Professional Documents
Culture Documents
Lidietta Giorno
Editors
Encyclopedia of
Membranes
1 3Reference
Encyclopedia of Membranes
Enrico Drioli Lidietta Giorno
Editors
Encyclopedia of
Membranes
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:
Other elds that will exploit more and more membranes unique features
include:
Sensors
Micromanufacturing
Responsive materials
Safety and security
Space
Enrico Drioli
Lidietta Giorno
Acknowledgments
Enrico Drioli
Lidietta Giorno
vii
Editors Biography
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)
xiii
Editorial Board
xv
Contributors
xvii
xviii 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
Council of Italy, ITM-CNR, Rende, Italy
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
Council of Italy, ITM-CNR, Rende, Italy
Contributors xxi
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
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
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
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
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
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
Product
Water
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
N N
N N
2-Amino-4,6-trichloro-s-triazine Cl OH
N N
Cl N NH2
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
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.
J Chromatogr A 952:1327
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
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)
Fc
150
50
50
0 5 10 15 20 25
Time (min)
Aged Carbon Membrane 23
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
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
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
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
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
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
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
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
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
150
100
1 1.5 2 2.5 3
Concentration ratio, CV
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)
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
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
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
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
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
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).
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
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)
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
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.
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
Reactants retentate
Reactants retentate
A
packed catalyst catalytic
permeate membrane permeate
membrane
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
CF3 CF3
Cross-References
Amorphous Perfluoropolymers
Glass Transition Temperature (Tg)
Marco Avataneo Hyon AD
RD&T, Solvay Specialty Polymers, Bollate (MI),
Italy
References
deformation
A
glassy state rubbery state plastic state
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
References
Amphipathic Molecules
Bernardo P, Drioli E, Golemme G (2009) Membrane gas
separation. Ind Eng Chem Res 48:46384663 Amphiphilic Molecules
Amphiphilic Membrane 69
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
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
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
Analytical Pervaporation
A
Wanqin Jin
State Key Laboratory of Materials-Oriented
Chemical Engineering, Nanjing University of
Technology, Nanjing, China
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
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
+
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
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-
ings: recent developments in the design of surfaces that
pound in an agar plate covered by a bacterial lawn.
prevent fouling by proteins, bacteria, and marine organ-
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
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
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
Retentate
MICROMOLECULES SEPARATION
A
Pre-
FEED MF UF/ED
treatment
MACROSCOPIC PRE-TREATMENT
NF
RO
OD/MD
CONCENTRATION
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
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).
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
Apple
References
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
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
c
a
3. Amino functionalized
Aptamer
100 nm 10 nm HN
b 120 O
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
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
Lp Pf V w =RT
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
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
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
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
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
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
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
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
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)
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.
Institute on Membrane Technology, National
The analytic potential energy functions are
Research Council of Italy, ITM-CNR, Rende,
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
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
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
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
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
Bacterial Biofilm Formation 141
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
Bacterial Biofilm Formation 143
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
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
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
148 Basic Electrochemical Relations in Membrane Processes
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 Relations in Membrane Processes 149
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,
Fig. 3 Schematic drawing
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
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
Basic Electrochemical a b --
Relations in Membrane
Processes: Electron
+
+ + +
Conductivity, Ion
--
Conductivity, Ohms
+
--
--
+
B
+
Law, and Coulombs Law, --
+
Fig. 6 Schematic drawing
+
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
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
Bulk
Flux (B)
B
B B
Surface
Surface
Surface
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
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.
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
kCounts
500
1 4
400
300
200
5
2
100 3
25 50 75 100 minutes
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
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)
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
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
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
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
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
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
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
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
1. Retentate recycle
CONVERSION REACTOR
Reactants
Biocatalyst Membrane
Permeate
2.
CONVERSION REACTOR
Reactants
Retentate
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)
Outer surface
186 Biocatalytic Hollow Fiber Membrane Reactor
Enzyme solution
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
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
Research Institute of Chemical and Process
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
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
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
H2N N OH
H NH2
HO OH
O NH2
OH
NH2
HO
HS OH
NH2
OH OH
HO
NH2 NH2
N NH2
OH
N NH2
H
196 Biocatalytic Membrane Reactors with Chemically Bound Enzyme
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
O O
Cl Cl
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
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
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
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
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
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
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
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).
