Original Report
High-Pressure/High-
Temperature Microreactors
for Nanostructure Synthesis
Samuel Marre,1 Jinyoung Baek,1 Jongnam Park,2 Moungi G. Bawendi,2
1
Department of Chemical Engineering, Massachusetts Institute of Technology,
2
Department of Chemistry, Massachusetts Institute of Technology, Cambridge, MA
hemically synthesized nanomaterials, such as
Keywords:
microfluidics,
C nanocrystalline quantum dots, are being considered
as the active element in many applications, including
high pressure, photovoltaics, displays, and biochem sensing. To realize
synthesis, the promise of these devices, it will be critical to have an
nanoparticles, efficient, reproducible synthesis technique of the
quantum dots
nanostructures. Currently, nanoparticles are synthesized
in a batch mode in small volumes, which is appropriate for
studying the fundamental properties of nanosized
structures and for developing proof of principle device
structures. However, batch synthesis suffers from control
of size, size distribution, and quality of the nanomaterial
from batch to batch. Moreover, there is an inherent
difficulty in scaling up to quantities more reasonable for
device development and optimization. Continuous-flow
reactors based on microfluidics (microreactors) integrated
with heaters and fluid control elements offer a solution to
these problems and additional advantages. We describe
continuous synthesis of nanostructures in microfluidic
systems consisting of multiple subemillimeter-sized
channels in which fluid flows continuously and chemical
reactions take place. The small reaction volumes
*Correspondence: Klavs F. Jensen, Ph.D., Department of Chemical
Engineering, Massachusetts Institute of Technology, Room 66-
350, 77 Massachusetts Avenue, Cambridge, MA 02139, USA;
Phone: 1.617.253.4589; Fax: 1.617.258.8992; E-mail:
kfjensen@mit.edu
1535-5535/$36.00
Copyright c
 2009 by The Association for Laboratory Automation
doi:10.1016/j.jala.2009.06.005
Klavs Jensen, Ph.D.
Massachusetts Institute of
Technology
Cambridge, MA
and Klavs F. Jensen1*
Cambridge, MA
combined with the high heat and mass transfer rates
enable reactions to be performed under more controlled
conditions with higher yields than can typically be achieved
with conventional reactors. Moreover, manipulation of
reaction parameters, while the reaction proceeds, allows
optimization of synthesis conditions. The ability to work at
elevated temperatures and pressures while confining
potentially toxic, high reactive starting materials will
become important for the synthesis of novel
nanostructured materials. ( JALA 2009;14:36773)
INTRODUCTION
Chemically synthesized nanomaterials are being
considered as the active elements in many applica-
tions, including photovoltaic, energy, displays,1 and
biochem sensing.2 To realize the promise of these
devices, it will be critical to have an efcient, repro-
ducible synthesis technique of the nanostructures.
Currently, nanoparticles are synthesized in a batch
mode in small volumes, which is appropriate for
studying the fundamental properties of nanosized
structures and for developing proof of principle
device structures. However, batch synthesis suffers
from irreproducibility of size, size distribution, and
quality of the nanomaterial from batch to batch, as
well as fast screening and optimization of the synthesis
conditions. Moreover, there is inherent difculty in
scaling up to quantities more reasonable for device
development and optimization.
JALA December 2009 367
Original Report
Continuous-ow reactors based on microuidics (micro-
reactors) integrated with heaters and uid control elements
oer a solution to these problems as well as additional ad-
vantages, including enhancement of mass and heat transfer,
feedback control of temperature and feed streams, reproduc-
ibility, potential for sensor integration for in situ reaction
monitoring, rapid screening of parameters, and low reagent
consumption during optimization.3e5 The small reaction vol-
umes combined with the high heat and mass transfer rates
