Microfluidic chip accomplishing self-fluid replacement using only capillary force and its bioanalytical application.

Abstract

A polymer microfluidic chip accomplishing automated sample flow and replacement without external controls and an application of the chip for bioanalytical reaction were described. All the fluidic operations in the chip were achieved by only natural capillary flow in a time-planned sequence. For the control of the capillary flow, the geometry of the channels and chambers in the chip was designed based on theoretical considerations and numerical simulations. The microfluidic chip was made by using polymer replication techniques, which were suitable for fast and cheap fabrication. The test for a biochemical analysis, employing an enzyme (HRP)-catalyzed precipitation reaction, exhibited a good performance using the developed chip. The presented microfluidic method would be applicable to biochemical lab-on-a-chips with integrated fluid replacement steps, such as affinity elution and solution exchange during biosensor signaling.

Full text

Page 1

Analytica Chimica Acta 585 (2007) 1–10
Microfluidic chip accomplishing self-fluid replacement using
only capillary force and its bioanalytical application
Kwang Hyo Chunga, Jung Woo Hongb, Dae-Sik Leea, Hyun C. Yoonb,∗
aBiosensor Team, ETRI, Daejeon 305-350, Republic of Korea
bDepartment of Molecular Science & Technology, Ajou University, Suwon 443-749, Republic of Korea
Received 3 November 2006; received in revised form 30 November 2006; accepted 5 December 2006
Available online 13 December 2006
Abstract
A polymer microfluidic chip accomplishing automated sample flow and replacement without external controls and an application of the chip for
bioanalyticalreactionweredescribed.Allthefluidicoperationsinthechipwereachievedbyonlynaturalcapillaryflowinatime-plannedsequence.
For the control of the capillary flow, the geometry of the channels and chambers in the chip was designed based on theoretical considerations
and numerical simulations. The microfluidic chip was made by using polymer replication techniques, which were suitable for fast and cheap
fabrication. The test for a biochemical analysis, employing an enzyme (HRP)-catalyzed precipitation reaction, exhibited a good performance using
the developed chip. The presented microfluidic method would be applicable to biochemical lab-on-a-chips with integrated fluid replacement steps,
such as affinity elution and solution exchange during biosensor signaling.
© 2006 Elsevier B.V. All rights reserved.
Keywords: Capillary force; Microfluidics; Polymer chip; Surface tension; Bioanalysis
1. Introduction
The concept of point-of-care-testing (POCT) came into the
spotlight for lifetime healthcare [1]. Recently, portable diag-
nostic systems using lab-on-a-chip technology have received
great interest. For the portability of such a system, it needs
to be small and light, thus all the components in the sys-
tem are required to be miniaturized. This concern is especially
significantinmicrofluidiccontrolsforlab-on-a-chips.Inthelab-
on-a-chips for biochemical applications, various microfluidic
handlings of liquid samples, e.g., pumping, stopping, mixing,
washing and so on, are required for each purpose. A number of
methods have been proposed for such microfluidic actuations,
in which various driving mechanisms, such as electric, mag-
netic,mechanical,chemical,andopticalones,havebeenutilized
[2–8].
However,thesemethodsformicrofluidicactuationsgenerally
includeexternalcontrolunits,whicharehardtobeminiaturized.
Therefore,toattainportablediagnosticsystemforPOCT,afluid
∗Corresponding author. Tel.: +82 31 219 2512; fax: +82 31 219 2394.
E-mail address: hcyoon@ajou.ac.kr (H.C. Yoon).
actuation method requiring smaller and lighter control unit is
necessary, and a method with no additional control unit would
be the best. This would be possible using a natural force for the
fluidic control. Surface tension is a natural force acting on the
free surface of liquid, and it generates capillary pressure when
the liquid contacts solid surface in a capillary. In a miniaturized
microfluidicsystem,thesurfacetensionbecomesdominantover
thebodyforce[5,9]andisimportantformicrofluidicactuations.
Arepresentativedeviceusingonlycapillaryflowforfluidichan-
dling is the strip-type diagnostic kits such as pregnancy test kit.
However, these kits utilize the capillary flow only for unidirec-
tional sample movement and mixing. Exact and complex fluidic
handlings cannot be achieved by using such kits, and it is neces-
saryandusefultofindatechniquetorealizecontrolledcapillary
flow.
The surface tension and accompanying capillary force have
pros and cons for fluid movement and control. As was men-
tioned above, it needs no external flow control unit because the
applied fluid flows naturally through the flow circuit. However,
the capillary force is very sensitive to surface conditions (e.g.,
contact angle) and fluidic circuit configurations. Thus, princi-
ples of capillary flow must be well understood and employed
for the design of target devices [10,11].
0003-2670/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.aca.2006.12.012

