Content uploaded by Vincent Studer
Author content
All content in this area was uploaded by Vincent Studer
Content may be subject to copyright.
Microfluidic stickers{
Denis Bartolo,*
a
Guillaume Degre´,
b
Philippe Nghe
b
and Vincent Studer*
c
Received 13th August 2007, Accepted 2nd November 2007
First published as an Advance Article on the web 22nd November 2007
DOI: 10.1039/b712368j
We present how to make and assemble micro-patterned stickers (mPS) to construct high
performance plastic microfluidic devices in a few minutes. We take advantage of soft UV imprint
techniques to tailor the geometry, the mechanical properties, and the surface chemistry of 2D and
3D microfluidic circuits. The resulting microfluidic stickers substantially overcome the actual
performance of the very popular PDMS devices for a wide range of applications, while sharing
their celebrated fast and easy processing. To highlight the intrinsic advantages of this method,
three important applications are detailed: (i) we show that both aqueous and organic droplets can
be produced and stored in stickers without any specific surface coating. (ii) We report on the
outstanding pressure resistance of the mPS, which open the way to the transport of viscous
complex fluids. (iii) Finally, a simple design strategy is proposed to generate complex flow
patterns in interconnected stacks of mPS.
Introduction
Since the pioneering work of Delamarche et al.,
1
elastomer
materials have been considered as excellent options for the fast
fabrication of microfluidic chips.
2
Microfluidic devices made
of PDMS (polydimethysiloxane) by replica molding are now
part of the standard instrumental toolbox of many research
laboratories, ranging from molecular and cell biology to
hydrodynamics and analytical chemistry. The reason for this
ubiquitous spreading is twofold: (i) replica molding allows for
submicron size pattern replication; yet it neither requires
expensive facilities and equipments, such as a clean room, nor
advanced skills in microfabrication.
3
(ii) PDMS combines
numerous excellent physical and chemical properties: it is
optically clear and has low autofluorescence,
4
it is permeable
to gases,
5
it has a low elastic modulus and thus allows for
simple integration of active elements, such as valves and
pumps.
6
Nevertheless, some of the properties of PDMS
strongly limit its range of applications. First, this silicon
elastomer swells in most organic solvents,
7
which mostly limits
its use to aqueous solutions. Second, permanent modification
of the chemical properties of PDMS surfaces is a challenging
task.
8
This narrows the range of possible formulations in
multiphase flows.
9
Third, due to the low elastic modulus of
PDMS, low pressures can significantly alter the geometry of
the channels, and the associated relaxation time can reach
hours for viscous fluids. This bounds the working pressure
typically below y1 bar.
10
In turn, this prevents the accurate
transport of viscous fluids, such as concentrated polymer
solutions making rheological measurements almost impossible
in PDMS devices.
11
Only few other simple prototyping strategies have been
proposed to circumvent these limitations. Desimone and
coworkers have replaced PDMS by low modulus solvent
resistant fluoropolymer.
12
So far this technique is restricted to
experts in polymer chemistry. Alternatively, the replacement of
replica molding by direct photolithographic patterning of
photoresist resins is a promising and straightforward method
for the fabrication of single layer polymeric microdevices.
13–15
But, so far these techniques are restrained to millifluidic
channels (typical size: 100 mm–1 mm).
In this paper, we present a simple prototyping method
to make micro-patterned stickers (mPS) and show how to
assemble them to form microfluidic circuits in a few minutes
(Fig. 1A). In a first part, we show how to pattern and handle
a
Laboratoire de physique et me´canique des milieux he´te´roge`ne, PMMH-
ESPCI-CNRS UMR 7636-P6-P7, Paris, France.
E-mail: denis.bartolo@espci.fr.
b
, Microfluidique MEMS et Nanostructures, ESPCI-CNRS UMR
Guliver 7083, Paris, France
c
Laboratoire de Neurobiologie et Diversite´ Cellulaire, ESPCI-CNRS
UMR 7637, Paris, France. E-mail: vincent.studer@espci.fr
{ The HTML version of this article has been enhanced with colour
images.
Fig. 1 (A) Microfluidic circuit made of a mPS sealed by a glass slide.