enable reactions to be performed under more aggressive con-
ditions with higher yields than can typically be achieved with
conventional reactors. Moreover, new reaction pathways
deemed too difcult to control in conventional macroscopic
equipment can be conducted safely because of the high heat
transfer and ease of conning small volumes. This ability to
work at elevated temperatures and pressures while conning
potentially toxic, high reactive starting materials could be-
come important for the synthesis of novel nanostructured
materials.
The development of high-temperature microuidic reac-
tors for inorganic nanocrystal (NC) synthesis has been dem-
onstrated in several studies with the continuous synthesis of
NCs at atmospheric pressure using a single-phase laminar
ow microreactor designs.6 However, one difculty of these
synthetic procedures is the requirement for solvents that
can both dissolve the precursors at ambient conditions and
also remain liquid over the entire operating temperature
range (25e350  C), signicantly limiting the set of solvents,
ligands, and precursors that are compatible with continu-
ous-ow systems. Furthermore, solvents that are available,
are typically very viscous, leading to slow mixing, broad res-
idence time distributions (RTDs), and as a consequence,
broad NC size distributions. Segmentation of the reacting
phase with an immiscible phase7 can address some of these
limitations by narrowing the RTD and improving reactant
mixing.8,9 Application of segmented ow for continuous syn-
thesis of inorganic NCs has been previously demonstrated
for liquidegas,4 and liquideliquid10 segmented ows.
Figure 1. Pressure/temperature diagram with typical batch process conditions and potential novel on chip processes using high-pressure/
high-temperature microfluidic systems.
368 JALA December 2009
However, even with ow segmentation, limitations on the
number of compatible chemistries and the limited number
of available high boiling point solvents have been a major
obstacle in the rapid adoption of microreactors as a universal
platform for nanostructure synthesis.
One avenue of improving synthesis in microreactors is to
perform experiments at high pressure (Fig. 1). High pressure
allows for wider range of chemistries and processes, because
at sufciently high pressure, virtually any common solvent,
precursor, and ligand will remain either liquid or become su-
percritical (sc) at the temperatures required for nanomaterial
synthesis. This particular homogeneous domain (supercriti-
cal) of the pressure/temperature phase diagram of compound
is located above the critical coordinates (Tc and pc) of the
considered uid or mixture. The supercritical regime allows
tuning the uid properties from liquid-like to gas-like, dis-
playing the high miscibility and fast diffusion rates typical
of gasses, but with sufciently high densities for solubilizing
a wide range of compounds of low-to-medium polarity,
typical of liquids.
The main challenge for adoption and development of high-
pressure/high-temperature (HP/HT) microuidic is the uidic
connections (packaging). Several authors have recently
reported several ways to connect tubing to chips enabling
medium pressure use in microreactors,11 but these packaging
techniques are not suitable for high-temperature processes,
though their use was demonstrated for low temperature pro-
cesses using supercritical CO2 as medium to perform catalysis
and chemical reactions.12 However, nanostructure synthesis
often requires higher temperature conditions, which are not
feasible in those microsystems.
In this contribution, we present our strategy to develop
high pressure microuidic devices, which can be further in-
terfaced with online characterization techniques coupled to
optimization software, allowing for fast screening and opti-
mization of processes, especially for inorganic nanostructure
synthesis. Proofs of concept are illustrated with the synthesis
of semiconductor NCs in various solvents.
 Original Report