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
Proteins Polysaccharides
Pullulan
Gluten Cellulose
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
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
microorganism microorganism
microorganism
CO2
extracellular low-molecular
enzymes CH4 H2O
weight
intermediates metabolic
products
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
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):
a b c External resistance
Membrane Cathode
Anode
Organics
Substitute
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
Cell recycle
Product B
Separation 1
Separation 2
Bioreactor
enriched
liquid
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
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
Gasification
B
Bio Gas
Biogas, Fig. 2 Possible applications of biogas as energy resource (Modied from Basu et al. 2010)
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
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
Rasi S, Veijanen A, Rintala J (2007) Trace compounds of
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
Rasi S, Veijanen A, Rintala J (2007) Trace compounds of
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
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
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
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
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
COOCH3
COOH
CH3O
CH3O
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
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
ABA triblock
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
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
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
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
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
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)
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
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.
Desalination 205:185191
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
Boron Removal by Reverse Osmosis
190:7178
References
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
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
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
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 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
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
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
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
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
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
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
J Membr Sci 487:6073
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)
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
Carbon Porous
Membranes Modified
with Enzymes,
Fig. 1 Photography of
carbon tube (photo IEM)
306 CarmanKozeny Equation
Dp KV o m 1 e2
2 2 (1)
L 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.
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
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
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
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)
Casting, Fig. 1 Procedure for casting at sheet membranes at lab scale and methods for inducing phase separation
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
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
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-
Mulder M (1991) Basic principles of membrane technol-
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 pro- of 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
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.
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
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
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
Catalytically Modified
(Activated) MIEC
Membranes,
Fig. 1 Different sections
involved in oxygen
transport during oxygen
permeation (Sunarso
et al. 2008)
Catalytically Modified
(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
Catalytically Modified
(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
branes. J Membr Sci 377:198205
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 pro- of 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
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
industry. Taylor & Francis, London
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
Institute on Membrane Technology, National
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
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 pro- of 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
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
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
O
R1 (R2, R3) = Cellulose acetate methyl carbonate
O
O
R1 (R2, R3) = Cellulose acetate methoxy acetate
O
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
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
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
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
(NitrosepTM, MTR Inc.)
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
Composition (% mol.)
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
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
Al2O3 support
5 m 5 m
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 365
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.
366 Characterization of Porous and Dense Membranes
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:
Characterization of Porous and Dense Membranes 367
Characterization force
of Porous and Dense
Membranes,
Fig. 3 Schematic drawing rp
that illustrates the capillary
effect of a membrane pore
wetting liquid
Characterization
of Porous and Dense minimum pore size
Membranes,
Fig. 4 Schematic drawing
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
370 Characterization of Porous and Dense Membranes
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 371
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
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)
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
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
Chemical and Electrochemical Equilibrium in Membrane Systems 377
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
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.
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 con- ical 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
Chemical Cleaning of Membranes 381
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
4d 8d 16d 32d
a
Feed flow
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 of Membranes 383
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.
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
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
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)
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
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+
392 Chemical Pretreatment
Inorganic coagulants
Inorganic polymeric coagulants
Organic polymeric coagulants
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
394 Chemical Pretreatment
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
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,
Fig. 1 Schematic diagram
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 mem- branes 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
(S)-isomer (acid)
Aqueous Phase (lumen side)
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
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
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
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:
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
- 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
- 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
408 Chromium Transport Through Ion Exchange Membranes
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
2 100
Aqueous
[Cr+6](mol/m3) stripping
80 Phase
1,5 Aqueous
[Cr+6](mol/m3)
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)
Chromium Transport Through Ion Exchange Membranes 409
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
(kLi)
KL,Chloride 0.317 . (Q1/3
2
)(m/h) Calculated
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 Transport Through Ion Exchange Membranes, Fig. 6 Optimal process conguration for the emulsion
phase
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 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.