Page 2

2
K.H. Chung et al. / Analytica Chimica Acta 585 (2007) 1–10
Some theoretical analyses and applications of capillary flow
within microfabricated devices can be found in the literatures
[12–27]. Examples are passive stopping using surface modi-
fication or geometrical modulation of flow channels [12,13],
controllingtimedurationofcapillaryflow[14–18],preventionof
clogging by bubble induction [19], design of chamber for com-
pletecapillaryfilling[20],capillarymixing[21],dispensingand
metering[22–24],electro-wetting[25,26],andtheMarangonian
flow pumping [27]. However, in these microfluidic devices, the
utilization of capillary force has been limited to simple move-
ment and stopping of liquid samples. Especially, to the best of
our knowledge, the fluid replacement (exchange) in a chamber
without involving active components like pump and external
control unit has not been reported. Many steps in biological and
chemicallab-on-a-chipsneedthereplacementofliquidinareac-
tion chamber. An example is the washing step after biospecific
bindingsforaffinityseparationorbiosensing[28].Inthecurrent
study,wetriedtorealizethefluidreplacementafterreactionina
chamberusingonlycapillaryforceinasinglepolymermicroflu-
idic chip. To achieve this goal, capillary actions like passive
stopping, confluence, and flow retardation/acceleration control
have been simultaneously considered with the aid of theoretical
equations, numerical simulations and microfluidic experiments.
For the application of the device in biochemical analysis, a
model enzyme-catalyzed precipitation reaction was conducted
on the microfluidic chip. The result of the application needed
to be simply observed with bare eyes for the possible extension
to POCT devices. In order to bring out the result in the desired
format,thehorseradishperoxidase(HRP)-catalyzedprecipitate
formation reaction with 4-chloro-1-naphtol (4-CN) was used
[29,30]. In the presence of HRP and hydrogen peroxide, soluble
4-CN substrate transforms into an insoluble blue precipitate,
benzo-4-chlorocyclohexadienone,thatispossibletobeobserved
visually (Scheme 1). This particular reaction was used because
it is simple, persistent and the result can be seen without special
instrumentation. The signaling performance from the chip can
be enhanced by employing a scalable detector such as light-
emitting diode (LED) and photodiode couple without affecting
significantly its size.
Here, we described the design, fabrication and microfluidic
experiments for a single-use self-fluid replacement chip. Theo-
reticalequationsforthecapillaryflowandnumericalsimulations
were utilized for the design of the microfluidic chip. Polymer
replication techniques were developed for the cheap fabrica-
tion of disposable microfluidic chips. Exemplary experiments
for the self-fluid replacement in the chip were conducted, and
an application of the fabricated microfluidic chip by using a
model enzyme reaction was also performed.
Scheme1. PrecipitationreactioninducedbyHRP.Thereactionproduct,benzo-
4-chlorocyclohexadienone, forms an insoluble blue precipitate.
2. Experimental
2.1. Contact angle measurement and simulation for chip
design
The static contact angle was measured by a commercial con-
tact angle measurement system (Phoenix 300, Image PRO 300
v.2.0, SEO), which consisted of CCD camera, micro-droplet
dispenser and image analyzing software. The temporal varia-
tionsofthecontactangleswereexaminedfortheliquidsamples
used in fluidic experiment and biochemical analysis. The liq-
uid samples of two colors were made by mixing deionized
(DI) water with blue and red dyes (1512, Monami) (0.1wt.%)
for fluidic experiment, and three liquid samples for biochem-
ical analysis (see, Section 2.3) were utilized. Base substrates
for contact angle measurement were poly(methyl methacrylate)
(PMMA) sheet (IF850, LG MMA) and a pressure sensitive
adhesive (PSA)-coated poly(ethyleneterephthalate) (PET) film
(Crystalex,GMP),whichwereusedforpolymerchipfabrication
inthisstudy.Theliquiddropletvolumeof8?Lwasusedforcon-
tact angle measurements, and data from five independent tests
wereaveragedaftereliminatingminimumandmaximumvalues.
All measurements were performed at room temperature (20◦C)
under ambient condition. The flows in the chip were observed
and characterized using a CCD camera and a camcorder.
Numerical simulations on the fluid flow in the designed chip
wereconductedusingCFD-ACE+(CFDRCCorp.),anadvanced
multi-physics solver. The volume of fluid (VOF) method was
used to find the shape of liquid meniscus of capillary flow. The
liquid properties except the contact angle were adopted from
the value of water. Because the VOF scheme is transient and
slow, the grid arrangement and the time step were optimized for
the stability and the speed of the simulations. The total num-
ber of grid was set to under 50,000. Parallel computation using
fourPentium-IVCPUswasemployedtoenhancethesimulation
speed.
2.2. Microfluidic chip fabrication
The designed chip was fabricated using a polymer because it
is suitable for disposable chip production. Microchannels were
molded on the polymer sheet by hot embossing (imprinting)
with a mold master made of Ni. The height of embossing pat-
terns on the Ni master was set to a constant value (100?m) to
ease the fabrication process. The fabrication steps and a resul-
tant master are shown in Fig. 1. For fabrication, a 4-in. Ni plate
of 3mm-thickness was polished by CMP (chemical-mechanical
polishing) process to make the plate to be uniform in thick-
ness. SU-8-100 photoresist (MicroChem Corp.) was patterned
ontheNiplateasthemaskinglayerforelectroplating.TheSU-8
process included spin-coating, soft baking, UV exposure, post-
exposure baking, development, hard baking and relaxation. We
followedtheprocessguidelineofmanufacturerasthefabrication
condition of SU-8. The reliable adhesion between the Ni-plate
and SU-8 was necessary during whole fabrication processes,
and it was achieved by relaxation steps after soft and hard bak-
ings. The electroplating was conducted using the Ni-plate as the