Figure 2. Top: High-pressure microreactors for nanomaterial synthesis: (a) 45-mL microreactor, (b) 100-mL microreactor, and (c) modular
high-pressure fluidic connections. Bottom: High-pressure experimental setup, including a high-pressure high-temperature microreactor,
a compression-cooling aluminum interface, two high-pressure syringe pumps, and a back-pressure regulator.

MICROFABRICATION AND EXPERIMENTAL SETUP
To overcome the above-mentioned HP/HT limitations, we
have developed high-pressure high-temperature continuous-
ow silicon-Pyrex microreactors (Fig. 2) for the synthesis of
nanomaterials. Well-established wet etching (with potassium
hydroxide) and dry etching (deep reactive ion etching) proce-
dures enable fabrication of microchannels with controlled
sidewall shape and channel dimensions from tens to hundreds
of micrometers.3 Silicon has excellent mechanical strength and
temperature characteristics, and it has good chemical compat-
ibility characteristics. Oxidation of silicon through
high-temperature wet oxidation process in a furnace leads to
the formation of a glass layer on the surface so that an oxidized
silicon microreactor becomes functionally equivalent to a glass
reactor. This allows for further surface functionalization with
the wide available silane chemistry for controlling the wettabil-
ity of the walls from hydrophilic to hydrophobic.13
After forming channels in the silicon and oxidizing the
surface, the patterned substrate was anodically bonded to

JALA December 2009 369

Original Report
Figure 3. (a) Example of characterization techniques that can be used with the high-pressure/high-temperature (HP/HT) microreactors.
(b) General scheme of an online HP/HT optimization setup.
Pyrex for easy optical access. The microreactors (Fig. 2a, b)
consisted of an injection zone maintained at room tempera-
ture and in a reaction zone under temperature. The two
zones were separated by a thermally isolating halo etch that
allowed for a temperature gradient of 25  C/mm. The pack-
aging was realized using high-pressure modular compression
uidic connections by compressing the microreactor between
two stainless steel parts using viton O-rings. In this congu-
ration, the setup enabled reaching high pressure (upto 300
bar) and temperature (upto 450  C in the heated section)
(Fig. 2c). Typical dimensions for the microchannels were
100e400 mm for the width and 100e300 mm for the depth,
370 JALA December 2009
with length 0.1e1 m, leading to volume ranging from w1
to 100 mL.
From this initial conguration, characterization techniques
can be further integrated, allowing control of reaction param-
eters as well as collecting time- and space-resolved informa-
tion. Potential characterization techniques include physical
measurements (thermocouples and piezo electric pressure
monitoring) and chemical measurements via integration of
optic bers for in situ spectroscopy ultraviolet-visible (UV-
vis), infrared (IR), Raman14 (Fig. 3). Conventional online
characterization techniques, well developed for microuidic
applications,15 could be used as well, such as microparticle
 Original Report

image velocimetry9 or dynamic light scattering,16 allowing

measurement of ow proles and particles size distribution
(PSD), respectively.
The use of this well characterized and controlled setup
with optimization software could provide useful laboratory
scale tools for processes optimization and fast screening of
reaction conditions, similarly to what can be done in conven-
tional microuidic conditions.17 In particular, pressure could
be screened as a new critical parameter for high-temperature
nanoparticle synthesis.

APPLICATIONS TO QUANTUM DOT SYNTHESIS

Applications of HP/HT microuidic systems are illustrated

with the synthesis of model nanomaterials, with CdSe quan-
tum dots (QDs) at conditions upto 300  C and 100 bar. Col-
loidal semiconductor QDs are interesting nanomaterials with
useful properties for biological imaging, electroluminescent
devices, and lasers due to their size tunable optical properties
and chemical stability. Most applications require highly crys-
talline samples with narrow size distributions, which are of-
ten difcult to achieve in a single step batch process with
its often-limited reaction conditions. Synthesis in microreac-
tor at high temperatures and pressures provides a mean for
continuous production and optimization of the QD proper-
ties beyond those achieved with typical batch processes.
In this study, synthesis conditions were adapted from
batch mode, replacing the usual high boiling point solvent
(squalane or octadecene) by cheaper solvents (hexane, tolu-
ene, octane, dodecane). CdSe QDs were synthesized by mix-
ing solutions of cadmium oleate (prepared from the reaction
of cadmium acetylacetonate with oleic acid at 120  C for
30 min) and trioctylphosphine selenide in the different sol-
vents at pressure ranging from 50 to 100 bar, while tuning
several parameters, such as temperature (180e350  C), resi-
dence time (12e3000 s), and concentration (3.2 103 M to
7.4 102 M). The easy manipulation of ows at high pressure
inside the microsystem allows for collectingmany samples in
a short period of time, leading to fast screening of the
operational parameters (Fig. 4a).
Photoluminescence (PL) from UV-excited QD solutions
reveals poor quality of QDs synthesized at low temperature
(T 180  C). The weak bluish-white emission results primar-
ily from electron-hole recombination at defects. These de-
fects on the QD surface stem from poor crystallization and Figure 4. (a) Optical characterization of different solutions of CdSe
unoptimized growth. Temperature (Fig. 4b) and residence under UV irradiation of the as-prepared CdSe quantum dots (QDs)
time (Fig. 4c) have a strong effect on the average size of in liquid Toluene (75 bar) for 180  C ! T ! 310  C, 12 s ! resi-
the CdSe NCs, as determined using the wavelength at the dence time ! 8.5 min at 50 bar with [Cd] [Se] 3.8  103 M.
emission maximum.18 The residence time effect tends to be (b) Evolution of the size of CdSe QDs synthesized in dodecane at
less pronounced at long values, with the size reaching a satu- 100 bar (a) as a function of the temperature (residence time:
ration value due to the nearly complete consumption of tR 30 s) and (c) as a function of residence time (T 270  C), with
precursors in the reaction media. additional optical images of the CdSe solutions under UV irradiation.
Starting from this demonstration of QD synthesis in HP/
HT microreactors, we explored the eect of solvent on the varying the length of the carbon chain (hexane, octane, do-
optical quality of the synthesized semiconductor NCs. We decane, squalane, and squalane with addition of polybutadi-
compared hydrocarbon solvents of dierent viscosities, ene [PB] to increase viscosity). All experiments were run