J Membr Sci 191:129142
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
Rinsing
Chemical cleaning 1
Chemical cleaning 2
Sanitizing
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
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-
Mulder M (1996) Basic principles of membrane technol-
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,
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
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
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
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
Membrane phase
+2 CoR2(HR)2 2H+ 2H+
Co
2H+
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
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
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
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
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
3100 3450
3000 3400
2 3 4 5 6 7 8 9
pH filler slurry
b 1650
400
H2O bending vibration / cm-1
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
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
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
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
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
xi
Bulk
Feed
(0, 1) Wang et al. (2002)
xFilm
1 Bulk i
C
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
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
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,
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
Zoltn Kovcs
Department of Food Engineering, Institue of
Bioengineering and Process Engineering, Szent
Istvan University, Budapest, Hungary
a b
Retentate Retentate
Diluant Diluant
Booster
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
Research Council, ITM-CNR, Rende, Italy
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
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
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
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
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
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
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
Flow equalization
Microfiltration
Ultrafiltration
Organic Matter,
Excess of Chlorine, Adsorption (activated carbon)
Metals
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
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.
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
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
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
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
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
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)
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
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
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
2nd edn. Wiley, Chichester
(Basu et al. 2010). Compared with the covalent
Mulder M (1997) Basic principles of membrane technol-
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.
J Membr Sci 155:6778
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
Cross-Linked Layer-by-
Layer Membranes,
Fig. 1 Schematic diagram
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
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
100
10-1
10-2 10-1 100 101 102 103 104
CO2 Permeability [Barrer]
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
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.
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
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
a b
Pressure
Free energy
c
Free energy
A form
B form
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.
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
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
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
Molecular C
weight 0.5 50 7000
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
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
n
d2 t dt
2
k (1)
dV dV
Intermediate 1 dt
dV Q1 f t Pore blocking and surface
blocking deposit
Complete blocking 2 dV
dt Q f V Pore blocking
Dealcoholization by Pervaporation 511
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
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
Institute on Membrane Technology, National
Research Council of Italy, ITM-CNR, Rende,
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
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 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
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
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
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
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
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
Mulder M (1996) Basic principles of membrane technol-
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
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
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-
Mulder M (1996) Basic principles of membrane technol-
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
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
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)
(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:
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).
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
DIALYSIS
FEED SOLUTION FLOW
FEED SOLUTION
MEMBRANE
DIALYSATE
DIALYSATE FLOW
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
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-
ogy. Kluwer, Dordrecht
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
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:
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 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.
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)
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
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
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
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
c c0 e1sV d =V s
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.
Institute on Membrane Technology, National
Heat and mass transfer phenomena occurring
Research Council of Italy, ITM-CNR, Rende,
in DCMD process are discussed in various recent
Italy
publications (Lawson and Lloyd 1997; Drioli
Department of Environmental and Chemical
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
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
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
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)
KHSO4
Solvent
Solvent DESALTING
recovery 2KHSO4 K2SO4+H2SO4
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
References
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)
0.2
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Current density / A cm2
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
a b
Interface
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:
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
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)
576 Donnan Equilibrium
Initial state
SideI SideII 50 120
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
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
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
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
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
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
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
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 reservoir
Adhesive
Skin Membrane
(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
Complutense de Madrid, Madrid, Spain
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
D
Spinneret Spinneret
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-
ciples and applications. Elsevier, Amsterdam
ent from that determined from water per-
Mulder M (1992) Basic principles of membrane technol-
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
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
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))
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.
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 (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)
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
(continued)
Dynamic Membrane Emulsification, Table 1 (continued)
Dynamic Membrane Emulsification
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
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
:
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
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
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
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
Fig. 1 Schematic diagram
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
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.
Desalination 28:3140
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
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
Electrodriven Membrane
Processes, Fig. 1 An
Electrodialysis Cell
Electrodriven Membrane
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,
638 Electrodriven Membrane Processes
Electrodriven Membrane
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.
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
640 Electrodriven Membrane Processes
Electrodriven Membrane
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
Electrodriven Membrane Processes 641
Electrodriven Membrane
Processes, Fig. 6 Electro-
Ion Substitution Cell
Electrodriven Membrane
Processes, Fig. 7 A Salt
Methathesis Cell
Electrodriven Membrane
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
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
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
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
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
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
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
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
Electroosmotic Flow
Cross-References
Electroosmosis, Overview of
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
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
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
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
(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
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.
J Membr Sci 376:241246
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
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
Proton-Exchange Membranes for Fuel Cells
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
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 mem- ber 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
Proton-Exchange Membranes for Fuel Cells
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
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 by Membrane Operations, Table 1 Different membrane operations used for the production emulsion
droplets
Advantages Disadvantages
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
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
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
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-
Institute on Membrane Technology, National
ionic functional groups at their head, such as
Research Council of Italy, ITM-CNR, Rende,
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
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 condi- tion 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
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.