Page 3

K.H. Chung et al. / Analytica Chimica Acta 585 (2007) 1–10
3
seed layer. The overgrown Ni and SU-8 layers were polished
by CMP to make the height of Ni pattern constant (100?m).
Finally, the masking layer of SU-8 was removed by dipping the
master in acetone solution for 24h. The fabricated Ni-master
is shown in Fig. 1b. The photo shows the selected pattern in
this study among several patterns on a Ni-master. We measured
the height of the pattern using an alpha-step profiler, and found
that the height was uniform over the entire region of the Ni-
master.
Fig. 2 shows the fabricated polymer microfluidic chip.
Microchannels were made by imprinting the channel pattern
Fig. 1. (a) Schematic presentation of the master fabrication steps and (b) the
fabricated nickel master for hot embossing. Several patterns were fabricated on
a nickel plate of 4-in.-diameter, and the picture shows the selected pattern used
in this study.
Fig. 2. Microfluidic chip for self-fluid replacement. Parts in the microfluidic
chip include (1) inlet of liquid 1, (2) inlet of liquid 2, (3) air vent, (4) reaction
chamber, (5) stop valve 1, (6) stop valve 2, (7) side time retardation channel,
and (8) waste chamber. The diameters of inlets are 10mm.
onto one face of a polymer plate (PMMA of thickness 1mm)
with the fabricated Ni-master, and then the imprinted plate was
covered and sealed with a polymer film (PSA-coated PET film
of thickness 100?m). A roller was used in the film lamination
step for a bubble-free sealing. Before covering the channel with
aPETfilmlid,circularholeswereformedintheimprintedpoly-
mer plate to introduce the samples (two inlets) and to vent air
(four air vent). Also, the imprinted polymer plate was sawed
around the boundary of chip section using mechanical sawing
machine. Then, it was rinsed with ethanol and DI, and dried by
airblowing.Thestructureandheightofpatternswascheckedby
measuring the height of embossed channel with an alpha-step
profiler. The final microfluidic chip containing all microfluidic
components has the size of 26mm×36mm×1.1mm. The flu-
idiccomponentsinthechipandtheirfunctionswillbeexplained
below (see Section 3.3).
2.3. Bioanalytical application of the microfluidic chip by
using enzyme-mediated precipitation reaction
Peroxidase (EC 1.11.1.7 from horse radish, HRP) was
purchased from Biochemika. Hydrogen peroxide 34.5% was
supplied by Kanto Chemical. 4-Chloro-1-naphtol (4-CN) and
polystyrene microbeads (mean particle size: 3?m) were
acquiredfromSigma.Allotherchemicalsusedwereofthehigh-
est quality available and purchased from regular sources. For
solutions, doubly distilled and deionized water with a specific
resistance over 18M?cm was used throughout the study.
2.3.1. Enzyme-mediated precipitation reaction
Priortothechipexperiment,thesubstratesolutionforenzyme
reaction was prepared by mixing 50?L of 4-CN (50mM) and
hydrogenperoxidein200?Lofphosphatebufferedsaline(PBS)

