A flexible sample introduction method for polymer microfluidic

MINIATURISATION FOR CHEMISTRY, BIOLOGY & BIOENGINEERING
A flexible sample introduction method for polymer microfluidic
chips using a push/pull pressure pump
Zhiyong Wu, Henrik Jensen, Jean Gamby, Xiaoxia Bai and Hubert H. Girault*
Laboratoire d’Electrochimie Physique et Analytique, Ecole Polytechnique Fe´de´rale de
Lausanne, CH-1015 Lausanne, Switzerland. E-mail: hubert.girault@epfl.ch;
Fax: 141.21.693.36.67; Tel: 141.21.693.31.51
Received 21st July 2003, Accepted 18th May 2004
First published as an Advance Article on the web 22nd July 2004
A push/pull sample introduction method based on push/pull pressure flow for microfluidic systems (cross,
double T and multichannel structures) is presented. This leads to well-defined and controllable sample plugs
even when dealing with long channels. By tuning the relative push/pull pressure, it is shown that the size of the
sample plug can effectively be controlled. Good signal reproducibility upon continued sample introduction and
subsequent chip electrophoresis employing fluorescence detection is demonstrated for different chip geometries
(i.e. short channels and long channels). Since the performance of the method is relatively insensitive to chip
geometry, it is particularly useful for polymeric prototype microchips as tedious optimization is not required.
Furthermore, the push/pull sample introduction is extended to multichannel chips thus demonstrating the
possibilities of applying the methodology for realizing single chip high throughput sample analysis.
Instrumentation
Introduction
The advent of micro Total Analysis Systems (mTAS) in
analytical chemistry has received a lot of attention in recent
years.1,2 Although much progress has been achieved in terms of
general chip design, such as chip electrophoresis for sample
separation and detector integration, a reliable and flexible
sample introduction remains as a challenging task. Most of the
current methods rely on electrodynamic injection, although in
some cases, this procedure has proven to be problematic, since
flooding or leakage may occur.3,4 Meanwhile, a certain sample
bias (depending on the analyte mobility) is introduced in
response to the high electric field used for the sample
introduction.3 This problem is particularly noticeable for
polymeric microfluidic chips as the surface stability and
properties are difficult to control. One solution to remove
this injection bias is to use the double-T injection design
together with a long enough injection time.5 Another effective
solution could be based on the pressure injection which is
insensitive to the channel surface and the sample properties.6–11
In this paper, a pressure induced sample introduction
procedure is presented. Basically the non-biased sample plugs
are introduced in a separation channel by applying a push and
pull pressure simultaneously at respectively the sample and
sample waste reservoirs. It is demonstrated that for polymer
chips the new method is comparable to the existing methods in
terms of reproducibility and furthermore the sample dilution
by the buffer is avoided. Finally, the setup is extended to the
multichannel chips which are suitable for general high
throughput sample analysis.
Experimental
DOI: 10.1039/b308405a
Chemicals and materials
512
The PET chip substrate (100 mm, Melinix S grade) was
purchased from DuPont (Switzerland). A 20 mM (pH ~ 7)
phosphate buffer (Fluka, Buchs Switzerland) was used as the
running buffer. Fluorescein sodium salt and fluorescein biotin
were obtained from Sigma (St. Louis, MO, USA). Deionized
water (18.2 MV) was prepared using a Milli-Q system from
Millipore (Bedford, MA, USA).
Lab Chip, 2004, 4, 512–515
Axiovert 200 (Zeiss) equipped with a CCD camera CF 8/4
(Kappa, Gleichen, Germany) was used for fluorescence
detection with a UV excitation source Lamp HPO 100/2 W
(Mercury lamp) and the filter was set for fluorescein (495/
519 nm). A high voltage source (Spellman CZE 1000R Power
Supply, Hauppauge, NY, USA) was used for electrophoresis.
Acquisition of fluorescence using an IMAQ PCI-1409 card
(National Instruments) and the high voltage was computer
controlled using a homemade LabView program (National
Instruments). The dual syringe pump (Model 260) was from
Kd Scientific (USA). The switch valve with plastic pump body
(CHEMINERT2) was from VICI (Valco Instruments,
Switzerland). Inner diameter of the tube for the peristaltic
pump was 0.13 mm. The internal diameter of the plastic
connection tube was 0.25 mm. The volume of the glass syringe
(Hamilton) was 100 ml.
