purication, and (bio)sensors (Xu 2015;
Waals forces and making the exfoliation easier Smith 2013; Gugliuzza 2014).
than micromechanical cleavage, wherein cohe- Graphene sheets have been recently proposed
sive forces between layers and long-distance as barrier membranes against water and gases for
interactions take place between graphene and applications in food and pharmaceutical packag-
substrate. The problem is a partial change in the ing. The related impermeable properties are
carbon hybridization from sp2 to sp3 due to the ascribable to the high density of delocalized elec-
oxidation of graphene. The latter is surely an trons, which prevent molecules from passing
interesting precursor of graphene derivate, but through the hexagonal rings. Recent density func-
its electronic properties are different from semi- tional theory studies have however demonstrated
metallic graphene produced according to the how the creation of functionalized sub-nanosized
other approaches. holes (around 3 ) in graphene sheets can
However, the use of graphene in membrane enhance the selectivity for the H2/CH4 gas pair
separation processes requires some expedients, (Tsetseris and Pantelides 2014). Specically,
including controlled nanoperforation, modulated values of selectivity of 108 for H2/CH4 may be
layers inter-distance, and chemical functiona- obtained through graphene sheets with nitrogen-
lization of the layers, in order to make this functionalized pores while selectivity of 1023 for
already excellent electronic material also an out- H2/CH4 by using graphene with all-hydrogen pas-
standing interface for selective gas separation, sivated pores (Fig. 3).
wrinkle
defects
functionality
Graphene Membranes, Fig. 3 Representative scheme of graphene with functionalized nanopores, defects, and
wrinklings responsible for selective behavior G
These pores represent a barrier (1.6 eV) for graphene with pores of the order of nanometers
CH4, whereas H2 (0.22 eV) can pass easily. There- are big enough to let water pass through, but
fore, the creation of nanoporous in graphene somewhat small to stop pollutants and salts at
together with the occurrence of local defects and the interface. Pore size, layer inter-distance, and
wrinkling and variation of the layer thickness, but functionality together with pore wettability are
also the insertion of functional groups in the pore again parameters to be considered for the optimi-
edges or between the stacked layers, have been zation of the productivity and efciency of water
simulated to make the graphene selective for gases purication. Concerning the selectivity of the
such as He, Ne, H2, CH4, and CO2 (Gadipelli and membrane process, rejection mechanisms based
Guo 2015). on size exclusion, chemical and electrostatic inter-
Electron beam, ultraviolet-induced oxidative, actions, and Donnans exclusion have been pro-
etching of holes on graphene surfaces together posed. Computational studies concerning, as an
with the use of molecular building blocks lithog- example, the separation of NaCl from water have
raphy, and tethering of graphene with carbon evidenced that small and hydrophobic pores
nanotubes have been proposed to produce enable at low pressure the rejection of NaCl
nanoporous graphene-based membranes through according to size exclusion mechanisms, while
the deposition of graphene sheets onto substrates, bigger ions in the absence of hydrogen bonding
including the polymeric ones. However, it should could increase the energy barrier to the ion pas-
be better to purify the gas mixtures before passing sage. Ionic exchange mechanisms have been also
the pores molded in graphene, since impurities, detected through graphene oxide membranes for
molecular residues, could obstruct severely the the separation of acid from FeCl3 solutions. The
diffusion pathways and affect the selectivity prop- transport of ion H+ has been estimated to be two
erties. Despite that various predictive studies indi- orders of magnitude greater than Fe3+ ions (Sun
cate the nanoporous graphene and derivate such as 2014).
graphene oxide as excellent materials for gas mix- Some hypotheses of mechanisms for water
ture separation, a lot of academic and industrial permeation through graphene and graphene
research is still necessary to develop on large scale oxide membranes have been also suggested,
products with suitable structural and chemical including capillary-driven ow between a water
fruitful features for scaled applications in gas monolayer and the pristine graphene layer and
separation. passage through porous microfractures formed in
Water purication and desalination represent the graphene sheets. The latter have been envis-
another challenge to reach by using graphene- aged as critical elements for enhancing water per-
based membranes as well (Xu 2015). Perforated meation, while hydrophilic groups attached to the
Graphene Membranes, Fig. 4 Scheme of an unstrained (a) and strain (a0 ) graphene layer over a cavity. Scheme of a
functional graphene sensor-based device (b)
graphene oxide have been regarded as responsible the design of well-dened structure can be
for slowing down the ux due to the establishment addressed at overcoming the performance of tra-
of hydrogen bonding between water and oxidized ditional membrane separations on larger scale.
regions of graphene. On this basis, well-sized Theoretical and experimental studies have also
interlayer gaps and broader nanochannels through conrmed the high potential of graphene for
graphene or graphene oxide are expected to well nanoelectromechanical system (NEMS) applica-
retain salt ions and undesired organic compounds, tions, where electrical conduction and outstanding
while the water permeation is expected to increase mechanical and optical properties (i.e., a transpar-
over three orders of magnitude with respect to tra- ency with a transmittance that changes from 95 %
ditional commercial reverse osmosis membranes. to 70 % for thickness ranging from 2 to 10 nm
Undoubtedly, membranes prepared from per- with an absorption of 2.3 % of white light) can be
forated and functionalized graphene represent an combined (Bunch 2007). As an example,
attractive solution to the incessant requirement of graphene sheets exhibit a pressure sensitivity of
highly performing gas and liquid separations. The two- to tenfolds superior to the standard silicon- or
outstanding mechanical properties together with carbon nanotube-based micro-electromechanical
the predicted ultrafast permeance and high- sensors (MEMSs). Despite that graphene has no
retention power make this class of membranes band gap, such the piezoelectrical properties
promising candidates for gas separation and cause changes in the electronic structure
water purication processes. However, a lot of understrain (up to 23 %) resulting in a shift of
explorative research is required to optimize mass the Fermi level (Smith 2013). When used as pres-
production of high-quality, durable, and energy- sure sensors, freely suspended graphene mem-
efcient graphene wherein the formation of branes can sense a gradient of pressure
nanopores and the insertion of functional moieties established between the external surrounding
between the stacked layers can be perfectly con- atmosphere and the unchanged internal environ-
trolled in number, shape, and size without yield- ment, which is regarded as the result of no perme-
ing mechanically vulnerable points. Also, the ability of graphene sheet to air. This gradient of
resistance of this kind of materials to swelling pressure yields a proportional force, thus causing
and fouling and ion clogging, which often take the expansion of graphene and, consequently, a
place during separation processes, needs to be variation of the resistance with enhancing of the
detected as well as the ability to work under high electrical charge transfer (Figs 4a-a'). This ability
pressure and harsh conditions. It is auspicial that of the graphene membranes makes them
computational and experimental studies can be extremely promising materials for controlling
reciprocally validated, thus casting new light on and monitoring actions in the eld of automobile,
predominant transport mechanisms. In this way, aerospace, textiles, and weather.
Despite those proof-of-concept demonstra- References

tions (Fig. 4b), graphene membranes have been
also proposed as chemical (e.g., small organic Bunch JS, van der Zande AM, Verbridge SS, Frank IW,
Tanenbaum DM, Parpia JM et al (2007) Electrome-
compounds and ions) and biological (e.g., sugars,
chanical resonators from graphene sheets. Science
proteins, DNAs) sensors enabling to reveal an 315:490493
increase in the current passage as an effect of Gadipelli S, Guo ZX (2015) Graphene-based materials:
doping generated by interactions between the synthesis and gas sorption, storage and separation.
Prog Mater Sci 69:160
functional graphene sheet and specic analyte
Gugliuzza A (2014) In: Gugliuzza A (ed) Smart mem-
(Liu 2012). As a chemical sensor, a recent membranes and sensors. Wiley/Scrivener Publishing, Salem
brane integrated with single-walled carbon Lau CN, Bao W, Velasco J Jr (2012) Properties of
nanotubes, which are regarded as rolled graphene suspended graphene membranes. Mater Today
15:238245
sheets, has been proposed for detection and regu-
Lee C, Wei X, Kysar JW, Hone J (2008) Measurement of
lation of surrounding humid atmosphere the elastic properties and intrinsic strength of mono-
(Pingitore 2015). Graphene membranes are layer graphene. Science 321:385388 G
expected to have enhanced performance. Liu Y, Dong X, Chen P (2012) Biological and chemical
sensors based on graphene materials. Chem Soc Rev
Similarly to gas and water separation, a deeper
41:22832307
understanding of the mechanisms which regulate Lu G, Yu K, Wen Z, Chen J (2013) Semiconducting
recognition and interactions and signal transmis- graphene: converting graphene from semimetal to
sion are necessary in order to promote technology semiconductor. Nanoscale 5:13531368
Pingitore V, Miriello D, Drioli E, Gugliuzza A (2015)
transfer and commercial application of devices
Integrated carboxylic carbon nanotube pathways with
realized according to reliable, reproducible, and membranes for voltage-activated humidity detection
scalable manufacturing processes. and microclimate regulation. Soft Matter
Overall, the use of graphene could lead to a 11:44614468
Smith AD, Vaziri S, Niklaus F, Fischer AC, Sterner M,
completely novel membrane with well-organized
Delin A, stling M, Lemme MC (2013) Pressure sen-
ultrathin nanocomposite architecture. They may sors based on suspended graphene membranes. Solid
promote very high selectivity thanks to precise State Electron 88:8994
pore morphology control as well as high perme- Soldano C, Mahmood A, Dujardin E (2010) Production,
properties and potential of graphene. Carbon
ability thanks to a very thin layer. Furthermore,
48:21272150
energy use can be minimized, thanks to facilitated Sun P, Wang K, Wei J, Zhong M, Wu D, Zhu H (2014)
transport promoted through the membrane by Effective recovery of acids from iron-based electrolytes
favorable surface interactions. using graphene oxide membrane lters. J Mater Chem
A 2:77347737
Production of ultrathin nanocomposite mem-
Tsetseris L, Pantelides ST (2014) Graphene: an imperme-
branes is challenging to produce, and therefore, it able or selectively permeable membrane for atomic
is difcult to predict the real commercial promise. species? Carbon 67:5863
However, the importance of energy reduction in Xu Q, Xu H, Chen J, Lv Y, Dong C, Sreeprasad TS
(2015) Graphene and graphene oxide: advanced mem-
massive production, such as seawater desalina-
branes for gas separation and water purication. Inorg
tion, provides a crucial driver, which will foster Chem Front 2:417424
the development of these novel membranes.
H
H2 Permeation Through Pd-Based used for hydrogen supply or withdrawal from

Membranes the reaction zone, thus providing improved yield
and high catalyst effectiveness (Pagliery and
V. V. Volkov Way 2002).
A.V. Topchiev Institute of Petrochemical Hydrogen permeation through solid palladium
Synthesis Russian Academy of Sciences proceeds via the solutiondiffusion mechanism,
(TIPS RAS), Moscow, Russian Federation and this mechanism involves the following steps:
(a) diffusion of molecular hydrogen to the surface
of the palladium-based membrane, (b) a reversible
Synonyms dissociative adsorption on the palladium surface,
(c) dissolution of atomic hydrogen in the bulk
Hydrogen permeation metal, (d) diffusion of atomic hydrogen through
the bulk metal, (e) association of hydrogen atom
Palladium is known to be one of the most effective on the palladium surface, (f) desorption of molec-
metals for hydrogen adsorption, dissociation, and ular hydrogen from the surface, and (g) diffusion
recombination even though the permeability of of molecular hydrogen away from the surface
this metal is by an order of magnitude lower (Yun and Oyama 2011). On the palladium sur-
than that of refractory metals such as tantalum, faces, the dissociative adsorption of hydrogen
vanadium, and niobium. molecules proceeds with a low or zero activation
In 1863, Sainte-Claire Deville and Troost dis- energy barrier. This is the rst step in bulk absorp-
covered that hydrogen is able to permeate through tion toward the formation of metal hydrides.
palladium; later, in 1866, Graham found that Generally, hydrogen permeation is described
palladium can absorb huge amounts of hydrogen by the following equation:
(several times with respect to its initial volume) at
room temperature. The palladium-based J Ppn h pn l =L;
membrane technology for hydrogen production
has come into life in the late 1950s. In the where J is the hydrogen ux, P is the permeability
1960s, pioneering studies by Gryaznov led to the coefcient, L is the thickness of the palladium
development of a breakthrough membrane layer, ph and pl are the partial pressures of hydro-
reactor when palladium-based membranes were gen on the high-pressure (feed) side and the

DOI 10.1007/978-3-662-44324-8
902 Hairy Superhydrophobic Surface
low-pressure (permeate) side, respectively, and

n is the pressure exponent. This exponent gener- Hairy Superhydrophobic Surface
ally ranges from 0.5 to 1 depending on the rate-
controlling step [steps (ag)]. According to Wolfgang M. Sigmund and Shu-Hau Hsu
Sieverts law, when the rate-controlling step is Department of Materials Science and
the bulk diffusion through the palladium layer Engineering, University of Florida, Gainesville,
[step (c)], the n value is equal to 0.5 because the FL, USA
diffusion rate is proportional to the concentration Department of Energy Engineering, Hanyang
of hydrogen atoms on the opposite sides of the University, Seoul, Republic of Korea
metal surface and this hydrogen concentration is
proportional to the square root of the hydrogen
pressure. When mass transport to or from the Superhydrophobicity not only exists on lotus-like
surface (a, g) or dissociative adsorption (b) or leaves but also on surfaces comprising high-
associative desorption (e) becomes rate- aspect ratio soft hairs. Although the piliferous
controlling steps, the expected value of n is exterior is found in almost all water-repellent
1 because these processes depend linearly on the arthropods and has been noted for over
concentration of molecular hydrogen. In the case 100 years, (Comstock 1887) compared to the
of the Pd-based composite membranes with thick Lotus effect, studies based on this structure are
(>5 mm) Pd layers, the n values are higher than relatively limited. The superhydrophobicity is
0.5, and this fact can be explained by defects or dominated by both surface roughness and chem-
pinholes through which substantial amounts of istry, so as the hairy superhydrophobic surfaces.
hydrogen can permeate. This process proceeds Surface roughness arises from the arrangement of
via the Knudsen or Poiseuille ow mechanisms, hairs, and, similar to the water-repellent leaves,
and the corresponding exponents are higher hairs are also covered with a layer of epicuticular
than 0.5. wax that renders the exterior hydrophobic and
In the case of binary palladium alloys, hydro- prevents the penetration of water. The wax layer
gen permeation is generally proportional to the is not only important in water repellency but also
average bond distance of the alloys. Permeability in water retention within the insects structure;
is the product of diffusivity and solubility. In the removal of the waxy layer will result in rapid
case of PdAg alloys, hydrogen solubility desiccation of the insects (Wigglesworth 1945).
increases with increasing Ag content up to The hairs in almost all water-repellent arthro-
2040 wt.%; hydrogen diffusivity decreases with pods are generally arranged as a series of G
increasing Ag content. Concomitant changes in shapes. Most of the hairs are not only inclined
solubility and diffusivity lead to higher hydrogen but curled with their tips parallel to the bodies,
permeability (1.7 times higher) as compared which in turn provides a greater resistance to
with that of pure Pd at 23 wt.% of Ag and at wetting. The array of tilted pillars increases the
623K (Ma et al. 2003). pressure required to penetrate into the grooves, as
the free space between the tilted pillars is reduced,
making them favorable to resisting water.
A horizontal array of cylindrical pillars also pro-
References
vides a greater water resistance (Extrand 2002).
Ma YH, Mardilovich IP, Engwal EE (2003) Thin compos- Overall, a series of inclined and parallel pillar
ite palladium and palladium/alloy membranes for arrays is energetically unfavorable for a liquid to
hydrogen separation. Ann N Y Acad Sci 984:346360 intrude between the pillars parallel to the
Paglieri SN, Way JD (2002) Innovations in palladium interface.
membrane research. Sep Purif Rev 31:1169
Yun S, Oyama ST (2011) Correlations in palladium mem- Figure 1 shows an example of typical hairy
branes for hydrogen separation: a review. J Membr Sci superhydrophobic surface, water striders body
375:2845 and its leg (Hsu 2010). Two different sets of
Hairy Superhydrophobic Surface 903
Hairy Superhydrophobic Surface, Fig. 1 SEM images water repellency. (a) The hair on striders abdomen is tilted
of the hair layer of the water striders body. The hair tips are at around 3045 relative to the body surface. (b) The hair
bent inward toward the body to enhance the degree of on striders leg (Hsu 2010) H
Hairy Superhydrophobic
Surface, Fig. 2 A water
droplet forms a spherical
bead on Ladys Mantle leaf
body hair can be usually found from the arthro- reaching over 160 , and it provides crucial living
pods. Longer hairs, termed macrotrichia, are functions for the species, such as underwater res-
less dense with lengths of 20 mm and diameters piration, walking on water, or surviving under rain
of around1mm, and shorter hairs, termed impact (Bush et al. 2008). More details on under-
microtrichia, are approximately 2 mm long and water respiration can be found in the section of
100 nm wide. Longer hairs taper to a point from plastron effect.
the base and incline at an angle of 3045 with Hairy leaves with a feature similar to lotus
their tips parallel to the body surface. Shorter hairs effect are also common. These plants, i.e.,
are however highly dense and are more randomly Ladys Mantle (Alchemilla vulgaris), exhibit a
oriented. The hair is also patterned with plurality of exible hairs, termed trichomes, on
nanoscaled grooves with widths of around their leaves and stems (Fig. 2). Water droplets on
400 nm, thus providing a structure with hierarchi- the fuzzy leaves can be suspended by the tri-
cal roughness. Many arthropods possessing hairy chomes, which effectively prevent the droplets
integument are reported to have contact angles from wetting the leaf. Although the leaf itself
904 Halar
shows an excellent hydrophobicity, those tri- Meltblown lters made by Halar ECTFE are
chomes are found to be hydrophilic (Otten and commercially available for pharmaceutical appli-
Herminghaus 2004). Water droplets are pinned by cations, solvent ltration, and semiconductor
the hydrophilic trichomes causing high contact applications where high ozone resistance is
angle but roll-off angle. needed. Other possible applications of Halar
ECTFE membranes are membrane contactors,
including membrane distillation, pervaporation,
Cross-References and industrial water treatment.
The following features make Halar ECTFE a
Lotus Effect suitable material for manufacturing membranes
Plastron Effect for specialty applications:
Youngs Model
Stability at pH levels ranging from 1 to 14
Outstanding resistance to organic solvents,
References ozone, and chlorine
Very good tensile properties
Bush JWM, Hu DL, Prakash M (2008) The integument of High temperature stability continuous use up
water-walking arthropods: form and function. Adv
to 150 C
Insect Physiol 34:117192
Comstock JH (1887) Note on the respiration of aquatic Melt processability, including TIPS (thermally
bugs. Am Nat 21:577578 induced phase separation) technology for
Extrand CW (2002) Model for contact angles and hyster- membrane manufacturing
esis on rough and ultraphobic surfaces. Langmuir
Low level of extractible
18:79917999
Hsu SH (2010) Biologically inspired hairy surfaces for FDA compliance for selected grades
liquid repellency. Doctoral dissertation, University of
Florida The general physical properties of the Halar
Otten A, Herminghaus S (2004) How plants keep dry: a
ECTFE grades available for membrane applica-
physicists point of view. Langmuir 20:24052408
Wigglesworth VB (1945) Transpiration through the cuticle tions are shown in Table 1.
of insects. J Exp Biol 21:97114 Because of its exceptional chemical resistance,
Halar ECTFE cannot be processed via solution
phase inversion. It must be processed at tempera-
tures close to its melting point (200240 C) using
a TIPS process for manufacturing hollow ber and
Halar at sheet membranes. Typical diluents used to pro-
cess this material are high boiling solvents such as
Anna Maria Bertasa acetyl tributyl citrate, glycerol triacetate (GTA),
Solvay Specialty Polymers, Bollate (MI), Italy dibutyl phthalate (DBP), and dioctyl phthalate
(DOP) (Simone et al. 2012). The polymer should
be melt blended with the selected solvent, casted or
Halar ECTFE (ethylene chlorotriuoroethylene) extruded into the desired form, and cooled at a
is a semicrystalline uoropolymer based on eth- temperature that promotes phase separation. As
ylene and chlorotriuoroethylene. It is currently nal step the diluent should be extracted from the
utilized in several demanding market segments membrane, typically with a low boiling solvent that
where high purity, outstanding resistance to is mixable with the diluent.
harsh environments, low permeation, and high Phase diagrams of Halar ECTFE are avail-
temperature stability are needed. Examples of able in the literature (Ramaswamy et al. 2002),
applications are chemical processing, semicon- together with examples of processing for
ductors, wire and cable, and the pharmaceutical manufacturing membranes and membrane prop-
industry. erties (Roh et al. 2010). For example, when using
Halophiles 905
Halar, Table 1 General physical properties of Halar ECTFE commercial grades

Melting point Melt ow index Tensile modulus
Halar grade Tm ( C) g/10 min at 275 C MPa Contact angle
Halar 901 242 1 @ 2.16 kg 1,6001,800 9095
Halar 902 222 1 @ 5 kg 1,6001,800 9095
DBP as solvent, it is possible to obtain a liquid-

liquid demixing for polymer concentrations Halophiles
below 25 % and solid-liquid demixing with poly-
mer crystallization for polymer content above Chiara Schiraldi and Mario De Rosa
25 %. It is possible as well to extend the liquid- Department of Experimental Medicine, Section of
liquid demixing region of the phase diagram by Biotechnology, Medical Histology and Molecular
changing the polarity of solvent, for instance by Biology, Second University of Naples, Napoli,
adding a non-solvent in the polymer mixture. Italy
It is possible to manufacture either at sheet H
either hollow ber membranes with Halar ECTF-
E. The equipment must be properly designed tak- Halophiles (i.e., salt loving) are organisms that
ing into consideration the needed processing live in high salt concentrations. In taxonomy most
temperature; its corrosion resistance in contact of them belong to the Archaea kingdom, but also
with uoropolymers must be veried before use. Halobacteria and some Eukarya were isolated.
Porosity of membrane can vary from MF to UF Often they are colored orange, red, or pink/purple;
according to processing parameters, such as poly- colonies are evidencing the presence of caroten-
mer concentration, type of solvent, processing oids and bacteriorhodopsin. The latter is used to
temperature, air gap, and cooling rate; it is possi- absorb light, which provides energy to create ATP.
ble as well to obtain symmetrical or asymmetrical Halobacteria also possess a second pigment,
structures and to ne tune the thickness of the skin halorhodopsin, which pumps chloride ions in the
layer (M uller 2006). cell in response to photons, creating a voltage
Halar ECTFE is a hydrophobic material due gradient and assisting in the production of energy
to its chemical composition. Surface properties from light. A number of species have been iso-
can be modied with posttreatments that are typ- lated in oceans but also in environment with salt
ically performed at industrial level for obtaining concentration much higher than marine ones like
desired performances of the nished membranes. in salty lakes (e.g., Dead Sea) or in evaporation
ponds. Halophiles, both aerobes and anaerobes,
may use a variety of nutrients and energy sources.
Anaerobic ones include phototrophic, fermenta-
References tive, sulfate-reducing, homoacetogenic, and
methanogenic species. Halophiles require NaCl
Muller H (2006) A new solvent resistant membrane based
on ECTFE. Desalination 199:191192
for growth, while halotolerant organisms can
Ramaswamy S, Greenberg A, Krantz W (2002) Fabrica- stand saline conditions up to 0.8 M NaCl, moder-
tion of poly (ECTFE) membranes via thermally ate halophiles prefer salt concentration in the
induced phase separation. J Membr Sci 210:175180 range 0.83.4 M, and nally extreme halophiles
Roh I, Ramaswamy S, Krantz W, Greenberg A (2010) Poly
(ethylene chlorotriuoroethylene) membrane forma-
grow at higher concentrations up to 5.1 M NaCl
tion via thermally induced phase separation (TIPS). (Oren 2002a). These microorganisms have
J Membr Sci 362:211220 evolved physiological and genetic characteristics
Simone S, Figoli A, Santoro S, Galiano F, Alfadul S, to cope with the increase in osmotic pressure in
Al-Harbi O, Drioli E (2012) Preparation and character-
ization of ECTFE solvent resistant membranes and
different ways. The extremely halophilic Archaea
their application in pervaporation of toluene/water mix- accumulate ions such as K+; other bacteria devel-
tures. Sep Purif Technol 90:147161 oped biosynthetic pathways to synthesized
906 Heat and Mass Transfer in Membrane Distillation
compatible solutes (e.g., glycine, betaine, sugars, nanodevices, including high-capacity computer
polyols, amino acids, and ectoines) in high con- storage; some subcellular components, such as
centration so as to make the cytoplasm isotonic polyhydroxyalkanoate granules, gas vesicles,
with respect to the surrounding medium; besides, and phospholipid vesicles, and bioactive com-
they often present efcient ion pumps, pounds like halocins are being developed for med-
UV-absorbing pigments, and acidic proteins, ical applications. Last but not least, important
which help them to resist the osmotic stress and metabolic activities of halophiles have also been
the denaturing effects of salts, as well as intense tapped for a variety of environmental applications
solar radiation. Many of these characteristics con- and specically for bioremediation of pollutants
stitute valuable resources for biotechnology and in saline wastewaters.
nanotechnology (Oren 2002b). Salty food
bioprocessing and fermentation (such as soy
sauce, Chinese fermented beans, salted cod, salted References
anchovies, sauerkraut, etc.) often involve
Halobacteria, as either essential ingredients or Lang YJ, Bai L, Ren YN, Zhang LH, Nagata S (2011)
Production of ectoine through a combined process that
accidental contaminants. Besides these food
uses both growing and resting cells of Halomonas
applications, compatible solutes, which are small salina DSM 5928T. Extremophiles 15(2):303310
biomolecules that also help to protect cells against Oren A (2002a) Molecular ecology of extremely halophilic
stresses like high temperature, desiccation, and Archaea and Bacteria. FEMS Microbiol Ecol 39:17
Oren A (2002b) Diversity of halophilic microorganisms:
freezing, have now been used for 15 years in
environments, phylogeny, physiology and applications.
cosmeceutical products. These peculiarities J Ind Microbiol Biotechnol 28:5663
improved performances of moisturizers and Schiraldi C, Maresca C, Catapano A, Galinski EA, De
other products, thus prompting the research on Rosa M (2006) High-yield cultivation of Marinococcus
M52 for production and recovery of hydroxyectoine.
bioprocess improvement. Recently, using
Res Microbiol 157(7):693699
ectoine-excreting strain Halomonas salina DSM
5928(T), a new process was developed by Lang
and co-authors (2011) for high-efciency produc-
tion of ectoine, which involved a combined Heat and Mass Transfer in Membrane
process of batch fermentation to grow cells Distillation
followed by a resting cell production phase.
Hydroxyectoines were efciently produced using Chunrui Wu
a microltration membrane bioreactor (Schiraldi State Key Laboratory of Hollow Fiber Membrane
et al. 2006). More recently also the production of Materials and Membrane Processes, Institute of
mannosylglycerate (MG) was established through Biological and Chemical Engineering, Tianjin
the milking process. In particular the trehalose- Polytechnic University, Tianjin, China
decient mutant of the strain Thermus
thermophilus RQ-1 was cultivated using upshock
fermentation and osmotic downshock, for the Membrane distillation (MD) is a membrane
effective production/excretion of MG. More than seperation process based on the principle of evap-
90 % of the intracellularly accumulated solute was oration which uses hydrophobic membrane as
recovered from the water fraction. Although cur- separating medium. Unlike the membrane process
rently established applications of halophiles are of ultraltration (UF) and reverse osmosis
limited primarily to compatible solutes in the cos- (RO) which only contain mass transfer, MD also
metic industry and carotene and hydrolases in the contains heat transfer and the heat transfer is a key
nutritional and food industries, many other novel factor affecting mass transfer efciency. So the
applications are under development. They include investigation of heat and mass transfer in MD
specialized applications for lms of bacteriorho- process is the key to explore the mechanism
dopsin from halophiles for a variety of optical of MD process. Many scientists have made
Heat and Mass Transfer in Membrane Distillation 907
Heat and Mass Transfer Conduction

in Membrane
Distillation, Fig. 1 Heat
transfer resistances in
membrane distillation Tb,f TI,f TI,p Tb,p
Thermal boundary Thermal boundary

layer layer
Latent heat of vaporisation
Heat and Mass Transfer Membrane

in Membrane
Distillation, Fig. 2 Direct Tf Permeate
contact membrane
distillation
T1
T2
Feed H
Tp
Mass flux (J)

Heat flux (q)
signicant progress in the study of heat and mass and permeate, respectively; DHv is the heat of
transfer in MD process. vaporization and DTm is the temperature drop
across the membrane.
For MD process, heat and mass transfer occur
Heat Transfer simultaneously and result in complex heat transfer
mechanisms. So the mass transfer can affect both
The heat transfer of MD consists of three steps, as heat transfer rates and heat transfer coefcients.
shown in Fig. 1. First, heat transfer by the way of The heat transfer model has been proposed to
convection in the feed boundary layer. Second, describe the heat transfer mechanisms in MD
heat transfer by the way of conduction through and facilitates the evaluation of the membrane
the membrane and the heat ows together with the surface temperatures. In most literatures, the heat
vapor through the membrane pores, i.e., the latent transfer model for MD was developed based on
heat of vaporization. Third, heat transfer by the the assumption of linear temperature prole and
way of convection in the permeate boundary isenthalpy ow of vapor (Figs. 2, 3, and 4).
layer. All types of MD contain the rst two The effective energy transfering from the feed
steps, while for the third step, other types of MD bulk to the membrane surface is:
do not contain it except DCMD.

At steady state, the overall heat transfer coef- Qf hf T f T fm
cient (H) of the MD process can be expressed by a
resistance in series model (M. Khayet et al. 2000): The energy transfering across the membrane is:
1 1

1

1 hm
H hf hm JDH v =DT m hp Qm JDH T fm T pm
d
where hf, hm, and hp are the individual heat trans- For the VMD process, the contribution
fer coefcients of liquid feed phase, membrane, of heat conduction in the membrane matrix
908 Heat and Mass Transfer in Membrane Distillation
Heat and Mass Transfer

in Membrane
Distillation, Fig. 3 Air
gap membrane distillation
Heat and Mass Transfer

in Membrane Liquidvapour
Distillation, interface Aqueous
Fig. 4 Vacuum membrane Retentate Vacuum &
solution
cooling
distillation
Vapour
Suction and
condenstation Porous membrane
Permeate
Aqueous solution (warm)
can be neglected due to the high vacuum degree in km

hf T f T f , m JDH T fm T pm
the vacuum side (Li and Sirkar 2005). d
At steady state, the overall
hp T p , m T p
heat transfer ux through the membrane is
given by: Q H Tf Tp
It is worth pointing out that the heat con-

Q Qf Qm Qp duction can be neglected for nonsported thin
Heat and Mass Transfer in Membrane Distillation 909
Heat and Mass Transfer Viscous

in Membrane
Distillation, Fig. 5 Mass
transfer resistances in
membrane distillation
xb,f xI,f xI,p xb,p
Concentration boundary Cocentration boundary

layer layer
Knudsen Molecular
membrane and high operating temperature as well In AGMD process, the molecular diffusion
(Phattaranawik and Jiraratananon 2001). More- theory is used to describe the transfer of vapor
over, the heat transfer by convection is ignored molecules through the membrane and the air gap.
in the MD process, except in AGMD (Curcio and A stagnant gas lm (air) is assumed to lie inside
Drioli 2005). the membrane at the air gap side. H
For VMD process, in order to remove air
trapped in the membrane pores, the deaeration of
Mass Transfer the feed solution or a continuous vacuum in the
permeate side should be applied. Consequently,
The mass ux (J) in MD process is assumed to be the ordinary molecular diffusion resistance is
proportional to the vapor pressure difference neglected. The Knudsen mechanism is used to
across the membrane and is given by: express the mass transfer (Bandini and Sarti
1999), Poisseille ow, or both together (Banat

J C m Pf P p and Simandl 1994).
where Cm is the membrane coefcient, Pf and Pp

are the vapor pressure at the membrane feed and
permeate surfaces, which can be calculated from References
the Antoine equation (Fig. 5). For different types
of MD processes, the obtainment of Cm is based Banat FA, Simandl J (1994) Theoretical and experimental
on different models: study in membrane distillation. Desalination
95(1):3952
For DCMD process, mass transfer through the Bandini S, Sarti GC (1999) Heat and mass transport resis-
membrane can be divided into three models. tances in vacuum membrane distillation per drop.
These models relate the mass transport with colli- AIChE J 45:14221433
sions between molecules, and/or molecules with Curcio E, Drioli E (2005) Membrane distillation and
related operations a review. Sep Purif Rev
membrane. Knudsen diffusion takes place 34(1):3586
when the pore size is too small, such that the Ding Z, Ma R, Fane AG (2003) A new model for mass
collision between the molecules and the inside transfer in direct contact membrane distillation. Desa-
walls of the membrane suitably expresses the lination 151(3):217227
Khayet M, Godino P, Mengual JI (2000) Theory and
mass transport and the collision between mole- experiments on sweeping gas membrane distillation.
cules can be ignored (Zhongwei et al. 2003). J Membr Sci 165:261272
Molecular diffusion occurs when the molecules Li B-A, Sirkar KK (2005) Novel membrane and device for
move corresponding to each other under the inu- vacuum membrane distillation-based desalination pro-
cess. J Membr Sci 257:6075
ence of concentration gradients. In Poiseuille ow Phattaranawik J, Jiraratananon R (2001) Direct contact
(viscous ow), the gas molecules act as a contin- membrane distillation: effect of mass transfer on heat
uous uid driven by a pressure gradient. transfer. J Membr Sci 188(1):137143
910 Heavy Metal Recovery by Chelating Agents and Membranes
Fe. Further, they show a low biodegradability,

Heavy Metal Recovery by Chelating and biodegradable chelating agents such as
Agents and Membranes EDDS ([S,S]-ethylenediaminedisuccinic acid),
IDSA (iminodisuccinic acid), MGDA (methyl-
Jose Sabate glycinediacetic acid), and citric acid have been
Department dEnginyeria Agroalimentaria i receiving increasing attention in recent years.
Biotecnologia, Universitat Politecnica de Because of their molecular size, chelates can be
Catalunya, Castelldefels, Spain concentrated by nanoltration, but previously, col-
loids that can clog the NF membranes have to be
removed by microltration or ultraltration.
Heavy metal concentration from wastewaters can Recycling the chelating agent improves the eco-
be carried out by several membrane operations. nomical competitiveness of the process and dimin-
Electrodialysis (ED), reversed osmosis (RO), and ishes the environmental impact, especially for
nanoltration (NF) membranes retain polyvalent nonbiodegradable chelating agents. That involves
cations. Alternatively, heavy metals can be con- separating metal and ligand in the contaminated
centrated by polymer-enhanced ultraltration washing solution. Ion exchange resins and selec-
(PEUF) as well. tive precipitation of metal or chelating agents by
Further, membrane separations, combined with addition of chemical agents have been proposed.
other process, can contribute to heavy metal Experiments at laboratory scale have showed that
recovery from groundwater, contaminated soils, some NF membranes are able to reject Cu2+ while
sediments, sludge, etc (Dermont et al. 2008; citric acid leaks. That can be achieved in acidic
Hashim et al. 2011; Tsang et al. 2012). First of conditions when the chelate is broken and NF
all, heavy metals must be extracted from the solid membrane is positively charged, which favor the
matrix using a washing solution containing a suit- rejection of Cu2+ coions and the leakage of neutral
able extractant. The huge amount of metal- or slightly negatively charged citric acid.
containing solution produced can be concentrated Some water-soluble polymers contain ligand
by membrane technologies with the concomitant groups that can form chelates with heavy metals
production of a clean permeate solution, which as well. They can be retained by ultraltration
ideally should be reused. membranes. At low pH, metal ions are release
Metal extraction efciency depends on several from the polymer and a new ultraltration opera-
factors such as extractant/metal afnity, tion allows one to separate them. The polymer can
extractant/metal ratio, leaching method, liquid/ be regenerated by alkali addition and reused. This
solid ratio, pH, and extraction time. Heavy metal process is named polymer-enhanced ultraltration
extraction has mainly been done by solutions of (PEUF). Recently, dendrimers, highly branched
strong acids, chelating agents, dendrimers, surfac- macromolecular compounds with suitable termi-
tants, or reducing or oxidizing agents (Dermont nal functional groups, have been tested for heavy
et al. 2008). All of them but acid extraction takes metal recovery from contaminated soil.
place under mild conditions preventing structural
changes in the matrix to be decontaminated.
The most used chelating agents are EDTA
(ethylenediaminetetracetic), NTA (nitrilotriacetic Cross-References
acid), and DTPA (diethylenetriaminepentaacetic
acid). They form very stables chelates and they Heavy Metal Recovery by Membrane
are very efcient in removing heavy metal ions Operations
bound to mineral and organic fraction of soil Heavy Metal Recovery by Polymer-Enhanced
(Hashim et al. 2011). One of their drawbacks is UF
their low selectivity, which involves high reagent Nanoltration
consumption due to the complexation of Ca and Ultraltration (UF)
Heavy Metal Recovery by Membrane Operations 911
References can be achieved with high efciency using various

membrane technologies: ultraltration (UF),
Dermont G, Bergeron M, Mercier G, Richer-Lache nanoltration (NF), reverse osmosis (RO), and
M (2008) Soil washing for metal removal: a review of
electrodialysis (ED) (Coman et al. 2013; Fu and
physical/chemical technologies and eld applications.
J Hazard Mater 152:131 Wang 2011; Kurniawan et al. 2006).
Hashim MA, Mukhopadhyay S, Sahu JN, Bhaskar UF membranes reject macromolecules, but
Sengupta B (2011) Remediation technologies for their pores are not small enough to retain heavy
heavy metal contaminated groundwater. J Environ
metal ions such as Cd2+, Cu2+, Hg2+, Ni2+, Ni3+,
Manage 92:23552388
Tsang DCW, Lo IMC, Surampalli RY (eds) (2012) Chelat- etc. Nevertheless, they can be concentrated with
ing agents for land decontamination technologies. the addition of a soluble polymer or surfactant that
American Society of Civil Engineers, Reston binds the metal ions. Those processes are named
polymer-enhanced ultraltration (PEUF) and
surfactant-enhanced ultraltration (SEUF)
(Coman et al. 2013; Fu and Wang 2011;
Heavy Metal Recovery by Membrane Kurniawan et al. 2006). At low pH, metal ions
Operations are released from the polymer and a new ultral- H
tration operation allows one to separate them. The
Jose Sabate polymer can be regenerated by alkali addition and
Department dEnginyeria Agroalimentaria i reused.
Biotecnologia, Universitat Politecnica de RO is used to remove almost everything from
Catalunya, Castelldefels, Spain the water which includes small molecules and
ions. It can produce drinking water from brackish
water or seawater. Therefore, it has been shown to
Several industries such as metal plating facilities, concentrate heavy metal ions such Cd2+, Cu2+,
mining operations, fertilizer industries, tanneries, and Ni2+. However, compared to other membrane
batteries, paper industries, pesticides, etc., pro- technologies, RO requires the highest pressure
duce heavy metal (Cd, Cr, Cu, Hg, Ni, Pb, Zn) and energy cost.
wastewaters. Heavy metals are not biodegradable Nanoltration (NF) is the intermediate process
and persist in the environment: waters, soils, sed- between UF and RO (Soldenhoff et al. 2005). NF
iments, etc. Further, their recovery could be attrac- membranes usually have an electrical charge, pos-
tive from the economical perspective, and several itive for pH values lower than the membrane iso-
technologies have been studied and proposed to electric point (IEP) and negative for higher pH
carry it out: chemical precipitation, ion exchange, values. NF yields high rejections of heavy metal
adsorption, membrane ltration, coagulation- ions, although NF membrane pores are larger than
occulation, otation, and electrochemical heavy metal ions in solution. That is due to the
methods (Coman et al. 2013; Fu and Wang contribution to the separation process of several
2011; Kurniawan et al. 2006). Membrane separa- mechanisms: steric (sieving), electric (Donan),
tion technologies yield a high efciency and allow and dielectric (born and image forces). Dielectric
continuous operation. However, to prevent mem- and Donan effects strongly favor the rejection of
brane clog, a pretreatment of the polluted efuent polyvalent ions. That makes NF very suitable for
to achieve a clean feed is required. Further, they heavy metal concentration, which can be done in
need careful operation and periodic washing to one step and at lower pressure and energy cost
prevent membrane damages and fouling. than RO.
Membrane separations can reduce the volume Electrodialysis (ED) is another membrane pro-
of contaminated water, making the next processes cess for the separation of ions that uses an electric
necessary to full recovery easier and cheaper. At eld as the driving force. It is an alternative to RO
the same time, a clean permeate that meets the for the production of drinking water. It has been
applicable efuent guidelines is obtained. That demonstrated to be able to concentrate hexavalent
912 Heavy Metal Recovery by Polymer-Enhanced UF
and trivalent chromium and Cd2+, Cu2+, Co2+, Ni2+, liquid streams. It is also called polymer-assisted
and Fe3+. ultraltration (PAF), complexation-ultraltration
Membrane technologies can contribute to process (CUFP), or polymer-supported ultraltra-
the heavy metal recovery from contaminated tion (PSU) (Rivas et al. 2009).
underground water, soils, sediments, sludge, Ultraltration membranes reject macromole-
etc. However, a previous extraction of metals is cules, but their pores are not small enough to
needed, which can be achieved with acids, chelat- retain heavy metal ions such as Cd2+, Cu2+, Ni2+,
ing agents, dendrimers, or surfactants. Ni3+, Hg2+, etc. Nevertheless, if a soluble polymer
able of binding polyvalent cations is added to a
solution containing heavy metal ions (F), the
Cross-References resulting metal-polymer complex will be retained
by UF membranes. Then, a clean permeate
Electrodialysis (P) can be obtained and the heavy metals will
Heavy Metal Recovery by Chelating Agents be concentrated in the retentate stream (R).
and Membranes See gure 1.
Heavy Metal Recovery by Polymer-Enhanced The polymers contain functional groups such
UF as amine, carboxylated, phosphonic, or sulfonate
Nanoltration which link to metals either through ionic or che-
Ultraltration (UF) lating bonds. The most used synthetic polymers
are polyacrylic acid (PAA) and polyethyleneimine
(PEI) (Rivas et al. 2009). The use of natural poly-
References
mers (chitosan, humic acids, etc.) and polymers
Coman V, Robotin B, Ilea P (2013) Nickel recovery/ anchored on the membrane represents growing
removal from industrial wastes: a review. Resour alternatives.
Conserv Recycl 73:229238 A suitable polymer should have a high afnity
Fu F, Wang Q (2011) Removal of heavy metal ions from for target ions and should be able to be recycled in
wastewaters: a review. J Environ Manage 92:407418
Kurniawan TA, Chan GYS, Lo WH, Babel S (2006) Phys- order to make the process economically feasible. It
icochemical treatment techniques for wastewater should be biodegradable to prevent new environ-
laden with heavy metals. Chem Eng J 118:8398 mental problems, but it cannot have a very short
Soldenhoff K, McCulloch J, Manis A, Macintosh P (2005) half-life in working conditions. As commercial or
Nanoltration in metal and acid recovery. In: Fane AG,
Schafer AI, Waite TD (eds) Nanoltration: principles natural polymers have a molecular weight distribu-
and applications. Elsevier, Amsterdam, pp 459477 tion, they must previously undergo a dialtration
process to remove short chains, which would bind
heavy metals but would leak through the UF mem-
brane. Further, its interactions with membrane
materials should be considered to minimize fouling.
Heavy Metal Recovery by Modeling PEUF has usually been carried out
Polymer-Enhanced UF through the classic complexation (or exchange
ions) equilibrium reactions. For instance, the
Jose Sabate union between a heavy metal ion, Mm+, and a
Department dEnginyeria Agroalimentaria i neutral ligand group of the polymer, L, can be
Biotecnologia, Universitat Politecnica de represented by the following chemical equilibri-
Catalunya, Castelldefels, Spain ums and their respective constants:
Polymer-enhanced ultraltration (PEUF) is one Mm nL MLn m bn (1)

of the membrane operations that can be used
to separate or concentrate heavy metal from For n=1, 2, 3. . .
Hemodialysis (HD) Membranes 913
polymer polymer
F F R R
polluted polluted P P
water water
Mm+ Mm+
Heavy Metal Recovery by Polymer-Enhanced UF, Fig. 1 Scheme of PEUF. (F: feed, R: retentate, P: permeate).
The protonation of the L group is described by: Cross-References

H
LH L H Ka (2) Heavy Metal Recovery by Membrane
Operations
High values of bn indicate a strong afnity Heavy Metal Recovery by Chelating Agents
between polymer and target ion. Generally, the and Membranes
basic forms of ligands (R-COO, R-NH2, etc.) Nanoltration
bind the metal ions, but the acid forms Ultraltration (UF)
(R-COOH, R-NH3+, etc.) do not do it. This fact
can be used for polymer recycling. When an acid
References
is added to the metal-polymer complex solution,
metal ions are shifted by protons from the ligand Mungray AA, Kulkarni SV, Mungray AK (2012) Removal
sites. A second ultraltration operation yields to a of heavy metals from wastewater using micellar
permeate containing free heavy metal ions and a enhanced ultraltration technique: a review. Cent Eur
retentate with polymer. The polymer can be J Chem 10:2746
Rivas BL, Pereira E, Maureira A (2009) Functional water-
regenerated by the alkali addition and therefore soluble polymers: polymermetal ion removal and bio-
reused. The global process requires the consump- cide properties. Polym Int 58:10931114
tion of an acid and an alkali, and it is a concentra-
tion operation as the permeate in acid medium
contains most of the heavy metal from the initial
contaminated water, but in a smaller volume. Hemodialysis (HD) Membranes
Nanoltration and reverse osmosis membranes
are able to retain and therefore concentrate heavy Michel Jaffrin
metal ions without the need of chemical addition. Department of Biological Engineering,
Nevertheless, they require higher pressures and Technological University of Compiegne,
energy consumption than UF. Selectivity is an BP 20529, Compiegne Cedex, France
additional advantage of PEUF as bn values are
particular for any metal-polymer pair.
Micellar-enhanced ultraltration (MEUF) is a Introduction
similar process to PEUF. Instead of a polymer, a
surfactant solution over the critical micellar con- Hemodialysis (HD) consists in recirculating the
centration is used. Micelles bound to metal ions patients blood into a hemodialyzer with an ultra-
and are large enough to be rejected by UF mem- ltration membrane which permits to remove ure-
branes (Mungray et al. 2012). mic toxins mainly by diffusion, while excess
914 Hemodialysis (HD) Membranes
water is removed by ultraltration. The treatment Hemodialysis (HD) and Hemodiafiltration (HDF)
of chronic renal disease requires three dialysis
sessions of 4 h per week, as well as dietetic HD HDF
restrictions to limit potassium intake. In 2008,
1.58 millions of patients were treated by HD Post dilution
worldwide. Without reuse, assuming a mean
membrane area of 0.9 m2 per hemodialyzer, they
would have consumed 213 106 m2 of membrane
per year. If 75 % of dialysis centers reuse dialyzers
2.54 times (Chuang et al. 2008), this consumption
is reduced to 116 106 m2, which is still the biggest
application of membranes. The rst disposable
hemodialyzers were commercialized in 1950, but
it is the arteriovenous shunt invented by Dr. Scriber
from Seattle which permitted modern dialysis by
connecting the dialyzer inlet to an arm artery and
the outlet with an arm vein, permitting a blood ow
rate of 200300 ml/min into the dialyzer. Pre dilution
Blood Dialysis
fluid Blood Dialysis
Physiological Requirements fluid
Membrane
Diffusion + Increased
Dialysis must reproduce the four main functions
Diffusion convective transport
of kidney:
Hemodialysis (HD) Membranes, Fig. 1 Difference
Elimination of uremic toxins, urea, creatinine, between HD and HF modes
uric acid, phosphates, and sulfates
Removal of excess body water by ultraltra-
tion, 13 L per run
Control of plasma electrolytes concentrations Dialysis Membranes
Na+, Cl, K+, Ca++, by adjusting dialysate They are divided into two categories, low and
ionic concentration middle permeability cellulosic membranes and
Maintaining blood pH to 7.2 through an acetate high permeability polymeric membranes, which
or bicarbonate buffer in dialysate are listed in Table 1. They are generally of hollow
ber geometry (200 mm diameter), but some at
In regular hemodialysis, only the excess water sheet membranes are also available. Cuprophan
is removed by ultraltration, while in membranes are now replaced by cellulose acetate
hemodialtration (HDF) about 1020 L of water or triacetate which is more biocompatible,
is removed from blood per session to eliminate even though they activate complement and
toxins and middle molecules by convection in leukocytes, leading to inammatory reaction
addition to diffusion, and it is replaced by a sterile (Bour and Vanholder 2004). Masking hydroxyl
dialysate except for the part corresponding to groups increases their biocompatibility. High
excess water. Figure 1 illustrates the difference permeability membranes permit to carry out
between HD and HDF. In predilution mode, the hemodialtration and hemoltration treatments
sterile dialysate is reinjected in blood at dialyzer with high water removal, partially replaced by
inlet, while it is reinjected at outlet in postdilution. sterile dialysate. In addition, polymeric mem-
Hemoltration (HF) is analogous to HDF with up branes are often more biocompatible than cellu-
to 25 L water removal, but without feeding dialy- losic ones. They are also superior for patient
sate in the dialyzer as toxin, removal by diffusion quality of life, reducing b2-M amyloidosis, infec-
is not necessary. tions, and loss of residual kidney function.
Heterogeneous Ion-Exchange Membranes 915
Hemodialysis (HD) Membranes, Table 1 Hemodialysis membranes

Type Plane Fiber Manufacturer
Membrane Dialyzer Permeability ml/h-mmHg m2
Low and middle permeability,
Cuprophan X X AKZO Baxter, Bellco, Gambro 23.5
Cuprophan HDF X AKZO Gambro 8
Cuprophan TAF X Terumo Terumo
Hemophan X AKZO CD Medical
Cellulose acetate X 5
Cellulose triacetate X Baxter, Nipro
High permeability
PMMA X Toray Toray 14
Polyacrylonitrile PAN X X Hospal Hospal, Asahi 25
PAN S X Hospal Hospal 40
Polysulfone X Fresenius Fresenius 3040
Polysulfone X Amicon (HF) Amicon 50 H
Polyamide X Gambro (HF) Gambro 30
High-ux membranes with larger pores increase References

removal of middle molecules such as b2-
microglobulin, cytokines, resistin, immunoglobu- Bour T, Vanholder R (2004) Which dialyser membrane to
choose? Nephrol Dial Transplant 19:293296
lins, parathyroid hormones, and dinucleoside
Chuang FR, Lee CH, Chuang HW, Lee CN, Chen TC
polyphosphates. Their removal protects the (2008) A quality and cost benet analysis of dialysers.
patient against cardiovascular problems and Ren Fail 30:521526
other side effects of uremic syndrome Vanholder R, Glorieux G, Van Biesen W (2011) Advan-
tages of new hemodialysis membranes and equipment.
(Vanholder et al. 2011).
Nephron Clin Pract 114:c165c170
Characteristics of Recent European

Hemodialyzers
Hospal Gambro (Meyzieu, France) commercial- Heterogeneous Ion-Exchange
izes three high-ux hollow bers dialyzers: the Membranes
Evodial with heparin grafted to AN membrane,
giving very good hemocompatibility; the Nephral Mitsuru Higa
ST with PEI, a biopolymer grafted on AN69; and Graduate School of Science and Engineering,
the Revaclear with a Poractron membrane Yamaguchi University, Ube, Yamaguchi, Japan
permitting to reduce its area. Gambro (Sweden)
sells, for HF and HDF, the polyux H with a
polyamidepolyacrylether sulfone membrane. Heterogeneous ion-exchange membrane consists
Fresenius Medical Care (Germany) commercial- of nely powdered ion exchanger and a binder
izes high-ux dialyzers with a Helixone mem- polymer, and, in many cases, the membrane is
brane, the FX60, 800, and 1,000 with urea reinforced by woven cloth or net to improve its
clearances of up to 278 ml/min at a blood ow mechanical properties. In general, heterogeneous
of 300 ml/min, and other dialyzers with ion-exchange membranes (IEMs) are prepared by
polysulfone membrane, both at high ux (F 40S the following method (Sata 2004): nely pow-
to F80S with area from 1 to 1.8 m2) and at low ux dered organic and/or inorganic ion exchanger is
(F4HPS to F8HPS with area from 0.7 to 2.4 m2 homogeneously mixed and heated with a thermo-
according to patient weight). plastic polymer such as poly(vinyl chloride),
916 Heterogenization or Immobilization of a Catalyst
Heterogeneous Ion-Exchange Membranes, Fig. 1 Schematic diagram of heterogeneous IEMs with various kinds of
shapes
Astom. Co., Ltd., Japan). EDCORE is a mem-

brane electrode apparatus, with smooth-surfaced
seamless and tubular IEMs (Fig. 2), and is used in
industries such as electro-deposition painting of
automobile, building materials, house appliance,
and other applications due to their high mechani-
cal strength and ease of handling.
References
Heterogeneous Ion-Exchange Membranes, http://www.astom-corp.jp/en/en-main2-edcore.html

Fig. 2 Tubular-type heterogeneous IEM, EDCORE Kiyono R, Koops GH, Wessling M, Strathmann H (2004)
Mixed matrix microporous hollow bers with
ion-exchange functionality, J Membr Sci 231: 109115
polyethylene, polypropylene, or other engineer- Sata T (2004) Ion exchangemembrane. The Royal Society
of Chemistry, Cambridge
ing plastics, and then the mixture is formed into Vermaas DA, Saakes M, Nijmeijer K (2011) Power
the membrane by pressing and/or heating. Hetero- generation using proled membranes in reverse elec-
geneous IEMs have slightly lower electrochemi- trodialysis, J Membr Sci 385-386. 234-242
cal properties: lower permselectivity for
counterions and/or higher membrane resistance
than homogeneous ion-exchange membrane. Heterogenization or Immobilization
However, heterogeneous IEMs are easily pre- of a Catalyst
pared and have high mechanical strength. More-
over, the IEMs of various kinds of shapes can be Enrica Fontananova
easily prepared by pressing and/or heating as Institute on Membrane Technology, National
shown in Fig. 1: (a) commercial at-sheet IEMs Research Council of Italy, ITM-CNR, Rende,
such as Ralex CMH and AMH (Mega a.s., Czech Italy
Republic), (b) proled IEM for reverse electrolyte
applications (Vermaas et al. 2011), (c) hollow
ber-type IEMs (Kiyono et al. 2004), and The heterogenization or immobilization of a
(d) commercial tubular-type IEMs (EDCORE, catalyst in/on a membrane consists in the
Heterogenization or Immobilization of a Catalyst 917
incorporation of a catalyst in/on a membrane by Leaching out from the membrane of the
chemical bonds or physical entrapment. catalyst
In the case of a homogeneous catalyst, the
incorporation constitutes a way to heterogenize it. However, catalytic membranes deserve special
When a catalyst is incorporated within or on advantages that in many cases justify these addi-
the surface of a membrane, the membrane com- tional efforts.
position (characteristics of the membrane mate- In the design of a catalytic membrane, major
rial, hydrophobic or hydrophilic, presence of issues in the polymer selection are the mechanical,
chemical groups with specic functionality, etc.) thermal, and chemical stability under reaction
and the membrane structure (dense or porous, conditions.
symmetric or asymmetric) can positively inu- What is fundamental is to realize a stable cat-
ence the catalyst performance, not only by the alyst immobilization in order to avoid its leaching
selective absorption and transport of out from the membrane. Different immobilization
reagents and/or products but also inuencing the strategies can be used in order to achieve these
catalyst activity by electronic and conformational goals: covalent bonds, electrostatic and weak
effect. interactions (van der Waals or hydrogen bonds) H
The main requirements that must be considered of the catalyst with the membrane, or catalyst
to produce an ideal catalyst are: low costs, high encapsulation.
selectivity, high stability under reaction condi- A good afnity for the catalyst is desirable in
tions, nontoxicity, green properties, and rst of order to avoid catalyst leaching and to have a good
all recoverability, i.e., the possibility to recover adhesion between the polymer and the catalyst,
and reuse the catalyst (Gladysz 2001; Hill 1999). with an optimal dispersion of the second one.
In this perspective, the heterogenization of cat- The afnity between a compound and a poly-
alysts has interesting implications because it meric material can be often anticipated by the
allows the reuse several times of the same catalyst. calculation of the afnity parameters (Van
Among the different heterogenization methods, Krevelen 1990; Barton 1990). These parameters
the entrapping of a catalyst in/on a membrane reect the ability to form hydrogen bond, polar,
offers new possibility for the design of new cata- and van der Waals interactions in condensed
lytic processes. phase.
In addition to the fundamental criteria that have The polymer/catalyst afnity can be also
to be considered in the selection of a membrane improved by an appropriate chemical functiona-
for a specic separation process, i.e.: lization of one or both components.
In the case of application in liquid phase, ide-
Stability (mechanical, thermal, and chemical) ally, the solvent used for the reaction needs to be a
Permeability and selectivity non-solvent for the catalyst, and the membrane
Fouling tendency has not to have an excessive swelling in it.
Cost (capital and operating) It is also possible in some cases to improve the
catalyst retention in the membrane by the catalyst
Extra technical complexities need to be con- enlargement as: dendrimers, hyperbranched poly-
sidered in the design and realization of catalytic mers, and catalyst bound to a soluble polymer
membranes: (Dioos et al. 2006).
Mass transfer of the reagents to the catalytic
Catalyst stability during the immobilization sites, and of the products away from them, should
procedure be fast enough in order to not limit the reaction,
Inuence of the catalyst particles on the mem- but, at the same time, the contact time catalyst/
brane properties (mechanical, transport, etc.) reagent should be also appropriate.
Effect of the catalyst loading, dimensions, and For a porous membrane, the choice of the
distribution, on the catalytic efciency polymer material is of less importance for
918 Heteropolyacid Composite Membranes
transport properties in comparison with a dense Chemicals Production, Catalysis by Metal Complexes
membrane, because permeation does not take 33, Springer, Berlin
Gladysz JA (2001) Recoverable catalysts. Ultimate goals,
place through the polymer matrix, but through criteria of evaluation, and the green chemistry interface.
the membrane pores (Vankelecom 2002). Pure Appl Chem 73:13191324
However, the membrane material is relevant Hill CL (1999) Homogeneous catalysis. Controlled green
for the membrane stability and surface properties, oxidation. Nature 401:436437
such as wettability and fouling tendency. to membrane science and technology. Publisher CNR,
The membrane preparation conditions depend Roma. ISBN 88-8080-063-9
on the membrane material and the desired struc- Van Krevelen DW (1990) Properties of polymers, 3rd edn.
ture and morphology. In the case of a catalytic Elsevier Science, Amsterdam
Vankelecom IFJ (2002) Polymeric membranes in catalytic
membrane, the incorporation of the catalyst com- reactors. Chem Rev 102:37793810
plicates the process because the catalyst should
be rmly entrapped in the membrane and the
catalyst structure and activity has to be preserved
during the membrane preparation procedure.
Moreover, the properties (including transport Heteropolyacid Composite
properties) of the catalytic membrane are usually Membranes
different than those of the reference membrane
prepared in the same conditions but without the Enrica Fontananova
catalyst. Institute on Membrane Technology, National
Different techniques for membrane preparation Research Council of Italy, ITM-CNR, Rende,
are today available and can be opportunely Italy
employed also to prepare catalytic membranes:
phase separation, coating, interfacial polymeriza-
tion, sintering, stretching, and track etching are Heteropolyacid composite membranes are
some examples (Strathmann et al. 2006). membranes functionalized with heteropolyacids
In some cases, the catalyst is entrapped in/on (HPAs).
the membrane already formed, by covalent Acid catalysis and catalytic oxidation are two
bonds, electrostatic interactions, absorption by of the main application areas of HPAs composite
weak interactions, or physical entrapment. In membranes. Moreover, interesting examples of
alternative, the catalyst immobilization can be HPAs immobilized in membranes are present in
carried out in the same time of the membrane the eld of the polymer electrolyte membrane fuel
formation, for example, by dispersing the cata- cells (PEMFCs) (Norby 1999).
lyst in the polymer casting solution and succes- There are many different structural types of
sive phase separation (Drioli and Fontananova HPAs, but the Keggin type (Mio et al. 2005) is
2010). the most common for the majority of the applica-
tions as catalysts or solid state proton conductors.
The HPAs discussed in the following text are
References molybdophosphoric acid (H3PMo12O40),
phosphotungstic acid (H3PW12O40), and
Barton AFM (1990) Handbook of polymer-liquid interac- silicotungstic acid (H4SiW12O40).
tion parameters and solubility parameters. CRC Press,
Catalysis by heteropoly acids has well-
Boca Raton
Dioos BML, Vankelecom IFJ, Jacobs PA (2006) Aspects of established industrial applications including het-
immobilisation of catalysts on polymeric supports. Adv erogeneous catalyzed selective oxidations
Synth Catal 348:14131446 (oxidation of methacrolein to methacrylic acid
Drioli E, Fontananova E (2010) Catalytic membranes
and ethylene to acetic acid) and acid-catalyzed
embedding selective catalysts: preparation and applica-
tions. In: P. Barbaro and F. Liguori (eds.), reactions (homogeneous liquid-phase hydration
Heterogenized homogeneous catalysts for ne of olens, biphasic polymerization of THF to
Heteropolyacid Composite Membranes 919
poly(tetramethylene glycol) and the gas-phase HPAs typically exist in hydrated phase with the
synthesis of ethyl acetate from ethylene and degree of hydration varying from 6 to 30 mole-
acetic) (Kozhevnikov 2007). cules of water, depending on the temperature and
Nevertheless, in many processes using HPA the relative humidity (e.g., six molecules of water
catalysts, there are relevant difculties concerning are present in the H3PW12O40 structure up to
catalyst separation and recovery and reaction 175230 C) (Mio et al. 2005). In Keggin-type
selectivity. In this perspective, the heteroge- HPAs, hydrogen bonds exist between each acid
nization of HPAs in membranes can offer impor- proton and two water molecules, are also present
tant advantages in catalysis by the integration of between water molecules and the terminal oxygen
the catalytic activity with the selective transport atoms of the HPAs, and moreover, involving dif-
through the membranes. ferent Keggin units (Moffat 1986).
Hydration of a-pinene to a-terpineol over If the HPA sizes are in the order of a few
molybdophosphoric acid entrapped in polyvinyl nanometers and uniformly distributed in the pro-
alcohol (PVA) membranes cross-linked with ton exchange membrane, the water hydration mol-
succinic acid, and molybdophosphoric acid ecules and/or dioxonium ions of the additive can
encaged in zeolites and dispersed in be used to better interconnect the ionic cluster of H
polydimethylsiloxane (PDMS) membranes, the polymeric matrix providing a better pathway
have been reported in literature (Castanheiro for proton hopping (Ramani et al. 2005).
et al. 2003). In both cases, the catalytic activity Composite membranes with higher glass tran-
of the membranes is quite good in successive runs sition temperature (Tg) and higher stability of the
although a partial catalyst leaching was observed. proton conductivity at high temperature have been
Catalytic membranes have been prepared obtained by incorporation of 60 wt.% of three
by linking phosphotungstic acid on the surface different HPAs (H3PW12O40, H3PMo12O40,
of plasma-modied membranes made of Na2HPW12O40) in SPEEK membranes with
polyvinylidene uoride (PVDF). The PVDF 70 %, 74 %, and 80 % degree of sulfonation
membranes was modied by plasma treatments (Zaidi et al. 2000).
on the surface, in order to graft N-containing Hybrid membranes prepared from sulfonated
polar groups able to act as binding sites with polyetherketone (SPEK), heteropoly acids, and an
phosphotungstic acid. The catalytic activity of inorganic network of ZrO2 or RSiO3/2 have been
the membranes has been evaluated in the aerobic also prepared obtaining a higher proton conduc-
degradation reaction of phenol in water tivity of the membranes (Ponce et al. 2003).
(Fontananova et al. 2006). Higher phenol degra- The HPAs have been also used to stabilize the
dation rate was observed for the HPA proton conductivity of the Naon membranes at
heterogenized on PVDF membrane compared to elevated temperature and low relative humidity.
homogeneous catalyst dissolved in solution. For example, H3PW12O40, also exchanged with
Moreover, the catalytic membranes have given larger cations such as Cs+, NH4+, Rb+, and Tl+,
proof of their complete stability under photooxi- has been heterogenized in Naon membranes
dation condition and their good recycle. observing better proton conductivity (Ramani
In numerous literature studies, it has been et al. 2005).
reported that the HPAs can reduce the resistance Tungstophosphoric acid, silicotungstic acid,
to the proton transport in PEMFCs. and phosphomolybdic acid have been used as
The reduction of the HPAs particles sizes and functional additives in composite membranes
their good dispersion inside the polymeric matrix made of the sulfonated derivative of an
resulted to be key factors in order to have a pos- amorphous-modied polyetheretherketone, called
itive effect on the proton conductivity because of SPEEK-WC (Fontananova et al. 2010).
the increase of the surface to volume ratio of the The embedding of the inorganic HPAs, in par-
additives and the possibility to have a more ef- ticular the H4SiW12O40, in SPEEK-WC mem-
cient proton hopping (Ramani et al. 2005). branes, enhanced the proton conductivity
920 Heteropolyacids
Heteropolyacid
Composite Membranes, O
the interactions by S O
hydrogen bonds between a O S
heteropoly acid (the Keggin
O H2O H3O+ O
structure is represented) H3O+ O
with the sulfonic group of a
cation exchange membrane, H2O
the water hydration H3O+ H2O
molecules, and
dioxonium ions O H3O+
O
S H2O
O
H3O+
H3O+ O
H2O H3O+ S
O
H2O H2O
O
because the HPAs improved the interconnection polyetherketone-heteropolyacid membranes. J Membr

of the ionic clusters in the polymeric matrix pro- Sci 217:515
Ramani V, Kunz HR, Fenton JM (2005) Effect of particle
viding a preferential pathway for proton hopping size reduction on the conductivity of Naon /
(Fig. 1). phosphotungstic acid composite membranes. J Membr
Sci 266:110114
Zaidi SMJ, Mikhailenko SD, Robertson GP, Guiver MD,
Kaliaguine S (2000) Proton conducting composite
membranes from polyether ether ketone and
References heteropolyacids for fuel cell applications. J Membr
Sci 173:1734
Castanheiro JE, Ramos AM, Fonseca I, Vital J (2003) The
acid-catalysed reaction of a-pinene over
molybdophosphoric acid immobilised in dense poly-
meric membranes. Catal Today 82:187193
Fontananova E, Donato L, Drioli E, Lopez L, Favia P,
dAgostino R (2006) Heterogenization of Heteropolyacids
polyoxometalates on the surface of plasma modied
polymeric membranes. Chem Mater 18:15611568 Ubavka B. Mio
Fontananova E, Trotta F, Jansen JC, Drioli E (2010) Prep-
aration and characterization of new non-uorinated Faculty of Physical Chemistry, University of
polymeric and composite membranes for PEMFCs. Belgrade, Belgrade, Republic of Serbia
Kozhevnikov IV (2007) Sustainable heterogeneous acid
catalysis by heteropoly acids. J Mol Catal A Chem
262:8692 The basic structural unit in heteropolyacids
Mio UC, Todorovi MR, Davidovi M, Colomban P, (HPAs) is the Keggins cluster (nd Keggins
Holclajtner-Antunovi I (2005) Heteropoly structure). Groups are distributed in a three-
compounds from proton conductors to biomedical dimensional lattice and are interconnected by
agents. Solid State Ionics 176:30053017
Moffat JB (1986) Cation-anion effects in heteropoly com- hydrogen-bonded water molecules or some other
pounds with Keggin structures. Polyhedron 5:261269 proton species (OH, H2O, H3O+, H5O2+)
Norby T (1999) Solid-state protonic conductors: princi- forming the secondary structure of Keggins
ples, properties, progress and prospects. Solid State anion. Water molecules ll channels in solid
Ionics 125:111
Ponce ML, Prado L, Richau K, Mohr R, Nunes SP HPAs, and next interactions, OH. . .H2O,
(2003) Reduction of methanol permeability in H2O. . .H2O and H3O+. . .H2O, are identied.
Heteropolyacids 921
Empirical RO. . .O distances, between 294 and

263 nm, correspond to weak and medium hydro-
gen bonds. a 14-29 H2O
Three best-known acids have been investigated:
H3PMo12O40 29H2O (MoPA), H3PW12O40
29H2O (WPA), and H4SiW12O40 30H2O
(WSiA). All three acids are isostructural. Among
HPAs 12-tungstophosphoric acid hydrates, charac-
a H5O2+ 6 H2O
terized by strong acidity, are classied into
superacids (Okuhara et al. 2004). From mother
solutions HPAs crystallize with high number of
water molecules, 2930, forming quasi liquid
state. Acids are very sensitive to ambient condi-
tions: relative humidity and temperature, and lose a H+ 0.2 H2O
very easily water molecules and change to lower
hydrates, Eq. 1 and Fig. 1. On the other hand, H
dehydrated acids are very hygroscopic.
Therefore it is interesting and necessary to see
how these compounds change in process of calci-
a 0 H2O
nation/dehydration and to determine structural
phase transition and temperature regions of their
stabilities, Fig. 2. Investigations have to be
multidisciplinary and different physicochemical Heteropolyacids, Fig. 1 Schematic survey of different
methods have to be used. Thermal analysis steps of hydration of HPAs
(TGA, DTA), XRPD, IR, and Raman spectros-
copy are recommended for the beginning (Pope
and Mi uller 1994; Mio et al. 1990, 1994, 2005).
Thanks to such investigations besides the phase At higher temperatures, about 400 C (the
transition to different crystal hydrates: 29, 21, broad peak) was dened as Tg anomaly on
14, and 6 hydrates, dehydrated phase, WPA-0, DTA curves. It corresponds to exothermic trans-
at about 230 C, was dened as cubic phase. It formation which is followed by the mass loss
was postulated that Keggins anions of WPA-0 (TGA) corresponding to about 0.50.3 water
phase are stabilized by protons free protons. molecules.
20-31C 31-35C 42-60C 60-175C
WPA-29, cubic WPA-21, orthorhombic WPA-14, triclinic WPA-6, cubic
(1)
175-230C 230-580C 580-604C 604-1000C
0-WPA, cubic 0-WPA/D-WPA, cubic D-WPA, cubic PWB
It was evident that these water molecules are Nowadays, this water is called protonated (Janik
formed from protons and few oxygen atoms et al. 2004). Metastable (D-phasedenuded phase)
from host lattice in endothermic process, which is stable up to 604 C, for WPA, when structural
is simultaneous with water dehydration. phase transformation to phosphate tungsten
922 Heteropolyacids
Heteropolyacids, a 100 12
Fig. 2 DTA and TGA
curves of WPA-29 in 10
98
Endo. DT Exo.
different temperature 8
regions. Scanning rates 96
(wt %)
6
were 1, 5, and 10 C/min
(mV )
94 4
2
92
0
90 2
10 20 30 40 50 60 70 80 90 100
t (C)
b 99.5 4
3
Endo. DT Exo.
98.5 2
1
97.5
(wt %)
(mV )
1
96.5 2
3
95.5 4
5
94.5 6
0 100 200 300 400 500 600 700

t (C)
c 100 25
98 20
Endo. DT Exo.
96
15
(wt %)
94
(mV )
10
92
5
90
0
88
86 5
0 200 400 600 800 1000
t (C)
DTA and TGA curves of 29-WPA
at temperatures up to 1000 C
bronzes (PWB) happen (Mio et al. 1994, 2005). He claims that a new Q phase is formed after
All WPA phases were indexed and corresponding the loss of constitutional water (Fournier
unit cell dimension calculated. Some of them for et al. 1992).
the rst time: WPA-14 as triclinic. All phases from These investigations have been followed by IR
WPA-6 up to structural phase transformation are and Raman spectroscopic study in situ, in temper-
cubic. Phosphate tungsten bronzes (PWB) ature range from temperature of liquid nitrogen
have also been obtained and dened for the 195 to 220 C, and with 18O-labeled com-
rst time using HPAs as precursor (Mio pounds. On the basis of the obtained spectra, few
et al. 1994). Fournier noted also the same large proton species were identied (H5O2+, H3O+,
peak for WPA and identied it as a new Q phase. H2O, and OH groups): terminal OH groups of
Hexavalent Chromium Separation by Supported Liquid Membranes 923
Keggins anions at about 3,550; OH groups from

water molecules, due to the most frequent inter- Hexavalent Chromium Separation
action H2O. . .H2O at about 3,490; and OH bands by Supported Liquid Membranes
belonging to H3O+. . .H2O, the strongest interac-
tion at about 3,292 and 3,225 and 3,140 cm1; Pietro Argurio
bands in the region 2,2502,050 were assigned to Department of Environmental and Chemical
n3 and/or n4 of H3O+ and water libration modes. Engineering, University of Calabria, Arcavacata
Dioxonium ion was identied also, thanks to the di Rende, Italy
band n4 at 1,720 and n2 at about 1,220 cm1.
Special attention in the process of calcination/
dehydration was paid to the intensity changes of The removal and/or recovery of heavy metals from
bands n4(H3O+) at 1,720 and n2(H2O) at industrial wastewater is a major topic of research.
1,610 cm1 (Mio et al. 1990, 1994, 2005). Chromium is a unique, toxic heavy metal released
Protonic entities and their dynamic equilibrium in aqueous environment in both +3 and +6 oxida-
inuence the host structure too. Identication of tion states. Hexavalent chromium (Cr(VI))
proton species and following of its dynamic equi- receives particular attention because of its muta-, H
librium is also very important for understanding terato- and carcinogenic properties. Cr(VI) exists
catalytic (formation of Brnsted acid centers) and as anionic species, such as HCrO4, Cr2O72, and
conductive processes (dening charge carriers, CrO42, which are highly mobile on subsurface
conductivity dependence of temperature, fre- environment (Kumbasar 2008). These anionic spe-
quency, and so on) as characteristics of these cies are bioaccumulative, and their oxidizing
compounds. potentials make them highly toxic to biological
systems. The major industries that contribute to
water pollution by chromium are mining, leather
References tanning, textile dyeing, electroplating, metal
nishing, and corrosion inhibition (Rajasimman
Fournier M, Feumi-Jantou C, Rabia C, Herve G, Launay and Karthic 2010).
S (1992) Polyoxometalates catalyst materials: X-ray The World Health Organization recommends
thermal stability study of phosphorus-containing the toxic limits of Cr(VI) in wastewaters at the
heteropolyacids H3+xPM12xVxO40 13-14H2O
level of 0.005 mg/L. Many countries have regula-
(M = Mo,W; x = 01). J Mater Chem 2:971978
Janik MJ, Davis RJ, Neurock M (2004) A rst principles tions of the maximum permissible concentration
analysis of the location and afnity of protons in the of Cr(VI) in natural or drinking water which typ-
secondary structure of phosphotungstic acid. J Phys ical tolerance limit for discharge into inland sur-
Chem B 108:1229212300
face waters is 0.1 mg/L and in potable water is
Mio U, Colomban P, Novak A (1990) Infrared and Raman
study of some heteropolyacid hydrates. J Mol Struct 0.05 mg/L. Usually Cr(VI) concentrations in
218:123128 industrial wastewaters range from 0.5 to
Mio UB, Dimitrijevi R, Davidovi M, Nedi ZP, 270,000 mg/L. Then chromium-bearing wastewa-
Mitrovi MM, Colomban P (1994) Thermally induced
ters must be discharged into aquatic environments
phase transformations of 12-tungstophosphoric acid
29-hydrate: synthesis and characterization of or onto land after appropriate treatments to dras-
PW8O26-type bronzes. J Mater Sci 29:37053718 tically reduce Cr(VI) content.
Mio UB, Todorovi MR, Davidovi M, Colomban P, The methodologies for Cr(VI) recovery, from
Holclajtner-Antunovi I (2005) Heteropoly
compounds-from proton conductors to biomedical
industrial wastewater, range from ion exchange to
agents. Solid State Ion 176:30053017 solvent extraction, non-dispersive solvent extrac-
Okuhara T, Nishimura T, Watanabe H, Misono M (2004) tion, precipitation, and adsorption. Solvent extrac-
Insoluble heteropoly compounds as highly active cata- tion has been widely used for the removal and/or
lysts for liquid-phase reaction. J Mol Catal 74:247256
recovery of chromium in hydrometallurgy since
Pope MT, Muller A (eds) (1994) Polyoxometalate: from
platonic solids to anti-retroviral activity. Kluwer, this technique allows the Cr ions recovery, but it
Dordrecht involves high capital and operating costs due to
924 Hexavalent Chromium Separation by Supported Liquid Membranes
large inventory of solvent, especially in the case of Feed Liquid membrane Strip
dilute solutions. Conventionally the treatment
chromium-bearing wastewaters consists of the HCrO4 H2CrO4 TOPO HCrO4
reduction of Cr(VI) to Cr(III) with an adequate
chemical-reducing agent. Cr(III) is then easily
precipitated by the addition of an alkali compound
(generally calcium hydroxide) to the liquid efu-
ent. The main drawback of this method is the H+ TOPO H+
production of a large amount of sludge-containing
chromium often in high concentration, which dis- HCrO4 HCrO4
posal/treatment is a very costly affair and it is not H+ H+
eco-friendly.
Hexavalent Chromium Separation by Supported Liq-
Supported liquid membrane (SLM) process uid Membranes, Fig. 1 Schematics of separation mech-
offers a technology with a low solvent/extractant anism through liquid membrane
consumption since it involves in a single stage the
extraction and stripping processes, which are gen-
erally performed in two separate steps in conven- The so-formed complex diffuses through the
tional solvent extraction. SLM consists of a LM membrane toward the stripping basic solution
phase, impregnated in the pores of a thin hydro- where the de-complexation reaction takes place
phobic microltration membrane, kept there by and HCrO4 and H+ ions are released. The
capillary forces. It combines the typical advan- so-regenerated carrier molecule diffuses back to
tages of liquid membrane with the mechanical the feed and the transport cycle begins again
resistance of solid membranes. The transport (Fig. 1). This transport mechanism is the
across a SLM is mediated by a mobile extractant so-called facilitated coupled co-transport, typical
(carrier) which binds very selectively the target when a basic carrier like amines or phosphates is
component in the donor phase (feed), transporting used to transport negatively charged species
it into the acceptor phase (strip), resulting in the (in this case HCrO4) and usually H+ as counter-
so-called facilitated transport (Molinari ion across the membrane in the same direction
et al. 2009a, b). The selection of an appropriate (Fig. 1).
carrier provides higher selectivity and enrichment Cr(VI) complexes can be efciently removed
factor as compared to the other separation from acidic chloride aqueous solutions by facili-
methodologies. tated transport with TOA into a basic (NaOH
Despite of their advantages with respect to the 0.1 M) acceptor phase (Kozlowski and
traditional separation techniques, SLM is not Walkowiak 2002). In agreement with the transport
widely applied in treatment of chromium-bearing mechanism that is similar to that one previously
wastewaters. Cr(VI) compounds could be reported for TOPO, the permeability coefcient
removed from dilute aqueous solutions by using and then initial ux values decrease linearly by
trioctylphosphine oxide (TOPO), Alamine increasing the feed pH. Cr(VI) concentration can
336, tri-n-octylamine (TOA), and tributyl phos- be successfully reduced in the feed phase from 1.0
phate (TBP) as ionic carrier. to 0.0028 mg/L, thus respecting the World Health
The extraction of HCrO4 with TOPO from Organizations recommendations.
acidic solutions could be expressed by the follow-
ing equation (Kumbasar 2009; Hasan et al. 2009):
References