Page 4

4
K.H. Chung et al. / Analytica Chimica Acta 585 (2007) 1–10
solution. Then, the solution was dispensed in one liquid inlet
and the HRP solution (1mgmL−1in PBS) was deposited in
another inlet. The precipitation reaction was initiated and the
change in blue color intensity in the reaction chamber was
registered.
2.3.2. Precipitation reaction with microbead-adsorbed
enzyme in reaction chamber
To avoid poor mixing of HRP and substrate solutions due
to the plug flow formation in reaction chamber, polystyrene
microbead was used for enzyme immobilization. HRP (1mg
mL−1in PBS) and microbeads were conjugated by being
stirred under 4◦C overnight. Then, 1?L solution of HRP-
microbead conjugate was smeared over the reaction chamber
of the PMMA plate and dried, before sealing it with the PET
film lid. The substrate solution was dispensed in one liquid inlet
and the PBS solution was deposited in another inlet. The color
change in the reaction chamber was recorded by using a cam-
corder. The color intensity was numericalized by Palette Build
software.
3. Results and discussion
The contact angle and the geometry of the chip become the
mostimportantparameterstoobtaintheintendedcapillaryflow.
Wecollectedthecontactangledataanddesignedthechipgeom-
etry using this contact angle data with the aid of theoretical
considerations. Then, we conducted numerical simulations to
checkthefunctionalityofthedesignedchip.Finally,flowexper-
imentandbioanalysiswiththefabricatedmicrofluidicchipwere
performed.
3.1. Contact angle evaluation and theoretical approach
First, the temporal variations of the contact angles were
monitored. Generally, the contact angle continues to decrease
as time elapses, which means that the capillary pressure in a
channel changes continuously. It should be noted that the tem-
poral variation of contact angles as well as the initial values
must be considered together to acquire the designed capillary
flows. For example, an initially stopped flow at the stop valve
could reflow after some time duration because of the change
in capillary pressure with time. On the other hand, the capil-
lary filling is more influenced by the initial values of contact
anglewhenthefillingisfasterthananotablevariationincontact
angle. Thus, temporal variations of contact angles have influ-
ence on the functions of each component in the microfluidic
chip.
We collected data regarding contact angle change with time
for four liquid samples used in this study (sample 1: dye-mixed
DI water; sample 2: HRP solution in PBS; sample 3: 4-CN+
H2O2solution; sample 4: mixture of samples 2 and 3) against
two polymer substrates used for chip fabrication (substrate 1:
PMMA plate; substrate 2: PSA-coated PET film). Fig. 3 shows
the changes in liquid contact angles. The data were registered
three times, i.e., at less than 10s, at 5min, and at 10min. The
figure indicates that the contact angles and their rate of change
Fig. 3. Temporal variation of contact angles (θ) on (a) PMMA and (b) PSA-
coated PET surfaces for reaction solutions including DI water+dye 0.1wt.%
(sample 1, rectangle), HRP 1mgmL−1in PBS (sample 2, circle), 4-CN+H2O2
(sample 3, up triangle), and the mixture of samples 2 and 3 (sample 4, down
triangle), respectively.
decreasedwithtimeingeneral.However,contactanglesofsam-
ple 1 for both substrates varied relatively little with time (3◦
during 10min). On the contrary, the values of sample 4 changed
significantly (40◦and 23.5◦during 10min against substrates 1
and 2).
Absolute values of contact angle against the substrate 2
(Fig. 3b) were always greater than those against the substrate
1 (Fig. 3a), regardless of samples. This suggested that the PSA
adhesive,whichwascoatedonPETfilm,generatedmorehydro-
phobicsurfacesthantheessentiallyhydrophilicPMMAsurface.
In principle, liquid flows forward in the channel when the con-
tact angle of channel surface is smaller than 90◦. Therefore,
surfaces of the microfluidic components should be maintained
hydrophilic for continuous liquid flow. On the contrary, a
hydrophobic surface is useful for the stopping of flow working
as a valve.
As valves to stop capillary flows at desired locations, we
utilized suddenly expanded channels in this study. By using
geometric changes, the control of surface hydrophobicity at dif-