JALA December 2009 371

Original Report
Figure 5. Photoluminescence curves of CdSe solutions under
UV irradiation (l 365 nm) prepared in supercritical hexane,
supercritcial octane, liquid dodecane, squalane, and squala-
ne polybutadiene (PB), respectively, at T 310  C and 50 bar
with [Cd] [Se] 3.8  103 M.
under the same conditions (T 310  C and p 50 bar,
[Cd] [Se] 3.8  103 M). Dodecane, squalane, and squa-
lane PB were in standard liquid phase, whereas octane
and hexane were under supercritical conditions (sc-Hexane:
Tc 234.7  C and pc 30.3 bar; sc-Octane: Tc 296.2  C
and pc 25 bar).
The PL curves from QDs synthesized in the dierent sol-
vents show two main eects: (1) a shift of the maximum emis-
sion going from viscous long carbon chain solvents to the
smaller solvent molecule and (2) narrowing of the PL curves
as measured by a decrease in the full width half maximum
(Fig. 5).
The shift in the maximum emission can be attributed to
the phase in which the reaction takes place. No dierence
is observed for standard liquid solvent conditions, whereas
the QDs synthesized in supercritical uids have smaller sizes,
and correspondingly, the maximum PL emission is shifted
Figure 6. TEM images of CdSe quantum dots synthesized from sc-hexane at T 250  C and 80 bar for a residence time of 30 s, with
[Cd] [Se] 3.8  103 M. The average size is 3.8  0.2 nm.
372 JALA December 2009
toward the blue. The concentration of nuclei in the media
was estimated from the optical absorption (not shown) and
PL emission spectra. These data demonstrate that higher
numbers of QDs are synthesized in sc-hexane than in sc-oc-
tane than in all the liquid-phase solvents. The smaller sizes
can be attributed to the larger number of NCs, assuming that
the reaction yield remains constant and independent of the
solvent system. This result is consistent with experience from
batch synthesis of nanomaterials in supercritical uids, where
high nucleation rates often are reached.19
The variation in width of the distribution results primarily
from a hydrodynamic eect. In microuidics systems, a broad
RTDdtypical of single-phase viscous owsdleads to a wide
PSD for the synthesized nanomaterial (and thus wide PL
peaks in the case of QDs). Being able to work at elevated
pressure allows for the use of low boiling point solvents,
which are typically less viscous20 with a correspondingly nar-
rower RTD. Additional transmission electron microscopy
(TEM) measurements performed on CdSe QDs synthesized
in sc-hexane (Fig. 6) further showed a narrow size distribu-
tion of the semiconductor NCs as a consequence of the prop-
erties of supercritical uid ows in microreactors.
CONCLUSION
Materials Science requires ecient methods of synthesizing
nanomaterials with controlled size, morphology, composi-
tion, and structure. We have demonstrated that the use of
high-pressure high-temperature microreactors can open
new routes for nanomaterial synthesis in microuidic devices
by enlarging the set of solvents and phases (liquid, gas, super-
critical) available toward new and/or better nanomaterials.
In addition, the supercritical microuidics provide methods
to obtain better control on the size distribution of the synthe-
sized materials compared with standard liquid-phase condi-
tions. The high-pressure microuidic devices can be
interfaced with online characterization techniques coupled
to optimization software, allowing for fast screening and

