C O M M U N I C A T I O N S
(62.5 mM) was brought into the hydrophilic region followed by
introduction of a solution of adipoyl chloride in xylenes (46.9 mM)
into the hydrophobic region. Interfacial polymerization occurred
immediately when the two phases made contact producing a
polymer film at the hydrophilic-hydrophobic boundary (Figure
3b,c). The polymerization proceeded at room temperature for 8 min
at which point the organic solution was flushed out of the channel
with xylenes and the aqueous solution was flushed out of the
channel with methanol. Both sides of the membrane were then
rinsed with 10 mL of methanol and dried with nitrogen. Membrane
permeability was studied on an Olympus fluorescent microscope
BX 60 using an aqueous suspension of 0.2 µm fluorescent
microspheres.14 The suspension was injected into the hydrophilic
region and was retained by the membrane under ambient conditions
(Figure 3c,d). When a pressure was applied to the suspension, water
gradually passed through the membrane while microspheres
remained behind and became concentrated in the vicinity of the
membrane (Figure 3e,f). This study indicated that the membrane’s
pore size is below 200 nm.
In conclusion, controlling the boundary between immiscible
liquids has been achieved by patterning surfaces inside microchan-
nels. The maximum pressures that liquid walls can sustain were
derived analytically and studied experimentally. The ability to
confine immiscible liquids in specified regions inside microchannels
opens a wide range of opportunities in microfluidic systems, as
exemplified here by fabrication of semipermeable membranes via
interfacial polymerization.
Figure 3. Fabrication and permeability study of a semipermeable polyamide
membrane. (a) Schematic illustration of a surface-patterned channel. (b)
Schematic illustration of a polymer membrane fabricated inside the channel
by interfacial polymerization. The red border indicates the region where
images (c), (d), (e), and (f) were recorded. Optical micrograph (c) and
fluorescent image (d) of an aqueous suspension of 0.2 µm fluorescent
microspheres confined to the hydrophilic region by the membrane. Optical
micrograph (e) and fluorescent image (f) of the aqueous suspension of 0.2
µm microspheres when a pressure is applied. An advancing water front
free of microspheres can be seen passing through the membrane.
Acknowledgment. This work was supported by grants from
the Defense Advanced Research Projects Agency-MTO F30602-
00-1-0570 and NSF DMR 01-06608.
Supporting Information Available: Synthesis of F-SAM; photo-
patterning procedures; substrate preparation and microchannel fabrica-
Obviously, θwater/org and θorg/water must be greater than 90° to confine
H2O in the hydrophilic region and organic liquids in the hydro-
phobic region. We have measured θwater/org on a F-SAM and θorg/water
on a UV-irradiated F-SAM. The experimentally determined maxi-
mum pressures that liquid walls can withstand are generally in good
agreement with the calculated values by use of the independently
measured values of θwater/org and θorg/water and the liquid-liquid
interfacial tensions.10
tion; measurement of maximum pressures and θwater/org and θorg/water
;
experimental results on maximum pressures and advancing contact
angles (PDF). This material is available free of charge via the Internet
References
(1) Kovacs, G. T. A. Micromachined Transducer Sourcebook; McGraw-
Hill: Boston, 1998.
In comparison to multistream laminar flow, the ability to confine
organic liquids to hydrophobic regions and aqueous solutions to
hydrophilic regions makes it practical to manipulate immiscible
liquids inside microchannels. For example, one liquid can remain
static while the other liquid is flowing and the boundary remains
constant. Moreover, we have shown that immiscible liquids flowing
in the same direction (concurrent flow) or in the opposite directions
(countercurrent flow) also maintain a stable, stationary boundary.
In contrast to laminar flow, the width ratio of two streams is not
determined by the relative volumetric flow velocities but rather by
the width ratio of hydrophobic and hydrophilic pathways provided
the pressure is subcritical. We should point out that countercurrent
flow is widely adopted in nature for efficient mass transfer such as
gas exchange between air and blood in bird lungs and between
water and blood in fish gills.13 Here we utilize liquid walls to
conduct an interfacial polymerization in a surface-patterned channel
to fabricate a semipermeable membrane.
(2) Weigl, B. H.; Yager, P. Science 1999, 283, 346.
(3) Kamholz, A. E.; Weigl, B. H.; Finlayson, B. A.; Yager, P. Anal. Chem.
1999, 71, 5340.
(4) Kenis, P. J. A.; Ismagilov, R. F.; Whitesides, G. M. Science 1999, 285,
83.
(5) Takayama, S.; Ostuni, E.; Qian, X.; McDonald, J. C.; Jiang, X.; LeDuc,
P.; Wu, M. H.; Ingber, D. E.; Whitesides, G. M. AdV. Mater. 2001, 13,
570.
(6) Zhao, B.; Moore, J. S.; Beebe, D. J. Science 2001, 291, 1023.
(7) Hisamoto H.; Horiuchi, T.; Uchiyama, K.; Tokeshi, M.; Hibara, A.;
Kitamori, T. Anal. Chem. 2001, 73, 5551.
(8) Hibara, A.; Tokeshi, M.; Uchiyama, K.; Hisamoto, H.; Kitamori, T. Anal.
Sci. 2001, 17, 89.
(9) Zhao, B.; Moore, J. S.; Beebe, D. J. Unpublished results.
(10) Details can be found in the Supporting Information.
(11) No external pressure is applied to the organic phase when a pressure is
applied to the aqueous phase, and vice versa.
(12) Adamson, A. W. Physical Chemistry of Surfaces, 5th ed.; Wiley: New
York, 1990.
(13) Schmidt-Nielsen, K. How Animals Work; Cambridge University Press:
London, 1972.
(14) The aqueous suspension of 0.2 µm carboxylate-modified microspheres
(yellow-green fluorescent, 505/515) was purchased from Molecular Probes,
and was diluted with two volumes of deionized water before used for
permeability study.
A F-SAM coated channel was photopatterned as illustrated in
Figure 3a. An aqueous solution containing hexamethylenediamine
JA025835J
9
J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002 5285