(B) mPS bound on a curved surface: the outer side of a 50 ml glass
beaker containing water and orange G dye and hexadecane. A hole
made with a sand blaster connects the inner side of the beaker to the
microcircuit. The liquid flow is driven by the hydrostatic pressure. (C)
SEM picture of the cross section of a mPS microfluidic channel
(NOA81) bound to a glass slide. Note the absence of sagging, even for
this very low aspect ratio channel (width: 350 mm, height: 7 mm).
PAPER www.rsc.org/loc | Lab on a Chip
274 |
Lab Chip
, 2008, 8, 274–279 This journal is
ß
The Royal Society of Chemistry 2008
thin plastic films with adhesive capabilities using soft imprint
lithography.
16
This technique offers the opportunity to tailor
not only the geometry but also the mechanical properties and
the surface chemistry of the mPS. We then explain how to stack
and seal the stickers to construct simple or complex 2D or 3D
microdevices that can combine high spatial resolution,
compatibility with organic solvents, and outstanding mechani-
cal resistance. In a second part, we present three important
applications in order to highlight some of the advantages of
the microfluidic stickers: (i) hydrophilic and hydrophobic
devices for digital microfluidics, (ii) pressure resistant micro-
channels for the transport of complex fluids, and (iii) multi-
layer 3D microfluidic circuits. Finally, we provide a thorough
description of the fabrication protocols of each device.
Fabrication
The stickers are made by soft imprint lithography, as sketched
in Fig. 2A. A drop of photocurable monomer (or prepolymer)
is deposited on a flat or patterned PDMS substrate. A
structured PDMS stamp is then gently pressed onto the drop.
To cure the polymer film, the liquid is insolated through the
transparent PDMS stamp. Since oxygen inhibits the free-
radical polymerization used here to build the polymer
network, the permeability to gas of the PDMS ensures that
an ultra thin superficial layer of liquid remains uncured.
16
Therefore, both sides of the resulting textured film retain
adhesive capabilities. For most imprint processes, the total
thickness of the sticker exceeds the height of the pattern due to
the usual residual layer remaining below the stamp (Fig. 2A).
This residual layer can be suppressed without any further
etching process, by matching its height with the thickness of
the uncured monomer layer. However, this requires a fine
tuning of the exposure dose. To obtain such stencil like
stickers, it is actually much more convenient to first contact the
two PDMS substrates and then fill the empty space in between
with the liquid prepolymer (Fig. 2B). In practice, the gap is
spontaneously filled due to capillary forces.
3
To speed up the
capillary spreading, the parts of the stamp surrounding the
pattern can be decorated with a dense network of pillars. This
porous structure enhances the local curvature of the liquid
meniscus and consequently the Laplace pressure which drives
the liquid motion becomes higher.
17
Assembly of microfluidic device: stacking and sealing
We now describe how to handle and assemble the plastic
stickers in order to construct microfluidic devices. The main
idea consists in never handling freestanding plastic films. Such
thin sheets would be almost impossible to manipulate because
of their very low bending modulus. A first sticker still lying on
one of its PDMS mold is gently pressed by hand using a rigid
substrate (glass slide, cover slip, metal plate etc.) and bound to
it by photocuring the remaining reactive layer (Fig. 2C). This
straightforward sealing of these adhesive mPS is one of the
main advantages of this technique. Excellent adhesion is
achieved without any specific chemical, thermal or mechanical
(high pressure) treatment. Furthermore, the high flexibility of
the thin stickers enables bonding on curved surfaces (Fig. 1B).
The PDMS mold is then removed, which makes the second
adhesive side of the sticker accessible. The process can be
reiterated to stack multiple stickers made of identical or
different polymers. Each layer constitutes a different micro-
fluidic network. These networks can be interconnected by
means of stencil like stickers. Eventually, a second rigid
substrate can be added on top of the upper layer to strengthen
the device. Each rigid substrate can include access holes for the
macroscopic fluidic connections.
Materials
In practice, the choice of the mPS materials is driven by the
application of the microfluidic circuits. The design strategy
described above is very general and can be applied to an
almost unbound number of combinations of polymers and
channel geometries. However, most of the stickers used in the
experiments described below have been made of the thiolene
based resin NOA 81 (Norland optical adhesive). This resin
offers an excellent combination of physical and optical pro-
perties complementary to the one of PDMS. (i) NOA 81 has
shown a greater solvent resistance to swelling than PDMS.