HCrO4 aq H aq TOPOorg
Hasan MA, Selim YT, Mohamed KM (2009) Removal of
, H2 CrO4 TOPOorg chromium from aqueous waste solution using liquid
emulsion membrane. J Hazard Mater 168:15371541
High Free Volume Polymer 925
Kozlowski CA, Walkowiak W (2002) Removal of chro- V 1:3V w

mium (VI) from aqueous solutions by polymer inclu- fv
V
sion membranes. Water Res 36:48704876
Kumbasar RA (2008) Studies on extraction of chromium
(VI) from acidic solutions containing various metal where V is the specic volume (reciprocal of den-
ions by emulsion liquid membrane using Alamine sity) and Vw is the specic van der Waals volume,
336 as extractant. J Membr Sci 325:460466 which for many polymers may be estimated using
Kumbasar RA (2009) Extraction of chromium (VI) from
group contribution methods. The factor 1.3 takes
multicomponent acidic solutions by emulsion liquid
membranes using TOPO as extractant. J Hazard Mater account of the fact that molecules cannot
167:11411147 completely ll space even in a perfectly ordered
Molinari R, Argurio P, Poerio T (2009a) Studies of various crystal at absolute zero and is an average value for
solid membrane supports to prepare stable sandwich
crystalline materials. By this denition, most
liquid membranes and testing copper(II) removal from
aqueous media. Sep Purif Technol 70:166172 glassy polymers have fv < 0.2. Some glassy poly-
Molinari R, Argurio P, Poerio T (2009b) Flux enhancement mers, however, have much higher fractional free
of stagnant sandwich compared to supported liquid volumes and consequently exhibit high perme-
membrane systems in the removal of Gembrozil
abilities to gases and vapors (Budd and McKeown
from waters. J Membr Sci 340:2634
Rajasimman M, Karthic P (2010) Application of response 2010; Yampolskii 2012; Pinnau and Toy 1996; H
surface methodology for the extraction of chromium Starannikova et al. 2008; Thomas et al. 2009).
(VI) by emulsion liquid membrane. J Taiwan Inst Chem The structures of some polymers with exception-
Eng 41:105110
ally high free volume are shown in Fig. 1.
A common feature of high free volume poly-
mers is that they have relatively inexible, twisted
High Free Volume Polymer backbones, which cannot change conformation in
order to ll space efciently. In substituted
Peter M. Budd polyacetylenes such as PTMSP, the bulky side
Organic Materials Innovation Centre, School of group inhibits rotation about single bonds in the
Chemistry, University of Manchester, backbone and forces the backbone into a twisted
Manchester, UK shape. In peruoropolymers such as Teon
AF2400, neighboring dioxolane rings cannot eas-
ily rotate past each other. In substituted
Free volume is the unoccupied space between polynorbornenes prepared by addition polymeri-
molecules. The concept of free volume is used to zation, such as PTMSN, ring structures and bulky
explain molecular motion in liquids and solids. In side groups restrict rotation about backbone
a liquid or rubber, free volume increases with bonds. The ultimate extension of this idea is
increasing temperature. A exible polymer will found in polymers of intrinsic microporosity,
ow or behave as a rubber at temperatures at such as PIM-1, which have no single bonds in
which there is sufcient free volume for large- the backbone about which rotation can occur but
scale movements of polymer segments but will which incorporate sites of contortion (the spiro
behave as a glass when the temperature is reduced centers in PIM-1) to force the backbone to twist
to the point where there is not enough free volume and turn.
for such movements. The polymers shown in Fig. 1 are soluble and
Different assumptions may be made about can readily be processed from solution to form
what constitutes occupied and what constitutes membranes. However, there are other high free
free volume in a material. Thus, in different volume polymers which cannot be prepared in
contexts, different values may be quoted for the soluble form. Sometimes, it is possible to form a
amount of free volume in a polymer. In membrane membrane from a soluble precursor and subse-
science, it is common to dene fractional free quently convert it to the desired structure through
volume as chemical or thermal treatment. For example,
926 High Free Volume Polymer
High Free Volume Polymer, Fig. 1 Molecular structures (PTMSN), and a polymer of intrinsic microporosity
and fractional free volumes of poly(1-trimethylsilyl-1- prepared from 5,50 ,6,60 -tetrahydroxy-3,3,30 ,30 -tetramethyl-1,
propyne) (PTMSP), a copolymer of 2,2-bistriuoromethyl- 10 -spirobisindane and 1,4-dicyanotetrauorobenzene
4,5-diuoro-1,3-dioxole and tetrauoroethylene (Teon (PIM-1)
AF2400), addition-type poly(trimethylsilyl norbornene)
polybenzoxazole structures can be prepared by material (pore size <2 nm) (Rutherford 2001), in
thermal rearrangement from aromatic hydroxyl- which case the apparent micropore distribution
containing polyimides (Park et al. 2007). can be determined from gas adsorption studies
The permeation of gases and vapors through a (McKeown and Budd 2010).
polymer depends not only on the amount of free
volume but also on the size, distribution, and
connectivity of free volume elements. Computer References
simulation is useful for visualizing the free vol-
ume distribution in amorphous polymers Budd PM, McKeown NB (2010) Highly permeable poly-
(Hofmann et al. 2002). A number of experimental mers for gas separation membranes. Polym Chem
1:6368
techniques, notably positron annihilation lifetime Hofmann D, Heuchel M, Yampolskii Y, Khotimskii V,
spectroscopy (PALS), have been employed to Shantarovich V (2002) Free volume distributions in
obtain information about the size and distribution ultrahigh and lower free volume polymers: comparison
of free volume elements. In a high free volume between molecular modeling and positron lifetime
studies. Macromolecules 35:21292140
polymer, there may be sufcient connectivity McKeown NB, Budd PM (2010) Exploitation of intrinsic
between free volume elements for the polymer to microporosity in polymer-based materials. Macromol-
behave like a molecular sieve or microporous ecules 43:51635176
High Selective Carbon Membranes 927
Park HB, Jung CH, Lee YM, Hill AJ, Pas SJ, Mudie ST, alcohol (PFA), a polypyrrolone, and poly
Van Wagner E, Freeman BD, Cookson DJ (2007) Poly- (2,6-dimethyl-1,4-phenylene oxide) (PPO).
mers with cavities tuned for fast selective transport of
small molecules and ions. Science 318:254258 Results for O2/N2 selectivity and O2 permeability
Pinnau I, Toy LG (1996) Gas and vapour transport prop- are sometimes used as benchmarks for characteriz-
erties of amorphous peruorinated copolymer mem- ing the properties of membranes as a function of
branes based on 2,2-bistriuoromethyl-4,5-diuoro- synthesis variables. It can be clearly seen that most
1,3-dioxole/tetrauoroethylene. J Membr Sci
109:125133 carbon membranes have O2/N2 selectivities that
Rutherford SW (2001) Mechanism of sorption and diffu- are roughly three times higher than the upper
sion in a high free volume polymer. Ind Eng Chem Res limit for polymer membranes when the selectivities
40:13701376 are compared at the same O2 permeability. Previ-
Starannikova L, Pilipenko M, Belov N, Yampolskii Y,
Gringolts M, Finkelshtein E (2008) Addition-type ous reports have shown that carbon membranes
polynorbornene with Si(CH3)3 side groups: detailed have considerable promise for commercial appli-
study of gas permeation and thermodynamic proper- cations, not only in O2/N2 separation but also in
ties. J Membr Sci 323:134143 hydrogen purication and the separation of CO2/
Thomas S, Pinnau I, Du N, Guiver MD (2009) Pure and
CH4, CO2/N2, or propene/propane mixtures, which
mixed-gas permeation properties of a microporous
spirobisindane-based ladder polymer (PIM-1). are all processes in which conventional polymer H
J Membr Sci 333:125131 membranes do not perform adequately.
Yampolskii Y (2012) Polymeric gas separation mem- Optimal conditions for the production of car-
branes. Macromolecules 45:32983311
bon membranes vary according to the nature of
the precursor polymers and their morphologies,
because the various polymers show different ther-
mal decomposition behaviors. Control of the con-
High Selective Carbon Membranes ditions for pyrolysis of the precursor polymer, as
discussed below, is the most important factor in
Miki Yoshimune increasing the gas selectivity of the resulting car-
Research Institute for Chemical Process bon membrane, because the nature of the pyroly-
Technology, National Institute of Advanced sis process has a considerable effect on the pore
Industrial Science and Technology (AIST), structures of the membrane.
Tsukuba, Japan
Pyrolysis Temperature
In general, carbon membranes prepared under
appropriate conditions show higher selectivities in The pyrolysis temperature is generally chosen to
gas separations than do conventional polymer be above the decomposition temperature of the
membranes (Ismail and David 2001; Sau and polymer, but below its graphitization temperature
Ismail 2004; Kita 2006; Ismail and Li 2008; Wil- (5001000 C). As a general rule, an increase in
liams and Koros 2008; Yoshimune and Haraya the pyrolysis temperature reduces the permeabil-
2011). Examples of O2/N2 separation perfor- ity of the resulting carbon membrane but increases
mances of carbon membranes, taken from litera- its selectivity. High pyrolysis temperatures pro-
ture sources, are plotted in Fig. 1, which also shows duce increased crystallinity and lower average
the upper limits for polymer membranes (Kita interplanar spacings in carbon membranes,
et al. 1997; Suda and Haraya 1997; Shiett and resulting in smaller micropores and sharper pore-
Foley 2000; Ghosal and Koros 2000; Kim size distributions. This behavior certainly depends
et al. 2004, 2005; Shao et al. 2004; Yoshimune on the physical properties of the polymer, so that a
et al. 2005; Xiao et al. 2005a, b; Robeson 2008). suitable carbonization temperature needs to be
The carbon membranes were prepared from chosen individually for a particular polymer pre-
various precursors, including various aromatic cursor to achieve optimal performance in gas
polyimides (Kapton, Matrimid, etc.), polyfurfuryl separations.
928 High Selective Carbon Membranes
High Selective Carbon 100 Kapton [a] Matrimid [b]

Membranes,
BTDA-ODA [c] 6FDA-DAI [d]
Fig. 1 Comparison of the
oxygen permeabilities and 6FDA/BPDA-DAM [e] BTDA-ODA/DBA [f ]
selectivities of various 6FDA/PMDA-TMMDA [g] PFA [h]
carbon membranes with the Polypyrrolone [i] PPO [j]
upper limit for polymer
membranes ([a] Suda and
Haraya 1997, [b] Xiao
et al. 2005a, [c] Kim O2 /N2 selectivity
et al. 2005, [d] Xiao
et al. 2005b, [e] Ghosal and 10
Koros 2000, [f] Kim
et al. 2004, [g] Shao
et al. 2004, [h] Shiett and
Foley 2000, [i] Kita
et al. 1997, [j] Yoshimune
et al. 2005, [k] Robeson
2008)
Upper bound of polymer membranes [k]

1
0.1 1 10 100 1000 10000
O2 Permeability (Barrer)
Thermal Soak Time standpoint, however, an optimal heating rate that

is not too low should be chosen, because low
The thermal soak time can have varied effects on heating rates increase the costs and time involved
the performance of the nal membrane. Changing in producing carbon membranes.
the thermal soak time, particularly at the nal Another method for tuning the gas selectivities
pyrolysis temperature, can be used effectively to of carbon membranes is the production of mixed-
ne-tune the permeation properties of a carbon matrix membranes (Yoshimune and Haraya 2011).
membrane. A prolonged thermal soak time Mixed-matrix carbon membranes are dened as
increases the selectivity of carbon membranes products of the pyrolysis of membranes consisting
but decreases their permeability. of a continuous polymer matrix containing nano-
sized inorganic particles dispersed throughout the
polymer. Metal ions or particles, silica particles,
Heating Rate zeolites, or carbons are normally used as inorganic
particles to improve the properties of the resulting
The heating rate determines the rate of evolution membranes. Such nano-sized particles increase the
of volatile components from a polymeric mem- polarity of the membranes, form interlayer spaces,
brane during pyrolysis and, consequently, affects or have an afnity toward target gases.
the nature of the pores that are formed in the Figure 2 shows the hydrogen permeability and
resulting carbon membrane. Lower heating rates selectivity of several types of mixed-matrix car-
(5 C/min or less) favor the formation of small bon membranes (Park et al. 2004; Yoda
pores and increase the crystallinity of the resulting et al. 2004; Park and Lee 2005; Liu et al. 2006;
carbon, thereby producing carbon membranes Yoshimune et al. 2006; Grainger and Hgg 2007).
with increased selectivities. From the practical Some of the mixed-matrix carbon membranes
High Selective Carbon Membranes 929
High Selective Carbon Cu-carbon (H2/CH4) [l]

Membranes, 100000
Pd-carbon (H2/N2) [m]
Fig. 2 Comparison of the
hydrogen permeabilities M-carbon (H2/N2) [n]
Pure Kapton carbon membrane for H2/N2 [a]
and selectivities of various Si-carbon (H2/N2) [o]
mixed-matrix carbon 10000
Si-carbon (H2/CH4) [p]
membranes with the upper
Z-carbon (H2/N2) [q]
limits for polymer
H2/N2 or H2/CH4 selectivity

membranes ([a] Suda and
Haraya 1997, [k] Robeson 1000
2008, [l] Grainger and Hgg
2007, [m] Yoda et al. 2004,
[n] Yoshimune et al. 2006,
[o] Park et al. 2004, [p] Park 100
and Lee 2005, [q] Liu
et al. 2006)
10
H2/N2 upper bound of polymer membranes [k]

H
1
1 10 100 1000 10000 100000
H2 Permeability (Barrer)
exhibit signicantly superior performances in References

comparison to that of the Kapton-derived carbon
membrane, which is shown here as the boundary Ghosal AS, Koros WJ (2000) Air separation properties of
at sheet homogeneous pyrolytic carbon membranes.
for pure-carbon membranes. This technique can
help to improve the trade-off between the perme- Grainger D, Hgg MB (2007) Evaluation of cellulose-
ability and the selectivity of carbon membranes, derived carbon molecular sieve membranes for hydro-
as well as provide novel functionalities in mem- gen separation from light hydrocarbons. J Membr Sci
306:307317
brane reactors.
Ismail AF, David LIB (2001) A review on the latest devel-
High selectivity by carbon membranes is gener- opment of carbon membranes for gas separation.
ally observed in the case of dense membranes in J Membr Sci 193:118
forms such as at sheets or hollow bers. For Ismail AF, Li K (2008) From polymeric precursors to
hollow ber carbon and ceramic membranes. In:
commercial applications, it is necessary to
Mallada R, Menendez M (eds) Inorganic membranes:
fabricate membranes into modules that have synthesis, characterization and applications. Elsevier,
both a high selectivity and a high permeability. Amsterdam, pp 81120
Thinner separation layers, such as asymmetric hol- Kim YK, Lee JM, Park HB, Lee YM (2004) The gas
separation properties of carbon molecular sieve mem-
low bers, in combination with adequate mechan-
branes derived from polyimides having carboxylic acid
ical strength are desirable; however, carbon groups. J Membr Sci 235:139146
membranes are brittle in handling, and even a Kim YK, Park HB, Lee YM (2005) Preparation and
small number of pinholes can ruin their gas selec- characterization of carbon molecular sieve
membranes derived from BTDA-ODA polyimide
tivity. This hinders the production of carbon mem-
and their separation properties. J Membr Sci
branes on an industrial scale. Although carbon 255:265273
membranes have considerable potential for use, Kita H (2006) Gas and vapor separation membranes based
not only in gas and vapor separations but also as on carbon membranes. In: Yampolskii Y, Pinnau I,
Freeman BD (eds) Material science of membranes for
membrane reactors, further investigations are
gas and vapor separation. Wiley, Chichester,
needed before they can be used commercially. pp 337354
930 High-Defined Porous Films
Kita H, Yoshino M, Tanaka K, Okamoto K (1997) Gas (phenylene oxide) and its derivatives for gas separation.
permselectivity of carbonized polypyrrolone mem- Chem Lett 34:958959
brane. J Chem Soc Chem Commun 11:10511052 Yoshimune M, Fujiwara I, Suda H, Haraya K (2006) Gas
Liu Q, Wang T, Liang C, Zhang B, Liu S, Cao Y, Qiu transport properties of carbon molecular sieve mem-
J (2006) Zeolite married to carbon: a new family of branes derived from metal containing sulfonated poly
membrane materials with excellent gas separation per- (phenylene oxide). Desalination 193:6672
formance. Chem Mater 18:62836288
Park HB, Lee YM (2005) Fabrication and characterization
of nanoporous carbon/silica membranes. Adv Mater
17:477483 High-Defined Porous Films
Park HB, Jung CH, Kim YK, Nam SY, Lee SY, Lee YM
(2004) Pyrolytic carbon membranes containing silica
derived from poly(imide siloxane): the effect of silox- Honeycomb Membrane Structure
ane chain length on gas transport behavior and a study
on the separation of mixed gases. J Membr Sci
235:8798
Robeson LM (2008) The upper bound revisited. J Membr
Sci 320:390400 Higher (C3+) Hydrocarbon Removal
Sau SM, Ismail AF (2004) Fabrication of carbon mem-
branes for gas separation a review. Carbon Arnaud Baudot
42:241259
Shao L, Chung TS, Wensley G, Goh SH, Pramoda KP Physics and Analysis Division, IFP Energies
(2004) Casting solvent effects on morphologies, gas Nouvelles, Solaize, France
transport properties of a novel 6FDA/PMDA-
TMMDA copolyimide membrane and its derived car-
bon membranes. J Membr Sci 244:7787
Shiett MB, Foley HC (2000) On the preparation of Rubbery polymer membrane materials have the
supported nanoporous carbon membranes. J Membr ability to permeate preferentially condensable
Sci 179:275282 vapors. Rubbery membranes were found to be
Suda H, Haraya K (1997) Gas permeation through micro- very protable for the recovery of high-value
pores of carbon molecular sieve membranes derived
from kapton polyimide. J Phys Chem monomer in petrochemical plants purge gases
B 101:39883994 (Table 1). In the case of large polymerization
Williams PJ, Koros WJ (2008) Gas separation by facilities, the value of purge monomers can reach
carbon membranes. In: Li NN, Fane AG, Ho amounts up to two million USD/year.
WSW, Matsuura T (eds) Advanced membrane
technology and applications. Wiley, Hoboken,
pp 599632 Higher (C3+) Hydrocarbon Removal,
Xiao Y, Dai Y, Chung TS, Guiver MD (2005a) Effects of Table 1 Membrane-based olefin recovery from polymer
brominating matrimid polyimide on the physical and resin degassing vent (MTR Inc. membrane modules) Baker
gas transport properties of derived carbon membranes. and Jacobs 1996; Baker et al. 2000
Macromolecules 38:1004210049
Olen ow
Xiao Y, Chung TS, Chng ML, Tamai S, Yamaguchi
Feed rate recovered
A (2005b) Structure and properties relationships for
Application composition (recovery yield)
aromatic polyimides and their derived carbon mem-
branes: experimental and simulation approaches. LLDPE/HDPE Butene 36 % C2 = 145 kg/h
J Phys Chem B 109:1874118748 reactor purge Ethylene 41 % (88 %)
Yoda S, Hasegawa A, Suda H, Uchimaru Y, Haraya K, Ethane 1 % nC4 = 142 kg/h
Tsuji T, Otake K (2004) Preparation of a platinum and Nitrogen 21 % (98 %)
palladium/polyimide nanocomposite lm as a precur- Hydrogen 1 %
sor of metal-doped carbon molecular sieve membrane HDPE resin 1-butene 57 % iC4 = 1160 kg/h
via supercritical impregnation. Chem Mater degassing CO2/O2 2 % (97 %)
16:23632368 Nitrogen 40 %
Yoshimune M, Haraya K (2011) Microporous carbon Water 1 %
membranes. In: Basile A, Gallucci F (eds) Polypropylene Propylene 14 % C3 = 500 kg/h
Membranes for membrane reactors. Wiley, Chichester, resin degassing Propane 0.3 % (91 %)
pp 6398 Nitrogen 84.4 %
Yoshimune M, Fujiwara I, Suda H, Haraya K (2005) Novel Hydrogen 1 %
carbon molecular sieve membranes derived from poly Water 0.3 %
Highly Permeable Polyimides 931
References most common monomers for the formation

of high-performance polyimides (Fig. 1),
Baker RW, Jacobs M (1996) Improve monomer recovery and 2,3,5,6-tetramethyl-1, 4-phenylenediamine
from polyolen resin degassing. Hydrocarbon Proc
(4MPDA). The hindrance of the four methyl
75:4951
Baker RW, Lokhandwala KA, Jacobs ML, Gottschlich DE groups of the di-aniline restricts the rotation
(2000) Recover feedstock and product from reactor around the imide linkage allowing the synthesis
vent streams. Chem Eng Progr 96:5157 of more porous (so less dense) polyimides.
The higher porosity facilitates the mass transport
of the gases through the pores of the membrane
with consequent increase of the permeability
Highly Permeable Polyimides (molecular sieving effect), typically over
100 Barrer for O2 and 400 Barrer for CO2
Mariolino Carta with selectivity aO2 =N2 3:5 and aH2 =N2 16:6
School of Chemistry, The University of (Lin et al. 2000).
Edinburgh, Edinburgh, Scotland, UK A different method to increase the FFV is
represented by the formation of hyperbranched H
polyimides. In this case a triamine monomer,
Polyimides are polymers produced by cycloimi- such as 1,3,5-tris(4-aminophenoxy)benzene
dization (formation of an imide linkage) between (TAPOB), is reacted with different di-anhydrides
di-anhydrides and diamines via step-growth poly- to obtain a highly branched structure (Tsai
merization (Koros and Fleming 1993; Ghosh and et al. 2003) (Fig. 2). The polyimides obtained
Mittal 1996). Among other useful applications with this technique are insoluble so, to be used
such as materials for electronics, coatings, foam, as membrane for gas separation, they must be
and bers, because of their excellent solubility in embedded in a support. Typically, it can be pre-
common low boiling point solvents, thermal sta- pared by the dispersion of the polyimide with
bility, and physical properties, they have been colloidal silica by solgel processes, to form a
extensively studied as gas separation membranes. composite material. The most remarkable charac-
A typical limitation for their use in this eld is due teristic of this hyperbranched polymer is that they
to the fact that they often exhibit high selectivity can be made out of a large variety of A3 + B2
but low permeability for important gas pairs such terminal groups so that the synthesis can afford
as O2 and CO2, usually far below 100 Barrer different multifunctional polymers.
(1 Barrer = 1010 cm3(STP) cm cm2 s1 Following the idea of restricting the rotation
cmHg1). The low permeability is strongly around the imide link, the concept of polymers of
related with the lack of fractional free volume of intrinsic microporosity (PIMs), which is based on
the material (FFV) because of the free rotation the polymerization of monomers that possess a
around the imide linkage that allows the polymer site of contortion, was applied to the synthesis of
to pack densely, limiting the mass transport. Initial polyimides. In this case, di-anhydrides and/or
successes in enhancing the permeability by di-anilines such as the spirobisindane A or the
increasing the FFV have been achieved with the ethanoanthracene B (as shown in Fig. 3) are
use of 4, 40 -(hexauoroisopropylidene)diphthalic employed. The site of contortion is represented
anhydride (6FDA), which is now one of the
Highly Permeable Polyimides, Fig. 1 6FDA + 4MPDA-based polyimide (Lin et al. 2000)
932 Highly Permeable Polyimides
Highly Permeable Polyimides, Fig. 2 Hyperbranched polyimides (Tsai et al. 2003)
Highly Permeable Polyimides, Fig. 3 Highly permeable polyimides of intrinsic microporosity (Rogan et al. 2014)
by the central quaternary carbon in case of A or (for this reason they are called polymers of intrin-
the methylene bridge for B (Rogan et al. 2014). sic microporosity).
Although there is still free rotation around the Studies on this class of polymers demonstrate
imide linkage, the big hindrance of the bulky that the systematic increase of the rigidity of the
substituent, combined with the reduced exibility structural backbone favors the molecular sieving
given by the site of contortion, allows the forma- effect of the material, enhancing both permeabil-
tion of high FFV polymers that cannot pack ity and selectivity for various gas pairs. The
space efciently in the solid state, leaving resulting PIM-polyimides (PIM-PIs) possess
interconnected pores of nano-/microdimension high molecular mass and elevated microporosity,
High-Pressure Electrolysis 933
with BET surface areas in the range of 600700 water and has two electrodes connected with an
m2g1. These features, associated with the high external power supply. Electric current causes
solubility in common organic solvents, allow the positively charged hydrogen ions to migrate to
preparation of robust thin-lm membranes with the negatively charged cathode, where a reduction
excellent performance toward important takes place in order to form hydrogen atoms. The
commercial gas pairs, such as O2/N2, H2/N2, and atoms formed then combine to form gaseous
CO2/CH4, with exceptional permeabilities (over hydrogen molecules (Eq. 1). On the other hand,
1000 Barrer for O2 and over 7000 Barrer for CO2) oxygen is formed at the other electrode (the pos-
accompanied by good selectivities aO2 =N2 3:5 itively charged anode) (Eq. 2). The amount of
aH2 =N2 11:5,and aCO2 =CH4 16:1: Remarkably, gases produced per unit time is directly related to
all data points lie above the Robeson upper the current that passes through the electrochemi-
bounds. With the appropriate choice of the mono- cal cell (Wendt and Kreysa 1999).
mers of both di-anhydrides and di-anilines, it is Oxygen and hydrogen gas can be generated at
possible to tune the physical properties of these noble metal electrodes by the electrolysis of
highly permeable polyimides to improve them water: Positive electrode (anode):
even further. H
4OH $ 2H 2 O O2 4e (1)
Negative electrode (cathode):

References
Ghosh M, Mittal KL (1996) Polyimides: fundamentals and 2H

aq 2e $ H 2gas (2)
applications. Marcel Dekker, New York
Koros WJ, Fleming GK (1993) J Membr Sci 83:l80
Lin WH, Vora RH, Chung TS (2000) J Polym Sci In a differential pressure conguration, only the
B 38:27032713 cathode (hydrogen) side is under pressure, and
Rogan Y, Malpass-Evans R, Carta M, Lee M, Jansen JC, this can eliminate the hazards related to handling
Bernardo P, Clarizia G, Tocci E, Friess K, Lanc M, pressurized oxygen. The high-pressure operation
McKeown NB (2014) J Mater Chem A 2:48744877
Tsai FY, Harding DR, Chen SH, Blanton TN (2003) Poly- of an electrolyser brings the advantage of deliver-
mer 44:9951001 ing hydrogen at a high pressure (sometimes called
electrochemical compression) for the end user,
thus requiring less energy to further compress
and store the hydrogen. It also diminishes the
volume of the gaseous phase at the electrodes
High-Pressure Electrolysis thus signicantly improving product gas removal
which follows Ficks law of diffusion (Carmo
Csar Valderrama et al. 2013).
Department of Chemical Engineering, Universitat A number of advantages can be gained from
Politcnica de Catalunya-Barcelona Tech (UPC), operating an electrolyser at high pressure; on the
Barcelona, Catalonia, Spain one hand, specic power consumption is
reduced, then the cost of gas compressors is
reduced or eliminated by delivery of gas pressure,
High-pressure electrolysis (also called high- and nally, the size of electrolysis cells is
pressure electrolysis HPE) is the water electroly- reduced. Though the reversible cell voltage
sis performed at pressures higher than ambient in increases theoretically with pressure, the overvolt-
which electrical energy is the driving force of the age is reduced by higher operating pressure and
water decomposition to produce oxygen (O2) and decreased gas volume, and a small overall reduc-
hydrogen (H2). The core of an electrolysis unit is tion in the cell voltage is usually found (Zoulias
an electrochemical cell, which is lled with pure et al. 2004).
934 High-Pressure Electrolysis
Types of Electrolysers membrane (Naon , fumapem ) provides high

proton conductivity, low gas crossover, compact
The electrolyser unit consists in the simplest prin- system design, and high-pressure operation
ciple of two electrodes separated by an electrolyte. (Carmo et al. 2013).
The role of the electrolyte is to close the electrical The reinforcement of the cells with metallic
circuit by allowing ions (but not electrons) to support on the anode side allows the generation
move between the electrodes. Moreover it keeps of hydrogen at high pressures (up to 160 bar com-
the produced gases separated. Three common mercial units) (Badwal et al. 2013). The energy
types of electrolysers are used: (i) alkaline penalty for generating hydrogen at higher pres-
electrolysers, (ii) polymer electrolyte membrane sures is very small, that is, an order of magnitude
(PEM) electrolysers, and (iii) solid oxide electrol- increase in hydrogen pressure requires only
ysis cells (SOECs). 30 mV per cell increase in voltage (as shown
Alkaline electrolysers (AEC) represent a very by Nernst equation).
mature technology that is the current standard for PEM electrolysis cells achieve high current
large-scale electrolysis. The anode and cathode densities already at low operational temperatures
materials in these systems are typically made of from 20 C to 100 C. PEM electrolysers are not
nickel-plated steel and steel, respectively. The utilized in market applications to the same extent
electrolyte in these systems is a liquid based on a as alkaline ones due to the high investment costs.
highly caustic KOH solution. The ionic charge One of the most important components of such a
carrier is the hydroxyl ion, OH-, and a membrane unit is a separating membrane that allows the
porous to hydroxyl ions, but not to H2, and O2 passage of ions, or electrons and not oxygen, or
provides gas separation. Key advantages of this hydrogen atoms. This membrane allows the gases
technology include its maturity and its durability. to be kept separate in order to avoid the risk of an
Key disadvantages are its use of a highly caustic explosive mixture being formed in the electrolysis
electrolyte and its inability to produce hydrogen at unit. The pressure increase minimizes the expan-
high pressures (Jensen et al. 2008). This inability sion and dehydration of the membrane, preserving
to produce high-pressure hydrogen for storage the integrity of the catalytic layer (Grigorev
results in the added need for an external compres- et al. 2001). The membrane technology enables
sor, which adds cost and complexity to the system. safe differential pressure operation, which in turn
Water electrolysis by solid oxide electrolysis allows the electrolysis unit to be operated in a
cells (SOECs) is principally performed at temper- load-following mode with 100 % turndown capa-
atures in between 700 C and 1000 C (see entry bility, which is very attractive for renewable
High-Temperature Electrolysis) due to the H2 energy applications.
electrolysis reaction which becomes increasingly The electrolyser utilizes pure water as the cir-
endothermic with temperature and the ohmic culating uid, which results in a simpler balance
losses in SOECs which decrease with increasing of plant. The differential pressure also results in
temperature. The high temperature allows hydro- the ability to operate the oxygen-water uid loop
gen production with signicantly reduced electri- at ambient pressure, which has signicant safety
cal energy demand (Larminie and Dicks 2003). advantages as well as use of plastic tubing rather
In an acidic polymer electrolyte membrane than expensive high-pressure piping.
(PEM) electrolysis cell, water splitting is The one disadvantage is the acidic environ-
performed according to the reactions shown by ment within the membrane, which limits the cat-
(Eqs. 1 and 2). Protons are formed at the anode alysts to noble metal oxides and requires
under water consumption, where the oxygen evo- semiprecious metals in the oxygen ow elds.
lution reaction (OER) takes place. The protons The ion exchange membrane is also more sensi-
migrate through the solid polymer electrolyte tive to impurities than the liquid electrolyte, since
(SPE) membrane to the cathode, where they are cations can bind to the proton-conducting sites,
reduced into hydrogen. The polymer electrolyte reducing conductivity (Zoulias et al. 2004).
High-Rejection Seawater Reverse Osmosis Membrane 935
References virtually impermeable barrier to salts and other

components. Over the years, it has evolved from
Badwal SPS, Giddey S, Munnings C (2013) Hydrogen the rst-generation asymmetric cellulose acetate
production via solid electrolytic routes. WIREs Energy
(CA) membrane to the current generation thin lm
Environ 2:473487. doi:10.1002/wene.50
Carmo M, Fritz DL, Mergel J, Stolten D (2013) composite (TFC) membranes with better salt
A comprehensive review on PEM water electrolysis. retention and higher water permeability, which
Int J Hydrog Energy 38:49014934 has signicantly improved the water quality and
Grigorev SA, Khaliullin MM, Kuleshov NV, Fateev VN
reduced the energy consumption and costs of sea-
(2001) Electrolysis of water in a system with a solid
polymer electrolyte at elevated pressure. Russ water desalination.
J Electrochem 37(8):819822
Jensen OJ, Jensen SH, Tophoj N, Bandur V, Ebbesen S, Structure of TFC RO Membrane
et al (2008) Pre-investigation of water electrolysis. Ris
and Separation Mechanism
DTU National Laboratory for Sustainable Energy
2008. http://130.226.56.153/rispubl/NEI/NEI-DK- TFC RO membrane, prepared using interfacial
5057.pdf. Accessed June 2014 polymerization technique, is composed of an
Larminie J, Dicks A (2003) Fuel cell systems explained, active skin layer (thickness 200 nm and pore
size of <0.1 nm) commonly made from cross- H
Wendt H, Kreysa G (1999) Electrochemical engineering:
science and technology in chemical and other indus- linked aromatic polyamides, a polysulfone porous
tries. Springer, New York support layer (thickness 50 mm), and a polyester
Zoulias E, Varkaraki E, Lymberopoulos N, Christodoulou nonwoven fabric substrate (thickness 100 mm),
CN, Karagiorgis GN (2004) A review on water elec-
as shown in Fig. 1 (Kurihara and Tomioka 2010;
trolysis. TCJST 4(2):4171
Uemura et al. 2012).
The active layer determines the separation
properties, whereas both the support layer and
substrate merely provide mechanical support for
High-Rejection Seawater Reverse the active layer to withstand high-operating pres-
Osmosis Membrane sures of up to 7 MPa. The physicochemical prop-
erties associated with the separation function are
Tzyy Haur Chong1, Rong Wang1 and characterized by pore size, surface morphology or
Anthony Gordon Fane2 roughness, surface charge, and hydrophobicity
1
Singapore Membrane Technology Centre, (Bellona et al. 2004). Retention of compounds is
Nanyang Environment and Water Research based on solute-membrane interaction mecha-
Institute; School of Civil and Environmental nisms, namely, steric hindrance, electrostatic
Engineering, Nanyang Technological University, repulsion for charged solutes, and hydrophobic-
Singapore, Singapore hydrophobic and dipole-dipole interactions. The
2
Singapore Membrane Technology Centre, transport mechanism for RO membranes can be
Nanyang Environment and Water Research described by the widely accepted solution-
Institute, Nanyang Technological University, diffusion model (Wijmans and Baker 1995).
Challenges for RO Membrane Desalination
The commercially available TFC RO membranes
Introduction have excellent salt retention properties, typically
greater than 99.5 % (Baker 2004). However,
Seawater desalination to convert seawater to pota- boron removal, which is usually less than 95 %,
ble water by reverse osmosis (RO) membranes is a major challenge to RO membranes to meet the
has become a pivotal technology to meet the recommended value of below 0.5 mg/L in drink-
increasing water demand. The RO membrane ing water by the World Health Organization
has a unique property that allows passage of (Jonsson and Macedonio 2010; Misdan
water with little hindrance while presenting a et al. 2012). Boron exists naturally in seawater
936 High-Rejection Seawater Reverse Osmosis Membrane
High-Rejection Seawater
Reverse Osmosis
Membrane,
Fig. 1 Structure of
composite RO membrane Active skin layer
(Uemura et al. 2012)
Support layer
Substrate
(pH 78) in the form of uncharged boric acid with wall, thus resulting in the increase of solute pas-
pKa value of 9.2 and at concentration of sage (Fane et al. 2009). Hence, a high-retention
5 mg/L. Below pH 9, the non-hydrated boric membrane needs to possess good antifouling
acid is poorly rejected due to the lack of electric properties to ensure consistent performance in
double layer repulsive force between the nega- long-term operation.
tively charged membrane surface and boric acid.
Current practice for boron removal in seawater Future Direction
desalination is either to install a second pass RO Incorporation of nano-materials and biomaterials
or ion exchange column to further purify the water into the active layer is seen to be the innovative
or to increase the pH to above 9 so that the mem- way to produce the next generation of high-
brane can easily reject the hydrated boric acid. permeability and high-retention seawater RO
The latter is least preferred as it can promote membranes, i.e., nanocomposite or biomimetic
membrane scaling by hardness in the water. membranes (Matsuura and Fane 2013). In addi-
Another concern is the incomplete retention of tion, a thorough understanding on the molecular
trace organic micro-pollutants such as endocrine- structures of membrane and its interaction with
disrupting compounds, pesticides, pharmaceuti- solutes via molecular simulation is necessary to
cally active compounds, etc. (Bellona allow better manipulation of surface properties
et al. 2004). In general, due to the nature of the such as the extent of cross-linking, chain-packing
membrane, charged/hydrophilic molecules have density, and chain stiffness in order to enhance the
better retention than uncharged/hydrophobic mol- performance of RO membranes.
ecules. However, predictive models to account for
the organic solute transport through the mem-
brane are still lacking.
References
In addition, true retention is an intrinsic prop-
erty of the membrane, though the observed reten- Baker RW (2004) Membrane technology and applications.
tion is strongly inuenced by the process Wiley, Chichester
parameters such as operating pressure, tempera- Bellona C, Drewes JE, Xu P, Amy G (2004) Factors affect-
ing the rejection of organic solutes during NF/RO
ture, pH, and membrane fouling. For instance, treatment a literature review. Water Res
when a membrane is fouled, the solutes no longer 38:27952809
freely move in the concentration polarization Fane AG, Chong TH, Le-Clech P (2009) Fouling in mem-
boundary layer, instead need to diffuse through brane processes. In: Drioli E, Giorno L (eds) Membrane
operations: innovative separations and transformations.
the tortuous paths in the fouling layer. Conse-
Wiley-VCH, Weinheim, pp 121138
quently, the hindered back-diffusion effect causes Jonsson G, Macedonio F (2010) Fundamentals in reverse
greater accumulation of solutes at the membrane osmosis. In: Drioli E, Giorno L (eds) Comprehensive
High-Temperature Electrolysis 937
membrane science and engineering, vol 2. Elsevier, negative electrode (cathode):

Oxford, pp 122
Kurihara M, Tomioka H (2010) Preparation of industrial
RO, NF membranes, and their membrane modules and 2H
aq 2e $ H 2gas (2)
applications. In: Drioli E, Giorno L (eds) Comprehen-
sive membrane science and engineering,
From the thermodynamic viewpoint of water
vol 2. Elsevier, Oxford, pp 2334
Matsuura T, Fane AG (2013) Editorial: new directions in decomposition, it is more advantageous to elec-
desalination. Desalination 308:1 trolyze water at high temperature (8001,000 C)
Misdan N, Lau WJ, Ismail AF (2012) Seawater reverse because the energy is supplied in mixed form of
osmosis (SWRO) desalination by thin lm composite
electricity and heat. The main advantage is that a
membrane current development, challenges and
future prospects. Desalination 287:228237 substantial part of the energy needed for the elec-
Uemura T, Kotera K, Henmi M, Tomioka H (2012) Mem- trolysis process is added as heat, which is much
brane technology in seawater desalination: history, cheaper than electric energy. In addition, the high
recent developments and future prospects. Desalin
temperature accelerates the reaction kinetics,
Water Treat 33(13):283288
Wijmans JG, Baker RW (1995) The solution-diffusion reducing the energy loss due to electrode polari-
zation, thus increasing the overall system ef-
model: a review. J Membr Sci 107:121
H
ciency (Zoulias et al. 2004).
The minimum necessary cell voltage for the
start-up of electrolysis, E0cell, is given under stan-
dard conditions (P, T constant) by the following
High-Temperature Electrolysis
equation (Allebrod 2013):
Csar Valderrama
Department of Chemical Engineering, Universitat DG0
Politcnica de Catalunya-Barcelona Tech (UPC), E0cell
nF
Barcelona, Catalonia, Spain
where DG0 is the change in the Gibbs free energy
under standard conditions, n is the number of
High-temperature electrolysis (also called steam electrons transferred, and F is the Faraday con-
electrolysis) is the water electrolysis at tempera- stant. The overall water electrolysis cell reaction,
tures that ranged between 700 and 1,000 C in E0 (25 C), is 1.23 V, the Gibbs free energy change
which electrical energy is the driving force of of the reaction is +237.2 kJ/mol, and the standard
water splitting to produce oxygen (O2) and hydro- reaction enthalpy is D0 = 285.8 (kJ/mol). The
gen (H2). The core of an electrolysis unit is an water splitting is an endothermic reaction, and the
electrochemical cell, which is lled with pure term DG decreases with increasing temperature
water and has two electrodes connected with an (Zeng and Zhang 2010).
external power supply. At a certain voltage, which Thermodynamic conditions of high-
is called critical voltage, between both electrodes, temperature electrolysis are more favorable in
the electrodes start to produce hydrogen gas at the the sense that the molar Gibbs energy of the reac-
negatively biased electrode (Eq. 1) and oxygen tion (DG) drops from ~1.23 V (237 kJ/mol) at
gas at the positively biased electrode (Eq. 2). The ambient temperature to ~0.95 V at 900 C
amount of gases produced per unit time is directly (183 kJ/ mol), while the molar enthalpy of the
related to the current that passes through the elec- reaction (DH) remains essentially unchanged
trochemical cell (Wendt and Kreysa 1999). (DH ~1.3 V or 249 kJ/mol at 900 C) (Larminie
Oxygen and hydrogen gas can be generated at and Dicks 2003). A signicant part of the energy
noble metal electrodes by the electrolysis of required for an ideal (loss free) high-temperature
water: positive electrode (anode): electrolysis can thus be provided by heat. Thus,
these systems can utilize waste heat from nuclear
4OH $ 2H 2 O O2 4e (1) (especially fast breeder reactors), chemical, and
938 High-Temperature Electrolysis
ceramic and conducts either oxide ions or protons.

Such systems are typically known as solid oxide
electrolysis cells (SOEC). Material choice for the
electrodes and operating temperature is essen-
tially dened by the composition and type of the
electrolyte material used with all other compo-
nents being matched to the electrolyte. The con-
ductivity of the electrolyte is a key parameter as it
is a large contributor to cell losses and denes the
design and operating temperature regime of the
electrolysis cell.
The most widely studied high-temperature
electrolysers are based on O2 conducting
ceramics such as fully or partially stabilized
(with Y2O3, Sc2O3) zirconia (referred to as
High-Temperature Electrolysis, Fig. 1 A schematic of YSZ or ScSZ). The advantages of these
solid oxide electrolysis cell (SOEC) for high-pressure elec- electrolysers are that they can be incorporated
trolysis (Brisse et al. 2008)
with low-grade waste heat from chemical or
power plants or solar thermal energy to reduce
coal- or natural gas-red power plants and reduce the electrical power input and increase efciency;
the electric power requirements. Alternatively, the however, severe limitations on construction mate-
heat may come from solar concentrators, where rials; signicant problems with system design,
high temperatures (above 800 C) can easily be materials degradation, and lifetimes; hydrogen
achieved. production on the anode side of the cell leading
Steam electrolysis or vapor electrolysis is to complex system design; slow start-up and shut-
performed using solid oxide electrolysis cell down; and limited thermal cycling capability are
(SOEC) as shown in Fig. 1. This may be viewed disadvantages of these conducting systems
in simple terms as the reverse operation of a solid (Badwal et al. 2013).
oxide fuel cell. Apart from O2 conducting electrolytes, sev-
The stream passes through the cathode side of eral proton-conducting systems have been studied
the solid electrolyte where hydrogen ions for the electrolysis of steam. In these systems,
are reduced to hydrogen, releasing oxide ions in pure hydrogen is produced on the cathode side
the process (Eq. 3). The oxide ions then of the electrolysis cell with the oxygen being
migrate through the electrolyte to the anode produced in the anode chamber where steam is
where they combine to form oxygen present. Proton conduction in perovskite-
molecules (Eq. 4), releasing an electron current structured SrCeO3, where Ce is partially
following back to the power source (Zeng and (510 %) substituted by Y, Yb, Mg, or Sc, was
Zhang 2010). rst developed and used for steam electrolysis in
The reactions on two electrodes are: the 1980s. Following this early work, a number of
other materials (based on BaCeO3, CaZrO3,
Cathode : H2 O 2e ! H2 O2 (3) SrZrO3, and BaZrO3) have been investigated,
and their performance in steam electrolysis cells
1 evaluated. Hydrogen is produced on the cathode
Anode : O2 ! O2 2e (4) side that simplies system design, but in addition
2
to the disadvantages of O2 conducting systems,
There are a number of materials that are used for very few electrolytes offer good proton conduc-
the fabrication of electrolysers that operate above tivity and stability in the operating regime
500 C. The electrolyte is typically an oxide (Badwal et al. 2013).
High-Throughput Membrane Technology 939
Hydrogen production via steam electrolysis Cross-References

may involve less electrical energy consumption
than conventional low-temperature water Hybrid Membranes
electrolysis, reecting the improved thermody-
namic and kinetic operating conditions at
References
elevated temperatures. For an average current
density of 7,000 Am2 and an inlet steam tem- Allebrod F (2013) High temperature and pressure alkaline
perature of 1,023 K, the predicted electrical electrolysis. PhD Dissertation. DTU Technical Univer-
energy consumption of the stack is around sity of Denmark, Department for Energy Conversion
3 kWh per normal m3 of hydrogen, which is Baglio V, Ornelas R, Matteucci F, Martina F, Ciccarella G,
Zama I, Arriaga LG, Antonucci V, Arico AS (2009)
signicantly smaller than 4.5 kWh of Solid polymer electrolyte water electrolyser based on
low-temperature alkaline water electrolysis cells naon-tio2 composite membrane for high temperature
commercially available today (Zeng and Zhang operation. Fuel Cells 3:247252
2010). However, this result does not include the Badwal SPS, Giddey S, Munnings C (2013) Hydrogen
production via solid electrolytic routes. WIREs Energy
heating energy circulation and loss. Furthermore,
construction materials, safety issues, and
Environ 2:473487. doi:10.1002/wene.50
Brisse A, Schefold J, Zahid M (2008) High temperature
H
strict temperature control have to be addressed water electrolysis in solid oxide cells. Int J Hydrog
as well. Typical high-temperature electrolyser Energy 33:53755382
Larminie J, Dicks A (2003) Fuel cell systems explained,
achieves 92 % electrical efciency, while 2nd edn. Wiley, New York
low-temperature electrolysers can reach at most Wendt H, Kreysa G (1999) Electrochemical engineering:
85 % efciency (Zoulias et al. 2004). Despite this science and technology in chemical and other indus-
high efciency with respect to electricity, the tries. Springer-Verlag, New York
Zeng K, Zhang D (2010) Recent progress in alkaline water
high-temperature system still produces hydrogen electrolysis for hydrogen production and applications.
at about four times the cost of the steam reforming Prog Energy Combust Sci 36:307326
hydrogen. Zoulias E, Varkaraki E, Lymberopoulos N, Christodoulou
Intermediate-temperature electrolysis: CN, Karagiorgis GN (2004) A review on water elec-
trolysis. TCJST 4(2):4171
A number of proton and O2 conductors are Xu W, Scott K, Basu S (2001) Performance of a high
available for potential use as the solid electrolyte temperature polymer electrolyte membrane water
for water electrolysis in the 100500 C tempera- electrolyser. J Power Sources 196:89188924
ture range. The peruorosulfonic acid polymer
membrane-based electrolysers (PEM) are
generally not operated above 90 C due to a
reduction in proton conductivity caused by its High-Throughput Membrane
dehydration. Approaches used to increase the Technology
operating temperature range of polymer
membranes (fuel cells and electrolysis cells) Sanne Hermans1 and Ivo Vankelecom2
1
include the use of Polybenzimidazole (PBI)/acid Centre for Surface Chemistry and Catalysis,
complex membranes, hybrid membranes Department of M2S, Faculty Bioscience
consisting of polymers, and proton-conducting Engineering, KU Leuven, Leuven, Belgium
2
oxides and composite membranes consisting of Centre for Surface Chemistry and Catalysis,
ceramic particles (SiO2, TiO2, etc.) impregnated KU Leuven, Leuven, Heverlee, Belgium
proton-conducting membranes. Mainly, SiO2- or
TiO2-impregnated peruorinated polymer-
based composite membranes (cells with 15 cm2 The term high-throughput (HT) refers to the
active area) have been employed in creation of a large output, i.e., a large amount of
investigations for temperatures up to 130 C data that is obtained from simultaneous, parallel
(Badwal et al. 2013; Baglio et al. 2009; Xu et al. experiments. In general, HT methodologies are
2011). used to speed up the formulation and discovery
940 High-Throughput Membrane Technology
of new materials or the optimization of processes. properties of multiple membranes in one aerated
They include a variety of advanced tools which MBR (Bilad et al. 2011). In addition, an HT cross
allow fast parallel experiments, leading to higher ow ltration system was adapted into an online
productivities and shorter time to market times. fouling monitoring system, allowing fouling
While HT thus refers to the number of experi- studies of multiple membranes with high-quality
ments, combinatorial strategies are related with imaging and a high reproducibility (Vanysacker
the experimental design. et al. 2013). Filter plate HT platforms were also
Besides automation (to whatever desired applied together with photoinduced graft poly-
degree), a faster screening and more rapid optimi- merization or atmospheric pressure plasma-
zation due to simultaneous testing, other advan- induced graft polymerization to synthesize and
tages of HT technologies are the miniaturization screen antifouling membranes (Zhou et al. 2009;
(leading to smaller sample sizes and reductions in Gu et al. 2013). An HT ionic conductivity appa-
costs and waste streams) and the reproducible ratus and an HT mechanical testing apparatus for
conditions between samples (as parallel tests are thin membranes were also reported (Zapata
performed under identical experimental condi- et al. 2009, 2010).
tions, e.g., pressure, temperature). When new HT devices are developed, their
Membrane technology can gain from the performance should be evaluated based on their
implementation of HT technologies on several reproducibility (of the same sample), interchange-
levels. As membrane performance is inuenced ability (of sample positions), and scalability of the
by a variety of parameters, an efcient HT screen- setup (compared to conventional methods) (Bilad
ing of multiple variables can lead to the rapid et al. 2011).
development of new membranes and membrane Together with the application of HT technol-
applications. ogy, the design of experiments (DoE) has
HT devices can be implemented during the become increasingly important. As membranes
membrane preparation, the membrane screen- can now be quickly synthesized and screened,
ing, and the optimization of the membrane pro- the implementation of appropriate experimental
cess parameters. The former relates to automated design is crucial to obtain membranes with opti-
devices developed for weighing powders, dis- mal performances. The aim is then to quickly
pensing liquids, homogenizing solutions, casting eliminate experimental areas in which limited suc-
these polymer solutions, dip coating, etc. More cesses are expected so combined experimentation
details can be found on www.html-membrane. can be focused on potentially promising areas. An
beand, www.porometer.com. efcient search strategy combined with HT tech-
HT devices for membrane screening are com- nology can rapidly lead to an optimum in a large
mercially available for pressure-driven liquid sep- parameter space, in contrast to the slower and less
aration systems with a variable number of effective traditional parameter-by-parameter
membranes and different membrane areas per approach. Two important aspects are the certainty
equipment (Vandezande et al. 2005; Cano-Odena of nding an optimum and a rapid progression of
et al. 2011; Van Doorslaer et al. 2010; www. convergence toward the optimal membrane
porometer.com 2014). Commercially available performance.
HT devices also exist for gas separations, which Several experimental design types were
were successfully applied to perform quality con- already applied in membrane technology.
trol tests at industrial-scale production and to Among them are articial neural networks
quickly analyze membrane performance in (ANNs) to predict membrane fouling and separa-
diverse studies (Khan et al. 2010, 2013; Basu tion performances in both liquid and gaseous
et al. 2011). The development of an HT membrane environments (Soleimani et al. 2013; Shokrian
bioreactor (MBR) was also reported to simulta- et al. 2010). Genetic algorithms (GAs) were
neously study the performance and fouling applied to obtain an optimum in membrane
High-Throughput Membrane Technology 941
composition, separation modules and operating Guria C, Bhattacharya PK, Gupta SK (2005) Multi-
conditions, gas separation systems, and desalina- objective optimization of reverse osmosis desalination
units using different adaptations of the non-dominated
tion units (Bulut et al. 2006; Vandezande sorting genetic algorithm (NSGA). Comput Chem Eng
et al. 2009; Yuen et al. 2000; Chang and Hou 29:19771995
2006; Guria et al. 2005). Genetic programming, Khan AL, Basu S, Cano-Odena A, Vankelecom IFJ
Plackett-Burman design, Taguchis method, (2010) Novel high throughput equipment for
membrane-based gas separations. J Membr Sci
response surface methodology, and fractional fac- 354:3239
torial design are other DoE types already Khan AL, Klaysom C, Gahlaut A, Khan AU, Vankelecom
employed in membrane technology. IFJ (2013) Mixed matrix membranes comprising
The combination of HT technology with of Matrimid and -SO3H functionalized
mesoporous MCM-41 for gas separation. J Membr
appropriate DoE clearly leads to the acceleration Sci 447:7379
of acquiring output, hence, to a better understand- Shokrian M, Sadrzadeh M, Mohammadi T (2010) C3H8
ing of the effects of parameters and the properties separation from CH4 and H2 using a synthesized
of new membrane materials and formulations. It PDMS membrane: Experimental and neural network
modeling. J Membr Sci 346:5970
can thus generate a signicant decrease in cost and
increase in efciency during the discovery and
Soleimani R, Shoushtari NA, Mirza B, Salahi A (2013)
Experimental investigation, modeling and optimization
H
optimization stages of membrane design and of membrane separation using articial neural network
development processes. and multi-objective optimization using genetic algo-
rithm. Chem Eng Res Des 91:883903
Van Doorslaer C, Glas D, Peeters A, Cano Odena A,
Vankelecom I, Binnemans K, Mertens P, De Vos
Cross-References D (2010) Product recovery from ionic liquids by
solvent-resistant nanoltration: application to ozona-
tion of acetals and methyl oleate. Green Chem
Design of Experiment (DOE) 12:17261733
Vandezande P, Gevers LEM, Paul JS, Vankelecom IFJ,
Jacobs PA (2005) High throughput screening for rapid
References development of membranes and membrane processes.
Vandezande P, Gevers LEM, Weyens N, Vankelecom IFJ
Basu S, Cano-Odena A, Vankelecom IFJ (2011) (2009) Compositional optimization of polyimide-
MOF-containing mixed-matrix membranes for CO2/ based SEPPI membranes using a genetic algorithm
CH4 and CO2/N2 binary gas mixture separations. Sep and high-throughput techniques. J Comb Chem
Purif Technol 81:3140 11:243251
Bilad MR, Declerck P, Piasecka A, Vanysacker L, Yan X, Vanysacker L, Declerck P, Vankelecom I (2013) Develop-
Vankelecom IFJ (2011) Development and validation of ment of a high throughput cross-ow ltration system
a high-throughput membrane bioreactor (HT-MBR). for detailed investigation of fouling processes. J Membr
J Membr Sci 379:146153 Sci 442:168176
Bulut M, Gevers LEM, Paul JS, Vankelecom IFJ, Jacobs Yuen CC, Aatmeeyata, Gupta SK, Ray AK (2000) Multi-
PA (2006) Directed development of high-performance objective optimization of membrane separation mod-
membranes via high-throughput and combinatorial ules using genetic algorithm. J Membr Sci
strategies. J Comb Chem 8:168173 176:177196
Cano-Odena A, Spilliers M, Dedroog T, De Grave K, Zapata P, Basak P, Carson Meredith J (2009)
Ramon J, Vankelecom IFJ (2011) Optimization of cel- High-throughput screening of ionic conductivity
lulose acetate nanoltration membranes for in polymer membranes. Electrochim Acta
micropollutant removal via genetic algorithms and 54:38993909
high throughput experimentation. J Membr Sci Zapata P, Mountz D, Meredith JC (2010) High-throughput
366:2532 characterization of novel PVDF/acrylic polyelectrolyte
Chang H, Hou W-C (2006) Optimization of membrane gas semi-interpenetrated network proton exchange mem-
separation systems using genetic algorithm. Chem Eng branes. Macromolecules 43:76257636
Sci 61:53555368 Zhou M, Liu H, Kilduff JE, Langer R, Anderson DG,
Gu M, Vegas AJ, Anderson DG, Langer RS, Kilduff JE, Belfort G (2009) High-throughput membrane surface
Belfort G (2013) Combinatorial synthesis with high modication to control NOM fouling. Environ Sci
throughput discovery of protein-resistant membrane Technol 43:38653871
surfaces. Biomaterials 34:61336138
942 Hollow Fiber
spinneret) with different congurations and

Hollow Fiber dimensions are used in spinning hollow bers
(Khayet and Matsuura 2011). The difference
Khayet Mohamed between capillary, tubular, and hollow ber is
Department of Applied Physics, Universidad simply a matter of dimensions. The approximate
Complutense de Madrid, Madrid, Spain diameter of a capillary membrane is between 0.5
and 5 mm, while that of a tubular membrane is
higher than 5 mm (Mulder 1992). Since hollow
In general a hollow ber is a cylindrical geometry bers are mechanically self-supporting, the ber
of any material having a hollow space with dimensions are very important. A hollow ber has
an external diameter approximately less an inner diameter (di) and outer diameter (do),
than 0.5 mm (Fig. 1). It may have different while the thickness (d) of the hollow ber wall is
cross-sectional shapes such as round, trilobal, the difference between the outer radius and the
pentagonal, octagonal, etc. Both smooth and inner radius. The primary process variables
corrugated-shaped hollow bers were fabricated establishing hollow ber diameters and thickness
from different micro-engineered or smart are spinneret geometry. Other spinning parame-
spinnerets (Fig. 2). Various types of spinnerets ters such as the bore liquid (uid) ow rate, the
(i.e., single or dual spinnerets, straight annular or dope extrusion rate, the take-up velocity and the
conical spinnerets, microstructured orice in air gap distance can modify the geometrical
Hollow Fiber, a Outer diameter

Fig. 1 Hollow ber (a) and (do): 500 m
cross-section of common c Polymer
spinnerets (b, c) used for
fabrication of hollow bers
b solution
Polymer
by (a) wet spinning and Thickness () solution
dry/wet spinning and (b) d =(do-di)/2
melt spinning and dry
spinning
Bore
liquid
Inner diameter (di)
Hollow Fiber, Fig. 2 Scanning electron microscopy and inner (b) microstructure (Reprinted from (Nijdam
(SEM) images of polyethersulfone/polyimide (PES/PI) et al. 2005). Copyright 2005, with kind permission from
blend hollow ber membranes with a corrugated outer (a) Elsevier)
Hollow Fiber 943
parameters of the spun hollow bers. and external coagulants and involves more con-
For example, Wu et al. (2007) found that the trolling parameters than the at sheet membrane.
thickness of the poly(vinylidene uoride), Shape and dimension of the spinneret, viscosity of
PVDF, hollow bers prepared by wet spinning the spinning dope (whether hollow bers can be
technique decreased with the increase of the inter- spun from the dope), temperature of the spinning
nal coagulant (bore liquid) ow rate. For example, dope, properties of the internal and external coag-
when the bore liquid ow rate was increased from ulants, ow rate of the bore liquid, dope extrusion
1.1 to 1.5 ml/min and further to 1.8 ml/min, the rate, length and humidity of the air gap, wind-up
corresponding thickness of the PVDF hollow speed, and ber take-up speed are some of those
bers decreased from 188 to 176 mm and further examples. As a consequence, the major parts of
to 155 mm. An enhancement of the inner diameter hollow bers that appeared in the literature with
and a reduction of the outer diameter of the hollow different structures and morphologies are based
bers with the increase of the bore liquid ow rate mostly on trial-and-error experiments.
were detected. Hollow bers are mechanically self-
The process of fabricating synthetic hollow supporting, have good exibility, and are easy to
ber membranes is more difcult than that of at assemble in module and to handle for different H
sheet membranes. In general, the process is car- applications. The feed is supplied to either the
ried out by extruding or pumping a dope solution inside or outside of the ber, and the permeate
through a spinneret (Fig. 1) following different passes through the ber wall to the other side of
methods such as melt spinning, dry spinning, the ber. The ber wall may have an asymmetric
wet spinning, or dry/wet spinning. In melt spin- structure being the active skin layer placed to the
ning and dry spinning, the dimensions of the feed side. A bundle of hollow bers are often
spinneret are not so crucial because the ber mounted in a shell and tube module, which is
dimensions are mainly determined by the ratio of featured by a very large packing density (i.e.,
the extrusion rate and tearing rate (Khayet and high surface area per unit module volume), and
Matsuura 2011). In wet spinning and wet/dry therefore the size of the hollow ber module is
spinning, the spinneret has a tube-in-orice struc- smaller than other modules for a given perfor-
ture with typical dimensions 0.51 mm inner mance capacity. The packing capacity of a hollow
diameter and 0.92 mm outer diameter. ber module may reach 5009,000 m2/m3
The rst hollow ber membranes were pat- resulting in a high productivity per unit volume.
ented by Mahon in late the 1960s (Mahon Spinning corrugated hollow bers is similar to
1966a, b). Various types of hollow ber mem- spinning smooth cylindrical bers; the only dif-
branes, hydrophobic or hydrophilic, dense or ference is the use of a modied spinneret (Cao
porous, single or double layered, have been pro- et al. 2004; ulfaz et al. 2010; Nijdam et al. 2005;
posed for different membrane separation pro- Wang et al. 2004). For example, to prepare a
cesses leading to a fast growth of synthetic hollow ber with a corrugated inner microstruc-
membrane technology. Since then signicant ture instead of a cylindrical needle, the spinneret is
efforts have been made to develop new hollow fabricated with a structured needle. To spin a
bers with desirable structures and morphologies. hollow ber with a corrugated outer surface, a
However, until now, various scientists admit that microstructured orice is incorporated to the spin-
understanding the mechanisms of hollow ber neret. Figure 2 shows SEM images of corrugated
formation is rather qualitative than quantitative. hollow ber membranes.
It is not easy to control the characteristics of The use of corrugated membrane surfaces is to
hollow bers because various important spinning enhance the heat and mass transfer by increasing
parameters are involved simultaneously. For the membrane surface area and the turbulence
example, the preparation of the hollow ber mem- near the membrane surface. Corrugations act as
brane by the dry/wet spinning or wet spinning turbulence promoters but also may lead to an
techniques requires both internal (bore liquid) increase of pressure drop in the ow channel. It
944 Hollow Fiber
Hollow Fiber, Fig. 3 Cross-sectional structure of PVDF 40 wt.% NMP aqueous solution as internal coagulant
hollow ber membranes prepared with wet spinning tech- (Reprinted from (Bonyadi et al. 2007). Copyright 2007,
nique and dry/wet spinning technique using 20 wt.% with kind permission from Elsevier)
PVDF in the solvent N-methylpyrrolidone (NMP) and
was observed that the size of the corrugations on their external or internal contours. These corruga-
the ber was very sensitive to the air gap length tions can be eliminated increasing the air gap
and even disappeared from the external surface for distance (Fig. 3). The number of corrugations in
high air gap lengths. When used in gas separation, the internal surface of the hollow ber membranes
the selectivity (oxygen/nitrogen) and permeance is reduced approaching a nal circular shape. The
of polyethersulfone/polyimide (PES/PI) blend hydrodynamic instability, which leads to
corrugated hollow ber membrane were found to nonuniform cross-section and wavy geometry,
be similar to those of smooth cylindrical bers might be due to the pressure induced in the
(i.e., circular-shaped bers) prepared under the nascent ber (as a result of diffusion/convection,
same operating conditions. However, the gas per- precipitation, densication, and shrinkage), which
meate ow was increased by 19 %, which is the will buckle the rigid elastic shell formed at the
increase of membrane surface area due to corru- interface between the bore uid and the spinning
gations (Nijdam et al. 2005). It is worth quoting dope solution (Bonyadi et al. 2007).
that an increase of membrane surface area of 89 %
can be obtained with a corrugated hollow ber
membrane (ulfaz et al. 2010). Furthermore, References
PES/PVP blend hollow ber membranes were
used in UF, and it was found that the water per- Bonyadi S, Chung TS, Krantz WB (2007) Investigation of
meability, molecular weight cutoff, and pore size corrugation phenomenon in the inner contour of hollow
bers during the non-solvent induced phase-separation
distribution of both the corrugated and circular- process. J Membr Sci 299:200210
shaped bers were similar (ulfaz et al. 2010). Cao C, Chung TS, Chen SB, Dong ZJ (2004) The study of
This indicates a higher permeate ow of the cor- elongation and shear rates in spinning process and its effect
rugated hollow ber compared to the circular- on gas separation performance of poly(ether sulfone) (PES)
hollow ber membranes. Chem Eng Sci 59:10531062
shaped ber and a similar UF separation factor. ulfaz PZ, Rolevink E, van Rijn C, Lammertink RGH,
Instabilities and irregularities may occur dur- Wessling M (2010) Microstructured hollow bers for
ing hollow ber spinning inducing corrugations at ultraltration. J Membr Sci 347:3241
Hollow Fiber Carbon Molecular Sieve Membranes 945

ciples and applications. Elsevier, Amsterdam Hollow Fiber Carbon Molecular Sieve
Mahon HI (1966a) Permeability separatory apparatus and
membrane element, method of making the same and Membranes
process utilizing the same. US Patent 3,228,876
Mahon HI (1966b) Permeability separatory apparatus and Xuezhong He
process using hollow bers. US Patent 3,228,877 Department of Chemical Engineering, Norwegian
ogy. Kluwer, Dordrecht University of Science and Technology,
Nijdam W, de Jong J, van Rijn CJM, Visser T, Versteeg L, Trondheim, Sr-Trndelag, Norway
Kapantaidakis G, Koops GH, Wessling M (2005) High
performance micro-engineered hollow ber mem-
branes by smart spinneret design. J Membr Sci
256:209215 Carbon molecular sieve (CMS) membranes have
Wang KY, Matsuura T, Chung TS, Guo WF (2004) The been studied in more than twenty years as prom-
effects of ow angle and shear rate within the spinneret ising and energy-efcient membranes for gas sep-
on the separation performance of poly(ethersulfone) aration. Due to their high permeability, selectivity,
(PES) ultraltration hollow ber membranes.
and also thermal and chemical stability under
Wu B, Li K, Teo WK (2007) Preparation and characteriza- adverse and harsh conditions, the CMS mem- H
tion of poly(vinylidene uoride) hollow ber membranes are becoming increasingly important for
branes for vacuum membrane distillation. J Appl separating gas mixtures with quite similar molec-
Polym Sci 106:14821495
ular kinetic diameter. Carbon molecular sieve
Hollow Fiber Carbon Molecular Sieve Membranes, Fig. 1 SEM images of cross section of precursor and HFCM-2.
(a, c) are the cross section of the precursor; (b, d) are the cross section of the prepared carbon membrane (He et al. 2011)
946 Hollow Fiber Carbon Molecular Sieve Membranes
Hollow Fiber Carbon Molecular Sieve Membranes, Table 1 Structural parameters of the cellulosic-based carbon
membranes (He and Hgg 2012)
W0 Average
Carbon True density, rs Bulk density, rb (cm3 E0 micropore width
membranes (g cm3)a (g cm3) g1) (kJ mol1) () Reference
HFCM-1 1.53 1.24 0.15 32.2 5.2 This work
HFCM-2 1.38 1.12 0.17 30.8 5.6 This work
CMSM1 1.6 1.1 0.28 31.6 5.5 (Lagorsse
et al. 2004)
CM-V823 0.16 22.75 (Lua and Su
2006)
a
Density of carbons is 1.31.8 g/cm3 as compared to 2.2 g/cm3 for graphite (Koresh and Soffer 1983)
membranes can be prepared by controlling the et al. (He and Hgg 2012), and the results were
carbonization procedure during carbonization of shown in Table 1, which will be helpful to under-
different polymeric materials such as polyimide, stand the transport mechanism through carbon
polyacrylonitrile (PAN), phenolic resin, poly membranes.
(phthalazinone ether sulfone ketone), poly The operating parameters such as feed pres-
(phenylene oxide), cellophane paper, cellulose, sure, temperature, and ow rate of retentate will
and deacetylated cellulose acetate. The carboniza- greatly affect the CMS membrane separation per-
tion parameters (e.g., purge gas, purge gas ow formances and need to be optimized for a specic
rate, nal temperature, heating rate, and nal soak application as reported by He et al. (He and Hgg
time) will affect the resulting carbon membrane 2011b). Considering the aging problem for most
performance, and the optimal condition was carbon membranes, different regeneration tech-
obtained on the basis of experimental design and niques can be used to recover the performances
statistical analysis to prepare the high- online or off-line (Lie 2005; Hagg and He 2011).
performance carbon membranes for a specic In order to ensure the efciency of the carbon
application as reported by He et al. (He and membrane separation process, a hollow ber
Hgg 2011a). Different techniques such as scan- membrane module should be designed and
ning electron microscopy (SEM), Fourier trans- constructed. Figure 2 shows a schematic and
form infrared spectroscopy (FTIR), and X-ray lab-scale hollow ber carbon membrane module
diffraction (XRD) have been used to characterize (He and Hgg 2011b), while Fig. 3 shows a small
the membranes structure and morphology. pilot-scale module (designed by one of
Figure 1 shows the cross section and inside sur- NanoGloWa project partner, HyGear B.V.)
face of the precursors and HFCMs, respectively (He 2011). Single-gas permeability tests at differ-
(He et al. 2011). The outer diameter and wall ent temperatures and pressures were mostly
thickness of the HFCMs were approximately conducted by using the lab-scale module, while
250 mm and 30 mm, respectively, indicating a the small pilot-scale module was employed to test
signicant radical shrinkage comparing to the the gas mixture at a higher permeate ux in order
precursor membranes with 400 mm and 50 mm. to determine the gas composition more precisely.
The pore size and pore-size distribution (PSD) Based on the experimental investigations and
can be estimated by gas gravimetric adsorption process simulations, the carbon membranes show
measurements equipped with a Rubotherm mag- high potential for selected industrial applications
netic suspension balance (MSB) (Belmabkhout such as CO2-CH4 separation for biogas upgrading
et al. 2009). The structural parameters of a (Hgg and Lie 2006), H2-CH4 separation wher-
cellulosic-based carbon membrane have been esti- ever relevant (Grainger and Hgg 2008), and CO2
mated from CO2 sorption measurements, and a capture from ue gases (He and Hgg 2011b; He
narrow pore-size distribution was found by He et al. 2009). Moreover, carbon molecular sieve
Hollow Fiber Carbon Molecular Sieve Membranes 947
Retentate Feed
Hollow fibre
carbon membranes
Epoxy Epoxy
Effective length Permeate
Feed
Retentate
Permeate
Hollow Fiber Carbon Molecular Sieve Membranes, Fig. 2 A schematic and lab-scale carbon membrane module
(He and Hgg 2011b)
Hollow Fiber Carbon

Molecular Sieve
Membranes, Fig. 3 A
small pilot-scale carbon
membrane module
(Designed by HyGear)
(He 2011)
membranes also show a great potential in air sep- in natural gas networks. Int J Hydrog Energy
aration, petrochemical, and high-temperature 33(9):23792388
Hagg M-B, He X (2011) Carbon molecular sieve mem-
membrane reactor applications. branes for gas separation. In: Drioli E, Barbieri G (eds)
Membrane engineering for the treatment of gases: Gas-
separation Problems Combined with Membrane Reac-
tors, The Royal Society of Chemistry, Cambridge, Vol.
References 2, pp 162191
Hgg M-B, Lie JA (2006) Upgrading of biogas by mem-
Belmabkhout Y, Serna-Guerrero R, Sayari A (2009) branes for vehicle fuel or injection into the European
Adsorption of CO2 from dry gases on MCM-41 silica Gas Network. ECI Conferences, Tomar
at ambient temperature and high pressure. 1: pure CO2 He X (2011) Development of Hollow Fiber Carbon
adsorption. Chem Eng Sci 64(17):37213728 Membranes for CO2 Separation [PhD]. Norwegian
Grainger D, Hgg MB (2008) The recovery by carbon University of Science and Technology, Trondheim,
molecular sieve membranes of hydrogen transmitted p 42
948 Hollow Fiber Enzymatic Reactor, Modeling of
He X, Hgg M-B (2011a) Optimization of carbonization should be considered: ow in lumen, within mem-
process for preparation of high performance hollow brane matrix, and in the extracapillary space
ber carbon membranes. Ind Eng Chem Res
50(13):80658072 (Nagy 2011, p. 179). Let us consider a steady-
He X, Hgg M-B (2011b) Hollow ber carbon membranes: state, laminar, incompressible, isothermal, fully
investigations for CO2 capture. J Membr Sci developed ow in vertically oriented tube; thus
378(1-2):19 the Navier-Stokes equations (Bird et al. 1960) will
He X, Hgg M-B (2012) Structural, kinetic and perfor-
mance characterization of hollow ber carbon mem- be as the continuity equation:
He X, Lie JA, Sheridan E, Hagg M-B (2009) CO2 capture 1 @ @u
by hollow bre carbon membranes: experiments and rur 0 (1)
r @r @z
process simulations. Energy Procedia 1(1):261268
He X, Lie JA, Sheridan E, Hgg M-B (2011) Preparation
and characterization of hollow ber carbon membranes where u is the axial convective velocity, m/s; ur is
from cellulose acetate precursors. Ind Eng Chem Res the transverse convective velocity, m/s; and z is
50(4):20802087 the axial space coordinate, m. The momentum
Koresh JE, Soffer A (1983) Molecular sieve carbon mem- equations are as
brane part I: presentation of a new device for gas
mixture separation. Sep Sci Technol 18:723734
Lagorsse S, Magalhes FD, Mendes A (2004) Carbon @u @u @p
r u u
molecular sieve membranes: sorption, kinetic and @r @z @z
structural characterization. J Membr Sci
241(2):275287 @ 1 @ ru @2u
m 2
Lie JA (2005) Synthesis, performance and regeneration of @r r @r @z
carbon membranes for biogas upgrading-a future
energy carrier [PhD]. Norwegian University of Science (2)
and technology, Trondheim
Lua AC, Su J (2006) Effects of carbonisation on pore @u @u @p
evolution and gas permeation properties of carbon r u u
@r @z @z
membranes from Kapton?polyimide. Carbon

44(14):29642972 @ 1 @ ru @2u
m 2
@r r @r @z
(3)
Hollow Fiber Enzymatic Reactor, and for solute concentration in case of constant
Modeling of density
Endre Nagy
@c @c @ @c 1 @ Dc
Research Institute of Chemical and Process u u D
Engineering, University of Pannonia, Veszprm, @z @r @r @r r @r

Hungary @ @c
D (4)
@z @z
Modeling of hollow ber enzymatic reactor. where r is the density, p is the pressure, m is
The complete description of the transport of uid dynamic viscosity, and g is gravitational
in hollow ber capillary membrane requires that acceleration.
the local transport equations that describe the ow Equations 2 and 3 may be reduced by consider-
and transport conditions should be solved. The ing the fact that for a small wall Reynolds number,
starting Navier-Stokes ow models (Bird the inertial terms can be negligible. In most viscous
et al. 1960) are very complex; signicant simpli- ows, normal stress effects, @ 2 u=@z2, are negligible
cation is needed to obtain applicable model for when the ratio, ro/L (L is length of capillary, ro is
engineering points of view. In a capillary mem- lumen radius), is less than 102, a condition that is
brane with permeable wall, three regions of ow satised in almost all hollow ber membrane
Hollow Fiber Liquid Membrane 949
devices. Depending on the transmembrane convec- From Eq. 11 the wall velocity, uw, can be given as
tive velocity, the axial velocity often can be much
larger than that of the permeation rate. Due to its r3o @ 2 pt
large values, the axial diffusion term can be uw (12)
16m @z2
neglected. The radial convective velocity is rather
low; accordingly, momentum equation of the trans- The radial convective velocity through the mem-
verse velocity can be neglected. Accordingly, Eqs. 2 brane can also be expressed by the next equation
to 4 can be simplied as Biocatalytic membrane reactor: mass transport
as
@p 1 @ @u
0 m r (5)
@z r @r @r k r o pt z pe z
uM r, z (13)
em r ln1 d=ro
The component balance will be as:
From equality of Eqs. 12 and 13, a second-order
@ci @ @c differential equation can be obtained neglecting
u
@z
D r
@r r
(6) H
the radial pressure gradient (Nagy 2011, p. 187).
Solving this equation the p(z) and uw(z) functions
When Eq. 6 is integrated twice with respect to r, can be obtained (Nagy 2011). This convection
one can get as velocity can then be replaced in the mass balance
equation for the membrane (see Biocatalytic
dp r2 Membrane Reactor, Mass Transport in).
u Alnr B (7)
dz 4
For the hydrodynamic analysis in the lumen and

in the shell, the continuity Eq. 1 and the momen- References
tum Eq. 5 have to be solved with suitable bound-
Bird RB, Stewart E, Lightfoot EN (1960) Transport phe-
ary conditions. In the lumen (t) these are nomena. Wiley, New York
Nagy E (2011) Basic equations of the mass transport
@u through a membrane layer. Elsevier, Amsterdam
at r 0 then 0, u 0; (8)
@z
at r r0 then u 0 (9)
Hollow Fiber Liquid Membrane
Consequently it can be get
Vladimir S. Kislik
2 ! Campus Givat Ram, Casali Institute of Applied
r2o r dpt
u 1 (10) Chemistry, The Hebrew University of Jerusalem,
4m ro dz Jerusalem, Israel
Replacing Eq. 10 into Eq. 5, one can get for the

radial convective velocity in the lumen side (for Hollow ber (HF) supported liquid membrane is a
distinction for the membranes velocity, f sub- three-phase liquid membrane system in which the
script denotes the uid phase here in the lumen membrane phase (liquid) is held by capillary
side) forces in the pores of microporous polymeric or
inorganic hollow bers. The design of the hollow
( )
r3o r 1 r 3 d2 pt ber SLM module consists of a certain number of
ut (11) thin bers, placed along the length of the shell.
m8 ro 2 ro dz2
The feed phase is pumped through the bers; the
950 Hollow Fiber Liquid Membrane
b Sweep out
a
F F F
S S S
S Feed Feed
out in
F F F
Pressure
source
S S
LM -
reservoir
Sweep in
Hollow Fiber Liquid Membrane, Fig. 1 Hollow ber contained liquid membrane (HFCLM): (a) conguration of
permeator shell and (b) side view of the permeator
Hollow Fiber Liquid RO R MO

M
Membrane, F FO
Fig. 2 Scheme of the
hollow ber contactor with
parallel ow of phases
distributed U-shaped
bundles of bers with 1
separated inlet and outlet
end chambers. [F feed
(donor) phase, M liquid 3
membrane phase,
A A'
R stripping solution,
1 hollow ber for the feed, 2 Cross-section AA':
2 glass tubulet, 3 glass R
1
tempering jacket of the
contactor] (From Schlosser
M
and Sabolova 2002)
2 F
3
pores in the bers are lled with the LM solution, of the HFLM transport was developed by Sirkar
and the receiving (strip) phase is forced out et al. (see in Sirkar 1996) and Schlosser et al. (see
through the sides of the shell. HF SLMs are sche- in Schlosser and Sabolova 2002).
matically presented in the Figs. 1, 2, and 3. Com- Different hollow ber modules can be classi-
pact and modular hollow ber devices can be used ed into several groups. One, termed as a hollow
with exceptionally high mass transfer area per unit ber contained liquid membrane (HFCLM)
equipment volume. Very often such HF systems (Majumdar and Sirkar 1992) conguration, is
are termed as contactor devices. Contactor shown in Fig. 1a. The process is achieved by
devices are used as construction units in many of packing thousands of microporous hollow bers
the SLM, HFCLM, and BLM systems. The theory in a permeator shell lled with stationary liquid
Hollow Fiber Liquid Membrane 951
Hollow Fiber Liquid R RO