Page 5

K.H. Chung et al. / Analytica Chimica Acta 585 (2007) 1–10
5
Fig. 4. (a) Schematic of capillary flow at expansion channel. (b) Theoretical
capillary pressure at the neck (Eq. (3)) of expansion channel with respect to
expansion angle β and the width of channel w. Here, the channel height was
fixed (h=100?m), and the contact angles were from the values for the chip of
PMMA channel substrate with PSA-coated PET lid and for the solution of DI
water+dye 0.1wt.% in Fig. 3 (θB=θL=θR=68◦, θT=87.5◦).
ferent positions was not required, which would be very hard
to achieve. To design the required flow geometry, we derived
theoretical equations for the capillary pressure in an expansion
channel. Fig. 4a illustrates the liquid movement in an expansion
channel and the parameters considered.
When the liquid interface is located at δ as shown in Fig. 4a,
the capillary pressure is
?sinα
?cosθT+ cosθB
P(δ) = σ
α
?
×
h
+cos(θL+ β) + cos(θR+ β)
w + 2sinβδ
?
(1)
.
Here,σ isthesurfacetensioncoefficient,αthecurvatureangle,β
theexpansionangle,δthewettedlengthafterneck,θ thecontact
angle, h the height of channel, and w is the width of channel.
The sub-indices T, B, L and R denote top, bottom, left, and right
channel. And, the curvature angle is
α = θ + β −π
2.
(2)
Atthechannelneck(δ=0),weobtainedthecapillarypressure
at the neck as
?sinα
?cosθT+ cosθB
Pn= lim
δ→0P(δ) = σ
α
?
+cos(θL+ β) + cos(θR+ β)
×
hw
?
.
(3)
In these equations, the liquid moves forward crossing the
channel neck when the capillary pressure at the neck (Pn) is
positive, and it stops at the neck when Pnis negative. In the
derivation of equation, we utilized the concept that pressure is
the force acting on a unit area, and assumed that the shape of
liquid interface is cylindrical.
Fig. 4b demonstrates the variation of Pnwith respect to the
changes of expansion angle β and channel width w. Here, the
channel height was fixed (h=100?m), and the contact angle
data were collected from the chip of PMMA substrate with
PSA-coated PET lid for the sample 1 (dye-mixed DI water,
σ =0.0725Nm−1, time 10s) in Fig. 3. The capillary stop pres-
sures (absolute values of negative Pn) were maximized at the
valves of β around 90◦for each width. As the width of chan-
nel increased, the maximum value of capillary stop pressure
decreased, exhibiting diminished stopping strength of the stop
valve. Because the capillary forces in flow direction acting on
topandbottomsurfacesbecamelargerforawiderchannel,total
force acting in opposite direction (stopping strength) decreased.
For example, when the width was 50?m, the maximum stop
pressure was 1.86kPa (Fig. 3b). In case w = 500?m, Pnwas
positive for all the expansion angles (β). When the expansion
angle β was less than 30◦, Pnwas always positive for all the
widths w. In these conditions, the expansion channel could not
stop the flow.
Fig. 5. Theoretical evaluation of the temporal variation in capillary pressure at
the neck of expansion channel. The contact angles are from the values for the
chip of PMMA channel substrate with PSA-coated PET lid and for the reaction
solutions including DI water+dye 0.1wt.% (rectangle), HRP 1mgmL−1in
PBS (circle), 4-CN+H2O2(up triangle), mixture of HRP 1mgmL−1in PBS
and 4-CN+H2O2(down triangle), respectively. Other parameters were fixed
(h=100?m, w = 100?m, β=90◦).