15
For exemple, we have flown apolar solvents, such as decaline
and hexadecane, in devices made of two NOA 81 stickers for
several hours. In PDMS channels of identical geometry, the
swelling was so strong that it completely prevented the filling
of the device. (ii) NOA 81 has an elastic modulus 3 orders
of magnitude higher than PDMS (typically 1 GPa). This
avoids sagging effects, even for very low aspect ratio shallow
Fig. 2 (A) Sketch of the fabrication process of a mPS. (B) Sketch of the fabrication process of a stencil like mPS. For both methods the two surfaces
of the sticker still have reactive sites after UV illumination. (C) Construction of microfluidic devices. (Left) One layer device: the circuit imprinted
on the mPS is sealed with a glass slide. (Right) Multilayer devices: the stencil like mPS is sealed with a sticker previously bound to a glass slide.
This journal is
ß
The Royal Society of Chemistry 2008
Lab Chip
, 2008, 8, 274–279 | 275
channels. To illustrate this important advantage, we show in
Fig. 1C a SEM picture of a straight channel with a 1 : 50
aspect ratio. Note that contrary to replica molding, the mPS
fabrication method is especially well suited for making and
handling thin polymer films with a high elastic modulus. (iii)
NOA 81 enables the replication of submicron features. (iv)
NOA 81 is a resin designed for optics applications. An
important advantage is its very low autofluorescence which,
for instance, makes possible the tracking of single quantum
dots in microfluidic stickers.
27
More quantitatively, we have
compared the fluorescence background of 170 mm films made
of PDMS, NOA 81 and PMMA. The excitation wavelength
was 488 nm and the emission was measured at 520 and 580 nm.
The fluorescence intensity was normalized by the intensity
measured with a glass cover slip of the same thickness. The
autofluorescence level measured for the NOA films is 4 (resp. 5)
times smaller than for the PDMS (resp. PMMA) for both
wavelengths. (iv) NOA 81 is biocompatible. Several living
cells, including neurons, have been observed over days in
microfluidic stickers.
27
(v) Contrary to a PDMS device, NOA
81 stickers are impermeable to air and water vapor, which
prevents the evaporation of the liquids flowing in the channels.
(vi) Like PDMS micromolding, the soft imprint of NOA 81
films enables the replication of (sub)micron size features.
For example, we managed to imprint arrays of cylindrical
microwells of radius 1 mm and 3 mm depth.
However, this resin has a high shear viscosity (y300 mPa?s
at 25 uC), this makes the fabrication of stencil like mPS
difficult. To circumvent this limitation, we also detail how
to use another photocuring mixture with lower viscosity:
tri(methylolpropane)triacrylate (TMPTA) + 1 wt% of photo
initiator (Darocur 1173, Sigma Aldrich).
Applications
We now present three microfluidic experiments done in devices
made of patterned stickers. These examples are intended to
highlight more accurately some of the advantage of the mPS
to make performance devices. For the sake of clarity, all the
detailed protocols used to fabricate the corresponding micro-
fluidic circuits are provided in the last section of this paper.
Hydrophobic and hydrophilic drop emitters
The stabilisation of emulsions or foams requires the slow down
of the coalescence of the droplets or bubbles. The basic
strategy consists of adding a third component (surfactants,
polymers or small particles) to kinetically stabilize the liquid
dispersion.
18
Given the very high surface over volume ratio of
microfluidic channels, another important requirement has to
be fulfilled to produce, transport, and store microfluidic
emulsions. It is indeed necessary to prevent the partial wetting
of the channels walls by the droplets.
9
For instance, it is
impossible to produce and store oil in water emulsions in
native hydrophobic PDMS channels, even by adding surfac-
tants to the aqueous phase. To overcome this strong limitation
to liquid formulation, the usual strategy consists of changing
the wetting properties of the channel walls by modifying their
surface chemistry. Again, in the case of PDMS, long term
stable surface modification remains challenging. For the most
conventional oxygen plasma treatment, the original hydro-
phobicity is typically regained after a few minutes.
19
Moreover, surface modifications are often incompatible with
the standard sealing methods.