Membrane,
hollow ber contactor of 6
three phases with crossow
FO F
of LM phase: (1) body of
A 1
element, (2, 3) hollow ber B B'
in downstream and SO 4
upstream part, (4, 5) inlet 7
and outlet chamber, (6,7) 2
inlet and outlet tube, (8)
5
owing head, (9) central
bafe; F feed, E LM phase, 9
R strip solution (From
Schlosser 2000) 3
8
H
E 2
5
Section AA:
A' 3
Section BB:
1 3 9 2
membrane solution. The bers are present in feed module with ber in ber units and crossow of
and strip sets very close to each other but with the one phase is described in Ref. Schlosser and
ends of each set being separated. If the bers are Sabolova (2002).
hydrophilic, the aqueous feed and strip ow A capillary membrane module was developed
through the lumen side of the bers and ll the (Teramoto et al. 2003) for gas separation. The
pores of the ber; if bers are hydrophobic, they concept of a capillary membrane apparatus is
are wetted by organic LM phase (Fig. 1b). So, shown in Fig. 4. Both a feed gas and a LM solu-
reader can see that this is a hybrid system with tion are supplied to the lumen side (high-pressure
combination of modied SLM and BLM feed side) of the capillary ultraltration membrane
congurations. module and ow upward. The LM solution, which
Hollow bers in tube pertractor (Schlosser contains dissolved solute gas (CO2), ows to the
2000) is another type of HF module (see Fig. 2) permeate side (low-pressure shell side), where the
with parallel ow consisting of one or two solution liberates dissolved gas to become a lean
intermixed U-shaped bundles of microporous solution. The lean solution is returned to the
polypropylene bers, inserted into polysulfon lumen of the capillary module by a pump. In the
(or glass) tubes. Pulsation of the LM phase shell side, dissolved gas is stripped from the liquid
along the bers is used for better mixing. owing down on the outer surface of the capillary
One more type of the HF modules is presented and discharged from a vacuum pump through a
schematically in Fig. 3. A crossow with pulsa- liquid reservoir.
tion perpendicularly to bers was used here. Some One more example of hollow ber LM separa-
plate and frame HF modules were proposed. The tion system is described as afnity dialysis (Devis
952 Hollow Fiber Liquid Membrane
a b
exhaust gas exhaust gas
capillary
membrane
to pressure 30mm
gauge permeable
stripped gas section capillary
(165mm) membrane
to vacuum
liquid
pump
permeation capillary
enriched capillary membrane
impermeable
CO2 membrane coated with
section
opoxy (265mm) adhesive
adhesive to vacuum
capilarty mem exhaust gas pump distributor
vacuum module
absorbent feed gas absorbent feed gas
pump
CO2/N2 tee tee
carrier feed gas
solution (a) Module A (b) Module B
Hollow Fiber Liquid Membrane, Fig. 4 Schematic dia- dioxide: (a) experimental capillary membrane apparatus
grams of facilitated transport of gas using capillary mem- and (b) capillary membrane modules with permeation of
brane module for removal and enrichment of carbon carrier solution (From Teramoto et al. 2003)
et al. 1988) Water-soluble polymer (WSP) solu- production of new or better foodstuffs, novel
tions as LM were immobilized into the pores of pharmaceutical products with well-dened enan-
hollow ber units (bers of 5,000 molecular tiomeric compositions, conversion of lactose into
weight cutoff from Spectra/Por were used). Selec- glucose and galactose, the isomerization of glu-
tive separation and concentration of both cations cose into fructose, and the separation of chiral
(Ca/Na, Cu/Zn) and anions (potassium chromate/ isomers of acylated racemic amino acids and
sodium chloride) were tested. 2.5 % w/v aqueous wastewater treatment.
solution of poly (2-acrylamido-2-methyl-1- The regeneration of degraded hollow ber liq-
propanesulfonic acid) (poly-AMPS) for cation uid membranes could be done by pumping the
separations and 5 % w/v solution of poly fresh LM at the lumen side of the support for a
(ethylenimine) (PEI) for anion separations were few minutes instead of the aqueous solution. Con-
used as LM. High concentration factor (5075) tinuous impregnation of the membrane pores is
and selectivity (3040) were obtained. possible as it is in the HFCLM devices or by
Hollow ber membrane catalytic reactors adding membrane phase as an emulsion to the
(HFMCR) and bioreactors (HFMBR) are rela- strip aqueous phases. It is termed the
tively new developments of SLM techniques. pseudoemulsion-based hollow ber strip disper-
The SLM reactors can be used to shift the equi- sion (PEHFSD) technique. The main drawback is
librium conversion (e.g., esterication reaction), pollution of the product with the membrane
to remove selectively products and by-products liquid.
from the reaction mixture, and to supply selec-
tively the reagents (e.g., oxygen for partial oxida-
tion reactions) (Li et al. 2003). Application of References
SLM reactors appears of particular interest in
hydrogen production and oxidation reactions. Cussler EL, Ding H (1995) Bioseparations, especially
Applications of biocatalytic membrane reactors using hollow bers. In: Singh RK, Rizvi SSH (eds)
Bioseparation processes in foods. Marcel Dekker,
(Cussler and Ding 1995) provide an ideal support New York, pp 115
for catalyst immobilization due to their wide Devis JC, Valus RJ, Lawrence EG (1988) Afnity dialysis-
available surface area per unit volume and include a method of continuous, rapid metal ion separation
Hollow Fiber Membrane Bioreactor for Cell Growth 953
using dialysis membranes and selective water-soluble are separated by the membranes. Hollow ber
polymers as extractants. Sep Sci Technol membranes of suitable MWCO separate cells
23(1011):10391066
Li N, Giorno L, Drioli E (2003) Effect of immobilization from the medium compartment allowing the com-
site and membrane materials on multiphasic munication between the compartments through
enantiocatalytic enzyme membrane reactors. Ann NY the pores of ber wall. Depending on the pore
Acad Sci 984:436452 sizes, the passage of nutrients, metabolites, and
Majumdar S, Sirkar KK (1992) Hollow-ber contained
liquid membrane. In: Ho WSW, Sirkar KK (eds) Mem- catabolites occurs and cells are retained. The prin-
brane handbook. Van Nostrand Reinhold, New York, ciple of a hollow ber bioreactor based on
pp 764808 hemodialyzer modules was rst described by
Schlosser S (2000) Pertraction through liquid and poly- Knazek (Knazek et al. 1972). Usually cells are
meric membranes. In: Bela-Bako K, Gubicza L,
Mulder M, (eds) Integration of membrane processes cultured in suspension or attached to the outer
into bioconversions. In: Proceedings of the 16th Euro- surface of the membranes in the extracapillary
pean Membrane Society Annual Summer School, space. Culture medium containing nutrients
Veszprem, Aug 1999. Kluwer/Plenum, New York, ows in the lumen of the bers and diffuses out
pp 73100
through the pores of the ber wall to the cells. The
Schlosser S, Sabolova E (2002) Three-phase contactor
with distributed U-shaped bundles of hollow-bers metabolic species together with waste molecules H
for pertraction. J Membr Sci 210:331347 produced by cells permeates back into the bers
Sirkar KK (1996) Hollow ber-contained liquid mem- and is removed from cell compartment (Fig. 1).
branes for separations: an overview. In: Bartsch RA,
Way JD. (eds) Chemical separations with liquid mem- As bioreactors, they are extensively used for
branes. American Chemical Society, Symposium production of biochemicals and for the develop-
Series 642, pp 222238 ment of bioarticial organs. Recently they found
Teramoto M, Ohnishi N, Takeuchi N, Kitada S, applications in tissue engineering applications for
Matsuyama H, Mano H (2003) Separation and enrich-
ment of carbon dioxide by capillary membrane module the development of tissue constructs, cell therapy,
with permeation of carrier solution. Sep Purif Technol and the expansion and transdifferentiation of stem
30:215227 cells. Hollow ber bioreactor offers several
advantages for the culture and growth of adherent
and suspension cells including: (i) an in vivo-like
microenvironment regarding nutrient supply and
Hollow Fiber Membrane metabolic waste removal as each ber replicates a
blood capillary, (ii) high surface area to volume
Hollow Fiber Membrane Module ratio, (iii) cells can be grown to high density,
(iv) cells are protected from shear stress, and
(v) support of three-dimensional growth of
cells for the development of tissue-engineered
Hollow Fiber Membrane Bioreactor constructs. Different types of polymers (poly-
for Cell Growth ethersulfone, polysulfone, polyacrylonitrile, mod-
ied polyetheretherketone, polyamide, cellulose
Loredana De Bartolo acetate, etc.), generally hydrophilic polymers or
Institute on Membrane Technology, National with moderate wettability, are used for the devel-
Research Council of Italy, ITM-CNR, Rende, opment of hollow ber bioreactors in order to
Italy favor the adhesion of cells or the exchange
of medium. Hydrophobic hollow bers (e.g.,
polypropylene) are included in some devices
Hollow ber membrane bioreactor consists of for a separate oxygenation of cells (Flendrig
hollow ber membranes parallel-assembled in an et al. 1997). A major drawback of the hollow
external shell, which is usually in the shape of a ber bioreactors is represented by the mass trans-
cylinder. The bers are potted in the shell, thus fer resistance due to the membrane because while
creating a medium and a cell compartment which it separates cells from the medium compartment,
954 Hollow Fiber Membrane Bioreactor for Cell Growth
a b
Hollow fiber Culture medium

membranes flow
Culture
cells medium flow Hollow fiber
membranes cells
Hollow Fiber Membrane Bioreactor for Cell Growth, ows in the lumen of bers; (b) cells are cultured in the
Fig. 1 Schematic diagram of hollow ber membrane bio- lumen of ber, and the medium ows in the extracapillary
reactor used for cell culture and growth: (a) cells are space
cultured in the extracapillary space, and the medium
it does reduce diffusion of nutrients and oxygen to have been developed for the culture of cells that
the cells. In addition, fouling due to the adsorption grow in adhesion and/or in suspension. Different
of plasma proteins or the development of biolms congurations have been investigated for the cul-
further increases the resistance to mass transport. ture and growth of primary cells (e.g., lympho-
In the membrane bioreactors, mass transfer is cytes, hepatocytes, pancreatic islets, neurons,
determined by the MWCO or pore diameter of osteoblasts, broblasts), cell lines, progenitor
membrane and occurs by diffusion and/or convec- cells, and stem cells (De Bartolo and Bader
tion in response to existing transmembrane con- 2013). Usually cells are cultured in adhesion or
centration or pressure gradients. Most of the suspension outside the bers, and the medium
bioreactors use membranes with MWCO ranging ows in the lumen of bers (De Bartolo
from 70 to 100 kDa that allow the transport of et al. 2007a). In other bioreactors cells are cultured
small molecules that are fed or produced by cells inside the lumen of bers, and the medium ows
but exclude proteins with high MW such as in the shell of ber (Fig. 1) (Nyberg et al. 1994).
immunoglobulins and cells. These membranes More complex devices have been explored for
provide immunoprotection, which is important the culture of hepatocytes, stem cells, and
when the devices are used in the organ replace- progenitor cells that include the use of different
ment therapy in order to avoid the passage of bers. A bioreactor with four interwoven
immunocompetent species present in the patients independent capillary membrane systems that
blood. Other bioreactors use microporous mem- serve different functions has been developed by
branes with a large pore diameter (0.2 mm) that Gerlach et al. (1994). The cells are cultured on
permit the free passage of proteins, toxins, and the outer surface and among the capillaries.
clotting factors, but they exclude the passage of Each ber type exhibits a different function:
cells. The advantage of using a membrane with silastic membranes for oxygen supply and
large pore diameter is that it increases the uid removal of carbon dioxide, polyamide ber for
convection improving the mass transfer condi- the plasma inow, polyethersulfone ber for the
tions. Several hollow ber membrane bioreactors plasma outow, and hydrophilic polypropylene
Hollow Fiber Membrane Module 955
membranes for the sinusoidal endothelial References

coculture. A crossed hollow ber membrane bio-
reactor was developed as an in vitro liver tissue De Bartolo L, Bader A (2013) Biomaterials for stem cell
therapy: state of art and vision for the future. CRC
model to study disease, drugs, and therapeutic
Press, Boca Raton
molecules alternative to animal experimentation De Bartolo L, Piscioneri A, Cotroneo G, Salerno S,
(De Bartolo et al. 2009). The bioreactor consists Tasselli F, Campana C, Morelli S, Rende M, Caroleo
of two types of bers with different MWCO and MC, Bossio M, Drioli E (2007a) Human lymphocyte
PEEK-WC hollow ber membrane bioreactor.
physicochemical properties cross-assembled in
J Biotechnol 132:6575
alternating manner: modied polyether- De Bartolo L, Morelli S, Rende M, Campana C, Salerno S,
etherketone and polyethersulfone used for the Quintiero N, Drioli E (2007b) Human hepatocyte mor-
medium inow and outow, respectively. The phology and functions in a multibore ber bioreactor.
Macromol Biosci 7:671680
combination of these two ber sets produces an
De Bartolo L, Salerno S, Curcio E, Piscioneri A, Rende M,
extracapillary network for the adhesion of cells Morelli S, Tasselli F, Bader A, Drioli E (2009) Human
and a high mass exchange through the crossow hepatocyte functions in a crossed hollow ber mem-
of culture medium. This bioreactor has been brane bioreactor. Biomaterials 30:25312543
developed for the culture of human hepatocytes Flendrig LM, La Soe JW, Joerning GGA et al (1997) In H
vitro evaluation of a novel bioreactor based on an
and expansion and differentiation of liver progen- integral oxygenator and a spirally wound nonwoven
itor cells. Multibore ber membrane bioreactor polyestermatrix for hepatocyte culture as small aggre-
consisting of modied polyethersulfone ber has gates. J Hepatol 26:13791392
Gerlach JC, Schnoy N, Smith MD, Neuhaus P (1994)
been developed for the long-term maintenance of
Hepatocyte culture between woven capillary networks:
human hepatocytes. Each ber is composed of a microscopy study. Artif Organs 18:226230
seven compartments that communicate through a Knazek RA, Gullino PM, Kohler PO, Dedrick RL
porous foamy support structure that is located in (1972) Cell culture on articial capillaries: an approach
to tissue growth in vitro. Science 178:6566
between the capillaries. Cells have been cultured
Nyberg SL, Remmel RP, Mann HJ et al (1994) Primary
inside the lumen of the multibore bers and lumen outperform hep G2 cells as source of biotransformation
in the shell compartment (De Bartolo functions in a bioarticial liver. Ann Surg 220:5967
et al. 2007b). Salerno S, Piscioneri A, Morelli S, Al-Fageeh MB,
Drioli E, De Bartolo L (2013) Membrane bioreactor
Hollow ber membrane bioreactors have a
for expansion and differentiation of embryonic liver
great potential for the expansion and differentia- cells. Ind Eng Chem Res 52(31):1038710395
tion of embryonic and adult stem cells, which
have ability to either self-renew or differentiate
into multiple cell lineages. These properties make
stem cells attractive as a cell source for cell ther- Hollow Fiber Membrane Module
apies, tissue engineering, and model systems for
drug screening. Depending on the application, Joerg Balster
hollow ber membrane bioreactors can be used Application Technology, Evonik Fibres GmbH,
for the production of expanded stem cells with Schoering, Austria
uniform properties and/or for a controlled and
reproducible differentiation into selected mature
cell types (Salerno et al. 2013). They are particu- Synonyms
larly advantageous with respect to the other cul-
ture systems because of their capacity to ensure a Hollow ber membrane; Hollow ber module
more homogeneous environment, to monitor key
culture parameters (pH, temperature, oxygen, The hollow ber membrane module as well as a
etc.), and to guide the transdifferentiation of cells capillary membrane module assembles as shell-
by engineering membranes with targeted and-tube heat exchanger. It consists of a large
properties. number of hollow bers assembled in a module,
956 Hollow Fiber Membrane Module
Hollow Fiber Membrane

Module, Fig. 1 Schematic Inside Feed
of a hollow ber module
with the feed on the inside
of the ber (bore side feed) Retentate
Feed
and the feed on the outside
of the bers (shell side feed)
(Taken from Scholz et al.,
2011)
Permeate
Outside Feed Feed
Permeate
Permeate
Retentate
as shown schematically in Fig. 1 (Scholz Hollow Fiber Membrane Module, Table 1 Properties
et al. 2011; Mulder 1997). of hollow fiber modules (Taken from Scholz et al. 2011)
The free ends of the bers are potted with Hollow ber module
agents such as epoxy resins, polyurethanes, sili- Structure Self-supporting
cone rubber, thermoplastics, thermosets, or inor- Active layer Inner/outer diameter
ganic cements. In some cases, they can also be Feed Lumen/shell side
fused by heating. The difference between hollow Inner diameter 40250 mm
Outer diameter 80400 mm
ber modules and capillary modules is a matter of
Packing density <10,000 m2/m3
dimensions while the module concepts are the
Allowed 100 bar shell side; 15 bar lumen side
same (Scholz et al. 2011; Mulder 1997). pressure
Hollow bers consist of a porous, nonselective Applications GP, RO, DL
support layer (about 200 mm) and an active layer
(<40 nm). The active layer is the actual mem-
brane, but due to its small thickness, it must be
supported by a thicker layer in order to obtain based on the application where parameters such as
mechanical strength, to withstand the pressure pressure, pressure drop, type of membrane avail-
difference between feed to permeate side (Scholz able, etc. are important. Depending on the concept
et al. 2011). chosen, asymmetric hollow bers are used with
The hollow ber membranes are self- their skin on the inside or on the outside (Mulder
supporting. Two types of module arrangement 1997).
can be distinguished: (i) where the feed passes The hollow ber module is the conguration
through the bore of the hollow ber (lumen) with the highest packing density, which can attain
whereas the permeate is collected on the outside values of 30,000 m2/m3 (Scholz et al. 2011). The
of the hollow bers (Fig. 1 Inside Feed) and (ii) typical properties and dimensions of hollow ber
where the feed enters the module on the shell side membrane modules are summarized in Table 1
of the hollow bers (external) and the permeate (Melin and Rautenbach 2007):
passes into the ber bore (Fig. 1 Outside Feed). The advantages of hollow ber modules are the
The choice between the two concepts is mainly following (Melin and Rautenbach 2007):
Hollow Fiber Modules for Desalination 957
Highest packing density and membrane area to multi-effect evaporation (MED), and multistage
module volume ratio ash (MSF) distillation. Although such conven-
Cheapest manufacturing costs tional processes are widely used, there is a need
for novel desalination techniques such as mem-
The disadvantages are the following: brane distillation (MD) which may be much
cheaper and have other distinct advantages. The
Mostly laminar ows (increased mass transfer membranes used in MD processes can be divided
limitations) into two forms: at sheet membrane and hollow
Lower pressure resistance ber-based membrane. Most of the work on
hydrodynamic improvement in MD studies has
Up to 80 % of the commercial gas separation focused on at sheet membrane modules. They
membranes are being formed into hollow ber have small membrane areas and thus are limited to
modules (Baker 2002). laboratory research. In industry, hollow ber-based
membrane modules are preferable due to their lager
membrane area per unit volume and reduced vul-
References nerability to temperature polarization. Based on the H
MD mechanism, the obtained ux depends both on
Baker RW (2002) Future directions of membrane the membrane permeation properties and the ow
gas-separation technology. Ind Eng Chem Res
geometry in the membrane modules. Therefore, the
41:13931411
Melin T, Rautenbach R (2007) research on the ux enhancement in MD can be
Membranverfahren Grundlagen der Modul- und divided into two large areas, the fabrication of
Anlagenauslegung. Springer, Berlin highly permeable membranes and designs of opti-
mized membrane modules.
ogy, 2nd edn. Kluwer Academic Publishers, Dordrecht
Scholz M, Wessling M, Balster J (2011) Design of mem- Recently, many designs of hollow ber mem-
brane modules for gas separations. In: Drioli E (ed) brane modules have been proposed in order to
Membrane engineering for the treatment of gases: suppress the undesirable temperature and concen-
gas-separation problems with membranes. RSC Pub-
tration polarization phenomena. The perfor-
lishing, Cambridge UK, p 5
mances of different polyvinylidene uoride
(PVDF) hollow ber membranes have been stud-
ied in different size modules (Fig. 1). The proper-
Hollow Fiber Module ties of the laboratory modules made, which
depend on the experiments requirements, can be
Hollow Fiber Membrane Module changed. The modules of PVDF have a stable
performance and can be used in reverse osmosis
drained wastewater treatment (Wu Chunrui
et al. 2010). The study in comparison of three
Hollow Fiber Modules membrane distillation congurations and seawa-
for Desalination ter desalination by vacuum membrane distillation
(VMD) showed that the shell-and-tube capillary
Chunrui Wu membrane module had unique advantages such as
State Key Laboratory of Hollow Fiber Membrane exible and changeable (Chen Huayan
Materials and Membrane Processes, Institute of et al. 2011).
Biological and Chemical Engineering, Tianjin Li et al. designed a novel membrane device
Polytechnic University, Tianjin, China made of novel membranes. To enhance the brine
side heat transfer coefcient substantially, they
employed cross ow of hot brine over the outside
The desalination of brine is commercially surface of hollow ber membranes. To achieve a
implemented by reverse osmosis (RO), compact device, they designed a hollow ber
958 Hollow Fiber Modules for Desalination
Hollow Fiber Modules

for Desalination,
Fig. 1 Modules of PVDF
hollow ber membrane

for Desalination,
Fig. 2 Rectangular cross
ow test module without
face plates
device built in a rectangular cross ow mode for processes (Gilron et al. 2007). The MD experi-
the hot brine to ow over the outside of the bers ments results have shown that the modules per-
(Fig. 2) (Li and Sirkar 2005). Simplied models formance was quite satisfactory and the design of
have been developed to describe the observed module enhanced the heat transfer as well as mass
performance of the modules in different MD transfer.
Hollow Fiber Modules for Desalination, Fig. 3 Different modules design and hollow ber geometries congurations
H
Different hollow ber congurations with (Yanga et al. 2011). The novel designs could be
wavy geometries (twisted and braided) for ux divided into structured-straight bers, curly bers,
enhancement in MD process have been developed central tubing for feeding, spacer-wrapped
by Teoh et al. in 2008 (Teoh et al. 2008). The bers, and spacer-knitted bers. The test results
different module designs, such as bafes and showed that the vapor permeate ux could be
spacers, were incorporated during the module enhanced from 53 % to 92 % compared with the
fabrication process, as illustrated in Fig. 3. A sin- conventional module, and the spacer-knitted
gle sieve spacer was placed about 7.5 cm away module had the best performance. Experiments
from both ends of the module in the spacer mod- reveal that improved ber geometries or
ule designs. And the distance between each spacer arrangements can provide a better ow distribu-
for the two or three-sieve spacers congurations tion, thus much lower pumping energy cost and
was 5 or 3.75 cm. Similarly, the helical bafe higher thermal efciency could be accomplished
introduced in the central of the module which (Fig. 4).
was surrounded by hollow bers was placed In 2005, Liu et al. prepared some coiled hollow
5 cm away from the both ends of the membrane ber membrane modules and experimentally
module. The results showed that the application of examined their mass transfer performances (Liu
bafes can increase the feed side heat transfer et al. 2005). The research results showed that,
coefcient. In addition, using spacers among the compared with the conventional straight module,
bers may increase the effective membrane area the mass transfer in both tube and shell side of the
1833 %. Lastly, the application of different hol- coiled hollow ber module could be remarkably
low ber congurations with wavy geometries enhanced. And the mass transfer coefcient of the
could lead to ux enhancements as high as 36 % coiled module at either side of membrane was
without inserting any external turbulent promoter. more than two times higher than that of the
In overall, the goal of greater ux enhancements straight module. The congurations of the coiled
with modied hollow ber membranes modules hollow ber membrane module and the conven-
was achieved. tional strait module are shown in Fig. 5. Different
Five types of novel hollow ber module con- geometric coiled modules were employed to study
gurations have been designed and constructed the inuences of the factors such as coiled diam-
by Yang et al. for the direct contact membrane eter, wind angle, and internal diameter of bers on
distillation (DCMD) process in 2011 mass transfer. It was found that wind angle had an
Hollow Fiber Modules for Desalination, Fig. 4 (continued)

Hollow Fiber Modules for Desalination, Fig. 4 Novel module design and fabrication: (a) structured-straight module;
(b) curly-ber module; (c) central-tubing module; (d) spacer-wrapped module; e spacer-knitted module
important inuence on the uid condition of the ber membranes and heat exchanged hollow
coiled module and its mass transfer. bers (Fig. 6). They investigated the inuences
Geng et al. developed a new air-gap membrane of operating factors such as feed ow rate, feed
distillation (AGMD) module with internal latent temperature, and feed initial concentration on
heat recovery conguration (Geng et al. 2014). AGMD process. Simultaneously they established
The novel module is consisted of parallel hollow a theoretical model based on the energy and mass

for Desalination,
Fig. 5 Schematic
representation of straight
module (a) and coiled
module (b)
Th
hollow fiber membrane
polypropylene net
heat exchange
hollow fiber Hot
Feed
Cold Cold
Feed Feed
Arrangements in cross section

of hollow fiber membranes
and heat exchange hollow
fibers Tc Tc
Hollow Fiber Modules for Desalination, Fig. 6 Schematic presentation of the novel AGMD module
balances of the hot feed side in membrane. References

The model was used to calculate the temperature
and the vapor permeate ux distributions along the Hongxin Geng, HaoyunWu, Pingli Li, Qingfeng He
(2014) Study on a new air-gap membrane distillation
hollow ber membrane. The performance of the
module for desalination. Desalination 334:2938
AGMD module was experimentally demonstrated Gilron J, Song L, Sirkar KK (2007) Design for cascade of
under different operating conditions, and the results cross ow direct contact membrane distillation. Ind
showed that the energy of AGMD process was Eng Chem Res 2007(46):23242334
Chen Huayan, Wu Chunrui, Jia Yue, Wang Xuan, Lu
recycled effectively, simultaneously, a high gained
Xiaolong (2011) Comparison of three membrane dis-
output ratio (GOR) of 5.7 was obtained. tillation congurations and seawater desalination by
Hollow Fiber Spinning 963
vacuum membrane distillation. Desalination Water vary from 0 (wet spinning) to more than 1 m. In
Treat 28:321327 wet spinning, the dope solution is extruded
Li B, Sirkar KK (2005) Novel membrane and device for
vacuum membrane distillation desalination process. directly into the precipitating liquid (external
J Membr Sci 2005(257):6075 coagulant). In this technique the spinneret is sub-
Liying Liu, Lijun Li, Zhongwei Ding, Runyu Ma, Zurong merged in the coagulation bath causing immediate
Yang (2005) Mass transfer enhancement in coiled hol- ber precipitation and solidication. In the
low ber membrane modules. J Membr Sci
264:113121. dry/wet spinning, coagulation of the internal sur-
Teoh MM, Bonyadi S, Chung T-S (2008) Investigation of face of the nascent ber starts immediately after
different hollow ber module designs for ux enhance- its extrusion from the spinneret, whereas the
ment in the membrane distillation process. J Membr Sci external surface experiences coalescence and ori-
2008(311):371379
Wu Chunrui, Jia Yue, Chen Huayan, Wang Xuan, Lu entation of polymer aggregates through the air gap
Xiaolong (2010) Membrane distillation and novel inte- before gelation in the external coagulation bath
grated membrane process for reverse osmosis drained takes place. After spinning, the nascent bers are
wastewater treatment. Desalination Water Treat 18:16 oriented by means of guiding wheels and nally
Yanga X, Wanga R, Fane AG (2011) Novel designs for
improving the performance of hollow ber membrane pulled into a collecting reservoir by a wind-up
distillation modules. J Membr Sci 2011(384):5262 drum. Only a ber with sufcient mechanical H
integrity can pass over the guiding wheels to the
collecting bath. During spinning, the take-up
velocity is generally kept at the same speed as
the free-falling velocity of the nascent ber to
Hollow Fiber Spinning prevent stretching of the ber (Khayet and
Matsuura 2011).
Khayet Mohamed Prior to the drying step, the spun hollow bers
Department of Applied Physics, Universidad usually go through the solvent exchange and other
Complutense de Madrid, Madrid, Spain posttreatment procedures in order to remove
residual solvents, prevent hollow ber shrinkage,
eliminate possible defects, and reduce pore col-
Hollow bers are fabricated by extruding or lapse. Conventional solvent exchange step before
pumping a dope solution through a spinneret fol- drying consists of immersion of the obtained hol-
lowing different methods such as melt spinning, low bers in one non-solvent or successive
dry spinning, wet spinning, or dry/wet spinning. immersion in several non-solvent mixtures
The spinneret has a tube-in-orice structure with containing components with lower surface ten-
typical dimensions 0.51 mm inner diameter and sions than water. Mixtures of water with increased
0.92 mm outer diameter. A simple experimental concentrations of lower surface tension compo-
device for the hollow ber spinning by dry/wet nents such as methanol and ethanol can be
spinning is shown in Fig. 1. In some spinning employed. Hexane can also be employed for fur-
systems, a circulation pump is employed instead ther solvent exchange since its surface tension is
of the gas cylinder (i.e., nitrogen) to drive the dope even lower than alcohols and it is volatile.
solution through the spinneret. The dope solution The hollow bers can be spun at room temper-
under gas pressure is forced to pass through the ature (2022 C) or at a temperature higher than
spinneret the annular space of the spinneret. The the ambient temperature. During spinning the
internal coagulant (bore liquid) is driven through take-up speed is normally maintained nearly the
the central tube either by the gravity force or by same as the dope extrusion speed so that no exter-
means of a circulation pump. The dope solution, nal elongational stresses, except gravity, can be
after being extruded from the spinneret, enters applied to the nascent hollow ber. A very large
into a coagulation bath before traveling a certain air gap distance may lead to ow instability of the
distance of air gap (i.e., the distance between the ber resulting in nonuniform ber dimensions.
spinneret and the coagulation bath), which may Examples of typical applied spinning parameters
964 Hollow Fiber Spinning
Hollow Fiber Spinning,

of a typical hollow ber
spinning system: 1 spinning
dope tank; 2 regulating
pressure valve; 3 pressure
gauge; 4 dope vessel;
5 dope valve; 6 bore liquid
vessel; 7 bore liquid pump;
8 spinneret; 9 air gap; 10
coagulation bath; 11 wind-
up drum; 12 ber collecting
reservoir; 13 wash water
(Reprinted from Khayet
et al. (2009). Copyright
2009, with kind permission
from Elsevier)
of different poly(vinylidene uoride), PVDF, used for solvent evaporation and solidication of
dope solutions are bore liquid ow rate the bers inuences to a signicant extent the
(0.53 ml/min), air gap distance (03 cm), take- quality of the produced bers.
up speed (free fall, 5 m/min), dope solution ow In melt spinning, the polymer is rst com-
rate (24 ml/min), applied pressure over the dope pressed, heated, and melted by an extrusion
solution (138207 kPa), and temperature of the screw, then forced through the annular space of
coagulants (15.525 C) (Wu et al. 2005, 2006, the spinneret. The polymer solidies by cooling
2007; Wang et al. 2008). after being extruded from the spinneret and then
In dry spinning, air or an inert gas is used collected by means of take-up wheels. The spin-
instead of the bore liquid so that solidication is ning rate in melt spinning (thousands of meter per
achieved through evaporation of the solvent. As the minute) is much higher than that used in the
dope solution exit the spinneret, air or an inert gas dry/wet spinning technique (meters per minute)
is used to evaporate the solvent(s) so that the ber (Mulder 1992).
solidies and collected on a take-up wheel. Spin- The spinneret geometry, the dope solution, and
ning is followed by stretching of the as-spun bers the spinning conditions (bore uid type and its
to orient the polymer chains along the gas gap (i.e., ow rate, dope solution extrusion rate, take-up
ber axis). The as-spun bers do not come into speed, air gap distance, external coagulant, spin
contact with a precipitating liquid. In dry spinning draw ratio, spinning temperature, cold stretching,
and melt spinning, the dimensions of the spinneret etc.) affect the hollow ber geometrical parame-
are not so crucial because the ber dimensions are ters (inner and outer diameters, thickness), its
mainly determined by the ratio of the extrusion rate internal structure, as well as the structural mor-
and tearing rate (Mulder 1992). The gas ow phology of the inner and outer surfaces.
Homogeneous Anion-Exchange Membranes 965
References
+ + + +
ciples and applications. Elsevier, Amsterdam 2 + 5 6
1
Khayet M, Garca-Payo MC, Qusay FA, Zubaidy MA 3
7
(2009) Structural and performance studies of poly
(vinyl chloride) hollow ber membranes prepared at + +
different air gap lengths. J Membr Sci 330:3039
Mulder M (1992) Basic principles of membrane technol- 5 7
Wang KY, Chung TS, Gryta M (2008) Hydrophobic PVDF
4
hollow ber membranes with narrow pore size distri-
bution and ultra-skin for the fresh water production
through membrane distillation. Chem Eng Sci
63:25872594 Homogeneous Anion-Exchange Membranes,
Wu B, Tan X, Teo WK, Li K (2005) Removal of benzene/ Fig. 1 A typical preparation method of AEMs (paste
toluene from water by vacuum membrane distillation in method). 1 Reserve roll of reinforcing fabric, 2 reserve
a PVDF hollow ber membrane module. Sep Sci roll of separating lm, 3 receiving role, 4 paste (mixture
of vinyl monomers, inert polymers, initiator, additives,
Technol 40:26792695
etc.) reservoir, 5 expand roll, 6 press roll, 7 guide roll
H
Wu B, Tan X, Li K, Teo WK (2006) Removal of 1,1,1-
trichloroethane from water using a poly(vinylidene
uoride) hollow ber membrane module: vacuum
membrane distillation operation. Sep Purif Technol (Sata 2004). A monomer mixture consisting of
52:301309 chloromethylstyrene (CMS)/DVB, acrylonitrile-
Wu B, Li K, Teo WK (2007) Preparation and characteriza-
tion of poly(vinylidene uoride) hollow ber mem- butadiene rubber, and benzoyl peroxide to initiate
branes for vacuum membrane distillation. J Appl polymerization was used to prepare the paste.
Polym Sci 106:14821495 A reinforcing material such as woven poly (vinyl
chloride) cloth, reserved to roll 1, was dipped in
the paste reservoir 4. A separating lm such as
poly(ethylene terephthalate) was fed from roll
Homogeneous Anion-Exchange 2. The reinforcing material coated with the paste
Membranes was wound on roll 3 in layers with the separating
lm. In this way, the paste (the monomer mixture)
Mitsuru Higa was continuously coated on the reinforcing mate-
Graduate School of Science and Engineering, rial and covered on both sides with the separating
Yamaguchi University, Ube, Yamaguchi, Japan lm. A precursor membrane was obtained by
heating the resultant composite at 80 C for 10 h
under a nitrogen atmosphere to polymerize the
Anion-exchange membranes (AEMs) are classi- monomers. An ion-exchange group was intro-
ed into two types by their microstructure: het- duced by quaternization by immersing the precur-
erogeneous and homogeneous. Homogeneous sor membrane in an aqueous solution of 1.0 mol/
AEMs consist of a uniform membrane matrix dm3 trimethylamine at 30 C for 20 h. The cross-
with anion-exchange groups (positively charged linker content (CLC) in the mixture was changed
groups) and, in many cases, reinforcing materials to control the network structure of the membranes.
such as woven cloth or net. Various methods The CLC was dened as follows:
have been reported for preparing homogeneous
AEMs; however, almost commercial homoge- CDVB
CLC (1)
neous AEMs are copolymer membranes com- CCMS CDVB
posed of styrene and divinylbenzene (DVB) with
benzyl trimethylammonium groups. Figure 1 where CDVB and CCMS are the content of DVB
shows a schematic diagram of a typical prepara- and CMS in the mixture. The value of CLC is
tion method (paste method) of the AEMs controlled to optimize the fundamental
966 Honeycomb Ceramic Membranes
physicochemical properties, transport number of widely used as catalyst supports and as particulate
anions and electrical resistance, to their applica- lter (Pronob 1997).
tions. In general, the transport number and elec- The honeycomb membrane is viewed as a type
trical resistance of AEMs increase with of multichannel membranes. It brings the element
increasing CLC. of ceramic honeycomb monoliths, with proprie-
Other preparation methods of homogeneous tary modications, as membrane supports. Such
AEMs have been reported such as: (a) modica- large honeycomb structures are characterized by
tion of aromatic polymer. An aromatic polymer very high membrane area, up to several hundred-
such as poly(ether-ether-ketone) is chloro- fold the amount of area in most commercially
methylated using chloromethyl ether in the pres- available ceramic membrane elements. Variations
ence of a catalyst, and then the chloromethylated such as adding the pore numbers and decreasing
polymer is quaternized using triethylamine. the diameter of every pore will improve mem-
The quaternized polymer solution is cast to form brane area. The loading density of ceramic mem-
an AEM. (b) Pore-lling method. AEMs are pre- brane is also improved with the honeycomb
pared by lling the pores of a porous lm (e.g., design. The cross section schematic diagram of
porous polyethylene) with a mixture of honeycomb ceramic membrane is shown in Fig. 1.
vinylbenzene-based monomers, a cross-linking However, there are disadvantages in honey-
agent, and an initiator. Then, thermal cross- comb ceramic membrane. The rst disadvantage
linking polymerization is performed. Finally, qua- is asymmetry distribution of uid. Because of the
ternary ammonium groups are introduced. honeycomb element of high sectional area, the
(c) Electron-beam-induced graft polymerization. volume of membrane is relatively large. It will
A polymer lm such as high-density polyethylene lead to the consequence that the velocity of uid
is grafted by electron-beam-induced graft poly- in honeycomb ceramic membrane which is coax-
merization with CMS and DVB. The quaternary ial to pipe of raw materials is higher than uids
ammonium groups are introduced by treating the from other uids. So it is hard to control concen-
grafted lm with trimethylamine. tration polarization in membrane as the asymme-
try distribution of uid. Another disadvantage is
high permeate pressure. With decrease of pore
References diameter and pore wall thickness, more permeate
liquid passes through inner pore of honeycomb
Sata T (2004) Ion exchange membrane. The Royal Society ceramic support. So the transmission resistance is
of Chemistry, Cambridge
improved, permeation ux is decreased. Using the
method to change parts of membrane pores to
permeation ux channel, or to slot ports, perme-
ation length is decreased and permeation ux is
Honeycomb Ceramic Membranes improved. But unit area of membrane decreases in
some extent. So post-processing of honeycomb
Wanqin Jin ceramic membrane is extremely important,
State Key Laboratory of Materials-Oriented despite of its difculties. The other disadvantage
Chemical Engineering, Nanjing University of is high cost with low yield. Since the preparation
Technology, Nanjing, China process is very complex, probabilities of mem-
brane defects are much higher. Those defects can
be improved with designed membrane modules.
Honeycomb structure body possesses a specied The use of honeycomb structure in ceramic
pore distribution on the surface of the partition membrane is promising. It has been used in
walls and an enlarged porosity in addition to thin reverse osmosis desalination for sea water.
partition walls of a prescribed value. Because of Exploratory and advanced developments have
the special structure, ceramic honeycombs are been accomplished in some companies. With
Honeycomb Membrane Structure 967
Honeycomb Ceramic
Membranes, Fig. 1 The
cross section schematic Raw material channel
diagram of honeycomb
ceramic membrane
Porous membrane
surface
Ceramic support body
mean pore size of 50 nm, honeycomb ceramic Honeycomb membrane structures are a competi-
membrane improves the quality of pure water in tive and concrete example of highly ordered poly- H
osmosis process to 86 %. And the permeation ux meric patterns, which can work as perm-selective
can be kept in about 150 L/(m2hbar) according to interfaces in advanced membrane processes
experiments mentioned in several articles. (Gugliuzza et al. 2008). They are prepared by
using a bioinspired nanotechnology, which
exploits the ability of water to condense from
References fog to cooled surfaces and self-assemble in semi-
crystalline lattices under dragging Marangoni
Pronob B (1997) Ceramic honeycomb lters and catalysts. convection and capillary forces (Gugliuzza
Curr Opin Solid State Mater Sci 2:577583
et al. 2009). When the solvent evaporates, a dif-
http://www.jiuwu.com/en_index.aspx
ference of temperature is established between the
cooled liquid lm and nearby humid atmosphere,
causing droplets to condense and rearrange in
honeycomb-packed geometries. The droplet lat-
Honeycomb Membrane tice works as a real pore builder through the poly-
meric solution, leading to the formation of air
Monolithic Membranes bubble arrays. Each single cavity can be regarded
as the result of the imprinting action of the drop-
lets. No coalescence is observed when a stable
polymeric protective layer is formed around each
single droplet during sinking (Fig. 1).
Honeycomb Membrane Structure Under suitable experimental conditions, drop-
let dynamics can be directed towards the forma-
Annarosa Gugliuzza tion of membranes with modulated spherical pore
Institute on Membrane Technology, National size, monodisperse pore distribution, long-range
Research Council of Italy, ITM-CNR, Rende, order, and high interconnectivity (Speranza
Italy et al. 2010). The typical hexagonally packed
geometry, resembling natural structures, produces
very high interfacial area necessary to promote
Synonyms large mass transfer through the membranes.
Also, the complete propagation of the porous
Air bubble geometry; Breath gure membranes; structure along the entire membrane thickness
High-dened porous lms reduces signicantly any resistance to transport,
968 Honeycomb Membrane Structure
Honeycomb Membrane Structure, Fig. 1 Representative mechanism for the formation of honeycomb membranes
yielding performance higher than that observed Chemical and structural features make the hon-
for commercial tubular modules when worked in eycomb membranes ideal interfaces for some
membrane distillation plants (Gugliuzza membrane contactors operations, including mem-
et al. 2008). brane distillation, membrane crystallization, and
The singular topography of honeycomb mem- membrane emulsication (in particular premix
branes has signicant consequences on wettabil- membrane emulsication), where large mass
ity as well. The hexagonally packed geometry transfer along with the establishment of durable
increases the surface roughness according to zig- interfaces at the entrance of each single pore are in
zag geometrical model (Thomas 1999), great demand. Moreover, the extraordinary regu-
preventing wetted pores owing to air pockets uni- larity and stability of 3D polymeric architecture
formly distributed through the surface. Indeed, the make these membranes suitable for working as
air is easily entrapped inside the big cavities molecular reservoirs, scaffolds for tissue engi-
according to the CassieBaxter model, thereby neering, and platforms for sensors.
minimizing the spreading of the liquid and Compared with traditional manufacturing pro-
improving signicantly the water resistance of cedures, the one-pot fabrication of micro- and
the membrane. nanostructured membranes, based on the self-
Hybrid Modeling of Membrane Biofilm Reactors 969
assembly of sacricial natural building blocks, is Hybrid membranes can be classied into two
regarded as a novel, exible, and eco-friendly main types, depending on the nature of the inter-
technology. This approach is expected to open action between components: (1) systems where
new routes to fabricate a large variety of ordered there are no covalent or iono-covalent bonds
functional membranes for industrial applications, between components, only Van der Waals, hydro-
but also for handling the miniaturization of light- gen bonding, or electrostatic forces, and (2) sys-
weight systems perceptively. tems where at least parts of the components are
linked through strong covalent or iono-covalent
bonds (Sanchez and Gmez-Romero 2004). The
References rst type can also be named as composites, or
nanocomposites, where at least one of the compo-
Gugliuzza A, Aceto MC, Macedonio F, Drioli E (2008) nents domains has a dimension ranging from a
Water droplets as template for next generation self-
few angstroms to several nanometers.
assembled poly-(etheretherketone) with Cardo mem-
branes. J Phys Chem B 112(34):1048310496 The main objective in the synthesis of hybrid
Gugliuzza A, Speranza V, Trotta F, Drioli E (2009) membranes is the performance improvement of
Bio-inspired membranes with well-dened channels. the material for different applications. It is obvi- H
Chem Eng Trans 17:15371542
ous that the properties of these membranes are not
Speranza V, Trotta F, Drioli E, Gugliuzza A (2010) High-
denition polymeric membranes: construction of 3D only the sum of the individual contributions of
lithographed channel arrays through controlling natural both components, and interfaces can play a
building blocks dynamics. ACS Appl Mater Interfaces signicant role.
2(2):459466
The high number of parameters involves in the
Thomas JR (1999) Rough surfaces. Imperial College Press,
London design and preparation of hybrid membranes: the
number of components, composition, components
ratio, size, shape, and kind of interaction between
components results in an almost innite number
Horizontally Aligned Carbon of combinations.
Nanotube
Aligned Carbon Nanotube References
Sanchez C, Gmez-Romero P (2004) Functional hybrid

materials. Wiley VCH, Weinheim
Hybrid Membranes
Mario Aparicio Hybrid Modeling of Membrane

Instituto de Ceramica y Vidrio (CSIC), Madrid, Biofilm Reactors
Spain
Endre Nagy
The term hybrid membrane more commonly Engineering, University of Pannonia, Veszprm,
refers to a membrane formed by at least two Hungary
components of different chemical nature
(considering chemical-bond modes) from the
groups of metals, organic materials and their poly- The immobilized whole cells in a membrane bio-
mers, and inorganic materials. reactor are grown within the bers and/or in the
970 Hybrid Modeling of Membrane Biofilm Reactors
extracapillary space-forming biolm. The reactor: mass transport (c = cO2 , YXO is yield
conventional biolm is increasingly used for coefcient):
wastewater purication, oxidation of organic
p
components, nitrication, etc. as they are environ- J k1 D tanh#co (2)
mentally friendly and less energy intensive. Bio-
lm formation is a complicated dynamical process with
governed by various physical and chemical prin- s
ciples and biological protocols. Biolms are k1 d2 1 vmax
W ; k1
agglomerations of microorganisms or bacteria D YXO kMO2
p
cd
whose structure can be homogeneous, in case of
k1 D o
sufcient substrate level, or heterogeneous when J c (3)
tanh# cosh#
the diffusion is rate limiting, due to the change of
the nutrient concentration in axial direction D 2 o
(Gavin and Gillian 2012). This circumstance can J # c (4)
d
alter the value of transport parameters (diffusion
coefcient, convective velocity) and can even with
alter the biochemical reaction rate (the consump-
s
tion rate of nutrient). The variation of cell density vmax d2
is also true in the biocatalyst membrane layer #
DY XO2
perpendicular to the inlet surface. Increasing dis-
tance from the surface can mean decreasing nutri-
D o #2
ent concentration. That is why the variability of J c cd (5)
the transport parameters should also be taken into d 2
account. Assuming simultaneous effect of two
I. n order to overcome the diffusive mass transport
substrates, namely, oxygen and carbon sources,
limitation of nutrients, especially of the sparingly
the mass balance for diffusive ow can be given
soluble oxygen, the substrate might be fed on both
for a sheet membrane layer with constant param-
sides of the biolm applying oxygen permeable
eters, as
support layer. The sum of the mass transfer rates
for rst- and zero-order reactions can be expressed
d 2 cO 2 1 mmaxO2 cO2 cA by Eqs. 6 and 7, respectively. The substrate con-
D 0:
dy 2 Y XO2 K MO2 cO2 K MA cA centrations of the two sides of the biolm layer are
(1) co1 and co2. The substrate concentration curve has
inection point at Y = Y1 (Y = y/d):
The quasi-analytical approach of solution of Eq. 1 p
is given, e.g., by Nagy (2011). Let us give here the k1 D 1
J 1 co
rate equations for limiting cases with constant cA tanh# cosh2 #Y 1 1
concentration in the biolm. Two important cases
1
can be distinguished, namely, when the outlet 1 co (6)
cosh2 #1 Y 1 2
(at y = d) diffusive ow is zero (dc/dy = 0, the
support membrane layer is impermeable) and with
dc/dy 6 0 at y = d (there is sweep phase on the
permeate side of a permeable membrane). The co1 cosh#Y 1

mass transfer rates are given for these cases, co2 cosh#1 Y 1
for rst-order reaction by Eqs. 2 and 4 and for
D o #2
zero-order reaction by Eqs. 5 and 7, respectively. J c c2 2c1 1 2Y 1 1 #
o o 2
The transfer rate in the presence of convective d 1 2

ows is given in entry biocatalytic membrane (7)
Hybrid Organic-Inorganic Nanostructured Membranes 971
with Improved and also new properties are expected

in this specic kind of hybrid membranes because

#2 of the combination of two very different materials
co1 1 2Y 1 1 and their chemical interactions. For example, it is
2
possible to obtain membranes with a high exi-
#2
c2 1 2Y 1 1 :
o
(8) bility and processability, such as a polymer, but
2 with the improved thermal and chemical stability,
and mechanical strength of inorganic materials
The Y1 inection point can be determined by (Sanchez et al. 2005).
Eq. 8 and then the mass transfer rate by Eq. 7 as The possible applications of these hybrid
well. Detailed solution of Eq. 1 is published by membranes are increasing continuously. For
Nagy (2011), recommended explicit expression example, an antimicrobial drug (substituted
for the mass transfer rate. The biolm can involve 1,3,4-oxadiazole) with functionalized silica was
several substrates with several parallel and/or con- successfully incorporated into an organic phase
secutive reactions as this is the case of aerobic by sol-gel to achieve a highly stable and anti-
wastewater treatment. System of the differential biofouling membrane for water treatment (Singh H
mass balance equations of the reactants should be et al. 2012). Hybrid organic-inorganic nanostruc-
solved in order to get their concentration distribu- tured membranes have also found applications in
tions and the mass transfer rates of the transferring the medical eld for advanced separation of heavy
substrates. metals from blood or other physiological liquids,
such as new polymeric-carbon nanotube compos-
ite membranes based on polysulfone with differ-
References ent types of nanotubes (Nechifor et al. 2009).
Another important area where these hybrid mem-
Gavin L, Gillian DL (2012) Microbial biolms: current branes may have a signicant relevance is the
research and application. Caister Academic Press,
Christchurch
energy sector, especially as new membranes for
Nagy E (2011) Basic equations of the mass transport proton exchange membrane fuel cells (PEMFC)
through a membrane layer. Elsevier, Amsterdam and solid-state Li-ion batteries. There membranes
share several common characteristics, such as
high ion conductivity (proton and lithium ion,
respectively), low electronic conductivity, high
thermal stability, and high chemical/electrochem-
Hybrid Organic-Inorganic ical stability. Decreasing the membrane thickness
Nanostructured Membranes but preserving the properties described above
would improve the performance of the systems.
Mario Aparicio Hybrid organic-inorganic nanostructured mem-
Instituto de Ceramica y Vidrio (CSIC), Madrid, branes can be designed to incorporate all these
Spain properties (Mosa and Aparicio 2012).
A hybrid organic-inorganic nanostructured mem-

References
brane is a specic kind of hybrid membrane
formed by at least two components from organic Mosa J, Aparicio M (2012) Hybrid materials for high ionic
materials and their polymers and inorganic mate- conductivity. In: Aparicio M, Jitianu A, Klein LC (eds)
rials. The components of this kind of hybrid mem- Solgel processing for conventional and alternative
energy, 1st edn. Springer, New York, pp 99122
branes have dimensions up to several nanometers,
Nechifor G, Voicu SI, Nechifor AC, Garea S (2009) Nano-
and components are linked through covalent or structured hybrid membrane polysulfone-carbon
iono-covalent bonds. nanotubes for hemodialysis. Desalination 241:342348
972 Hybrid Processes
Sanchez C, Julian B, Belleville P, Popall M (2005) Appli- separation equipment can be combined.
cations of hybrid organicinorganic nanocomposites. Depending on the process, membranes are applied
J Mater Chem 15:35593592
Singh AK, Singh P, Mishrab S, Shahi VK (2012) Anti- for both, to perform the bulk separation and to
biofouling organicinorganic hybrid membrane for polish the respective products. Hybrid processes
water treatment. J Mater Chem 22:18341844 are particularly interesting when membrane units
are added to existing separation equipment.
Hence, debottlenecking of processes or research
activities can easily be performed without shutting
Hybrid Processes down the separation process (Hmmerich and
Rautenbach 1998).
Marco Scholz
Aachener Verfahrenstechnik Chemical Process
Engineering, RWTH Aachen University, Aachen, Examples of Hybrid Processes
Germany
Skiborowski et al. analyzed a hybrid process in
which seawater is desalinated (Skiborowski
In membrane science, hybrid processes are pro- et al. 2012). Reverse osmosis modules are com-
cesses, in which membrane technology is combined with forward-feed multi-effect distillation.
bined with conventional separation equipment to For low energy costs, they identied the reverse
perform a separation more efciently. The objec- osmosis network as the most protable plant. For
tive is to combine the positive characteristics of moderate energy costs and for high energy costs,
each individual technology to obtain a process the hybrid process and the multi-effect distillation
which is better than the individual process. are the optimal process conguration,
Reverse osmosis, gas permeation, and in par- respectively.
ticular pervaporation are the most prominent Baker and Lokhandwala investigated natural
membrane processes which are combined with gas processing in detail (Baker and Lokhandwala
conventional separation equipment to form hybrid 2008). Here, typically membrane technology and
processes (Melin and Rautenbach 2007). Conven- amine scrubbing compete. The carbon dioxide
tional separation equipment is either limited by content in the feed gas and the feed gas ow rate
equilibrium and kinetics (e.g., amine) or by ther- determine which separation technology should be
modynamic equilibrium only (e.g., cryogenic). applied. Interestingly, for high carbon dioxide
However, the membrane separation is controlled mole fractions and for high feed ow rates, the
by kinetics. Hence, combining both the equilib- combination of membrane technology and amine
rium and the kinetic limited separation can have absorption is the most efcient process
outstanding advantages. For instance, combining conguration.
a distillation column and a pervaporation module The vapor recovery of gasoline vapors from
is protable as the pervaporation unit breaks the storage tanks can be done by applying rubbery gas
azeotrope and the bulk separation is performed by permeation membranes and condensation equip-
distillation. ment (Rautenbach et al. 1996). Here, the bulk
Combining membrane technology and con- separation is done by condensation and the gas
ventional separation technology can signicantly is subsequently polished in a gas permeation unit.
reduce the equipment size. By applying hybrid Both, the recovery of gasoline and environmental
processes, thermodynamic limitations can be standards for the exhaust gas, have to be consid-
overcome or more efcient processes can be ered in the design of such a plant.
designed regarding the capital as well as the oper- In the MTBE (methyl tert-butyl ether) produc-
ational costs. tion methanol, MTBE and n-C4 have to be sepa-
There are many different process congura- rated. The separation is complex due to the
tions, in which membranes and conventional formation of azeotropes (Hmmerich and
Hybrid Regenerated Cellulose/Loaded Lipid Nanoparticle Membranes 973
Rautenbach 1998). Two hybrid processes can be (Muller et al. 2000). Among them, lipid
applied: one in which the mixture is fed to a nanoparticles (LNPs) prepared using biocompati-
pervaporation module to remove the methanol ble components and with tunable properties are of
while the MTBE and the n-C4 fraction are sepa- signicant interest (Gupta and Kompella 2006;
rated in a distillation column. In the other process, Huynh et al. 2009); although, their use is still
the raw mixture is fed to a distillation column limited due to stability problems during contact
while the pervaporation module operates at a with biological uids, storage, or administration
side draw. This latter hybrid process outclasses (Korting and SchferKorting 2010). To over-
the conventional process, in which different dis- come these limitations, polymeric nanosphere
tillation columns are applied. gels were proposed for topical delivery of lipo-
philic molecules (Martins et al. 2007). In this
context, the inclusion of functionalized lipid
References nanoparticles (FLNPs) in support to biocompati-
ble membranes (such as regenerated cellulose
Baker RW, Lokhandwala K (2008) Natural gas processing membranes) offers an attractive route for con-
with membranes: an overview. Ind Eng Chem Res
trolled release of pharmacologic agents. H
47:21092121
Hmmerich U, Rautenbach R (1998) Design and The LNPs used in this study were prepared by
optimization of combined pervaporation/distillation the ultrasound method (Feng and Huang 2001)
processes for the production of MTBE. J Membr Sci using L-a-phosphatidylcholine and Tween
146:5364
80 as surfactants and glyceryl tristearate as the
Melin T, Rautenbach R (2007) Membranverfahren-
Grundlagen der Modul-und Anlagenauslegung. main lipid component, while sunscreen DHB
Springer, Berlin (2,4-dihydroxybenzophenone) was the active
Rautenbach R, Knauf R, Struck A, Vier J (1996) Simula- organic component. The hybrid membrane was
tion and design of membrane plants with Aspen Plus.
obtained by embedding the DHBLNPs in a
Chem Eng Technol 19:391397
Skiborowski M, Mhamdi A, Kraemer K, Marquardt dense highly hydrophilic regenerated cellulose
W (2012) Model-based structural optimization of sea- (RC) support by immersion in a water dispersion
water desalination plants. Desalination 292:3044 of DHBLNPs (membrane RC/DHBLNPs). The
incorporation of the loaded LNPs was character-
ized by AFM, brilliant eld microscopy (Fig. 1a),
TEM (Fig. 1b), and Raman spectroscopy
(Vzquez et al. 2011; Hierrezuelo et al. 2012).
Hybrid Regenerated Cellulose/ Reductions in material characteristic parameters
Loaded Lipid Nanoparticle (conductivity and dielectric constant) for dry
Membranes: Preparation RC/DHBLNP samples when compared with the
and Characterization original RC and changes associated to thermal
effects were also obtained.
Juana Benavente1 and Juan Manuel The stability of the hybrid membrane as a
Lpez-Romero2 result of both contact time with NaCl solutions
1
Departamento de Fisica Aplicada I, Universidad and osmotic pressure gradients was established by
de Malaga, Facultad de Ciencias, Malaga, Spain comparing diffusional permeability (Ps) for
2
Departamento de Quimica Organica, original RC and RC/LNP membranes (Vzquez
Universidad de Malaga, Facultad de Ciencias, et al. 2011; Hierrezuelo et al. 2012). The presence
Malaga, Spain of the DHBLNPs reduces in approximately 20 %
Ps values for the whole interval of feed
concentrations studied (0.001 Cf(M) 0.4),
New membrane systems related to medical appli- and a constancy in the NaCl ow for at
cations (drug-release devices, mimetic mem- least 20 h was also observed. The cellulose struc-
branes, or patches) are receiving great attention ture also reduces the DHB delivery from the
974 Hybrid Silica Membranes
Hybrid Regenerated Cellulose/Loaded Lipid Nanoparticle Membranes: Preparation and Characterization,

Fig. 1 Hybrid membrane micrographs: (a) brilliant led microscopy (b) TEM
loaded LNPs, being the value for the hybrid

membrane 5 % of that for the DHBLNPs (after Hybrid Silica Membranes
30 min).
Toshinori Tsuru
Department of Chemical Engineering, Hiroshima
References University, Higashi-Hiroshima, Japan
Feng SS, Huang G (2001) Effects of emulsiers on the

controlled release of paclitaxel (Taxol ) from A great deal of attention has been paid to silica
nanospheres of biodegradable polymers. J Control
Release 71:5369
membranes, which can accomplish separation
Gupta RB, Kompella UB (2006) Nanoparticle technology tasks even under harsh conditions due to their
for drug delivery. Taylor and Francis/CRC Press, New high-separation performance, chemical inertness,
York and thermal stability (de Vos and Verweij 1998;
Hierrezuelo J, Benavente J, LpezRomero JM, Martnez
de Yuso MV, RodrguezCastelln E (2012) Prepara-
Tsuru 2008). Two strategies have been proposed to
tion, chemical and electrical characterizations of lipid prepare amorphous molecular sieving silica mem-
nanoparticles loaded with dihydroxybenzophenone. branes: the solgel process and the chemical vapor
Med Chem 8:541548 deposition (CVD) method. Tetraethoxysilane
Huynh NT, Passirani C, Saulnier P, Benoit JP (2009) Lipid
nanocapsules: a new platform for nanomedicine. Int
(TEOS), which is a common Si precursor, has
J Pharm 379:201219 been widely adopted in both approaches, and the
Korting HC, SchferKorting M (2010) Carriers in the resultant membranes have been extensively inves-
topical treatment of skin disease. Handb Exp tigated for gas separation, pervaporation, and
Pharmacol 197:435468
Martins S, Sarmento B, Ferreira DC, Souto EB
membrane reaction. However, traditional TEOS-
(2007) Lipidbased colloidal carriers for peptide and derived silica membranes suffer from poor hydro-
protein deliveryliposomes versus lipid nanoparticles. thermal stability, which greatly hampers their prac-
Int J Nanomedicine 2:595607 tical application. Inspired by the distinctive
Muller RH, Mder K, Gohla S (2000) Solid lipid
nanoparticles for controlled drug delivery a review
features of organicinorganic hybrid silica mate-
of the state of the art. Eur J Pharm Biopharm rials with respect to exibility, hydrothermal sta-
50:161177 bility, and tunable pore sizes, researchers have
Vzquez MI, Pelez L, Benavente J, Lpez-Romero JM, concentrated on the preparation of hybrid silica
Rico R, Hierrezuelo J, Guilln E, Lpez-Ramrez MR
(2011) Functionalized lipid nanoparticles-cellophane
membranes with improved hydrothermal stability
hybrid lms for molecular delivery. J Pharm Sci and controlled pore sizes, using organoalkox-
100:48154822 ysilanes instead of TEOS.
Hybrid Silica Membranes 975
Organic/inorganic hybrid membranes can be (triuoropropyl)triethoxysilane (TFPTES),

prepared either by the post-modication of hydro- ethylenetriethoxysilane (ETES), or
philic ceramic membranes (Tsuru et al. 2004) or 3-methacryloxypropyltrimethoxysilane (MPS)
by the in-situ preparation using hydrophobic sols. (Raman and Brinker 1995; de Vos et al. 1999;
In post-modication, hydrophilic porous mem- Tsuru et al. 2011). Without burning out organic
branes, the pore surface of which is covered with pendant groups, hybrid hydrophobic silica mem-
hydrophilic hydroxyl groups such as -SiOH, were branes could be obtained via the solgel process,
modied with silane coupling agents such as which was rst reported by de Vos et al. using
trimethylchlorosilane in either a gas or a liquid TEOS and MTES as co-precursors (de Vos
phase, to convert the hydroxyl groups to a et al. 1999). The hybrid membranes prepared via
trimethylated surface. However, this method has this route also showed good hydrothermal stabil-
several disadvantages; post-modication always ity and have been used for both gas separation and
reduces the effective pore sizes, leading to a pervaporation. Considerable research on silica
decrease in ux. Moreover, it is difcult to modify membranes prepared by mono-silicon alkoxides
membrane pores to less than several nanometers with pendant groups has led to a great deal of
since the silane coupling agents cannot enter such progress in understanding both the solgel pro- H
small pores (Tsuru et al. 2004). cess and the membrane characteristics.
In-situ preparation of hydrophobic silica mem-
branes was reported in solgel processing. Two
typical kinds of organoalkoxysilanes have been
References
used as Si precursors to prepare silica membranes
via a solgel process. One is a bridged alkoxide, Castricum HL, Sah A, Kreiter R, Blank DHA, Vente JF, ten
that is, a di-silicon alkoxide with a bridged group Elshof JE (2008a) Hybrid ceramic nanosieves: stabiliz-
between the two silicon atoms, such as 1, 2-bis ing nanopores with organic links. Chem Commun
(triethoxysilyl)ethane(BTESE) (CH3CH2O)3Si 110371105. DOI: 10.1039/B718082A
Castricum HL, Kreiter R, van Veen HM, Blank DHA,
CH2CH2Si(OCH2CH3)3 and bis(triethoxysilyl) Vente JF, ten Elshof JE (2008b) High-
methane (BTESM) (CH3CH2O)3SiCH2Si performance hybrid pervaporation membranes with
(OCH2CH3)3 (Kanezashi et al. 2009; Castricum superior hydrothermal and acid stability. J Membr Sci
et al. 2008a, b). Castricum and coworkers (2008a; 324:111
de Vos RM, Verweij H (1998) High-selectivity, high ux
Castricum et al. 2008b) rst reported on the silica membranes for gas separation. Science 279:1710
BTESE-derived silica membrane and applied it de Vos RM, Maier WF, Verweij H (1999) Hydrophobic
to pervaporation dehydration of solvents of silica membranes for gas separation. J Membr Sci
95 wt% n-butanol-5 wt% water in an attempt to 158:277
Kanezashi M, Yada K, Yoshioka T, Tsuru T (2009) Design
increase hydrothermal stability. The membrane of silica networks for development of highly permeable
was stable after 1.5 years, even though the test hydrogen separation membranes with hydrothermal
temperature was as high as 150 C, while TEOS- stability. J Am Chem Soc 131:414
derived silica membrane completely lost its sepa- Raman NK, Brinker CJ (1995) Organic template
approach to molecular sieving silica membranes.
ration ability within weeks at a much lower tem- J Membr Sci 105:273
perature of 95 C. The excellent hydrothermal Tsuru T (2008) Nano/subnano-tuning of porous ceramic
stability endowed by the organicinorganic struc- membranes for molecular separation. J Solgel Sci
tures appears to be very promising for practical Technol 46:349
Tsuru T, Kondo H, Yoshioka T, Asaeda M (2004) Perme-
application. The formation of a looser structure ation of nonaqueous solution through organic/inor-
and improved hydrothermal stability was due to ganic hybrid nanoporous membranes. AIChE
the presence of a SiCCSi minimum unit in the J 50:10801087
silica networks. Tsuru T, Nakasuji T, Oka M, Kanezashi M, Yoshioka
T (2011) Preparation of hydrophobic nanoporous meth-
Another typical kind of organoalkoxysilane ylated SiO2 membranes and application to
is a mono-silicon alkoxide with a pendant nanoltration of hexane solutions. J Membr Sci
group, such as methyltriethoxysilane (MTES), 384:149156
976 Hydraulic Resistance due to Reversible Fouling
At the top of the column, a head stream rich in

Hydraulic Resistance due to dibranched parafns with six carbon atoms,
Reversible Fouling isopentane, and normal pentane. This stream,
which corresponds to 90 % by weight of the
Reversible Fouling Resistance fresh feed, has an RON (research octane num-
ber) of 87 though it contains about 16 % by
weight of normal pentane, which has a low
RON of 61.
Hydrocarbon Branching With current distillation processes, it is not

economically feasible to separate normal pentane
Arnaud Baudot from the other components in the top stream of the
Physics and Analysis Division, IFP Energies deisohexanizer since their respective boiling
nouvelles, Solaize, France points are very close. This type of separation,
however, can be achieved with molecular sieves,
such as zeolites, implemented in cyclic adsorption
For more than 20 years now, regulations have processes such as a simulated moving bed
imposed increasingly tight limitations on the con- (UOP Molex process) or cycled pressurization/
tent in gasoline of aromatic octane number depressurization (IFP IPSORB process or
boosters produced by the reforming of straight- ExxonMobil ISOSIEVE process). With such pro-
run gasoline (alkanes). Among the available alter- cesses, normal parafns are preferentially
native technologies designed to enhance the adsorbed inside the microporosity of the zeolites
octane number of straight-run gasoline, and therefore separated from their branched iso-
hydroisomerization is a catalytic technology that mers. Though offering excellent separation per-
upgrades low-octane-number linear parafns into formance, this type of technology exhibits several
higher-octane-number branched parafns. drawbacks: high investment costs, sophisticated
Since the rate of conversion of linear parafns sequential operation (adsorption-desorption
in the isomerization units is limited by a thermo- cycle), the use of large quantities of solvents
dynamic equilibrium, an option for increasing the (desorbants), and lack of modularity.
production yield of dibranched parafns consists Since the beginning of the 1990s, much atten-
of separating the linear and monobranched paraf- tion has been paid to overcoming the drawbacks of
ns from the isomerization unit efuent and conventional zeolite adsorbents through the devel-
recycling them in the input of the isomerization opment of zeolite membranes that combine the
reactor. The more conventional solution consists technical advantages of membranes (modularity,
of fractionating the output stream from the isom- continuous operation) with the high separation
erization reactor through a continuous distillation performances of zeolites (due to their sieving prop-
column (a deisohexanizer or DIH) into three erties). Most of the published R&D work on that
efuents: topic was carried out at lab scale mainly by aca-
demic laboratories, though a few companies, like
A sidestream, mainly containing unconverted ExxonMobil or NGK Insulators, have also
normal hexane and the monobranched paraf- published results on that topic. One of the most
ns with six carbon atoms. This stream is studied topics in this research eld was the separa-
recycled to the isomerization unit. tion of normal short (C4C6) parafns from their
The bottom stream, containing the heaviest branched isomers through MFI-type zeolite mem-
alkanes (parafns) with six carbon atoms and branes (Arruebo et al. 2006; Bakker et al. 1996;
naphthenes (cyclic parafns) with six carbon Coronas et al. 1998). Indeed, MFI zeolites are
atoms, with an RON of 82, is sent directly to crystalline aluminosilicates with a microporous
the gasoline pool. structure that is composed of two intricate
Hydrogel 977
micropore networks: elliptical straight channels

with openings of 0.51 0.55 nm and zigzag Hydrogel
channels that are almost cylindrical with a diameter
of 0.53 0.56 nm, as measured by X-ray diffrac- Fiore Pasquale Nicoletta1, Giovanni De Filpo2
tion at ambient temperature (Flanigen et al. 1978). and Patrizia Formoso1
1
In such a conned porous system wherein the Dipartimento di Farmacia e Scienze della Salute
diameter of the micropores and the kinetic diame- e della Nutrizione, Universit della Calabria,
ter of the diffusing molecules are close, the higher Arcavacata di Rende, Italy
2
the kinetic diameter of a permeating molecule, the Dipartimento di Chimica e Tecnologie
higher the friction of the molecule alongside the Chimiche, Universit della Calabria, Arcavacata
micropore wall and therefore the lower its diffu- di Rende, Italy
sion coefcient inside the microporosity of the
MFI zeolite. Therefore, the diffusion coefcient
of normal alkanes in MFI zeolites is higher Hydrogel
(Courthial et al. 2008) than the diffusion coef-
cient of their monobranched isomers. Moreover, Hydrogels are cross-linked polymer systems able H
these materials prove to be hardly permeable to to retain large amounts of water by swelling. The
dibranched parafns. As an illustration, it was presence in the polymeric chain of polar func-
shown experimentally that a MFI zeolite mem- tional groups makes hydrogel hydrophilic sys-
brane operated under close to industrial operating tems; however, the cross-linking of the polymer
conditions (24 bar feed pressure, membrane tem- matrix makes them insoluble in water (Fig. 1).
perature between 200 C and 400 C) was able to The cross-linked structure can be achieved
produce a permeate composed of 95 % normal through two types of processes: chemical gelation
pentane and 5 % isopentane from a vapor feed and physical gelation. Gels with chemical bonds
composed of 20 % 2,2-dimethylbutane, 55 % occur through the formation of covalent bonds
isopentane, and only 25 % normal pentane among the structural units of the polymer chains
(Baudot and Bournay 2009). and cross-linking agents (Peppas et al. 2006). The
three main processes of chemical gelation are
addition polymerization, condensation, and vul-
References canization. In contrast, the polymer chains in
physical gels are connected via van der Waals
Arruebo M, Falconer JL, Noble RD (2006) Separation of forces, electrostatic attractions, and hydrogen
binary C-5 and C-6 hydrocarbon mixtures bonds. Pectin, cellulose, agarose, collagen, and
through MF1 zeolite membranes. J Membr Sci
269:171176
gelatin form physical gels.
Bakker WJW, Kapteijn F, Poppe J, Moulijn JA (1996) Per-
meation characteristics of a metal-supported silicalite-1
zeolite membrane. J Membr Sci 117:5778 Chemical and Physical Hydrogels
Baudot A, Bournay L (2009) Integration of MFI zeolite
membranes in the light gasoline isomerisation process.
Oil Gas Sci Technol 64:759771 Initiation, propagation, and termination are the
Coronas J, Noble RD, Falconer JL (1998) Separations of three steps of a polymerization process, i.e., the
C-4 and C-6 isomers in ZSM-5 tubular membranes. Ind reaction among monomer molecules to form poly-
Eng Chem Res 37:166176
Courthial L, Baudot A, Tayakout-Fayolle M, Jallut C
mer chains or three-dimensional networks. Dur-
(2008) A transient permeation-based method for coming the rst step of polymerization, a suitable
posite zeolite membrane characterization. AIChE J amount of initiators, R, form by light irradiation
54:25272538 or heat free radicals, R., which trigger the poly-
Flanigen EM, Bennett JM, Grose RW, Cohen JP, Patton
RL, Kirchner RM, Smith JV (1978) Silicalite, a new
merization reaction as schematically shown in the
:
hydrophobic crystalline silica molecular-sieve. Nature following equation: R hv=D
! R . These radicals can
271:512516 interact with a rst monomer, M, and lead to the
978 Hydrogel
Hydrogel, Fig. 1 (a) Structure of a hydrogel. Mc is the appearance of a hydrogel. The blue color is due to the
average molecular weight of chains between the cross- ability of hydrogels to absorb both water and chemical
links and x the average mesh size; (b) macroscopic compounds (in this particular case, methylene blue dye)
formation of monomers in propagation, RM., the polymer matrix. A polymer system initially
as shown in the following: R: M ! RM: . free of water is dened as xerogel; the presence of
Propagation consists in the growth of the hydrophilic groups and the degree of cross-
polymer chain by the addition of monomers linking establish the ability of the xerogel to
on the active site, as reported: store water. The greater the number of hydrophilic
RM . . . M: M ! RM . . . MM: . Termination groups, the greater will be the ability to absorb
can halt propagation by either a disproportion- water. The higher the degree of cross-linking, the
ation reaction or a combination reaction between lower the elasticity of the polymer network and
two radical species. the absorption of water.
There are different methods for the synthesis of
physical gels: heating and cooling of polymer
solutions (e.g., gelatin); ionic interactions (e.g., Swelling Parameters
alginate); coacervation complexes by the interac-
tion of polyanions with polycations, which form The degree of swelling is due to the equilibrium
soluble complexes or insoluble depending on their between the osmotic force, which favors the vol-
concentrations and pH; hydrogen bonds, which ume increase of polymer chains in the network,
lower the polymer solubility leading to the forma- and the elastic forces arising from physical or
tion of elastic hydrogels (e.g., carboxymethyl cel- covalent bonds present among the polymer
lulose); aging/maturation (e.g., arabic gum); and chains. The swelling ratio, Sw, is a measure of
freeze-thaw. Gels with chemical bonds are pro- how much water is retained by hydrogels and
vided by either polymerization of mixtures of can be expressed as either the ratio between the
monomers and cross-linkers or by cross-linking weight of swollen gel at equilibrium, Ws, and the
of preformed polymers. In the rst method, a weight of dry gel, Wd, i.e., Sw = Ws/Wd, or the
water solution of monomer, cross-linker, initiator, ratio between the volume of swollen gel at equi-
and catalyst is heated or irradiated to trigger the librium, Vs, and the volume of dry gel, Vd, i.e.,
polymerization reaction. In the second method, Sv = Vs/Vd. The percentage of swelling, %Sw, is
preformed hydrophilic polymers are cross-linked dened as %Sw = (Ws Wd)/Wd 100. As a
by either the use of cross-linking agents (e.g., function of their swelling behavior, hydrogels can
glutaraldehyde) or grafting, which functionalizes be classied into conventional hydrogels, able to
the polymer backbone with reactive species. Due absorb water when they are inserted in an
to their three-dimensional structure and cross- aqueous media, and their equilibrium swelling
linking centers, hydrogels swell rather than dis- is not affected by environmental parameters
solve due to the diffusion of water molecules into and stimuli-responsive hydrogels, polymeric
Hydrogel 979
networks which have a different equilibrium increasing the amount of cross-linker compared to
swelling in relation to environmental stimuli the quantity of monomer, there is a reduction of
such as changes in pH, temperature, ionic pore volumes and, consequently, of mesh sizes.
strength, and electromagnetic radiation. In fact,
every change in the water afnity of functional
groups, such as a protonation or deprotonation Applications
due to a pH variation or a thermal change of
hydration, leads to a reversible change in the According to the particular application, hydrogels
degree of swelling. For example, the can be prepared as bulk systems, sheets, mem-
pH-sensitive hydrogels are widely used in branes, and nanoparticles (Nicoletta et al. 2012).
the pharmaceutical eld for administering active There are several interesting applications of
drugs irritating to the gastric mucosa or to protect hydrogels in many elds including biochemical
an acid-labile substance from the stomach acid engineering, microelectronics, environmental sci-
pH, i.e., the formulation is in a shrunk state at ence, agro-food and packaging, regenerative med-
low pH and in a swollen state at high pH values. icine and drug delivery, and gas separation.
Hydrogels can withdraw solutes from aqueous Examples of applications are: H
solutions and can drop them in a controlled man-
ner (Fig. 1b). It is possible to dene the partition 1. Immobilization of enzymes. The immobiliza-
coefcient, K, as the ratio between the solute tion is the restriction of enzyme mobility in a
concentration in the gel (per unit volume of gel), xed space. This technique, compared to the
Cgel, and the solute concentration in the aqueous use of enzymes in solution, has the advantage
solution in equilibrium with the external gel, of making easier their recovery. The enzymes
Cbulk, i.e., K = Cgel/Cbulk. The changes in the can be synthesized and reused reducing costs
hydrogel structure are often related to the average of manufacture; it is possible to optimize the
mesh size, x, which can be estimated by the Flory- microenvironment in order to increase the
Rehner theory (Peppas and Merril 1977) enzymatic stability and the product quality.
according to the following equation: The main methods for the immobilization of
1=2 enzymes are adsorption (enzyme is linked on
1=3 2Cn Mc
x v2, s Mr l, where Cn is the Flory char-
the surface of the substrate by physical bonds),
acteristic ratio, l the length of the bond along the retention on the support surface by covalent
polymer backbone, Mr the molecular weight of bonds, and entrapment (the enzyme solution
the repeating units from which the polymer chain is mixed with the prepolymer solution before
is formed, and Mc is given by the polymerization process).