Fabrication of the microchips
The microchip of PET substrate was fabricated using an
excimer laser (Argon Fluor Excimer Laser operating at 193 nm
(200 mJ, 50Hz) from Lambda Physik LPX 2051, Go¨ttingen,
Germany). The fabrication process is similar to that reported
elsewhere.12 At first, a main separation channel (longitudinal)
was photoablated with two cavities (reservoirs) at both ends.
Then, two transversal sample channels (1 cm long), upward
and downward with respect to the main channel, are
photoablated vertically with a typical center-to-center distance
of 200 mm between them to form the double-T injection design.
In order to reduce flow resistance (back pressure), the sampling
channels are 4 times wider than the main channel. In the case of
chip with multi separation channels, the upward sample
channel was aligned with respect to the bottom separation
channel, while the downward sample channel to the top one. It
has to be pointed out that in this case, the junctions between the
separation channels and the sample channels have a double
depth. However, no negative effect of this depth difference on
the separation has been observed. Finally, the microchannel is
thermally bonded with PE using a lamination apparatus
(Morane pneumatic Senator Laminator, Oxford, UK).
This journal is ß The Royal Society of Chemistry 2004
Fig. 1 Sample introduction by pressure-induced push/pull flow using
a switch valve. The solid line connection on the valve represents the
sampling state and the dotted line the stand-by state.
Sampling process
First the sample flow in the sampling loop was initiated to
make sure the double-T region in the main channel is filled.
Then the flow in the separation channel was started
immediately after the sampling loop was switched to the
stand-by position (Fig. 1). The polymer electrophoresis chip
was fixed with plastic blocks through which the plastic tubes
were connected with the valve and the pump. Safety warning:
attention should be paid to the sampling loop since high
voltage leakage may occur through the sample solution and
connection tubes even though the valve body and connection
tubes are all made of plastic.
Results and discussion
The advantages of using synchronized push and pull pressures
is that sample flooding in the separation channel can be
minimized and the sample dilution by the running buffer is
avoided. When a well-defined sample plug has been formed
(Fig. 2a), the injection valve is switched to equalize the pressure
on both sides of the double-T and a high voltage field is applied
along the separation channel. An electroosmotic flow is
established which drives the sample plug along the separation
channel (Fig. 2b and c). It can be observed that a substantial
dilution takes place in the sample waste channel following the
separation step; this phenomenon is caused by a pressure
gradient between the open reservoir in the separation channel
and the waste reservoir in the sample introduction channel.
An important parameter for the shape of the sample plug is
the push/pull flow ratio, r. As shown schematically in Fig. 3
three situations can be encountered. For r v 1 the plug size is
reduced as part of the running buffer in the main channel is
pulled out of the channel (Fig. 3a). In Fig. 3b the push and pull
pressures are equal (r ~ 1) and the buffer in the separation
channel is left unchanged. Finally, for r w 1 (Fig. 3c) the
sample plug increases over time as more sample is pushed into
the channel than pulled out of it. Depending on the value of r,
different sample amounts are introduced into the separation
channel. It is thus in principle possible to effectively control
sample shape and amount by tuning the push/pull pressure
ratio. In practice, it is usually advantageous to work with a
small degree of self-pinching (r v 1) to prevent for sample
Fig. 2 Images of sample introduction by push/pull flows. a, b and c show
the injection and switch process respectively. The sample flow direction is
from bottom to top and migrated to the right. The microchannels are
50 mm wide and 50 mm deep. Fluorescein was used as indicator dye.
Fig. 3 Three sample plug patterns corresponding to the push/pull
sample introduction at different flow ratios. Sample flow direction from
down to top; a: r v 1; b: r ~ 1; c: r w 1, where r is the flow rate ratio of
push and pull flows respectively.
flooding into the separation channel and thereby a time
independent sample plug.
Electro- and push/pull sample introduction were compared
under similar experimental conditions (i.e. using the same chip
and separation voltage). The separation of fluorescein and
fluorescein biotin present in the sample was carried out. For
electroinjection, the injection voltage is 400 V. For both
injection methods, the injection time was 60 s and a separation
voltage of 800 V is applied. The EOF in the channel was in the
direction of the electric field (i.e. positive to negative along the
separation channel) fluorescein therefore reaches the detection
point after fluorescein biotin since their electrophoretic
migration and the EOF have opposite directions. The
migration time, t, is give by eqn. (1):
t~
L
L
~
vep zveof (uep zueof )E
(1)
where vep and veof are the electrophoretic velocity and the
electroosmotic velocity of the sample respectively, uep and ueof
are the electrophoretic and the electroosmotic mobility
respectively, E is the electric field strength and L is the detector
position.