Page 6

6
K.H. Chung et al. / Analytica Chimica Acta 585 (2007) 1–10
Asmentionedabove,itisalsoimportanttoknowhowlongthe
stop valve works, because the contact angle varies continuously
with time. Fig. 5 shows the temporal variation of the capillary
stop pressure at an expansion channel. The contact angles val-
ues for four samples were selected from Fig. 3. The chip was
assumed to be made of a PMMA substrate and a PSA-coated
PET film as a lid. The theoretical equation (3) for capillary stop
pressure was used. Here, the channel dimensions were fixed
(h=100?m, w = 100?m, β=90◦). From Fig. 5, data of Pn
were negative for four samples during 10min except the case
of sample 4 around 10min, supporting that the stop valves were
effective for flow control during the test. Based on the result, we
designed stop valves with this dimension.
3.2. Microfluidic chip design
With the aid of contact angle data and theoretical consid-
erations, we designed the geometry of compartments in the
microfluidic chip. The chip was comprised of a microfluidic
substrate made of PMMA plate and a sealing lid made of PSA-
coated PET film. The microfluidic substrate had all the required
fluidic components, including two inlet ports, a reaction cham-
ber, two stop valves, four air vents, a side retardation channel
andawastechamber.Thenotationsofcompartmentsofthechip
were shown in Fig. 2. The main objective of the chip was a self-
liquid replacement in the reaction chamber after predetermined
time duration. Two liquid samples were introduced into two
respective inlet ports, and consecutive fluidic actions including
capillary filling, stopping, flow retardation, merging, and liquid
replacementwereinducedbyonlycapillaryaction.Liquid1was
introduced from the inlet 1 and filled the reaction chamber first,
andtheliquid2wasintroducedfromtheinlet2andreplacedthe
liquid 1 after predetermined reaction time duration. The reac-
tion time was controlled by manipulating the flowing time at the
side flow retardation channel. This filling time could be modu-
latedbychangingthelength,shape,andcontactangleoftheside
channel. Two expansion valves stop liquid flows at the top and
bottom of the reaction chamber. The stop valves opened when
the stopped liquid surface was merged with another liquid. The
merging of two liquid streams at valves alters the capillary pres-
sure (Pn) to a positive value. The waste chamber was designed
to give the liquid flow an enough momentum for the complete
Fig. 6. The computational fluid dynamics (CFD) simulation result on the filling and stopping of fluid 1 at the stop valve 1. The simulation was started from the initial
condition shown in the top-left panel (0s) and the evolution of flow was displayed with indicated times. The computational conditions are set for the chip of PMMA
channel substrate with PSA-coated PET lid and for the solution of DI water+dye 0.1wt.%.