v=V

ln 1v , v
, w v 2
1 2, s
2. Drug delivery systems (DDS). A drug delivery
Mn
1 2 1 2 s 2 s
Mc 1=3 v2, s
, where n is
v2, s system is able to carry the drug in the body and
v2 , r v 2, r
2v2, r control its release maintaining the plasma con-
the specic volume of polymer; V1 is the solvent centration within the therapeutic window. With
molar volume; w1 is the polymer-solvent a DDS it is possible to administer lower doses
interaction parameter; v2,s and v2,r are the of drugs and limit side effects. As an example,
polymer volume fraction in the swollen and polyacrylamide hydrogels were used as a DDS
relaxed states, respectively; and Mn is the in the sustained release of insulin and delivery
molecular weight of the polymer chains prepared of other bioactive compounds. Molecular dif-
in the absence of cross-linking agent. The mesh fusion and carrier chemically controlled ero-
size can be controlled by several factors, sion are generally the major mechanisms of
mainly by the water quantity before cross-linking drug release.
and the amount and chain length of cross-linker. 3. Fillers in aesthetic and reconstructive surgery.
Both theory and experimental data show that by The use of minimally invasive techniques is
980 Hydrogel Capsules
now well known in plastic and cosmetic sur- References

gery. An ideal ller should be safe, effective,
biocompatible, nonimmunogenic, easily Di Proo G, Polino M, Nicoletta FP, Belviso BD,
Caliandro R, Fontananova E, De Filpo G, Curcio E,
obtainable, nonabsorbable, low-cost, and long
Drioli E (2014) Tailored hydrogel membranes for ef-
term storable, properties that are fullled by cient protein crystallization. Adv Funct Mater
many hydrogels, e.g., polyacrylamide is well 24:15821590
tolerated in the subcutaneous tissue and glan- Nicoletta FP, Cupelli D, Formoso P, De Filpo G, Colella V,
Gugliuzza A (2012) Light responsive polymer mem-
dular tissue in the breast. Another widespread
branes: a review. Membranes 2:134197
application of hydrogel is the fabrication of Peppas NA, Merril EW (1977) Crosslinked poly(vinyl
contact lens for the most common ametropias alcohol) hydrogels as swollen elastic networks. J Appl
or ocular drug release. Polym Sci 21:17631770
Peppas NA, Hilt JZ, Khademhosseini A, Langer R (2006)
4. Wound synthetic dressing. Hydrogel dressing is
Hydrogels in biology and medicine: from molecular
able to create a moist environment around the principles to bionanotechnology. Adv Mater
wounds and facilitate wound healing, such as 18:13451360
burns or necrotic wounds. Hydrogels can keep
clean wounds by removing necrotic and infected
tissues or by release of drugs. They are commer-
cially available in sheet forms, plasters, gauzes, Hydrogel Capsules
and gels that can be directly applied to wounds.
5. Membrane-supported hydrogels (Di Proo Emma Piacentini
et al. 2014), which combine the general advan- Institute on Membrane Technology, National
tages of crystallization in gels (reproducibility, Research Council of Italy, ITM-CNR, Rende,
size increase, and mechanical stability of crys- Italy
tals) with those of membrane-assisted crystalli-
zation (process control, optimal conditions,
reduced times, and specic targets). Capsules are particles consisting of an inner core
Membrane-supported hydrogels are excellent and a shell that covers and protects the core,
crystallization platforms as supersaturation is prepared using polymeric hydrogels. Hydrogels
generated by gradual solvent removal in vapor constitute a three-dimensional network structure
phase through the porous structure of both obtained from a class of synthetic and/or natural
membrane support and hydrogel layer. The ben- polymers which can absorb and retain signicant
ecial effect of the gels, made from mixtures of amount of water. Water holding capacity, perme-
acrylamide (monomer), poly(ethylene glycol ability, and biocompatibility are the most impor-
dimethacrylate) (cross-linker), N,N,N0 ,N- tant characteristic property required by the
0
-tetramethylethylenediamine (catalyst), and hydrogel. The cross-links between the different
photoinitiator, in crystallization processes by poly- polymer chains give a gel its structure (hardness)
propylene membranes was conrmed on two and elasticity and contribute to its stickiness. The
model proteins, hens egg white lysozyme and hydrogel chemistry can be modied by control-
concanavalin A. By varying the ratio between ling their polarity, surface properties, mechanical
monomer and cross-linker, it was possible to properties, and swelling behavior. Hydrogels can
change the cross-linking degree and, conse- also be stimuli sensitive and respond to surround-
quently, the pore size tuning in the most adequate ing environment like temperature, pH, and pres-
way the crystallization properties. Membrane- ence of electrolyte (Nho et al. 2005). They can
supported hydrogels were found to increase the display changes in their swelling behavior of the
efciency of the crystallization process: crystals network structure according to the external
were produced at lower protein concentrations environments.
and with enhanced diffraction properties than con- The main physical approaches used for
ventional techniques. hydrogel capsule manufacturing involve gelling
Hydrogel Capsules 981
Hydrogel Capsules, Table 1 Cross-linking methods to produce hydrogels and types of polymer
Cross-linking method Polymer
Physical cross-linking
Heating/cooling Gelatin, carrageenan
Helix-formation, association of the helices, and junction zones formation Hot
polymeric solution is cooled down
Ionic interaction Alginate, chitosan
Ionic polymers can be cross-linked by the addition of di- or trivalent counterions
Complex coacervation Polyanionic xanthan with
Complex coacervate gels can be formed by mixing of a polyanion with a polycation. polycationic chitosan
Polymers with opposite charges stick together and form soluble and insoluble Gum arabic with gelatin
complexes depending on the concentration and pH of the respective solutions
H-bonding Carboxymethyl cellulose
H-bonded hydrogel can be obtained by lowering the pH of aqueous solution of
polymers carrying carboxyl groups
Maturation (heat-induced aggregation) Gum arabic
Aggregation of the proteinaceous components, induced by heat treatment,
increases the molecular weight and subsequently produces a hydrogel form with H
enhanced mechanical properties and water binding capability
Freeze-thawing Polyvinyl alcohol
Microcrystals in the structure are formed due to freeze-thawing Xanthan
Chemical cross-linking
Chemical cross-linkers Polyvinyl alcohol
Cross-linked chains are obtained by introducing new molecules (such as Carboxymethyl cellulose
glutaraldehyde, epichlorohydrin, 1, 3-diaminopropane) between the polymeric chains
Grafting Starch
Polymer chains are activated by the action of chemical reagents, or high-energy Polyvinyl alcohol
radiation treatment. The growth of functional monomers on activated
macroradicals leads to branching and further to cross-linking
Radiation cross-linking
Free radicals are produced in the polymer following the exposure to the high- Carboxymethyl cellulose
energy source such as gamma ray, x-ray, or electron beam Chitin, chitosan, and derivatives
Starch and derivatives
Alginate
Hyaluronan and hyaluronic acid
the droplet phase of an emulsion or spray retaining the biocompatibility of the biopolymer.
(Shewan and Stokes 2013). Emulsication can Also, the modication and sterilization can be
be used for many combinations of biopolymers achieved in single step, and hence, it is a cost-
and oils. Cross-linked networks of synthetic and effective process to modify biopolymers having
natural polymers can be obtained by physical their end use specically in biomedical
cross-linking, chemical cross-linking, grafting application.
polymerization, and radiation cross-linking Hydrogel capsules of many synthetic and nat-
(Table 1). The physical cross-linking method is ural polymers have been produced with their end
receiving an increased interest due to relative use mainly in tissue engineering, pharmaceutical,
ease of production and the advantage of not and biomedical elds (Fernandez-Neives
using cross-linking agents. These agents affect et al. 2011; Lima et al. 2012). Due to their high
the integrity of substances to be entrapped (e.g., water absorption capacity and biocompatibility,
cell, proteins, etc.) as well as the need for their they have been used in wound dressing, drug
removal before application. Radiation cross- delivery, injectable polymeric systems, ophthal-
linking is a widely used technique since it does mic applications, and hybrid-type organs
not involve the use of chemical agents, therefore (encapsulated living cells).
982 Hydrogel Membranes
References variations and limited mechanical properties. In

contrast, hydrogel membranes by synthetic poly-
Fernandez-Neives A, Wyss H, Mattson J, Weitz DA mers can be prepared with precisely controlled
(2011) Microgel suspensions fundamentals and appli-
structures and functions. However, many syn-
cations. Wiley-VCH Verlag, Weinheim
Lima AC, Sher P, Mano JF (2012) Production methodolo- thetic polymers do not degrade in physiological
gies of polymeric and hydrogel particles for drug deliv- conditions and have to be puried from eventually
ery applications. Expert Opin Drug Deliv 9:231248 present toxic chemicals. Hydrogels are extremely
Shewan HM, Stokes JR (2013) Review of techniques to
useful in biomedical and pharmaceutical applica-
manufacture micro-hydrogel particles for the food
industry and their applications. J Food Eng tions for their high water content and rubbery
119:781792 nature, which make them similar to natural tissue
Nho Y.-C, Park S.-E, Kim H.-I, Hwang T.-S (2005) Oral and with a high degree of biocompatibility.
delivery of insulin using pHsensitive hydrogels based
Hydrogels can protect drugs from hostile environ-
on polyvinyl alcohol grafted with acrylic acid/
methacrylic acid by radiation. Nucl Instrum Meth B ments, such as enzymes and low pHs. Drugs
236:283288 loaded into the gel matrix can be released at a
predesigned rate in a desired site. The mechanism
of drug release from hydrogels is generally pas-
sive diffusion of molecules, which freely diffuse
Hydrogel Membranes out of the hydrogel matrix, but is possible to
control the drug release by changing the gel struc-
Fiore Pasquale Nicoletta1, Giovanni De Filpo2 ture in response to environmental stimuli, such as
and Patrizia Formoso1 pH, temperature, ionic strength, and electric/mag-
1
Dipartimento di Farmacia e Scienze della Salute netic eld.
e della Nutrizione, Universit della Calabria,
Arcavacata di Rende, Italy
2
Dipartimento di Chimica e Tecnologie Tissue Remediation
Chimiche, Universit della Calabria, Arcavacata
di Rende, Italy In addition to drug delivery applications, hydrogel
membranes are also relevant in tissue engineering
and remediation, i.e., for repairing and
Hydrogel Membranes regenerating a wide variety of tissues and organs.
Hydrogel dressing is a modern synthetic dressing
A membrane can be dened as a nanostructured/ of wounds. Its essential characteristics are the
functionalized system which separates two differ- rehydratation of tissues, the cleaning of wounds
ent phases and can control the exchange of chem- by removing necrotic and infected tissues, the
ical species and energy between them. The delivery of drugs, and the maintenance of a
transport is driven by the gradient of chemical moist environment in order to facilitate the
potential and the application of external forces, healing of wounds, such as burns, necrotic sores,
and the ow is regulated by uid and membrane and open and chronic wounds. Such dressings are
properties. Hydrogel membranes are selective available in a variety of forms and shapes, such as
barriers made by cross-linked polymers able to sheets, pads, and gels that can be directly applied
retain large amounts of water by swelling to wounds. Commercially available hydrogel
(Peppas et al. 2006). They allow the ow of cer- sheets do not need a secondary dressing and can
tain compounds (permeate) and retain some other be cut to t around the wounds. Hydrogel dress-
compounds (retentate) present in uid mixtures. ings do not stick to wounds and, consequently, can
Hydrogel membranes made by naturally derived be easily removed, causing less pain during
(biologic) molecules have obvious biocompatibil- changes and increasing the compliance of patients
ity, cell-controlled degradability, and intrinsic cel- and, in particular, of children. In the eld of tissue
lular interaction, but they may exhibit batch engineering, hydrogels are used as a scaffold
Hydrogel Membranes 983
material for the growth and replacement of vari- spiropyran, which can be entrapped, cross-linked,
ous tissues as they structurally resemble the extra- or introduced as a side chain or part of the main
cellular matrix of the human body. Other chain in polymer matrices (Nicoletta et al. 2012).
applications of hydrogels include tissue- Upon irradiation with UV-Vis light, the conforma-
engineered matrices for skin replacement and tional changes of the photoreactive groups result in
non-load-bearing bone tissue engineering, i.e., photoinduced effects on hydrogel matrix: pore
articial skin grafts, cartilage tissues, and tissue- average size changes as a consequence of expan-
engineered products. Hydrogel membranes can be sion and contraction of the photoreactive groups.
prepared with different mesh sizes in order to Some applications include the control of ion per-
modulate their transport properties such as ux meation through membrane gates (a kind of light-
and selectivity. Mesh size can be varied by chang- valved membranes that change the permeability of
ing material concentrations, and, in particular, the ions under UV irradiation) and light-controllable
amount of cross-linking agent. Thanks to its resis- microuidics. In fact, as hydrogel membranes
tance to protein adsorption, biocompatibility, and expand in response to environmental conditions
non-immunogenicity, poly(ethylene glycol), such as pH, temperature, photoirradiation, or sol-
PEG, plays a signicant role as a hydrophilic vent, they can facilitate the permeation of solutes, H
polymer for the preparation of hydrogel mem- while their shrinking can suppress permeation. In
branes used in regenerative medicine. PEG has a particular, hydrogel brushes cast on porous mem-
chemical structure which can be converted into branes can open or close pores in response to
other functional groups (methyloxyl, carboxyl, environmental stimuli. An interesting application
acetylene, amine, azide, vinyl sulfone, thiol, and is the polyacrylamide hydrogel containing
acrylate) for conjugation of biomolecules. Cross- bis-[4-{dimethylamino}phenyl]{4-vinyl-phenyl}
linked PEG hydrogels can be obtained by free methyl leucohydroxide, which has been studied as
radical photopolymerization of PEG acrylates, photoresponsive and ion-selective membrane.
Michael-type addition, native chemical ligation, Upon irradiation, the production of photocations
and enzymatic reaction. Among these approaches, increased the transport of anionic species, but had
the synthesis of PEG hydrogels through no effect on neutral species.
photopolymerization is the most common
approach which uses light to convert PEG
prepolymer solutions into cross-linked insoluble Membrane-Supported Hydrogels
hydrogels. The major type of macromers used for
photopolymerization includes PEG diacrylates, Crystallization is, today, a fundamental operation in
multiarm PEG acrylate, and PEG dimethacrylate. industry, especially pharmaceutics, health care, and
PEG hydrogels can be modied to be degradable biotechnology. Efciency, simplicity, exibility,
through incorporation of degradable segments, selectivity, permeability, compatibility, economy,
such as polyester, polypropylene fumarate, acetal, stability, eco-friendship, control, and scale-up are
and disulde. some characteristics of membrane operations,
which are important tools for the preparation, sepa-
ration, crystallization, and purication of chemicals.
Photoresponsive Hydrogel Membranes Membrane-assisted crystallization, or shortly mem-
brane crystallization, is an innovative technology,
The category of responsive membranes hosting which allows crystal nucleation and growth under
reactive functions includes several systems more uniform and predictable conditions favoring,
such as at and hollow ber membranes, in particular, compound purity, polymorphic selec-
nanocomposite layers, microcapsules, switchable tivity, crystal size, and process upscale. Membrane-
interfaces, core-shell structures, and free-standing assisted crystallization improves signicantly prod-
hydrogels. Typical photoreactive groups in poly- uct characteristics and enhances process efciency
mers are azobenzene, triphenylmethane, and with less waste, energy, and resources. Costs are
984 Hydrogel Membranes
Hydrogel Membranes,
Fig. 1 Scanning electron
microscope image of a
membrane-supported
hydrogel. A hydrogel layer
is polymerized on a porous
membrane of
polypropylene. Unit bar is
equal to 10 mm
reduced consequently. An interesting approach is optimization, reduced times, and specic targets)
represented by membranes in combination with (Di Proo et al. 2014) has been shown. Membrane-
functionalized layers of hydrogels. This strategy supported hydrogels were found to be excellent
includes the use of porous polymeric membranes crystallization platforms in which supersaturation
and their functionalization with a thin hydrogel is generated by gradual solvent removal in vapor
layer with an opportune mesh size (Peppas and phase through the porous structure of both
Merril 1977) and specic chemical composition. membrane support and hydrogel layer (Fig. 1). The
Cross-linked hydrogels have shown their ability to benecial effect of the gels, made from mixtures
concentrate solute molecules via thermodynamic of acrylamide (monomer), poly(ethylene
partitioning driven by favorable polymer-solute glycoldimethacrylate) (crosslinker), N,N,N0 ,
0
interactions. Gels generally give rise to crystals N -tetramethylethylenediamine (catalyst), and
with greater sizes, fewer defects, and enhanced dif- photoinitiator in membrane crystallization, was con-
fraction properties than crystals grown in solution rmed on two model proteins, chicken hen egg
because of the suppression of convection currents, white lysozyme and concanavalin A. By varying
sedimentation, and reduced random collisions the ratio between monomer and crosslinker, it was
between molecules. Nevertheless, the main draw- possible to change the cross-linking degree and,
back of gels is their handling difculty, which can consequently, the pore size tuning in the most ade-
be passed by using adequate supports such as a quate way with the crystallization properties.
membrane. In fact, earlier researches on Membrane-supported hydrogels were found to
membrane-supported hydrogels were mainly related increase the efciency of the crystallization process:
to gel-lled pore membranes with the aim of enhanc- crystals were produced at lower protein concentra-
ing mechanical stability, modulating transport of tions and with enhanced diffraction properties than
specic components as sketched in section conventional techniques.
Photoresponsive Hydrogel Membrane and reduc-
ing fouling. More recently, the possibility to fabri-
cate membrane-supported hydrogel displaying References
controlled chemical composition and nanostructure
to combine the general advantages of crystallization Di Proo G, Polino M, Nicoletta FP, Belviso BD,
Caliandro R, Fontananova E, De Filpo G, Curcio E,
in gels (reproducibility, size increase, and mechani-
Drioli E (2014) Tailored hydrogel membranes for ef-
cal stability of crystals) with those of membrane- cient protein crystallization. Adv Funct Mater
assisted crystallization (process control, extended 24:15821590
Hydrogel Nanoparticles 985
Nicoletta FP, Cupelli D, Formoso P, De Filpo G, Colella V, (e.g., adsorption, reaction rates, etc.) than hydro-
Gugliuzza A (2012) Light responsive polymer mem- gel membranes. Consequently, there has been a
branes: a review. Membranes 2:134197
Peppas NA, Merril EW (1977) Crosslinked poly(vinyl very rapid spread of nanoparticles in many elds
alcohol) hydrogels as swollen elastic networks. J Appl of daily life including llers, cosmetics, catalysts,
Polym Sci 21:17631770 pharmaceuticals, sensors, lubricants, electronic
Peppas NA, Hilt JZ, Khademhosseini A, Langer R (2006) devices, actuators, waste (dyes, heavy metals,
Hydrogels in biology and medicine: from molecular
principles to bionanotechnology. Adv Mater and organic pollutants) remediation, advanced
18:13451360 materials, regenerative medicine, and imaging
tools (Chidichimo et al. 2013). Stimulus-
responsive nanogels offer the opportunity to
develop systems for applications in particular
elds, such as drug delivery, biological sensing,
Hydrogel Nanoparticles and separation and purication technologies,
thanks to their ability to undergo reversible vol-
Fiore Pasquale Nicoletta1, Giovanni De Filpo2 ume phase transitions in response to environmen-
and Patrizia Formoso1 tal stimuli, such as pH, temperature, ionic H
1
Dipartimento di Farmacia e Scienze della Salute strength, solvent quality, enzymes, and
e della Nutrizione, Universit della Calabria, external electromagnetic elds (Fig. 1; Nicoletta
Arcavacata di Rende, Italy et al. 2012). Nanogels are obtained by chemical or
2
Dipartimento di Chimica e Tecnologie physical cross-linking of either synthetic or natu-
Chimiche, Universit della Calabria, Arcavacata ral polymers and can be swollen/shrunk in the
di Rende, Italy presence/absence of water or under the change
of environmental conditions. Nanogels can be
lled with molecules which can be released in
Hydrogel Nanoparticles opportune sites depending on the swelling/shrink-
ing properties of nanoparticles, the erosion/degra-
Hydrogel nanoparticles are nanometer-sized par- dation of polymer networks, and the reaction of
ticles composed of cross-linked hydrophilic poly- chemical groups to external stimuli. These prop-
mer networks (hydrogels) (Lyon and Serpe 2012; erties promote nanogels as the major carrier of
Peppas et al. 2006). They are also known as drugs and contrast agents in human body. Nowa-
nanogels, where the term nano refers to materials, days researchers are interested in the synthesis of
one dimension of which is less than 100 nm. The new hydrogel nanoparticles for applications as
very low dimensions of nanomaterials confer drug delivery systems, thermosensitive capsules,
them physicalchemical properties which can dif- and carriers of agents for therapy and diagnostics.
fer from those of the same bulk materials. As all The main drawbacks for the use of hydrogels in
other nanoparticles, nanogels are characterized by many applications are the gel shrinking/swelling
a high surface/volume ratio which gives them speed and mechanical strength. The formation of a
larger activities in surface-related phenomena skin layer at the very initial stage of gel shrinking
Hydrogel Nanoparticles,
Fig. 1 Volume phase
transitions in nanogels in
response to an external
stimulus
986 Hydrogel Nanoparticles
lowers shrinking rate as it reduces the water dif- mixed to the oil phase and surfactant to give
fusion from inside the gel to the external environ- thermodynamically stable micelles. Polymeriza-
ment. Possible solutions to these drawbacks are tion is initiated by initiators either in the aqueous
the synthesis of new monomers, the modication phase or in the oil phase. In both oil-in-water and
of polymer chains (e.g., by grafting), and the use water-in-oil emulsion polymerization, the size of
of additives (plasticizers, llers, chemical cata- the nanogels can be controlled by varying the
lysts, retarders, surfactants, etc.) in monomer amount of surfactant. Smaller nanogels can be
solutions. In addition to nontoxicity, the response obtained by higher surfactant concentrations.
of a tissue to gel material is an important item that However, surfactants used in emulsion polymeri-
has to be tested carefully. In fact, different tissues zation have to be removed at the end of the poly-
(subcutis, arteries, tendons, muscles, nerves, etc.) merization reaction by dialysis or desorption, with
can respond in a different way to the contact with possible coagulation or occulation of
polymer and induce its degradation with time. nanoparticles. A way to solve this problem is the
Other goals in nanogel research are the integration use of surfactant-free emulsion polymerization.
of multiple functions, the sensitivity to different Noncovalent interactions, such as hydrogen
stimuli, and a better control of responses. Hybrid bonds, van der Waals forces, and electrostatic
hydrogel nanoparticles are nanometric particles and hydrophobic interactions, allow preparing
with their surface properties modied by cross- nanogels. A physical self-assembly of interactive
linked hydrophilic polymer networks (hydrogels) polymers consists in amphiphilic block copoly-
in order to increase efciency and selectivity mers that self-assemble into micelles via
toward particular functions or stealth nature of noncovalent interactions. During the self-
nanoparticles to the reticuloendothelial system or assembly process, small active molecules can be
macrophage system. incorporated into the nanostructures. Nanoparticle
size can be tailored by changing the block nature,
block lengths, and copolymer composition. In
Chemically and Physically Cross-Linked addition, the change of other parameters, such as
Nanogels temperature, solvent, and additives, allows tuning
the properties of aggregates. In particular, surface
Nanogels can be classied as either chemically modications and core functionalities of nanogels
cross-linked nanogels or physically cross-linked can allow the recognition of specic cells and an
nanogels. Chemically cross-linked nanogels are improved drug encapsulation. However, it is dif-
formed by cross-linking of polymers via covalent cult to obtain stable physically cross-linked
bonding. In contrast, noncovalent bonds (i.e., nanogels with controlled sizes using noncovalent
hydrogen bonds, ionic bonds, and hydrophobic interactions because the bonds are
interactions) link polymers in physically cross- relatively weak.
linked nanogels. The most common approaches
for the preparation of responsive nanogels are
(i) emulsion polymerization and (ii) physical Responsive Nanogels
self-assembly of polymers. Emulsion polymeriza-
tion generally allows controlling size, stability, A responsive nanogel is a nanoscopic three-
and surface properties of nanogels. It involves dimensional gel network consisting of cross-
the emulsication of hydrophobic/hydrophilic linked macromolecular chains that can undergo
monomers in water by an oil-in-water/water-in- volume phase transitions in response to an exter-
oil emulsier and the initiation of reaction with nal stimulus. The thermally induced swelling or
either a water- or oil-soluble initiator (oil/water or shrinking of a nanogel is caused by the thermally
water/oil emulsions). Water-in-oil emulsion poly- driven coil-to-globule transition of polymer
merization is the best technique for water-soluble chains (conformational change) between two
monomers: an aqueous solution of monomers is neighboring cross-linking points inside the gel
Hydrogen from Bioethanol 987
network. Thermally sensitive charged nanogels technology for remediation. In: Piemonte V, De Falco
can be used in protein adsorption and separation BA (eds) Sustainable development in chemical engi-
neering: innovative technologies. Blackwell Science,
as they can form with protein complexes with a Oxford, pp 297348
coreshell structure characterized by different Lyon LA, Serpe MJ (2012) Hydrogel micro and
properties. Interesting monomers for the synthesis nanoparticles. Wiley-VCH, Weinheim
of thermally responsive nanogels are those with a Nicoletta FP, Cupelli D, Formoso P, De Filpo G, Colella V,
Gugliuzza A (2012) Light responsive polymer mem-
lower critical solution temperature in water branes: a review. Membranes 2:134197
around that of human body for their obvious Peppas NA, Hilt JZ, Khademhosseini A, Langer R (2006)
applications as drug delivery systems. Poly Hydrogels in biology and medicine: from molecular
(N-vinylcaprolactam)-based gels have a volume principles to bionanotechnology. Adv Mater
18:13451360
phase transition temperature around 3334 C,
which can be modied by copolymerization of
N-vinylcaprolactam with other monomers. In fact,
copolymerization changes the hydrophilic/hydro-
phobic balance of the resulting polymer and con- Hydrogen from Bioethanol
sequently its critical solution temperature. In H
addition, copolymerization of N-vinylcaprolactam Jose M. Sousa
with monomers/cross-linkers allows tuning the Chemistry Department, School of Life and
chemical structure, size, charge, and swelling Environment Sciences, University of
degree of nanoparticles. The biocompatibility and Tras-os-Montes e Alto Douro, Vila Real,
solubility in water solutions make poly Portugal
(N-vinylcaprolactam)-based gels suitable for bio- LEPABE Laboratory for Process Engineering,
medical applications. Unlike the densely packed Environment, Biotechnology and Energy,
nanocarriers usually used for drug delivery sys- Chemical Engineering Department, Faculty of
tems, nanogels can bind within their networks a Engineering, University of Porto, Rua Roberto
large amount of water molecules and bioactive Frias, Porto, Portugal
compounds including drugs, proteins, and DNA
in a relatively stable manner. These compounds
can be released by the gel structure changes occur- The production of hydrogen from bioethanol has
ring during the volume phase transition. been considered an attractive way for exploring
Alkali-swellable nanogels containing carbox- sustainable renewable energy sources, from an
ylic acids display extreme pH sensitivity and rhe- environmentally friendly point of view.
ological changes in water. In fact, the Bioethanol consists of an aqueous solution
neutralization of such aqueous dispersion causes containing 812 wt.% of ethanol, besides other
the swelling of the nanogels with a consequent by-products such as glycerol, acetaldehyde,
dramatic increase in viscosity and shear thinning diethyl ether, methanol, etc. (Ni et al. 2007;
of the dispersion. On increasing pH, the swelling Iulianelli and Basile 2011).
increases due to the ionizing of carboxylic acid Hydrogen production from ethanol is essen-
groups. In addition, the increase in ionic strength tially carried out by steam reforming, according
screens the internal repulsion or reduces the to the main reaction described by C2 H5 OH 3H2
osmotic pressure difference between inside and O ! 6H2 2CO2 , DH 298 K 348 kJ mol1 .
outside of the nanogel particles. Ethanol steam reforming is a very attractive way
to locally produce hydrogen, comparatively to
other fuels such as methanol, glycerol, acetic
References acid, diethyl ether, etc. For example, ethanol is
easily obtained by fermentation from renewable
Chidichimo G, Cupelli D, De Filpo G, Formoso P, sources such as sugars and starches (e.g., sugar-
Nicoletta FP (2013) Nanoparticles as a smart cane and corn rst-generation bioethanol) and
988
CH3CH2OH H2O CO2 CO2
O O O O O O O O O O
O O O O Co O O
Co Co
ce Support (metal oxide) Support (metal oxide) O Support (metal oxide) O

rfa
c su scheme 1 scheme 9 scheme 13
idi
Ac
H2 O O O O O
C C CO2
H
CH3CH2O
C2H4 2 H2O H
O O OH OH O O O O O O O O
O O O O Co O O
Co Co
H H H H
Co O Support (metal oxide) O
Support (metal oxide) Support (metal oxide)
scheme 3 scheme 8 scheme 12
Support (metal oxide)
scheme 2 CH CHO
3
Route 1
CH4CHO CH3COCH3
H2 O O
HC
CH3CH2O
H H
Route 2
CH3CH O
OH OH OH OH OH + O O O O O
O O O O O Co O O
Co CH4 CO H2
Deactivation O O Support (metal oxide) O
Cx Cx C Support (metal oxide)
Cx
OC
x scheme 4 OH H3C OH + OH scheme 11
x
OH OH OH O O
x
Cx
O Co
C C
O
Co
H2 CO
Support (metal oxide) O H2
Support (metal oxide) O
H3CC
H2 scheme 7 O
scheme 6 H H-CHx
OC
OH O O OH OH OH OH OH + OH
O O O O O Co O O
Deficient oxygen Co
or Co site accessiblity Sufficient oxygen accessiblity
Support (metal oxide) O Support (metal oxide) O
scheme 5 scheme 10
Hydrogen from Bioethanol, Fig. 1 Proposed reaction mechanism for ethanol steam reforming over supported Co catalysts (Song 2012)
Hydrogen from Bioethanol
Hydrogen from Bioethanol 989
from lignocelluloses (agricultural, industrial, and dehydrogenation. In addition to H2 and CO2,

forest residues second-generation bioethanol), reformate stream may contain also CO, methane,
among other sources; ethanol is easy to transport aldehydes, ketones, ethylene, ethane, and high
and store, it is biodegradable and shows low tox- alcohols, among others (Vizcano et al. 2007;
icity; ethanol is relatively easy to dehydrogenate Song 2012) (Fig. 1).
by steam reforming, and it does not contain cata- The main drawback of using bioethanol to
lyst poison such as sulfur (Ni et al. 2007). produce hydrogen via steam reforming is the
Steam reforming of ethanol is a highly endo- high cost associated to the downstream distillation
thermic reaction, which limits its industrial appli- and purication steps of the crude ethanol
cation for hydrogen production. The oxidative obtained from fermentation. Feeding directly the
reforming, oxidation of a fraction of ethanol to crude bioethanol to the reformer would reduce
provide part of the energetic needs, is a possible drastically the costs of the produced hydrogen.
way to minimize such impact. If the amount of Besides the unnecessary expensive distillation
oxygen is sufcient for balancing the reforming process for water and other compounds elimina-
enthalpy needs, the process is named autothermal tion, the reforming of other oxygenated hydrocar-
reforming (Song 2012). Other ways to produce bons contained in the fermentation broth could H
hydrogen from ethanol are fermentation processes contribute to generate extra hydrogen. The main
using metabolically engineered microorganisms, challenge for the implementation of this approach
solar photocatalytic processes using suitable at an industrial level remains in the tolerance and
semiconductors, CO2 dry reforming, plasma stability of the catalyst to the impurities present in
reforming, partial oxidation, and aqueous phase the crude ethanol solution, especially high linear
reforming (Song 2012). and branched alcohols (Le Valant et al. 2011;
Concerning the catalysts for the hydrogen pro- Song 2012).
duction from (bio)ethanol, current state of the art In the viewpoint of hydrogen production for
includes the noble metal-based (Pt, Pd, Ru, Rh) supplying polymer electrolyte membrane fuel
and non-noble metal-based (Cu, Co, Zn, Ni) ones cells (PEMFCs), the reformate stream, which
(Iulianelli and Basile 2011). The ethanol conver- comprises a complex mixture of compounds,
sion and hydrogen yield and selectivity are needs a separation/purication, especially due to
strongly dependent on the catalyst type, its sup- the maximum allowed CO concentration
port (e.g., ZnO, MgO, Al2O3, SiO2, La2O3, (0.2 ppm). A conventional steam reformer system
CeO2), and the reaction conditions (Costa-Serra is composed by the reformer (xed bed reactor
et al. 2010; Song 2012). (FBR)), two water gas shift reactors (high- and
Besides the ethanol steam reforming main low-temperature WGS), a CO partial oxidation
reaction referred above, the process follows a reactor (PROX), and pressure swing adsorption
complex reaction system with several possible (PSA) units (Fig. 2).
consecutive parallel reactions, such as partial This complex process may be replaced by a
reforming to CO, water gas shift, methanation, much simpler membrane reactor (MR) holding
coke formation from intermediate products, hydrogen permselective membranes. This new
Boudouard reaction, CO reduction, methane reactor is able to perform both the steam
cracking, dehydration/hydrogenation, and reforming of bioethanol and the separation/
Water Air
Feed H2
FBR HT WGS LT WGS CO PROX H2Purification
(PSA)
Cooler Cooler Cooler
Hydrogen from Bioethanol, Fig. 2 High-purity hydrogen production in a conventional multistage system
990 Hydrogen Permeation
Hydrogen from
Bioethanol, Fig. 3 High-
purity hydrogen production
in a membrane reactor with
a Pd-based membrane
purication of the produced hydrogen in the same

device (Fig. 3). Hydrogen Permeation
Moreover, this kind of MR makes possible the
in situ removal of hydrogen from the reaction side, H2 Permeation Through Pd-Based Membranes
allowing the conversion to overcome the thermo-
dynamic equilibrium value (which is not possible
in the FBR). Furthermore, if Pd or Pd-based mem-
branes are used, a pure hydrogen stream is col- Hydrogen Production by Membrane
lected in the permeate side, suitable for direct Reactors
PEMFC supplying (Iulianelli and Basile 2011).
Giuseppe Barbieri
References Research Council of Italy, ITM-CNR, Rende,
Italy
Costa-Serra JF, Guil-Lpez R, Chica A (2010) Co/ZnO and
Ni/ZnO catalysts for hydrogen production by
bioethanol steam reforming. Inuence of ZnO support
morphology on the catalytic properties of Co and Ni Hydrogen, the energetic vector of the future, is
active phases. Int J Hydrog Energy 35:67096716 expected to become more and more important
Iulianelli A, Basile A (2011) Hydrogen production from
ethanol via inorganic membrane reactors technology: a according to leading energy scenarios as also
review. Catal Sci Technol 1:366379 reported in the EU-White Paper (http://ec.europa.
Le Valant A, Garron A, Bion N, Duprez D, Epron F (2011) eu/research/energy/pdf/efchp_hydrogen13.pdf) on
Effect of higher alcohols on the performances of a 1 % renewable and alternative energy. Moreover, BP
Rh/MgAl2O4/Al2O3 catalyst for hydrogen production
by crude bioethanol steam reforming. Int J Hydrog and GE announced plans to develop jointly up to
Energy 36:311318 15 new hydrogen power plants to generate elec-
Ni M, Leung DYC, Leung MKH (2007) A review on tricity over the coming decade (http://www.bp.
reforming bio-ethanol for hydrogen production. Int com/content/dam/bp/pdf/investors/BP_Annual_
J Hydrog Energy 32:32383247
Song H (2012) Catalytic hydrogen production from Report_and_Form_20F_2014.pdf). The hydro-
bioethanol. In: Lima MAP, Natalense APP (eds) gen required is planned to be derived from fossil
Bioethanol. InTech Publishing, Janeza Trdine 9, fuels, including coal and natural gas, while devel-
Rijeka, Croatia oping, in the meantime, new plants equipped for
Vizcano AJ, Carrero A, Calles JA (2007) Hydrogen pro-
duction by ethanol steam reforming over CuNi carbon capture technologies capable of reducing
supported catalysts. Int J Hydrog Energy 32:14501461 CO2 emissions by 90 % (http://www.worldwatch.
org/node/4516). If one considers that, at present,
Further Reading 96 % hydrogen is directly produced from fossil
Curcio S (2011) Membranes for advanced biofuels produc- fuels and ca. 4 % indirectly by using electricity
tion. In: Basile A, Nunes S (eds) Advanced membrane generated through them (Kothari et al. 2008), the
science and technology for sustainable energy and
environmental applications. Woodhead Publishing need to nd new efcient and competitive tech-
Limited, Cambridge, UK nologies able to maximize hydrogen production
Hydrogen Production by Membrane Reactors 991
from fossil fuels while lowering CO2emissions

CO H2 O CO2 H2 DH298
and other pollutants appears evident.
In addition, up to now, steam is utilized for most 41 kJ mol1
H2 obtained industrially from natural gas or light
hydrocarbons. Conventional technologies can also CH4 2H2 O CO2 4H2 DH298
operate with alternative fuel sources such as, for
instance, bioalcohols and other biomasses, having 165 kJ mol1
attracted great attention owing to their being
renewable sources (Abanades et al. 2013; Vizcano High temperature and low pressure thermody-
et al. 2012; Sanz et al. 2014; Iulianelli et al. 2014). namically favor the process because the reaction
Along with these technologies,which can be inte- is endothermic and also characterized by an
grated with CO2 separation units, membrane reac- increase in the number of moles. This reaction is
tion and separation units are becoming more and thermodynamically limited and the selective
more promising and signicant in hydrogen pro- hydrogen removal will increase its conversion.
duction. Membrane reactors couple conversion The development of ceramic and supported
and separation in the same operating unit and this metallic (e.g., palladium-based) membranes has H
leads to a conversion higher than the thermody- promoted the investigation of the membrane
namic one of conventional reactors, a lower reformer for a high-temperature reaction. Dense
reaction volume, a pure (permeate) hydrogen palladiumbased membranes are characterized by
stream, and a concentrate in CO2 (retentate) stream. innite hydrogen selectivity which prevents the
The permeate hydrogen stream has a pure grade permeation of other species through the metallic
that can be directly fed without any further treat- structure of the palladium.
ment to a polymeric electrolyte membrane fuel cell The operating conditions (temperature,
and no performance loss is observed (Brunetti pressure, ow rate, composition, etc.) have to be
et al. 2008). completely rethought since the permeation
Thus, the new model reactor has several advan- signicantly affects the reforming process/
tages with respect to conventional technology, reaction.
i.e., the production of a new carrier and CO2 Methane steam reforming can be operated at a
capture, since pure hydrogen and CO2 concen- signicantly lower temperature (400500 C)
trated streams exit this reactor. since the reaction pressure favors hydrogen per-
Membrane reactors can operate in various meation. The higher the feed pressure, the higher
sections of the hydrogen production plants: the hydrogen removal and as a consequence the
replacing both steam reforming (SR) as well as methane conversion (see Fig. 1).
high- and low-temperature water gas shift (WGS) The stream produced by reformers and/or coal
reactors. gasication plants contains around 50 % molar of
Steam reforming is the reaction between hydrogen (on a dry basis) and between 40 % and
hydrocarbons and water vapor. The steam 45 % molar of CO, which is usually reduced in an
reforming of methane is the best known of this upgrading stage producing more hydrogen at the
reaction class owing to its industrial importance in same time, by WGS reaction. The upgrading stage
hydrogen and syngas production (Scholz 1993). of traditional processes consists of a multistage
A methane steam reformer is industrially operated CO-shift process based on two catalytic reactors:
at 7001100 C and 2030 bar over Ni-based the rst operates at high temperatures (about
catalyst; the main species and reactions involved 350400 C) to take advantage of the faster reac-
are (Xu and Froment 1989a, b) tion rates, whereas the other operates at low tem-
peratures (around 220300 C) to rene the
carbon monoxide conversion, thus allowing a
CH4 H2 O CO 3 H2 DH298
lower nal CO concentration (less than 1 %
206 kJ mol1 molar) (Raggio et al. 2005).
992 Hydrogen Production by Membrane Reactors
1 thermodynamics constraints on the reaction

(Barbieri et al. 2008; Dittmeyer et al. 2001; Mendes
et al. 2010, 2011; Paglieri and Way 2002; Pinacci
r
ba
MSR
et al. 2010). Only few works in the literature have
10
ar
explored the combination of a high-temperature
5b
ar
X CH (-)
WGS catalyst with Pd-based membranes
1b
4
r
0.5 (Augustine et al. 2011; Bi et al. 2009; Brunetti
ba
1
r et al. 2012, 2015; Catalano et al. 2013; Chein
ba
5
r et al. 2013; Iyoha et al. 2007; Pinacci et al. 2010;
I =10 1 0b
a
Uemiya et al. 1991). An interesting review on the
recent advances of this technology for high-
I = 0 (TR) temperature applications is proposed by Cornaglia
0 et al. (2015).
200 400 600 Rather than the limitations of thermodynam-
T (C) ics, fast kinetics and promoted permeation are
the main advantages offered by the high temper-
Hydrogen Production by Membrane Reactors, ature. However, a fundamental role is played by
Fig. 1 Equilibrium conversion of Methane as function of the characteristics of the membrane used in the
the temperature for both membrane and traditional reactors
reactor. In fact, until now one of the main hur-
at differrent feed pressure values for H2O/CH4 initial ratio
of 3. For membrane reactor, the permeate pressure is 1 bar dles limiting the large-scale development of
and the sweep contained in the permeate chamber is the these MRs is the high cost of Pd-based mem-
times higher than the methane amount in the reaction branes, partly owing to the cost of palladium
volume. (Marigliano et al. 2001, 2003)
that could further increase if Pd membranes are
exploited on a large industrial scale (Helmi
CO H2 O CO2 H2 et al. 2014). For this reason, great efforts are
being made to obtain thin Pd layers on appro-
DH0 298 41 kJ mol1 priate supports in order to maximize the perme-
ate ux, minimize the support inuence, and
The H2-rich stream coming out from the last reac- reduce the cost related to Pd (PachecoTanaka
tor is fed to a pressure swing adsorption (PSA) et al. 2005; Zhang et al. 2009).
unit for H2 separation from other gases. Most
often another reaction unit is added for oxidizing
CO to CO2, to meet the purity targets for fuel cell References
uses (CO concentration lower than 1020 ppm).
Additionally, the reduction of the percentage of Abanades A, Rubbia C, Salmieri D (2013) Thermal crack-
CO in the system reduces coking of metal catalyst ing of methane into hydrogen for a CO2-free utilization
surfaces, implying a longer lifetime. This means of natural gas. Int J Hydrog Energy 38(20):84918496
Augustine AS, Ma YH, Kazantzis NK (2011) High pres-
an intensied process with a reduced plant size sure palladium membrane reactor for the high temper-
and a higher yield. In some cases, the interesting ature watergas shift reaction. Int J Hydrog Energy
results achieved at laboratory level thanks to this 36:53505360
effect have even led to several patents (Deckman Barbieri G, Brunetti A, Tricoli G, Drioli E (2008) An
innovative conguration of a Pd-based membrane reac-
et al. 2005; Gummalla et al. 2010; Lamm tor for production of pure hydrogen. Experimental
et al. 2005; Tsay et al. 2007; Wei 2009; Willms analysis of water gas shift. J Power Sources
and Birdsell 2000). 182:160167
However, most of the studies on WGS in MRs Bi Y, Xu H, Li W, Goldbach A (2009) Water-gas shift
reaction in a Pd membrane reactor over Pt/Ce0.6Zr0.4O2
were carried out in a low and medium temperature catalyst. Int J Hydrog Energy 34:29652971
range (180250 C and 250320 C, respectively) Brunetti A, Barbieri G, Drioli E (2008) A PEM-FC and H2
by using CuO-based catalysts, owing to the membrane purication integrated plant. Chem Eng
Hydrogen Production by Membrane Reactors 993
Process: Process Intesif 47(7):10811089. methane steam reforming process. Catal

doi:10.1016/j.cep.2007.03.015, special issue Today 67(13):8599. doi:10.1016/S0920-5861(01)
Euromembrane 2006 00268-1
Brunetti A, Drioli E, Barbieri G (2012) Medium/high tem- Marigliano G, Barbieri G, Drioli E (2003) Equilibrium
perature water gas shift reaction in a Pd-Ag membrane conversion for a palladium membrane reactor. Depen-
reactor: an experimental investigation. RSC Adv dence of the temperature and pressure, Chem. Eng. and
2(1):226233 Processing, 42, 231236; DOI:10.1016/S0255-2701
Brunetti A, Caravella A, Fernandez E, Pacheco Tanaka (02)00092-2.
DA, Gallucci F, Drioli E, Curcio E, Viviente JL, Mendes D, Chibante V, Zheng J-M, Tosti S, Borgognoni F,
Barbieri G (2015) Syngas upgrading in a membrane Mendes A, Madeira LM (2010) Enhancing the produc-
reactor with thin Pd-alloy supported membrane. Int tion of hydrogen via watergas shift reaction using
J Hydrog Energy 40(34):1088310893. doi:10.1016/j. Pd-based membrane reactors. Int J Hydrog Energy
ijhydene.2015.07.002 35:1259612608
Catalano J, Guazzone F, Mardilovich IP, Kazantzis NK, Mendes D, S S, Tosti S, Sousa JM, Madeira LM, Mendes
Ma YH (2013) Hydrogen production in a large scale A (2011) Experimental and modeling studies on the
water gas shift Pd based catalytic membrane reactor. low-temperature water-gas shift reaction in a dense
Ind Eng Chem Res 52:10421055 PdAg packed-bed membrane reactor. Chem Eng Sci
Chein R, Chen Y-C, Chung JN (2013) Parametric study of 66:23562367
membrane reactors for hydrogen production via high-
temperature water gas shift reaction. Int J Hydrog
PachecoTanaka DA, LlosaTanco MA, Niwa S, Wakui Y,
Mizukami F, Nambra T et al (2005) Preparation of
H
Energy 38:22922305 palladium and silver alloy membrane on a porous
Cornaglia L, Munera J, Lombardo E (2015) Recent a-alumina tube via simultaneous electroless plating.
advances in catalysts, palladium alloys and high tem- J Membr Sci 247(1e2):2127
perature WGS membrane reactors. A review. Int Paglieri SN, Way JD (2002) Innovations in palladium
J Hydrog Energy 40:34233437 membrane research. Sep Purif Methods 31:1169
Deckman HW, Fulton JW, Grenda JM, Hershkowitz Pinacci P, Broglia M, Valli C, Capannelli G, Comite
F (2005) Electric power generation with heat A (2010) Evaluation of the water gas shift reaction in
exchanged membrane reactor. EP Patent 1,294,637 a palladium membrane reactor. Catal Today
Dittmeyer R, Hollein V, Daub K (2001) Membrane reac- 156:165172
tors for hydrogenation and dehydrogenation processes Raggio G, Pettinau A, Orsini A, Fadda M, Cocco D,
based on supported palladium. J Mol Catal A Chem Deiana P, Pelizza ML, Marenco M, Second interna-
173:135184 tional conference on clean coal technologies for our
Gummalla M, Vanderspurt TH, She Y, Dardas Z, future, Castiadas, Sardinia, Italy, 1012 May 2005
Olsommer B (2010) Power plant with membrane Sanz R, Calles JA, Alique D, Furones L (2014) H2 produc-
water gas shift reactor system. US Patent tion via water gas shift in a composite Pd membrane
20,100,104,903 reactor prepared by the pore-plating Method. Int
Helmi A, Gallucci F, van SintAnnaland M (2014) Resource J Hydrog Energy 39:47394748
scarcity in palladium membrane applications for carbon Scholz WH (1993) Process for industrial of hydrogen and
capture in integrated gasication combined cycle units. associated environmental effects. Gas Sep Purif
Int J Hydrog Energy 39:1049810506 7(3):131139
Iulianelli A, Pirola C, Comazzi A, Galli F, Manenti F, Tsay D, Weiss SE, Tsay TC, (2007) Single stage membrane
Basile A (2014) Chapter 1: Water gas shift membrane reactor for high purity hydrogen production. US Patent
reactors. In: Basile A, Hai F, Di Paola L, Piemonte 20,070,157,517
V (eds) Membrane reactors for energy applications Uemiya S, Sato N, Ando H, Kikuchi E (1991) The water
and basic chemical production. Woodhead (in press) gas shift reaction assisted by a palladium membrane
Langford Lane, Kidlington, OX5 1GB, UK reactor. Ind Eng Chem Res 30:589591
Iyoha O, Enick R, Killmeyer R, Howard B, Ciocco M, Vizcano AJ, Lindo M, Carrero A, Calles JA (2012) Hydro-
Morreale B (2007) H2 production from simulated coal gen production by steam reforming of ethanol using
syngas containing H2S in multi-tubular Pd and 80 wt% Ni catalysts based on ternary mixed oxides prepared
Pd-20 wt% Cu membrane reactors at 1173 K. J Membr by coprecipitation. Int J Hydrog Energy
Sci 306:103115 37(2):19851992
Kothari R, Buddhi D, Sawhney RL (2008) Comparison of Wei W (2009) Polygeneration systems. US Patent
environmental and economic aspects of various hydro- 20,090,084,035
gen production methods. Ren Sust En Rev 12:553563 Willms RS, Birdsell SA (2000) Apparatus and method for
Lamm A, Poschmann T, Schaefer J (2005) Apparatus for simultaneous recovery of hydrogen from water and
generating virtually pure hydrogen for fuel cells. US from hydrocarbons. US Patent 6,165,438
Patent 20,050,039,401 Xu J, Froment GF (1989a) Methane steam reforming,
Marigliano G, Barbieri G, Drioli E (2001) Effect of energy methanation and water gas shift: I. Intrinsic kinetics.
transport in a palladium based membrane reactor for AIChE J 35:8896
994 Hydrogen Production by Water Splitting
Xu J, Froment GF (1989b) Methane steam reforming: hydrogen economy (Fujishima and Honda
II. Diffusional limitations and reactor simulation. 1972). Presently, the solar-to-hydrogen energy
AIChE J 35:97103
Zhang K, Gao H, Rui Z, Liu P, Li Y, Lin YS (2009) High- conversion efciency is too low for the technol-
temperature stability of palladium membranes on ogy to be economically sound. The main barriers
porous metal supports with different intermediate are the rapid recombination of photo-generated
layers. Ind Eng Chem Res 48:18801886 electron/hole pairs and the backward reactions
and the poor activation of TiO2 by visible light
(only 35 % of photons present in the solar light
can photoactivate it). In response to these de-
Hydrogen Production by Water ciencies, some investigators studied the effects
Splitting of addition of sacricial reagents and carbonate
salts to prohibit rapid recombination of electron/
Raffaele Molinari hole pairs and backward reactions. Other research
Dipartimento di Ingegneria per lAmbiente e il focused on the enhancement of photocatalysis by
Territorio e Ingegneria Chimica (DIATIC), modication of TiO2 by means of metal loading,
Universit della Calabria, Rende, Italy metal ion doping, dye sensitization (ORegan and
Grtzel 1991), composite semiconductor, anion
doping, and metal ion implantation. Metal ion
Hydrogen is a promising energy vector in many implantation and dye sensitization are very
industries such as metallurgical industry, thermal effective methods to extend the activating
treatments, petrochemical industry, food industry, spectrum to the visible range. Therefore, they
electrical energy production, electronic industry, play an important role in the development of
and mechanical industry because of the zero efcient photocatalytic hydrogen production
environmental impact of the combustion products (Ni et al. 2007). The metal ion-implanted TiO2
(only water). Main sources of hydrogen are fossil could efciently work as a photocatalyst under
(reforming of petroleum, reforming of natural gas, visible light irradiation. Some eld tests under
carbon gasication) and nuclear thermochemical solar light irradiation clearly revealed that the Cr
and renewable processes (electrolysis from eolic ion- or V ion-implanted TiO2 samples showed 23
and photovoltaic electrical energy, solar thermo- times higher photocatalytic reactivity than the
chemical processes, biomass gasication). Steam un-implanted TiO2. Instead, a visible light-
reforming of hydrocarbons is the dominant tech- responsive TiO2 thin-lm photocatalyst, developed
nology, but today an interesting renewable source by a single process using a radio frequency mag-
for hydrogen production is water splitting. The netron sputtering (RF-MS) deposition method,
theoretical energy to split water to produce H2 showed high photocatalytic reactivity for various
and O2 (e.g., under solar light) is reactions such as reduction of NOx, degradation of
organic compounds, and splitting of H2O under
H2 O hn ! H2 12 O2 visible and/or solar light irradiations (Takeuchi
et al. 2012). Cation- or anion-doped metal oxides
DG 237 kJ mol1 or metal oxynitride was used to prepare visible
light-responsive TiO2 thin lms by a radio fre-
Energy for water splitting can be supplied by quency magnetron sputtering method that were
various processes such as electrolysis, photo- applied for the separate evolution of H2 and O2
chemical, photocatalytic, and thermal from water under visible or solar light irradiation
decomposition. (Matsuoka et al. 2007). A photochemical water-
Photocatalytic water-splitting technology splitting system, composed by two reactors divided
using nano-sized TiO2 has great potential for by a proton-conducting membrane in which
low-cost, environmentally friendly solar- photocatalytic half reactions of water reduction
hydrogen production to support the future and water oxidation took place, was proposed by
Hydrogen Production by Water Splitting 995
Zamrescu et al. Complex molecular devices perovskite membranes toward a continuous and
based on ruthenium(bipyridine)(3)(2+) photosen- isothermal operation. At steady state and
sitizers were dissolved in both reactors, which 900 C, 25 7 cm3(STP)H2 m2 min1 is pro-
generate electrons or holes when exposed to duced in puried state (Nalbandian et al. 2009).
high-energy photonic radiation, and acted as cata- In a perovskite hollow-ber membrane, the oxy-
lysts for water splitting. These molecular devices gen permeation ux increased 3.5 times more
for water reduction have a unique property to than that of the blank membrane when feeding
enhance the existence time of photoelectrons, reactive CH4 onto the permeation side of the
such that the likelihood of generated electron membrane. The membrane was used as a reactor
pairs to produce a molecule of hydrogen is to shift the equilibrium of thermal water disso-
increased (Zamrescu et al. 2011). Thermochemi- ciation for hydrogen production because it
cal cycles using water as raw material and nuclear/ allows the selective removal of the produced
renewable energies as sources of energy are oxygen from the water dissociation system. It
believed to be a safe, stable, and sustainable route was found that the hydrogen production rate
of hydrogen production. Among the well-studied increased from 0.7 to 2.1 mLH2 min1 cm2 at
thermochemical cycles, the sulfur-iodine (S-I) 950 C after depositing a perovskite Pd porous H
cycle is capable of achieving an energy efciency layer onto the perovskite membrane (Jiang
of 50 %, making it one of the most efcient cycles et al. 2010).
among all water-splitting processes (Kar
et al. 2012).
The S-I cycle is characterized by three basic
References
reactions as shown below:
Fujishima A, Honda K (1972) Electrochemical photolysis
1. I2 SO2 2H2 O ! 2HI H2 SO4 120 C of water at a semiconductor electrode. Nature
2. 2H2 SO4 ! 2SO2 2H2 O O2 830 C 238:3738
Jiang HQ, Liang FY, Czuprat O, Emov K, Feldhoff A,
3. 2HI ! I2 H2 450 C
Schirrmeister S, Schiestel T, Wang HH, Caro J (2010)
Hydrogen production by water dissociation in surface-
In order to overcome the low efciency due modied BaCoxFeyZr1-x-yO3-delta hollow-ber mem-
to the poor equilibrium decomposition of HI in brane reactor with improved oxygen permeation. Chem
Eur J 16:78987903
the third reaction, ongoing research is dedicated
Kar S, Binclal RC, Prabhakar S, Tewari PK (2012) The
toward the development of a hydrogen- application of membrane reactor technology in hydro-
permselective membrane reactor. Proper identi- gen production using S-I thermochemical process: a
cation of suitable membranes (e.g., asymmet- roadmap. Int J Hydrog Energy 37:36123620
Matsuoka M, Kitano M, Takeuchi M, Tsujimaru K,
ric silica membrane) and introduction of
Anpo M, Thomas JM (2007) Photocatalysis for new
membrane reactor is proposed to improve the energy production recent advances in photocatalytic
efciency of the overall cycle and make hydro- water splitting reactions for hydrogen production. Catal
gen production more economical. The chal- Today 122:5161
Nalbandian L, Evdou A, Zaspalis V (2009) La(1-x)Sr(x)
lenges are associated toward the development
MO(3) (M = Mn, Fe) perovskites as materials for ther-
of a membrane reactor which can be applied in mochemical hydrogen production in conventional and
highly corrosive environment like HI under a membrane reactors. Int J Hydrog Energy
high temperature of about 500 C (Kar 34:71627172
Ni M, Leung MKH, Leung DYC, Sumathy K (2007)
et al. 2012). Perovskites are investigated as A review and recent developments in photocatalytic
potential redox catalyst materials for the ther- water-splitting using TiO2 for hydrogen production.
mochemical production of hydrogen where Renew Sustain Energy Rev 11:401425
water is dissociated giving rise to the production ORegan B, Grtzel M (1991) A low-cost, high-efciency
solar cell based on dye-sensitized colloidal TiO2 lms.
of pure hydrogen during the oxidation step. The
Nature 353(6346):737740
oxidation and reduction steps can be combined Takeuchi M, Matsuoka M, Anpo M (2012) Ion engineering
in a membrane reactor constructed from dense techniques for the preparation of the highly effective
996 Hydrogen Removal by Membranes
TiO2 photocatalysts operating under visible light irra- of hydrogen-selective hollow-ber membrane
diation. Res Chem Intermed 38:12611277 systems for the in-process recycling of hydrogen
Zamrescu C, Dincer I, Naterer GF (2011) Analysis of a
photochemical water splitting reactor with supramolec- from ammonia purge gases has triggered large-
ular catalysts and a proton exchange membrane. Int scale application of the membrane gas separation.
J Hydrogen Energy 36:1127311281 This technology has been extended to other situ-
ations for recovery of hydrogen from gas mixtures
(H2/CO or H2/CH4 ratio adjustment for syngas
production) and has been successfully competing
Hydrogen Removal by Membranes with cryogenic distillation and PSA processes. In
the petrochemical industry, hydrogen recovery
V. V. Volkov from renery streams is an emerging eld for
A.V. Topchiev Institute of Petrochemical membrane separation; it is a key approach to
Synthesis Russian Academy of Sciences meet the increased demand of hydrogen (for
(TIPS RAS), Moscow, Russian Federation hydrotreating, hydrocracking, or hydrodesul-
furization processes) owing to new environmental
regulations (Bernardo et al. 2009).
Pure hydrogen is known to be a valuable indus- The main liability of hydrogen-selective mem-
trial material, and its annual consumption comes branes is that recompression of the permeated
to billions of cubic meters. Over 90 % of hydrogen hydrogen is usually required. Therefore, an alter-
is generated from fossil fuel sources (mainly, native method involves contaminate preferential
steam reforming of natural gas) and the other permeation, and this method is treated as a new
10 % is produced by water electrolysis. Hydrogen approach for hydrogen purication, primarily by
is widely used in diverse large-scale processes in using the carbon-based membranes. For example,
metallurgical, chemical, petrochemical, pharma- hydrogen production by natural gas steam
ceutical, and textile industries for the production reforming yields a gaseous mixture containing
of a wide range of products from semiconduc- hydrogen and also carbon dioxide and carbon
tors and steel alloys to vitamins and raw chemical monoxide. Solubility-controlled membranes
materials such as ammonia, methanol, and hydro- (e.g., carbon-based membranes) are preferentially
gen peroxide. Recently, hydrogen has attracted a permeable for big-sized gas molecules (e.g., CO2)
keen attention as an alternative energy carrier for in relation to hydrogen. This peculiarity offers
the solution of environmental problems (so-called evident economic advantages that maintaining/
concept of hydrogen economy). In contrast to collecting hydrogen in the retentate at nearly
fossil fuels, hydrogen combustion does not gen- bulk feed pressure mitigates the demands for
erate carbon dioxide but only water vapors. Large- costly hydrogen recompression steps, even
scale hydrogen production requires signicant though a CO2 compression step is required
capital investments for separation and purication (Ockwig and Nenoff 2007).
processes; thus, the cost of hydrogen markedly Hydrogen-separating membranes made of pal-
increases. Conventional technologies for hydro- ladium alloys have been developed over the past
gen separation include solvent absorption, pres- 50 years into a technology that in some instances
sure swing adsorption (PSA), fractional/cryogenic is used in practice. In an early work in the United
distillation, and membrane separation. Membrane States and in the former Soviet Union, relatively
separation seems to be the most promising method thick-walled tubes were used. This design has
because of its low energy consumption, continu- been advanced for ultra-purication of hydrogen
ous operation, low investment costs, and easy and for its application in the semiconductor
operation (Ockwig and Nenoff 2007). manufacturing processes and in the hydrogen gen-
Among all basic large-scale applications of erators for remote or small-scale usage. Current
polymer membranes, hydrogen recovery is the research topics in this area are concerned with
most important. In the 1970s, commercial success membrane reactors for hydrogen gas supply for
Hydrogen Selective Membranes 997
fuel cells or for the chemical process industry. Hydrogen selective membranes are designed
Palladium-alloy diffusers present a key compo- such that hydrogen concentration increases in
nent for the recovery of hydrogen radioisotopes the permeate. Based on the materials used, hydro-
which are used and produced by the nuclear s- gen selective membranes can be classied into
sion in the fusion reactors. Hydrogen recovery four types: (i) polymer (organic), (ii) metallic,
from waste gases or purge streams (e.g., (iii) carbon, and (iv) ceramic. Metallic, carbon,
hydrotreater off-gas) presents a potentially large and ceramic membranes are referred to as inor-
application of the palladium-based membrane ganic membranes. Depending on the type of the
technology. Coal gasication or natural gas raw material, inorganic membranes can be classi-
reforming coupled to a palladium membrane reac- ed into two groups: metal membranes and
tor can offer an alternative huge source of hydro- ceramic membranes. In addition, they could be
gen. Production of pure hydrogen for its use in divided into porous (microporous) and nonporous
fuel cells also seems to be another important mis- (dense) membranes (Adhikari and Fernando
sion of the palladium-based membrane reactor 2006). Microporous inorganic membranes
(Paglieri and Way 2002). include carbon and ceramic (amorphous or crys-
talline) membranes, and their separation H
characteristics are governed by the molecular
sieving transport mechanism. Usually, dense inor-
References
ganic membranes are based on a metal or on a
Bernardo P, Drioli E, Golemme G (2009) Membrane gas polycrystalline ceramic. In these dense mem-
separation: a review/state of the art. Ind Eng Chem Res branes, the fundamental operating mechanism
48:46384663 involves the conduction of free electrons and
Ockwig NW, Nenoff TM (2007) Membranes for hydrogen
the presence of specic catalytic surfaces in
separation. Chem Rev 107:40784110
Paglieri SN, Way JD (2002) Innovations in palladium order to dissociate H2 on the raw feed stream
membrane research. Sep Purif Rev 31:1169 side and reassociate both protons and electrons
on the product side. In such systems, hydrogen
selectivity is typically very high because their
dense structure prevents the transport of larger
atoms and molecules (e.g., CO, CO2, O2, N2,
Hydrogen Selective Membranes etc.). This high selectivity provides the production
of high-purity hydrogen (Ockwig and Nenoff
V. V. Volkov 2007).
A.V. Topchiev Institute of Petrochemical Historically, hydrogen separation is accom-
Synthesis Russian Academy of Sciences plished on the Pd-based membranes because
(TIPS RAS), Moscow, Russian Federation they catalyze surface dissociation/reassociation
processes and they are highly permeable to
hydrogen. Palladium and Pd-alloy membranes
As compared with other known fuels, hydrogen is can produce hydrogen gas on the ppb impurity
abundantly available in the universe and pos- levels. Since the late 1950s, small-scale
sesses the highest energy content per weight Pd-based membrane modules have produced
unit. Moreover, in contrast to fossil fuels, the use high-purity hydrogen at remote sites and for
of hydrogen as an energy source yields water as industrial, laboratory, or military purposes. There
the only by-product. Hence, in recent years, the are various types of metallic membrane
demand for hydrogen energy and production has materials for hydrogen separation: (i) pure metals
been growing. Membrane separation process (Pd, V, Ta, Nb, and Ti); (ii) binary alloys of Pd
offers an attractive alternative to mature technol- such as Pd-Cu, Pd-Ag, Pd-Y, and also Pd-based
ogies such as pressure swing adsorption (PSA) alloys with Ni, Au, Ce, and Fe; (iii)
and cryogenic distillation. complex alloys (Pd alloyed with three to ve
998 Hydrogenation Contactors and Reactors
other metals); (iv) amorphous alloys (typically References

Group IV and Group V metals); and (v) coated
metals such as Pd on Ta, V, etc. (Adhikari Adhikari S, Fernando S (2006) Hydrogen membrane sep-
aration techniques. Ind Eng Chem Res 45:875888
and Fernando 2006). Dense ceramic membranes
Ockwig NW, Nenoff TM (2007) Membranes for hydrogen
(e.g., perovskite, bismuth oxide, and solid electroseparation. Chem Rev 107:40784110
lyte) have been developed and commercialized. Ritter JA, Ebner AD (2007) State-of-the-art adsorption and
Dense inorganic membranes are generally membrane separation processes for hydrogen produc-
tion in the chemical and petrochemical industries. Sep
designed as a thin lm on a porous inorganic
support. Yampolskii YP, Volkov VV (1991) Studies in gas perme-
To overcome the relatively low permeance and ability and membrane gas separation in the Soviet
high cost of dense metallic membranes, Union. J Membr Sci 64:191228
researchers are exploring the utility of high-
permeance, less costly, and less selective micro-
porous inorganic (silica or carbon or zeolite)
membranes. The structure of these membranes Hydrogenation Contactors
presents a network of interconnected micropores and Reactors
with a diameter of ~0.5 nm. In general, micropo-
rous silicas show the highest hydrogen selectiv- V. V. Volkov
ities, and they exhibit the best H2/N2 selectivity A.V. Topchiev Institute of Petrochemical
which exceeds 10,000 for the membranes pre- Synthesis Russian Academy of Sciences
pared by chemical vapor deposition (Ritter and (TIPS RAS), Moscow, Russian Federation
Ebner 2007).
Since the 1970s, gas separation polymer
membranes have been used industrially for Among all known catalysts for hydrogenation,
hydrogen separation from gaseous mixtures in metals of platinum family (Pt, Pd, Ni) are known
the United States (Ockwig and Nenoff 2007) to be most efcient and advantageous. Palladium
and in the former Soviet Union (Yampolskii and catalysts for selective hydrogenation hold the
Volkov 1991). Membranes made of glassy unique position. Palladium and its alloys show
polymers (i.e., polymers with glass transition an excellent potential to dissolve huge amounts
temperatures above the operating temperature) of hydrogen, and as membrane materials, dense
are used for removing hydrogen from gas palladium and palladium alloy membranes are
mixtures. Examples of membrane materials for permeable only for hydrogen.
commercial polymer membranes include Pioneering studies by Gryaznov and his
polyimide and polysulfone (asymmetric coworkers have encouraged the use of palladium
hollow ber membranes) or polyvinyl and palladium alloy membranes in catalytic mem-
trimethylsilane and cellulose acetate (asymmetric brane reactors, including fundamental studies on
at membranes). Such polymer membranes oper- selective hydrogenation of various unsaturated
ate according to the solution-diffusion hydrocarbons, for example, acetylene,
mechanism. 1,3-pentadiene, and cyclopentadiene. The general
In general, hydrogen selectivity is low, concept can be formulated as follows: hydrogen is
moderate, and very high for polymeric, supplied to one side of the Pd-based catalytic mem-
microporous inorganic, and metallic membranes, brane, and then it diffuses selectively through the
respectively. Polymeric membranes primarily membrane and approaches the other side of the
operate at temperatures below 373 K, whereas membrane in its highly reactive form. Fundamental
carbon, silica, and dense ceramic membranes concepts on membrane catalysis and catalytic mem-
can function at higher temperatures brane reactors have been advanced in the 1960s, and
(7731,173 K). Metallic membranes can be used the phenomenon of conjugating of chemical reac-
at ~573873 K. tions with the evolution (dehydrogenation reaction)
Hydrogenation Contactors and Reactors 999
H < 0
Substrate
Liquid Hydrogenated products
Palladium q
H ~ 50 100 m
alloy
Gas R (opt.)
H2 or R-H2
H > 0
Hydrogenation Contactors and Reactors, Dehydrogenation of RH2 occurs on the gas-side surface
Fig. 1 Principle of a dense and self-supporting palladium and this process is endothermic; both reactions are accom- H
alloy membrane (a thin-walled tube or a foil) for liquid panied by hydrogen transfer. Due to excellent thermal
phase hydrogenation. Hydrogen or hydrogen-containing conductivity of the membrane, the heat released by hydro-
source RH2 is supplied to the gas side; a to-be- genation can be utilized to facilitate endothermic dehydro-
hydrogenated substrate is supplied to the liquid side. genation (autothermic operation) (Dittmeyer et al. 2004)
and consumption (hydrogenation reaction) of hydro- of dissolved oxygen (DO) from water (Volkov
gen on palladium membranes has been discovered in et al. 2011). Hydrogen and DO-containing water
1964 (Gryaznov et al. 2003). This concept is illus- are supplied to the opposite sides of a hydropho-
trated in Fig. 1. bic porous catalytic membrane. In this case, mem-
Early studies in this direction have been focused brane material should be non-wettable for water
on the use of relatively thick nonporous Pd or (hydrophobic) in order to provide liquid-free
Pd-alloy membranes. Their thickness can be pores and high-rate hydrogen transmembrane
reduced by preparation of composite membranes mass transfer. The Pd catalyst is placed onto the
on porous or nonporous supports. A critically new water-contacting membrane surface. Water deox-
step providing reduced consumption of expensive ygenation by catalytic membranes, including
noble metals is concerned with the development of palladium-loaded hollow bers, is accomplished
catalytic membranes based on the metalpolymer by the chemical reaction between dissolved oxy-
composites. In this case, nonporous polymer gen and hydrogen in the presence of the palladium
matrix, for example, silicon rubber, is loaded with catalyst.
palladium nanoparticles (Gryaznov et al. 2003). In a ow-through contactor, reactants are
Pd-based catalytic nanoparticles can be forced to ow through a porous catalytic mem-
immobilized on the surface of porous inorganic brane, i.e., through the Pd-loaded pores. This
or polymeric membranes, thus providing a new enables to control and rule residence time of
type of catalytic membrane reactor, called cata- reacting species in the active zone of the catalytic
lytic contactor. According to the proposed classi- membrane.
cation (Miachon and Dalmon 2004), depending
on the mode of supply of reagents, two types of
catalytic contactors exist: an interfacial contactor References
and a ow-through contactor. In this case, a key
point is concerned with localization of reactants in Dittmeyer R, Svajda K, Reif M (2004) A review of cata-
lytic membrane layers for gas/liquid reactions. Top
the very volume where the catalyst is deposited.
Catal 29(12):327
The typical example of hydrogenation on an Gryaznov VM, Ermilova MM, Orekhova NV, Tereschenko
interfacial catalytic contactor is related to removal GF (2003) Reactors with metal and metal-containing
1000 Hydrophilic Membrane
membranes. In: Cybulski A, Moulijn JA (eds) Struc- solution systems such as drinking water produc-
tured catalysts and reactors, 2nd edn. Taylor & Francis, tion, wastewater treatment, and bio-separation
New York, pp 579614
Miachon S, Dalmon J-A (2004) Catalysis in membrane due to its hydrophobic nature. The critical prob-
reactors: what about the catalyst. Top Catal 29:5965 lems of PVDF membrane lie in:
Volkov VV, Petrova IV, Lebedeva VI, Roldughin VI,
Tereshchenko GF (2011) Palladium-loaded polymeric 1. The surface energy of PVDF, which is very
membranes for hydrogenation in catalytic membrane
reactors. In: Basile A, Gallucci F (eds) Membranes for low, and the critical surface tension (gc) of
membrane reactors: preparation, optimization and PVDF, CF2, and CH2, which is
selection. Wiley, New York, pp 531548 25 mN/m, 18 mN/m, and gc = 31 mN/m,
respectively. This results in the poor wettability
of the PVDF membrane. Therefore, the pure
water ux is usually low due to the strong
Hydrophilic Membrane hydrophobic nature of PVDF.
2. The hydrophobic PVDF membrane is suscep-
Fu Liu tible to fouling while treating aqueous solution
Ningbo Institute of Materials Technology and containing natural organic matters, e.g., pro-
Engineering, Chinese Academy of Sciences, teins. Proteins are prone to be easily absorbed
Ningbo, China onto the membrane surface or block the surface
pores, decreasing the permeability and the nal
separation performance, depressing the life-
The hydrophilic term characterizes a membrane in time of the membranes and subsequently caus-
relation to its behavior in an aqueous environ- ing more operation costs of the replacement
ment. The membranes interaction with water is and maintenance of the membrane modules.
determined by its chemical composition and
corresponding surface. Hydrophilic membranes One major possible reason for the protein foul-
have an afnity to water. Their surface chemistry ing on the hydrophobic PVDF membrane is that
allows these materials to be wetted immediately. there are almost no hydrogen bonding interactions
Hydrophilic membrane is extremely imperative in in the boundary layer between the PVDF mem-
water treatment to avoid the fouling of organic brane interface and water. The repulsion of water
matter on the membrane surface. Diverse methods molecules away from hydrophobic PVDF mem-
have been developed to render the hydrophobic brane surface is a spontaneous process with an
polymeric membranes with improved hydrophi- increasing entropy, and therefore, protein mole-
licity including surface modication and blending cules have a tendency to adsorb onto membrane
modication. surface and dominate the boundary layer. In con-
trast, the membrane with the hydrophilic layer
possesses a high surface tension and is able to
form the hydrogen bonds with surrounding water
Hydrophilic Modification molecules to reconstruct a thin water boundary
between the membrane and bulk solutions. It is
Fu Liu difcult for hydrophobic solutes, like some pro-
Ningbo Institute of Materials Technology and teins, to approach the water boundary and break
Engineering, Chinese Academy of Sciences, the orderly structure because an increase of energy
Ningbo, China would be required to remove the water boundary
and expose the PVDF membrane surface.
Therefore, hydrophilic modication is neces-
The developments and applications of PVDF, PSf, sary to improve the hydrophilicity of PVDF mem-
or other polymeric membranes are limited in the branes through versatile methods to enhance the
areas of purication and separation of aqueous lifetime and reduce the operation cost of PVDF
Hydrophobic Membranes 1001
membranes. Different modication ways have surface to become wet or to absorb water. Gener-
been explored to tailor the surface engineering of ally speaking, when a surface shows hydrophilic-
the PVDF membrane in recent years. Two straight- ity, the molecules of water form bonds with the
forward ways are usually used to catalogue the molecules of the surface.
diverse hydrophilic and fouling resistant modica- In most cases, hydrophilicity refers to how
tion techniques, which can be mainly classied into readily a surface reacts with water molecules. It
surface modication and blending modication. is also possible to measure the hydrophilicity of a
Surface modication is usually achieved by coating surface when it comes into contact with other
or grafting a functional layer on the prepared mem- liquids. Understanding how hydrophilic a surface
brane surface, in which most of the modied sites can help engineers to develop materials that show
occurred on the top and/or bottom surface of the the proper amount of hydrophilicity for their
membrane, excluding the pores inside the mem- intended use.
brane, due to the limited diffusion ability of the There are ways of determining hydrophilicity
modifying agents into the membrane pores. Blend- via utilizing precise mathematical formulas. If
ing modication is usually used to achieve the droplets of water remain formed on the surface
desired functional properties along with the mem- and roll off easily when the surface is angled H
brane preparation; therefore, the preparation and toward the ground, the material is said to be par-
modication process can be accomplished in a sin- ticularly hydrophobic. If, on the other hand,
gle step. Both surfaces and inside pores of the mem- water is absorbed into the surface and that
brane have a window of opportunity to be modied surface becomes wet, it is hydrophilic. The water
synchronously through the synergy effect between contact angle of a hydrophilic membrane is
PVDF and compatible additives (Fu Liu 2011). usually below 90 .
References
Hydrophobic Membranes
Liu F, Hashim NA, Liu Y, Abed MRM, Li K (2011) Pro-
gress in production and modication of PVDF mem- Fu Liu
brane. J Membr Sci 375:127 Ningbo Institute of Materials Technology and
Engineering, Chinese Academy of Sciences,
Ningbo, China
Hydrophilic-Hydrophobic Membrane
The hydrophobic term characterizes a membrane
Amphiphilic Membrane in relation to its behavior in an aqueous environ-
ment. The membranes interaction with water is
determined by its chemical composition and
corresponding surface. Hydrophobic membranes
Hydrophilicity have little or no tendency to absorb water, so that
a droplet remains on the surface. The water con-
Fu Liu tact angle of a hydrophobic membrane is usually
Ningbo Institute of Materials Technology and over 90 .
Engineering, Chinese Academy of Sciences, Most polymeric membranes can be regarded as
Ningbo, China hydrophobic membranes, e.g., PVDF, PTFE, PP,
PE, and PSf membranes. Hydrophobic mem-
branes usually possess an unique advantage in
Hydrophilicity refers to a membrane with hydro- membrane distillation, oil/water separation or
philic property. Hydrophilicity is the tendency of a membrane contactor.
1002 Hydrophobic-Hydrophilic Interaction
Hydrophobic-Hydrophilic Hydrophobicity
Interaction
Wolfgang M. Sigmund and Shu-Hau Hsu
Fu Liu Department of Materials Science and
Ningbo Institute of Materials Technology and Engineering, University of Florida, Gainesville,
Engineering, Chinese Academy of Sciences, FL, USA
Ningbo, China Department of Energy Engineering, Hanyang
University, Seoul, Republic of Korea
The hydrophobic effect is the observed tendency

of nonpolar substances to aggregate in aqueous Hydrophobicity is the term frequently used in
solution and exclude water molecules (IUPAC surface chemistry describing the tendency of
1997; Interfaces and the Driving Force 2005). moving away from water or aqueous environ-
The name, literally meaning water fearing, ment. The name, the combination of hydro- and
describes the segregation and apparent repulsion phobos, simply stands for water and fear in Attic
between water and nonpolar substances. The Greek. A hydrophobic surface tends not to be
hydrophobic effect explains the separation of a dissolved or wetted by water. Water forms distinct
mixture of oil and water into its two components droplets on a hydrophobic surface, and their con-
and the beading of water on nonpolar surfaces tact angles increase as its hydrophobicity
such as waxy leaves. At the molecular level, the increases. For more details, please refer to the
hydrophobic effect is important in driving protein section of Youngs Model.
folding (The Binding of Benzoarylsulfonamide
2003), formation of lipid bilayers and micelles,
insertion of membrane proteins into the nonpolar
lipid environment, and protein-small molecule Hydroprocessor Purge Gases
interactions. The hydrophobic surface is usually
dened as the water drop contact angle is over Arnaud Baudot
90 . Physics and Analysis Division, IFP Energies
Hydrophilicity is the tendency of a surface to nouvelles, Solaize, France
become wet or to absorb water. Generally speak-
ing, when a surface shows hydrophilicity, the
molecules of water form bonds with the molecules The demand for hydrogen is constantly increasing
of the surface via highly dynamic hydrogen in reneries due to more and more stringent sulfur
bonds. The hydrophilic surface is usually dened content specications for fuels (leading to an
as the water drop contact angle is below 90 . increasing hydrogen consumption in hydrodesul-
furization processes) and a growing heavy crude
consumption, which results in a higher worldwide
demand for highly hydrogen-consuming
References upgrading processes, such as hydrocracking. In
this changing landscape, permeation membranes
(2003) The binding of benzoarylsulfonamide ligands to
constitute an elegant option for the recovery of
human carbonic anhydrase is insensitive to formal uo-
rination of the ligand. Angew Chem Int Ed hydrogen that is nowadays wasted in an array of
52(30):77147717 renery off-gases, such as fuel gas, PSA tail gas,
(2005) Interfaces and the driving force of hydrophobic FCC gas, catalytic reformer off-gases, or hydro-
assembly. Nature 437(7059):640647
cracker/hydrotreater off-gases.
IUPAC (1997) Compendium of chemical terminology, 2nd
edn (the Gold Book). Online corrected version: Nowadays, three main membrane providers
(2006) hydrophobic interaction offer hydrogen purication permeators:
Hydroprocessor Purge Gases 1003
Air Products with the Prism silicon-coated Hydroprocessor Purge Gases, Table 1 Ube Industries
polysulfone membranes (issued from membrane material compatibility against contaminants
(Ube Industries website)
Monsanto). Air Products claims that the life-
time of the Prism modules can be more than Contaminants Maximum allowable content
15 years. Water vapor Up to saturation
Ube Industries with polyimide hollow ber H2S 3 % vol
NH3 and amines 100 ppm vol
membranes.
Methanol 5 % vol
Air Liquide with the MEDAL polyimide and
Methyl ether 5 % vol
high selectivity polyaramide membranes.
Benzene 1 % vol
Toluene 2000 ppm vol
All those membranes are based on glassy poly- C5+ hydrocarbons Up to saturation
mers and offer a diffusion-based hydrogen selec-
tivity. Todays membranes have high H2/CH4
selectivities (from 35 to 200 at 80 C (Roman used in order to remove poisonous com-
et al. 2001)). For instance, Air Products claims pounds: coalescing lters in order to remove
that a single-stage array of Prism modules is able aerosols, sorbent beds, or even complete
H
to raise the concentration of gases from 1030 % PSA (pressure swing adsorption) or TSA
up to 7090 % (Air Products website), while (temperature swing adsorption) units
MEDAL membranes are able to raise the concen- (Monereau 2003). This can lead to a signicant
tration of a gas at 51 bar from 86 % in hydrogen up increase of investment and operating costs. It
to 98 % with a permeate pressure at 30 bar and a should be reminded here that recent hydrogen
residue containing 52 % of hydrogen at 50 bar purication membranes are still relatively tol-
(Medal website). erant to contaminants (Table 1). This is partic-
In 2003, it was reported that more than ularly true if membrane modules are operated
400 hydrogen permeators were installed world- at higher temperatures (from 80 C to 110 C),
wide (Monereau 2003) while approximately which results in lowering of the sorption of
100 were operated in reneries (Baker 2002). As contaminants.
there are more than 500 reneries in the world, it The membranes mechanical integrity can be
is clear that the potential market for this type of damaged in transient operating conditions,
membrane applications is far from saturated. Nev- especially in the case of an emergency blow-
ertheless, three main limits still hinder the wide down of the membrane-based process. In cer-
acceptance of permeation-based hydrogen puri- tain cases, the membrane module can be sub-
cation in the rening industry: mitted to pressures differences larger than its
mechanical tolerance. Solution is nowadays
The puried hydrogen is recovered at low pres- proposed by membrane providers in order to
sure in the permeate side and requires compres- monitor automatically the pressure balance
sion in order to feed it back to reactors. As between the feed and the permeate compart-
such, PSA (pressure swing adsorption) is a ment when operating condition limits are
more attractive process, as the produced puri- reached (Monereau 2003).
ed hydrogen is delivered directly at high
pressure.
The sensitivity of membrane to contaminants, References
such as water vapor, higher hydrocarbons, or
acid gases. The installation of permeators gen- Air Products website. http://www. airproducts.
erally requires at rst a very detailed analysis com/Products/Equipment/PRISMMembranes/page08.
htm
of the contaminants present in the feed, even at Baker RW (2002) Future directions of membrane gas sep-
traces level, in order to design pretreatment aration technology. Ind Eng Chem Res 41:13931411
operations. A wide array of solutions can be Medal Internet website. www.medal.airliquide.com
1004 Hydrothermal Stability of Zeolite
Monereau C (2003) Permation hydrogne: de la priphrie at the edge of the crystal, which would be
des procd vers le coeur des procds. Intgration des extracted easily by calcination and steaming treat-
membranes dans les procds 2:89:275282
Roman IC, Ubersax RW, Fleming GK (2001) New direc- ments. In general it could be concluded that there
tions in membrane for gas separation. Chim Industria is a trade-off in the aluminum content of the
83:13 zeolite, as it increases acidity and catalytic activ-
Ube Industries website. http://www.ube.com/content.php? ity; however, hydrothermal stability decreases.
pageid=45
Similarly some problems related to hydrother-
mal stability of zeolite A, Si/Al = 1 in dehydra-
tion of organic solvent by pervaporation, has been
reported. The damage effects of water on LTA
Hydrothermal Stability of Zeolite zeolite membranes were veried and occur at
water concentrations higher than 15 % and tem-
Reyes Mallada peratures of 70 C. The effects of water content in
Department of Chemical and Environmental the feed and PV temperature on the membrane
Engineering and Aragon Nanoscience Institute, stability were recently investigated Li
Universidad de Zaragoza, Zaragoza, Spain et al. (2007). The preliminary results show that
the damage of LTA zeolite membrane by water
mainly occurs in the grain boundary regions. The
Zeolites are microporous crystalline aluminosili- amorphous-like structure of the grain boundary is
cates, and its Si to Al ratio and ion exchange probably the principal reason for the low hydro-
determines its main properties. The most impor- thermal stability of LTA zeolite membrane at high
tant application of zeolites is catalysis, and in water concentration, as the case of mesoporous
particular zeolite Y and ZSM-5 are the catalysts molecular sieving materials. Solutions to this
in the uid catalytic cracking (FCC) process for problem include the use of a zeolite with a higher
gasoline production in reneries. The acidity of Si/Al ratio, such as CHA membrane (Sato
the zeolites, related to its aluminum content and et al. 2012), and posttreatments in zeolite LTA
the shape selectivity given by the microporous such as silylation.
network of cavities, are the main properties that
contribute to the success of the zeolites in the
cracking reactions. The production of coke in the References
FCC implies a continuous regeneration step at
Li Y, Zhou H, Zhu G, Liu J, Yang W (2007) Hydrothermal
temperatures up to 800 C in the presence of stability of LTA zeolite membranes in pervaporation.
steam. Under these severe hydrothermal condi- J Membr Sci 297(12):1015
tions, the activity and stability of zeolite is Martinez A (2008) Chapter 10: J. Cejka, J. Perz-Pariente
affected mainly due to the dealumination of the and W. J. Roth (eds) application of zeolites in the
production of petrochemical intermediates. In Zeolites:
zeolite. The preparation of USY, ultrastable from model materials to industrial catalysts,
zeolite Y, consists of increasing the Si/Al ratio of 1st ed. Transworld Research Network, Kerala,
the zeolite, decreasing the activity but increasing pp 227261
hydrothermal stability. Similarly the addition of Sato K, Sugimoto K, Shimotsuma N, Kikuchi T, Kyotani T,
Kurata T (2012) Development of practically available
phosphorous also stabilizes the aluminum frame- up-scaled high-silica CHA-type zeolite membranes for
work in the zeolite (Martinez 2008). Finally the industrial purpose in dehydration of N-methyl
creation of the mesopores in the zeolites for FCC pyrrolidone solution. J Membr Sci 409410:8295
is also desired to crack larger molecules. The
introduction of mesopores is accomplished by
alkali treatment, but the thermal and hydrothermal
stability again decreases, since the desilication Hydrothermal Synthesis
effect of the NaOH solution, and this made some
Al sites in the interior crystal become the Al sites Perovskite Powder Preparation Methods
Hydrothermal Synthesis of Zeolite 1005
membranes being the most important: MFI,