As the mPS can be made of hydrophilic or hydrophobic
polymers, the wetting properties of the channels walls can be
chosen to produce either direct or reverse emulsions. We show
in Fig. 3, two identical flow focusing drop emitters made in a
hydrophilic (NOA 81) (Fig. 3A), and in a hydrophobic resin
(TMPTA) (Fig. 3C). A monodisperse hexadecane in water
emulsion is generated in the first device (Fig. 3B), the reverse
water in hexadecane emulsion is produced in the second one
(Fig. 3D). This demonstrates that specific surface modifica-
tions can be bypassed by an appropriate choice of the stickers
materials to produce and transport aqueous and organic drops.
Pressure resistant devices for the transport of complex fluids
Microfluidics is a new tool to investigate the flow of complex
fluids, such as concentrated polymer or surfactant solu-
tions.
11,20,21
However, pressure drops induced by the flow of
such viscous fluids (up to several Pa?s) in microfluidic devices
can rapidly reach several bars. In this high pressure range,
mechanical deformation of the channels occurs when working
with low elastic modulus material.
10
As a consequence, the
change in the geometry and the long transient regimes
associated with the mechanical relaxation are two major
limitations of PDMS microdevices. In this paragraph, we
compare the mechanical performances of PDMS and mPS
channels.
Fig. 3 (A) Picture of a flow focusing drop emitter made with a single
NOA 81 sticker. Monodisperse droplets of hexadecane are produced in
a solution of orange G dye and 1 wt% sodium dodecyl sulfate in water.
(B) Magnified view of the output serpentine channel transporting the
direct emulsion. Channel width: 200 mm, channel height: 80 mm. (C)
Picture of a flow focusing drop emitter made with a single tri(methylol
propane) triacrylate sticker. Monodisperse plugs of an aqueous
solution of orange G dye are produced in a hexadecane and 1 wt%
span 80 solution. (D) Magnified view of the output serpentine channel
transporting the reverse emulsion. Channel width: 200 mm, channel
height: 80 mm.
276 |
Lab Chip
, 2008, 8, 274–279 This journal is
ß
The Royal Society of Chemistry 2008
Resistance to deformation. We flow a Newtonian solution in
a straight channel (2 cm 6 200 mm 6 16 mm) and measure
the relationship between the imposed flow rate Q and the
measured pressure drop DP along the microchannel (Fig. 4A).
For PDMS microchannels (squares), the DPvsQcurve shows
non linear variations for pressures higher than a few hundreds
of millibars. The hydrodynamic resistance (i.e. the slope of the
DPvsQcurve) decreases as the pressure increases due to
the elastic deformation of the channel. On the contrary, with
the mPS channel (circles), the pressure increases linearly
with the flow rate, even for pressures as high as 15 bars, as
expected for a Newtonian fluid in a channel with a constant
cross section.
Relaxation dynamic. To study the transient response of the
device, we impose a constant pressure flow in the same micro-
channel filled with a glycerine–water solution (shear viscosity:
50 mPa?s) containing fluorescent tracers (Fluospheres 500 nm,
Molecular Probes). The pressure difference is dropped from
3 bars to 0 bar. PIV measurements are performed to measure
the temporal evolution of the flow rate.
21
The flow in the mPS
device stops after a few tens of milliseconds, whereas this
relaxation time is three orders of magnitude higher in the
PDMS channel (Fig. 4B).
These two experiments demonstrate that mPS circuits are
efficient tools for the study of high pressure flows. Thanks to
mPS devices, we have been able to investigate the rheology of
semi-dilute micellar fluids with zero shear viscosity of several
tens of Pa?s.
11
3D microfluidic circuits and flow patterning
An important drawback of microfluidic devices made using
standard lithographic methods is their intrinsic planar
geometry. This strongly limits the architectural and the
functional complexity of the fluidic circuits. For instance, the
crossing of independent channels cannot be achieved in a
single layer circuit. So far, only very few methods have been
proposed to microfabricate 3D fluidic networks.
14,22,23
The
strategies proposed to assemble and connect multilayer devices
are often time consuming and technically involved. As a con-
sequence, most of the microfluidic applications are restrained
to single layer devices (or of stacks of such devices).