The push/pull sample introduction (the flow rate was
0.4 ml min21 in the sampling channel) results in well-defined
peaks illustrating an effective separation of fluorescein and
fluorescein biotin as shown in Fig. 4. For six sequential
injections of 2 mM fluorescein, a relative standard deviation
(RSD) of 4.8% and 1.5% have been obtained corresponding to
the peak height and the migration time of the signal
respectively, which implies a satisfactory reproducibility.
Meanwhile, the same experiment using the electroinjection
Fig. 4 Fluorescein biotin and fluorescein electropherograms obtained
by chip capillary electrophoresis with a push/pull flow rate of
0.4 ml min21 (syringe pump). The running buffer was a phosphate
buffer (20 mM, pH ~ 7); separation voltage is 800 V over 22 mm long
channel (100 mm wide, 45 mm deep; double-T injector with a center-tocenter distance of 200 mm long).
Lab Chip, 2004, 4, 512–515
513
Fig. 5 Electrophoregram of fluorescein and fluorescein biotin
obtained in a long microchannel (13 cm long with a width of 50 mm
and a depth of 50 mm). The applied voltage was 4 kV over 13 cm long
channel with an effective separation channel length of 11 cm (a). The
‘‘Swiss roll’’ structure is shown in (a) and the separation of fluorescein
and fluorescein biotin (both 0.3 mM in 20 mM phosphate buffer of
pH ~ 7) is shown in (b).
has shown a lower signal sensitivity due to the sample dilution
effect and a poor reproducibility (i.e. RSD of migration time up
to 20%). The difference in migration times can be ascribed to
differences in the EOF, which could be due to different degree
of surface oxidation in the laser ablation process.13,14 The
reproducibility in the case of electroinjection can, at least
partially, be related to a high voltage induced sample bias
during injection, but other factors also plays a role.7 For
instance electroinjection is also dependent on the nature of the
channel surface (i.e. the zeta potential), non-constant Laplace
pressures, siphoning effects and solvent evaporation from the
sample reservoirs.7
One of the advantages of photoablation microfabrication is
that the length of the separation channel can be increased without
significantly increasing the fabrication costs. Fig. 5b shows the
separation of fluorescein and fluorescein biotin using the push/
pull sample introduction method in a long channel, i.e. with a
total length of 13 cm. In this case, the setup is a bit different as the
flexible polymer substrate is bend in a roll structure so that the
inlet and outlet are fixed on the same base as shown in Fig. 5a.
The three-dimensional roll structure is practical when the
separation chip is to be incorporated into an actual device.
Upon six repeated experiments, a RSD of 11.5% and of 2.4% are
obtained corresponding to the fluorescein signal height and to its
migration time respectively, which illustrates that the push/pull
sample introduction procedure is also feasible in combination
with a long separation channel. The main point to emphasize is
that even for the relatively long separation time, no sample
leakage into the separation channel was observed.
In order to explore the possibilities of using the push/pull
method in a high throughput format, a separation chip
featuring multi-parallel channels was fabricated (Fig. 6a; the
parallel separation channels are separated by 200 mm). By
loading the sample across the channel array, well-defined plugs
could be formed as demonstrated by the fluorescent image in
Fig. 6b. Similar to that is shown in Fig. 2c, when the valve is
switched, the sample is injected in the multichannel and then
separated, as shown in Fig. 6c. The array channels shared the
same inlet and outlet reservoirs; the same electrical field for
separation was thus applied for all the channels. As for the
single channel polymer microchips, effective separation of
fluorescein and fluorescein biotin has been observed with a
RSD of 2.8% (peak height) and 6.0% (migration time) between
the four separation channels. The multichannel chip thus
demonstrates that simultaneous parallel analysis is in principle
possible using the present sample injection method.
In a further developed system, it can be envisioned that each
channel is optimized to separate and analyze a specific analyte.
In this way a complete sample analysis can be conducted
employing only one microchip thus significantly reducing
analysis time and required sample amount.