Hydrothermal Synthesis of Zeolite LTA, FAU, CHA, and MOR.
A hydrothermal treatment implies the use of an
Reyes Mallada aqueous solution, synthesis gel, which is heated in
Department of Chemical and Environmental an autoclave at a certain temperature under autog-
Engineering and Aragon Nanoscience Institute, enous pressure. The chemicals in the synthesis gel
Universidad de Zaragoza, Zaragoza, Spain include the silicon and aluminum sources in the
presence of a mineralizing agent commonly
sodium hydroxide or alternatively uoride ion
Zeolites are crystalline, hydrated aluminosilicates and in some cases a structure-directing agent
having microporous, regular structures that occur (SDA) that facilitates the synthesis of the desired
naturally. The synthetic zeolites early developed zeolite structure.
by Barrer and Milton in the late 1940s (Cundy and In a conventional hydrothermal synthesis all
Cox 2005) were prepared using a hydrothermal the reactants are mixed together and dissolved in
treatment at high temperature and pressure mim- the highly alkaline synthesis gel and heated up in
icking the conditions for the synthesis of these the autoclave up to the selected temperature. After H
minerals in nature. There are more than 200 frame- an induction period crystalline zeolite nuclei
work types, recognized by the International Zeo- could be detected later, the nuclei start growing
lite Association (IZA, http://www.iza-structure. and nally all the amorphous material is
org), but only a few have been synthesized as converted into zeolite. There are several
Soluble silicates and aluminosilicates
Gel
Depolimerization of aluminosilicate
chains
Primary Building
Units
Secondary Building
Units
Polyhedra
Zeolite
Hydrothermal Synthesis of Zeolite, Fig. 1 Steps in Hydrothermal Synthesis of Zeolites

1006 Hydrotropes
mechanisms proposed for the hydrothermal syn- time of zeolites and zeolite membranes (Li and
thesis of zeolites (Cundy and Cox 2005), but in Yang 2008) due to a more efcient volumetric
general they comprised the dissolution of the ionic heating compared to conventional heating.
species, organization of the aluminum and silicon
tetraeda species around the structure-directing
agents forming initial primary building units, References
that result in secondary building units, that nally
form the polyhedra that comprise the zeolite struc- Cundy CS, Cox PA (2005) The hydrothermal synthesis of
zeolites: precursors, intermediates and reaction mecha-
ture, see Fig. 1.
nism. Microporous Mesoporous Mater 82:178
The most probable mechanistic pathways in Li Y, Yang W (2008) Microwave synthesis of zeolite
zeolite formation are described in sequence: membranes: a review. J Membr Sci 316:317
induction period, nucleation, and crystal growth.
The nucleation and crystal growth could be sepa-
rated, giving rise to the seeded hydrothermal
synthesis method of zeolite membranes. Hydrotropes
The composition of the synthesis gel inu-
ences the crystallization and the zeolite formed. Amphiphilic Molecules
Firstly the Si/Al ratio will dene the ion exchange
capacity, as one atom of Si is replaced by Al a
charge defect is created. Also the hydrophylicity
of the zeolite increases as the Al content increases, Hydroxylation of Benzene to Phenol
being zeolite A with a Si/Al = 1 the most hydro- by Membrane Contactors
philic and silicalite with Si/Al = 1 the most
hydrophobic. The solvent, as water (and some- Teresa Poerio
times alcohols) has a direct effect on the concen- Institute on Membrane Technology, National
tration of reactants, and therefore on the Research Council of Italy, ITM-CNR, Rende,
nucleation and growth rates. Alcohols are often Italy
added to retard the hydrolysis of the silicon alk-
oxide which allows controlling the crystal size by
reducing the polymerization rate of the silica. The The search for new routes for phenol production
amount and nature of the solvent will also affect based on the one-step direct benzene oxidation
the viscosity of the synthesis gel, and therefore the became more intensive because this reaction is a
diffusion rates of reactants during synthesis. The little selective, being the phenol more reactive
presence of the structure-directing agent (SDA), than benzene, and over-oxygenated by-products
facilitates zeolite synthesis, but in zeolite layers, occur (Molinari and Poerio 2009). To slow down
and especially in membranes, may cause compli- consecutive catalytic reactions, a membrane
cations associated to template removal. The min- contactor has been employed. This system permits
eralizing agents they also have a direct effect on to combine membrane separation and catalytic
the synthesis dynamics and in the hydrophilic/ reaction in one unit operation. The separation of
hydrophobic character of the resulting a product from the reaction mixture is one of the
membranes. advantages in using a membrane in a reactor and
The synthesis time of zeolites could vary from permits to obtain improvements in terms of yield
several hours to days in conventional ovens at and selectivity in equilibrium-limited reactions
temperatures ranging from 90 C to 200 C. It is and in consecutive catalytic reactions (Armor
important to mention that in the last decade the 1998).
microwave hydrothermal assisted method (see The benzene hydroxylation to phenol was car-
Microwave synthesis of zeolite membranes) ried out in a biphasic membrane contactor (Fig. 1)
has contributed to shorten considerably synthesis in which a at-sheet membrane separates two
Hydroxylation of Benzene to Phenol by Membrane Contactors 1007
Hydrophobic Hydrophobic
Membrane membrane
Aqueous
phase Organic
phase
C6H6
Aqueous Organic
phase phase C6H6+H2O2+cat C6H5OH+H2O+cat
C6H5OH
Stirrers
Hydroxylation of Benzene to Phenol by Membrane Contactors, Fig. 1 Scheme of the membrane contactor
compartments containing two immiscible phases: reducing the phenol productivity (0.64 gh1). H
an acidic aqueous phase (containing an iron cata- The use of different operating pH values and
lyst and hydrogen peroxide as the oxidant) and an different organic acids delayed and in some tests
organic phase (only benzene). avoided tar formation but also gave a reduction of
The benzene phase, in the double task of strip- the phenol selectivity and then productivity. Dif-
ping solution and reagent, avoided the use of other ferent catalysts, such as vanadium-based vanadyl
organic solvents with benet in product recovery. (IV) acetylacetonate (VAAC) and vanadium(III)
In this system the benzene permeates across the chloride (VC), were also tested, and the obtained
membrane from the organic phase to aqueous results evidenced improved productivity (0.97 vs
membrane interface where the interfacial oxida- 0.78 gphgcat1h1) using VC compared to VAA-
tion reaction takes place. Then the formed phenol C. No tars formation has been observed, but the
permeates back across the membrane to the system productivity was signicantly lower than
organic phase where it takes shelter from over- that obtained using an iron(II) catalyst (Molinari
oxidation reaction. The high value of phenol et al. 2012).
selectivity (98 %) in the organic phase and inter- A direct synthesis of phenol from benzene was
esting phenol productivity (3.60 gphgcat1 h1) also studied using a photocatalytic membrane
were obtained, thanks to its extraction in the contactor (PMC), using TiO as suspended catalyst
organic phase, avoiding further contact between (Molinari et al. 2009).
phenol and the catalyst, which was soluble in the The PMC allowed to obtain the phenol produc-
aqueous phase (Molinari et al. 2006). The draw- tion and its separation, although the formation of
back of this system was the low rate of phenol intermediate oxidation by-products, like benzo-
extraction in the organic phase. Indeed, phenol quinone, hydroquinone, and other oxidized mole-
that does not cross rapidly to the membrane reacts cules, was observed. Nevertheless, the acidic
further to generate over-oxidation products such condition allowed to control the selectivity toward
as benzoquinone, biphenyl as trace, and tars. In the intermediates, permitting to obtain a lower
order to avoid/reduce tar formation and promote formation and extraction of the over-oxidized
phenol extraction in the organic phase, different products.
strategies have been applied (Molinari and Poerio
2011). For example, the use of dissolved salts
References
such as sodium sulfate in the aqueous reacting
phase increased phenol extraction in the organic Armor JN (1998) Applications of catalytic inorganic mem-
phase but also increased reaction kinetics, thus brane reactors to renery products. J Mem Sci
promoting black solid (tar) formation and 147:217233
1008 Hydroxypropyl Cellulose (HPC) Membrane
Molinari R, Poerio T (2009) Remarks on studies for direct et al. 2008). Figure 1 shows the synthetic route
production of phenol in conventional and membrane for making hydroxypropyl cellulose. The
reactors. Asia Pac J Chem Eng 5:191206
Molinari R, Poerio T (2011) Selectivity control of benzene hydroxyl groups on the glucose unit of cellulose
conversion to phenol using dissolved salts in a mem- can be partially or completely substituted by
brane contactor. Appl Catal A Gen 393:340 hydroxypropyl groups through an etherication
Molinari R, Poerio T, Argurio P (2006) One-step produc- reaction. The maximum molar substitution is
tion of phenol by selective oxidation of benzene in a
biphasic system. Catal Today 118:5256 3, and consequently, the viscosity of the polymer
Molinari R, Caruso A, Poerio T (2009) Direct benzene increases with the degree of polymerization. The
conversion to phenol in a hybrid photocatalytic mem- broad range of accessible viscosity leads to differ-
brane reactor. Catal Today 144:8186 ent applications in pharmaceutical formulation,
Molinari R, Argurio P, Poerio T (2012) Vanadium(III) and
vanadium(IV) catalysts in a membrane reactor for ben- e.g., hydroxypropyl cellulose with low viscosity
zene hydroxylation to phenol and study of membrane has been used as tablet binder, while those with
material resistance. Appl Catal A Gen medium and high viscosities have been used in the
437438:131138 formulation of sustained-release matrices, espe-
cially in oral pharmaceutical products (Guo
et al. 1998).
Due to the excellent water solubility,
hydroxypropyl cellulose can be mixed with other
Hydroxypropyl Cellulose (HPC) water-soluble polymers, such as chitosan, to fab-
Membrane ricate pervaporation membranes after cross-
linking. The composite membrane has been
Hongyang Ma, Benjamin S. Hsiao and employed to dehydrate isopropanol, with excel-
Benjamin Chu lent ux and selectivity improvement, when com-
Department of Chemistry, Stony Brook pared with pure chitosan membranes (Veerapur
University, Stony Brook, NY, USA et al. 2007).
Biocompatible hydroxypropyl cellulose has
been used as a backbone in gene delivery vectors
Hydroxypropyl cellulose (HPC) is a water- via grafting modications, where the comb-shape
soluble polymer, often used as a thickening copolymer exhibits low cytotoxicity and high
agent, a tablet binder, and a lm-coating material transfection capability in a gene delivery process
for drug release. Combining with ethyl cellulose (Xu et al. 2009). Similarly, membranes composed
(EC) or cross-linked with polyacrylic acid, HPC of a mixture of hydroxypropyl cellulose and ethyl
can be used as a mucoadhesive delivery system or cellulose can offer special release proles in drug
oral delivery system, where the drug release rate delivery application (Sakellariou and Rowe
and the drug dosage can be controlled (Kamel 1995).
OH OR
HO OH O RO OR
O O
O O O
O Etherification
O RO O
HO OH OR
OH n OR n
R = CH2CH(OH)CH3 or H
Hydroxypropyl Cellulose (HPC) Membrane, Fig. 1 Etherication of cellulose to hydroxypropyl cellulose (HPC)
Hyflon AD 1009
References dioxole (TTD) (Colaianna et al. 1999). The chem-

ical structure of these completely uorinated (i.e.,
Guo JH, Skinner GW, Harcum WW, Barnum PE peruoro-) copolymers is shown in Fig. 1.
(1998) Pharmaceutical applications of naturally occur-
Hyon AD and the monomers (Navarrini
ring water-soluble polymers. Pharm Sci Technol Today
1:254261 et al. 1999) used for its polymerization are
Kamel S, Ali N, Jahangir K, Sham SM, El-Gendy AA manufactured by Solvay.
(2008) Pharmaceutical signicance of cellulose: a The presence of TTD units in the copolymer
review. eXPRESS Polym Lett 2:758778
prevents the formation of crystalline phases
Sakellariou P, Rowe RC (1995) Interactions in cellulose
derivative lms for oral drug delivery. Prog Polym Sci (Milani et al. 2008) and completely amorphous
20:889942 materials are obtained when its content is higher
Veerapur RS, Gudasi KB, Aminabhavi TM than approximately 20 % (Arcella et al. 2003).
(2007) Pervaporation dehydration of isopropanol
The glass transition temperature (Tg) of Hyon
using blend membranes of chitosan and hydroxypropyl
cellulose. J Membr Sci 304:102111 AD copolymers increases with the content of the
Xu FJ, Ping Y, Ma J, Tang GP, Yang WT, Li J, Kang ET, cyclic monomer up to 192 C which corresponds
Neoh KG (2009) Comb-shaped copolymers composed to value of the TTD homopolymer.
of hydroxypropyl cellulose backbones and cationic
The commercial grades of this amorphous H
poly((2-dimethyl amino)ethyl methacrylate) side
chains for gene delivery. Bioconjug Chem peruoropolymer are Hyon AD 40 and
20:14491458 Hyon AD60, whose density, Tg, and refractive
index are shown in Fig. 2. The numbers close to
the Hyon AD trade name indicates the molar
content of TTD which is then equal to 40 % for
Hyflon AD AD40 and 60 % for AD60.
Hyon AD 40 is available both as high molec-
Marco Avataneo ular weight and a low molecular weight grade,
RD&T, Solvay Specialty Polymers, Bollate (MI), whereas Hyon AD 60 is produced only in one
Italy grade.
All Hyon AD copolymers display high sol-
ubility in uorinated solvents and, in particular, in
Hyon AD is a family of amorphous peruorinated solvents like Galden peruoropo-
peruorinated copolymers of tetrauoroethylene lyethers (from Solvay) and Fluorinert FC40
(TFE) and 2,2,4-triuoro-5-triuoromethoxy-1,3- peruorotributylamine (from 3M). This property,
combined with low solution viscosity (Arcella
et al. 1999), makes Hyon AD an ideal material
OCF3 for solution process technologies: uniform and
CF C CF2 CF2 thin lms or coatings, down to submicron thick-
m nesses, can be easily obtained by casting or
O O coating techniques followed by solvent evapora-
CF2 tions at temperatures well below the typical
n
baking temperatures of semicrystalline
Hyflon AD, Fig. 1 Hyon AD chemical structure peruoropolymers.
Hyflon AD,
Fig. 2 Density, glass
transition, and refractive
index for the two
commercial grades of
Hyon AD
1010 Hyflon AD Membranes
Hyon AD copolymers combine excellent Membrane Swelling

optical properties (high optical transparency, low O2/N2 Separation
refractive index) and a low dielectric constant
with outstanding chemical and temperature resis-
References
tance. They are therefore choice materials for
active or protective thin layers into electronics, Arcella V, Colaianna P, Maccone P, Sanguineti A,
such as displays, wafer handling equipment, and Giordano A, Clarizia G, Drioli E (1999) A study on
other applications where very high optical trans- a peruoropolymer purication and its
parency, surface contamination repellency, or application to membrane formation. J Membr Sci
163:203209
static protection is required. Arcella V, Ghielmi A, Tommasi G (2003) High perfor-
High transparency and high purity are the main mance peruoropolymer lms and membranes. Ann
criteria for pellicles and for electronics applica- N Y Acad Sci 984:226244
tions. To satisfy these demanding specications, Bernardo P, Drioli E, Golemme G (2009) Membrane gas
separation: a review/state of the art. Ind Eng Chem Res
ultrapure grades of this material are manufactured 48:46384663
by submitting Hyon AD to a photochemical Colaianna P, Brinati G, Arcella V (1999) US Patent
uorination process (Tortelli and Calini 2002) in 5,883,177
order to completely eliminate the functional poly- Merkel TC, Pinnau I, Prabhakar R, Freeman BD
(2006) Gas and vapour properties of peruoro-
mer end groups and any other residual contami- polymers. In: Yampolskii Y, Pinnau I, Freeman BD
nants that can be derived from the polymerization (eds) Materials science of membranes for gas and
process. vapour separation. Wiley, New York, pp 251270
Gas separation membranes based on Hyon Milani A, Tommasini M, Castiglioni C, Zerbi G, Radice S,
Canil G, Toniolo P, Triulzi F, Colaianna P (2008) Spec-
AD are also of increasing interest in industrial troscopic studies and rst-principles modelling of
applications (Merkel et al. 2006; Bernardo 2,2,4-triuoro-5-triuoromethoxy-1,3-dioxole (TTD)
et al. 2009) due to a remarkable combination of and TTD-TFE copolymers (Hyon AD). Polymer
high chemical resistance, low hydrocarbon-vapor 49:18121822
Navarrini W, Tortelli V, Russo A, Corti S (1999) Organic
sorption, and high resistance toward swelling and hypouorites and their new role in industrial uorine
plasticization. The diffusivity of a gas through an chemistry
amorphous peruoropolymer is well described by Tortelli V, Calini P (2002) US 2002/0183459
its fractional free volumes (FFV), a common mea-
sure of the free space available, in a polymer
matrix, for molecular transport, which is directly
correlated to the Tg of the material. On the con-
trary, the higher the permeability the lower the Hyflon AD Membranes
selectivity is. Hyon AD60 having a Tg
(125 C) is a good compromise between moder- Marco Avataneo
ately high selectivity and good permeability. Typ- RD&T, Solvay Specialty Polymers, Bollate (MI),
ical applications are O2/N2 enrichment, He/H2 Italy
separation in space industry, and purication of
natural gas.
Hyon AD membranes are made of amorphous
peruorinated copolymers produced by Solvay
Cross-References and composed by tetrauoroethylene and 2,2,4-
triuoro-5-triuoromethoxy-1,3-dioxole units
Amorphous Peruoropolymers
(Colaianna et al. 1999). Hyon AD40 high-low
Carbon Dioxide (CO2) Separation by molecular weight, Hyon AD40 high molecular
Membranes weight and Hyon AD60 are three commercial
Gas Separation grades of this family. All of them are suitable for
Glass Transition Temperature (Tg) the preparation of membranes. Even if Hyon
Hyflon AD Membranes 1011
AD can be melt processed by molding or extru- Hyon AD membranes shows not only an
sion, the solution-based technologies are the pre- extremely hydrophobic character (contact angle
ferred methods as they allow the preparation of vs. water 120 ) but also an oleophobic character,
thinner and more uniform membranes: spin- being the contact angle vs hexadecane ca. 60 , a
coating deposition and casting are used to obtain value which implies fouling resistance.
ultrathin, uniform thickness coating on at sur- Separation membranes based on Hyon AD
faces, whereas nonplanar surfaces are more pref- are of increasing interest in industrial applications
erably coated by dipping. due to a remarkable combination of high chemical
By selecting the proper coating procedures, resistance, low hydrocarbon vapor sorption, and
dense symmetric, asymmetric, or composite high resistance toward swelling and plasticiza-
membranes can be obtained (Gordano tion. The permeability of a gas through an amor-
et al. 1999; Arcella et al. 2003, 2009, 2010). phous peruoropolymer is linearly correlated to
Dense membranes are prepared by deposition of its glass transition (Tg) (Arcella et al. 1999), and
a thin lm of the uoropolymer solution followed therefore, the higher the Tg, the higher the perme-
by solvent evaporation, whereas asymmetric ability, assuming a constant solubility coefcient.
membranes are obtained by phase inversion tech- High Tg amorphous peruoropolymers, like Tef- H
nique: after deposition, the support coated by the lon AF (160 C and 240 C, two grades), offer
uoropolymer solution is dipped into a coagula- high permeability to gases but also low selectivity.
tion bath containing a non-solvent, for example, On the contrary, Hyon AD40 (Tg = 95 ) and
n-pentane. Composite membranes, made of a Cytop (Tg = 108 C) show lower permeability
layer of the peruorinated polymer on a but higher selectivity. Among all the commercial
porous support, like polyethersulfones and amorphous peruoropolymers, Hyon AD60,
polyvinylidene uoride (Gugliuzza and Drioli which has an intermediate Tg (125 C), represents
2007; Gordano et al. 2001; Gugliuzza the best compromise between selectivity and
et al. 2006), can be prepared by coating the permeability.
solution on a porous at surface or hollow ber Gas permeability data for symmetric dense
(Jansen et al. 2006), followed by solvent membranes of Hyon AD40 and Hyon
evaporation. AD60 at 35 C and 50 psig are reported in Table 1.
Hyon AD solutions are prepared, even at As previously indicated, Hyon AD 60 shows
concentration higher than 10 %, by dissolving permeability higher than that of Hyon AD40.
the uoropolymer powder at room temperature As diffusivity is attributed to the presence of voids
(or above) in Galden peruoropolyethers. at the molecular scale, the permeability tends to
These latter are a family of uorosolvents diminish with the increase of the volume of the
manufactured by Solvay in several grades and gas molecule. Carbon dioxide, due to its para-
characterized by different boiling points: those mount solubility in uoromaterials (Kirby and
preferred for the preparation of membranes have McHugh 1999; McHugh et al. 1996), does not
a boiling point between 55 C and 170 C. Alter- follow this rule being its permeability higher
native solvents are peruorocarbons, peruorotri- than that expected on the basis of its critical
buthylamine, and hydrouoroethers. volume.
The combination of excellent solubility and The combination of high permeability with
low solution viscosity, in particular for the grades good selectivity and with high chemical and
AD40 low molecular weight and AD60, greatly swelling resistance makes these materials partic-
simplify the purication of the uoropolymer ularly useful for several applications such as gas
solutions by ltration (Arcella et al. 1999). In separation, membrane contractors, membrane dis-
fact, to ensure the preparation of membranes tillation, and pervaporation processes (Sanders
with no defects, it is essential to avoid the pres- et al. 2013; Bernardo et al. 2009; Gordano
ence of both suspended and dissolved et al. 2001; Gugliuzza et al. 2006). Typical appli-
contaminants. cations are purication of natural gas, O2/N2
1012 Hyperbranched Polyimide Membranes, Preparation and Characterization of
Hyflon AD Membranes, Table 1 Gas permeability for Arcella V, Ghielmi A, Tommasi G (2003) High perfor-
symmetric membranes of Hyflon AD40 and Hyflon mance peruoropolymer lms and membranes. Ann
AD60 at 35 C and 50 psig N Y Acad Sci 984:226244
Arcella V, Toniolo P, Avataneo M, Ghielmi A, Marchionni
Permeability Permeability
G (2010) Amorphous peruoropolymer membranes. In:
(1010 cm3 (1010 cm3
Drioli E, Giorno L Comprehensive membrane science and
(STP) (STP)
engeneering, vol 1. Oxford Academic Press, pp 147158
cm/cm2 s cm/cm2 s
Bernardo P, Drioli E, Golemme G (2009) Membrane gas
Critical cmHg) cmHg)
separation: a review/state of the art. Ind Eng Chem Res
volume Hyon AD Hyon AD
48:46384663
Gas (Helium = 1) 40 60
Colaianna P, Brinati G, Arcella V (1999) US 5883177
He 1,00 211 387 Gordano A, Clarizia G, Tocci E, Drioli E (1999) Hydro-
H2 1,13 83 182 phobic membranes of tetrauoroethylene and 2,2,4
O2 1,28 25 57 triuoro 5 triuorometoxy 1,3 dioxole. Korean
Ar 1,30 15 33 Membr J 1:5058
Gordano A, Clarizia G, De Santo M, Arcella V, Drioli
N2 1,55 8 20
E (2001) Composite membranes from amorphous
CO2 1,62 53 128 peruoropolymers for novel applications in mem-
CH4 1,73 4 10 branes processes. Polym Mater Sci Eng 84:10611062
C2H4 2,25 3 6 Gugliuzza A, Drioli E (2007) PVDF and Hyon AD
C2H6 2,21 2.5 3.3 membranes: ideal interfaces for contactor applications.
C3H8 3,50 0.94 1.1 J Membr Sci 300:5162
Gugliuzza A, Ricca F, Drioli E (2006) Controlled pore size,
thickness and surface free energy of super-hydrophobic
PVDF and Hyon AD membranes. Desal 200:2628
Huang Y, Ly J, Nguyen D, Baker RW (2010) Ethanol
enrichment, He/H2 separation in space industry,
dehydration using hydrophobic polymer membranes.
and ethanol dehydration (Huang et al. 2010). Ind Eng Chem Res 49:1206712073
Jansen JC, Tasselli F, Tocci E, Drioli E (2006) High-ux
composite peruorinated gas separation membranes of
Hyon AD on a hollow bre ultraltration membrane
Cross-References support. Desalination 192:207213
Kirby CF, McHugh MA (1999) Phase behavior of polymers
Amorphous Peruoropolymers in supercritical uid solvents. Chem Rev 99:565602
Asymmetric Hollow Fiber Membranes McHugh MA, Rindeisch F, Di Noia TP (1996) Solubility
of polymers and copolymers in supercritical CO2.
Asymmetric Membrane
J Phys Chem 100:1558115587
Coagulation Bath Sanders DF, Smith ZP, Guo R, Robeson LM, McGrath JE,
EthanolWater Mixtures: Separation by Paul DR, Freeman BD (2013) Energy efcient poly-
Pervaporation meric gas separation membranes for a sustainable
future: a review. Polymer (article in press, accepted
Gas Separation
manuscript)
Glass Transition Temperature (Tg)
Hydrophobic Membranes
Hydrophobicity
Ideal Gas Selectivity
Membrane Distillation (MD)
Hyperbranched Polyimide
Membrane Swelling
Membranes, Preparation
O2/N2 Separation
and Characterization of
Pervaporation Membrane
Petr Sysel
Department of Polymers, Institute of Chemical
References Technology, Prague, Czech Republic
Arcella V, Colaianna P, Maccone P, Sanguineti A,

Gordano A, Clarizia G, Drioli E (1999) Study on a
peruoropolymer purication and its application to Due to their structure, the macromolecular com-
membrane formation. J Membr Sci 163:203209 pounds can be classied as linear, branched, and
Hyperbranched Polyimide Membranes, Preparation and Characterization of 1013
cross-linked polymers. The dendritic topology which can react with each other but cannot
(dendrimers, hyperbranched polymers) has undergo self-reaction.) Many kinds of hyper-
recently been recognized as a new class of mac- branched polymers, e.g., polyesters, poly(ether
romolecular architecture (Tomalia 2005). ketone)s and polyamides, have been investigated
Hyperbranched polymers are highly branched as novel dendritic macromolecules so far (Gao
macromolecules. The highly branched structure and Yan 2004). An increasing attention has been
and a large number of terminal functional groups also devoted to hyperbranched polyimides due to
are their important structural features. a potentially possible connection of the known
Hyperbranched polymers often can be simply advantages of linear or cross-linked polyimides
prepared by a direct one-step polymerization of with those of hyperbranched polymers (Jikei and
multifunctional monomers using a single- Kakimoto 2004).
monomer (ABx, most frequently AB2 monomer) Polyimides exhibit very good chemical,
or double-monomer (A2+B3) methodology. mechanical, and dielectric stability at tempera-
(A and B represent two kinds of functional groups tures from 150 C to 250 C. These rigid
Hyperbranched H2N NH2

H
O O
Polyimide Membranes,
Preparation O H
and Characterization of, O O + C
Fig. 1 Preparation of
hyperbranched polyimide O O
based on 4,40 -oxydiphthalic
anhydride and 4,40 ,400 -
triaminotriphenylmethane
NH2
(via a polyimide precursor
(polyamic acid))
O O H H
O N N
H
HO OH C
O O
O O
O
N N
H
O O C
1014 Hyperbranched Polyimides
polymers with a high glass transition temperature References

are mostly used in (micro)electronics, aircraft
industry, and as polymeric separation membranes. Gao C, Yan D (2004) Hyperbranched polymers: from
synthesis to applications. Prog Polym Sci 29:183275
Linear polyimides are traditionally prepared by
Hergenrother PM (2003) The use, design, synthesis, and
two-step polymerization. The polyimide precursor, properties of high performance/high temperature poly-
polyamic acid (the most often a solution in mers: an overview. High Perform Polym 15:345
1-methyl-2-pyrrolidone), is prepared from an aro- Jikei M, Kakimoto M (2004) Dendritic aromatic polyam-
ides and polyimides. J Polym Sci A Polym Chem
matic dianhydride and aromatic diamine. This pre-
42:12931309
cursor is transformed into a polyimide using thermal Minko E, Sysel P, Spergl M, Slapakova P (2012)
or chemical imidization (Hergenrother 2003). It is Hyperbranched polyimides prepared from 4,40 ,400 -
difcult to prepare hyperbranched polyimides from triaminotriphenylmethane and mixed matrix materials
based on them. In: Abadie MJM (ed) High performance
ABx monomers due to the high reactivities of A and
polymers polyimides based from chemistry to
B groups. Therefore, they are often prepared by the applications. InTech, pp 3762. ISBN 978-953-51-
combination of bifunctional (A2) and trifunctional 0899-3
(B3) monomers. The ratio of anhydride and amine Suzuki T, Yamada Y, Tsujita Y (2004) Gas transport prop-
erties of 6FDA-TAPOB hyperbranched polyimide
component determines the kind and the ratio of end
membranes. Polymer 45:71677171
groups. Nevertheless, their synthesis from Tomalia DA (2005) Birth of new macromolecular archi-
trifunctional monomers often a rather complex tecture: dendrimers as quantized building blocks for
structure requires special and controlled reaction nanoscale synthetic polymer chemistry. Prog Polym
Sci 30:294324
conditions to avoid gel formation. As such triamines
were used, e.g., tris(4-aminophenyl)amine or 1,3,5-
tris(4-aminophenoxy)benzene (Jikei and Kakimoto
(2004)).
Recently, hyperbranched polyimides were aus- Hyperbranched Polyimides
piciously tested as polymeric membranes for gas
separation. Gas transport properties of the mem- Ryohei Shindo, Shinji Kanehashi and
branes made of 1,3,5-tris(4-aminophenoxy)ben- Kazukiyo Nagai
zene and 4,40 -(hexauoroisopropylidene) Department of Applied Chemistry, Meiji
diphthalic anhydride were monitored and these University, Tama-ku, Kawasaki, Japan
properties were compared with those for linear
polyimides with a similar chemical composition.
The permeability, diffusivity, and solubility of the Hyperbranched polyimides are formed by
gases were higher in the membranes based on repeated division of branches of comb
hyperbranched polyimides (Suzuki et al. 2004). polyimides. Hyperbranched polymers are synthe-
Similarly, the permeability coefcients of hydro- sized by polymerization of AB2-type monomers
gen, carbon dioxide, oxygen, nitrogen, and meth- and consist of branched structures and mixed
ane in the membrane from the hyperbranched straight chains (Fig. 1). Rigorous structural anal-
polyimide based on commercially available ysis cannot be performed because branching does
4,40 ,400 -triaminotriphenylmethane and 4,40 - not regularly occur. However, hyperbranched
oxydiphthalic anhydride (Fig. 1) were 23.5 polymers show properties differing from a normal
times higher than those in the membrane linear polymer because entanglement of the
from linear polyimide based on 4,40 -diaminodiphe- intermolecular chains is difcult. Flory (1952)
nylmethane at comparable selectivities (e.g., oxy- showed that gelation of polymerization of ABx-
gen/nitrogen about 6) (Minko et al. 2012). It seems type monomer cannot be statistically performed.
that the favorable effect of both the chain character Kricheldorf et al. (1982) reported the use of AB2-
(rigidity, interactions, arrangements, higher-free type monomer as one component in copolymers,
volume) and selective gas sorption (cavities, but their results lacked research attention.
end-capping groups) contributes to this increase. Hyperbranched polymers containing various
Hyperbranched Polyimides 1015
Hyperbranched
Polyimides,
Fig. 1 Architecture of
polymers
repeating units have been reported since the syn- 4-methylphthalimide as end groups show low
thesis of hyperbranched polyphenylene by dielectric constant, birefringence, and high optical
one-step polymerization of AB2-type monomer transparency. These qualities result from the
was reported as a simple synthesis method of improvement in isotropy of molecular chains by
polymers, which were similar in structure to the introduction of a multiple branching structure
dendrimers by Kim and Webster (1990) of Du and inhibition of the formation of charge-transfer
Pont. However, dendrimers as structurally- complex, causing coloration on linear polyimides.
controlled polymers have been actively studied in For aromatic amine to react easily with acid anhy-
the eld of medicine and for use in catalytic reac- dride at room temperature, isolation of ABx-type
tion and photoreaction. Hyperbranched polymers monomer with these functional groups in a mole-
have an obvious advantage in synthesis compared cule is difcult because of their instability. Poly
with dendrimers. Thus, these polymers can be used (amic acid ester) is synthesized by AB2-type mono-
as an alternative to dendrimers and a low-viscosity mer, which has a carboxylic acid ester and two
polymer in a wide range of areas. Polyimide is a amino groups in a molecule and a condensation
condensation polymer synthesized from dicarbox- agent. Hyperbranched polyimides are synthesized
ylic anhydride and primary diamine. Aromatic het- by chemical imidization of poly(amic acid ester)
erocyclic polyimides show good mechanical (Yamanaka et al. 2000). Hyperbranched
property and superior thermal and oxidation stabil- polyimides can be synthesized by self-
ity. These polyimides are widely used in place of polycondensation of ABx-type monomers with
metal and glass. They are also used for high- imide ring in the monomer framework. Thompson
functional application in electrical engineering, et al. (1999) reported that hyperbranched
electronic engineering, automobiles, aircraft, and polyetherimides are synthesized by thermal poly-
packaging industry. Linear aromatic polyimides condensation of AB2-type monomers with uorine
are known as polymers, which have poor workabil- (A functional group) and silylated phenolic
ity because they are insoluble and infusible in the hydroxyl group (B functional group), which can
rigidity of the main chain structure. However, sol- be detached in a molecule. When A2- and B3-type
ubility for organic solvent can remarkably improve monomers are used as starting materials in poly-
by the introduction of a multiple branching merization, AB2-type monomers need not be syn-
structure. Hyperbranched polyimides with thesized. Various hyperbranched polyimides are
1016 Hyperthermophiles
synthesized because they can be obtained by poly- geothermal power plants and sewage sludge sys-
merization using a commercial A2-type monomer tems), and deep-sea biotopes present also high
and synthesized B3-type monomer. Hyperbranched hydrostatic pressures (up to 360 atm); thus, the
polyimides have attracted attention as materials for isolated species are even barophilic (Vieille and
gas separation membranes since the early 2000s, Zeikus 2001). The most heat-resistant microor-
and most of them are synthesized by A2- and ganisms isolated up to date are the anaerobic
B3-type monomers. The gas permeability of archaea. The rst to be identied (1960s),
hyperbranched polyimides is almost equal to or Sulfolobus acidocaldarius, a hyperthermophile
higher than that of other glassy polymers such as and an acidophile, was found in an acidic hot
polysulfone or polycarbonate. spring in Yellowstone National Park, Wyoming.
Since then, more than 50 hyperthermophiles have
been isolated. For example, Pyrolobus fumarii,
References isolated in proximity marine hydrothermal vents,
is a nitrate-reducing chemolithotroph growing in
Flory PJ (1952) Molecular size distribution in three- the temperature range of 90113 C. The upper
dimensional polymers. VI. Branched polymer
temperature at which life is possible is still
containing A-R-Bf-1-type units. J Am Chem Soc
74:27182723 unknown, but above 110 C, amino acids and
Kim YH, Webster OW (1990) Water soluble metabolites become highly unstable. The mecha-
hyperbranched polyphenylene: a unimolecular nism at the base of adaptation and survival of
micelle. J Am Chem Soc 112:45924593
microbes to high temperatures recently proved to
Kricheldorf HR, Zang QZ, Schwarz G (1982) New poly-
mer syntheses. 6. Linear and branched poly involve DNA transfer (van Wolferen et al. 2013),
(3-hydroxybenzoates). Polymer 23:18211829 and DNA repair through this mechanism, between
Thompson DS, Markoski LJ, Moore JS (1999) Rapid syn- the domains of bacteria and archaea, seems to
thesis of hyperbranched aromatic polyetherimides.
have played a major role, also via homologous
Macromolecules 32:47644768
Yamanaka K, Jikei M, Kakimoto M (2000) Synthesis of recombination Figs. 1 and 2. These microorgan-
hyperbranched aromatic polyimides via polyamic acid isms, because of the extreme niches in which they
methyl ester precursor. Macromolecules 33:11111114 thrive, are adapted to distinct environmental fac-
tors including composition of minerals and gas-
ses, pH, redox potential, and salinity; most
biotopes of hyperthermophiles are essentially
Hyperthermophiles anaerobic. Interestingly, at ambient temperatures,
although unable to grow, hyperthermophiles can
Chiara Schiraldi and Mario De Rosa survive for many years by freezing their meta-
Department of Experimental Medicine, Section of bolic activities. Hyperthermophile communities
Biotechnology, Medical Histology and Molecular are complex systems of primary producers and
Biology, Second University of Naples, Napoli, decomposers of organic matter. The former are
Italy chemolithoautotrophs (i.e., sulfur oxidizers, sul-
fur reducers, and methanogens). Because of the
extremely low organic matter content of their
Hyperthermophilic bacteria and archaea have submarine environments, hyperthermophilic het-
been isolated in water-containing volcanically erotrophs typically obtain their energy and carbon
and geothermally heated environments situated from complex mixtures of peptides derived from
mainly along terrestrial and submarine tectonic the decomposition of primary producers. A few
fracture zones and represent the organisms at the species are able to use polysaccharides (e.g.,
upper temperature border of life. In fact they starch, pectin, glycogen, and chitin). In
reproduce and grow between 80 and 115 C. laboratory-scale cultivation, they needed complex
Hyperthermophiles have also been isolated from medium components; hyperthermophiles reached
hot industrial environments (e.g., the outow of very low cell densities, unless a specic
Hyperthermophiles 1017
Hyperthermophiles,
Fig. 1 Membrane Water solvatation shell
organization of
hyperthermophilic archaea
10A betagalpbetaglop
isoprenoid tetraether
with one cyclopentane ring
39.3A (isoprenoid methyls)
10A Pmyoinositol
Water solvatation shell
membrane bioreactor was employed. Enzymes of the entropy of unfolding, and intersubunit inter-
synthesized by hyperthermophiles, also called actions. Among hyperthermophilic enzymes,
hyperthermophilic enzymes, are typically thermo- there are many of potential industrial interest and
stable (i.e., resistant to irreversible inactivation at few that already reached commercialization. For
high temperatures) and are optimally active at instance, a number of hyperthermophilic prote-
high temperatures. When cloned and expressed ases are now used in molecular biology and bio-
in mesophilic hosts, they usually retain their ther- chemistry procedures (e.g., protease S from
mal properties, indicating that these properties are P. furiosus due to the broad specicity) and
genetically encoded (Schiraldi et al. 2000). How- numerous thermophilic restriction endonucleases.
ever, both sequence alignments and amino acid Thermophilic amylopullulanases have been
content comparisons show close similarities with suggested, as alternative enzymes to replace
the mesophilic counterparts suggesting that it is a-amylases during starch liquefaction for produc-
not a single mechanism, but a small number of ing fermentation syrups, rich in maltose,
highly specic alterations responsible for the maltotriose, and maltotetraose (DP2 to DP4)
remarkable stability (Cobucci-Ponzano (Cobucci-Ponzano et al. 2012). However, malto-
et al. 2012). The molecular mechanisms involved dextrin biotransformation at high temperature pre-
in protein thermostabilization reported to date sents various advantages. Besides the established
include ion pairs, hydrogen bonds, hydrophobic biotechnological applications, few other poten-
interactions, disulde bridges, packing, decrease tially interesting enzymes are isolated, cloned,
1018 Hyperthermophiles
Hyperthermophiles, Fig. 2 Liposomes made from archaeal lipids
and characterized, among this S. solfataricus Sulfolobus solfataricus: cloning, overexpression and
paraoxonase. That proved very stable may be properties. Extremophiles 9(4):297305
Schiraldi C, Martino A, Acone M, Di Lernia I, Di
used for detoxication of chemical warfare agents Lazzaro A, Marulli F, Generoso M, Carten M, De
and agricultural pesticides (Merone et al. 2005). Rosa M (2000) Effective production of a thermostable
alpha-glucosidase from Sulfolobus solfataricus in
Escherichia coli exploiting a microltration bioreactor.
Biotechnol Bioeng 70(6):670676
References van Wolferen M, Ajon M, Driessen AJ, Albers SV
(2013) How hyperthermophiles adapt to change their
Cobucci-Ponzano B, Perugino G, Strazzulli A, Rossi M, lives: DNA exchange in extreme conditions.
Moracci M (2012) Thermophilic glycosynthases for Extremophiles 17(4):545563
oligosaccharides synthesis. Methods Enzymol Vieille C, Zeikus GJ (2001) Hyperthermophilic enzymes:
510:273300 sources, uses, and molecular mechanisms for thermo-
Merone L, Mandrich L, Rossi M, Manco G (2005) stability. Microbiol Mol Biol Rev 65(1):143
A thermostable phosphotriesterase from the archaeon
I
Ideal Gas Selectivity permeability is the product of the diffusivity and

the solubility:
Johannes Carolus Jansen
Institute on Membrane Technology, National P D S (2)
Italy where D is the diffusion coefcient and S is the
solubility coefcient. Similarly, the selectivity can
be expressed in a diffusion term and a solubility
The ideal gas selectivity of a membrane, aij, is term:
dened as the ratio of the permeability of two pure
gases, measured separately under the same D i Si
aij (3)
conditions: D j Sj
Pi The ideal gas selectivity is an intrinsic property,

aij (1) specic for the membrane material and the partic-
Pj
ular gas pair. However, it is not a constant but it
depends on the operation conditions temperature
where Pi and Pj are the permeability (or the and pressure because both D and S depend on the
permeance) of the two pure gases, respectively, temperature and on the operating pressure.
with i being the most permeable gas. Rarely the
real selectivity is equal to the ideal gas selectivity.
Most commonly the ideal selectivity of a mem-
brane is lower than the real selectivity, especially Ideal Separation Factor
when the more permeable gas species plasticizes
the polymer matrix, making it relatively more Johannes Carolus Jansen
permeable for the slower species. In some cases, Institute on Membrane Technology, National
in particular in high free-volume polymers, strong Research Council of Italy, ITM-CNR, Rende,
sorption of the more permeable species may Italy
obstruct the transport of the less permeable spe-
cies, making the mixed gas selectivity higher than
the ideal selectivity. The separation factor, SF, is a measure of the
In dense membranes, where transport occurs efciency of the separation process and is deter-
by the solution-diffusion mechanism, the mined from the ratio of the concentrations of the
DOI 10.1007/978-3-662-44324-8
1020 IgG Purification
more permeable gas species i and the less perme- leakage that contaminates the therapeutic product,
able gas species j in the permeate divided by the and also they are expensive and difcult to handle,
ratio of the same gases i and j in the feed stream: sterilize, and preserve (Fuglistaller 1989).
Pseudo-specic ligands, such as histidine, trypto-
xi, p =xj, p phan, phenylalanine, etc., can be used for the
SF (1)
xi, f =xj, f IgG purication. They are small molecules with
high physical and chemical stability and low cost
where xi,p and xj,p are the fractions of components (Altnta and Denizli 2009; Turkmen et al. 2008).
i and j in the permeate and xi,f and xj,f are the The interaction of histidine with IgG has been
fractions of components i and j in the feed. The shown to be water mediated involving the com-
separation factor is not a material property, but it bined electrostatic, hydrophobic, and charge-
also depends on the conditions of the separation transfer interactions between histidine and the
process. It depends both on the membrane prop- specic amino acid residues available on the pro-
erties and on the driving force, which in turn tein surface (Bhattacharyya et al. 2003). In
depends on the pressure and on, for instance, the immobilized metal ion afnity chromatography
presence of concentration polarization phenom- (IMAC), the separation is based on the interaction
ena, nonideal behaviour such as plasticization, of a Lewis acid (electron pair acceptor), i.e., a
coupling effect, etc. chelated metal ion, with electron donor atoms
Analogously, the ideal separation factor is the (N, O, and S) on the side groups of the surface
separation factor under ideal conditions. It can be histidine, tryptophan, and cysteine of the protein.
calculated from the pure gas permeabilities. Histidine-rich sequence-containing IgGs show an
innate afnity for metal ions, and IMAC allows
one-step separation of IgG (Altnta et al. 2007).
Textile dyes bind proteins in a selective and
reversible manner and can be used for antibody
IgG Purification purication (Denizli and Pikin 2001). Dye
ligands can engage in ionic, hydrophobic,
Nilay Bereli, Deniz Turkmen, Handan Yavuz and charge-transfer, and hydrogen bonding with pro-
Adil Denizli teins. In the thiophilic adsorption of proteins,
Department of Chemistry, Biochemistry Division, electron donor-acceptor interactions between
Hacettepe University, Ankara, Turkey both functional groups present in the ligand struc-
ture and the adjacent sulfone group are the driving
force for selective recognition (Bakhspour et al.
Owing to their use in the treatment of various 2014). In general, specicity, rapid processing,
diseases, such as primary and secondary immune mild operation conditions, conventional equip-
deciencies, infections, and inammatory and ment, and reusability determine which technique
autoimmune diseases, large number of immuno- to be used for IgG purication. The use of mem-
globulin G (IgG) products is under clinical devel- branes has become indispensible for chromato-
opment. This requires certain protocols for graphic applications in both research and
purication and standardization. Afnity chroma- industry area for the last few decades due to their
tography is the most popular technique to reach relatively wide conguration for the size-, charge-,
these requirements (Low et al. 2007). Staphylo- and afnity-based protein separation and puri-
coccal protein A is one of the rst afnity ligands cation. The pressure drop across the membranes is
with a very high specicity for IgG purication. It very low due to the large pore size. Owing to the
interacts with IgG through hydrophobic interac- continuous pore structure, mass transport occurs
tions and some hydrogen bonds and electrostatic by convection rather than by diffusion. Chromato-
interactions. Main disadvantages of protein graphic membranes are generally cost effective,
A-containing carriers are the possible ligand and their scale-up is easier than the packed-bed
Immersion Casting 1021
chromatography (Charcosset 1998). Membrane

operations including ultraltration (Mohanty and Immersed Membrane Bioreactor
Ghosh 2008; Rosenberg et al. 2009), dialysis (IMBR)
(Bruce et al. 2002), and afnity membrane chro-
matography (Boi et al. 2009) have been demon- Submerged Membrane Bioreactor
strated for their potential for IgG purication.
Immobilization of Enzymes
References
Biocatalytic Membrane
Altnta EB, Denizli A (2009) Monosize magnetic hydro- Enzyme Compartmentalization
phobic beads for lysozyme purication under magnetic Enzymes Immobilized on Lumen
eld. Mater Sci Eng C 29:1627
Altnta EB, Tuzmen N, Uzun L, Denizli A (2007)
Immobilized metal afnity adsorption for antibody
depletion from human serum with monosize beads.
Ind Eng Chem Res 46:7802 Immersion Casting
Bakhspour M, Bereli N, enel S (2014) Preparation and I
characterization of thiophilic cryogels with
2-mercaptoethanol as the ligand for IgG purication. Francesco Galiano
Colloid Surf B 113:261 Institute on Membrane Technology, National
Bhattacharyya R, Saha RP, Samana U, Chakrabarti Research Council of Italy, ITM-CNR, Rende,
P (2003) Geometry of interaction of the histidine ring
with other planar and basic residues. J Proteome Res
Italy
2:255
Boi C, Busini V, Salvalaglio M, Cavallotti C, Sarti GC
(2009) Understanding ligand-protein interactions in Immersion casting is one of the most widely used
afnity membrane chromatography for antibody puri-
methods in preparation of polymeric membranes.
cation. J Chromatogr A 1216:86878696
Bruce MP, Boyd V, Duch C, White JR (2002) Dialysis- In this process, known as the nonsolvent-induced
based bioreactor systems for the production of mono- phase separation (NIPS) technique, the cast lm is
clonal antibodies-alternatives to ascites production in immersed in a coagulation bath, containing a
mice. J Immunol Methods 264:5968
nonsolvent, where the phase separation process
Charcosset C (1998) Purication of proteins by
membrane chromatography. J Chem Technol Biotechnol takes place (Fig. 1). The most common nonsolvent
71:95 is generally water, but aqueous solutions or pure
Denizli A, Pikin E (2001) Dye-ligand afnity systems. organic solvents such as ethanol, isopropanol, or
J Biochem Biophys Methods 49:391
butanol can be also used. The exchange
F
uglistaller P (1989) Comparison of immunoglobulin
binding capacities and ligand leakage using eight dif- (or demixing) between the solvent, contained in
ferent protein A afnity matrices. J Immunol Methods the cast lm, and the nonsolvent determines mem-
124:171 brane formation by precipitation of the polymer.
Low D, OLeary R, Pujar NS (2007) Future of antibody
The cast lm, after the immersion, separates into
purication. J Chromatogr B 848:48
Mohanty K, Ghosh R (2008) Novel tangential-ow coun- two phases: one polymer-rich phase forming the
tercurrent cascade ultraltration conguration for con- membrane matrix and one solvent-rich phase
tinuous purication of humanized monoclonal forming the pores of the membrane. Immersion
antibody. J Membr Sci 307:117125
casting technique is a process normally used for
Rosenberg E, Hepbildikler S, Kuhne W, Winter G (2009)
Ultraltration concentration of monoclonal preparing membranes with a porous and asym-
antibody solutions: development of an optimized metric structure. Generally, at lm surface, the
method minimizing aggregation. J Membr Sci pores are smaller since the phase separation occurs
342:5059
quickly; while at the bottom side, pores are bigger
Turkmen D, zturk N, Elkak A, Akgl S, Denizli A (2008)
Phenylalanine containing hydrophobic nanospheres for since the phase separation occurs slowly due to the
antibody purication. Biotechnol Prog 24:1297 tardy nonsolvent penetration. As a consequence,
1022 Immobilized Enzyme
Immersion Casting, Fig. 1 Cast lm immersed into a coagulation medium where solvent and nonsolvent exchange
occurs
the dense top layer acts as a selective barrier while The specicity of the binding makes this tech-
the porous sublayer acts as a support giving to the nique a very useful tool for the applications in
membrane the mechanical resistance. The porous which selective and strong antigen-antibody bind-
sublayer can exhibit different morphologies, such ing is advantageous. Immunoadsorption, in gen-
as spongelike or ngerlike structures, depending eral, can be used for the purpose of therapy as well
on the type of demixing. as preparative chromatography. Normally, the
A combination of factors such as the polymer human immune system works to recognize,
concentration, the nonsolvent, and the precipita- respond, and destroy pathogenic substances.
tion temperature are responsible for different When the ability of the immune systems to recog-
membrane structures. By immersion casting a nize foreign antigens versus healthy cells or tis-
range of polymeric membranes such as asymmet- sues is failed, arising immune complexes,
ric porous ultraltration membranes and asym- so-called autoantibodies, cause many kinds of
metric reverse osmosis and gas separation autoimmune diseases (Massey and McPherson
membranes, in which the top layer is completely 2007). For example, myasthenia gravis, autoim-
dense, can be produced. mune hemolytic anemia and immune thrombocy-
topenic purpura, rheumatoid arthritis, systemic
lupus erythematosus, thyroiditis, and insulin-
dependent diabetes mellitus are such diseases.
Immobilized Enzyme The immunoadsorption columns have been used
for the treatment of immune diseases since the
Biocatalytic Membrane Reactors with mid-1970s, in a study performed for the removal
Chemically Bound Enzyme of DNA antibodies (Terman et al. 1974). Since
then immunoadsorption therapy with afnity
adsorbents using target specic antibodies has
been increasingly utilized to remove pathogenic
Immunoaffinity Membranes autoantibodies from patients plasma (Uzun
et al. 2010). Besides their use in the treatment of
Nilay Bereli, Handan Yavuz and Adil Denizli autoimmune diseases, immunoafnity mem-
Department of Chemistry, Biochemistry Division, branes can be used for the purication of anti-
Hacettepe University, Ankara, Turkey bodies or antigens with a high purity and also
used for the selective and specic removal of
toxic substances from human plasma (Denizli
Immunoafnity chromatography is a process in 2002). In such applications, membrane-based col-
which the specic binding of an antigen to its umns have advantages over traditional columns in
specic antibody is utilized (Subramanian 2002). terms of compressibility of the particles, the
Impedance Spectroscopy 1023
fouling and slow ow rate through the column. Technique Description

Especially in contact with blood, stacked mem-
brane system is desirable because of high convec- Impedance spectroscopy (IS) technique measures
tive transport rates without cell damage. The other the impedance (analogous to the resistance in the
desirable properties of afnity membranes are ideal resistor that follows Ohms law) of a system
high porosity; large internal surface area; high over a wide range of frequencies. It is usually
chemical, biological, and mechanical stabilities; measured by applying an AC potential
hydrophilicity; low nonspecic adsorption of (or current) to an electrochemical cell and then
blood proteins; and the presence of functional measuring the current (or potential) response
groups for derivatization (Denizli 2011). through the system. In a linear system, the current
response to a sinusoidal potential will be a sinu-
soid at the same frequency but shifted in phase.
The corresponding expression representing the
References system impedance (Z) is indicated in Eq. 1, where
o is the radial frequency (2pf) and F is the phase
Denizli A (2002) Preparation of immuno-afnity mem- shift. This equation can be graphically represented
branes for cholesterol removal from human plasma.
in a Bode plot, see Fig. 1. In a Bode plot, a
J Chromatogr B 772:357 I
Denizli A (2011) Autoimmune diseases and logarithmic scale is used for the impedance and
immunoadsorption therapy. Hacettepe J Biol Chem frequency magnitudes:
39(3):213
Massey HD, McPherson RA (2007) Human leukocyte
sin ot
antigen: the major histocompatibility complex of man. Z Z0 (1)
In: Henrys clinical diagnosis and management by lab- sin ot f
oratory methods. 21st edn, McPherson RA, Pincus AR
(eds); Saunders-Elsevier publishes, Philadelphia pp. Also, the impedance is commonly expressed as a
876-893
Subramanian A (2002) Immunoafnity chromatography. complex function:
Mol Biotechnol 20:41
Terman DS, Stewart I, Hofmann A, Carr R, Harbeck Zo Z 0 cos f j sin f (2)
R (1974) Specic removal of DNA antibody with an
immunoadsorbent. Experientia 30:1493
Uzun L, Yavuz H, Osman B, elik H, Denizli A (2010)
This complex expression is usually represented in
PHEMA based afnity membranes for in-vitro removal a Nyquist plot. It displays the imaginary, Im(Z),
of anti-dsDNA antibodies from SLE plasma. Int J Biol and the real, Re(Z), contribution to the complex
Macromol 47:44 impedance. Figure 2 shows a Nyquist plot of the
corresponding parallel RC circuit shown as well
in the gure. The vector from the zero-point to any
Impedance Spectroscopy
Veronica Silva
Dow Water and Process Solution division at The
Dow Chemical Company, Boulevard Cecilia
Grierson 355, Buenos Aires, Argentina
Synonyms
Dielectric spectroscopy (DS); Electrochemical

impedance spectroscopy (EIS) Impedance Spectroscopy, Fig. 1 Bode plot
1024 Impedance Spectroscopy
Application to Membrane
Characterization
Impedance spectroscopy is a powerful technique

for membrane characterization. It is currently being
Z im
used with different objectives: dielectric character-

ization (Montalvillo et al. 2011, 2014), membrane
resistance determination for ion-exchange mem-
branes (Park et al. 2006), and fouling characteriza-
Abs(Z)
tion (Hu et al. 2014; Kavanagh et al. 2009), among
other specic applications.
Z real
In general, for the measurements, a membrane is
Impedance Spectroscopy, Fig. 2 Nyquist plot placed in a cell with an electrolyte solution of specic
concentration. In the cell, two electrodes are located
at each side of the membrane surface, and an alter-
nating current at a specic frequency is applied to the
system which leads to an electrical potential devel-
point of the curve denotes the magnitude |Z| and oped through the membrane. The same procedure is
the phase shift, F. In Nyquist plot the frequency repeated for a wide frequency range. In order to
values are absent, but it decreases from right to left extract the membrane electric properties from these
in the plot; each point in the plot corresponds to a data it is necessary to associate the system to an
frequency. The maximum of the semicircle equals equivalent electric circuit. For more complex sys-
occurs at such a frequency that oRC = 1, with tems, the difculty relies on setting the proper equiv-
RC = t being the corresponding relaxation time. alent circuit in order to obtain the electric parameters.
More complex systems usually present a dis- Following this procedure, the resistances and
tribution of relaxation times, and the resulting capacitances of the membrane are obtained and the
Nyquist plot appears as a depressed semicircle, thickness of the membrane can be evaluated. Also,
which is represented by a nonideal capacitor or by considering the dielectric constant of the solution
constant phase element (CPE). and the membrane polymer, many membrane prop-
When alternating current (AC) is used in the erties can be obtained as the resistance and capaci-
measurements some other parasite currents can tance values, and thus porosities and thicknesses can
appear and disturb the experiment. This is why it be obtained for each sub-layer separately as well as
is necessary to subtract these currents from the dielectric constant. The membrane is normally not
measure for each frequency. The open correc- damaged during EIS measurements (Benavente
tion is made by measuring with the cell without et al. 2005; Caas et al. 2001; Caas and Benavente
any membrane system, with the cell containing air 2002; Coster et al. 1996).
at the atmospheric pressure. The short correc-
tion has been made by using the same cell and
arrangement, but the sample is a spiral cupper Cross-References
wire in perfect contact with the electrodes without
any AgCl coating. Finally, the load correction is Impedance Spectroscopy, Membrane
obtained by using a sample with a known resis- Characterization by
tance. This open/short/load correction should be
recorded and then taken into account during the
measurements with the membrane system, in
References
order to correct the values for each frequency Benavente J, Zhang X, Valls RG (2005) Modication of
(Agilent Impedance Measurement Handbook polysulfone membranes with polyethylene glycol and
2009). lignosulfate: electrical characterization by impedance
Impedance Spectroscopy, Membrane Characterization by 1025
spectroscopy measurements. J Colloid Interface Sci The impedance spectroscopy (IS), or electro-
285:273 chemical impedance spectroscopy (EIS), is a
Caas A, Benavente J (2002) Electrochemical characterisation
of an asymmetric nanoltration membrane with NaCl and powerful technique for characterizing electrical
KCl solutions: inuence of membrane asymmetry on behavior of systems in which coupled electrical
transport parameters. J Colloid Interface Sci 246:328 processes occur at different rates (Barsoukov and
Caas A, Ariza MJ, Benavente J (2001) Characterization Macdonald 2005). The IS technique is able to
of active and porous sublayers of a composite reverse
osmosis membrane by impedance spectroscopy, measure quantitatively the electrical resistance in
streaming and membrane potentials, salt diffusion and the bulk and interfacial regions of solid and liquid
X-ray photoelectron spectroscopy measurements. electrolyte materials, including membranes.
J Membr Sci 183:135 IS technique is widely used in the in situ non-
Coster HGL, Chilcott TC, Coster ACF (1996) Impedance
spectroscopy of interfaces, membranes and ultrastruc- destructive characterization of membranes, as
tures, a review. Bioelectrochem Bioenerg 40:79 well as to investigate concentration polarization
Hu Z, Antony A, Leslie G, Le-Clech P (2014) Real-time and fouling phenomena (Fontananova et al. 2012;
monitoring of scale formation in reverse osmosis using Antony et al. 2013).
electrical impedance spectroscopy. J Membr Sci
453:320327 In EIS experiments, a sinusoidal electrical
Kavanagh JM, Hussain S, Chilcott TC, Coster HGL stimulus (voltage or current) is applied over a
(2009) Fouling of reverse osmosis membranes using frequency range to a pair of electrodes, and the
electrical impedance spectroscopy: measurements and I
response of the system under investigation is
simulations. Desalination 236:187193
Makoto Honda. Agilent Technologies, Inc. 2009-2013. observed by the same or different electrodes. In
Agilent impedance measurement handbook. In: the rst case, the conguration is indicated as a
Agilent Technologies (ed) A guide to measurement two-probe (or two electrodes) type (Fig. 1a). If
technology and techniques. USA two additional electrodes are used to collect the
Montalvillo M, Silva V, Palacio L, Hernndez A, Prdanos P
(2011) Dielectric properties of electrolyte solutions in response of the system, the conguration is indi-
polymeric nanoltration membranes. Desalination cated as four-probe (or four electrodes) type
Water Treat 27:2530 (Fig. 1b). Another possible conguration uses
Montalvillo M, Silva V, Palacio L, Calvo JI, Carmona FJ, three electrodes, but it is usually employed to
Hernndez A, Prdanos P (2014) Charge and dielectric
characterization of nanoltration membranes by characterize only one half of an electrochemical
impedance spectroscopy. J Membr Sci 454:163173 cell, or phenomena occurring on an electrode and
Park J-S, Choi J-H, Yeon K-H, Moon S-H (2006) An it will not be discussed here. The two-probe con-
approach to fouling characterization of an guration is usually applied when the membrane
ion-exchange membrane using currentvoltage relation
and electrical impedance spectroscopy. J Colloid Inter- is pressed between two solid conductive elec-
face Sci 294:129138 trodes, like in the case of the membrane electrode
assembly (MEA) for fuel cells (Fontananova
et al. 2012).
On the contrary, if the membrane is in contact
Impedance Spectroscopy, Membrane with a liquid electrolyte, the four-probe congu-
Characterization by ration is the most convenient and commonly used
(Antony et al. 2013). This second conguration
Enrica Fontananova has the advantage to eliminate the contribution of
Institute on Membrane Technology, National the electrode injecting stimulus/electrolyte charge
Research Council of Italy, ITM-CNR, Rende, transfer resistance, from the impedance spectra,
Italy focusing the probing on the membrane and its
interfaces.
The sinusoidal electrical stimulus is injected
Synonyms through two planar electrodes (working and coun-
ter electrode), and the response of the system to
Electrochemical impedance spectroscopy; Imped- the sine wave perturbation is measured by two
ance spectroscopy reference electrodes (indicated as sense and
1026 Impedance Spectroscopy, Membrane Characterization by
Impedance Spectroscopy, Membrane Characterization by, Fig. 1 Schematic view of the experimental setup used
for membrane characterization by impedance spectroscopy using two-probes (a) and four-probes conguration (b)
reference) using an impedance analyzer (usually a Z0 jZ j cos (5)

potentiostat/galvanostat combined with a fre-
quency response analyzer) which measures volt- and the imaginary part:
age, current, and phase shift (Fig. 1).
In analogy to Ohms law, the impedance is Z00 jZj sin (6)
dened as (Barsoukov and Macdonald 2005):
The real part of the impedance is the resistance
U o (Z0 ); the imaginary part is called reactance (Z00 ).
Zo (1)
I o On the contrary of the measurement in direct
current (DC), using an alternate current (AC) over
where Z(o) [] is the impedance, U(o) [V] is the a frequency range, it is possible to distinguish
voltage drop, I(o) [A] is the current, and they phenomena proceeding at different rates, like
depend on the circular velocity or circular fre- bipolar concentration polarization of an ion
quency o [rad/sec] as follows: exchange membrane (IEM) in contact with an
electrolyte solution, which induces the formation
U o Uo sin ot (2) of an electrical double layer (EDL) and a diffusion
boundary layer (DBL) (Sang et al. 2008).
I o I o sin ot (3) As a consequence, the total electrical resistance
is the sum of the contribution for the transport
where o = 2pu and u [sec1] is the frequency, through the membrane and the interfaces.
t [sec] is the time, [ ] is the phase shift between The bipolar concentration polarization during an
voltage and current, and the subscript refers to AC cycle, caused by the buildup and depletion of
the amplitude of voltage and current
p
in phase; j is ions at the interfaces, is time dependent. The contri-
the imaginary number j 1. bution of the interfacial ionic charge transfer through
The impedance can be also rearranged as the interfaces layers at low frequencies is greater
follows: than at high frequencies, because at high frequencies
there is insufcient time for their formation.
Zo jZj cos jjZ j sin (4) The experimental impedance data can be tted
with equivalent electrical circuit models able to
The Eq. 4 indicates that the impedance is com- represent the physical system and phenomena
posed of two parts, i.e., the real part: under investigation (Zhang and Spichiger 2000).
Imprinted Composite Membranes 1027
a b
Rct Rm+s Redl Rdbl
Rm
C C
Cct Cedl CPEdbl

Z () Z ()
Rm Rm+Rct Rm+s Rm+s+Redl Rm+s+Redl+Rdbl

Z () Z ()
Impedance Spectroscopy, Membrane Characteriza- models are also shown. The resistor is indicated as R, the
tion by, Fig. 2 Impedance spectra reported as imaginary capacitor as C, and the constant phase element (a nonideal
(Z00 ) versus real part of the impedance (Z0 ) for an ion capacitor) as CPE. The subscript m is referring to the
exchange membrane: pressed between two solid electrodes membrane, ct to the charge transfer between electrode
(a) and separating two electrolyte solutions (b). The spec- and membrane, m+s to membrane plus solution, edl to
tra are registered respectively with the two- and four-probe the electrical double layer, and dbl to the diffusion I
conguration, and the corresponding equivalent circuit boundary layer
The typical shape of the impedance spectra (Z00 the preparation conditions on the properties of poly-
vs. Z0 ) in the case of an IEM pressed between two meric and hybrid cation exchange membranes.
Electrochim Acta 66:164172
solid electrodes exhibits a distinct arc (Fig. 2a). In Sang S, Wu Q, Huang K (2008) A discussion on ion
the case of an IEM separating two liquid electro- conductivity at cation exchange membrane/solution
lyte solutions, additional semicircles appear in the interface. Colloids Surf A 320:4348
spectra (Fig. 2b). The corresponding equivalent Zhang Z, Spichiger UE (2000) An impedance study on
Mg2+ selective membrane. Electrochim Acta
circuits are also shown in Fig. 2. 45:22592266
At high frequencies (o!1) the intercept on
the real axis gives the membrane (Rm) or mem-
brane plus solution resistance (Rm+s). Of course,
the solution resistance can be obtained by blank Imprinted Composite Membranes
experiments (without the membrane) and its con-
tribution can be subtracted to obtain the pure Laura Donato
membrane resistance. Institute on Membrane Technology, National
At low frequencies (o!0), the intercept on the Research Council of Italy, ITM-CNR, Rende,
real axis gives the sum of membrane (plus solu- Italy
tion, if present) and the interface resistances.
Imprinted composite membranes represent one of

References the various kinds of molecularly imprinted mem-
branes. They are produced in a form of thin lms or
Antony A, Chilcott T, Coster H, Leslie G (2013) In situ
structural and functional characterization of reverse hollow bers from at least two different materials.
osmosis membranes using impedance spectroscopy. The preparation of imprinted composite membranes
J Membr Sci 425426:8997 involves the surface imprinting of a preexisting sup-
Barsoukov E, Macdonald JR (2005) Impedance spectros- port membrane with a thin imprinted polymer layer
copy. Theory, experiment, and applications, 2nd edn.
Wiley, New Jersey (Xu et al. 2009). This method allows to introduce
Fontananova E, Cucunato V, Curcio E, Trotta F, specic recognition sites into the support membrane
Biasizzo M, Drioli E, Barbieri G (2012) Inuence of and therefore to increase its transport selectivity. In
1028 Imprinted Composite Membranes
Imprinted Composite Membranes, Fig. 1 Schematic representation of the two surface imprinting strategies
addition, the mechanical integrity of the original exhibit higher performance with respect to mem-
membrane is preserved. The preparation of these branes directly prepared with the imprinted poly-
membranes is mainly performed by exploiting two mer (Donato et al. 2013).
different polymerization strategies: the surface The rst imprinted composite membrane was
grafting via photopolymerization and the surface prepared in 1997 (Wang et al. 1997) using as
coating via thermopolymerization. The rst route support matrix a poly(acrylonitrile) membrane.
allows the copolymerization of two or more A molecularly imprinted layer on the membrane
photopolymerizable monomers using the light as surface was formed by means of the UV-initiated
energy source. In the case of the thermopoly- photo-copolymerization of acrylic acid and N,N
merization, the energy source is represented by the methylenebisacrylamide. Theophylline was used
heat. Depending upon the followed route, photo- or as template molecule.
redox initiators are used to activate the original Due to the mechanical and chemical stability in
membrane surface through the formation of radical combination with high performance, imprinted
sites. The polymerization proceeds by the addition composite membranes have a large potential
of monomer molecules to the free radical ends of the application, like in biosensor technology, in solid
growing polymer chains. phase extraction, in microltration, and in enan-
Figure 1 reports a schematic representation of tiomeric separation.
the two surface imprinting strategies.
Composite thin imprinted membranes are also
prepared by simple deposition or of imprinted
polymer particles on the surface of microltration
References
membranes. Cross-ow ltration can be also used Donato L, Tasselli F, De Luca G, Del Blanco SG, Drioli E
to form the surface polymer layer. (2013) Novel hybrid molecularly imprinted mem-
A particular type of composite membranes is branes for targeted 4,4-methylendianiline. Sep Purif
prepared by means of the hybrid molecular Technol 116:184191
Ulbricht M (2004) Membrane separations using molecularly
imprinting. This approach envisages the incor- imprinted polymers. J Chromatogr B 804:113125
poration of a previously synthesized imprinted Wang HY, Kobayashi T, Fuji N (1997) Surface molecular
polymer particles into a typically used polymer imprinting on photosensitive dithio-carbamoyl polyac-
matrix and the subsequent membrane formation rylonitrile membrane using photo graft polymerization.
J Chem Technol Biotechnol 70:355362
via the phase inversion technique (Ulbricht 2004). Xu Z, Huang X, Wan L (2009) Surface engineering of
These kinds of membranes (also called hybrid polymer membranes. Zhejiang Universitys/Springer,
imprinted membranes) are highly selective and Berlin
Imprinted Polysulfone-Aldehyde Derivatized Nanofiber Membranes 1029
imprinting so that molecular recognition sites,

Imprinted Polysulfone-Aldehyde which specically incorporate target substrate
Derivatized Nanofiber Membranes into the membrane, can be introduced into a
given membrane. However, the enhancement
Masakazu Yoshikawa of ux without a concurrent reduction in
Department of Biomolecular Engineering, Kyoto permselectivity is perceived to be an unsolved
Institute of Technology, Kyoto, Japan problem or an unsolvable problem in mem-
brane separation. In other words, ux and
permselectivity often show a trade-off relation-
The functionalization of aromatic polysulfones for ship. Membranes with high surface area and
tailoring properties in membrane applications is of high porosity would be expected to break
great interest. Polysulfone has overall thermal and such a trade-off relationship between ux and
chemical stability combined with good mechanical permselectivity. Nanober membranes are
and membrane-making qualities. Polysulfone is a expected to simultaneously give both high ux
stable platform for functional group attachment and and high permselectivity. To this end, nanober
a good candidate polymeric material for mem- membranes with molecular recognition sites,
branes with tailored functionalities (Guiver which are called molecularly imprinted nanober
I
et al. 1999). To this end, modied polysulfones membranes, were fabricated by simultaneously
have been intensively studied in connection with applying an alternative molecular imprinting and
chiral separation (Yoshikawa et al. 1998, 2005, an electrospray deposition (Sueyoshi et al. 2010;
2006, 2007; Mizushima et al. 2011; Sueyoshi Yoshikawa et al. 2007). Those studies revealed
et al. 2012), pervaporation separation (Yoshikawa that molecularly imprinted nanober membranes
et al. 1992a, b, 1999), and selective separation of gave high ux without a concurrent reduction in
CO2 (Yoshikawa et al. 2000). permselectivity. A breakthrough in membrane sep-
In membrane separation, both ux and aration was attained; in other words, membrane
permselectivity are important factors. The morphology in the form of molecularly imprinted
enhancement of permselectivity would be rela- nanober fabric was one of the suitable membrane
tively easily attained by application of molecular forms to solve a trade-off relationship in membrane
Imprinted Polysulfone-
Aldehyde Derivatized
Nanofiber Membranes,
Fig. 1 Chemical structures
of polysulfone with
aldehyde group (PSf-CHO)
and print molecule (Z-D-Glu
or Z-L-Glu) (Cited from
Sueyoshi et al. 2012 with
permission. Copyright 2012
Elsevier Inc)
1030 Imprinted Polysulfone-Aldehyde Derivatized Nanofiber Membranes
separation. Molecularly imprinted nanober mem- PSf-CHO with degree of substitution of 0.50
branes and usual molecularly imprinted mem- and 1.00 were adopted as candidate materials,
branes were fabricated from polysulfone with and the derivative of D- or L-glutamic acid
aldehyde group (PSf-CHO) and print molecules, was applied as a print molecule to obtain molec-
and their membrane performances, such as adsorp- ularly imprinted nanober membranes and
tion selectivity, permselectivity, and ux, were usual molecularly imprinted membranes for
studied (Sueyoshi et al. 2012). optical resolution (Figs. 1 and 2).
Imprinted Polysulfone-
Aldehyde Derivatized
Nanofiber Membranes,
Fig. 2 Schematic
illustration for the
fabrication of molecularly
imprinted nanober
membranes, where
PSf-CHO and Z-Glu were
simultaneously
electrosprayed
2.0
D-Glu
L-Glu
105 [Glu]R / mol dm3
1.5 NaN3 NaN3
L R
membrane
1.0
a L/D = 1.20
0.5
D-Glu
L-Glu
0
0 2.0 4.0 6.0 8.0 10.0
Time / h
Imprinted Polysulfone-Aldehyde Derivatized Imprinted Polysulfone-Aldehyde Derivatized

Nanofiber Membranes, Fig. 3 Adsorption isotherm of Nanofiber Membranes, Fig. 4 Time-transport curves
D-Glu and L-Glu in the nanober membrane molecularly of racemic Glus through the nanober membrane molecu-
imprinted by Z-D-Glu. (PSf-CHO-10 was adopted as a larly imprinted by Z-D-Glu (PSf-CHO-10 was adopted as a
candidate material) (Cited from Sueyoshi et al. 2012 with candidate material) (Cited from Sueyoshi et al. 2012 with
permission. Copyright 2012 Elsevier Inc) permission. Copyright 2012 Elsevier Inc)
Imprinted Polysulfone-Aldehyde Derivatized Nanofiber Membranes 1031
Imprinted Polysulfone-Aldehyde Derivatized Nanofiber Membranes, Table 1 Results of chiral separation with
molecularly imprinted nanofiber (MINFMs) and molecularly imprinted (MIPMs) membranes
Z-D-Glu imprinted membrane Z-L-Glu imprinted membrane
Mmebrane aL/D u0 aD/L u0
MINFM-10a 1.24 1.15 109 (28) 1.20 1.67 109 (41)
MIPM-10a 1.20 4.20 1011 (1) 1.20 4.10 1011 (1)
MINFM-05b 1.12 7.00 109 (231) 1.20 2.20 109 (72)
MIPM-05b 1.25 6.64 1011 (2.2) 1.16 3.05 1011 (1)
a
Figures in parentheses are the relative values; the U Value for MIPM-10 imprinted by Z-L-Glu being set as unity
b
Figures in parentheses are the relative values; the U Value for MIPM-05 imprinted by Z-L-Glu being set as unity
c
U = (J/C)/(dm/dx) [{(mol cm cm2 h1)/(mol cm3)}/(J mol1 cm1) = mol cm cm2 J1 h1].
(Cited from ref. Sueyoshi et al. 2012 with permission. Copyright 2012 Elsevier lnc.)
The membranes molecularly imprinted by As proved in the previous studies (Sueyoshi

Z-D-Glu incorporated the D-isomer of glutamic et al. 2010; Yoshikawa et al. 2007), the present
acid (Glu) in preference to the corresponding study revealed that molecularly imprinted
L-isomer and vice versa. In other words, the mem- nanober membranes gave high ux without
brane imprinted by the L-isomer selectively depression of permselectivity. The emergence of
I
adsorbed the L-isomer. molecularly imprinted nanober membrane
Figure 3 shows the adsorption isotherms of would solve a trade-off relationship in membrane
D-Glu and L-Glu for the Z-D-Glu molecularly separation (Yoshikawa et al. 2011).
imprinted nanober membrane as an example of
adsorption isotherms. The adsorption isotherm of
D-Glu, which was preferentially adsorbed in the
membrane, shows a dual adsorption isotherm. It
consists of nonspecic adsorption and specic
References
adsorption on the specic recognition site, which
Guiver MD, Robertson GP, Yoshikawa M, Tan CM (1999)
was constructed by the presence of a print mole- Functionalized polysulfones: methods for chemical
cule during the membrane preparation process. modication and membrane applications. In: Pinnau I,
Contrary to this, L-Glu, which was nonspecically Freeman BD (eds) Membrane formation and modica-
tion. ACS symposium series, vol 744. ACS, Washing-
adsorbed in the membrane, gives a straight line
ton, DC, pp 137161
passing through the origin. Mizushima H, Yoshikawa M, Robertson GP, Guiver MD
Time-transport curves of racemic Glu through (2011) Optical resolution membranes from
the D-isomer molecularly imprinted nanober polysulfones bearing alanine derivatives as chiral selec-
tors. Makromol Mater Eng 296:562567
membrane are shown in Fig. 4. As often observed,
Sueyoshi Y, Fukushima C, Yoshiakwa M (2010) Molecu-
the transport of the enantiomer preferentially incor- larly imprinted nanober membranes from cellulose
porated into the membrane was retarded by a rela- acetate aimed for chiral separation. J Membr Sci
tively strong interaction between the enantiomer 357:9097
Sueyoshi Y, Utsunomiya A, Yoshiakwa M, Robertson GP,
and the membrane. As a result, the antipode was
Guiver MD (2012) Chiral separation with molecularly
selectively transported. Such discrepancy between imprinted polysulfone-aldehyde derivatized nanober
adsorption selectivity and permselectivity is often membranes. J Membr Sci 401402:8996
observed in chiral separation. Yoshikawa M, Hara H, Tanigaki M, Guiver M, Matsuura T
(1992a) Modied polysulfone membranes: 1.
Table 1 summarizes membrane performances Pervaporation of water/alcohol mixtures through mod-
for two types of molecularly imprinted mem- ied polysulfone membranes having methyl ester moi-
brane. As can be seen, the uxes through the ety. Polymer 33:48054813
molecularly imprinted nanober membranes Yoshikawa M, Hara H, Tanigaki M, Guiver M, Matsuura T
(1992b) Modied polysulfone membranes II.
gave one to two orders of magnitude higher than
Pervaporation of aqueous ethanol solution through
those of usual molecularly imprinted membranes modied polysulfone membranes bearing various
without depression of permselectivity. hydroxyl groups. Polym J 24:10491055
1032 Imprinting
Yoshikawa M, Izumi J, Ooi T, Kitao T, Guiver MD, Rob- molecule, is memorized in the polymeric
ertson GP (1998) Carboxylated polysulfone mem- membranes (materials) for the recognition or
branes having a chiral recognition site induced by an
alternative molecular imprinting technique. Polym Bull separation of target molecule from others during
40:517524 the formation of polymeric membranes
Yoshikawa M, Tsubouchi K, Guiver MD, Robertson GP (materials).
(1999) Modied polysulfone membranes. III. Such molecularly imprinted materials are pre-
Pervaporation separation of benzene-cyclohexane mix-
tures through carboxylated polysulfone membranes. pared by adopting two ways; one is molecular
J Appl Polym Sci 74:407412 imprinting, the other alternative molecular
Yoshikawa M, Niimi A, Guiver MD, Robertson GP (2000) imprinting. The former is a pioneering method
Modied polysulfone membranes. IV. Gas separation to prepare polymeric materials with molecular
with aminated polysulfone membranes. Seni
Gakkaishi 56:272281 recognition sites from functional monomer,
Yoshikawa M, Hanaoka K, Guiver MD, Robertson GP crosslinker, and print molecule (template) (Wulff
(2005) Chiral separation of racemic amino acids and Sarhan 1972; Arshady and Mosbach 1981);
through membranes derived from modied polysulfone the molecular imprinting is further divided into
having perillaldehyde moiety as a side group. Mem-
brane 30:219225 two methods, covalent molecular imprinting
Yoshikawa M, Murakoshi K, Kogita T, Hanaoka K, (Wulff and Sarhan 1972) and non-covalent molec-
Guiver MD, Robertson GP (2006) Chiral separation ular imprinting (Arshady and Mosbach 1981) as
membranes from modied polysulfone having schematically shown in Fig. 1. The latter is an
myrtenal-derived terpenoid side groups. Eur Polym J
42:25322539 alternative way to obtain polymeric membranes
Yoshikawa M, Nakai K, Matsumoto H, Tanioka A, Guiver bearing molecular recognition sites directly from
MD, Robertson GP (2007) Molecularly imprinted candidate polymeric materials and print molecule
nanober membranes from carboxylated polysulfone (Yoshikawa et al. 1995, 2011). The scheme of the
by electrospray deposition. Macromol Rapid Commun
28:21002105 alternative molecular imprinting is shown in
Yoshikawa M, Tanioka A, Matsumoto H (2011) Molecu- Fig. 2. In Step 1, the polymeric material, which
larly imprinted nanober membranes. Curr Opin Chem is a candidate material to construct molecular
Eng 1:1826 recognition sites, is interacted with a print mole-
cule by specic interaction before and during the
formation process of molecular recognition mate-
rials so that molecular memory can be introduced
into the polymeric materials. In Step 2, the print
Imprinting molecule is extracted from the molecularly
imprinted materials. When the molecularly
Masakazu Yoshikawa1 and Kalsang Tharpa2 imprinted material thus constructed is in contact
1
Department of Biomolecular Engineering, Kyoto with the print molecule or print molecule ana-
Institute of Technology, Kyoto, Japan logue, the molecular recognition sites preferen-
2
Department of Chemistry, University of Mysore, tially interact with them or incorporate them into
Mysore, India the molecular recognition sites (Step 3 and Step
4). Contrary to the pioneering molecular imprint-
ing method, molecular recognition sites are
formed at the same time as the molecularly
Imprinting, which is often called molecular imprinted materials are prepared from polymer
imprinting, is a facile way to introduce molecular solution or polymer melt. In other words, any
recognition sites into polymeric membranes polymeric materials, such as synthetic polymers,
(materials) (Sellergren 2001; Komiyama oligopeptide derivatives, derivatives of natural
et al. 2003; Alexander et al. 2006). In other polymer, and natural polymers, can be directly
words, the molecular memory, such as a shape of converted into molecular recognition material by
the target molecule and an alignment of the applying the alternative molecular imprinting
functional moieties to interact with those in target (Yoshikawa 2001).
Imprinting 1033
Covalent Molecular Imprinting
Synthesis of Removal by
Polymerizable Chemical Molecular
Print Molecule Polymerization Cleavage Recognition
Noncovalent Molecular Imprinting