J Membr Sci 84:201211
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
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.
688 Emulsion Pertraction Technology for Zinc Recovery
Emulsion Pertraction
Technology for Zinc
Recovery,
Fig. 2 Characteristics of
hollow ber contactors for a
bench scale EPT process
Parameter/Characteristic Value/Description
Thickness 30 m
Porosity 40%
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
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
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
690 Emulsion Pertraction Technology for Zinc Recovery
0
0 30 60 90 120 150 180
t (min)
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]
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
References
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)
Feed UF
Process
tank
tank
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
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
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
Pirkle W, Bowen W (1994) Preparative separation of enan-
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
enantioselective afnity toward amino acids. Yoshikawa M, Izumi J, Kitano T (1995) Molecularly
Cyclodextrins, bovine serum albumin (BSA), imprinted polymeric membranes for optical resolution.
J Membr Sci 108:171175
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
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
ln 1 c 1 eep macromolecules.
E
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
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 (EOR), Fig. 1 EOR example of permeation application (Spillman 1989)
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
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
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
industry. Taylor & Francis, London
(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
Permeate :
1.0 XCO2 SCF + Dispersed molecules
SCF
Retentate Pressure
Viscous
SCF (recycled)
liquid Membrane 20 MPa
injection
separation
chamber
Permeate :
Viscous liquid + Dissolved SCF
SCF
0.2 XCO2
Viscous liquid
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
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
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
Electroosmosis, Overview of
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
Sci Technol 46:234246
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-
mization of preparation conditions for polydimethyl-
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
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
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 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
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
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 perme- ation 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
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
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)
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
Institute on Membrane Technology, National
Research Council of Italy, ITM-CNR, Rende,
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)
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
Extracorporeal Blood Oxygenation Devices, Membranes for 739
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
740 Extracorporeal Blood Oxygenation Devices, Membranes for
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
742 Extracorporeal Blood Oxygenation Devices, Membranes for
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
Extracorporeal Blood Oxygenation Devices, Membranes for 743
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.
744 Extracorporeal Blood Oxygenation Devices, Membranes for
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
Extracorporeal Blood Oxygenation Devices, Membranes for 745
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
746 Extracorporeal Blood Oxygenation Devices, Membranes for
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
Extracorporeal Blood Oxygenation Devices, Membranes for 747
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).
748 Extracorporeal Blood Oxygenation Devices, Membranes for
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,
Extracorporeal Blood Oxygenation Devices, Membranes for 749
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).
Extracorporeal Blood Oxygenation Devices, Membranes for 751
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
Extracorporeal Blood Oxygenation Devices, Membranes for 753
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
754 Extracorporeal Blood Oxygenation Devices, Membranes for
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
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
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
E
F
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
References
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
@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
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
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
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
J Membr Sci 175:181196
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 con- tains 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
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)
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
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.
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
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 inter- phase 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
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
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
Forward Osmosis (FO) 799
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.
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
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
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
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
membrane fouling
time
Fouling 809
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. 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:
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
agitating agitating
agitating
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)
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
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
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
Institute on Membrane Technology, National
also demonstrated the addition of germanium into
Research Council of Italy, ITM-CNR, Rende,
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
F
828 Free Protons in Solid Heteropoly Compounds
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 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,
0.2
2 3 4 5 6 7 8
R,
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)
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
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
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
mization of preparation conditions for polydimethyl-
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
E E E E E
E Physical Adsorption
E E E E E E
E
Covalent binding
Entrapment E E
Cross-linking
PEI GA
E GA: Glutaraldehyde;
GA
PEI: Polyethyelenimine; E: Enzyme
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
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)
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)
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
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
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
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 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:
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
866 Gas/Vapor Transport
Transport in Heterogeneous
and Homogeneous Mixtures
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
870 Gas/Vapor Transport
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
Cliquid
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
Genotoxin Removal,
Fig. 1 Scheme of
genotoxin removal by OSN
with potential reverse
osmosis for solvent
recycling
References
Germanium Recovery by Ion Exchange Membranes, Fig. 1 Germanium recovery system using catechol complex-
ation and membrane permeation
Germanium Recovery by Ion Exchange Membranes, Fig. 3 Breakthrough curves of germanium recovery at pH of
11.0 and 2.0
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
Cathode Anode
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
Grain Models,
Fig. 1 Ballistic
reconstruction followed by
partial sintering
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
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
Granular Media Pretreatment Filtration 891
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.