In conclusion, a new sample introduction method based on
push/pull pressure induced flow has been developed and tested
on polymeric microfluidic chips. The new method has a number
of advantages compared with the more conventional procedure
Fig. 6 Channel-array chip with a double-T injector structure for push/pull pressure flow injection; the 4 separation channels share the same inlet
and outlet reservoir through which the separation voltage was applied (a). (b) shows the fluorescent image of the sample loading employing the pushpull procedure, and (c) represents the image of the separation.
514
Lab Chip, 2004, 4, 512–515
based on electro-injection. First of all, no sample bias due to
charge discriminations is introduced during sample injection and
no disturbance of the buffer in the separation channel occurs. Due
to the use of simultaneous push and pull flows, the sample dilution,
which is observed for pinch injection mode, is avoided. The quality
of the signal is independent of injection time and the reproducibility upon repeated sample injections is comparable to the results
obtained using electroinjection under similar conditions. In
addition, it is possible to control the plug size and introduce
sample pinching by tuning the push/pull pressure ratios.
When using a long separation channel, the advantages of
using a flexible polymeric substrate is clearly illustrated as the
chip can be bend without affecting the separation performance.
The present method was also extended to a novel microchip
design featuring a series of parallel separation channels, which
could be utilized for studying complex (for instance biological)
samples. The advantage is that only one sample introduction is
needed for carrying out simultaneous independent multianalyte
detection, thus reducing dead volume (and thereby required
sample amount) as well as the analysis time.
Although the present investigation has only focused on
electroosmotic flow in the separation channel, it is easy to envisage
that the method will also work for pressure-induced flow as
encountered in flow injection analysis or chip chromatography.8 It
can thus be stated that the flexible sample introduction method
described in this paper has the potential to find wide general
applications for analytical devices incorporating microfluidic chips.
Acknowledgements
Dr Jacques Josserand (EPFL) and Dr Joel S. Rossier
(Diagnoswiss) are thanked for fruitful discussions and Valerie
Devaud is gratefully acknowledged for technical support. The
EU is acknowledged for financial support through the project
‘‘MicroProteoMics’’ (grant no. QLG2-CT-2001-01903 / OFES
01.0182-2).
References
1 D. R. Reyes, D. Lossifidis, P.-A. Auroux and A. Manz, Anal.
Chem., 2002, 74, 2623–2636.
2 P.-A. Auroux, D. Lossifidis, D. R. Reyes and A. Manz, Anal.
Chem., 2002, 74, 2637–2652.
3 J. P. Alarie, S. C. Jacobson and J. M. Ramsey, Electrophoresis,
2001, 22, 312–317.
4 L. M. Fu, R. J. Yang, G. B. Lee and H. H. Liu, Anal. Chem., 2002,
74, 5084–5091.
5 S. C. Jacobson, R. Hergenroder, L. B. Koutny, R. J. Warmack and
J. M. Ramsey, Anal. Chem., 1994, 66, 1107–1113.
6 X. Bai, H. J. Lee, J. S. Rossier, F. Reymond, H. Schafer,
M. Wossner and H. H. Girault, Lab Chip, 2002, 2, 45–49.
7 H. J. Crabtree, E. C. S. Cheong, D. A. Tilroe and C. J. Backhouse,
Anal. Chem., 2001, 73, 4079–4086.
8 E. Chmela, M. T. Blom, J. G. E. Gardeniers, A. van den Berg and
R. Tijssen, Lab Chip, 2002, 2, 235–241.
9 Y. H. Lin, G. B. Lee, C. W. Li, G. R. Huang and S. H. Chen,
J. Chromatogr., A, 2001, 937, 115–125.
10 Z. L. Fang and Q. Fang, Fresenius’ J. Anal. Chem., 2001, 370, 978–
983.
11 D. Solignac and M. A. M. Gijs, Anal. Chem., 2003, 75, 1652–
1657.
12 M. A. Roberts, J. S. Rossier, P. Bercier and H. H. Girault, Anal.
Chem., 1997, 69, 2035.
13 F. Bianchi, F. Wagner, P. Hoffmann and H. H. Girault, Anal.
Chem., 2001, 73, 829–836.
14 F. Bianchi, Y. Chevolot, H. J. Mathieu and H. H. Girault, Anal.
Chem., 2001, 73, 3845–3853.
Lab Chip, 2004, 4, 512–515
515