optimization of processes, making this tool powerful for de-
veloping high-quality nanoparticles for a wide range of
promising applications, including displays, medical imaging,
and solar cells.
ACKNOWLEDGMENTS
This research was funded in part by the NSF and the US army through the
Institute for Soldier Nanotechnology. J. Park was supported by the Korean
research Foundation Grant funded by the Korean Government (KRF-2006-
214-D00038).
REFERENCES
1. Coe, S.; Woo, W. K.; Bawendi, M. G.; Bulovic, V. Electroluminescence
from single monolayers of nanocrystals in molecular organic devices.
Nature 2002, 420, 800e803.
2. (a). Bruschez, M. Jr.; Moronne, M.; Gin, P.; Weiss, S.; Alivisatos, A. P.
Semiconductor nanocrystals as uorescent biological labels. Science
1998, 281, 2013e2016.
(b). Liu, W.; Howarth, M.; Greytak, A. B.; Zheng, Y.; Nocera, D. G.; Ting,
A. Y.; Bawendi, M. G. Compact Biocompatible Quantum Dots Func-
tionalized for Cellular Imaging. J. Am. Chem. Soc. 1998, 130(4),
1274e1284.
3. Jensen, K. F. Silicon based microchemical systemsdCharacteristics and
applications. Mater. Res. Soc. Bull. 2006, 31, 101e107.
4. Yen, B. K.; Gunther, A.; Schmidt, M. A.; Jensen, K. F.; Bawendi, M. G.
Microfabricated gas-liquid segmented ow reactor for high temperature
synthesis: the case of CdSe quantum dots. Ang. Chem. Int. Ed. 2005, 44,
5447e5451.
5. Khan, S. A.; Jensen, K. F. Microuidic synthesis of titania shells on col-
loidal silica. Adv. Mater. 2007, 19, 2556e2560.
6. (a). Yen, B. K. H.; Stott, N. E.; Jensen, K. F.; Bawendi, M. G. A con-
tinuous-ow microcapillary reactor for the preparation of a size se-
ries of CdSe nanocrystals. Adv. Mater. 2003, 15, 1858e1861
(b). Nakamura, H.; Yamaguchi, Y.; Miyasaki, M.; Uehara, M.; Mae-
da, H.; Mulvaney, P. Preparation of CdSe nanocrystals in a micro-
ow reactor. Chem. Commun. 2003, 7(23), 2844e2845.
7. Song, H.; Tice, J. D.; Ismagilov, R. F. Microuidic system for controlling re-
action networks in time. Ang. Chem. Int. Ed. 2003, 42, 768e772.
8. Gunther, A.; Khan, S.; Thalmann, M.; Trachsel, F.; Jensen, K. F. Trans-
port and reaction in microscale segmented gas liquid ow. Lab Chip 2004,
4, 278.
9. Gunther, A.; Jhunjhunwala, M.; Thalmann, M.; Schmidt, M. A.; Jensen,
K. F. Micromixing of Miscible Liquids in Segmented Gas-Liquid Flow.
Langmuir 2005, 21, 1547.
10. Chan, E. M.; Alivisatos, A. P.; Mathies, R. A. High-
temperature microuidic synthesis of CdSe nanocrystals in nanoliter
droplets. J. Am. Chem. Soc. 2005, 127, 13854e13861.
11. (a). Pattekar, A. V.; Kothare, M. V. Novel microuidic interconnectors
for high temperature and pressure applications. J. Micromech. Mi-
croeng. 2003, 13(2), 337e345
Original Report
(b). Tiggelaar, R. M.; Benito-Lopez, F.; Hermes, D. C.; Rathgen, H.;
Egberink, R. J. M.; Mugele, F. G.; Reinhoudt, D. N.; van den Berg,