Thanks to the replication technique presented here, 3D
fluidic circuits, including an arbitrarily high number of
connections in the three directions, can be made in a few
minutes. Stencil like mPS makes it possible to fabricate vias,
where size and geometry are only limited by the lithographic
technique used to make the primary molds. We show in Fig. 5A
a simple 3D flow patterning circuit made of three mPS. A first
sticker, including a single straight channel, is covered with a
stencil sticker including five connection holes (Fig. 5B). A
second straight channel, perpendicular to the first one, is
included in the third sticker that seals the circuits. The
direction of flow of the two liquids is schematized in Fig. 5C. A
1 wt% SDS (sodium dodecyl sulfate) solution (liquid A) is
injected in the first straight channel. A 1 wt% SDS aqueous
solution colored with solvent blue dye (Sigma Aldrich)
(liquid B) is injected in the second one. One can observe on
Fig. 5B that, downstream of the vias, we manage to impose an
alternated pattern of liquid A and B (ABABABABABA) with
only two fluid inlets and one outlet. Obviously, this simple
example suggests the enormous advantage of such a 3D device
for the spatiotemporal patterning of multiphase flows. More
specifically, in the case of two fluid flows, the number of
maxima of the solute concentrations in 2D circuits scales with
the number of inlets. Conversely, using the same strategy as in
Fig. 4 (A) Pressure drop, DP, versus flow rate, Q, curve for 2 cm long
(200 6 16 mm cross-section) PDMS microchannels (squares) and NOA
81 microchannels (circles). (B) Evolution of the normalized flow rate of
a glycerin and water mixture after a step of pressure from 3 bars to
0 bar applied at time t =0.
Fig. 5 (A) Picture of a 3D microfluidic device made of 3 stacked mPS. This device has 2 inlets and one outlet. It allows for generating complex flow
patterns. (B) Complex stream of blue dye in water solution generated by the 3D microfluidic device. Scale bar 500 mm. (C) Schematic of the 3D
device. The stream of dye is represented by dashed arrows.
This journal is
ß
The Royal Society of Chemistry 2008
Lab Chip
, 2008, 8, 274–279 | 277
Fig. 5, only two inlets are required, whatever the complexity
of the imposed patterns. A detailed discussion of the
spatiotemporal flow patterns obtained with three layer devices
will be presented elsewhere.
24
Conclusions
We have presented a fast and simple fabrication method to
make both 2D and 3D fluidic networks by stacking micro-
patterned stickers. This method offers the opportunity to make
high performance devices in a few minutes. We show that a
wide range of physical and chemical properties of the mPS
enable numerous novel fast prototyping applications. The
range of accessible working pressure, combined with the high
spatial resolution of the mPS, should extend the range of
microfluidic applications in separation techniques
25
and
complex fluid rheology.
11
Furthermore, the mPS can be stuck
in aqueous media, which offers an interesting perspective in
flow patterning for cell and molecular biology.
26
Experimental
Hydrophilic flow focusing device (Fig. 3A)
This device has been made by sealing a microfluidic sticker
with a glass slide. The sticker is made of NOA 81. A y5mm
thick PDMS mold is made by the replica molding of a primary
master obtained by photolithography.
3
A y100 ml drop of
NOA 81 is sandwiched between the PDMS mold and a flat
PDMS sheet. The liquid is then cured with a collimated UV
light source (Hamamatsu LC8), exposure time: 15 s, exposure
power: 25 mW cm
22
. The mold is then removed and the plastic
sticker is pressed by hand against the glass surface in which
connection holes have been made with a sand blaster. A
second UV exposure (60 s, 25 mW cm
22
) permanently seals the
device. The fluid is then injected by connecting plastic tubing
(Tygon S54HL) to plastic connectors (upchurch nanoport
N-333) aligned with the holes on the glass substrate. The drop
emitter shown in Fig. 3 has been fabricated in less than 10 min.