Removal by
Solvent Molecular
Self-assembly Polymerization Extraction Recognition
Imprinting, Fig. 1 Schemes of the covalent and non-covalent molecular imprinting (Cited from Yoshikawa et al. 2011
with permission. Copyright 2012 Elsevier Inc.)
I
Imprinting, Alternative Molecular Imprinting
alternative molecular
imprinting (Cited from
Yoshikawa et al. 2011 with
permission. Copyright 2012
Elsevier Inc.)
Step 2
Step 1 Removal of print molecule
Interaction between print
molecule and candidate
material
Step 4 Step 3
Recognition of print molecule Formation of recognition
(print molecule analogue) site and permeation path
The similar approach was proposed by Since then, various molecularly imprinted mem-
Michaels et al. in 1962 (Michaels et al. 1962). branes were studied by adopting non-covalent
This study is the rst report on the alternative molecular imprinting. A wet phase inversion
molecular imprinting and the rst application of process was applied to an alternative molecular
molecularly imprinted polymeric membranes to imprinting to prepare asymmetric membranes
membrane separation. Michaels paper is the (Trotta et al. 2002).
commemorable paper in molecular imprinting As described above, applying molecular
and membrane separation. In addition to this, imprinting, such as conventional molecular
molecularly imprinted polymeric membranes imprinting or alternative molecular imprinting,
prepared by non-covalent molecular imprinting molecular recognition sites are easily introduced
was reported in 1990 (Piletskii et al. 1990). into polymeric membranes (Ulbricht 2004). From
1034 Industrial/Tannery Wastewater Treatment Using Membrane Bioreactors
this, it is easy to enhance permselectivity of a Yoshikawa M (2001) Molecularly imprinted polymeric

given membrane by applying those molecular membranes. Bioseparation 10:277286
Yoshikawa M, Izumi J, Kitao T, Koya S, Sakamoto S
imprinting techniques. The enhancement of ux (1995) Molecularly imprinted polymeric membranes
without a reduction in permselectivity is indis- for optical resolution. J Membr Sci 108:171175
pensable so that molecularly imprinted mem- Yoshikawa M, Tanioka A, Matsumoto H (2011) Molecu-
branes can be applicable to industries. larly imprinted nanober membranes. Curr Opin Chem
Eng 1:1826
Molecularly imprinted membranes with a
higher surface area and a higher porosity are
required to give higher ux and permselectivity.
Electrospun nanober membranes with molecular
recognition sites is a suitable or the best mem-
Industrial/Tannery Wastewater
brane morphology to attain high ux and high
Treatment Using Membrane
permselectivity. Possibility of the enhancement
Bioreactors
of ux without a concurrent reduction in
Membrane Bioreactors for Treatment of
permselectivity was proved by molecularly
Tannery Efuents
imprinted nanober membranes, which were fab-
ricated by simultaneously applying an
electrospray deposition and an alternative molec-
ular imprinting (Yoshikawa et al. 2011).
Inert Membrane
References Rune Bredesen

Sustainable Energy Technology, SINTEF
Allexander C, Andersson HS, Andersson LI, Ansell RJ, Materials and Chemistry, Oslo, Norway
Kirsch N, Nicholls IA, OMahony J, Whitcombe MJ
(2006) Molecular imprinting science and technology: a
survey of the literature for the years up to and including
2003. J Mol Recognit 19:106180 The term inert membrane denotes that no change
Arshady R, Mosbach K (1981) Synthesis of substrate- in chemical reaction occurs due to contact
selective polymers by host-guest polymerization. between the membrane material and the surround-
Makromol Chem 182:687692
Komiyama M, Takeuchi T, Mukawa T, Asanuma H (2003)
ing constituents. A reaction occurring between
Molecular imprinting. Wiley-VCH, Weinheim constituents A and B to form C and D can be
Michaels AS, Baddour RF, Bixler HJ, Choo CT (1962) used as an example:
Conditioned polyethylene as a permselective mem-
brane. Separation of isomeric xylenes. Ind Eng Chem
Process Des Dev 1:1425
ABCD (1)
Piletskii SA, Dubei IY, Fedoryak DM, Kukhar VP (1990)
Substrate-selective polymeric membranes Selective If the membrane is catalytically inactive with
transfer of nucleic acid components. Biopolim Kletka respect to the reaction, the term inert membrane
6:5558
Sellergren B (ed) (2001) Molecularly imprinted polymers
is used to describe the membrane. The presence of
man made mimics of anti bodies and their applications a catalyst, deemed not to be part of the membrane
in analytical chemistry. Elsevier, Amsterdam material, may yield a change in reaction 1, and the
Trotta F, Drioli E, Baggiani C, Lacopo D (2002) J Membr integration of such a combination of membrane
Sci 201:7784
and catalyst is referred to as an inert membrane
Ulbricht M (2004) Membrane separations using molecu-
larly imprinted polymers. J Chromatogr reactor (IMR). The term inert membrane is there-
B 804:113125 fore commonly used in connection with mem-
Wulff G, Sarhan A (1972) The use of polymers with brane reactors (Koros et al. 1996) to distinguish
enzyme-analogous structures for the resolution of race-
mates. Angew Chem Int Ed 14:341 [ber die
the membrane properties from those of a catalytic
Anwendung von enzymanalog gebauten Polymeren membrane, the latter being catalytically active. In
zur Racemattrennung. Angew Chem 84: 364] such reactors, the inert membrane may be used for
Inert Membrane Reactors 1035
selectively separating reaction products from the inert membrane or bio-inert membrane. These
reaction for which the catalyst serves to activate. expressions refer to a specic property of the
Alternatively, the inert membrane may be membrane such as chemical stability or compati-
employed as a distributor of reactants to ensure bility in the case of chemical inertness or biolog-
controlled delivery to the catalytic reaction site. ical inactivity in the case of bio-inertness. In these
Classication of a membrane as inert may also cases the term inert membrane has a somewhat
depend on the operation conditions and surface different meaning than that related to the IMRs.
properties since the actual catalytic activity
depends on parameters such as temperature, sur-
face area, and surrounding chemical composition. References
Thus the same membrane may, or may not, be an
inert membrane depending on the conditions Koros WJ, Ma YH, Shimidzu T (1996) Terminology for
membranes and membrane processes. Pure Appl Chem
under which it is operated.
68:14791489
The interactions between an inert membrane Mulder M (1996) Basic principles of membrane technol-
and its surroundings typically involve surface ogy. Kluwer Academic Publishers, Dordrecht, The
adsorption/desorption reactions, which may be Netherlands
Sirman J (2006) The evolution of materials and architec-
followed by other reactions necessary for trans-
ture for oxygen transport membranes. In: Sammells AF, I
port of matter within the membrane. For example, Mundschau MV (eds) Nonporous inorganic mem-
in the case of dense polymeric membranes, incor- branes. Wiley-VCH, Weinheim, pp 165184
poration of the permeant gas molecule is required Ward TL, Dao T (1999) Model of hydrogen permeation
behavior in palladium membranes. J Membr Sci
on the feed side, while the reverse process is
153:211231
required on the permeate side (Mulder 1996).
Dense inorganic membranes, in addition, require
transformation of the adsorbed molecule to atomic
(in the case of metal membranes (Ward and Dao
1999)) or ionic and electronic (in the case of Inert Membrane Reactors
ceramic membranes (Sirman 2006) species at the
feed side, which are then able to diffuse through Rune Bredesen
the bulk membrane phase. At the permeate side, Sustainable Energy Technology, SINTEF
the recombination of species to the same molecu- Materials and Chemistry, Oslo, Norway
lar form as on the feed side takes place before
desorption to the gas phase. To enhance the trans-
formation of adsorbed gas molecules to diffusing A membrane, dened as a barrier between two
species within the membrane and, thus, contribute phases through which transport of one or several
to higher uxes, catalytic surface properties are species occurs, can be made from virtually any
usually required. Nevertheless, such membranes solid or liquid material or combinations of both.
which although they incorporate catalytic surface Membranes are commonly divided into inorganic,
reactions, they are regarded as inert membranes polymeric, hybrid inorganic and polymeric, or as
since the reaction is present solely as a means of dual phase consisting of a solid phase and a liquid
sustaining transport of the gas molecule from one phase. Furthermore, the membrane may either be
membrane side to the other. Another example is dense or porous with a continuous network of
inert membranes for liquid separation applica- pores.
tions, where surface hydrophilicity and hydropho- A membrane reactor (MR) is a device for
bicity may completely determine the membrane simultaneously carrying out a reaction and
transport properties. The term inert membrane is membrane-based separation in the same physical
rarely used for conventional membrane separation enclosure (Koros et al. 1996). In an inert mem-
processes; however, one may nd the term inert brane reactor (IMR), the inert membrane (link)
membrane used in such phrases as chemically and the catalyst in the form of a separate solid or
1036 Inert Membrane Reactors
Inert Membrane Reactors, Fig. 1 Inert membrane (a) with solid catalyst, (b) with liquid containing the catalyst, (c)
encapsulating solid catalyst, (d) encapsulating liquid containing the catalyst
Inert Membrane Reactors, Fig. 2 (a) Membrane extractor operation, (b) membrane distributor operation
a liquid phase are contained in the reactor reactions). Extraction of hydrogen from hydrocar-
(Fontananova and Drioli 2010; Julbe et al. 2001; bon dehydrogenation, reforming, or water gas
Coronas and Santamaria 1999). Figure 1 illus- shift reactions, employing hydrogen selective
trates different catalyst and inert membrane com- membranes in combination with packed or uid-
binations in an IMR. Due to the separation of ized catalyst beds, has been widely studied
membrane and catalyst in the IMR, the (Sanchez Marcano and Tsotsis 2002). Dehydro-
membrane-based process and catalytic reactions genation and reforming are typically equilibrium-
occur in sequence. This decoupling of the pro- limited endothermic reactions, and membrane
cesses can be advantageous with respect to oper- reactor applications may benet from the use of
ation and replacement of membrane and/or lower operating temperature and/or higher pres-
catalyst compared to MRs where the membrane sure without sacricing yield. Esterication is yet
serves as both catalyst and membrane. The two another example of catalyzed equilibrium-limited
main functions of the membrane applied in IMRs reactions where water extraction by an inert mem-
are as an extractor or a distributor (Julbe brane is used to increase yield (Van der Bruggen
et al. 2001; Dalmon 1997) (see Fig. 2). As an 2010). As a distributor, the inert membrane
extractor, the inert membrane selectively sepa- delivers the reactant in a controlled manner to
rates a reaction product or intermediate product. the catalyzed reaction taking place in the reactor
The advantage may be higher conversion compartment (Julbe et al. 2001). The typical aim
(equilibrium-limited reactions) or/and higher is to enhance selectivity through careful addition
selectivity (e.g., via extraction of an intermediate and temperature control of exothermic oxidation
that would otherwise lead to subsequent unwanted and hydrogenation reactions. Some common
Inhibition Coefficient (IC) 1037
examples are IMR with packed catalyst bed for Sanchez Marcano JG, Tsotsis TT (2002) Catalytic mem-
oxidative coupling of methane to C2, or oxidative branes and membrane reactors. Wiley-VCH, Weinheim
Van der Bruggen B (2010) Pervaporation membrane reac-
dehydrogenation of hydrocarbons, by addition of tors. In: Drioli E, Giorno L (eds) Comprehensive mem-
oxygen and hydrogenation of alkenes by hydro- brane science and engineering, vol 3. Elsevier,
gen addition. For some oxidation reactions Amsterdam, pp 135163
employing IMRs, the separation of the bulk of
the reactants by the membrane wall lowers the
explosion potential (Coronas and Santamaria
1999). Inhibition Coefficient (IC)
As most membrane materials are inert, the
incorporation of many different types into IMRs Alessio Caravella
has been studied. Cheap polymeric membranes National Institute of Advanced Industrial Science
are advantageous with respect to capital cost, and Technology Research Institute for
high packing density, and simple sealing technol- Innovation in Sustainable Chemistry, ISC-AIST,
ogy in modules. However, their low operating Tsukuba, Ibaraki, Japan
temperature, typically less than 100 C, and lim-
ited chemical stability narrow the range of appli-
In membrane technology, the inhibition coef-
I
cations. More expensive inorganic membranes
enable high temperature operation, but since sta- cient (IC) refers to a coefcient indicating quanti-
bility and transport properties are usually very tatively the reduction of permeability (or,
temperature dependent, the different inorganic equivalently, reduction of permeance or ux) of
membranes have a limited window of operation. a permeating species because of one or more
Several decades of research have shown that a inhibiting species (see the entry Inhibition to
number of challenges still exist with respect to Permeation for denition of inhibition). The
high temperature applications, and commercial most general denition of inhibition coefcient
use of IMRs is still a future prospect. To aid can be provided as follows:
reaching this future prospect, mathematical
modeling and simulation is required to develop
FluxInhib
i 1 ICi FluxClean
i (1)
IMR design and provide an understanding of the
optimal operating conditions.
Equation 1 simply states that the ux in the pres-
ence of inhibition (indicated by the superscript
Inhib) is equal to the clean (i.e., inhibition-
References
free) ux reduced by a certain factor whose
Coronas J, Santamaria J (1999) Catalytic reactors based value depends on the inhibition degree.
on porous ceramic membranes. Catal Today According to this denition, ICi approaches the
51:377389 unity value (ux tending to zero) for maximum
Dalmon JA (1997) Catalytic membrane reactors. In: Ertl G,
inhibition, whereas it approaches zero for mini-
Knzinger H, Weitkamp J (eds) Handbook of hetero-
geneous catalysis, vol 3. Wiley-VCH Weinheim, Ger- mum inhibition (ux equal to clean ux). If the
many, pp 13871398 clean permeation ux of the generic species i in
Fontananova E, Drioli E (2010) Catalytic membranes and mixture can be described by Eq. 2:
membrane reactors. In: Drioli E, Giorno L (eds) Com-
prehensive membrane science and engineering,
vol 3. Elsevier, Amsterdam, pp 109133 PeClean
Julbe A, Farrusseng D, Guizard C (2001) Porous ceramic
FluxClean
i i
Driving Forcei (2)
d
membranes for catalytic reactors overview and new
ideas. J Membr Sci 181:320
Koros WJ, Ma YH, Shimidzu T (1996) Terminology for
and, at the same time, the surface coverage of the
membranes and membrane processes. Pure Appl Chem inhibiting species is sufciently low to fulll the
68:14791489 hypotheses leading to Henrys law, then IC can be
1038 Inhibition of Hydrogen Permeation Through Pd-Based Membranes by CO
dened for the species i as reported in Eqs. 3 and 4

(Caravella et al. 2010b): Inhibition of Hydrogen Permeation
Through Pd-Based Membranes by CO
PeClean
FluxInhib
i 1 ICi i
Alessio Caravella
d
National Institute of Advanced Industrial Science
Driving Forcei (3) and Technology Research Institute for
Innovation in Sustainable Chemistry, ISC-AIST,
PeInhib Tsukuba, Ibaraki, Japan
FluxInhib
i i
Driving Forcei (4)
d
where the permeability reduction is explicitly rec- Synonyms

ognized as:
CO inhibition to Pd-based membrane
PeInhib
i 1 IC PeInhib
i (5) For the general concept of inhibition, see the entry
Inhibition of Permeation. Concerning the more
For hydrogen permeation through Pd-based mem- specic case of inhibition of hydrogen permeation
branes, Eq. 2 becomes Eq. 6: through Pd-based membranes by CO which is of
crucial importance for PEM-FC applications
PeClean Feed n Perm n requiring a CO content in mixture lower than
FluxClean
H2 A H2
PH2 PH2 10 ppm (Fuel Cell Handbook, 2004) the CO
d
(6) adsorption on Pd-based surface does not involve
thermal dissociation and, thus, is completely
where PFeed Perm
H2 [Pa] and PH2 [Pa] are the hydrogen reversible (Conrad et al. 1974). Therefore, prop-
partial pressures in feed and permeate side and d erly speaking, this kind of adsorption is some-
[m] the membrane thickness. In Eq. 6, the charac- times referred to as physical adsorption or
teristic driving force is expressed by the difference physisorption (Conrad et al. 1974; Davies and
of partial pressure to the power of n, which is an Lambert 1981; Prigge et al. 1982; Ortega
empirical exponent whose value represents the et al. 1982), different from the chemical adsorp-
macroscopic effect of all the mechanisms tion (or chemisorption), which involves, on the
governing the hydrogen permeation process with- contrary, a chemical bond between species and
out inhibiting species (Caravella et al. 2010a; surface. This characteristics cause inhibition to
Flanagan and Wang 2010). decrease with increasing temperature, as adsorp-
tion and desorption are faster at higher
temperatures.
With the hypotheses of validity of Henrys law
References
for CO (i.e., sufciently low CO surface cover-
Caravella A, Scura F, Barbieri G, Drioli E (2010a) Sieverts age), the hydrogen permeation ux through
law empirical exponent for Pd-based membranes: crit- Pd-based membranes can be written as reported
ical analysis in pure H2 permeation. J Phys Chem in Eqs. 1 and 2:
B 114:60336047
Caravella A, Scura F, Barbieri G, Drioli E (2010b) Inhibi-
tion by CO and polarization in Pd-based membranes: a PeClean
H2, n
novel permeation reduction coefcient. J Phys Chem H2 1 ICH2, n
J Inhib
d
B 114:1226412276 Feed n Perm n
Flanagan TB, Wang D (2010) Exponents for the pressure PH2 PH2 (1)
dependence of hydrogen permeation through Pd PeInhib
H2, n n n
and Pd-Ag alloy membranes. J Phys Chem C PFeed
H2 PPerm
H2
114:1448214488 d
Inhibition to Permeation 1039
Clean
H2, n 1 ICH2, n PeH2, n
PeInhib (2) Ortega A, Huffman FM, Bradshaw AM (1982) The adsorp-
tion of CO on Pd(100) studied by IR reection absorp-
tion spectroscopy. Surf Sci 119:7994
where ICH2,n [-] is the inhibition coefcient, Pe is Prigge D, Schlenk W, Bauer E (1982) The adsorption of
membrane permeability, and n is an empirical CO and O2 on ultrathin Pd layers on W(110). Surf Sci
exponent whose value is an indication of the 123:L698L702
interactions among the mechanisms controlling
the hydrogen permeation. The subscript
n indicates that both membrane permeability and
inhibition coefcient depend on the value of n. Inhibition to Permeation
In the particular case in which the original
Sieverts law can be applied, the pressure expo- Alessio Caravella
nent n is equal to 0.5 (indicated by the superscript National Institute of Advanced Industrial Science
Siev), and Eq. 1 becomes (Caravella and Technology Research Institute for
et al. 2010): Innovation in Sustainable Chemistry, ISC-AIST,
Tsukuba, Ibaraki, Japan
PeCleanH2, Siev
J Inhib 1 IC H2, Siev
H2
d
q q
I
In physical chemistry and chemical engineering,
H2
PFeed PPerm
H2 (3)
inhibition mostly refers to a phenomenon for
which one or more species prevents other species
In all the reported expressions, the values of the from reaching active sites of a surface by a compet-
partial pressures must be those adjacent to the itive reversible site occupation. The reversibility of
membrane surface. In the absence of appreciable this phenomenon makes it different from poisoning,
external mass transfer resistance (concentration which affects irreversibly the considered surface.
polarization), these values coincide with the bulk Historically speaking, inhibition has rstly
values. However, the presence of inhibition regarded the activity reduction of catalysts. How-
always includes a certain level of concentration ever, with the development of membrane technol-
polarization, as the inhibiting species do not per- ogy, the concept of catalytic surface has become
meate through membrane, generating in this way wider, this more generally referring also to
an additional resistance not only on membrane surfaces interacting with chemical species in
surface but also in the uid phase of the feed side. processes of adsorption/desorption and solubili-
zation/desolubilization from membrane surface to
membrane bulk and vice versa (i.e., not necessar-
ily involving net chemical reactions).
References If the driving force promoting permeation
through membrane can be represented by the pres-
Caravella A, Scura F, Barbieri G, Drioli E (2010) Inhibition sure difference of a generic species i, the transmem-
by CO and polarization in Pd-based membranes: a brane ow rate Fi (mol s1) and ux Ji
novel permeation reduction coefcient. J Phys Chem
B 114:1226412276
(mol m2 s1) are expressed by Eq. 1 in absence
Conrad H, Ertl G, Koch J, Latta EE (1974) Adsorption of inhibition (indicated by the superscript Clean):
of CO on Pd single crystal surfaces. Surf Sci
43:462480 FClean AClean Clean
Davies PW, Lambert RM (1981) Structural inuences on i Mem J i
the adsorption-desorption properties of the palladium/ PeClean Feed
CO system. Surf Sci 111:L671L674 AClean
Mem
i
Pi PPerm
i (1)
EG&G Technical Services, Inc. (2004) Fuel cell handbook,
d
7th edn. U.S. Department of Energy, Ofce of Fossil
Energy, National Energy Technology Laboratory, where PiFeed (Pa) and PiPerm (Pa) are the partial
Morgantown pressure of hydrogen in feed and permeate side,
1040 Inorganic Acids Recovery by Nanofiltration
d (m) is the membrane thickness, and AMem (m2) is

the membrane area. In general, for membranes as Inorganic Acids Recovery
well as for catalysts, the reduction of active sites by Nanofiltration
directly corresponds to a reduction of the effective
catalytic surface. From this point of view, the B. S. Ooi
intrinsic membrane permeability PeiClean School of Chemical Engineering, Universiti Sains
(mol m1 s1 Pa1) i.e., the permeability eval- Malaysia, Nibong Tebal, Pulau Pinang, Malaysia
uated in clean (inhibition-free) conditions is not
affected by inhibiting species. Nanoltration (NF) is a type of membrane which
However, as it is extremely difcult to know functions in a similar way to reverse osmosis, but
the exact amount of surface reduction due to inhi- is generally used to retain divalent ions and allow
bition (and, thus, not available to permeation), in monovalent ions to pass through. Nanoltration
some cases the inhibition effect can be dealt with has been used to recover the metal or acid value
by considering a reduction of permeability (Eq. 2) from the mining and metal nishing industries. In
rather than a reduction of surface (Eq. 3). recent years, nanoltration has become an emerg-
ing technology for the recovery of strong
FInhib
i AClean Inhib
Mem J i acids inorganic (mineral) acidic solutions such
Feed as hydrochloric acid (HCl), sulfuric acid (H2SO4),
PeInhib
AClean
Mem
i
Pi PPerm
i (2) phosphoric acid (H3PO4), nitric acid (HNO3), and
d
hydrouoric acid (HF). These acids are com-
FInhib
i AInhib Clean
Mem J i
monly used in the winning, electroplating, acidic
pickling, and etching process. The acid strength of
PeClean Feed
AInhib
Mem
i
Pi PPerm
i (3) the solutions can be varied up to 33 wt.% H2SO4,
d 35 wt.% H3PO4, and 10 wt.% HNO3.
Nanoltration is increasingly applied to sepa-
The expressions 2 and 3 can be used under the
rate the metal ions and at the same time recycling
hypothesis of low surface coverage of inhibiting
the acidic solutions back to the plating bath. The
species (i.e., within Henrys law hypotheses)
performance of nanoltration in acid recovery is
(Caravella et al. 2010). For higher values of surface
complicated by their exclusion mechanism of
coverage due to inhibiting species, the driving force
nanoltration which includes adsorption, size
is gradually less suitable to be described by a simple
exclusion, coulombic interactions, and dielectric
function of feed and permeate partial pressure, it
exclusion. Due to the monovalent of proton (1+)
involving, on the contrary, more complex expres-
and its smaller ionic radius (hydrated oxonium
sions of the partial pressure of all the species in
ion) than water, nanoltration has relatively
mixture. The overall form of these expressions can-
lower acid (proton) rejection. However, the trans-
not be easily generalized, as they strongly depend
port of acids across the membrane also depends on
on the functionality of the adsorption/desorption
their counterions. The solution pH will affect the
mechanism on the surface as well as on that of the
membrane surface zeta potential and acid recov-
diffusion mechanism through membrane.
ery performance. For example, if the isoelectric
point of the membrane is 4, then above pH 4, the
membrane will be negatively charged. Under this
References
pH, sulfate ion will be repelled and less proton
Caravella A, Barbieri G, Drioli E (2009) Concentration permeation is expected in order to maintain the
polarization analysis in self-supported Pd-based mem- Donnan Equilibrium in the feed stream. On the
branes. Sep Pur Technol 66:613624 other hand, at low pH, the membrane will have net
Caravella A, Scura F, Barbieri G, Drioli E (2010) Inhibition
positive charge. Permeation of multivalence coins
by CO and Polarization in Pd-Based Membranes: A
Novel Permeation Reduction Coefcient. J Phys Chem (SO42) will be favorable; thus proton permeation
B 114:1226412276 will be enhanced to maintain electroneutrality.
Inorganic Catalytic Membrane 1041
Inorganic Acids Recovery by Nanofiltration, acid-pickling-solutions-using-nanoltration.html.

Table 1 Comparative rejection dataa Accessed 12 Aug 2013
Soldenhoff K, McCulloch J, Manis A, Macintosh P (2005)
Acid RO(%) NF(%)
Nanoltration in metal and acid recovery. In: Schafer
H2SO4 ~98 010 AI, Fane AG, Waite TD (eds) Nanoltration principles
HCl ~90 ~0 and applications, 1st edn. Elsevier Advanced Technol-
a
Performance depends on the feed concentration ogy, Oxford, p 464
Another important concerning factor in acid/ Inorganic Catalytic Membrane

metal separation is the chemical speciation prob-
lem as weak acids form neutral species under low Rune Bredesen
acidic conditions. Separation of neutral acid using Sustainable Energy Technology, SINTEF
NF therefore will solely depend on size exclusion. Materials and Chemistry, Oslo, Norway
Besides, the permeate ux is strongly inuenced
by acid viscosity and osmotic pressure. The per-
meate ux decreased as the acid concentration An inorganic catalytic membrane is an inorganic
increased due to the acid viscosity. membrane that is catalytically active. Inorganic
Membrane stability is always a concerning I
membranes are made from metals, ceramics,
issue for their application in acid recovery. Com- glass, and carbon and may be porous or dense.
mercial polymeric nanoltration membranes are The term inorganic catalytic membrane is typi-
typically stable in the pH range of 212. Mem- cally meant in the context of membrane reactors
branes of interest include CK (pH 28)/DK where yield is enhanced by combining a mem-
(pH 210) (Osmonics), SeIRO MPF-34 (Koch, brane separation and a catalyzed chemical reac-
pH range 014), NF45 (Dow, pH 310), PVD-1 tion. Inorganic catalytic membranes can either be
(Hydranautics), and NF-PES-10 (Nadir) composed of an inherently catalytically active
(Soldenhoff et al. 2005). The report from Cleaner material or an inert membrane structure to which
Production Germany shows that nanoltration is a catalytically active phase is added on the outer
suitable to be used for the treatment of sulfuric surface or inside pores (Specchia et al. 2006), see
pickling solutions at plant trial scale. Metal con- Fig. 1. In the latter case, integration of catalyst
centrations of 80 g/l were achieved in the concen- material and inorganic membrane structure is car-
trate (99 % rejection), and sulfuric acid was able to ried out by common deposition and impregnation
pass through the membrane at a ux of 30 l/m2.h methods such as wet (electro)chemical deposi-
to give ltrate (permeate) amounting to 250 g/l. In tion, chemical vapor deposition, or physical
this process, up to 99 % of the acid residue and the vapor deposition. Since the membrane is catalytic,
water can be returned to the pickling process reaction and separation occur in parallel a feature
(Cleaner Production Germany 2004). The general which distinguishes it from the sequential pro-
rejection data of acid using RO and NF mem- cesses of inert membrane reactors where the
branes are generalized in Table 1. It shows that inert membrane and catalyst material are sepa-
NF is relatively poor in retaining acid compared to rated. The function of the membrane may serve
the RO membrane which makes it very interesting as extractor, distributor, or contactor depending
in metal/acid separation. on application (Dalmon 1997) (see Fig. 2) and
the reaction mechanism can be (electro)catalytic,
photo-catalytic, or bio-catalytic. The many differ-
References ent concepts and envisaged applications are
described comprehensively in various text books
Cleaner Production Germany (2004) Regeneration of
sulphuric acid pickling solutions using nanoltration.
and scientic papers (Gryaznov et al. 2006;
http://www.cleaner-production.de/en/projects-publica Sanchez Marcano and Tsotsis 2002; Caro 2010;
tions/projects/metal/regeneration-of-sulphuric- Fontananova and Drioli 2010).
1042 Inorganic Catalytic Membrane
a b
c
d e
Inorganic Catalytic Membrane, Fig. 1 Inherently catalytic (a) porous, (b) dense membrane. Inert membrane with
catalyst deposited (c) inside pores, on outer surface of (d) porous, and (e) dense membrane
Inorganic Catalytic Membrane, Fig. 2 Inorganic catalytic membrane in (a) extractor operation, (b) distributor
operation, (c) ow-through contactor operation, (d) liquidgas contactor operation
Inorganic catalytic membranes can operate at adsorption, poisoning, clogging, and fouling are
higher temperatures than their polymeric counter- all critical and membrane lifetime may suffer
part, which opens a broader window of operation when trying to achieve the optimal trade-off
as high temperature is also required for many between membrane and catalyst properties. Com-
important reactions which are limited by low mon examples of applications include hydrocar-
yield in traditional reactors. The assembling of bon dehydrogenation, hydrocarbon reforming,
inorganic catalytic membranes in membrane mod- and hydrogenation of alkenes for which hydrogen
ules may, however, give too low a catalytic sur- selective inorganic catalytic membranes can be
face area and limited conversion and therefore used (Sanchez Marcano and Tsotsis 2002). Com-
render operation less cost efcient compared to bination of endothermic (e.g., water splitting to
more conventional reactors. To mitigate this produce hydrogen) and exothermic (e.g., partial
problem, additional catalyst can be added to the oxidation of methane) reactions for integrated
reactor volume. In order to increase efciency, mass and heat transport has been demonstrated
some current development activities are aimed by using catalytically active oxygen permeable
at manufacturing inorganic thin capillary, membranes (Caro 2010). By feeding oxygen in a
multichannel, and hollow ber membranes with controlled manner, partial oxidation of hydrocar-
high surface area/volume ratio. bons to form synthesis gas, oxidative coupling of
The dual operational nature of inorganic cata- methane, and oxidative dehydrogenation reac-
lytic membranes is particularly challenging tions have been demonstrated using dense and
since both membrane and catalytic properties porous inorganic catalytic membranes (Sanchez
must work satisfactory. Commonly encountered Marcano and Tsotsis 2002; Coronas and
issues related to degradation processes such as Santamaria 1999; Sammells et al. 2006). When
Inorganic Composite Membrane Preparation 1043
porous inorganic catalytic membranes are used as

ow-through contactors, a forced close spatial Inorganic Composite Membrane
contact between reactants and catalyst can give Preparation
improved conversion efciency compared to reac-
tions in catalyst powder beds (Westermann and Lidietta Giorno1 and Heiner Strathmann2
1
Melin 2009). Additionally, higher selectivity may Institute on Membrane Technology, National
be achieved due to the short and well-dened Research Council of Italy, ITM-CNR, Rende,
contact time between reactants and membrane Italy
2
catalyst. Yet in another contactor mode, the Universitt Stuttgart, Institute of Chemical
porous inorganic catalytic membrane provides a Process Engineering Stuttgart, Baden-
dened catalytic region for reaction and controls Wurttemberg, Germany
transport of reactants from both sides of the mem-
brane. For example, in liquidgas catalytic
contactors, both hydrogenation and oxidative Ceramic membranes to be used in micro-, ultra-,
reactions have been demonstrated. and nanoltration as well as in pervaporation and
gas separation are also prepared as multiple-layer
composite structures (Burggraaf et al. 1996).
I
References A typical three-layer structure of an ultra- or
nanoltration membrane is shown schematically
Caro J (2010) Basic aspects of membrane reactors. In: in Fig. 1a and as a scanning electron micrograph
Drioli E, Giorno L (eds) Comprehensive membrane
in Fig. 1b indicating the materials and the pore
science and engineering, vol 3. Elsevier, Amsterdam,
pp 124 size of the different layers. The rst support layer
Coronas J, Santamaria J (1999) Catalytic reactors based on is generally prepared from a suspension of Al203
porous ceramic membranes. Catal Today 51:377389 powder using an organic polymer such as cellu-
Dalmon JA (1997) Catalytic membrane reactors. In: Ertl G,
Knzinger H, Weitkamp J (eds) Handbook of hetero-
lose or polyvinyl alcohol as a binder and to
geneous catalysis, vol 3. Wiley-VCH, Weinheim, increase the viscosity of the suspension. The sus-
pp 13871398 pension can be processed by slip casting or extru-
Fontananova E, Drioli E (2010) Catalytic membranes and sion into at sheets, hollow bers, or tubes
membrane reactors. In: Drioli E, Giorno L (eds) Com-
forming the so-called green structure which is
prehensive membrane science and engineering,
vol 3. Elsevier, Amsterdam, pp 109133 dried and converted into the nal membrane by
Gryaznov VM, Ermilova MM, Orekhova NV, sintering at ca. 1500 C resulting in a porous
Teresschenko GF (2006) Reactors with metal and structure of a-Al2O3 with an average pore diame-
metal-containing membranes. In: Cybulski A, Moulijn
ter of 5 or more mm and an overall porosity of
JA (eds) Structured catalysts and reactors. Chemical
industries, 2nd edn. Taylor and Francis Group, Boca about 35 %. The process is indicated in Fig. 2.
Raton, pp 579614 A second layer prepared by a so-called solgel
Sammells AF, White JH, Makay R (2006) Membranes for process is deposited on the support layer by dip
promoting partial oxidation chemistries. In: Sammells
coating or ltration. The sol used in this process
AF, Mundschau MV (eds) Nonporous inorganic mem-
branes. Wiley-VCH, Weinheim, pp 185214 consists of colloidal particles as indicated in the
Sanchez Marcano JG, Tsotsis TT (2002) Catalytic mem- schematic drawing of Fig. 2. The material used for
branes and membrane reactors. Wiley-VCH, Weinheim the preparation of the second layer is usually
Specchia S, Fino D, Saracco G, Specchia V (2006) Reac-
g-Al2O3 or TiO2. The pore size of this layer is in
tors with metal and metal-containing membranes. In:
Cybulski A, Moulijn JA (eds) Structured catalysts and the range of 510 nm. Thus, the two-layer mem-
reactors. Chemical industries, 2nd edn. Taylor and brane can be used for micro- and ultraltration
Francis Group, Boca Raton, pp 615661 applications. For the separation of lower molecu-
Westermann T, Melin T (2009) Flow-through catalytic
membrane reactors-principles and applications. Chem
lar weight components in nanoltration and
Eng Process 48:1728 pervaporation a third layer is deposited. This
1044 Inorganic Composite Membrane Preparation
b
a polymeric sol (Ti-oxide,
pore size 1-2 nm)
colloidal sol -Al2O3TiO2

pore size 5-10 nm
partical suspension -Al2O3

pore size 1-5 mm
Inorganic Composite Membrane Preparation, Fig. 1 (a) Schematic drawing illustrating the structure of a three-layer
inorganic composite membrane, (b) Scanning electron micrograph of a three-layer ceramic membrane
Inorganic Composite
Membrane Preparation,
ceramic powder
water
Fig. 2 Porous inorganic
polymer binder
membrane preparation
scheme based on slip- surfactant
suspension
casting of an ceramic
powder and sintering at
15001800 C
membrane casting
hybrid organic inorganic

drying
"green" membrane
inorganic porous
sintering membrane
layer is also prepared by the solgel process. dried, and sintered again. Sometimes, more than
However, the sol contains an inorganic polymer one coating is applied to obtain the desired pore
as indicated in Fig. 2 which forms a polymer gel size at the surface of the membrane. The process
consisting of a three-dimensional network. As preparing a multilayer membrane by suspensions
material for the inorganic structure TiO2, ZrO2, of different particle sizes is referred to as slip
or SiO2 is often used (Liebau 1988). coating. With the slip-coating process membranes
A suspension is a dispersion of solid particles with pore sizes down to 20 nm to be used in
in a liquid; a sol is the suspension of colloidal micro- ultraltration can be made. To obtain
particles or polymers. The main difference membranes with smaller pores a solgel process
between suspension and sol is the size of the is applied and a colloidal or polymeric gel is used
dispersed particles. Suspensions are prepared which is prepared by controlled hydrolysis of a
from a ne powder of aluminum oxide with an metal alkoxide to a hydroxide. The process can
average particle diameter of 5 10 mm. The dis- follow two different process paths. One path is
persion is mixed with an organic polymer such as referred to as the colloidal and the other the poly-
polyvinyl alcohol or a cellulose derivative and meric process path. Both pathways are illustrated
poured in a mold or extruded in the desired in Fig. 3.
shape as a at sheet, tube, or capillary; dried; In the colloidal process path a metal oxide such
and sintered at 1500 1800 C. The process as aluminum oxide is dissolved in alcohol and
scheme is shown in Fig. 2. then hydrolyzed and precipitated by addition of
One surface of the support structure is then excess water at elevated temperature to form a
coated again with a suspension of ner particles, stable colloidal solution. The solution is cooled
Inorganic Fouling 1045
Inorganic Composite colloidal route polymeric route

Membrane Preparation, alkoxide in alcohol
water organic
illustrating the colloidal and colloidal sol
polymeric route of the polymeric sol
solgel process used in the
preparation of composite
inorganic membranes
membrane coating
colloidal gel polymeric sol
drying
hybrid organic inorganic"green" membrane
sintering
I
inorganic membrane
to form a gel, which is then coated on an appro- Hydrolysis of the metal alkoxide
priate support and sintered at ca. 800 C. The Polymerization and coating
average pore diameter obtained in the colloidal Sintering and cross-linking
solgel process is in the range of 25 nm.
The overall process consists of three steps:
References
Metal alkoxideformation: Al2 O3
alcohol ! AlOR3 Burggraaf AJ, Cot L (1996) Fundamentals of inorganic
membrane science and technology, vol 4, Membrane
Precipitation: AlOR3 H2 O ! AlOH3 science and technology series. Elsevier Science B. V,
! g Al2 O3 H2 O Bhmite Amsterdam
Coating and sintering: g Al2 O3 H2 O Liebau F (1988) Structural similarities and dissimilarities
! g Al2 O3 between SiO2 and H2O. In: Devine RAB (ed) In the
physics and technology of amorphous SiO2. Plenum
Press, New York
In the polymeric solgel process the metal
alkoxide is only partially hydrolyzed in an alcohol
solution by a controlled addition of a minimum of
water. The hydroxyl groups of the alkoxide react
Inorganic Fillers
with each other and form an inorganic/
organic polymer, which forms a clear gel that is
Filler in Membranes
coated on the support structure. It is dried and
sintered at 500 800 C and further cross-linked
to form a porous structure with a pore diameter of
less than 1 nm. For the polymer solgel process
often a titanium alkoxide is used which is Inorganic Fouling
converted into a cross-linked inorganic polymer
in three steps: Inorganic Scaling
1046 Inorganic Scaling
crystal cake causes cake-enhanced concentration

Inorganic Scaling polarization (CECP) (see Irreversible Fouling
Resistance), and the increased CP of the scale
Tony Fane formers leads to precipitation on the membrane
School of Chemical Engineering, The University and within the cake.
of New South Wales, Sydney, NSW, Australia There are several strategies to control scaling
Singapore Membrane Technology Centre, in membrane systems. Bulk crystallization can be
Nanyang Technological University, Singapore, mitigated by controlling the supersaturation level
Singapore of the salts by limiting the maximum recovery in
the RO system. The effect of CP on surface con-
centration can be manipulated by the ratio of ux
Synonyms to crossow velocity, where lower CP is favored
by modest ux and raised crossow velocity.
Inorganic fouling Most scale formers are more soluble at lower
pH, so the injection of acid into the feed stream
Inorganic scaling occurs in reverse osmosis can provide some control. The exception to this is
(RO) applications and is caused by the retention silica scaling which is more serious at lower pH and
of sparingly soluble mineral salts such as calcium requires an increase in pH. If silica and calcium
carbonate, calcium sulfate, calcium phosphate, scalants are both present, it may be necessary to
barium sulfate, magnesium salts, silica, etc. Due soften the feed by ion exchange to remove the
to water removal and concentration polarization calcium and then raise the pH. Scale formation
(CP), the concentrations build up and can exceed can also be controlled by the use of antiscalant
the saturation level and cause scaling of the mem- chemicals. These are proprietary materials, but
brane surface. Membrane scaling has two path- include various molecular weight polycarboxylates
ways, namely, surface crystallization and bulk and polyacrylates, usually dosed at a few ppm
crystallization. For surface crystallization, CP level. Care needs to be taken with antiscalant selec-
causes the sparingly soluble salt concentration at tion as some tend to promote biofouling.
the membrane surface to exceed the solubility
limit. Consequently, the surface is gradually cov-
ered by the lateral growth of a crystal deposit that References
reduces the effective area for permeation. For
xed imposed ux, the local ux increases in Fane AG, Chong TH, Le-Clech P (2009) Fouling in mem-
brane processes, Chapter 6. In: Drioli E, Giorno L (eds)
order to compensate for loss of area, and this
Membrane operations, innovative separations and
increases the local CP level which exacerbates transformations. Wiley VCH. ISBN 978-3-527-
the scale formation. In the other pathways, crys- 32038-7
tals form in the bulk phase when concentrations
exceed saturation level due to high recovery
(dened as fraction of feedwater removed as per-
meate). Crystals in suspension are then deposited Integrated Membrane Processes
on the membrane and eventually form a porous for Biofuel Production
cake layer, which provides a hydraulic resistance.
There are many membrane elements connected in Lan Ying Jiang
series in a RO plant, and concentration builds up School of Metallurgy and Environment, Central
from the feed to the tail end. Scale formation is South University, Changsha, China
therefore more prevalent in the tail end elements
of the plant. It is likely that both scaling mecha-
nisms, surface and bulk, could occur simulta- Biofuels are important solutions to worldwide
neously in a RO system. This is because a energy crisis and greenhouse effect. Their
Integrated Membrane Processes for Biofuel Production 1047
production through chemical, physiochemical, or value-added chemicals have already been reviewed
biochemical conversions is one of the many pro- (Schugerl 2000). Other metabolites such as
cedures involved in biorenery. Due to the com- carbon dioxide (CO2) and ammonia (NH3) can
plexity of the conversion mechanisms and be removed easily by aeration stripping.
feedstocks, what are delivered in the fuel genera- Pervaporation or membrane distillation has the
tion procedures are not highly puried fuels but a dual function of rejecting the cells and preferen-
mixture of a series of organics containing the fuels tially removing the VOCs, simultaneously. But
in some concentration range. In fermentation pro- more often, a microltration membrane is used
cess to produce liquid bioalcohols, the situation is to hold back the microbes, so that the hydrophobic
further complicated by the existence of bacterial membrane, the more costly unit, will not be
cells inside the yeast cultures. In this regard, inte- fouled. With the above separation procedures,
gration of separation technologies to recover the the fermentation process can be carried out con-
biofuels, remove other components, or enhance tinuously and obtain improved productivity and
the concentration of biofuels is indispensible to conversion rate. The recovered VOCs (ethanol,
biofuel production. The separation technologies acetone, butanol, 2-propanol) are directly reused
used to rene and upgrade biofuels include distil- within other processes or further separated. In
lation, membrane separation, gas stripping, principle, gas stripping or adsorption is more
I
liquidliquid extraction, etc. (Vane 2005). The accurate in discriminating these VOCs (Schugerl
reasons to consider membrane separation include 2000). Upon obtaining the ethanol-enriched aque-
moderate temperature, space efciency, and easy ous solution, it is dehydrated by hybrid system
operation. incorporating distillation and pervaporation.
Several types of integrated processes using Intensive dehydration by adsorption might be
membrane as key units are employed in biofuel performed on condition that the purity is required
production. In bioethanol production, an inte- to be very high. Similar design of integrated pro-
grated membrane process as illustrated in Fig. 1 cess can be used in biobutanol production.
is employed (Huang et al. 2008). In achieving The most common way to produce biodiesel is
recovery of volatile organic compounds (VOCs) transesterication that transforms triglyceride
containing the alcohols from the fermentation extracted from vegetable oil, animal fats, grease,
process, hydrophobic pervaporation and mem- or algae into fatty acid alkyl esters (FAME) in the
brane distillation are used. There also exist in the presence of alcohol (e.g., methanol) and a catalyst
fermentation broth some other chemicals with (e.g., acid, base) (Shuit and Subhash 2012). Two
higher molecular weight such as phenolics, car- major membrane separation procedures are inte-
boxylic acids, and amino acids. Various tech- grated in the above process. For the real-time
niques for recovery or extraction of these separation of products from the feed mixture,
Integrated Membrane Hydrophilic

CO2 Retentate
Processes for Biofuel
Bleed (dehydrated PV
Production,
ethanol)
Fig. 1 Integrated
membrane process for
Retentate
bioethanol production
(Adapted from Huang
et al. 2008) Feed H2 O
Ethanol
enriched enriched
permeate permeate
Hydrophobic
Bioreactor Microfiltration (MF)
(Fermenter) PV
1048 Integrated Membrane Processes for Biofuel Production
porous membrane is used to retain the oil droplet on the purity and quality of biodiesel as fuel for
larger than the pore size while allowing the alco- use in compressionignition engines.
hol, glycerol, and biodiesel to pass through. High-temperature gasication and anaerobic
Thereafter, glycerol is removed to obtain puried digestion are major approaches for producing bio-
biodiesel also by using porous membrane (e.g., gas containing H2, carbon monoxide (CO), carbon
MF), wherein the reversed micelles binding glyc- dioxide (CO2), methane (CH4), etc. The inte-
erol are rejected (Wang et al. 2009). If dense grated process for biogas upgrading includes two
membrane (e.g., pervaporation) selectively trans- parts (Shuit and Subhash 2012). The pretreatment
fers glycerol that is used, one-step separation can shown in the top part removes impurities includ-
replace the above two-step separation. Following ing H2S, H2O, other VOCs, and siloxanes mainly
removal of glycerol, biodiesel is further puried to using adsorption and temperature swing adsorp-
remove impurities such as tri-, di-, and monoglyc- tion. The treated feed is delivered to the mem-
erides, catalyst, soap, and traces of alcohol. The brane system that consists of three blocks as
removal of these impurities is essential as the shown in Fig. 2. In block 1, the feed is separated
presence of these impurities poses great impact into a CO2-enriched permeate and a CH4-enriched
Integrated Membrane
Processes for Biofuel
Production,
Valve
Fig. 2 Integrated process
for upgrading biogas from
anaerobic digestion Valve
Water Sep.
(Adapted from Shao Desul.
et al. 2002) TSA
Valve
Valve
Gas Blower
Feed (CO2 40%) Desul.

TSA
CO2 (99%)
Cooling
Compressor
A3
Compressor
A1
CH4 (97%)
A2
Water Sep.
Intrinsic Activation Energy 1049
retentate. They are further enriched to 99.0 vol.% to its original volume with a diluent. This
in block 2 and 97.0 vol.% in block 3, respectively. concentration-dilution sequence may be
repeated several times to obtain the desired degree
of separation of macrosolutes from microsolutes.
As a nal step, an additional post-concentration
References may be applied to achieve the required nal
volume. A mathematical analysis of intermittent
Huang HJ, Ramaswamy S, Tschirner UW, Ramarao BV
(2008) A review of separation technologies in feed dialtration can be found in (Wang et al.
current and future bioreneries. Sep Purif Technol 2007).
62:121 Note that the processing scheme of intermittent
Schugerl K (2000) Research review paper-integrated feed dialtration is basically a slight modication
processing of biotechnology products. Biotechnol
Adv 18:581599 of the method employed by sequential dilution
Shao P, Dal-Cin M, Kumar A, Li H, Singh DP dialtration (also called discontinuous
(2002) Design and economics of a hybrid dialtration by sequential dilution).
membranetemperature swing adsorption process for
upgrading biogas. J Membr Sci 413414:1728
Shuit SH, Subhash B (2012) Membrane technology as a
promising alternative in biodiesel production: a review.
Biotechnol Adv 30(6):13641380
References I
Vane LM (2005) A review of pervaporation for product
recovery from biomass fermentation processes. J Chem Wang L, Yang G, Xing W, Xu N (2007) Mathematic model
Technol Biotechnol 80:603629 of the yield for dialtration processes. Sep Purif
Wang Y, Wang X, Liu Y, Ou S, Tan Y, Tang S (2009) Technol 59:206213. doi:10.1016/j.
Rening of biodiesel by ceramic membrane separation. seppur.2007.06.007
Fuel Process Technol 90:422427
Intrinsic Activation Energy

Integration of Membrane Separation
with Biotechnology Renzo Di Felice
Membrane Biotechnology
Transport phenomena and chemical reaction rates

are inuenced by temperature. This effect is
Intermittent Feed Diafiltration reected on the numerical value of the coefcient
rate, which therefore depends on the variable tem-
Zoltn Kovcs perature. Both for mass transfer coefcient and
Department of Food Engineering, Institue of chemical reaction coefcient, various theoretical
Bioengineering and Process Engineering, Szent approaches have been suggested in order to pre-
Istvan University, Budapest, Hungary dict and quantify their dependency on the temper-
ature. However, these approaches have had
limited success, and the effect of the temperature
Intermittent feed dialtration is characterized is generally obtained through experimental evi-
as a batch dialtration process in which the dence. A generally accepted relationship linking
processed liquor is rst concentrated to a kinetic coefcient and temperature is the Arrhe-
predetermined volume and then diluted back nius law, which is written as
1050 Ion Beam Irradiation
E the membrane, they lose energy to their surround-

k k0 exp ings (membrane structure) by two main processes:
RT
interacting with target nuclei (nuclear stopping)
where k is the generic kinetic coefcient, E is the and interacting with target electrons (electronic
activation energy, R is the universal gas constant, stopping; Lee 1999).
and T is the absolute pressure. Nuclear stopping energy losses arise from col-
Arrhenius law is basically an empirical rela- lisions between energetic particles and target
tionship which however has been found to repro- nuclei (Davenas et al. 1989; Lee 1999). Atomic
duce with the satisfactory accuracy of displacement occurs when the colliding ion
experimental data and the effect of temperature. imparts energy greater than certain displacement
A small value of the activation energy threshold energy. If the energy is not great enough
E corresponds to a weak inuence of the temper- for displacement, the energy dissipates as atomic
ature on the kinetic constant, whereas a large vibrations known as phonons. The threshold
value of the activation energy E is representative energy is the energy that a recoil requires to over-
of kinetic constant strongly inuenced by temper- come binding forces and to move more than one
ature. Just as a reference, for gas phase chemical atomic spacing away from its original site. The
reaction, the activation energy is normally high, in interaction of an ion with a target nucleus is
the range of 50250 kJ/mol, while that for diffu- treated as the scattering of two screened particles
sion coefcient is less than 10 kJ/mol). since the nuclear collision occurs between two
The classic way to estimate the activation coef- atoms with electrons around protons and neu-
cient from experimental data is to utilize a plot trons. Nuclear stopping varies with ion velocity
having 1/T on the x-axis and ln(k) on the y-axis. as well as the charges of two colliding atoms, as it
The slope of the best tting curve passing through is derived with consideration of the momentum
the experimental points will yield the value of the transfer from ion to target atom and the
ratio (E/R) and consequently that of E. interatomic potential between two atoms.
Electronic stopping energy losses arise from
electromagnetic interaction between the posi-
tively charged ions and the target electrons (Lee
1999). There are two mechanisms: one is called
Ion Beam Irradiation glancing collisions (inelastic scattering, distant
resonant collisions with small momentum trans-
Isabel C. Escobar fer), which are quite frequent, but each collision
University of Kentucky, Lexington, KY, USA involves a small energy loss (<100 eV), and the
other is called knock-on collisions (elastic scatter-
ing, close collisions with large momentum trans-
Ion beam irradiation has long been recognized as fer), which are very infrequent, but each collision
an effective method for the synthesis and modi- imparts a large energy to a target electron
cation of diverse materials, including polymers (>100 eV). Both collisions transfer energy in
(Davenas et al. 1989; Xu et al. 1995a, b). Ion two ways: electronic excitation and ionization.
beam irradiation is the bombardment of a sub- Excitation is the process in which an electron
stance with energetic ions. When the ions pene- jumps to a higher energy level, while in ionization
trate through the surface of a membrane, they may an orbital electron is ejected from the atom.
eliminate the tall peaks and deep valleys, resulting Various gaseous molecular species are released
in an overall reduction in surface roughness during irradiation. The most prominent emission
(Xu and Coleman 1997). As the ions penetrate is hydrogen, followed by less abundant heavier
Ion Beam Irradiation 1051
Ion Beam Irradiation, H H

Fig. 1 Typical
C* C Excitation Phonon
consequences induced by
ion irradiation, which H H
include electronic
excitation, phonons, H H
ionization, ion pair +
C C + e Ionization
formation, radical
formation, and chain H H
scission
H H H _
E
C C C C + H+ Ion Pair
H H H H
H

C C + H Radical Pair
H H
I
H H
C + C Scission
H H
molecular species, which are scission products four events can, in turn, lead to membrane micro-
from the pendant side groups and chain-end seg- structure alterations.
ments, and their reaction products (Lee 1999).
Figure 1 illustrates various functional chemical
entities created by irradiation. Cross-linking References
occurs when two free dangling ion or radical
pairs on neighboring chains combine, whereas Davenas J, Xu XL, Boiteux G, Sage D (1989) Relation between
structure and electronic properties of ion irradiation poly-
double or triple bonds are formed if two neigh-
mers. Nucl Instrum Methods Phys Res 89:754763
boring radicals in the same chain combine. It has Lee EH (1999) Ion-beam modication of polymeric mate-
been well established that mechanical, physical, rials: fundamental principles and applications. Nucl
and chemical property changes in polymers are Instrum Methods Phys Res B 151:2941
Lee EH, Rao GR, Mansur LK (1996) Trends Polym Sci
determined by the magnitude of cross-linking and
4:229
scission and that cross-linking enhances mechan- Xu XL, Coleman MR (1997) Atomic force microscopy
ical stability while scission degrades mechanical images of ion-implanted 6FDA-pMDA polyimide
strength (Lee et al. 1996). The intensive energy lms. J Appl Polym Sci 66:459469
Xu XL, Dolveck JY, Boiteux G, Escoubes M,
deposition in polymers can lead to the following:
Monchanin M, Dupin JP, Davenas J (1995a) A new
(1) formation of volatile molecules and free radi- approach to microporous materials application of ion
cals which leave defects in the polymer matrix, beam technology to polyimide membranes. Mater Res
(2) creation of additional cross-linking between Soc Symp Proc 354:351356
Xu XL, Dolveck JY, Boiteux G, Escoubes M,
polymer chains, (3) formation of new chemical
Monchanin M, Dupin JP, Davenas J (1995b) Ion
bonds, and (4) chemical reactions with chemical beam irradiation effect on gas permeation properties
atmosphere such as oxidation (Lee 1999). These of polyimide lms. J Appl Polym Sci 55:99107
1052 Ion-Exchange Membrane Characterization
spot on top of an anion-exchange membrane and

Ion-Exchange Membrane a deep blue spot on top of a cation-exchange mem-
Characterization brane, respectively. For a quantitative characteriza-
tion, more complex procedures such as the
Lidietta Giorno1, Enrico Drioli1,2,3,4 and determination of the ion-exchange capacity or the
Heiner Strathmann5 electrical resistance and transport properties for
1
Institute on Membrane Technology, National different ions of the membrane are applied. Addi-
Research Council of Italy, ITM-CNR, Rende, tional information can be obtained by impedance
Italy spectroscopy and chronopotentiometric measure-
2
Department of Environmental and Chemical ments or from the determination of osmotic and
Engineering, University of Calabria, Rende, electroosmotic water transfer. Bipolar membranes
Italy are further characterized in terms of the water dis-
3
WCU Energy Engineering Department, sociation capability and salt leakage under operat-
Hanyang University, Seoul, S. Korea ing conditions (Strathmann et al. 1993).
4
Center of Excellence in Desalination
Technology, King Abdulaziz University, Jeddah,
Saudi Arabia Hydraulic Permeability of Ion-Exchange
5
Universitt Stuttgart, Institute of Chemical Membranes
Process Engineering Stuttgart, Baden-
W urttemberg, Germany Hydraulic permeability measurements provide
information on the diffusive or convective trans-
port of components through a membrane under a
Characterization of ion-exchange membranes is hydrostatic pressure driving force. In most practi-
closely related to their preparation. The most cal applications of electrodialysis, the hydraulic
interesting properties of ion-exchange membranes permeability due to diffusion plays a minor role
are (Spiegler 1958): for the overall performance of the membrane in a
given process. However, for the determination of
he electrical resistance in different electrolyte pinholes in ion-exchange membranes will affect
solutions the performance of a membrane severely in appli-
The type and density of xed charges and their cations with a hydrostatic pressure differences
distribution in the membrane matrix across the membrane. Pinholes can be detected
The permselectivity of the membrane for dif- quickly by placing a wet membrane sheet on a
ferent ions sheet of white absorbent paper. Then, a 0.2 %
he transport rate of neutral components, espe- solution of methylene blue for an anion-
cially water exchange membrane or a 0.2 % solution of eryth-
The mechanical and chemical stability and the rosin B for a cation-exchange membrane is spread
swelling in different electrolyte solutions over the entire surface of the membrane. If no
spots of the dye can be observed on the paper,
A series of procedures and techniques are the membrane is free of large pinholes. The
applied to determine the different membrane prop- hydraulic permeability of the membrane is deter-
erties. Microscopic examination yields information mined at room temperature using deionized water
on heterogeneity of the membrane structure and in and a hydrostatic pressure driving force in a con-
case of reinforced membranes the type of rein- ventional ltration cell as used in reverse osmosis
forcement used. The electrical charge of an or ultraltration experiments. The permeability
ion-exchange membrane can be determined quali- can then be calculated from the volumetric
tatively by using indicator solutions. A drop of a ow rate.
0.05 % solution of methylene blue and methyl The economics of processes using
orange on a membrane sample stains a yellow ion-exchange membranes in different applications
Ion-Exchange Membrane Characterization 1053
is determined to a large extent by the chemical water. The ion-exchange capacity of the samples
stability of the membranes under process condi- is determined by back titration with 1N NaOH or
tions. Generally, the changes in membrane prop- 1N HCl, respectively. Weak base anion-exchange
erties are time and temperature dependent. membranes are characterized by equilibration in
Reliable information about the long-term stability 1N NaCl and titration with standardized 0.1N
of the membranes is of importance for a process AgNO3 solution. The samples are then dried, and
design and cost calculation. For a fast evaluation, the ion-exchange capacity is calculated based on
membrane samples are exposed to oxidizing the dry membrane. The accuracy of the measure-
agents or acids and bases and other chemical ment depends on the complete exchange of ions in
components in much higher concentration and at the membrane which can take some time. The
higher temperature than expected in practical use. concentration of the xed ion in commercial
Structural mechanical changes are determined by membranes is in the range between 1 and
visual or microscopic investigation and standard 3 milliequivalent per gram. The experimentally
mechanical and electrochemical tests comparing measured ion-exchange capacity of a membrane
exposed with unexposed membrane samples. is an average value assuming that the ions are
Since the useful life of membranes under process homogeneously distributed over the entire polymer
conditions is generally in the range of a couple of matrix. In many heterogeneous membranes, such
I
years, membrane stability tests are very time con- as most of the peruorinated cation-exchange
suming. However, in many cases long-term pilot membranes, the xed ions are clustered, and the
plant tests are mandatory to get a reliable estima- local xed ion concentration may be signicantly
tion of the life of the membranes in practical higher than the average value.
applications.
The performance of ion-exchange membranes
in the various processes is determined to a very Determination of the Electrical
large extent by their electrochemical properties. Resistance of Ion-Exchange Membranes
Therefore, a major task in characterizing ion-
exchange membranes is the determination of The electrical resistance of ion-exchange mem-
their electrochemical properties such as the xed branes is an important parameter determining to
charge density, the electrical resistance, the ion the energy consumption in electrodialysis
permselectivity, and the transport of nonionic (Strathmann et al. 2006). The specic membrane
components such as water or of other neutral resistance is usually reported as O cm or O
molecules under operating conditions. m. From the engineering point of view, the mem-
brane area resistance in units of O cm2 or O m2 is
more useful and generally given in the literature
The Ion-Exchange Capacity of a describing commercial products. The electrical
Membrane resistance of a membrane is determined by the
ion-exchange capacity and the mobility of the
The ion-exchange capacity of membranes is a cru- ion within the membrane matrix. The ion mobility
cial parameter which affects almost all other mem- in the membrane depends strongly on the nature
brane properties. It is usually expressed in of the mobile and xed ion species. It is further-
milliequivalent per gram of dry membrane. Exper- more affected by the temperature.
imentally, the ion-exchange capacity of strong The area resistance of ion-exchange mem-
acidic or strong basic ion-exchange membranes is branes can be determined by direct current
readily determined by titration with NaOH or HCl, (DC) or alternating current (AC) measurements.
respectively. For these tests, cation- and anion- In DC measurements the membrane is installed in
exchange membranes are equilibrated in 1N HCl a cell which consists of two chambers containing
or 1N NaOH, respectively, and then rinsed free the test solution separated by the test membrane as
from chloride or sodium ions with deionized indicated in Fig. 1.
volt meter
calomel electrode
1 n KCl-solution Haber-Luggin capillary
electrode electrode
solution A A C solution
gas gas
(O2) (H2)
anode cathode
+ porous plug -
- - - +
test solution test solution
Ion-Exchange Membrane Characterization, Fig. 1 Schematic drawing of a test cell for determining the resistance of
ion-exchange membranes with direct current
The two electrodes used to provide the electri- The area resistance is given by:
cal potential driving force to obtain a certain cur-
rent are separated by membranes from the test U
r ms RAm (1)
solution to avoid gas bubbles to disturb the mea- i
surement. As test solution a 1 or 0.5 mol/L
Na2SO4 solution is used. The same solution is The membrane resistance is:
also used to rinse the electrode chambers and
remove the gases produced at the electrodes. The r m r ms r s (2)
actual potential drop in DC measurements across
the membrane is determined with calomel elec- Here R is the resistance, Am is the area of the
trodes attached to Haber-Luggin capillaries membrane, U is the voltage drop measured
placed with their tips close to the membrane sur- between the Haber-Luggin capillaries and I is
face. The potential drop between the Haber- the current density, rm+s and rs are the area resis-
Luggin capillaries is measured with and without tances of the cell with and without the membrane
the membrane in the test cell as a function of the between the Haber-Luggin capillaries, and rm is
current density passing through the cell. The resis- the area resistance of the membrane.
tance is given by the slope of the current versus The determination of the membrane resistance
the voltage drop curve. To obtain the membrane by direct current measurements is quite accurate
resistance, the resistance of the cell without the as long as the position of the Haber-Luggin capil-
membrane is subtracted from the resistance of the laries is as close as possible to the membrane and
cell with the membrane. kept identical in the test with and without the
Ion-Exchange Membrane Characterization 1055
Ion-Exchange Membrane A U
Characterization,
of a test cell used for
membrane resistance
measurements using
alternating current
membrane
dm
I
electrolyte solution
membrane. To avoid boundary layer effects, the Rm

bulk solution must be well mixed. But even if Rs
the boundary layer effects are well controlled,
the direct current measurements are always
Cm
connected with a transport of ions resulting in a
concentration change in the electrolyte solution at Ion-Exchange Membrane Characterization,
the membrane surface which affects the accuracy Fig. 3 Equivalent resistance-capacitance circuit for a sim-
of the measurement. ple membrane in an electrolyte solution neglecting elec-
trode resistances and capacities
Concentration changes at the membrane sur-
face are avoided in resistance measurements with
alternating current. A typical test cell used for the thickness of the membrane and the distance
resistance measurements with alternating current between the electrodes.
is shown in Fig. 3. The membrane resistance is The specic resistance r is:
determined from resistance measurements in a
cell with and without membrane. The resistance Am
rR (4)
of the membrane can be calculated from the fol- d
lowing relation:
The area resistance rm is related to the specic Here Am is the cross-sectional area of the cell as
resistance by: indicated in Fig. 2 and R is the resistance mea-
sured between the electrodes.
For an accurate determination of the membrane
r m rms d m d rs d (3) resistance, it is important that the difference
between the measurement of the cell resistance
Here rm are the area resistances of the membrane, with and without a membrane is as large as possi-
rm+s and rs are the specic resistances of the cell ble. Since the specic resistances of the mem-
with and without the membrane, and dm and d are brane and the electrolyte solution are in the same
order of magnitude and the membrane thickness Ion-exchange membranes display a variety of
dm is small, the distance d has to be as small as electrical properties that suggest that they can be
possible. modeled as a combination of electrical analog
Comparing the direct and alternating current components. For example, an applied potential
membrane characterization processes often signif- can produce a current ow through the membrane,
icant differences in the measured resistances of and this is indication that the membrane has a
more than 50 % is often observed. The alternating nite resistance. In addition, the charged groups
current measurements give a much lower mem- at the membrane/solution interface provide sites
brane resistance than the direct current measure- for the redistribution of charges at the membrane
ments. This is a general nding. However, the surface when the potential is changed. This charge
magnitude of the difference in the results depends redistribution is similar to that observed for a
on the membrane structure. In very heterogeneous capacitor in which a potential change produces a
membranes, the difference is larger than in more change in the net charges. The electrolyte solu-
homogeneous membrane structures. The reason tions and the electrodes on both sides of the mem-
for the different results is that in direct current brane also have a certain resistance. The
measurements, the ions are physically transported properties of the ion-exchange membrane, the
through the membrane matrix, whereas in alter- solution, and the electrodes can be described elec-
nating current measurements, the ions are only trically by an equivalent circuit as indicated in
oscillating around their locations. Assuming the Fig. 3, where Cm and Rm are the membrane capac-
structure of an ion-exchange membrane consists itance and resistance, respectively, and Rs is the
of a microscopic scale of a solid neutral polymer resistance of electrodes and the electrolyte
phase and electrolyte-lled pores, then the mobil- solution.
ity of the ions in the pore liquid is high. But their To determine membrane resistances by imped-
transport through the membrane is hindered by the ance spectroscopy, the entire system, i.e., the
solid polymer structure which acts as a bottleneck membrane, the electrolyte, and the electrodes, is
between pores. This type of structure is typical for treated as a black box. An alternating sinusoidal
uorocarbon membranes. In alternating current voltage of a given frequency and amplitude is
measurements, the transport of the ions on a very applied to the system. The resulting current is
small scale, i.e., in a pore, is measured. In direct measured, and then the phase shift compared to
current measurements, the transport of the ions is the input signal is determined. The procedure is
measured on a macroscopic scale which includes repeated at different frequencies (see entry
also the passage through a bottleneck between Impedance Spectroscopy, Membrane Character-
electrolyte-lled pores. The direct current and ization by).
alternating current difference is treated in detail
in the literature (Zabolotsky and Nikonenko
1993).
A rather simple and reliable method of mea- References
suring resistances is based on impedance spec-
troscopy. The difference between the alternating Spiegler KS (1958) Transport process in ionic membranes.
Trans Faraday Soc 54:14081428
current resistance measurements and the imped- Strathmann H, Bauer B, Rapp HJ (1993) Better bipolar
ance spectroscopy is that in the rst case, the membranes. Chemtech 23:17
frequency of the alternating current is kept con- Strathmann H, Giorno L, Drioli E (2006) An introduction
stant, while in impedance spectroscopy the fre- to membrane science and technology. CNR Publisher,
Roma. ISBN 88-8080-063-9
quency of the alternating current is changed, and Zabolotsky VI, Nikonenko VV (1993) Effect of structural
the response to the changing frequency is deter- membrane inhomogeneity on transport properties.
mined by a spectrometer. J Membr Sci 79:181198
Ion-Exchange Membrane for Fuel Cells 1057
Grotthus mechanism of proton transport and do

Ion-Exchange Membrane for Fuel not require water (Wang et al.1995). These mem-
Cells branes operate at about 180200 C. Whereas
composite peruorosulfonic acid membranes or
Antonino Salvatore Arico sulfonated hydrocarbon including inorganic llers
CNR-ITAE Institute, Messina, Italy such as silica rely on the water-assisted mecha-
nism, they can operate up to 145 C, under par-
ticular conditions (3 bar abs pressure), due to the
Low-temperature fuel cells can be equipped with enhanced water retention of the ller (Aric
a proton or anion exchange membrane in alterna- et al. 2003).
tive to liquid electrolytes (Aric et al. 2001). The Composite recast Naon membranes containing
core of a polymer electrolyte fuel cell (PEMFC) is inorganic llers have been employed in high
the ion exchange membrane. The electrodes temperature (150 C) direct alcohol (Aric
(anode and cathode) are in intimate contact with et al. 1998) and H2-air fuel cells (Watanabe
the membrane faces. The membrane determines et al.1996). These composite membranes were
the fuel cell resistance and the fuel permeation originally developed for reduced humidication
rate, and it inuences the reaction rate. It is well operation in polymer electrolyte fuel cells
I
known that the use of non-noble metal catalysts is (Watanabe et al.1996) due to the enhanced water
possible in the presence of alkaline electrolytes. retention inside the membrane by the effect of the
Protons conducting electrolytes have been pre- inorganic ller (Aric et al. 1998). A further
ferred to alkaline electrolytes for several decades advantage of composite membranes relies in the
for practical reasons, e.g., to avoid carbonation. barrier effect given by the inorganic ller for
The standard electrolyte membrane is usually a methanol crossover (Ren et al. 1996).
peruorosulfonic acid membrane such as Naon.
Most of the electrolytes alternative to Naon both
proton conducting and alkaline type, e.g., hydro-
carbon type, are signicantly cheaper, and in
References
some cases, they are also characterized by smaller Aric AS, Cret P, Antonucci PL, Antonucci V (1998)
hydrogen and methanol crossover. However, life- Comparison of ethanol and methanol oxidation in
time characteristics similar to those shown by a liquid-feed solid polymer electrolyte fuel cell at
Naon-type membranes in fuel cells have not yet high temperature. Electrochem Solid-State Lett
1:6668
been demonstrated for the alternative membranes.
Aric AS, Srinivasan S, Antonucci V (2001) DMFCs: from
Concerning with the conductivity, only recently, fundamental aspects to technology development. Fuel
membrane alternative to Naon type have shown Cells 1:133
similar levels of performance. One critical aspect Aric AS, Baglio V, Di Blasi A, Creti P, Antonucci PL,
Antonucci V (2003) Inuence of the acidbase charac-
is related to the fact that the presence of water is a teristics of inorganic llers on the high temperature
requirement of low-temperature PEMFCs and performance of composite membranes in direct metha-
DMFCs for the occurrence of the electrochemical nol fuel cells. Solid State Ionics 161:251265
reactions and to promote ion conductivity (water- Ren X, Wilson MS, Gottesfeld S (1996) High performance
direct methanol polymer electrolyte fuel cells. J
assisted conductivity mechanism). High ionic
Electrochem Soc 143:L12
conductivities are often associated to the presence Wang J, Wasmus S, Savinell RF (1995) Evaluation of
of large water uptake by the membrane, but this ethanol, 1-propanol, and 2-propanol in a direct oxida-
property often causes poor mechanical character- tion polymer-electrolyte fuel cell. J ElectrochemSoc
142:4218
istics such as large swelling and relevant cross-
Watanabe M, Uchida H, Seki Y, Emori M, Stonehart P
over (especially methanol). Phosphoric acid- (1996) Self-humidifying polymer electrolyte mem-
doped polybenzoimidazole membranes use a branes for fuel cells. J Electrochem Soc 143:3847
1058 Ion-Exchange Membrane Preparation
A polymeric base membrane is formed as a sheet

Ion-Exchange Membrane by casting a polymer solution or polymerization of
Preparation monomers appropriately. Polymerization is carried
out by heat and/or UV radiation with initiators.
Seung-Hyeon Moon Various fabrication processes have been developed
School of Environmental Science and such as sandwich method, latex method, block
Engineering, Gwangju Institute of Science and copolymerization, paste method, and graft poly-
Technology (GIST), Gwangju, South Korea merization (Tanaka 2007). The thickness of a
base membrane is about 100200 mm, thinner
membranes being developed for low electric
resistance. When an ion exchange membrane is
Definition used under harsh conditions such as extreme pH
or high electron activity, hydrogen atom in the
An ion exchange membrane is a membrane polymer backbone is easily attacked, causing deg-
through which ions permeate selectively radation of the membrane. Common approaches to
between two aqueous phases by diffusion or eliminate the a-hydrogen are uses of uorinated
electromigration. Ion exchange membranes are polymers, aromatic chain polymers, and sulfone
prepared with polymeric materials, inorganic group.
salts, or composite materials. This entry describes
the preparation of ion exchange membranes
mostly based on polymeric materials. Like other Functionalization
membranes, an ion exchange membrane should
have a high ux and a high selectivity, often The base membrane is to be functionalized by
characterized by electrochemical properties sulfonation for a cation exchange membrane
representing transport of ions. The desired prop- (CEM) or by amination for an anion exchange
erties of ion exchange membranes are (i) high membrane (AEM). Frequently used sulfonating
ionic conductivity, (ii) high transport number, reagents are sulfuric acid, chlorosulfuric acid,
(iii) good mechanical stability, and (iv) high a mixture of sulfuric acid and chlorosulfuric
chemical and thermal stability. Ion exchange acid, dioxane-SO3 complex, and pyridine-
membranes are prepared by binding ion exchange SO3 complex. Also a base membrane is
functional groups on the polymer backbone with functionalized with aminating agents such as
or without supporting matrix. Therefore, the main trimethylamine (TMA), triethylamine (TEA),
components of ion exchange membrane fabrica- 1,4-diazabicyclo[2.2.2]octane (DABCO), N,N,N,
tion are polymer material for a base membrane, N-tetramethylethylenediamine (TMEDA), and
functional groups, and supporting matrix. tris(2,4,6-trimethoxyphenyl)-phosphine (TTMPP)
to make an AEM (Toshikatsu et al. 2004). Typical
functional groups on polymer backbones for CEMs
Base Membrane are SO3H, -COOH, -PO3H2, phenolic OH,
C(CF2)3OH, and CF2SO2NHR, and those for
The polymers used for ion exchange mem- AEMs are NH2, -NH, -aniline (NH2),
branes are polyethylene, polystyrene, poly(viny- -N(CH3)3OH, -S(CH3)2OH, and -P(CH3)3OH.
limidazole), poly(acrylonitrile), polyphosphazene, Some functional groups are attached by graft
polysulfone, poly(ether ether ketone), poly methods with the help of plasma or gray radiation.
(phenylene oxide), etc. (Sata 2004). The polymer An amphoteric membrane contains both cationic
is selected based on possibility of functionalization, and anionic functional groups in the polymer
polymer processability, chemical and mechanical backbone which are interacting with counter ions
stability, and compatibility with supporting matrix. differently depending on the environmental pH.
Ion-Exchange Membrane Preparation 1059
A mosaic membrane is prepared by blending Heterogeneous Membrane

cation exchange polymer and anion exchange
polymer. Also ion exchange membranes can Heterogeneous membranes are fabricated by
be prepared by polymerization of a monomer binding ion exchange particles and cast on a
that contains a cationic or an anionic moiety plate or lm. Ion exchange resin beads are ground,
with neutral polymer (Strathmann 2004) called by a ball mill or a high speed blender, in preparing
pre-functionalization. Monovalent selective ion ne particles to be uniformly mixed with binder in
exchange membranes are prepared by coating solvent. A heterogeneous ion exchange mem-
regular ion exchange membranes with highly brane is thicker (5001,000 mm) than a homoge-
cross-linked polymer material or oppositely neous ion exchange membrane and advantageous
charged ion exchange layer. in mechanical strength, but exhibiting a high elec-
tric resistance.
Reinforcement
Bipolar Membrane
Often a polymeric base membrane is hydrophobic
without ion exchange capacity. When the base A bipolar membrane is a sandwich-type ion
I
membrane turns to an ion exchange membrane, exchange membrane made with a sheet of cation
the membrane becomes hydrophilic by ionic exchange membrane and that of anion exchange
solvation as ion exchange capacity increases. membrane. Practically a bipolar membrane is pre-
The hydrophilicity leads to high conductivity pared by applying an anion exchange layer on a
and adverse effects such as swelling and mechan- cation exchange membrane, or vice versa. When a
ical weakness of the membrane. To improve the bipolar membrane is prepared with prefabricated
stability of the hydrophilic ion exchange ion exchange membranes, the contacting sides
membrane, reinforcing material is to be intro- should be slightly roughened to apply binder, so
duced. Chemically inert powder (e.g., PVC, that the interface resistance may be minimized.
SiO2, TiO2), nonwoven or woven fabrics (e.g., Since the membrane is used for water dissociation
PVC, bacterial cellulose, electrospun ber), under an electric eld at an overvoltage, catalyst
and porous sheets (e.g., PE, PP, PTFE) are for the water dissociation reaction should be
employed as reinforcing material. This type of placed in the interface between the cation
membranes is called hybrid or composite mem- exchange layer and the anion exchange layer.
branes. Inorganic particles such as SiO2 rather Typical catalysts are soluble metal ions or insolu-
increase the hydrophilicity while passage of ble metal oxides.
non-charged compounds is blocked effectively.
Another way to strengthen the membrane is
cross-linking the polymer chain, since a linear Porous Membrane
polymer chain is exible and readily absorbs
solvent including water. A frequently used Usually ion exchange membranes are considered
cross-linking agent is divinylbenzene. In some as nonporous since their pores are not controlled
cases polymer chains are cross-linked by conden- and ions pass through ion channel with minimized
sation reaction, and functional groups are convective ow. The diameter of ion channel is
involved in the reaction at a high temperature. known as 15 nm which is lled with pendant-
Generally electrochemical properties, i.e., ion type functional groups in an aqueous phase. While
exchange capacity and conductivity, should be nonporous ion exchange membranes are domi-
compromised with chemical and mechanical sta- nantly utilized, porous ion exchange membranes
bility in the membrane preparation with are investigated for limited applications where a
reinforcing materials. high ux transport is required at a slightly lower
1060 Ion-Exchange Membranes
selectivity. Such porous ion exchange membranes hydrocarbon or uorinated hydrocarbon polymers
are prepared by phase inversion to control nano- which carry positively or negatively charged ions
meter size pores, followed by functionalization. xed to the polymer structure (Xu 2005; Bergsma
Alternatively pre-functionalized polymer and Kruissink 1961; Molau 1981). The type and
dissolved in solvent is cast and dried in air or the concentration of the xed ions in a membrane
nonsolvent (e.g., water) to form pores in the structure determine permselectivity and the elec-
membrane. trical resistance of a membrane, while the chemi-
cal stability and the mechanical properties of the
membrane are determined mainly by the matrix
References polymer. There are three types of ion-exchange
membranes:
Sata (2004) Chapter 3: Preparation of ion exchange
membranes. In: Ion exchange membranes: preparation,
characterization, modication, and application. RSC, Cation-exchange membranes which contain
Cambridge, UK xed negatively charged ions and which have
Strathmann (2004) Chapter 3: Preparation and characteri-
zation of ion-exchange membranes. In:Ion-exchange a selective permeability for cations
membrane separation processes. Elsevier, Amsterdam, Anion-exchange membranes which contain
The Neterlands xed positively charged ions and which have
Tanaka Y (2007) Chapter 1: Preparation of ion exchange a selective permeability for anions
membranes. In: Ion exchange membranes fundamen-
tals and applications. Elsevier, Oxford Bipolar membranes which consist of a cation-
and an anion-exchange membrane laminated
together
Ion-Exchange Membranes The ions often used as xed charges in cation-

exchange membranes are SO 3 and COO .