A.; Werboom, W.; Gardeniers, H. J. G. E. Fabrication, mechanical
testing and application of high-pressure glass microreactor chips.
Chem. Eng. J. 2003, 131, 163e170
(c). Murphy, E. R.; Inoue, T.; Sahoo, H. R.; Zaborenko, N.; Jensen, K.
F. Solder-based chip-to-tube and chip-to-chip packaging for micro-
uidic devices. Lab Chip 2003, 7, 1309e1314
(d). Trachsel, F.; Hutter, C.; Rudolf von Rohr, P. Transparent silicon/
glass microreactor for high-pressure and high-temperature reac-
tions. Chem. Eng. J. 2003, 135S, S309eS316.
12. (a). Benito-Lopez, F.; Tiggelaar, R. M.; Sabut, K.; Huskens, J.; Egber-
ink, R. J. M.; Reinhoudt, D. N.; Gardeniers, H. J. G. E.; Verboom,
W. Substantial rate enhancements of the esterication reaction of
phthalic anhydride with methanol at high pressure and using super-
critical CO2 as a co-solvent in a glass microreactor. Lab Chip 2007,
7, 1345e1351
(b). Trachsel, F.; Tidona, B.; Desportes, S.; Rudolph von Rohr, P. Solid
catalyzed hydrogenation in a Si/glass microreactor using supercrit-
ical CO2 as the reaction solvent. J. Supercrit. Fluids 2007, 48(2),
146e153.
13. Srinivasan, U.; Houston, M. R.; Howe, R. T.; Maboudian, R. Alkyltri-
chlorosilane-based self-assembled monolayer lms for stiction reduction
in silicon micromachines. J. Microelectromech. Syst. 1998, 7(2), 252e260.
14. Benito-Lopez, F.; Verboom, W.; Kakuta, M.; Gardeniers, H. J. G. E.;
Egberink, R. J. M.; Ooseterbroeck, E. R.; Van der Berg, A.; Reinhoudt,
D. N. Optical ber-based on-line UV/Vis spectroscopic monitoring of
chemical reaction kinetics under high pressure in a capillary microreac-
tor. Chem. Commun. 2005, 2857e2859.
15. (a). Tong, L. M.; Rhigini, M.; Gonzalez, M. U.; Quidant, R.; Kall, M.
Optical aggregation of metal nanoparticles in a microuidic channel
for surface-enhanced Raman scattering analysis. Lab Chip 2009,
9(2), 193e195
(b). Pradere, C.; Joanicot, M.; Batsale, J. C.; Toutain, J.; Gourdon, C.
Processing of temperature elds in chemical microreactors with in-
frared temperature elds. QIRT J. 2009, 3, 117e135.
16. Chastek, T. Q.; Iida, K.; Amis, E. J.; Fasolka, M. J.; Beers, K. L. A mi-
crouidic platform for integrated synthesis and dynamic light scattering
measurement of block copolymer micelles. Lab Chip 2008, 8(6),
950e957.
17. Krishnadasan, S.; Brown, R. J. C.; deMello, A. J.; deMello, J. C. Intel-
ligent routes to the controlled synthesis of nanoparticles. Lab Chip 2007,
7, 1434e1441.
18. Murray, C. B.; Norris, D. J.; Bawendi, M. G. Synthesis and character-
ization of nearly monodisperse CdE (E sulfur, selenium, tellurium)
semiconductor nanocrystallites. J. Am. Chem. Soc. 1993, 115,
8706e8715.
19. Debenedetti, P. G. Homogeneous nucleation in supercritical uids.
AIChE J. 1990, 36(9), 1289e1298.
20. Nissan, A. H. Boiling point and viscosity. Nature 1943, 152, 630.
JALA December 2009 373