892 Granular Media Pretreatment Filtration
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
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
896 Graphene Membranes
b
a
pz orbital pz orbital
!120!
bond
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).
Graphene Membranes 897
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
898 Graphene Membranes
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.
Graphene Membranes 899
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
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
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
Vapour
Suction and
condenstation Porous membrane
Permeate
Aqueous solution (warm)
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).
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 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).
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
Heterogeneous Ion-Exchange Membranes, Fig. 1 Schematic diagram of heterogeneous IEMs with various kinds of
shapes
References
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
Strathmann H, Giorno L, Drioli E (2006) An introduction
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
Fig. 1 Schematic view of
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
(1)
175-230C 230-580C 580-604C 604-1000C
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
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
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
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
10
1
1 10 100 1000 10000 100000
H2 Permeability (Barrer)
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
Highly Permeable Polyimides, Fig. 1 6FDA + 4MPDA-based polyimide (Lin et al. 2000)
932 Highly Permeable Polyimides
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)
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
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.
J Membr Sci 250:305310
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
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
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
Feed
Retentate
Permeate
Hollow Fiber Carbon Molecular Sieve Membranes, Fig. 2 A schematic and lab-scale carbon membrane module
(He and Hgg 2011b)
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:
branes. J Membr Sci 390391:2331
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
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
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
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
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
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
Permeate
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-
Mulder M (1997) Basic principles of membrane technol-
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
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 959
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
960 Hollow Fiber Modules for Desalination
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
962 Hollow Fiber Modules for Desalination
Th
hollow fiber membrane
polypropylene net
heat exchange
hollow fiber Hot
Feed
Cold Cold
Feed Feed
Hollow Fiber Modules for Desalination, Fig. 6 Schematic presentation of the novel AGMD module
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
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
+ + + +
Khayet M, Matsuura T (2011) Membrane distillation: prin-
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
ogy. Kluwer, Dordrecht
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
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
Hybrid Membranes
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
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 com- bined 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
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
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
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
O O O O O O O O O O
O O O O Co O O
Co Co
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
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
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 (-)
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
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
H < 0
Substrate
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
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
Gel
Depolimerization of aluminosilicate
chains
Primary Building
Units
Secondary Building
Units
Polyhedra
Zeolite
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
Hyflon AD,
Fig. 2 Density, glass
transition, and refractive
index for the two
commercial grades of
Hyon AD
1010 Hyflon AD Membranes
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
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
H
HO OH C
O O
O O
O
N N
H
O O C
1014 Hyperbranched Polyimides
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
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
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
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
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
Impedance Spectroscopy
Veronica Silva
Dow Water and Process Solution division at The
Dow Chemical Company, Boulevard Cecilia
Grierson 355, Buenos Aires, Argentina
Synonyms
Application to Membrane
Characterization
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)
a b
Rct Rm+s Redl Rdbl
Rm
C C
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, 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
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
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 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.)
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
Synthesis of Removal by
Polymerizable Chemical Molecular
Print Molecule Polymerization Cleavage 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
Fig. 2 Scheme of the
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
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
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
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
b
a polymeric sol (Ti-oxide,
pore size 1-2 nm)
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
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
drying
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
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,
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
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
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
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.
1054 Ion-Exchange Membrane Characterization
volt meter
calomel electrode
electrode electrode
solution A A C solution
gas gas
(O2) (H2)
anode cathode
+ porous plug -
- - - +
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,
Fig. 2 Schematic drawing
of a test cell used for
membrane resistance
measurements using
alternating current
membrane
dm
I
electrolyte solution
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
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,
Fig. 1 Schematic drawing
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
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
2 300
250
[Cm STP cm/cm cmHg s]
4
3
200
Water (PPO membrane)
20
2
150
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
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
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 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
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)
Membrane operations, innovative separations and
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)
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, 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
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
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:
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,
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
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
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
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
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
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)
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
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
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
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
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
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
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)
Organic Fouling
Precipitate Fouling
Colloidal Fouling
Membrane fouling
Biofouling
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
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
L
M
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
*
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
dead-end
cross-flow with optimum
surface renewal