Heiner Strathmann Fixed charges used in anion-exchange mem-

Universitt Stuttgart, Institute of Chemical branes are N H R2 and N R3 .
Process Engineering Stuttgart, Baden- The structure of a cation-exchange membrane
Wurttemberg, Germany is illustrated in Fig. 1 which shows the polymer
matrix with the xed negative ions and the mobile
counterions as well as their pathway through the
Ion-exchange membranes. Most of todays ion- membrane.
exchange membranes used in commercially The most desired properties for ion-exchange
relevant processes are 0.21 mm thin sheet of membranes are:
Ion-Exchange
Membranes,
illustrating the structure of a
cation-exchange membrane
showing xed negative ions
and mobile positive
counterions in the polymer
matrix
Ion-Exchange Membranes in Chlor-Alkali Process 1061
High permselectivity an ion-exchange References

membrane should be highly permeable for
counterions, but should be impermeable to Bergsma F, Kruissink CH (1961) Ion-exchange mem-
branes. Fortschr Hochpolym Forsch 21:307362
co-ions.
Molau GE (1981) Heterogeneous ion-exchange mem-
Low electrical resistance the permeability of branes. J Membr Sci 8:309330
an ion-exchange membrane for the counterions Simons R (1993) Preparation of high performance bipolar
under the driving force of an electrical poten- membranes. J Membr Sci 78:1323
Xu TW (2005) Ion-exchange membranes: state of their
tial gradient should be as high as possible.
development and perspective. J Membr Sci 263:129
Good mechanical and form stability the
membrane should be mechanically strong and
should have a low degree of swelling or shrink-
ing in transition from dilute to concentrated ion
solutions. Ion-Exchange Membranes
High chemical stability the membrane should in Chlor-Alkali Process
be stable over the entire pH range from 1 to
14 and in the presence of oxidizing agents. Csar Valderrama
Department of Chemical Engineering, Universitat
Politcnica de Catalunya-Barcelona Tech (UPC),
I
For the practical preparation of ion-exchange
membranes, two rather different procedures are Barcelona, Catalonia, Spain
used. The rst procedure results in a homoge-
neous ion-exchange membrane structure and is
closely related to the preparation of ion-exchange Ion-exchange membranes in chlor-alkali
resins. Homogeneous ion-exchange membranes process (also called ion exchange membranes in
are produced by either a polymerization of mono- chlor-alkali membrane cell process) are the syn-
mers that carry anionic or cationic moieties or by thetic membranes employed to produce chlorine
introducing these moieties into a polymer which (Cl2) and alkali, sodium hydroxide (NaOH) or
may be in an appropriate solution or a solid potassium hydroxide (KOH) by electrolysis of a
preformed lm. The second widely used salt solution. These membranes are permeable to
ion-exchange membrane preparation technique either positively or negatively charged ions in
which leads to a rather heterogeneous structure aqueous solution used in the chlor-alkali industry.
is based on mixing an ion-exchange resin powder The membranes that are selectively permeable to
with a binder polymer, such as polyvinylchloride positively charged ions are usually named cation-
or polyethylene, and extruding the mixture as a exchange membranes (CEM), and membranes
lm at a temperature close to the melting point of selectively permeable for negatively charged
the polymer. ions are called anion membranes (AEM). The
A cation-exchange membrane with excep- selectivity occurs due to high concentration of
tional good chemical and thermal stability immobile (xed) ions within the membrane body.
which is widely used in the electrolytic chlorine- Cation membranes have negatively charged
alkaline production and as polymer electrolyte xed ions usually sulfonic or carboxyl groups
in low-temperature fuel cell consists of a chemically bound with the membranes matrix.
polyuorocarbon material. This membrane is Their charge is neutralized by positively charged
often referred to as Naon membrane which ions (counterions). An anion membrane would
is the trade name of DuPont. There are several have positively charged xed ions, usually qua-
more ion-exchange membranes with special prop- ternary ammonium groups, and negatively
erties commercially available such as the bipolar charged counterions. Fixed ions and counterions
membrane which is composed of an anion- and a are connected by ionic bounds in a dry membrane.
cation-exchange layer laminated together In the swollen membrane, this bond would be
(Simons 1993). dissociated. Therefore, the counterion is mobile
1062 Ion-Exchange Membranes in Chlor-Alkali Process
Ion-Exchange
Membranes
in Chlor-Alkali Process,
Fig. 1 Membrane cell
process in chlor-alkali
industry
and can be replaced by another ion. Hence the a separator by forming an alloy of sodium and
membrane would be permeable to ions with the mercury (sodium amalgam) which is subse-
charge sign opposite to the charge of the xed ion. quently reacted with water to form sodium
This unique property makes ion-exchange mem- hydroxide and hydrogen in a separate reactor
brane applications very attractive for chemical (OBrien et al. 2005).
industry, because it allows for the removal, addi- The membrane cell technology has several
tion, substitution, depletion, or concentration of advantages compared to the other processes:
ions in process solutions (Baker 2000; Nunes and high energy efcient, high purity of caustic, and
Peinemann 2001). low environmental impact; the membrane process
The chlor-alkali (also called chlorine- does not use highly toxic materials such as mer-
caustic) industry is one of the largest electro- cury and asbestos. Disadvantages of the mem-
chemical technologies in the world. It is an brane process are that the caustic soda produced
energy-intensive process and is the second largest may need to be evaporated to increase concentra-
consumer of electricity (2400 billion kWh) among tion and, for some applications, the chlorine gas
electrolytic industries. The chlor-alkali industry produced needs to be processed to remove oxy-
produces chlorine and caustic solution (sodium gen. Furthermore, the brine entering a membrane
or potassium hydroxide) simultaneously by cell must be of a very high purity, which often
means of decomposition of a saturated sodium requires costly, additional purication steps prior
chloride solution often called brine. Along to electrolysis. Nowadays, the membrane cell pro-
with the chlorine and the caustic solution, hydro- cess is the most used in the world, and in Europe,
gen is produced. There are three basic processes complete conversion to the membrane technology
for the electrolytic production of chlorine. These is expected in the next years due to the mandatory
three processes are the diaphragm cell process, the deadline (2020) to phase out the mercury cells
mercury cell process, and the membrane cell pro- (Euro Chlor 2011; de Bastos Vidal 2013).
cess. The distinguishing difference between the In a membrane cell, an ion-exchange mem-
technologies relies in the manner by which the brane separates the anode and cathode compart-
anolyte and the catholyte streams are prevented ments (Fig. 1). The separator is generally a bilayer
from mixing with each other. Separation is membrane made of peruorocarboxylic and
achieved in a diaphragm cell by a separator and peruorosulfonic acid-based lms, intercalated
in a membrane cell by an ion-exchange mem- between the anode and the cathode. The saturated
brane. In mercury cells, the cathode itself acts as brine is fed to the anode compartment where
Ion-Exchange Membranes in Chlor-Alkali Process 1063
Ion-Exchange
Membranes
in Chlor-Alkali Process,
Fig. 2 Illustrative scheme
of the structure of different
commercial membranes
(de Bastos Vidal 2013;
Vogel and Meier-Haack
2014)
chloride ions are oxidized to chlorine gas. The was quite fast due to impurities in precipitation in
sodium ions migrate through the membrane to the membrane structure. Later on, the develop-
the cathode compartment. The demineralized ment of a laminated membrane structure with
water added to the catholyte circuit is hydrolyzed, two different functional groups revolutionized
releasing hydrogen gas and hydroxide ions. The the chlor-alkali membrane cell process by improv-
sodium and hydroxide ions combine to produce ing the current efciency and reducing the oper-
I
caustic soda which is typically brought to a con- ating potential (Navin 2002; de Bastos Vidal
centration of 3235 % by recirculating the solu- 2013).
tion before it is discharged from the cell. The Nowadays, the commercially available mem-
reactions evolved in this process are the following branes in the chlor-alkali industry may have from
(IPPC 2014): one up to three layers but generally consist of two
At the anode: layers with a polymeric matrix made of tetrauor-
oethylene. Sulfonic and carboxylic groups are
2Cl ! Cl2 2e (1) used as xed ionic groups on the anode and cath-
ode side of the membrane, respectively (Fig. 2).
At the cathode: The use of different ionic groups, as illus-
trated in Fig. 2, has signicant effects on the
2H2 O 2e ! H2 2OH (2) permselectivity and conductivity of each layer.
Furthermore, the sulfonic groups are excellent
Depleted brine is discharged from the anode com- proton conductors while the carboxylic groups
partment and re-saturated with salt. If needed, to are highly permselective to cations due to their
reach a concentration of 50 % caustic soda, the lower water content (Sata 2000). The membranes
caustic liquor produced has to be concentrated by must remain stable while being exposed to chlo-
evaporation (using steam). rine on one side and a strong caustic solution on
The rst generation of commercial membranes the other. The mechanical properties of the mem-
(1975) was able to directly produce sodium branes (tensile and tear strength) are greatly
hydroxide solutions of 240 wt %. They were improved by reinforcements; a PTFE woven
made from uoropolymers functionalized with cloth is used to reinforce the membrane structure
sulfonic groups. These membranes were rela- to prevent damage or tearing of the membrane.
tively thick with heavy reinforcement cloths and There are different types of cloths to reinforce the
the electrode gap was 9 mm. Due to these mem- membranes and in some cases sacricial bers are
brane characteristics, the typical operating poten- also used. More and thinner bers increase
tials were high (3.84.0 V at 2.5 kA/m2) that mechanical strength besides providing low volt-
combined with low current efciency give a DC age and high current efciency. To prevent the
energy consumption of 3300 kWh/ton of caustic adherence of bubbles to the membrane surface,
soda. Additionally, the current efciency decline the membrane surfaces are made hydrophilic. The
1064 Ion-Exchange Mosaic Membranes
general economic lifetime of chlor-alkali mem- Nunes SP, Peinemann KV (2001) Membrane technology in
branes is approximately 3 years but ranges the chemical industry. Wiley-VCH Verlag GmbH,
Weinheim/New York
between 2 and 5 years (Strathmannm 2004; de OBrien TF, Bommaraju TV, Hine F (2005) Handbook of
Bastos Vidal 2013). chlor-alkali technology. Springer, New York
The peruorinated membranes developed for Sata T (2000) Membranes process: preparation and char-
use in chlor-alkali cells exhibit super selectivity acterization of ion-exchange membranes. In: Darwish
MK, Al-Gobaisi DMK (eds) Encyclopedia of desalina-
and high thermal and chemical stability, which tion and water resources (DESWARE). EOLSS,
are not found in other classes of polymeric Oxford
ion-selective membranes. There are four main Strathmannm H (2004) Ion-exchange membrane separa-
manufacturers of ion-exchange membranes to tion processes, vol 9, 1st edn, Membrane science and
technology. Elsevier, Amsterdam/Boston
the chlor-alkali industry: DuPont (Naon mem- Vogel C, Meier-Haack J (2014) Preparation of
branes), Asahi Glass (Flemion membranes), ion-exchange materials and membranes. Desalination
Solvay Plastics (formerly Dow membranes), and 342:156174
Asahi Kasei (Aciex membranes). The differ-
ences between suppliers are related to the modi-
cations made on the polymer and on the functional
groups, the type of reinforcement used, and other Ion-Exchange Mosaic Membranes
chemical and mechanical details (OBrien
et al. 2005). Mosaic Membranes
The membranes commercially available for the
chlor-alkali industry can be divided in two classes
regarding their characteristics: high performance
and high mechanical strength membranes. Iron Oxides: Polymer Composite
Low-resistance membranes are more adequate to Membranes
narrow gap electrolyzers due to their low voltage.
On the other hand, high- strength membranes are Zbigniew J. Grzywna and Gabriela Dudek
used in more robust electrolyzers where the main Department of Physical Chemistry and
requirement is high physical strength. Today, Technology of Polymers, Silesian University of
membranes operate at higher current densities Technology (SUT), Gliwice, Poland
with current efciencies of about 97 % producing
a caustic soda stream of 32 wt %. The typical DC
energy consumption values are as low as One of the best modication techniques for
2100 kWh/ton of caustic soda. enhancing membrane performance is blending
with inorganic nanoparticles such as iron oxides
(Dudek et al. 2012, 2013, 2014), ZrO2 (Gestel
et al. 2006), TiO2 (Gestel et al. 2006; Hong
References et al. 2013), Al2O3 (Hong et al. 2013), and carbon
nanotubes (Gethard et al. 2011). In hybrid mem-
Baker RW (2000) Membrane technology and applications,
2nd edn. Wiley, Chichester branes, inorganic components are dispersed in a
de Bastos Vidal AC (2013) Chlor-alkali membrane cell polymer matrix. These membranes benet in
process: study and characterization. PhD dissertation, the selectivity of active components and the
University of Porto
simplicity of polymeric membrane processing.
Euro Chlor (2011) Chlorine industry review 20102011.
Brussels. www.eurochlor.org. Accessed June 2014 Particularly, the usage of inorganic-organic com-
IPPC (2014) Best available techniques (BAT) reference posite materials is becoming increasingly impor-
document for the production of chlor-alkali, Denitive tant due to their extraordinary properties, which
document. Integrated Pollution Prevention Control
arise from the synergism between the properties
(IPPC), Seville
Navin T (2002). Membrane cell technology state of the of the individual components as well as their
art industry. Eltech Systems, Chardon interactions with the base matrix materials
Iron Oxides: Polymer Composite Membranes 1065
(Dudek et al. 2012). The separation properties of upon the synthesis method and conditions, iron
the nal composite membranes depend on com- oxide can be obtained in one of several phases,
ponent compatibility and dispersibility of the inor- such as FeO (wustite, antiferromagnetic), Fe3O4
ganic phase in the polymer matrix. Polymer (magnetite), Fe(II)Fe(III)2O4 (ferromagnetic or
grafting on the surface of the inorganic superparamagnetic when the size is less than
nanoparticles seems to be the best modication 15 nm), a-Fe2O3 (hematite, weakly ferromagnetic
method for the dispersibility improving of the or antiferromagnetic), g-Fe2O3 (maghemite, ferri-
inorganic phase in the polymer matrix. Another magnetic), e-Fe2O3, and b-Fe2O3 (Si et al. 2008),
surface modiers often used for the stabilization among which magnetite and maghemite are the
of particles are fatty acids (oleic, lauric, palmitic, very promising and widely tested candidates for
or stearic). Using those modiers, the inorganic separation application.
components could be well dispersed in polymer The presence of magnetic particles in polymer
solutions forming uniform dispersion (Liu and Su matrix changes the transport mechanism. Changes
2005). In the case of higher concentration of inor- in hydrophobicity/hydrophilicity balance of
ganic phase in polymer solution, sedimentation nanocomposite membrane containing metal
phenomenon can be observed and further aggre- oxide were reported by Balta et al. They noticed
gation and inhomogeneous distribution, what was that even ultralow concentration of embedded
I
discussed in Dudek et al. (2012). ZnO nanoparticles (0.035 w/w %) into PES
In recent years, the nanocomposite mem- matrix had an important impact on the increasing
branes, as a special functional material, have of membrane hydrophilicity (Balta et al. 2012).
attracted much attention and been applied in load- Dudek et al. (2012, 2013, 2014) investigated the
ing of biomolecules (Nicols et al. 2013), biosen- inuence of iron oxide nanoparticles on chitosan
sor (Chena et al. 2012), metal adsorption (Stopa and PPO (poly(2,6-dimethyl-1,4-phenylene
and Yamaura 2010), and others (Tsai et al. 2010). oxide)) membrane separation properties. In both
Especially, Fe3O4 ferrite magnetic nanoparticles cases, the addition of magnetite nanoparticles to
(Dudek et al. 2012, 2013, 2014) have been rising the polymer matrix has an impact on separation of
as a signicant useful material due to their specic ethanol/water mixtures. The main important
properties such as superparamagnetic, nontoxic, changes in transport properties are observed com-
great surface/area ratio, small size, etc. These paring membranes without and with the iron
materials exhibit novel macroscopic magnetic oxide nanoparticles. The diffusion coefcient
properties due to the combined action of increases for one component and decreases for
quantum-sized effects, strain and surface effects, the other with a larger amount of magnetite pow-
interface between nanostructures and the der. For the hydrophobic PPO membrane, the
matrix, matrix properties, and morphologies of component preferentially passing through the
nanostructures. Compared to the traditional membrane is ethanol. It has little tendency to
micro-sized magnetic supports used in separation absorb water, so that a droplet remains on the
process, nano-sized magnetic carriers possess surface. The opposite situation occurs for the
quite a good performance due to higher specic hydrophilic membranes. Hydrophilic membranes
surface area and lower internal diffusion resis- have an afnity for water. Their surface chemistry
tance (Dudek et al. 2012, 2013, 2014). Iron allows these materials to be wetted immediately.
oxide nanoparticles are being extensively For such membranes, water preferably passes
exploited in several scientic and technological through the membrane. Comparing the separation
applications such as separation processes (Dudek properties of both discussed membranes on the
et al. 2012, 2013, 2014), targeted drug delivery basis of permeability coefcients, it can be
(Saboktakin et al. 2010), magnetic resonance observed that the poly(2,6-dimethyl-1,4-
imaging (Andreas et al. 2012), hyperthermia and phenylene oxide) with magnetite addition has
cancer therapy (Santhosh and Ulrih 2013), water better separation properties than the hybrid
remediation (Crane et al. 2011), etc. Depending chitosan membrane. It is demonstrated by the
1066 Iron Oxides: Polymer Composite Membranes
6
450
EtOH (PPO membrane)
400
20
Permeation coefficients for water and ethanol P 10
P 10 [Cm STP cm/cm cmHg s]

350
5

2 300
250
[Cm STP cm/cm cmHg s]
4
3
200
Water (PPO membrane)
20
2
150

Water (chitozan membrane)

100
EtOH (chitozan membrane)

3 50
3
0
0 2 4 6 8 10 12
0
0 2 4 6 8 10 12 14 16
Amount of magnetic particles [% w/w]
Iron Oxides: Polymer Composite Membranes, Fig. 1 The comparison of permeability coefcients of water and
ethanol for PPO and chitosan membranes
greater difference between the received perme- Theoretical Background

ability coefcient for water and ethanol (Fig. 1).
The uxes and diffusion coefcient values also The basic law of a phenomenological, parabolic
conrm this relation. Large difference between diffusion through a slab membrane of unit cross
ux and permeation coefcient, for the same section is the Ficks law:
amount of magnetite, gives a chance for effective @c
J D (1)
and efcient separation of ethanol/water mixture @x
and gives a chance to concentrate ethanol solution
by faster permeation of water from the feed. The where J is the diffusion ux, c stands for the
reason of bigger difference between the water and concentration of the diffusing species, and D is
ethanol permeability coefcient for hybrid PPO diffusion coefcient (D dot stands for the group
membrane is probably the fact that the iron oxide of variables x, t, c).
powder, which is added to the polymer matrix, has When the external force eld acts, we rewrite
hydrophobic properties. The PPO matrix has also Eq. 1 in the form:
hydrophobic properties, so in this case the inter- @c
action between polymer matrix and ferroferric J D wc (2)
@x
oxide particles is very good. The opposite situa-
tion is for chitosan membranes with hydrophilic where w is a drift coefcient (Strzelewicz and
properties. For such membranes, the polymer Grzywna 2007).
matrix and iron oxide powder give an opposite Crank and Park in (1968) reviewed the more
effect. As a consequence permeability coefcients useful experimental techniques for measuring dif-
for the component preferentially transported fusion coefcient. The most fundamental quan-
through the membrane are not able to reach as tity, from many possible, is an average diffusion
high values of transport coefcients as in the case coefcient D calculated from the stationary per-
of PPO membrane. meation according to the evident formula, i.e.:
Iron Oxides: Polymer Composite Membranes 1067
Iron Oxides: Polymer

Composite Membranes, Q a (I,t)
downstream absorption
permeation curve
stationary
permeation
transient late time zone

transient early time zone
t
Ico /6 La(1) 6 La(I)
Js l Gadomski et al. 1993; Grzywna 1996; Grzywna

D cm2 =s (3)
Dc0 and Stolarczyk 2005; Strzelewicz and Grzywna
2007, 2008; Rybak et al. 2009, 2012).
where Js is a diffusive mass ux in a stationary I
state, l is the thickness of membrane, and Dc0 can
D2 2 2ew=Dl 2Dwl w2 l2
be obtained from an intercept of the asymptote to L^ l
a
the stationary permeation curve with the Qa (l, t) 2w3 l

axis (total ow of penetrant) (see Fig. 2). (5)
Comparing D with the value of diffusion coef-
cient DL calculated from the downstream Based on the rst Ficks law, we obtain the equa-
absorption time lag (Grzywna and uczka 1991; tion for permeation coefcient determination
Gadomski et al. 1993; Grzywna 1996; Grzywna (Dudek et al. 2013):
and Stolarczyk 2005; Strzelewicz and Grzywna
2007, 2008; Rybak et al. 2009, 2012) (Fig. 2) Js l
according to the following equation: P (6)
Dp
l2
DL cm2 =s (4) where:
6La l
P permeation coefcient [Barrer]
we can get some insight into the nature of the l membrane thickness [cm]
transport process in question, namely, if D DL , Dp difference of gas pressure at both sides of the
the diffusion is supposed to be an ideal Fickian membrane [cmHg]
unless there are some hidden processes, like a
Js diffusive mass ux cm3STP =cm2 s
drift, which are beyond the reach of this simple
protocol (Grzywna and uczka 1991; Gadomski Diffusive mass ux is dened as:
et al. 1993; Grzywna 1996; Grzywna and
Stolarczyk 2005; Strzelewicz and Grzywna
2007, 2008; Rybak et al. 2009, 2012). In fact, in QSTP
Js (7)
the case of the magnetic membranes, the pres- A
ence of a drift is almost certain. To calculate the
realistic value of drift and the average diffusion where:
coefcient, we should use a proper theory from QSTP ow rate at standard condition
which the following formula for time lag L ^ a l [cm3STP/s]
can be derived (Grzywna and uczka 1991; A membrane active area [cm2]
1068 Iron Oxides: Polymer Composite Membranes
Flow rate at standard condition is dened as hydroxide solution (24 %). The decanted precip-
(Dudek et al. 2013): itate was washed twice by mixing with a diluted
ammonium hydroxide solution (5 ml of ammo-
T STP p nium hydroxide in 95 ml of water). After the
QSTP Q (8)
T pSTP second washing, the volume of precipitate slurry
was adjusted to 50 ml and 4 g of dodecanoic acid
where: (lauric) was added. The next mixture was contin-
uously stirred and heated for 2 min using 550 W
Q measured ow rate [cm3/s] laboratory heating plate. Obtained stable magne-
p atmospheric pressure [Pa] tite uid, after ltration in a solid state, was used
T atmospheric temperature [K] for preparation of membrane.
pSTP standard pressure, pSTP = 1.013105 Pa Hybrid polymer membranes were prepared by
TSTP standard temperature, TSTP = 298,15 K cast method. For this purpose polymer (powder)
was dispersed and dissolved in appropriate sol-
The solubility coefcient is a measure of the vent (Dudek et al. 2012, 2013). After dissolution,
size of the sorption in membranes. It is character- the magnetite nanoparticles were added to the poly-
ized by division of the penetrant between the mer solution. For proper dispersion and homoge-
membrane and the outer phase in equilibrium. nization of iron oxide particles in the polymer
This parameter is calculated from the relation: solution, the mixture was sonicated using high-
intensity ultrasound radiation. Sonication time
P was 30 min. After mixing, the solutions were
S (9)
D poured into the Petri dishes, placed on the leveled
plate, and then evaporated at room temperature for
where: 24 h. In the case of hydrophilicity membranes, the
h i suitable cross-linking agent was applied.
cm3STP
S solubility coefcient (distribution) cm3 cmHg
Ideal selectivity coefcient could be given by

quotient of given permeation coefcients: References
Pi Andreas K, Georgieva R, Ladwig M, Mueller S, Notter M,

ai=j (10) Sittinger M, Ringe J (2012) Highly efcient magnetic
Pj stem cell labeling with citrate-coated superpara-
magnetic iron oxide nanoparticles for MRI tracking.
where Pi, Pj is the permeation coefcient of Biomaterials 33(18):45154525
separated components. Balta S, Sotto A, Luis P, Benea L, Van der Bruggen B,
Kim J (2012) A new outlook on membrane enhance-
ment with nanoparticles: the alternative of ZnO.
Membrane Preparation Chena H, Zhanga Z, Luoa L, Yaoc S (2012) Surface-
imprinted chitosan-coated magnetic nanoparticles
Polymer membranes containing iron oxide modied multi-walled carbon nanotubes biosensor for
nanoparticles were prepared using suitable poly- detection of bovine serum albumin. Sensors Actuators
B 163:7683
mer as a matrix, but rst of all ferroferric oxide
Crane RA, Dickinson M, Popescu IC, Scott TB
particles were prepared using the precipitation (2011) Magnetite and zero-valent iron nanoparticles
method described elsewhere (Khalafalla and for the remediation of uranium contaminated environ-
Reimers 1980). For example, 12 g of FeCl24H2O mental water. Water Res 45(9):29312942
Crank J, Park G (1968) Diffusion in polymers. Academic,
and 24 g of FeCl36H2O were each dissolved in
New York
50 ml of water and combined into a 600 ml beaker Dudek G, Turczyn R, Strzelewicz A, Rybak A,
with addition of 50 ml of concentrated ammonium Krasowska M, Grzywna ZJ (2012) Preparation and
Irreversible Flux Decline 1069
characterization of iron oxides-polymer composite Santhosh PB, Ulrih NP (2013) Multifunctional superpar-
membranes. Sep Sci Technol 47:13901394 amagnetic iron oxide nanoparticles: promising tools in
Dudek G, Turczyn R, Strzelewicz A, Krasowska M, cancer theranostics. Cancer Lett 336(1):817
Rybak A, Grzywna ZJ (2013) Studies of separation of Si H, Zhou C, Wang H, Lou S, Li S, Du Z, Li LS
vapours and gases through composite membranes with (2008) Controlled synthesis of different types iron
ferroferric oxide magnetic nanoparticles. Sep Purif oxides nanocrystals in parafn oil. J Colloid Interface
Technol 109:5563 Sci 327:466471
Dudek G, Strzelewicz A, Turczyn R, Krasowska M, Rybak Stopa LCB, Yamaura M (2010) Uranium removal by
A (2014) The Study of Ethanol/Water Vapors Perme- chitosan impregnated with magnetite nanoparticles:
ation through Sulfuric Acid Cross-Linked Chitosan adsorption and desorption. Int J Nucl Energ Sci
Magnetic Membranes. Sep Sci Technol 49 Technol 5(4):283289
(11):17611767 Strzelewicz A, Grzywna ZJ (2007) Studies on the air
Gadomski A, Grzywna ZJ, uczka J (1993) The growing membrane separation in the presence of a magnetic
processes in diffusive and convective elds. Chem Eng eld. J Membr Sci 294(12):6067
Sci 48(21):37133721 Strzelewicz A, Grzywna ZJ (2008) On the permeation time
Gestel TV, Kruidhof H, Blank DHA, Bouwmeester HJM lag for different transport equations by Frisch method.
(2006) ZrO2 and TiO2 membranes for nanoltration J Membr Sci 322(2):460465
and pervaporation. Part 1. Preparation and characteri- Tsai Z, Wang J, Kuo H, Shen C, Wang J, Yen T (2010) In
zation of a corrosion-resistant ZrO2 nanoltration situ preparation of high relaxivity iron oxide
membrane with a MWCO <300. J Membr Sci nanoparticles by coating with chitosan: a potential
284:128136 MRI contrast agent useful for cell tracking. J Magn
Gethard K, Sae-Khow O, Mitra S (2011) Water desalina- Magn Mater 322(2):208213 I
tion using carbon-nanotube-enhanced membrane distil-
lation. ACS Appl Mater Interfaces 3(2):110114
Grzywna ZJ (1996) Scaling in diffusive transport through
membranes. Chem Eng Sci 51(17):41154125
Grzywna ZJ, uczka J (1991) Parabolic versus hyperbolic
diffusion in membrane-transport. Acta Pharm Jugosl
41(4):327344 Irreversible Flux Decline
Grzywna ZJ, Stolarczyk JK (2005) Diffusion in glassy
polymers from random walks to partial differential
Tony Fane
equations. Acta Phys Polon B 36(5):15951611
Hong S, Hua S, Yan L (2013) Preparation and characteri- School of Chemical Engineering, The University
zation of Al2O3/TiO2/PVDF polymer composites ultra- of New South Wales, Sydney, NSW, Australia
ltration membrane. Appl Mech Mater Singapore Membrane Technology Centre,
253255:865870
Nanyang Technological University, Singapore,
Khalafalla S, Reimers G (1980) Preparation of dilution-
stable aqueous magnetic uids. IEEE Trans Magn Singapore
16:178183
Liu P, Su ZX (2005) Thermal stabilities of polystyrene/
silica hybrid nanocomposites via microwave-assisted
Membranes are either operated at constant pres-
insitu polymerization. Mater Chem Phys 94:412416
Nicols P, Saleta M, Troiani H, Zysler R, Lassalle V, sure (P) or constant ux (J); the former is more
Ferreira ML (2013) Preparation of iron oxide typical of lab scale studies and the latter of com-
nanoparticles stabilized with biomolecules: experimen- mercial operation. Fouling leads to ux decline
tal and mechanistic issues. Acta Biomater 9:47544762
under constant pressure or a rise in the required
Rybak A, Krasowska M, Strzelewicz A, Grzywna ZJ
(2009) Smoluchowski type equations for modelling transmembrane pressure DP for constant ux.
of air separation by membranes with various structure. These trends are illustrated in Fig. 1. It should be
Acta Phys Polon B 40(5):14471454 noted that fouling under constant pressure
Rybak A, Strzelewicz A, Krasowska M, Dudek G,
becomes self-limiting; the lower the ux, the
Grzywna ZJ (2012) Inuence of various parameters
on the air separation process by magnetic membranes. slower the fouling. For constant ux operation,
Sep Sci Technol 47(9):13951404 the fouling can be self-accelerating, often leading
Saboktakin MR, Tabatabaie RM, Maharramov A, to a sudden rise in TMP (DP) as shown for case
Ramazanov MA (2010) A synthetic macromolecule
(ii) in Fig. 1b. This TMP jump could be associ-
as MRI detectable drug carriers: aminodextran-coated
iron oxide nanoparticles. Carbohydr Polym ated with a critical consolidation of the fouling
80(3):695698 layer and rapid rise in fouling resistance.
1070 Irreversible Fouling
a b
M1
FLUX
TMP
(i)
M2 (ii)
Time Time
Flux vs time for Constant Pressure operation Pressure vs time for Constant Flux operation
Irreversible Flux Decline, Fig. 1 Constant pressure and constant ux operation: fouling trends
Irreversible ux decline is the drop in ux due to

irreversible fouling. The effect of irreversible Irreversible Fouling
fouling can also be expressed as irreversible per-
meability decline (given by the change in J/DP), Tony Fane
and this can be used for both constant pressure and School of Chemical Engineering, The University
constant ux operation. However, it is important of New South Wales, Sydney, NSW, Australia
to note that fouling is ux driven. As a result it is Singapore Membrane Technology Centre,
more appropriate to compare membranes at the Nanyang Technological University, Singapore,
same initial ux for constant pressure tests. Fig- Singapore
ure 1a illustrates two membranes, M1 and M2,
where M1 has an initially higher permeability and
starts at a higher ux. Fluxes for M1 and M2 both
decline due to the increasing fouling resistance Membrane fouling is the accumulation of material
(see Irreversible Fouling Resistance), and both on the surface of or within the membrane struc-
membranes asymptote to similar declined uxes. ture. The foulant may provide a cake or surface
This is due to the eventual dominance of the deposit layer, and if the membrane is micropo-
fouling resistance (RF) over the membrane resis- rous, the foulant could cause pore restriction/clo-
tance (Rm) (see Eqs. 1 and 2 in Irreversible sure or pore plugging. These fouling mechanisms
Fouling Resistance). If membranes M1 and M2 are depicted in Fig. 1. Membrane fouling is prev-
were tested at the same constant ux, it would be alent in all the liquid-phase membrane processes.
easier to identify if one was intrinsically less prone Fouling species can be organic macromolecules
to fouling. (proteins, polysaccharides, etc.), inorganics
(scale precipitates), and colloidal or biological
(biolms). Fouling is linked to, but different
from, concentration polarization (CP). CP is a
References buildup of retained species (solutes) adjacent to
the membrane surface and can be depicted as a
Fane AG, Chong TH, Le-Clech P (2009) Fouling in mem- boundary layer concentration prole based on a
balance between convection (ux driven) and
transformations. Wiley-VCH, Weinheim. ISBN back transport (diffusion). CP increases with ux
978-3-527-32038-7 but is totally reversible when ux is ceased;
Irreversible Fouling Resistance 1071
1 Pore Closure 2 Pore Plugging 3 Cake (surface)

deposition
Irreversible Fouling, Fig. 1 Membrane fouling mechanisms
fouling is not totally reversible. In many situations References

fouling has both reversible and irreversible com-
ponents. The reversible component may be Fane AG, Chong TH, Le-Clech P (2009) Fouling in mem-
removed by hydrodynamic means, such as back-
wash or increased shear at the membrane surface. transformations, Wiley-VCH Weinheim. ISBN:
This type of foulant is usually loosely deposited or 978-3-527-32038-7
bound on the membrane and subject to removal by
raised liquid velocities. However, it is not I
removed by simply dropping ux to zero. The
irreversible fouling is material that is tightly held
Irreversible Fouling Resistance
by the membrane. It could be colloids plugging
pores, macrosolutes bound by adsorption, insolu-
Tony Fane
ble salts spreading as crystals, and biolms com-
School of Chemical Engineering, The University
prising bacteria held in a matrix of extracellular
of New South Wales, Sydney, NSW, Australia
polymeric substances (EPS). Irreversible fouling
Singapore Membrane Technology Centre,
can usually be mitigated by feed pretreatment,
Nanyang Technological University, Singapore,
suitable hydrodynamics, and careful membrane
Singapore
selection. For example, pretreatment could
remove colloids or bacteria, or involve the addi-
tion of antiscalants. Hydrodynamic control could
Flux can be related to the driving force (DPDP)
match crossow velocity to the imposed ux to
and the overall resistance, RT, by,
operate below the critical ux, where the critical
ux is the ux below which fouling is negligible.
DP DP
Membrane selection would consider pore size, J (1)
m R T
surface charge, hydrophobicity/hydrophilicity,
etc. Irreversible fouling is usually partially remov-
The osmotic pressure term (DP) can be ignored
able by cleaning regimes, which could range from
for non-osmotic operation (i.e., MF applications).
daily to monthly, with chemical or physical
The resistance components are assumed to be in
agents, depending on the situation. However, the
series, so,
gradual buildup of residual irreversible fouling,
over multiple cleaning cycles, may eventually
reach a critical level, and membrane replacement RT Rm R F (2)
is required.
For more information on membrane fouling, where Rm is the membrane resistance and RF is the
see Irreversible Flux Decline and Irre- fouling resistance, which may comprise a revers-
versible Fouling Resistance. ible (Rrr) and irreversible (Rir) component, i.e.,
1072 Isotherm Palladium Phase Diagram
Irreversible Fouling a CB b CB
Resistance, Fig. 1 (a)
Concentration
polarization clean
membrane. (b) Cake-
enhanced concentration
Shear
polarization fouled Cake
membrane
CW1 CW2
Membrane
CP1 CP2
RF Rrr Rir (3) References
For a given DP, the ux vs. time data gives the Hoek EMV, Elimelech M (2003) Cake enhanced concen-
tration polarization: a new fouling mechanism for salt-
various resistances. The initial clean water ux rejecting membranes. Environ Sci Technol
gives Rm, the nal fouled ux gives RT, and 37:55815588
RT Rm gives RF. A water ush should remove
Rrr, leaving resistances Rm + Rir. Chemical Further Reading
cleaning should remove Rir, but usually a residual Fane AG, Chong TH, Le-Clech P (2009) Fouling in mem-
DRir remains as explained in irreversible fouling. brane processes, chapter 6. In: Drioli E, Giorno L (eds)
For constant ux, the DP data can be used to transformations. Wiley VCH, Weinheim. ISBN
estimate the various resistances. 978-3-527-32038-7
For processes like reverse osmosis, the osmotic
pressure term in Eq. 1 is important. The magni-
tude of DP is increased in practice by a factor M,
the polarization modulus; this is due to concentra-
tion polarization. Equation 1 becomes: Isotherm Palladium Phase Diagram
DP MDP Ekain Fernandez
J (4)
mRT Energy and Environment, TECNALIA,
Donostia-San Sebastian, Gipuzkoa, Spain
M is the ratio of solute concentration at the mem-
brane surface to the solute in the bulk, and
in a well-operated RO desalination process, Isotherm palladium phase diagram represents
without fouling, M = 1.11.2. The polarization the palladiumhydrogen system as a function of
modulus, M, can be increased if there is a cake pressure, temperature, and H/Pd concentration as
layer on the RO membrane surface. The layer shown in Fig. 1.
provides an unstirred region where the solute The isotherm palladium diagram is used to
(salt) concentration builds up. Figure 1 depicts show the nature of one of the biggest problems
concentration polarization for a clean and of pure palladium membranes: hydrogen embrit-
cake fouled RO membrane. The increased polar- tlement. As shown in Fig. 1, the absorption of
ization (Fig. 1b) is cake-enhanced concentra- hydrogen below its critical point of 571 K
tion polarization (CECP), and the increased (298 C) and 2 MPa produces two different phases
osmotic pressure is the cake-enhanced osmotic (a and b), both of which retain the pure palladium
pressure (CEOP) (Hoek and Elimelech 2003). face-centered cubic (fcc) lattice but with the crys-
Both fouling resistance and CEOP contribute to tal unit cell lattice parameter increasing from
loss of ux (for xed DP) or increased DP for 0.3890 nm for hydrogen-free palladium to
xed ux. 0.3895 nm for the a-phase and up to 0.410 nm
Isotope Separation, Membrane for 1073
Isotherm Palladium
Phase Diagram,
Fig. 1 Pressure
concentrationtemperature
phase diagram of the
palladiumhydrogen
system (Lewis 1967)
for the b-phase at room temperature (Lewis 1967; References

Ohira et al. 1996). The a-phase is obtained at low
H/Pd atomic ratios and becomes the dominant Lewis FA (1967) Palladium hydrogen system. Academic,
New York
phase at high temperature. The b-phase is formed
Ohira K, Sakamoto Y, Flanagan TB (1996) Thermody-
at high H/Pd atomic ratios and coexists with the namic properties for solution of hydrogen in Pd-Ag-
a-phase at low temperature (Fig. 1). The hydrogen Ni ternary alloys. J Alloys Compd 236:4249
vapor pressure is constant in the region of phase Paglieri SN, Way JD (2002) Innovations in palladium
membrane research. Sep Purif Rev 31:1169
coexistence and is bounded by an envelope den-
Shu J, Grandjean BPA, Van Neste A, Kaliaguine S (1991)
ing amax and bmin, the compositions of maximum Catalytic palladium-based membrane reactors: a
and minimum H/Pd ratio for pure palladium (Yun review. Can J Chem Eng 69:10361060
and Oyama 2011). The change in volume accom- Yun S, Oyama ST (2011) Correlations in palladium mem-
branes for hydrogen separation: a review. J Membr Sci
panying the phase transformation can give rise to
375:2845
strain and, consequently, after a few cycles to a
metal lattice distortion (Shu et al. 1991), and
micro-failures inside the metallic bulk are
induced. Isotope Separation, Membrane for
The alloying of palladium with other metals,
lowering the coexistence temperature of the two Grayna Zakrzewska-Kotuniewicz
crystalline phases, reduces the internal strength Institute of Nuclear Chemistry and Technology
generated and avoids micro-failure formation and Instytut Chemii i Techniki Jdrowej,
inside the alloy bulk. Silver is the most used Warszawa, Poland
metal for avoiding hydrogen embrittlement phe-
nomenon in Pdalloy membranes. The critical
temperature for the formation of b-phase of The use of membranes for the separation of iso-
PdAg alloy with a composition of 77 and topes is one of the rst practical applications for
23 (%wt) is around room temperature (Paglieri membrane processes. The rst application was
and Way 2002). based on the effects that were described in 1846
1074 Isotope Separation, Membrane for
by Thomas Graham, as a phenomenon of molec- the process was uranium hexauoride, UF6. The
ular effusion, i.e., differentiated ow rate of gases membrane material had to be chemically resistant
through small holes. When a mixture of two gases in this case, since the uranium hexauoride used
in a closed vessel is allowed to permeate through to produce enriched uranium-235 is highly corro-
the porous barrier, assuming an equal concentra- sive (easily converted into hydrouoric acid and
tion of the two gases, the molecules of the lighter harmful oxyuorides as UO2F2). The important
component strike the barrier more frequently than parameters of the molecular effusion barrier are
the molecules of the heavier gas. If the holes of the the pore diameter and the permeation rate since
barrier are sufciently small to avoid convection according to the theory of molecular effusion, the
ow (ow by volume) of gas and large enough to diffusion transport through the membrane is selec-
allow passage of the individual molecules, then tive when the pore size is commensurate with the
the lighter molecules escape more readily through mean free path of the molecules under the given
the pores and the permeate is enriched in the light conditions.
component. The efciency of partitioning is Also the large plants employing the molecular
described by separation factor a expressed as the effusion (which is also called molecular diffusion
ratio of mole fractions of separated components or gaseous diffusion) were built in the United
after and before passing through the porous bar- States with the most famous K-25 plant in Oak
rier or as the square root of the inverse ratio of the Ridge, Tennessee. The other systems were
corresponding molecular weights: constructed in 1950s at Oak Ridge, Paducah, Ken-
tucky, and Portsmouth, Ohio, in the United States,
r Capenhurst in England, and at Pierrelatte and
y1 x M2
a Tricastin in France. There are also some reports
x1 y M1 on gaseous diffusion plants in the former USSR,
Great Britain, Argentina, and in the Peoples
where x and y are mole fractions of two separated Republic of China (Kirk-Othmer 1997). Today,
isotopes and M1 and M2 are molecular weights of the gaseous diffusion is an important method for
these components of the isotopic mixture. uranium-235 enrichment, despite the competitive-
In 1895 Rayleigh and Ramsey have separated ness of the centrifugal methods. This is used in the
argon and nitrogen using this method, and in 1920 United States and France, including Eurodif, led
Aston employed molecular effusion for enrich- by France, Italy, Spain, and Belgium, which holds
ment of neon-22 isotope. Large improvement of the plants in Tricastin and Pierrelatte (France).
the technology was achieved by Gustav Ludwig For preparation of membranes, different
Hertz who developed the separating cascade, methods are applied: etching with strong acids
which was composed of several diffusion stages. (alloys of gold and silver), sintering and extrusion
They also used of the 24-stage cascade in 1932, of ne metal powders and oxides of metals
which almost completely separated isotopes of (nickel, aluminum oxides), and the anodic oxida-
neon, and by applying of the 48-stage cascade, tion (aluminum oxides). Typical gas diffusion
they have separated isotopes of hydrogen. In 1936 membrane has a pore size of 0.0060.04 mm
Wooldridge, Jenkins, and Smythe have enriched and operates at temperature of 60200 C and at
isotopes of neon, hydrogen, and carbon in meth- pressure of 0.43.3 bar. Support layer of the mem-
ane with 34-stage cascade. The most spectacular brane is a porous material with a pore size of
application of molecular effusion was enrichment 210 mm. Ceramic and metallic microporous
of uranium-235 in natural uranium for military membranes were applied for separation of iso-
purposes within the frames of Manhattan Project topes, including other than uranium, such as iso-
during the Second World War. The all of the topes of argon, neon, and hydrogen. Although
mentioned processes of separation of isotopes information about the materials used for
were carried out in the gas phase. In case of manufacturing gaseous diffusion membranes are
uranium isotope separation, the gas applied in very scarce, it is known that French technology
Isotope Separation, Membrane for 1075
used sintered anode and aluminum, alloys of gold used in the hybrid process, were developed based
and silver, nickel, Teon, porcelain (Burggraaf on DTPA (see also Chitry et al. 2001).
and Cot 1996), and zirconium (Hsieh 1996).
Attempts of using the polymeric membranes
for the separation of isotopes of hydrogen with
References
membranes made of polymers, such as polyethyl-
ene terephthalate (PET), polyethylene (PE), Burggraaf AJ, Cot L (1996) Membrane science and tech-
polytetrauoroethylene (PTFE), cellulose acetate nology series, vol 4, Fundamentals of inorganic mem-
(CA), and plasticized polyvinyl chloride (PVC), brane science and technology. Elsevier, Amsterdam/
Lausanne/New York/Oxford/Shannon/Tokyo
were taken by Marcea (1983). Although the selec-
Campbell DJ (1985) Fractionation of stable chlorine iso-
tivity of polymers is lower than metallic mem- topes during transport through semi-permeable mem-
branes, it is worthy to point out, since the branes. M.Sc. thesis, University of Arizona
numerous benets of their use, such as the Chitry F, Pellet-Rostaing S, Vigneau O, Lemaire M (2001)
Nanoltration-complexation: a new method for isoto-
low-cost of membranes, the possibility of forma-
pic separation of heavy metals. Chem Lett
tion of capillaries resulting in large areas for per- 30(8):770771
meation, and no need of high pressures and Chmielewski AG, Zakrzewska-Trznadel G, Miljevi N,
temperatures. Van Hook WA (1991) 16O/18O and H/D separation
factors for liquid/vapour permeation of water through I
The studies on separation of hydrogen and
an hydrophobic membrane. J Membr Sci 60:319329
oxygen in the liquid phase (water) were carried Fritz SJ, Hinz DL, Grossman EL (1987) Hyperltration-
out in Pacic Northwest Laboratory, USA induced fractionation of carbon isotopes. Geochim
(Nelson et al. 1996). Polymeric membranes were Cosmochim Acta 51:11211134
Hsieh HP (1996) Membrane science and technology series,
used in Institute of Nuclear Chemistry and Tech-
vol 3, Inorganic membranes for separation and reac-
nology, (Poland). They were made of regenerated tion. Elsevier, Amsterdam/Lausanne/New York/
cellulose, cellulose acetate (Chmielewski et al. Oxford/Shannon/Tokyo
1991), and polytetrauoroethylene (Zakrzewska- Kim J, Park SE, Kim TS, Jeong DY, Ko KH (2004) Isotopic
water separation using AGMD and VEMD. Nuklonika
Trznadel et al. 1996). The economic assessment
49(4):137142
showed that membrane permeation is competitive Kirk-Othmer A (1997) Diffusion separation methods. In:
with other methods of hydrogen and oxygen Ruthven DM (ed) Encyclopedia of separation technol-
isotopes separation. The demand for heavy oxy- ogy, vol 1. Wiley, New York/Chichester/Weinheim/
Brisbane/Singapore/Toronto
gen water, due to the development of PET diag-
Marcea P (1983) Permeation of H2 and D2 through poly-
nostic method, has intensied the interest in mers. Isotopenpraxis 19(5):153156
new processes of separating isotopes of oxygen, Nelson DA, Duncan JB, Jensen GA, Burton SD
among others membrane method (Kim (1996) Isotopomeric water separations with supported
polyphosphazene membranes. J Membr Sci
et al. 2004).
112:105113
The following isotopes of chlorine (Campbell Okamoto M, Oi T, Fujii Y, Kakihana H (1980) Dependen-
1985), carbon (Fritz et al. 1987), lithium cies of pH and current density on the isotope effect
(Whitworth et al. 1994) in aqueous solutions, occured by the electromigration of uranyl ions in a cation
exchange membrane. Isotopenpraxis 16(2):5861
and uranium (Okamoto et al. 1980) in ammonia
Whitworth TM, Marias BJ, Fritz SJ (1994) Isotopic frac-
were separated by means of various membrane tionation and overall permeation of lithium by a thin-
techniques and by using of different membranes. lm composite polyamide reverse osmosis membrane.
Gadolinium and neodymium isotopes were sepa- J Membr Sci 88:231241
Zakrzewska-Trznadel G, Chmielewski AG, Miljevi N
rated by means of a hybrid complexation/
(1996) Separation of protium/deuterium and oxygen-
nanoltration (Chitry et al. 2001) processes. The 16/oxygen-18 by membrane distillation process.
new complexing agents and ligands, which were J Membr Sci 113:337342
K
Keggin Structure heteroatoms are phosphorus (P) and silicon (Si);

nevertheless, some other metals such as V, Nb, Ta,
Ubavka B. Mio and Zoran P. Nedi and U also show comparable behavior, but to a
Faculty of Physical Chemistry, University of more limited extent. Bands in IR and Raman
Belgrade, Republic of Serbia spectra, characteristic for host lattice, and used
for quick and ease identication of POMs, are
assigned to PO43 group (at about 1,003,
Heteropoly compounds (HPCs) have an almost 988, and 983) -n1 stretching and at 516 cm1 are
two centuries-long history. Since 1826, when the assigned to PO43 asymmetric bending vibra-
rst polyoxometalate (POM), (NH4)3PMo12O40, tions. Some other bends are assigned to groups
has been synthesized (Berzelius 1826), they are of addenda atoms: about 805 to WOW asym-
well known as analytical reagents, catalysts, and metric stretching, at 536 to WOW symmetric
biochemical and biomedical active compounds stretching, and 273 cm1 WOW bending vibra-
(Rhule et al. 1998; Kunti et al. 2006). The best tions. Below 450 cm1 lattice modes are expected
known group of POMs are those where hetero- (Rocchiccioli-Deltcheff et al. 1983; Mio
atoms are surrounded by a cage of transition et al. 1994). Keggin anions form channels
metals (M=Mo and W), as a result of Mo and full with water and hydrated protons (cf.
W acids polymerization. Two structure types pre- Heteropoly Acids).
dominate in this group of compounds: Keggin POMs of Mo and W are of two types,
ions [Xn+M12O40](8n) synthesized in 1934 isopolyoxometalate and heteropolyoxometalate,
(Keggin 1934) and Dawson or Wells-Dawson which contain one or two atoms of another ele-
ions (Xn+)2M18O62(162n), with structure deter- ment (heteroatom) besides molybdenum or tung-
mined in 1953 (Pope 1983) (Fig. 1). It is evident sten, oxygen, and hydrogen. More than
from Fig. 1a how Keggin anion is formed. On 65 elements from all groups of the periodic table
each corner of central tetrahedron, three WO6 can be found incorporated into polyoxoanions as
octahedra are bonded sharing corners and sides. heteroatom. Covalent attachment of amino acids
Thus, such group could be presented as an M3O13 as glycine, alanine, and lysine to POMs is also
group. It is possible to distinguish four oxygen possible. Such compounds are interesting espe-
atoms in such structure: O1 (terminal), O2 (edge cially for the investigation of biochemical and
bridging), O3 (corner bridging), and O4 atoms biomedical activity of HPCs (Rhule et al. 1998;
shared with central tetrahedron. Typically, Kunti et al. 2006).

DOI 10.1007/978-3-662-44324-8
1078 Keggin Structure
Keggin Structure, Fig. 1 Three predominate structure of POMs: (a) group of three WO6 octahedra around central PO4
tetrahedron of Keggin anion; (b) primary structure of Keggin anion; (c) Dawsons structure
After the discovery of their superionic proton References

conductivity (s = (1100) 103 S/cm) at
room temperature in 1979 by Nakamura Berzelius JJ (1826) The preparation of the
phosphomolybdate ion [PMo12O40]3-. Pogg Ann
et al. (1979, 1981), the chemical and electrochem-
6:369371
ical conductive properties of these selected Colomban P (ed) (1992) Proton conductors, solids, mem-
compounds have become very interesting as branes and gels. Cambridge University Press,
new materials, potentially useful for different Cambridge
Holclajtner-Antunovi I, Mio UB, Todorovi M,
solid-state devices: as sensors, proton exchange
Jovanovi Z, Davidovi M, Bajuk-Bogdanovi
membranes (PEM) in H2/O2 fuel cells, D (2010) Characterization of potassium salts of
electrochromic displays, and ion-selective mem- 12-tungstophosphoric acid. Mater Res Bull
branes (Colomban 1992). Proton conductivity 45:16791684
Keggin JF (1934) Structure and formula of
(not lower than 1 104 S/cm) is necessary to
12-phosphotungstic acid. Proc R Soc A144:75100
enable the use POMs as PEM in fuel cells, or in Kunti V, Stanojevi M, Holclajtner-Antunovi I,
batteries (Mio et al. 2005; Holclajtner-Antunovi Uskokovi-Markovi S, Mio U, Todorovi M,
et al. 2010; Mentus et al. 2010). But regretfully Jovanovi T, Vukojevi V (2006) Synthesis, character-
ization, and biological activity of amino acid deriva-
they also have some deciencies. They are
tives of the heteropoly tungstophosphoric acid.
sensitive to ambient conditions: relative Monatsh Chem/Chem Mon 137:803810
humidity (RH) and temperature. Also, they Mentus SV, Gavrilov N, Nedi Z, Mio UB (2010) Appli-
have low specic surface area (2030 m2/g) and cation of Li doped phosphate bronzes as anode mate-
rials in Li-ion batteries. In: Mio UB, Davidovi
are washable from PEM. These deciencies
M (eds) Fast proton-ion transport compounds.
may be avoided if HPCs are incorporated in Transworld Research Network, Kerala, pp 179192
composites, and/or polymers, and also if POMs Mio UB, Dimitrijevi R, Davidovi M, Nedi ZP,
are used as acid insoluble salts of heteropoly Mitrovi MM, Colomban P (1994) Thermally induced
phase transformations of 12-tungstophosphoric acid
acids, or some other compounds, as bronzes,
29-hydrate: synthesis and characterization of
where HPCs have been used as precursors PW8O26-type bronzes. J Mater Sci 29:37053718
(Mentus et al. 2010). Mio UB, Todorovi MR, Davidovi M, Colomban P,
The applications of POMs are associated with Holclajtner-Antunovi I (2005) Heteropoly
compounds-from proton conductors to biomedical
their size, high ionic weight and redox character-
agents. Solid State Ion 176:30053017
istics, polarity, surface charge distribution, elec- Nakamura O, Kodama T, Ogino I, Miyaka Y (1979) High-
tron and proton transfer/storage ability, the conductivity solid proton conductors: dodecamolybdo-
formation of high Brnsted acid centers, and phosphoric acid and dodecatungstophosphoric acid
crystals. Chem Lett 8:1718
so on.
Kinetic Study in Bulk Liquid Membranes 1079
Nakamura O, Ogino I, Kodama T (1981) Temperature and mechanisms by using different types of carriers
humidity ranges of some hydrates of high-proton- and organic diluents in the membrane phase and
conductive dodecamolybdophosphoric acid and
dodecatungstophosphoric acid crystals under an atmo- different stripping agents in the product phase.
sphere of hydrogen or either oxygen or air. Solid State Dimensionless reduced concentrations of the tar-
Ion 34:347351 get species in the feed (Rf), membrane (Rm,), and
Pope MT (1983) Heteropoly and isopoly oxometalates. product phases (Rp) were used (Rf=Cft/C0;
Springer, Berlin
Rhule JT, Hill CI, Judd DA (1998) Polyoxometalates in Rm=Cmt/C0; Rp=Cpt/C0; the sum Rf +Rm+Rp
medicine. Chem Rev 98:327358 obviously being unity).
Rocchiccioli-Deltcheff C, Fournier M, Franck R, In most of these studies, the variation with time
Thouvenot R (1983) Vibrational investigations of of the reduced concentrations of target species
polyoxometalates. 2. Evidence for anion-anion interac-
tions in molybdenum(VI) and tungsten(VI) compounds decreased monoexponentially in the feed phase,
related to the Keggin structure. Inorg Chem followed an increasing sigmoided-type curve in
22:207216 the product phase and presented a maximum in
the membrane phase (Fig. 1).
All these results suggest that target species
transport through bulk liquid membranes by
Kinetic Study in Bulk Liquid using a carried mediated transport obeys the
Membranes kinetic laws of two consecutive irreversible rst-
order reactions, i.e., the extraction (rate constant,
Gerardo Len kex) and the stripping (rate constant, kst) reactions
Departamento de Ingeniera Qumica y (Szpakowska and Nagy 1991; He et al. 2000; K
Ambiental, Universidad Politcnica de Cartagena, Len and Guzmn 2005). The irreversibility is to
Cartagena, Spain be expected because Rf and Rm tend to be zero,
while Rp tends to be one, that is, the transport from
the feed phase to the product phase seems to be
Kinetic study of the transport of different chemi- virtually complete. This kinetic behavior can be
cal species through bulk liquid membranes has mathematically described according to the follow-
been widely studied, including metallic ions ing equations:
(Cu+2, Co+2, Cr+6, Hg+2, Na+, Pd+2, U+6), anionic
species (CN, SCN, Cl), dyes (methylene blue, dRf dRm
kex Rf Jf=m kex Rf kst Rm
cibacron red FN-R), amino acids (valine, phenyl- dt dt
dRp
alanine), and other compounds. Most of these kst Rm Jm=p
studies have employed carrier-mediated transport dt
Kinetic Study in Bulk 1

Liquid Membranes, Rf Rm Rp
Fig. 1 Typical curves of
time dependence of Rf in 0,8
feed phase, Rm in
membrane phase, and Rp in
0,6
the product phase, in the
transport of a chemical R
specie through a bulk liquid 0,4
membrane by using a carrier
facilitated transport
0,2
0
0 4 8 12 16 20 24
Time (hours)
1080 Knudsen Diffusion
Integration of those differential equations gives ratio was lower than 1, the process followed the
Rf expkex t kinetic laws of two consecutive irreversible rst-
order reactions, with reduced concentrations of
kex
Rm expkex t expkst t Cu(II) decreasing monoexponentially in the feed
kst kex phase, following an increasing sigmoided-type
1
Rp 1 kst expkex t curve in the product phase and presenting a max-
kst kex imum in the membrane phase. Between these two
kex expkst t ratio ranges, the kinetics was of intermediate char-
acter. These results show the possibility of chang-
These equations show that the time dependence of ing the transport kinetics by varying the [Target
Rf is monoexponential and the time dependence of specie]/[Carrier] ratio.
both Rm and Rp is biexponential. kst
kex kst
Rm has a maximum Rm max kkexst at

kex Cross-References
ln
k
time tmax kex kstst Carrier-Mediated Transport
Maximum uxes can also be obtained:
k kexk References

dRf kex ex st
kex Jmax
f
dt max kst He D, Ma M, Zhao Z (2000) Transport of cadmium ions
through a liquid membrane containing amine
dRm
0 extractants as carriers. J Membr Sci 169:5359
dt max Len G, Guzmn MA (2005) Kinetic study of the effect of
k kk st carrier and stripping agent concentrations on the facil-
dRp kex ex st
kst Jmax
p
itated transport of cobalt through bulk liquid mem-
dt max kst branes. Desalination 184:7987
Szpakowska M, Nagy OB (1991) Membrane material
effect on copper coupled transport through liquid mem-
dRf dRp
) Jmax
f Jmax
p
dt max dt max Szpakowska M, Nagy OB (2000) Inuence of physico-
chemical control and of membrane composition on
Moreover, in many cases, the activation energy the coupled transport of copper(II) ions through bulk
liquid membranes. J Membr Sci 168:183186
values of the processes were obtained from the
Arrhenius equation, lnk lnA RT Ea ,
by using
the kex and kst values at different temperatures.
Notwithstanding the previous comments,
Knudsen Diffusion
kinetic studies on Cu(II) transport through bulk
liquid membrane containing 2-hydroxy-5-nonyl-
Renzo Di Felice
benzaldoxyme (Acorga P-50) as carrier in
n-octane and sulphuric acid as stripping agent
(Szpakowska and Nagy 2000) described that the
[Cu(II)]/[Carrier] ratio inuences the kinetic
Synonyms
regime. It was found in these studies that for
[Cu(II)]/[Carrier] ratio higher than 5, both the
Knudsen transport modeling
extraction and the stripping steps were of zeroth
order with respect to copper, with remaining cop- Whenever a species is diffusing inside a conned
per concentration in the membrane small and space, such as cylindrical pore, an interaction
practically constant. When [Cu(II)]/[Carrier] between the diffusive molecules and the pore
Knudsen Transport Modeling 1081
wall is to be expected. One possibility is given by where d is the pore diameter in cm, T the absolute
the pore wall physically interacting with the dif- temperature, and M the molecular weight of the
fusive molecules and hindering, as a consequence, diffusive species. Unlike diffusion in unbounded
their movement. The phenomenon, called Knud- systems, Knudsen diffusion coefcient is inde-
sen diffusion, is relevant whenever the distance pendent of pressure and of the molecular weight
between molecular collisions is greater than the of the second species.
pore diameter. This interaction can be the
governing factor of the diffusion rates.
Liquid have a relatively small mean-free path, Cross-References
few angstroms, so that Knudsen diffusion in not
important in practical cases, but on the other hand, Principle of Gas Separation
for gas this parameter can assume values larger
than 1,000 and therefore be signicant.
The diffusion coefcient, DKn, in the Knudsen
regime can be estimated from gas kinetic theory,
Knudsen Transport Modeling
and it is given by
r Knudsen Diffusion
T
DKn 4850d
M
K
L
Lamellar Copolymers There are numerous methods for inducing the

orientation and order of large-scale morphologies
Mario Malinconico of the copolymers blocks, such as lamellae. Some
Research Director, Institute for Polymers, main strategies used are reported:
Composites and Biomaterials (IPCB-CNR),
Pozzuoli, Naples, Italy Fields of mechanical ow (extrusion (Keller
et al. 1970; Folkes et al. 1973), compression
(Konas and Cohen 1995; Drzal et al. 2001;
The presence of covalent bonds between the Quiram et al. 1998; van Asselen et al. 2004),
blocks in block copolymers leads to severe restric- ows involving oscillating (Skoulios 1977;
tions on the state of local segregation between Hadziioannou et al. 1979, p. 15, 1979, p. 136;
them allowing the formation of few and dened Morrison and Winter 1989; Morrison
morphologies. Comparing the phase separation et al. 1990; Wiesner 1997; Scott Pinheiro and
that occurs in a pure copolymer AB and into the Winey 1998; Leist et al. 1999; Hermel
corresponding mixture of homopolymers HA/HB et al. 2002; Stangler and Abetz 2003; Wu
and assuming that in both cases strips of different et al. 2004) or stationary (Sebastian
phases are formed, parallel to the surface of the et al. 2002;;Angelescu et al. 2004, 2005; Luo
system, it is observed that in the mixture the size and Yang 2004) shear, or techniques that com-
of the homopolymeric lamellae parallel to the bine different ow elds (Albalak and Thomas
surface and perpendicular to it are both much 1993, 1994; Honeker et al. 2000; Dair
greater than the radius of gyration (Rg) of the et al. 2000; Villar et al. 2002)) were success-
polymer. Instead, in the case of the block copoly- fully used to induce alignment in block
mer, the presence of the bond between the blocks copolymers.
allows the formation of lamellae whose dimen- Magnetic elds induce orientation in liquid
sions parallel to the surface are much greater than crystalline block copolymers (Osuji
the radius of gyration, but the size perpendicular et al. 2004; Tomikawa et al. 2005) and block
to the surface has dimensions comparable to the copolymers with a crystallizable block through
radius of gyration. On the basis of this different an accurate control of the crystallization pro-
structural feature, the domains of different com- cess (Grigorova et al. 2005).
positions which are formed in a mixture of homo- Static electric elds have been widely used to
polymers are named macrophases, while in the guide macroscopically cylindrical and lamellar
case of a copolymer are spoken of microphases morphologies in the melt of such copolymers
or even, sometimes, of nanophases (Fig. 1). (Morkved et al. 1996; Onuki and Fukuda 1995;
DOI 10.1007/978-3-662-44324-8
1084 Lamellar Copolymers
Lamellar Copolymers,
Fig. 1 Phase separation in
block copolymers (a) and in
homopolymers blend (b)
(Schulz and Bates 1996)
Thurn-Albrecht et al. 2000, 2002; Elhadj order spreads rapidly from the surface of the
et al. 2003; Xu et al. 2004; DeRouchey lm to the substrate. The consequent decrease
et al. 2004; Xiang et al. 2004). of Tg below room temperature, for at least one
Solvent evaporation allows the orientation of block, freezes the structures that, due to the
lamellar microdomains perpendicularly to the high directionality of the solvent gradient,
lm surface (Turturro et al. 1995; Kim and organize into domains with lamellar axes per-
Libera 1998, p. 2569, 1998, p. 2670). After pendicular to the surface (Kim et al. 2004;
the deposition, the glass transition temperature Kimura et al. 2003; Lin et al. 2002; Temple
(Tg) of the lm swollen with a solvent is below et al. 2003).
the ambient temperature, allowing free mobil-
ity of the chains. With increasing concentration A great potential of the block copolymers is
of the solvent, the copolymer undergoes a tran- represented by the possibility of using the ordered
sition from the disordered state to the ordered. nanostructures formed by self-assembly as a
The diffusion of the solvent in the thin lm matrix (host) for the inclusion of guest molecules
produces a concentration gradient so that (guest) and dispersion of different types of
Lamellar Copolymers 1085
Lamellar Copolymers,
Fig. 2 Bright-eld TEM
image of a thin lm of a
lamellar block copolymer
polystyrene-poly(ethylene-
co-propylene) in which
gold nanoparticles suitably
passivated are included
only in the layers of
polystyrene (Bockstaller
et al. 2005)
nanoparticles, in order to obtain nanocomposites Albalak RJ, Thomas EL (1994) Roll casting of BCPs and
with special physical properties. The nano- BCP-homopolymer blends. J Polym Sci Polym Phys
32:341
composites represent a class of composite mate- Angelescu DE, Waller JH, Adamson DH, Deshpande P,
rials consisting of matrix and a polymeric particle Chou SY, Register RA, Chaikin PM (2004) Macro-
reinforcement having at least one dimension of scopic orientation of block copolymer cylinders in sin-
the order of nano; such particle reinforcements are gle-layer lms by shearing. Adv Mater 16:1736
Angelescu DE, Waller JH, Register RA, Chaikin PM
called nanollers or nanoreinforcement. (2005) Shear-Induced alignment in thin lms of spher-
The different microdomain nanostructures ical nanodomains. Adv Mater 17:1878
generated by block copolymers (lamellae in par- Bockstaller M, Mickievic RA, Thomas EL (2005) Block
ticular, but also spheres or cylinders) act as host to copolymer nanocomposites: perspectives for tailored
functional materials. Adv Mater 17:1331
sequester selectively nanollers (guest) with a Dair BJ, Avgeropoulos A, Hadjichristidis N, Capel M,
L
suitable chemical and geometrical afnity. Thomas EL (2000) Oriented double gyroid lms via
The great innovation of these studies consists roll casting. Polymer 41:6231
in the fact that the use as a matrix of block copol- DeRouchey J, Thurn-Albrecht T, Russell TP, Kolb
R (2004) Block copolymer domain reorientation in an
ymers, then nanostructured materials already, electric eld: An in-situ small-angle X-ray scattering
instead of homopolymers, offers an important study. Macromolecules 37:2538
opportunity to control the distribution of spatial Drzal PL, Barnes JD, Konas P (2001) Path dependent
and orientational nanollers. The nanoparticles, microstructure orientation during strain compression of
semicrystalline block copolymers. Polymer 42:5633
which induce specic properties, are not, in fact, Elhadj S, Woody JW, Niu VS, Saraf RF (2003) Orientation
randomly distributed in the polymer matrix but of self-assembled block copolymer cylinders perpen-
are sequestered in ordered microdomains and, dicular to electric eld in mesoscale lm. Appl Phys
thus, distributed in an orderly manner in the Lett 82:871
Folkes MJ, Keller A, Scalisi FP (1973) Extrusion tech-
matrix. This allows a greater control of the nal nique for the preparation of single crystals of block
physical properties of the nanocomposites. copolymers. Colloid Polym Sci 251:1
A TEM image of a thin lm of a nanocomposite Grigorova T, Pispas S, Hadjichristidis N, Thurn-Albrecht
based on a lamellar block copolymer in which T (2005) Magnetic eld induced orientation in diblock
copolymers with one crystallizable block. Macromole-
gold nanoparticles are distributed in such an cules 38:7430
ordered way only in specic layers of the structure Hadziioannou G, Mathis A, Skoulios A (1979)
is shown in Fig. 2 (Bockstaller et al. 2005). Monocristaux de copolymres trisquencs styrne/
isoprne/styrne prsentant la structure cylindrique: I.
tude de l'orientation par diffraction des rayons X aux
petits angles. Colloid Polym Sci 257:15, 136
Hermel TJ, Wu LF, Hahn SF, Lodge TP, Bates FS (2002)
References Shear-induced lamellae alignment in matched triblock
and pentablock copolymers. Macromolecules 35:4685
Albalak RJ, Thomas EL (1993) Microphase separation Honeker CC, Thomas EL, Albalak RJ, Hajduk DA, Gruner
of BCP solutions in a ow eld. J Polym Sci Polym SM, Capel MC (2000) Perpendicular deformation of a
Phys 31:37 near-single-crystal triblock copolymer with a
1086 Lamellar Copolymers
cylindrical morphology. 1. Synchrotron SAXS. Macro- copolymer systems correlated to the linear viscoelastic
molecules 33:9395 properties of the parallel and perpendicular morphol-
Keller A, Pedemonte E, Willmouth FM (1970) Macro- ogies. Macromolecules 31:4447
lattice from Segregated Amorphous Phases of a Three Sebastian JM, Graessley WW, Register RA (2002)
Block Copolymer. Nature (London) 225:538 Steady-shear rheology of block copolymer melts and
Kim G, Libera M (1998) Morphological development in concentrated solutions: Defect-mediated ow at low
solvent-cast polystyrenepolybutadienepolystyrene stresses in body-centered-cubic systems. J Rheol
(SBS) triblock copolymer thin lms. Macromolecules 46:863
31:2569, 2670 Skoulios A (1977) Properties of oriented block copoly-
Kim SH, Misner MJ, Xu T, Kimura M, Russell TP (2004) mers. J Polym Sci Polym Symp 58:369
Highly oriented and ordered arrays from block copoly- Stangler S, Abetz V (2003) Orientation behavior of AB and
mers via solvent evaporation. Adv Mater 16:226 ABC block copolymers under large amplitude oscilla-
Kimura M, Misner MJ, Xu T, Kim SH, Russell TP (2003) tory shear ow. Rheol Acta 42:569
Long-Range ordering of diblock copolymers induced Temple K, Kulbaba K, Power-Billard KN, Manners I,
by droplet pinning. Langmuir 19:9910 Leach KA, Xu T, Russell TP, Hawker CJ (2003)
Konas P, Cohen RE (1995) Melt processing of semicrystal- Spontaneous vertical ordering and pyrolytic
line E/EP/E triblock copolymers near the order-disorder formation of nanoscopic ceramic patterns from poly
transition. Macromolecules 28:336l (styrene-b-ferrocenylsilane) Advanced Materials. 15:
Leist H, Maring D, Thurn-Albrecht T, Wiesner U (1999) 297-300
Double ip of orientation for a lamellar diblock copol- Thurn-Albrecht T, Steiner R, DeRouchey J, Stafford CM,
ymer under shear. J Chem Phys 110:8225 Huang E, Bal M, Tuominen M, Hawker CJ, Russell TP
Lin Z, Kim DH, Wu X, Boosahda L, Stone D, LaRose L, (2000) Nanoscopic templates from oriented block
Russell TP (2002) A rapid route to arrays of copolymer lms. Adv Mater 12:787
nanostructures in thin lms. Adv Mater 14:1373 Thurn-Albrecht T, DeRouchey J, Russell TP, Kolb
Luo KF, Yang YL (2004) Orientation phase transition in R (2002) Pathways toward electric eld induced
the hexagonal phase and rheological properties of alignment of block copolymers. Macromolecules
diblock copolymer under a simple shear ow. Polymer 35:8106
45:6745 Tomikawa N, Lu ZB, Itoh T, Imrie CT, Adachi M,
Morkved TL, Lu M, Urbas AM, Ehrichs EE, Jaeger HM, Tokita M, Watanabe J (2005) Orientation of
Mansky P, Russell TP (1996) Local control of microphase-segregated cylinders in liquid crystalline
microdomain orientation in diblock copolymer thin diblock copolymer by magnetic eld. Jpn J Appl Phys
lms with electric elds. Science 273:931 2 44:L711
Morrison FA, Winter HH (1989) Effect of unidirectional Turturro A, Gattiglia E, Vacca P, Viola GT (1995)
shear on the structure of triblock copolymers. 1. Poly- Free surface morphology of block copolymers: 1.
styrene-polybutadiene-polystyrene. Macromolecules Styrene-butadiene diblock copolymers. Polymer
22:3533 21:3987
Morrison FA, Winter HH, Gronski W, Barnes JD (1990) van Asselen OLJ, van Casteren IA,
Effect of unidirectional shear on the structure of Goossens JGP, Meijer HEH (2004) Deformation
triblock copolymers. 2. Polystyrene-polyisoprene- behavior of triblock copolymers based on
polystyrene. Macromolecules 23:7200 polystyrene: an FT-IR spectroscopy study. Macromol
Onuki A, Fukuda J (1995) Electric eld effects and form Symp 205:85
birefringence in diblock copolymers. Macromolecules Villar MA, Rueda DR, Ania F, Thomas EL (2002) Study of
28:8788 oriented block copolymers lms obtained by roll-cast-
Osuji C, Ferreira PJ, Mao G, Ober CK, Vander Sande JB, ing. Polymer 43:5139
Thomas EL (2004) Alignment of self-assembled Wiesner U (1997) Lamellar diblock copolymer under large
hierarchical microstructure in liquid crystalline diblock amplitude oscillatory shear ow: order and dynamics.
copolymers using high magnetic elds. Macromol Chem Phys 198:3319
Macromolecules 37:9903 Wu L, Lodge TP, Bates FS (2004) SANS determination of
Quiram DJ, Register RA, Marchand GR, Adamson DH chain conformation in perpendicular-aligned
(1998) Chain orientation in block copolymers undecablock copolymer lamellae. Macromolecules
exhibiting cylindrically conned crystallization. 37:8184
Macromolecules 31:4891 Xiang H, Lin Y, Russell TP, Kolb R (2004) Electrically
Schulz MF and Bates FS (1996) In Physical properties of induced patterning in block copolymer lms. Macro-
polymers handbook, American Institute of Physics. pp molecules 37:5358
427433 Xu T, Zhu Y, Gido SP, Russell TP (2004) Electric eld
Scott Pinheiro B, Winey KI (1998) Mixed alignment of symmetric diblock copolymer thin lms.
parallelperpendicular morphologies in diblock Macromolecules 37:2625
Langerhans Islet 1087
insulin-producing beta cells by autoimmune

Langerhans Islet responses. Thus, sustaining and/or replacing
malfunctioning islets is the only therapeutic
Loredana De Bartolo and Antonietta Messina approach in order to induce the partial recovery
Institute on Membrane Technology, National of the metabolic functions in the patients (Patel
Research Council of Italy, ITM-CNR, Rende, et al. 2014). Before 1893 and the discovery of the
Italy pancreas involved in the income of the diabetes,
starving was the only treatment possible for peo-
ple suffering for diabetes. Then, rst attempts in
The islets of Langerhans are endocrine cell clus- the development of grafts were made using
ter well ordered in lobules, of 0.30.7 mm in deceased donors islets serving as substitutes for
diameter, containing from 3,000 to 4,000 cells, the malfunctioning pancreas. Many patients
and located mainly in the tail region of the pan- showed a partial recovery; however, after trans-
creatic gland. Discovered in 1869, though plantation, these islet grafts were inevitably
representing only 2 % of the whole organ, they destroyed due to the immune and inammatory
serve to coordinate the physiological control of reaction in the hosts. In truth, despite the risk/
the glycemic values in the blood. Five different benet trade being quite high for the patients, the
hormone-secreting cell types constitute the islets transplantation still remains the only therapeuti-
cytoarchitecture: alpha cells producing glucagon, cally approved approach to gain the patients
beta cells producing insulin and amylin, delta cells recovery of the pancreatic activity. Allo and
producing somatostatin, PP cells (gamma cells) autotransplantation are nowadays performed
producing the pancreatic polypeptide, and epsilon through the infusion of puried islets into the
cells producing ghrelin. In human beings, the portal vein of the liver, and 50 % of the receivers L
alpha and beta cells are in close relationship with are able to regain the pancreatic functionality and
each other. Representing nearly 90 % of the total recover from the systemic complications in 1 year.
islet cells, their secretive activity is modulated by However, the number of available and compatible
the blood glucose levels: when low concentrations donors is quite limited, and the lifetime immuno-
of sugar are detected, the islets respond with the suppression therapy exposes the patient to infec-
glucagon secretion to induce the glucose release tions, diseases, and harmful side effects. Thus,
from the hepatic deposits into the bloodstream. new techniques labeled as tissue engineering
After a meal the high concentration of the blood approaches have been developed to overcome
sugars causes the beta cell activation determining these disadvantages and foresee the use of poly-
the insulin-induced glucose metabolism and the meric membrane systems to produce engineered
production of the energy requested by the body. device containing islets of insulin-producing
Since the interaction between the hormone- cells, to control the release of the cell products
secreting cells of the islets and the paracrine, from the islets, their sustainment, and their pro-
electric, and nervous-regulating factors is highly tection from the immunoreactive species in the
complex, the tiniest decompensation in the islet site of implantation. Coordinated by local drug
cells functions, modulation, and survival can delivery systems, consisting in boles of low-dose
result in the onset of severe pathologies and anti-rejection drugs able to minimize the immune
organic disorders, which leads patients to severe response as well as any inammatory effects at the
complications and a high risk of death. Actually, site of implantation, the encapsulation technique
98 % of the diseases involving the pancreas is the most promising method. The islets are
referred to the loss of the ability in modulating wrapped in a tight coating made of biocompatible
the food sugars conversion into energy, due polymers acting as a perm-selectivity membrane
to the malfunctioning or destruction of the that serves to mimic the heterogeneous and safe
1088 Langerhans Islet
environment where oxygen and nutrient perfusion replacement strategies are available for the
is provided continuously and appropriate space implantation: All-In/Biodegrade-Out replacement,
and ECM-like physical and mechanical support, made of biodegradable components easily
essential for the correct insulin production and absorbed by the body; All-In/All-Out replacement,
secretion and the islets survival, are provided. designed to not react with the host tissues as the
Different classes of encapsulating devices are cytotherapeutics devices, the islet sheet medical
available and classied depending on their size devices, and the beta-O2 device; and Flush/Reload
in macro-, micro-, and nano-devices. Classied as replacement used for the macrodevices designed in
extravascular diffusion devices when placed sub- order to grow into the host to provide a vascular
cutaneously or in the peritoneal cavity and intra- interface (David and Marchetti 2014). New cell
vascular diffusion system, three macrodevices sources. To overcome the shortage of donor organs
have been approved for the clinical trials: the and the side effect linked to the autoimmune
TheraCyte system, made up of two composite destruction of the patient islets as for diabetes
Teon-based membranes in between the islets type I, alternative sources of beta cells have been
are placed; the Islet Sheet device, constituted by researched. The adult stem cells or progenitor cells
supported alginate sheets containing the pancre- appear to be the most promising. Through
atic islets; and the Beta O2 device able to increase reprogramming cell strategy (transdifferentiation),
the oxygen exchange in the implantation site pancreatic precursor cells are reprogrammed in
being made of oxygen-generating biomaterial insulin-producing cells. Moreover, the own
able to deliver locally the oxygen to the cells. patients pancreas exocrine cells and the skin cells
Additional macrodevices consist of tubular that can undergo to the reprogrammation technique
porous membranes lled with islets generally representing a promising new source for the beta
placed into the vessels, where adequate kinetic cells supply, being not needed the immunosuppres-
properties allowed the insulin release straight sion therapy after the implant (Domnguez-
into the blood stream. Micro devices constituted Bendala et al. 2013). Therefore, another potential
by single cluster of beta cells incorporate into source of beta cells may be the xenotransplantation
semipermeable polymeric membranes, and with pig pancreas cells. Interestingly, human and
nano-devices are nowadays under investigation porcine insulin differ only for one amino acid, and
in a few animal models. They differ generally on insulin extracted from porcine pancreas has been
the processing methods used for the encapsulation used for the treatment of patients with diabetes
of the islets and both allow to increase the insulin before the development of recombinant human
release rate and the oxygen exchange into the insulin technology. However, several problems
islets. The micro systems are processed in a spher- need to be overcome for porcine islet transplanta-
ical system of alginate hydrogels coated with tion to become a viable clinical option; however,
poly-ornithine and polylysine or by cross-linking studies in rodents and large animals have
a thin hydrogel sheet onto an islet cluster by the shown great promise that justies cautious opti-
conformal coating methods, whereas the nano- mism for the near future. In the last years, another
devices (1100 mm) consist of islets encapsulated chance to improve the tissue engineering of the
by alternating charged polylysine, polyglutamic pancreas appears to be the prevention of the inam-
acid, and PEG-biotin polymers over the cell mation at the site of the transplant, co-transplanting
aggregate surface through the layer-by-layer mesenchymal stem cells. The MSCs that have been
method. The decreased thickness of the capsules demonstrated offer natural defenses by blocking
obtained by covering the islets with a single sheet the signals of the immune responses, limiting
of molecules increases and improves the insulin inammation, stimulating the blood vessel growth,
release kinetics after implantation and resulting and promoting the long-term function of islets due
as well in higher diffusion of nutrients in the to their trophic effects.
core of the cell clusters. Independently on the Up until now, though the most important pro-
size of the device, three different encapsulated gress in the pancreatic islet transplantation is
Layer-by-Layer (LbL) Method 1089
settled in preclinical trials and only one lead (with both hydrophobic and hydrophilic parts) is
research project has been presented in the last spread over at the air-water interface. The hydro-
years on baboons, 95 % of the patients who philic parts allow the spreading of the
received the capsule implantation resulted in con- molecules while the hydrophobic parts do not let
siderable improvements of the glucose metabo- the molecules dissolve into subphase. Movable
lism with the reverse of the main diabetes side barriers are used to compress the monolayer. The
effects at 1 year and more than 80 % at 5 years presence of the monolayer can be detected easily
after implants opening new hopes in the further by measuring the surface pressure (p), which is
improvement of new strategies for the diabetes the decrease in surface tension, as the monolayer
treatment in the nearest future. is compressed, thus leading to pressure-area
isotherms.
In the 1930s, Katharine Blodgett then an
assistant to Langmuir developed the procedures
References
to transfer the monolayer onto a solid substrate
David SW, Marchetti P (2014) Encapsulated islets for (Blodgett and Langmuir 1937). The deposited
diabetes therapy: history, current progress, and critical lms are now termed Langmuir-Blodgett lms.
issues requiring solution. Adv Drug Deliv Rev Basically, monolayers can be transferred by
6768:3573 immersing (or withdrawing) the substrate into
Domnguez-Bendala J, Pileggi A, Ricordi C (2013) Islet
cell therapy and pancreatic stem cells, Chapter 70. In: (or from) the liquid while the monolayer is
Atala A, Lanza R (eds) Handbook of stem cells, adsorbed homogeneously at the substrate. Thus,
2nd edn. Academic, New York, pp 835853 well-ordered lms with very accurate thickness
Patel T, Salvatori M, Hemal S et al (2014) Pancreatic can be formed.
islets regeneration: the bioengineering approach, L
Chapter 41. In: Orlando G, Lerut JP, Soker S, Stratta
RJ (eds) Regenerative medicine applications in organ
transplantation, 1st edn. Academic, New York,
pp 599607 References
Blodgett KD, Langmuir I (1937) Built-up lms of barium

stearate and their optical properties. Phys Rev
51:09640982
Langmuir I (1917) The constitution and fundamental prop-
erties of solids and liquids. II Liquids. J Am Chem Soc
Langmuir and LangmuirBlodgett 39:18481906
(LB) Films
Osvaldo N. Oliveira Jr and Diogo Volpati