Hydrophobic flow focusing device (Fig. 3B)
A y100 ml drop of TMPTA and 1 wt% of photo initiator
(Darocur 1173, Sigma Aldrich) is sandwiched between the
same PDMS stamp as above and a flat PDMS sheet. The
liquid is cured after a 30 s UV exposure (exposure power:
12.5 mW cm
22
). The PDMS mold is then removed and the
plastic sticker is pressed by hand against a flat plastic substrate
(petri dish) covered by a thin film of the same polymer. This
film is prepared by pressing a drop of monomer with a
flat PDMS surface. The resulting film is photopolymerized
(exposure time: 30 s, exposure power: 12.5 mW cm
22
). A
second UV exposure (exposure time: 60 s, exposure power:
25 mW cm
22
) permanently seals the device. For this device, the
four walls of the channels are made of the same hydrophobic
material and thus have identical chemical properties. Since the
TMPTA has a low shear viscosity, the thickness of the residual
layer is very thin. To strengthen the device, a second plastic
substrate has been bound to the plastic sticker after the sealing
step. The same tubing and connector are used to inject the
liquid and connection holes have been drilled in the plastic
substrate prior to sealing.
Pressure resistant device (Fig. 4)
We use a quite similar protocol to make this straight
microchannel. Here, a drop of NOA 81 is imprinted directly
on a glass slide (15 s, 25 mW cm
22
). After removing the PDMS
mold, the channel is sealed with a glass cover slip pressed by
hand on the sticker, in order to obtain a perfectly rigid chip
(made of one glass slide and one glass cover slip sandwiching a
plastic sticker). To improve the adhesion to the glass surfaces,
the second UV illumination (60 s, 25 mW cm
22
) is followed by
a heating of the whole system during 1 h at 150 uC. Note that if
the connection holes are to be on the glass slide, the residual
layer of adhesive polymer must be punched through with a
needle after the first UV curing and before the device is sealed.
3D microfluidic circuit (Fig. 5)
The 3D microfluidic circuit presented in Fig. 5 is composed of
3 different mPS. One mPS is fabricated for each of the two
rectangular channels. For the large rectangular channel (2 mm
wide, 50 mm high and 20 mm long), a y5 mm thick PDMS
mold is made by the replica molding of a primary master
obtained by photolithography.
3
A y100 ml drop of NOA 81 is
sandwiched between the PDMS mold and a glass microscope
slide, in which 3 connection holes have been made with a sand
blaster. The liquid polymer is then cured with a collimated UV
light source (Hamamatsu LC8), exposure time: 15 s, exposure
power: 25 mW cm
22
. The mold is removed and the connecting
holes are re-punched with a gauge 21 needle. The second
channel (500 mm wide, 50 mm high and 10 mm long) is
fabricated the same way on a clean 22 6 32 mm #1 cover
slip (ESCO, USA) without holes. The mold is removed.
Eventually, a stencil like mPS containing the 5 cylindrical
(diameter 200 mm) vias connecting the two perpendicular
channels is fabricated: a PDMS mold decorated with a dense
network of 600 mm diameters pillars and 5 cylinders (for the
vias) is placed over a flat sheet of PDMS. The gap between the
mold and the flat PDMS sheet is filled by capillarity with a
y100 ml drop of TMPTA and 1 wt% of photo initiator
(Darocur 1173, Sigma Aldrich). The liquid is cured after a 30 s
UV exposure (exposure power: 12.5 mW cm
22
) (Fig. 2B). The
flat PDMS sheet is removed and the stencil like mPS is aligned
and pressed with hand against the second channel. The mold
of the stencil mPS is removed, as shown on Fig. 2C (right). The
two stacked mPS are then aligned and pressed by hand against
the first channel. The entire stack is sealed with a collimated
UV light source (Hamamatsu LC8), exposure time: 60 s,
exposure power: 25 mW cm
22
. Three pierced cubes of PDMS
(5 mm 6 5mm6 5 mm) are plasma sealed on each hole of the
microscope slide for macroscopic fluidic connections (Fig. 5A).
Acknowledgements
We thank Pascal Silberzan and Maxime Dahan for insightful
comments. Adrien Rennesson and Charlie Gosse (LPN-CNRS
UPR 20 Marcoussis France) are gratefully acknowledged for
help with the experiments and the SEM imaging, respectively.
278 |
Lab Chip
, 2008, 8, 274–279 This journal is
ß
The Royal Society of Chemistry 2008
All the primary molds used in the experiments have been made
at the ESPCI clean room facility. VS and DB benefit from
BQR ESPCI financial support.
References
1E.Delamarche,A.Bernard,H.Schmid,B.Micheland
H. Biebuyck, Science, 1997, 276, 779.
2 G. M. Whitesides, Nature, 2006, 442, 368.
3 Y. N. Xia and G. M. Whitesides, Angew. Chem., Int. Ed., 1998, 37,
551.
4 D. Psaltis, S. R. Quake and C. Yang, Nature, 2006, 442, 381.
5 G. C. Randall and P. S. Doyle, Proc. Natl. Acad. Sci. U. S. A.,
2005, 102, 1081; E. Verneuil, A. Buguin and P. Silberzan,
Europhysics Letters, 68, 412; J. Leng, B. Lonetti, P. Tabeling,
M. Joanicot and A. Ajdari, Phys. Rev. Lett., 2006, 96, 084503.
6 M. A. Unger, H. P. Chou, T. Thorsen, A. Scherer and S. R. Quake,
Science, 2000, 288, 113.
7 J. Ng Lee, C. Park and G. M. Whitesides, Anal. Chem., 2003, 75,
6544.
8 H. Makamba, J. H. Kim, K. Lim, N. Park and J. H. Ha,
Electrophoresis, 2003, 24, 3607.
9 R. Dreyfus, P. Tabeling and H. Willaime, Phys. Rev. Lett., 2003,
90, 144505.
10 T. Gervais, J. El-Alib, A. Guntherb and K. F. Jensen, Lab Chip,
2006, 6, 500.
11 G. Degre´ and P. Nghe, Proc. mTAS 2007 Conference, 2007,
pp. 1480–1482; G. Degre´, P. Nghe, P. Tabeling and A. Ajdari, to
be published.
12 J. P. Rolland, R. M. Van Dam, D. A. Schorzman, S. R. Quake and
M. DeSimone, J. Am. Chem. Soc., 2004, 126, 2322.
13 D. J. Beebe, J. S. Moore, Q. Yu, R. H. Liu, M. L. Kraft, B. H. Jo
and C. Devadoss, Proc. Natl. Acad. Sci. U. S. A., 2000, 97, 13488.
14 J. B. Hutchison, K. T. Haraldsson, B. T. Good, R. P. Sebra,
N. Luo, K. S. Anseth and C. N. Bowman, Lab Chip, 2004, 4,
658–662.
15 Z. T. Cygan, J. T. Cabral, K. L. Beers and E. J. Amis, Langmuir,
2005, 21, 3629.
16 J. P. Rolland, E. C. Hagberg, G. M. Denison, K. R. Carter and
J. M. De Simone, Angew. Chem., Int. Ed.
, 2004, 43, 5796–5799.
17 D. Dendukuri, D. C. Pregibon, J. Collins, T. Alan Hatton and
P. S. Doyle, Nat. Mater., 2006, 5, 365.
18 See alsoM. Zimmermann, H. Schmid, P. Hunzikerb and
E. Delamarche, Lab Chip, 2006, 7, 119–125 for another micro-
fluidic application of artificial porous structure.
19 J. Bibette, F. Leal-Calderon, V. Schmitt and P. Poulin, Emulsion
Science,Springer-Verlag, New York, 2007.
20 D. C. Duffy, J. C. McDonald, O. J. A. Schueller and G. M.
Whitesides, Anal. Chem., 1998, 70, 4974.
21 G. H. McKinley, W. P. Raiford, R. A. Brown and R. C. Armstrong,
J. Fluid. Mech., 1991, 223, 411–456.
22 G. Degre´, P. Joseph, P. Tabeling, S. Lerouge, M. Cloitre and
A. Ajdari, Appl. Phys. Lett., 2006, 89,2.
23 J. R. Anderson, D. T. Chiu, R. J. Jackman, O. Cherniavskaya,
J. C. McDonald, H. K. Wu, S. H. Whitesides and G. M. Whitesides,
Anal. Chem., 2000, 72, 3158–3164.
24 E. P. Kartalov, C. Walker, C. R. Taylor, W. F. Anderson and
A. Scherer, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 12280–12284.
25 D. Bartolo and V. Studer, to be published.
26 L. R. Huang, E. C. Cox, R. H. Austin and J. C. Sturm, Science,
2004, 304, 987.
27 M. Morel, D. Bartolo, M. Dahan and V. Studer, Proc. mTAS 2007
Conference, 2007, pp. 161–163.
This journal is
ß
The Royal Society of Chemistry 2008
Lab Chip
, 2008, 8, 274–279 | 279