Sao Carlos Institute of Physics IFSC, University
of Sao Paulo USP, Sao Paulo, Brazil Layer-by-Layer (LbL) Method
Saren Qi1 and Chuyang Y. Tang2

1
An ultrathin, resistant layer can be formed at the Singapore Membrane Technology Centre,
air-water interface by spreading a single drop of Nanyang Technological University, Singapore,
oil. This phenomenon was discovered centuries Singapore
2
ago, but it was Irving Langmuir that in the early Department of Civil Engineering, The University
twentieth century established sound theoretical of Hong Kong, Hong Kong, Hong Kong SAR
and experimental basis for their study (Langmuir
1917). The now-called Langmuir monolayers are
formed at air-water interfaces using an experimen- Layer-by-layer (LbL) method (also known as LbL
tal system known as Langmuir trough. Basically assembly, electrostatic self-assembly (ESA), or
a solution containing amphiphilic molecules polyelectrolyte multilayer (PEM) technology) is
1090 Layer-by-Layer (LbL) Method
Layer-by-Layer (LbL)
Method,
Fig. 1 Simplied
schematic of the lm
deposition process using
LbL method
a versatile method to form mutilayers on a solid et al. 2008), and pressure-induced LbL process
support. The self-assembly method is driven by (Zhang et al. 2011). Characteristics of the LbL
the electrostatic force between the precursor mate- lms are known to be highly affected by the
rials. Precursor materials to form mutilayers can preparation procedures and conditions. In addi-
be chosen from polyelectrolytes, biopolymers, tion to the substrate and multilayer materials,
clays, metal complex and their derivates, other main factors affecting LbL lm
nanoparticles, dendrimers, and so on (Bertrand properties include ionic strength of coating solu-
et al. 2000). The advantages of this method are tion, solution pH, coating time, and the use of
the relatively simple process, independence of cross-linkers.
substrate properties (i.e., size, topology, materials,
and so on), and formation of fuzzy structure.
Therefore, LbL method has a potential to be
applied in various applications including forming Cross-References
multicomposite lms, nanoparticle modication,
enzyme process, optic development, sensors, Cross-Linked Layer-by-Layer Membranes
membrane for water treatment, fuel cell, and
so on.
The basic formation mechanism of a typical References
LbL process is shown in the schematic drawing
in Fig. 1. Step (1) shows the absorption of Bertrand P, Jonas A, Laschewsky A, Legras R (2000)
Ultrathin polymer coatings by complexation of poly-
polycation by a negatively charged substrate. electrolytes at interfaces: suitable materials, structure
After washing with deionized water (step (2)), and properties. Macromol Rapid Commun
the substrate was soaked in a polyanion solution 21(7):319348
(step (3)). Then, it is washed with deionized water Chiarelli PA, Johal MS, Casson JL, Roberts JB, Robinson
JM, Wang HL (2001) Controlled fabrication of poly-
(step (4)). These procedures can be repeated to electrolyte multilayer thin lms using spin-assembly.
prepare multilayers on the substrate. Other varia- Adv Mater 13(15):1167
tions to this basic coating method includes spray Deng HY, Xu YY, Zhu BK, Wei XZ, Liu F, Cui ZY
(Schlenoff et al. 2000), spin (Chiarelli et al. 2001), (2008) Polyelectrolyte membranes prepared by
dynamic self-assembly of poly (4-styrenesulfonic
hydrodynamic LbL deposition (Deng et al. 2008), acid-co-maleic acid) sodium salt (PSSMA) for
electric eld-induced LbL process (Zhang nanoltration (I). J Membr Sci 323(1):125133
Layer-by-Layer Self-Assembly Membrane 1091
Schlenoff JB, Dubas ST, Farhat T (2000) Sprayed poly- The LBL growth, structure, and properties of
electrolyte multilayers. Langmuir 16(26):99689969 multilayer assemblies are inuenced by several
Zhang P, Qian J, Yang Y, An Q, Liu X, Gui Z (2008)
Polyelectrolyte layer-by-layer self-assembly enhanced physicochemical parameters such as pH, ionic
by electric eld and their multilayer membranes for strength and electrolyte species, temperature, sol-
separating isopropanol-water mixtures. J Membr Sci vent quality, adsorption time, and the intrinsic
320(12):7377 properties of polyelectrolytes such as charge den-
Zhang G, Dai L, Ji S (2011) Dynamic pressure-driven
covalent assembly of inner skin hollow ber multilayer sity, concentration, architecture, and molecular
membrane. AIChE J 57(10):27462754 weight.
Despite the several methods available for LBL
assembly of multilayer lms, the concept of LBL
assembly based on multiple intermolecular inter-
Layer-by-Layer Self-Assembly actions opens for a great variety of materials with
Membrane well-dened properties in terms of thickness,
compositions, structure, wettability, building,
Emma Piacentini and Lidietta Giorno or charged/uncharged for a wide range of
Institute on Membrane Technology, National applications.
Research Council of Italy, ITM-CNR, Rende, Potential applications of LbL membranes are
Italy in biomedical eld such as drug delivery, gene
transfection, and biosensors (Yoon et al. 2014).
The controlled drug delivery can be achieved using
hydrolytically degradable polymers such as poly
Layer-by-layer (LbL) self-assembly membranes (b-amino ester). Gene transfection is obtained by
are multilayered nanoarchitectures obtained by using plasmid DNA as an anionic layer to form L
depositing oppositely charged polyelectrolytes or LbL lms. Biosensors can be produced by using
molecules presenting mutually interacting bind- charged biomolecules such as glucose oxidase
ing sites (Table 1) (Borges and Mano 2014). The (able to catalyze glucose oxidation with simulta-
LbL method can not only be applied to polymers neous production of hydrogen peroxide) used as
but also on the combinations of polymers with glucose sensor. LbL membranes can also be used in
particles. tissue engineering by introducing functional
The LBL assembly based on electrostatic inter- groups onto the surface of LbL membrane able to
action is the most used and investigated promote cell proliferation and differentiation as
method for multilayered nanoarchitecture con- substrate or by using cells as a template of LbL
struction. In this case, LbL self-assembly mem- systems.
branes are produced by simply repeating the The LbL self-assembly technique has been
alternate immersion of a charged substrate in the explored in the development of electrochemical
corresponding polyionic solution of opposite energy conversion and storage devices from
charge (Fig. 1). To speed up the deposition fuel cells to supercapacitors (Xiang et al. 2012).
process, otherwise time consuming as a conse- Other recent applications include water
quence of the necessity to wait several minutes treatment like water softening, desalination, and
between each step of adsorption and rinsing, recovery of certain ions (Sanyal and Lee 2014).
spray deposition and spin coating have been Advantages of LbL self-assembly technology
demonstrated to be suitable. The method is very include
attractive due to its simplicity and efciency in
lm preparation; however, the multilayer Its simplicity and low cost
assemblies are less stable and robust than the Variability of the applicable materials
ones obtained by hydrogen or covalent bonding. (conventional polyelectrolytes, water-soluble
1092 Layer-by-Layer Self-Assembly Membrane
Layer-by-Layer Self-Assembly Membrane, Table 1 Materials available for Layer-by-layer (LbL) self-assembly
membranes
Attractive Force
Molecules Type Examples
Polyelectrolyte polymers Electrostatic Poly(styrenesulfonate) (PSS)
interaction Poly(allylamine hydrochloride) (PAH)
Poly(L-lysine)/alginate (PLL/ALG)
PLL/hyaluronan (PLL/HA)
PLL/poly(L-glutamic acid) (PLL/ PGA),
PLL/poly(acrylic acid) (PLL/PAA),
PGA/PAH, chitosan/HA (CHT/HA), poly
(ethylenimine)/PAA (PEI/ PAA)
Polymers carrying hydrogen donor and Hydrogen bonding Poly(vinylpyrrolidone) (PVP)
acceptor moieties Poly(vinyl alcohol) (PVA)
Poly(acrylamide) (PAA)
Poly(ethylene oxide) (PEO)
Poly(vinylpyrrolidone) (PVPON)
Poly(methacrylic acid), PMAA
Uncharged molecules Hydrophobic Alkanethiols-proteins
interactions
Molecules having complementary functional Covalent bonding Poly(vinylamine-co-N-vinylisobutyramide)
groups [Poly(VAm-co-NVIBA)] and PAA. Carboxyl
group of
PAA activated by 1-ethyl-3-
(3-(dimethylamino)propyl)-
Carbodiimide hydrochloride (EDC) and
amine groups of poly(vinylamine-co-
N-vinylisobutyramide)
Poly(VAm-co-NVIBA)
Stereoregular polymers carrying sites of Stereocomplexation (It-st)poly(methyl methacrylate) (PMMA)
opposite chirality
Host (e.g., cyclodextrins, cucurbiturils, Host-guest Cyclodextrin (b-CD) and adamantane
calixarenes, pillararenes, crown ethers, interactions
porphyrins) and guest (e.g., ferrocene,
adamantane, azobenzene) molecules.
Nonionic molecules, which Charge-transfer Y[2-(9-carbazolyl)ethyl methacrylate]
present electron-accepting and electron- interaction (PCzEMA, electron-donor polymer) and poly[2-
donating groups, respectively, in the side chains. [(3,5-dinitrobenzoyl)oxy]ethyl methacrylate]
(PDNBMA, electron-acceptor polymer)
Avidin and biotin-labeled polymers or Biologically Biotin protein conjugate and polymerized
Antibody Antigen conjugated molecules specic interaction streptavidin on hydrophobic silica surfaces
or lectin carbohydrate or DNA
hybridization
Metal ions Coordination Poly(copper styrene 4-sulfonate) (PSS(Cu)1/2),
and organic ligand chemistry and PVP
interactions
proteins, charged polysaccharides, as well as Great control over the composition of the lm
charged inorganic substances, including colloi- along the vertical direction by changing the
dal nanoparticles, heteropolyacids, and metal nature of the building blocks that are deposited
nanoparticles) during each step
Precise microstructural control and high Extensive possibility in wide range applications
repeatability Its nontoxic and environment-friendly nature
Leather Industry 1093
1Step: Immersion of the charged 2Step: washing 3Step: Immersion of the charged 4Step: washing
substrate in the polyionic solution substrate in the polyionic solution
- - + - + -
- - + - + -
- 1Step - + 3Step - + -
- 2Step - + 4Step - + -
- - + - + -
- - + - + -
- - + - + -
- - + - + -
Layer-by-Layer Self-Assembly Membrane, Fig. 1 LbL self-assembled membranes buildup via electrostatic
interactions
References exible material, through a series of chemical

and mechanical treatments.
Borges J, Mano JF (2014) Molecular interactions driving The steps in the production of leather between L
the layer-by-layer assembly of multilayers. Chem Rev
curing (a salt treatment to prevent putrefaction
114:88838942
Sanyal O, Lee I (2014) Recent progress in the applications of the collagen) and tanning (a treatment of
of layer-by-layer assembly to the preparation of nano- leather stabilization with vegetable or mineral
structured ion-rejecting water purication membranes. substances) are collectively referred to as
J Nanosci Nanotechnol 14:21782189
beamhouse operations. They include, in order,
Xiang Y, Lu S, Jiang SP (2012) Layer-by-layer
self-assembly in the development of electro- soaking, unhairing/liming, deliming/bating,
chemical energy conversion and storage devices from and pickling. All these processes are realized
fuel cells to supercapacitors. Chem Soc Rev in chemical reactors (tumblers) in which the
41:72917321
skins react with different chemicals (acids,
Yoon H, Dell EJ, Freyer JL, Campos LM, Jang W (2014)
Polymeric supramolecular assemblies based on multi- alkalis, chromium salts, tannins, solvents, sul-
valent ionic interactions for biomedical applications. des, dyes, auxiliaries, etc.) in aqueous solutions
Polymer 55:453e464 (Fig. 1).
Wastewaters coming from different
tanning operations contain high concentrations
of organic and inorganic substances causing
Leather Industry signicant pollution phenomena. Organic
pollutants come from the skins (it is calculated
Alfredo Cassano that the raw skin has 30 % loss of organic
Institute on Membrane Technology, National material during the working cycle), or they
Research Council of Italy, ITM-CNR, Rende, are introduced during the working cycle (e.g.,
Italy tanning). Inorganic pollutants are a residual of
the used chemicals (i.e., chromium salts) that
are not completely xed by the skins owing to
In the leather industry, the animal rawhide is the low efciency of the operations (Cassano
converted into nished leather, a durable and et al. 2001).
1094 Leather Industry, Degreasing
Leather Industry, water,

Fig. 1 Scheme of tumbler skins,
chemicals
treated skins,
exhausted liquor
permeate
water,
surfactants, Retentate
enzymes bath (emulsified fat for fat liquoring)
feed
tank UF
Leather Industry, Degreasing, Fig. 1 Scheme of aqueous degreasing combined with ultraltration process
References allow the penetration of various tanning materials

and dyes. This operation is necessary especially
Cassano A, Molinari R, Romano M, Drioli E (2001) Treat- for sheepskins where the percentage of fat sub-
ment of aqueous efuents of the leather industry by mem-
stances on raw weight is of about 3040 %.
brane processes. A review. J Membr Sci 181:111126
Enzymatic degreasing is a better way of carry-
ing out degreasing than the use of solvents and
detergents. Lipases are much safer and less toxic
Leather Industry, Degreasing to workers and the environment.
Ultraltration (UF) can be used to treat the
Alfredo Cassano exhausted bath from the degreasing operation in
Institute on Membrane Technology, National order to recover surfactants in the permeate stream
Research Council of Italy, ITM-CNR, Rende, which can be recycled to the degreasing step
Italy leading to a reduction in raw material costs. Fat
substances removed from the skins can be con-
centrated in the retentate stream and reused, after
In the degreasing step, fats and grease are physical and chemical treatments, in the fat
removed from the interbrillary spaces with the liquoring step with signicant reduction of the
use of lipases, detergents, or solvents in order to wastewater treatment costs (Koltuniewicz 2010).
Leather Industry, Unhairing-Liming 1095
In another approach the UF process can be salt compounds with dirt (blood, excrement,
combined to an enzymatic degreasing step earth) which are attached to the skins.
(Fig. 1) with a continuous recycling of the perme- The soaking-exhausted efuent-containing
ate stream in the drum (Cassano et al. 1998). excrements, salts, and chemical additives are nor-
The proposed methodology permits to obtain a mally discharged into a water-treatment plant
high removal efciency of fatty substances from (Sharphouse 1983).
the degreasing bath and a reduction of washing UF membranes can be used to concentrate
cycles normally employed to remove the lipidic organic components in the feed tank of the UF
substances from skins and, consequently, of water plant. A clear permeate, enriched in salt compounds,
consumption. Polysulfone membranes with could be reused in the pickling step after adjustment
molecular weight cutoff of 20 kDa, in spiral- of the salt concentration with NaCl (Fig. 1). Prelim-
wound conguration, exhibited rejections toward inary treatments are necessary in order to remove the
chemical oxygen demand (COD) and fat sub- suspended materials; a sedimentation step allows to
stances higher than 97 %. reduce the suspended solids of 90 %; then steel spring
lters (200300 mm net size) could be employed to
References remove large particles avoiding clogging phenomena
of UF membranes (Cassano et al. 2001).
Cassano A, Drioli E, Molinari R (1998) Integration of
ultraltration into unhairing and degreasing operations.
J Soc Leather Technol Chem 82:130135
References
Koltuniewicz A (2010) Integrated membrane operations in
various industrial sectors. In: Drioli E, Giorno L (eds) Cassano A, Molinari R, Romano M, Drioli E (2001) Treat-
Comprehensive membrane science and engineering. ment of aqueous efuents of the leather industry by
Elsevier, Kidlington, pp 109164 membrane processes. A review. J Membr Sci
181:111126
L
Sharphouse JH (1983) Leather technicians handbook.
Leather Producers Association, Northampton
Leather Industry, Soaking
Alfredo Cassano Leather Industry, Unhairing-Liming

Research Council of Italy, ITM-CNR, Rende, Alfredo Cassano
Italy Institute on Membrane Technology, National
Italy
In this operation, raw skins are exposed to water
and chemicals (small quantities of imbibing sub-
stances) in order to hydrate proteins and bers, to The aim of the unhairing-liming step is to remove
solubilize the denatured proteins, and to eliminate from the raw skin all the components which are
Leather Industry, water,

Soaking, imbibing
Fig. 1 Treatment of substances, purging
exhausted efuents of the skins
soaking step by UF process
exhausted
bath
UF
pickling
1096 Leather Industry, Unhairing-Liming
proteic flour
desulfidation
desalination
drying
water,
Ca(OH)2, concentrated
Na2S proteic fractions
skins
exhausted
bath
UF
Permeate
(low M.W. species, sulfide)
preparation of new liming baths
Leather Industry, Unhairing-Liming, Fig. 1 Treatment of exhausted liming efuent by ultraltration
not transformed into leather, such as the super- An innovating system based on the use of
cial epidermis including the hair and the subcuta- enzymes and small quantities of sodium sulde
neous adipose layer. The liming step introduces (1.5 % instead of 10 %) has been proposed and
chemicals such as lime (Ca(OH)2) and sodium tested on industrial scale. In this approach the
sulde (Na2S) or sodium hydrosulde (NaHS) enzymatic unhairing was combined to the contin-
which open the brous structure of the skin. Con- uous cross-ow UF of the bath permitting to keep
sequently, exhausted efuents are highly polluted constant the sulde concentration in the unhairing
for the presence of sulde, amines, organic matter bath due to its permeation through the UF mem-
coming from degradation of hairs and epidermis, brane. The organic components (products of deg-
and high concentration of alkalis. The COD of radation of the keratin and of the interbrillar
exhausted efuents ranges between 20,000 and proteins, fat substances, etc.) are concentrated in
40,000 mg/L of consumed oxygen. the retentate stream. This new approach is safer
The UF treatment of the exhausted unhairing for workers and characterized by a reduced envi-
liquor has been one of the rst membrane ronmental impact because hairs are not dissolved
approaches introduced in the leather industry and can be removed separately (Cassano
(Molinari 1995). It can be exploited to recover et al. 1998).
sulde and solubilized lime together with low
molecular weight proteic substances in the
permeate stream in order to reuse the puried solu- References
tion for the preparation of a new liming bath. High
Cassano A, Drioli E, Molinari R (1998) Integration of
molecular weight proteic components, coming ultraltration into unhairing and degreasing operations.
from chemical degradation of hairs and epidermis, J Soc Leath Technol Chem 82:130135
are concentrated in the retentate stream (Fig. 1). Molinari R (1995) Application of membrane separation
Considering that 6065 % of the initial sulde techniques to the treatment of tanneries wastewaters.
In: Caetano A, De Pinho MN, Drioli E, Muntau H (eds)
remains in the exhausted liquor and 510 % is lost Membrane technology: application to industrial
in the retentate, the quantity of sulde that is wastewater treatment. Kluwer, The Netherlands,
possible to recycle with an UF system is 5560 %. pp 101122
Leather Processing, Chromium Recovery 1097
initial salt content and it is normally sent to a

Leather Processing, Chromium cleaning-up plant. Here chromium salts end up
Recovery into the sludges creating serious problems for
their disposal.
Alfredo Cassano Chromium recovery from tanning exhausted
Institute on Membrane Technology, National baths represents a signicant economical advan-
Research Council of Italy, ITM-CNR, Rende, tage for the leather industry in terms of its reuse
Italy and for the simplication of the depolluting pro-
cess of global efuents.
An integrated process based on a preliminary
ultraltration (UF) of the spent liquor followed by
In the tannage operation tanning agents are used in a nanoltration (NF) treatment of the UF permeate
order to prevent the leather from chemical and has been proposed as a technically viable method
thermal degradation. The most common tanning for recovering chromium salts from spent tanning
agent is the chromium basic sulfate. It enters the liquors (Cassano et al. 1996, 2007). The UF pro-
pores of the skin by a diffusion process to react cess allows a marked reduction of suspended
with the collagen carboxyl groups and to form solids and fat substances. The concentrated chro-
inter- and intramolecular cross-linking which mium solution obtained in the NF process can be
results in physical, chemical, and biological reused for the preparation of new tanning baths.
stability. The NF permeate can be reused in the pickling
The exhausted bath coming from the step because of its high content of chlorides
chromium tannage contains about 30 % of the (Fig. 1).
L
Watar Skins
NaCl
Cr 2(SO4)3
NaHCO 3
UF
Equalization tank
Exhausted permeate
solution
CHROMIUM TANNAGE
Skins
Water
NaCl
NF acids
PICKEL
Leather Processing, Chromium Recovery, Fig. 1 Proposed process scheme for the chromium recovery from
exhausted chromium baths
1098 Leather Processing, Deliming-Bating
References and from the ammonium sulfate used as chemical

auxiliary.
Cassano A, Drioli E, Molinari R, Bertolutti C (1996) Qual- In order to reduce the nitrogen concentration in
ity improvement of recycled chromium in the tanning
the deliming/bating exhausted bath, the replacement
operation by membrane processes. Desalination
108:193203 of ammonium salts by carbon dioxide (CO2) and the
Cassano A, Della Pietra L, Drioli E (2007) Integrated reuse of wastewater and chemicals after membrane
membrane process for the recovery of chromium salts ltration (MF or UF) of the exhausted liquor has
from tannery efuents. Ind Eng Chem Res
been proposed (Gallego-Molina et al. 2013).
46:68256830
MF or UF membranes lead to a remarkable
reduction of COD and fat substances of the
exhausted liquor (Cassano et al. 2001). The per-
meate solution can be reused for the preparation of
Leather Processing, Deliming-Bating
new bating baths or as washing water (Fig. 1),
providing environmental and economic benets
Alfredo Cassano
due to the water consumption reduction and the
reduction in nitrogen and salt discharge.
Italy
Cross-References
The deliming step is carried out to reduce the
excess of liming agents used in the previous Durability of Membrane (Lifetime of
unhairing operation by using acids and/or acidic Membrane)
salts. Since the pH must be slowly lowered,
ammonium salts are commonly used for this pur-
pose. In the bating operation, skins are treated
References
with proteolytic enzymes in order to open the Cassano A, Molinari R, Romano M, Drioli E (2001) Treat-
brous structure of skins, increasing their ment of aqueous efuents of the leather industry by mem-
softness. brane processes. A review. J Membr Sci 181:111126
Generally, deliming and bating operations are Gallego-Molina A, Mendoza-Roca JA, Aguado D,
Galiana-Aleixandre MV (2013) Reducing pollution
performed in the same drum. Wastewaters from from the delimingbating operation in a tannery.
these operations are characterized by high nitro- Wastewater reuse by microltration membranes.
gen content, coming from both the hide structure Chem Eng Res Des 91:369376
water,
acids,
ammonium salts/CO2, concentrate
surfactants, (reject stream to management)
enzymes
exhausted
bath
MF / UF
washing solution
preparation of new bating baths
Leather Processing, Deliming-Bating, Fig. 1 Proposed process scheme for the treatment of exhausted deliming baths
Lipid Coverage of a Regenerated Cellulose Membrane: Effect on Ion Transport 1099
on RC supports was already described

Lipid Coverage of a Regenerated (Benavente et al. 2011).
Cellulose Membrane: Effect on Ion Impedance is an a.c. technique for determining
Transport membrane electrical resistance and capacitance
by measurements performed at different frequen-
Juana Benavente and Virginia Romero cies (Benavente 2009). As was already indicated
Departamento de Fisica Aplicada I, Universidad (Benavente et al. 2011), IS measurements for dry
de Malaga, Facultad de Ciencias, Malaga, Spain LL-modied membranes showed a reduction in
conductivity and dielectric constant when com-
pared with original regenerated cellulose support,
The modication of solid substrates and mem- estimating a thickness of 2 mm for the
branes using biomaterials is an attractive eld of LL. Figure 1a shows a comparison of Z vs fre-
research for developing new devices such as bio- quency plots for hydrated RC and lipid-layer
sensors, structured multilayers, biomembrane RC/LL membranes, where differences associated
mimetics, biomaterial coatings for therapeutical with electrode-membrane interface (f <5 kHz)
applications, etc. (Goodsell 2004; Sackmann and bulk membrane can be observed as an indica-
1996). Particularly, lipid-layer deposition on tion of lipid coverage of RC support and its effect
membrane supports might improve their on the reduction of water (or electrolyte) adsorp-
biocompatibility, but it might also affect mem- tion by the cellulose structure.
brane transport parameters and, consequently, its Membrane potential (Dmeb) is the electrical
functionality. potential at both sides of membranes in contact
Changes in electrical and transport parameters with different electrolyte concentrations, and
for a hydrophilic regenerated cellulose membrane Fig. 1b shows Dmeb values vs NaCl concentrtion L
(RC sample) as a result of lipid surface deposition ratio cv/cc. Two different Dmeb-ln(cv/cv) relation-
(RC/LL sample) were studied by measuring ships are observed depending on the con-
impedance spectroscopy (IS) and membrane centration ratio values: (i) a Donnan-potential
potential. Lipid-layer (glyceryl tristearate as contribution for 0.2 cv/cc 10, which is asso-
main lipid with L-a-phosphatidylcholine and ciated with co-ion (Cl) exclusion due to the weak
sodium taurocholate as surfactants) deposition membrane effective charge (Xef 0.009 M);
104
a b
0
103
DFmeb (mv)
bulk membrane
Z (W)
10 2 10
101
20
100
101 102 103 104 105 106 2 1 0 1 2
f(Hz) In(Cv/Cf)
Lipid Coverage of a Regenerated Cellulose Membrane: Effect on Ion Transport, Fig. 1 (a) Impedance plot. (b)
Membrane potential dependence on NaCl concentration ratio. RC (), RC/LL (), and FTC (D)
1100 Lipid Monolayer/Bilayer
(ii) when 2> cv/cc >10, the higher contribution

seems to correspond to the diffusion potentials as Lipid Monolayer/Bilayer
a result of the different transport of counterions
(t+) and co-ions (t) in the membrane. Lipid layer Amphiphilic Membrane
does not seem to affect membrane xed charge,
according to the similar Dmeb values obtained
with RC and RC/LL membranes for the Donnan-
potential branch (i), but differences in ion trans- Lipids/Phospholipids
port number should exist due to their different
diffusion-potential contribution (<t+RC > = Amphiphilic Molecules
0.74, <t+RC/LL > = 0.69) (ii). For comparison,
Dmbr values for tomato fruit cutin (FTC), a mem-
brane with lipids as major component (Heredia-
Guerrero et al. 2012), are also drawn in Fig. 1b, Liquid Chromatography: Organic
and its comparison with RC/LL results indicates Carbon Detection (LC-OCD)
lower xed charge (XfFTC 5104M) but rather
similar transport parameters (<t+FTC > = 0.68) Loreen O. Villacorte
supporting the reliability of RC/LL values. FMC Technologies, Separation Innovation and
These results show small changes in transport Research Center, Arnhem, The Netherlands
parameters associated with LL deposition and
open the scope to study natural membranes.
Liquid chromatography organic carbon detec-
tion (LC-OCD) is an analytical technique for
identication and quantication of natural organic
References
matter (NOM) constituents in aquatic environ-
Benavente J (2009) Electrical characterization of mem- ments and water-soluble synthetic organic matter
branes. In: C. G uell, M. Ferrando, F. Lpez (eds), in technical waters. This technique has several
Monitoring and visualizing membrane-based process. specic applications including NOM investiga-
ISBN: 978-3-527-32006-6. Wiley-VCH tion in drinking water, wastewater, and marine
Benavente J, Vzquez MI, Hierrezuelo J, Rico R, Lpez-
Romero JM, Lpez-Ramirez R (2010) Modication of waters and quality control monitoring of ultrapure
regenerated cellulose membrane with lipid water used in power plants and the semiconductor
nanoparticles and layers. Nanoparticle preparation, industry (Huber and Frimmel 1994; Huber
morphological and physicochemical characterization et al. 2011). It is widely applied in membrane-
of nanoparticles and modied membranes. J Membr
Sci 355:4552 based water treatment to characterize the different
Benavente J, Vzquez MI, Hierrezuelo J, Lpez-Romero NOM constituents in the source waters (e.g.,
JM (2011) Physicochemical and transport parameters Kennedy et al. 2005; Amy et al. 2011; Villacorte
for a lipid coated regenerated cellulose membrane, et al. 2012), to assess the organic removal ef-
Vacuum, 85:10671070
Goodsell DS (2004) Bionanotechnology. Wiley-Liss, New ciency of pretreatment and membrane ltration
Jersey processes (e.g., Frimmel et al. 2004; Villacorte
Heredia-Guerrero JA, Lara R, Dominguez E, Heredia A, et al. 2009, 2010; Huang et al. 2011; Zheng
Benavente J, Benitez JJ (2012) Chemical-physical et al. 2010), and to identify the NOM constituents
characterisation isolated plant cuticles subjected to
low-dose g-irradiation. Chem Phys Lipids 165:803808 that cause fouling in MF/UF and NF/RO systems
Modication of regenerated cellulose membrane with lipid (e.g., Huber 1998; Henderson et al. 2011;
nanoparticles and layers. Nanoparticle preparation, Kennedy et al. 2008; Jiang et al. 2010; Batsch
morphological and physicochemical characterization et al. 2005; Zheng et al. 2009).
of nanoparticles and modied membranes. J Membr
Sci 355:4552 The principle behind NOM fractionation by
Sackmann S (1996) Supported membranes: Scientic & LC-OCD is based on three separation processes,
Practical Applications. Science 271:4348 namely, size exclusion, ion interaction, and
Liquid Chromatography: Organic Carbon Detection (LC-OCD) 1101
hydrophobic interaction. Since NOM constituents based on steric interactions or physical sieving
are highly heterogeneous in terms of size and where the difference in speed of diffusion for
majority of which are hydrophilic and weakly smaller and larger molecules is used to identify
acidic, size exclusion is considered as the domi- the different NOM fractions in the mobile phase
nant mechanism of separation (DOC-Labor (e.g., buffered water sample). The stationary
2006). Size exclusion chromatography (SEC) is phase is a packing of porous beads which allows
smaller molecules to diffuse into the bead interior
while preventing the larger molecules to diffuse
Liquid Chromatography: Organic Carbon Detection
through. As a consequence, larger molecules have
(LC-OCD), Table 1 Characteristics of different constitu-
ents of NOM identifiable by LC-OCD (www.doc-labor.de; less volume to traverse and travel faster through
Huber et al. 2011; Batsch et al. 2005) the chromatogram column (shorter elution time)
NOM Typical than smaller molecules.
fraction size (Da) Typical composition The modern LC-OCD technology is mainly
Biopolymers >20,000 Polysaccharides, proteins, attributed to the works of Stefan Huber in the
amino sugars, early 1990s when he successfully improved the
polypeptides, TEPs, EPS
sensitivity of the LC/DOC (predecessor of
Aquatic 1000 Humic and fulvic acids
humics
LC-OCD) developed earlier at the Engler-Bunte
Building 300450 Weathering and oxidation Institute in Karlsruhe, Germany (http://www.doc-
blocks products of humics labor.de/). The current LC-OCD system has an
LMW <350 Mono-oligosaccharides, online organic carbon detector (OCD), UV detec-
neutrals alcohols, aldehydes, tor (UVD), and organic nitrogen detector (OND)
ketones, amino acids to continuously measure the relative signal
LMW acids <350 All monoprotic organic
response of organic carbon, UV, and organic L
acids
nitrogen, respectively, at different retention
12
OCD: organic carbon detection
Building Blocks UVD: UV detection at l=254 nm
Acids and OND: organic nitrogen detection
Humics LMW Humics
10
8
rel. Signal Response
Biopolymers
LMW-Neutrals
6
OCD
4
Inorganic
Colloids
2
UVD
Nitrate
OND
0
20 30 40 50 60 70 80
Retention Time in Minutes
Liquid Chromatography: Organic Carbon Detection (LC-OCD), Fig. 1 Typical LC-OCD chromatogram of NOM in
surface water (DOC-Labor 2006)
1102 Liquid Chromatography: Organic Carbon Detection (LC-OCD)
times. In principle, the analysis technique is as Henderson RK, Subhi N, Antony A, Khan SJ, Murphy KR,
follows: (1) injection of buffered particle-free Leslie GL, Chen V, Stuetz RM, Le-Clech P (2011)
Evaluation of efuent organic matter fouling in ultra-
water sample to a chromatographic column to ltration treatment using advanced organic characteri-
separate fractions of NOM, (2) nondestructive sation techniques. J Membr Sci 382:5059
UV detection at 254 nm wavelength, (3) organic Huang G, Meng F, Zheng X, Wang Y, Wang Z, Liu H, Jekel
carbon detection based on high-sensitivity TOC M (2011) Biodegradation behavior of natural organic
matter (NOM) in a biological aerated lter (BAF) as a
analysis, and (4) simultaneous detection of pretreatment for ultraltration (UF) of river water. Appl
organic nitrogen in bypassed samples after the Microbiol Biotechnol 90:17951803
UV detector (see detailed specications by Huber SA (1998) Evidence for membrane fouling by spe-
Huber et al. 2011). The chromatogram data gen- cic TOC constituents. Desalination 119:229234
Huber SA, Frimmel FH (1991) Flow injection analysis for
erated by the three detectors are processed using a organic and inorganic carbon in the low-ppb range.
customized software program (ChromCALC, Anal Chem 63:21222130
DOC-Labor, Karlsruhe) to calculate organic car- Huber SA, Frimmel FH (1994) Direct gel chromatographic
bon concentrations of biopolymers, humic sub- characterization and quantication of marine dissolved
organic carbon using high-sensitivity DOC detection.
stances, building blocks, low molecular weight Environ Sci Technol 28:11941197
(LMW) acids, and neutrals fractions of NOM Huber SA, Balz A, Abert M, Pronk W (2011) Characteri-
based on area integration of the fractional peaks. sation of aquatic humic and non-humic matter with
The lower limit of detection of this technique was size-exclusion chromatography organic carbon
detection organic nitrogen detection
reported to be in the low-ppb range for individual (LC-OCD-OND). Water Res 45:879885
fractions (Huber and Frimmel 1991). The typical Jiang T, Kennedy MD, DeSchepper V, Nam S-N, Nopens I,
size ranges and chromatograms of the different Vanrolleghem PA, Amy G (2010) Characterization of sol-
NOM fractions detectable by LC-OCD are uble microbial products and their fouling impacts in mem-
brane bioreactors. Environ Sci Technol 44:66426648
shown in Table 1 and Fig. 1, respectively. Kennedy MD, Chun HK, Yangali-Quintanilla VA,
Heijman BGJ, Schippers JC (2005) Natural organic
matter (NOM) fouling of ultraltration membranes:
Cross-References fractionation of NOM in surface water and characteri-
sation by LC-OCD. Desalination 178:7383
Kennedy MD, Kamanyi J, Heijman BGJ, Amy G (2008)
Extracellular Polymeric Substance (EPS) Colloidal organic matter fouling of UF membranes: role
Natural Organic Matter (NOM) of NOM composition %26 size. Desalination 220:200213
Transparent Exopolymer Particle Villacorte LO, Kennedy MD, Amy G, Schippers JC
(2009) The fate of transparent exopolymer particles
(TEP) in integrated membrane systems: removal
through pretreatment processes and deposition on
References reverse osmosis membranes. Water Res 43:50395052
Villacorte LO, Schurer R, Kennedy MD, Amy G,
Amy GL, Salinas-Rodriguez SG, Kennedy MD, Schippers Schippers JC (2010) The fate of transparent
JC, Rapenne S, Remize P-J, Barbe C, Manes CLDO, exopolymer particles in integrated membrane systems:
West NJ, Lebaron P, Kooij DVD, Veenendaal H, a pilot plant study in Zeeland, The Netherlands. Desa-
Schaule G, Petrowski K, Huber S, Sim LN, Ye Y, lination Water Treat 13:109119
Chen V, Fane AG (2011) Water quality assessment Villacorte LO, Ekowati Y, Winters H, Amy GL, Schippers
tools. In: Drioli E, Criscuoli A, Macedonio F (eds) JC, Kennedy MD (2012) Characterisation of transpar-
Membrane-based desalination an integrated approach ent exopolymer particles (TEP) produced during algal
(MEDINA). IWA Publishing, New York, pp 332 bloom: a membrane treatment perspective. Desalin
Batsch A, Tyszler D, Br ugger A, Panglisch S, Thomas Water Treat. doi:10.1080/19443994.2012.699359
M (2005) Foulant analysis of modied and unmodied Zheng X, Ernst M, Jekel M (2009) Identication and
membranes for water and wastewater treatment with quantication of major organic foulants in treated
LC-OCD. Desalination 178:6372 domestic wastewater affecting lterability in dead-end
DOC-Labor (2006) LC-OCD Liquid chromatography ultraltration. Water Res 43:238244
Organic Carbon Detection. Information Brochure Zheng X, Ernst M, Jekel M (2010) Pilot-scale investigation
1/2006 on the removal of organic foulants in secondary efuent
Frimmel FH, Saravia F, Goreno A (2004) NOM removal by slow sand ltration prior to ultraltration. Water Res
from different raw waters by membrane ltration. 44:32033213
Water Sci Technol Water Supply 4:165174
Liquid Crystal Polymer Membranes 1103
membranes that have a large surface area to vol-

Liquid Crystal Polymer Membranes ume ratio are of great current interest for their
potential application in ltration, separation, ion
P. C. van H. Kuringen, A. P. H. J. Schenning and conductivity, drug delivery, and catalysis. The
D. J. Broer small pore size in these materials (less than
Department of Chemical Engineering and 1 nm) makes discrimination between molecules
Chemistry, Group Functional Organic and ions based on size and shape possible.
Materials and Devices (SFD), Eindhoven LCs can self-assemble in a variety of
University of Technology, Eindhoven, The phases that have orientation order and, in some
Netherlands cases, positional order. For the construction of
nanoporous membranes, various phases have
been used. Well-known examples are hexagonal
Liquid crystals (LCs) combine properties of both or columnar, lamellar or smectic, and cubic phases
liquids and crystals and can be organized in a (Fig. 1).
variety of nanostructured polymer lms with a One-dimensional pores are made from LCs
monolithic structure. Therefore, polymer-based with either a hexagonal or a columnar phase.
LCs are ultimately suited as membrane with an These phases are assumed by disk-shaped LCs:
accurately controlled nanoporosity. Nanoporous the polar regions are located at the center of the
a O
O crosslinker
O
( )
O ( ) ) ( O
O O O O O O O N+
L
O
( )
O 2 3 C12H25O
1 O BF4-
( )
O
O
O
O ( )O
( ) O O
O O ( ) Na+
O O
O-Na+
O-
1D 2D 3D
Liquid Crystal Polymer Membranes, Fig. 1 (a) Exam- materials to form nanostructured materials, with respec-
ples of chemical structures of the LCs used in the construc- tively one, two, or three dimensional pores. The red part
tion of nanoporous membranes. One columnar or represents the pore while the blue fraction is the molecular
hexagonal LC, two smectic or lamellar LC, and three region (Adapted with permission from reference (Kato
bicontinuous cubic LC. (b) The self-assembly of these 2010). Copyright Wiley, VCH)
1104 Liquid Crystal Polymer Membranes
columns and the hydrocarbon tails form the sur- larger molecules (contaminants) are rejected. The
rounding matrix (Fig. 1-a1). If the molecules are presence of charges in the pores of the membranes
well aligned, these one-dimensional pores are enables the selective permeation (Gin et al. 2006;
interesting for their low tortuosity. The straight Henmi et al. 2012) of ions with a remarkable
pores result in a short path length through the selectivity: for example, a membrane with pore
thickness of the material. The cylindrical pores size of approximately 0.6 nm was shown to trans-
must be continuous through the membrane and port more divalent sulfate ions than monovalent
uniformly aligned perpendicular to the lm sur- chloride ions (Henmi et al. 2012). Furthermore,
face for optimum transport. the pore size can be tailored to control the passage
Two-dimensional pores can be created by of molecules, making them suitable for
smectic LCs (Fig. 1-a2). These rod-like LCs programmed drug release systems or the conned
self-assemble in lamellar-like structures where environment within the material can be used to
the molecules are orientated perpendicular to the enhance chemical reactions.
layers. A lamellar structure is created with alter- Nanoporous LC polymer membranes can be
nating layers of a two-dimensional polymer sheets produced with pore sizes of 1 nm and below.
and two-dimensional pores. However, these Unique features can be realized, but because of
sheets do not connect to each other. To circumvent the length of the pores exceeding micrometers,
disruption of the lm, crosslinkers are used to the permeability of liquids is generally low with
keep the layers together. To some extent the pore high pressures necessary to achieve sufcient
size of the materials can be controlled by the ow across the membrane. The challenge in
length of the crosslinker. this eld is to make thinner membrane lms
A bicontinuous cubic phase is used to make of a mechanically robust material that can with-
nanostructured LC membranes with pores in three stand these high pressures. Research on using
dimensions (Fig. 1-a3). Generally, the molecular mesoporous supporting layers to endure the
structure of the LC differs only slightly from the force while the LC polymer membrane deter-
LCs used for the columnar phase. However, the mines the separation characteristics is being
self-assembly is different. Now, no alignment of undertaken.
the pores is needed, since the pores are orientated
in three dimensions.
These nanostructured LC polymer networks
are used for different membrane applications, References
such as (anisotropic) ion conductivity (Schenning
et al. 2011; Kato 2010) and gas separation Gin DL, Lu X, Nemade PR, Pecinovsky CS, Xu Y, Zhou
(Schenning et al. 2011; Gin et al. 2006). Perme- M (2006) Recent advances in the design of
ation is the product of diffusion and solubility. polymerizable lyotropic liquid-crystal assemblies for
heterogeneous catalysis and selective separations.
Therefore the solubility values of the ions and Adv Funct Mater 16:865878
gases are very important. The local organization Henmi M, Nakatsuji K, Ichikawa T, Tomioka H,
of the nanostructured material can promote the Sakamoto T, Yosho M, Kato T (2012) Self-organized
solubility of certain species, whereas other mole- liquid-crystalline nanostructured membranes for water
treatment: selective permeation of ions. Adv Mater
cules are less soluble. This results in higher perme- 24:22382241
ation and selectivity of the desired molecules Kato T (2010) From nanostructured liquid crystals to
compared to isotropic material of the same compo- polymer-based electrolytes. Angew Chem Int Ed
sition. Another application is the nanoltration of 49:78477848
Schenning APHJ, Gonzalez-Lemus YC, Shishmanova IK,
water (Schenning et al. 2011; Gin et al. 2006; Broer DJ (2011) Nanoporous membranes based on
Henmi et al. 2012) Molecules smaller than the liquid crystalline polymers. Liq Cryst 38(1112):
pore size can pass through the membrane, while 16271639
Liquid Membrane Separation 1105
References
Liquid Entry Pressure (LEP or LEPW)
El Bourawi S, Ding Z, Ma R, Khayet M (2006)
A framework for better understanding membrane dis-
Renzo Di Felice
tillation separation processes. J Membr Sci 285:429
University of Genova, Genova, Italy Kim B-S, Harriot P (1987) Critical entry pressure for liquids in
hydrophobic membranes. J Colloid Interface Sci 115:18
Smolders K, Franken ACM (1989) Terminology for mem-
brane distillation. Desalination 72:249262
In membrane distillation it is of paramount impor-
tance that liquid does not ll the membrane pores,
in order to keep membrane efciency to the
highest possible value. To this end hydrophobic Liquid Membrane Separation
membranes are normally used. The hydrophobic-
ity can be quantied through the so-called liquid Vladimir S. Kislik
entry pressure (LEP) parameter, dened as the Campus Givat Ram, Casali Institute of Applied
pressure difference at which liquid penetrates Chemistry, The Hebrew University of Jerusalem,
into the membranes pores. Jerusalem, Israel
The pressure required to force water through the
structure is inversely proportional to the opening
size and dependent on the polymeric properties of Liquid membrane separation combines the sol-
the membrane. A theoretical expression, based on vent extraction and stripping processes
Young-Laplace equation, enables the estimation of (re-extraction) in a single step.
LEP (see, e.g., El Bourawi et al. 2006): This entry has the objective of introducing the
reader to the basic denitions of the liquid mem- L
2gcosy brane eld, with classication.
DP
r The term liquid membrane transport includes
where g is the liquid surface tension, y the contact processes incorporating liquid-liquid extraction
angle between liquid and membrane, and r is the (LLX) and membrane separation in one continu-
pore radius. ously operating device. It utilizes an extracting
Given the practical difculty of obtaining the reagent solution, immiscible with water,
exact value of the above physical parameters, it is stagnant or owing between two aqueous solu-
not uncommon to estimate LEP experimentally, tions (or gases), the source or feed, and receiving
and Smolders and Franken (1989) described a or strip phases. In most cases, the source and
detailed procedure which should be followed in receiving phases are aqueous, and the membrane
this case. Basically pressure difference is gradu- is organic, but the reverse conguration can also
ally applied on both side of the membrane until be used. A polymeric or inorganic microporous
the rst drop of the liquid appears on the permeate support (membrane) may be used as bearer (as in
side (more specic description is given in the SLM) or barrier (as in BLM technologies) or not
suggested reference). Typical values of LEP used, as in ELM and layered BLM (see respective
range from tenth to tens of bars (0.120 bar): for entries: SLM, Emulsion Liquid Membrane
example, with liquid water at ambient condition, a (ELM), Silver Recovery by Bulk Liquid Mem-
PTFE membrane with 95 % porosity, 25 mm thick- brane (BLM)).
ness, and 3 mm nominal pore size has an LEP of The great potential for energy saving, low cap-
0.13 bar, whereas a PTFE supported membrane ital and operating cost, and the possibility to use
with 75 % porosity, 120 mm thickness, and 0.2 expensive extractants, due to the small amounts of
nominal pore size has an LEP of 4.0 bar (Kim and the membrane phase, make LMs an area deserv-
Harriot 1987). ing special attention. Liquid membrane systems
1106 Liquid Membrane Separation
are being studied extensively by researchers in considered (see Fig. 1): bulk (BLM), supported
such elds as analytical, inorganic, and organic or immobilized (SLM or ILM), and emulsion
chemistry; chemical engineering, biotechnology, (ELM) liquid membrane transport. Some
and biomedical engineering; and wastewater treat- authors add to these denitions polymeric inclu-
ment. Research and development activities within sion membranes, gel membranes, and
these disciplines involve diverse applications of dual-module hollow-ber membranes, but, to my
liquid membrane technology, such as gas separa- opinion, the rst two types are the modications
tions, recovery of valued or toxic metals, removal of the SLM, and the third is the modication of
of organic compounds, development of sensing BLM.
devices, recovery of fermentation products, and According to the transport mechanisms, the
some other biological systems. LM techniques may be divided into simple trans-
The general properties of liquid membrane port, facilitated or carrier-mediated transport,
systems have been a subject of extensive theoret- coupled counter- or cotransport, and active
ical and experimental studies. Some general char- transport.
acteristics of LM processes are as follows: According to applications, the LM techniques
may be divided into (1) metal separation-
1. Liquid membrane separation is a rate process, concentration, (2) biotechnological product
and the separation occurs due to a chemical recovery-separation, (3) pharmaceutical product
potential gradient, not by equilibrium between recovery-separation, (4) organic compound
phases. separation and organic pollutant recovery from
2. LM is dened based on its function, not the wastewaters, (5) gas separations, (6) fermentation
material used in fabrication. or enzymatic conversion-recovery-separation
(bioreactors), (7) analytical applications, and (8)
Permeation is a general term for the wastewater treatment including biodegradation-
concentration-driven membrane-based mass trans- separation techniques.
port process. Differences in the permeability pro- Classication according to carrier type is as
duce a separation between solutes at constant follows: (1) water-immiscible, organic carriers,
driving force. Because the diffusion coefcients (2) water-soluble polymers, (3) electrostatic,
in liquids are typically orders of magnitude higher ion-exchange carriers, and (4) neutral, but polar-
than in polymers, a larger ux can be obtained with izable carriers.
liquid membranes. Application of a pressure dif- Classication according to membrane support
ference, an electric eld, or temperature consider- type is as follows: (1) neutral hydrophobic, hydro-
ably intensies the process. philic membranes, (2) charged (ion-exchange)
There are several different directions in LM membranes, (3) at sheet, spiral module mem-
separation classications: according to module branes, (4) hollow-ber membranes, and (5) cap-
design congurations (see SLM, Emulsion Liq- illary hollow-ber membranes.
uid Membrane (ELM), Silver Recovery by Module design congurations are used as a
Bulk Liquid Membrane (BLM) entries), rule as basic classication.
according to transport mechanisms (see Trans- Practically, there are a lot of opportunities for
port Mechanisms with Liquid Membranes), liquid membrane separation processes in many
according to applications, according to carrier areas of industry. The most interesting develop-
type, and according to membrane support type. ments for industrial membrane technologies are
Below, these types of classications are described related to the possibility of integrating various
and discussed briey. membrane operations in the same industrial
According to conguration denition, three cycle, with overall important benets in terms of
groups of liquid membranes are usually product quality and plant compactness.
Liquid Permeability 1107
Liquid Membrane
Separation, Fig. 1 Three BLM
congurations of liquid
membrane systems: bulk
(BLM), supported R
F E
(immobilized) (SLM or
ILM), and emulsion (ELM).
F is the source or feed Porous
Porous
phase, E is the liquid Support
Support
membrane, and R is the
receiving phase
SLM
E
F
R
Porous
Support
ELM
E R
L
F F
as micro-, ultra-, and nanoltration, reverse and

Liquid Permeability forward osmosis, or hybrid processes like mem-
branes for fuel cell. For commercial membranes
Alexey Volkov liquid permeability is usually given by manufac-
A.V. Topchiev Institute of Petrochemical turers along with value of molecular weight cutoff
Synthesis Russian Academy of Sciences (MWCO); the last one is characterized the ability
(TIPS RAS), Moscow, Russia of the membrane to reject the solutes with certain
molecular weight. Since it is already normalized
Liquid permeability is a liquid ux normalized by by driving force of the process, liquid permeabil-
driving force applied pressure difference across ity shall be most likely constant value for specic
the membrane. Liquid permeability can be membrane in certain range of operated pressures
expressed in volumetric or gravimetric at constant feed composition and temperature.
units [l/(m2hbar)] or [kg/(m2hbar)], respec- Due to its uniformity, this parameter allows to
tively. This parameter is used to characterize the compare performance of diverse number of mem-
membrane productivity and passage of pure sol- branes between each other, tested in accordance
vent or solution, in liquid separation process such with different ltration protocols.
1108 Liquid-Liquid Membrane Extraction
In some cases, the liquid permeability can be a

function of transmembrane pressure, especially if Lotus Effect
the membrane material of top layer has a notice-
able afnity to the solvent like for organic solvent Wolfgang M. Sigmund1 and Shu-Hau Hsu2
1
nanoltration. Then, increase of pressure differ- Department of Materials Science and
ence across the membrane could lead to partial Engineering, University of Florida, Gainesville,
compaction of swollen polymer matrix resulting FL, USA
2
in some decline of liquid permeability (Volkov Department of Energy Engineering, Hanyang
et al. 2008). This is typically a reversible effect, University, Seoul, Republic of Korea
and the membrane recovers its transport proper-
ties, while the pressure is reduced again. How-
ever, if it is accompanied by collapsing of the Lotus leaf (Nelumbo nucifera) has become the
membrane porous structure or transformation of epitome of natural superhydrophobic surfaces and
membrane morphology (e.g., as a result of chem- has long been considered as a sacred symbol of
ical cleaning), the change in liquid permeability is purity for thousand years in oriental culture due to
irreversible. In a case of membrane fouling, there its impressive self-cleaning feature, where leaves
could be a combination of reversible and irrevers- remain unsmudged even being immersed into
ible decline in liquid permeability, and partial muddy water. Water contact angle on lotus leaf is
recovery of membrane performance is possible reported above 160 with few degrees of roll-off
after regeneration stage (Drioli and Giorno 2010). angle. Therefore, lotus effect is sometimes a syno-
Liquid permeability can be determined from nym for superhydrophobicity or self-cleaning now-
ltration experiments in dead-end or cross-ow adays. Although the effect has long been noticed for
regimes. Ideally, no difference between two several generations, a systematically detailed inves-
regimes shall be expected for ltration of pure tigation was not carried out until 1997 where more
solvents or dilute solutions of solutes with no than 200 water-repellent plants were studied via
specic interaction with membrane material. scanning electron microscopy (Neinhuis and
Meanwhile, it is preferable to carry out the ltra- Barthlott 1997). The study reveals the secret of
tion test in cross-ow membrane cell with high lotus leaf, which, not surprisingly, attributes to its
linear velocity in order to minimize concentration combination of surface roughness and chemical sub-
polarization effect or membrane fouling during stances. Hydrophobicity and self-cleaning of lotus
the ltration of real mixtures (e.g., ground or leaf are believed as a mechanism to resist harmful
wastewater). microorganism bounding to the leaf surface, since
water is usually required for the germination.
The lotus leaf is covered by small protrusions
(Fig. 1a) called papillae with their average diameter
References and height about 10 mm. The papillae are further
covered by an additional layer of epicuticular
Drioli E, Giorno L (2010) Comprehensive membrane sci- waxes, generated from epidermal cells (Fig. 1b).
ence and engineering, vol 1. Elsevier, Kidlington
Volkov AV, Korneeva GA, Tereshchenko GF
These wax crystals are presented in submicron
(2008) Organic solvent nanoltration: prospects and size and in crystalline tubules with water contact
applications. Russ Chem Rev 77:983993 angles of about 95110 , which is considered
hydrophobic. The epicuticular waxes play a practi-
cally important role as they are not only to provide
hydrophobicity but to generate an additional rough-
ness in a smaller length of scale other than micron-
Liquid-Liquid Membrane Extraction sized bumps. The absence of wax crystals, i.e.,
dropping hot water onto the leaf, will totally elimi-
Membrane Based Solvent Extraction nate the superhydrophobicity (Liu et al. 2009). The
Low Fouling Membranes 1109
Lotus Effect, Fig. 1 SEM micrographs of lotus leaf showing its relatively rough surface covered by small micron-sized
protrusions (a) and submicron-sized wax crystals (b) (Hsu 2010)
kind of hierarchical roughness on superhydrophobic separation membrane technologies, including

surfaces seems to play a crucial role, but the detailed microltration, ultraltration, nanoltration, and
mechanism is not yet completely clear. A general osmosis processes diminishes process produc-
benet suggested is to repel both macro- and micro- tivity, raises operating costs, and shortens mem-
scope water droplets (Nosonovsky and Bhushan brane life span. Fouling is the deposition of
2007). Surfaces with only one scale of roughness retained particles, colloids, macromolecules,
repelled macroscopic droplets fairly well, while the salts, etc., at the membrane surface or inside the L
condensation may easily form microscopic droplets pore, and at the pore walls. Fouling reduces the
between the grooves of the surface structure. membrane ux either temporarily or permanently.
The decrease of permeate ux or irreversible
References membrane fouling is recognized as the main prob-
lem in the application of membrane ltration tech-
Hsu SH (2010) Biologically inspired hairy surfaces for liquid nologies. Several types of membrane fouling have
repellency. Doctoral dissertation, University of Florida been introduced including inorganic fouling or
Liu Y, Chen X, Xin JH (2009) Can superhydrophobic scaling, colloidal fouling, organic fouling, and
surfaces repel hot water? J Mater Chem 19:56025611
Neinhuis C, Barthlott W (1997) Characterization and dis-
biofouling. A schematic representation is given
tribution of water-repellent, self-cleaning plant sur- in Fig. 1. Among them, the formation of biolms
faces. Ann Bot 79:667677 on the membrane surfaces or membrane biofoul-
Nosonovsky M, Bhushan B (2007) Multiscale friction ing has been regarded as the most serious
mechanisms and hierarchical surfaces in nano- and
bio-tribology. Mater Sci Eng R 58:162193
problem. The biofouling or biolm formation
starts with the adsorption of proteins and humic
substances on membrane surfaces which serves as
nutrient for microorganisms (Banerjee et al. 2011;
Low Fouling Membranes
Lejars et al. 2012). The two major approaches to
combat surface fouling are based on either
Bijay P. Tripathi
preventing foulants from attaching or degrading
Department of Nanostructured Materials, Leibniz
them. It is generally accepted that an increase in
Institute of Polymer Research Dresden, Dresden,
hydrophilicity offers better fouling resistance
Germany
because protein and many other foulants are
hydrophobic in nature (Rana and Matsuura
Membrane fouling a common hindrance to 2010). The high hydrophilicity leads to a high
the advancement of water treatment and surface hydration which is generally considered
1110 Low Fouling Membranes
Organic Fouling
Precipitate Fouling
Colloidal Fouling
Membrane fouling
Biofouling
Low Fouling Membranes, Fig. 1 Schematic representation of different types of fouling
the key to nonspecic protein adsorption resis- polydopamine coatings showed additional resis-
tance, organic fouling, microorganisms attach- tance to protein and bacterial adhesion.
ment, etc., since a tightly bound water layer Compared with the traditional hydrophilic
forms a physical and energetic barrier to prevent antifouling membrane surfaces, amphiphilic
the adsorption to the surface. While protein- membrane surfaces engineered by the forced
resistant coatings may also resist bacterial attach- surface segregation method were demonstrated
ment and subsequent biolm formation, in order to exhibit ultralow membrane fouling. The amphi-
to overcome the fouling-mediated risk of bacterial philic surfaces bearing mixed brush architecture
infection, it is highly desirable to design coatings comprised of both hydrophilic blocks and low
that are bactericidal. More recently, hydrophilic surface energy blocks are also suggested to be
coatings using 3,4-dihydroxyphenylalanine potential antifouling agents. Hydrophilic blocks
(DOPA) and dopamine have been used to generated fouling-resistant hydration layers,
modify the surface of microporous polyethylene whereas low surface energy blocks which formed
(PE), polyvinylidene uoride (PVDF), and amounts of nonadhesive microdomains played a
polytetrauoroethylene (PTFE), polysulfone, major role in repulsing the foulants away from the
etc. membranes. The strong adhesive properties surfaces. Surface modication with zwitterionic
of these compounds formed lms that attached molecules and polymers is believed to be the most
well to the hydrophobic membranes and decreased effective method to counter all types of fouling
the contact angle from 50 to 30 . PEG-modied (Tripathi et al. 2013). Zwitterionic surfaces and
LTA Zeolite Membranes 1111
membranes can be obtained by reacting different

sultones with amine group containing polymers LTA Zeolite Membranes
such as chitosan, polystyrene-poly-4-vinylpyridine
block copolymer, etc. Incorporation of nanoparticles Yanshuo Li
into polymeric ltration membranes such as silica, State Key Laboratory of Catalysis, Dalian
Al2O3, and TiO2 etc. also reported to enhance the Institute of Chemical Physics, CAS, Dalian,
antifouling performance of the membranes because China
the presence of hydrophilic nanoparticles in the
polymer matrix enhances its overall hydrophilic
behavior. Nano-sized TiO2 has received signicant LTA zeolite membrane is one of the most well-
interest for its high hydrophilicity, chemical stability, studied zeolite membranes for the separation of
and antibacterial properties (Mansouri et al. 2010). light gases and dewatering of organics. LTA
In addition, anatase TiO2 could serve as a (Linde type A) zeolite has a three-dimensional
photocatalyst to degrade pollutants during wastewa- pore structure, consisting of secondary building
ter treatment while having an antimicrobial effect in units (SBUs) 4, 6, 8, and 4-4 (Meier et al.). An
the presence of UV. example material is NaA zeolite (Linde type
A sodium form) with chemical formula of
Na12Al12Si12O48 27H2O. The unit cell of NaA
zeolite is cubic (a = 24.61 ) with Fm-3c sym-
metry. The pore diameter of NaA zeolite is dened
References
by an eight-member oxygen ring with size around
Banerjee I, Pangule RC, Kane RS (2011) Antifouling coat- 4.3 . Generally, LTA zeolite membranes exist in
ings: recent developments in the design of surfaces that the form of composite membranes, for which L
prevent fouling by proteins, bacteria, and marine organ- porous ceramic supports, typically alumina
isms. Adv Mater 23:690718
tubes, are usually used as substrates. The effective
Lejars M, Margaillan A, Bressy C (2012) Fouling release
coatings: a nontoxic alternative to biocidal antifouling LTA zeolite separating layer is usually in the
coatings. Chem Rev 112:43474390 thickness of several microns (Fig. 1).
Mansouri J, Harrisson S, Chen V (2010) Strategies for The most used strategies for the synthesis of
controlling biofouling in membrane ltration systems:
LTA zeolite membranes are in situ synthesis and
challenges and opportunities. J Mater Chem
20:45674586 secondary (seeded) growth. Microwave tech-
Rana D, Matsuura T (2010) Surface modications for niques combining with these strategies have
antifouling membranes. Chem Rev 110:24482471 been applied for the synthesis of LTA zeolite
Tripathi BP, Dubey NC, Choudhury S, Simon F, Stamm M
membranes (Li et al. 2008). As compared with
(2013) Antifouling and antibiofouling pH responsive
block copolymer based membranes by selective surface conventional hydrothermal synthesis, microwave
modication. J Mater Chem B 1:33973409 synthesis of zeolite membranes has the
LTA Zeolite Membranes, Fig. 1 (a) Framework structure of LTA zeolite, (b) typical SEM image of a NaA zeolite single
crystal, (c, d) typical SEM image of a LTA zeolite membrane
1112 LTA Zeolite Membranes in Pervaporation
advantages of shorter synthesis time, better mem- LTA Zeolite Membranes in Pervaporation,
brane performance, and easier to scale up. Table 1 Typical PV/VP performances of LTA zeolite
membrane for dewatering organics
LTA zeolite membrane has been extensively
Feed solution Water
studied for the separation of light gas mixtures, (10 wt.% of Temperature ux Separation
e.g., H2/CH4, O2/N2, and n-C4H10/i-C4H10 gas water) [ C] (kg/m2h) factor
pairs. The observed separation factors beyond Water/methanol 65 (PV) 0.6 116
the selectivities predicted from Knudsen diffusion Water/ethanol 65 (PV) 1.0 10,000
are attributed to the so-called molecular sieving Water/ethanol 120 (VP) 3.1 10,000
Water/n-propanol 110 (VP) 2.5 10,000
function. Another important application of LTA
Water/i-propanol 65 (PV) 1.8 10,000
(particularly NaA) zeolite membranes is for Water/ethyl 65 (PV) 1.2 10,000
removing water from various organics via acetate
pervaporation (PV) or vapor permeation (VP), Water/acetone 50 (PV) 0.9 10,000
thanks to its ultramicroporous channel structure Water/i- 105 (VP) 3.2 10,000
propylamine
and its hydrophilic nature. Nowadays, VP
dewatering technology based on NaA zeolite
membranes has been widely used in the
industries of chemistry, bioenergy, electronics, shows a high efciency for the separation of azeo-
pharmaceutics, etc. Compared with azeotropic tropes and close-boiling mixtures (Huang 1991).
distillation, pressure swing adsorption (PSA), In general, zeolites with low Si/Al ratio can be
and polymeric membranes, zeolite membrane adopted to produce hydrophilic pervaporative
can remarkably reduce the energy consumption membranes for dewatering of organics, as exem-
for dewatering of organics (Van hoof et al. 2004). plied by LTA, FAU, SOD, and T-type zeolites
(Bowen et al. 2004). Among those, LTA zeolite
membranes exhibit the highest water to alcohol
References selectivity up to present, owing to its lowest Si/Al
ratio and proper pore size (0.4 nm) that can block
Li YS (2008) Microwave synthesis of zeolite membranes:
a review. J Membr Sci 316:317 organic solvent molecules bulkier than water. LTA
Meier WM, Olson DH, Baerlocher C, Atlas of zeolite zeolite membranes have been widely used for
structure types, ISBN-13: 978-0444100153, Excerpta dewatering various organics, including alcohols,
Medica
aldehydes, ketones, ethers, esters, amines, and
Van Hoof V et al (2004) Economic comparison between
azeotropic distillation and different hybrid systems com- amides. Typical performances of LTA membrane
bining distillation with pervaporation for the dehydration are listed in Table 1.
of isopropanol. Sep Purif Technol 37:3349 LTA zeolite membrane is the rst zeolite mem-
brane to be used in large-scale industries for the
dehydration of alcohols (Morigami et al.
2001). Nowadays, commercial LTA zeolite mem-
LTA Zeolite Membranes in branes (in NaA form) can be bought from the
Pervaporation Japanese company, the Nano-Research Institute
Inc. (XNRI), a 100 % subsidiary of Mitsui &
Yanshuo Li Co., the European alliance between Smart
State Key Laboratory of Catalysis, Dalian (UK) and Inocermic, and most recently the Chi-
Institute of Chemical Physics, CAS, Dalian, nese company, HST Co. Ltd., located at Dalian.
China Compared with azeotropic distillation and
pressure swing adsorption (PSA), LTA zeolite
membrane can remarkably reduce the energy con-
LTA Zeolite Membrane in Pervaporation. sumption for dewatering of organics. Taking the
Pervaporation (PV) or vapor permeation (VP) is production of anhydrous ethanol as an example,
a membrane-based separation process which LTA zeolite membrane-based VP process can
LTA Zeolite Membranes in Pervaporation 1113
reduce the consumption of steam by 40 % and Huang RYM (1991) Pervaporation membrane separation
cooling water by 30 % compared with PSA, processes, Elsevier. ISBN 0444882278
Morigami Y, Kondo M, Abe J, Kita H, Okamoto K (2001)
respectively. The rst large-scale pervaporation plant using tubular-
type module with zeolite NaA membrane. Sep Purif
Tech 25:251260
References
Bowen TC, Noble RD, Falconer JL (2004) Fundamentals

and applications of pervaporation through zeolite mem-
L
M
Macrocapsules wall (i.e., light) (Bogdanowicz et al. 2013), which

can cause its breaking and the immediate release
Carles Torras Font of the macrocapsule contents (thus, convection
Bioenergy and Biofuels, Catalonia Institute for being the predominant mechanism).
Energy Research (IREC), Tarragona, Catalonia, Often, macrocapsules are manufactured from
Spain polymer due to the simplicity of the process,
capacity to control the porous structure of the
wall, and cost (specially compared with inorganic
Solid material with a circular cross section materials). Immersion precipitation (a type of
(or with circularities near 1). Typical sizes of phase inversion) can be used to obtain
macrocapsules (also named capsules) are between macrocapsules. First, the polymer is dissolved in
hundreds of micrometers to millimeters. a solvent. Second, drops are obtained (for
Macrocapsules have a thick porous wall with a instance, with a needle). Finally, drops are
small empty core (or nonexistent empty core) immersed in a bath containing the non-solvent,
(Pena et al. 2009). This characteristic causes the which makes the polymer precipitate and obtain
differentiation between them and macrospheres, the macrocapsules (Gumi et al. 2009). Conditions
which are similar but have a dense and thin wall of this process allow controlling the porous struc-
and a large empty volume inside (Lakshmi ture of the wall. Several variables exist: type of
et al. 2012). polymer, polymer concentration, type of solvent,
A main use of macrocapsules is to store com- type and composition of the non-solvent, bath
pounds inside, which are released in a controlled temperature, etc. The compound to be stored is
way and from different parameters (Pena usually added when the macrocapsule is being
et al. 2009). First of all, there should be a driven fabricated and when the polymeric uid is
force that usually is concentration gradient. Sec- obtained.
ondly, permeability occurs through the wall due to Two main characterizations are performed with
convection and diffusion. Depending on the type macrocapsules: release experiments and physical
of process, each of them can be the predominant. characterization of the macrocapsule to check
For instance, considering macrocapsules with their porous structure, hydrophilicity, etc. (Zhang
dense and thick wall, where only a concentration et al. 2005). The second one is performed using
gradient occurs, diffusion is predominant and the same methods as for at sheet membranes:
allows releases that can last several days mercury porosimetry, scanning electron micros-
(Zhang et al. 2005). On the other hand, the release copy (SEM), contact angle, etc. To check the wall
process can start after an external stimulus to the porous structure with SEM, the macrocapsule
DOI 10.1007/978-3-662-44324-8
1116 Macroporous, Mesoporous, and Microporous Membranes
Eng Chem Res 48(3):15621565. doi:10.1021/

ie801133f
Torras C, Pitol L, Garcia-Valls R (2007) Two methods for
morphological characterization of internal microcap-
sule structures. J Membr Sci 305:14. doi:10.1016/j.
memsci.2007.06.056
Zhang J, Yao S-J, Guan Y-X (2005) Preparation of
macroporous sodium cellulose sulphate/poly(dimethyl-
diallylammonium chloride) capsules and their charac-
teristics. J Membr Sci 255(12):8998. doi:10.1016/j.
memsci.2005.01.025
Macroporous, Mesoporous,
and Microporous Membranes
Macrocapsules, Fig. 1 SEM cross section micrograph of
a broken macrocapsule, showing the wall porous structure
Lidietta Giorno, Emma Piacentini and
Fabio Bazzarelli
should be broken (not cut to avoid wall damage). Institute on Membrane Technology, National
It can be performed using a cryotome or breaking Research Council of Italy, ITM-CNR, Rende,
the capsules within liquid nitrogen after swelling Italy
(i.e., ethanol) (Zhang et al. 2005; Torras
et al. 2007) (Fig. 1). An option to check visually
the porous structure of the macrocapsule without
breaking it is by using a confocal laser scanning Macroporous membranes consist of a solid matrix
microscopy after swelling the macrocapsule with with dened holes or pores which have diameters
a proper uorescent agent (i.e., rhodamine) larger than 50 nm (Strathmann et al. 2006). These
(Lamprecht et al. 2000). types of membranes are used in driving force
pressure process such as microltration
(MF) and ultraltration (UF). They can be classi-
References ed according to:
Bogdanowicz KA, Tylkowski B, Giamberini M (2013) 1. The type of materials: organic (polymers) and
Preparation and characterization of light-sensitive
microcapsules based on a liquid crystalline polyester.
inorganic (glasses, metals, and ceramics)
Langmuir: ACS J Surf Colloids 29(5):16011608. 2. Their structure: symmetric and asymmetric
doi:10.1021/la3038878 3. Their conguration: at sheet, hollow ber
Gumi T, Gascon S, Torras C, Garcia-Valls R (2009) membrane, capillary, and tubular
Vanillin release from macrocapsules. Desalination
245(13):769775. doi:10.1016/j.desal.2009.02.050
Lakshmi DS, Figoli A, Buonomenna MG, Golemme G, The separation of solutes by macroporous
Drioli E (2012) Preparation and characterization of membranes is mainly a function of molecular
porous and nonporous polymeric microspheres by the size and pore size distribution. The methods are
phase inversion process. Adv Polym Technol
31(3):231241. doi:10.1002/adv.21281
used for making membranes are phase inversion,
Lamprecht A, Schfer UF, Lehr CM (2000) Characteriza- sintering, stretching, track etching, template
tion of microcapsules by confocal laser scanning leaching, slip casting, and sol-gel process.
microscopy: structure, capsule wall composition and Mesoporous membranes consist of a solid
encapsulation rate. Eur J Pharm Biopharm 49(1):19.
doi:10.1016/S0939-6411(99)00063-6
matrix with dened holes or pores which have
Pena B, Casals M, Torras C, Gumi T, Garcia-Valls R (2009) diameters in intermediate range between 2 and
Vanillin release from polysulfone macrocapsules. Ind 50 nm (Strathmann et al. 2006).
Macrosolute 1117
Membrane
macro-porous meso-porous micro-porous non-porous
type
Pore size (nm) > 50 2 - 50 <2 no permanent pores

Separated * *
*
components *
* * *
*
suspended solids viruses bacteria colloids proteins sugars soluble salts gases/vapours
Macroporous, Mesoporous, and Microporous Membranes, Fig. 1 Schematic classication of membranes, related
processes, and separated components
Mesoporous membranes are used mainly for The transport is based on diffusion model. The
nanoltration (NF) process. A hydrostatic pres- driving force for gas separation and dialysis is
sure is applied to transport molecule mixture mass represented by pressure and concentration gradi-
across the membrane. They can be classied ent respectively.
according to: The methods are used for making membranes
are phase inversion, sintering, stretching, track
1. The type of materials: organic (polymers) and etching, template leaching, slip casting, and
inorganic (glasses, metals, and ceramics) sol-gel process.
2. Their structure: symmetric and asymmetric The schematic representation of such classi-
M
3. Their conguration: at sheet, hollow ber cation is illustrated in Fig. 1.
membrane, capillary, and tubular
The methods are used for making membranes References

are phase inversion, sintering, stretching, track
etching, template leaching, slip casting, and Strathmann H, Giorno L, Drioli E (2006) An introduction
to membrane science and technology. Betmultimedia,
sol-gel process.
CNR Rome
The separation of solutes by mesoporous mem-
branes is a function of molecular size, pore size,
distribution and electrostatic interaction.
Microporous membranes consist of a solid
matrix with dened holes or pores which have Macrosolute
diameters smaller than 2 nm (Strathmann
et al. 2006). They can be classied according to: Radoslav Paulen
Department of Biochemical and Chemical
1. The type of materials: organic (polymers) and Engineering, Technische Universitt Dortmund,
inorganic (glasses, metals, and ceramics) Dortmund, Germany
2. Their structure: symmetric and asymmetric
3. Their conguration: at sheet, hollow ber
membrane, capillary, and tubular. Macrosolute represents an operational term from
the eld of membrane ltration which denotes
This type of membrane can be used for gas solute(s) of particle size(s) larger than ones that
separation and dialysis. pass through a membrane (microsolute) of a
1118 Magnetic Membranes
Macrosolute, Fig. 1 1,50E-04

Comparison of cross-ow
CONCENTRATION OF THE SOLUTE IN

and dead-end ltration in
terms of permeate
THE PERMEATE [mM/ml]

concentration
dead-end
cross-flow with optimum
surface renewal

Enrico Drioli Lidietta Giorno Editors Enc Yclopedia of Membranes

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Enrico Drioli Lidietta Giorno Editors Enc Yclopedia of Membranes

Uploaded by

Copyright:

Available Formats

Enrico Drioli

With 968 Figures and 185 Tables

ISBN 978-3-662-44323-1 ISBN 978-3-662-44324-8 (eBook)

# Springer-Verlag Berlin Heidelberg 2016

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The Encyclopedia of Membranes is a reference work that covers all aspects of

Water treatment and purication

The large spectrum of challenges membrane operations contribute to face

We are very pleased to acknowledge the important help and contribution

Enrico Drioli Emeritus Professor, University of Calabria, Rende, Italy

Advisory Board of Journal of Membrane Science, Polish Journal of Chem-

Lidietta Giorno is serving as the Director of the Institute on Membrane

Bhattacharyya Dibakar University Alumni Professor, Lexington, KY

Adele Brunetti ITM-CNR c/o University of Calabria, Rende, Italy

Francesca Macedonio ITM-CNR c/o University of Calabria, Rende, Italy

A. Achilli Environmental Resources Engineering Department, Humboldt

He Bai Momentive Performance Materials Inc, Friendly, WV, USA

Renbi Bai Department of Civil and Environmental Engineering, National

Malini Balakrishnan The Energy and Resources Institute (TERI), New

Joerg Balster Application Technology, Evonik Fibres GmbH, Schoering,

Mohamed Abou El-Fetouh Barakat Department of Environmental Sci-

Giuseppe Barbieri Institute on Membrane Technology, National Research

Onita D. Basu Department of Civil and Environmental Engineering, Carle-

Arnaud Baudot Physics and Analysis Division, IFP Energies Nouvelles,

Laurent Bazinet Department of Food Sciences, Universit Laval, Quebec,

Fabio Bazzarelli Institute on Membrane Technology, National Research

Katalin Bela-Bako Research Institute of Bioengineering, Membrane Tech-

Marie-Pierre Belleville Institut Europen des Membranes (IEM): UMR

Benny Danilo Belviso Istituto di Cristallograa, Consiglio Nazionale delle

Juana Benavente Departamento de Fisica Aplicada I, Universidad de

Nilay Bereli Department of Chemistry, Biochemistry Division, Hacettepe

Samuel Bernard Institut Europen des Membranes, ENSCM-UM2-CNRS,

A. Moura Bernardes Universidade Federal do Rio Grande do Sul

Paola Bernardo Institute on Membrane Technology, National Research

Richard Bernewitz Food Process Engineering, Karlsruhe Institute of Tech-

Marc Cretin Institut Europen des Membranes, ENSCM-UM-CNRS (UMR

Winnie Dejonghe Separation and Conversion Technology, Flemish Institute

Piotr Dydo Silesian University of Technology, Gliwice, Poland

Filomena Freitas UCIBIO-REQUIMTE, Departamento de Qumica,

Karel Friess Department of Physical Chemistry, University of Chemistry

Francesco Galiano Institute on Membrane Technology, National Research

Claudia F. Galinha REQUIMTE-CQFB, Chemistry Dept. Faculdade de

Fausto Gallucci Chemical Process Intensication, Department of Chemical

Volker Gaukel Food Process Engineering, Karlsruhe Institute of Technology

Abaynesh Yihdego Gebreyohannes Institute on Membrane Technology,

Geoffrey Geise Department of Chemical Engineering, University of

Lidietta Giorno Institute on Membrane Technology, National Research

Zbigniew J. Grzywna Department of Physical Chemistry and Technology of

Annarosa Gugliuzza Institute on Membrane Technology, National Research

Gemma Gutirrez Department of Chemical and Environmental Engineer-

May-Britt Hgg Department of Chemical Engineering, Norwegian Univer-

Vu Pha Hai Greentech Environment Joint Stock Company, Ho Chi Minh

Nobuo Hara Research Institute for Chemical Process Technology, National

Tao He Lab for Membrane Materials and Separation Technology, Shanghai

Xuezhong He Department of Chemical Engineering, Norwegian University

Arash Helmi Department of Chemical Engineering and Chemistry, Chemi-

Juhana Jaafar Faculty of Chemical and Energy Engineering, Universiti

Chalida Klaysom Centre for Surface Chemistry and Catalysis, KU Leuven,

P. Knauth Madirel UMR 7246, Aix Marseille University-CNRS, Marseille

Isao Kobayashi National Food Research Institute, National Agriculture and

N. M. Kocherginsky Biomime, Urbana, IL, USA

G. Kolb Product Sector Energy and Chemistry, Fraunhofer ICT-IMM,

A. B. Koltuniewicz Biotechnology and Bioprocess Engineering Division,