Page 7

K.H. Chung et al. / Analytica Chimica Acta 585 (2007) 1–10
7
replacement. The air vents prevented the clogging of the flow
by air trapping.
To check liquid flows in the designed chip, computational
fluid dynamics (CFD) simulation using CFD-ACE+ was con-
ducted. Fig. 6 demonstrates the simulation results on the sample
filling into the reaction chamber and the stopping of sample
at the stop valve 1. Figure shows a time-course sample flow
at enlarged section of the designed pattern near stop valve 1
(expansion channel with h=100?m, w = 100?m, β=100◦).
The computational conditions were set for the chip of PMMA
channel substrate with PSA-coated PET lid and for the dye-
mixed DI water (sample 1). From simulation, the liquid filled
thechannelandchambersuccessfullywithoutairbubbles.After
0.453s from start, the liquid arrived at the stop valve 1 and
stopped. These simulation results supported that the design pro-
cess explained above was effective to the fabrication of the
present microfluidic device.
3.3. Self-fluid replacement with the fabricated microfluidic
chip
Flow experiments were conducted using the fabricated chip
as shown in Fig. 7. First, the liquid 1 (red) and the liquid 2
(blue) were dropped at inlet ports 1 and 2, respectively. The
diameters of circular inlet ports were large enough (10mm) to
minimize the capillary pressure acting backwards. The volume
of each liquid was 100?L. The liquid 2 stopped at the stop
valve 2 (Fig. 7a). The liquid 1 moved into the reaction chamber
upward by capillary action, while it stopped at the stop valve 1
that was located at the bottom of the chamber (Fig. 7b). After
complete filling the chamber, the liquid 1 approached the stop
valve 2 laterally and joined with the liquid 2 that was stopped
previously. The junction of two liquids made the stop valve 2
finished its role (Fig. 7c). Because the capillary stop pressure
wasformedthroughsurfacetension,thedestructionofmeniscus
by liquid junction resulted in the elimination of the stop valve.
This principle was also used for the removal of the stop valve
1. The merged liquid at the stop valve 2 moved through the side
channel on the left. The side channel played a role of retarding
the capillary flow to ensure the necessary reaction time in the
reactionchamber.Theliquid1flowedintothesidechannelfirst,
anditcontinuedtobereplacedbytheliquid2astimeproceeded.
Theflowofliquid2inthesidechannelwaspossiblebecausethe
viscous drag of the flow from the inlet port 2 to the side channel
was smaller than those from other paths.
After the time retardation, the liquid moving along the side
channel joined the liquid 1 that was stopped at the stop valve
1, eliminating the stop valve 1 (Fig. 7d). Then, the merged liq-
Fig. 7. Pictorial results for flow sequence in the self-fluid replacing microfluidic chip. (a) The introduction of liquid 1 (red) and 2 (blue) with liquid 2 stop at stop
valve 2, (b) the stopping of liquid 1 at stop valve 1 and filling of liquid 1 in the reaction chamber, (c) the merging of liquids 1 and 2 at stop valve 2, followed
by the movement of merged liquid into the side time retardation channel, (d) the second merging of liquids 1 and 2 at stop valve 1 after flow retardation, (e) the
flow of merged liquid into waste chamber and the replacement of liquid 1 in the reaction chamber with liquid 2, and (f) the completion of liquid replacement. (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

Page 8

8
K.H. Chung et al. / Analytica Chimica Acta 585 (2007) 1–10
uid moved into the waste chamber. As the liquid filled the waste
chamber,themomentumofliquidincreasedanddrovetheliquid
2fromtheinletport2throughthereactionchambertothewaste
chamber (Fig. 7e). In the meantime, the liquid 1 in the reaction
chamber was replaced by the liquid 2 like a washing action. The
width of the waste chamber increased along with the flow direc-
tionsothatthespeedandthevolumeoftheliquidflowincreased
(Fig. 7f). The resulting momentum of liquid, which was propor-
tionaltothevolumeandthesquareofvelocity,increasedrapidly.
In addition, it should be noted that the liquid filling the waste
chamber came mainly from the inlet port 2 through the reaction
chamber (blue color), confirming the efficient fluid replacement
in the reaction chamber. This was possible from the design that
the channel junction between the reaction chamber and the inlet
port 1 was located at the bottom and the width of connecting
path was much smaller than the width of the reaction chamber.
Thisconfigurationincreasedtheviscousdragandminimizedthe
undesired mixing problem.
This observation was also confirmed with numerical simula-
tions. Fig. 8 shows the simulation results for the flow progress
from Fig. 7d–f. We set the initial conditions in Fig. 8a with the
conditions of Fig. 7d. The side retardation channel was filled
partly with blue and red liquids, and the reaction chamber was
filledwithredliquid.Thediffusivityofthedyeswassettoavalue
of 1.0×10−8m2s−1, and the average velocity at the outlet was
¯V = 0.5mms−1. The simulations resulted in a good agreement
with the experimental results. In the replacement process, the
dominant stream was found at the reaction chamber, i.e., the
flow rates (Q1) at the side retardation channel and at the con-
necting channel to inlet 1 were negligible comparing with the
rate (Q2) at the reaction chamber (Q1/Q2=0.035).
3.4. Application to bioanalysis using enzymatic
precipitation reaction
Liquid flow could be maneuvered through the predicted path
of the fluidic chip successfully. At this point, a model study
employing bioanalytical technique was conducted by using the
fabricated chip. An enzyme-catalyzed precipitation reaction
was employed, consisting of HRP, which is frequently used as
enzyme labels in immunoassays and affinity chromatographs,
andaprecipitationreagent4-CN.Afterdroppingsamplestoinlet
ports, the solution of precipitation substrate in inlet 2 stopped
successfully at the stop valve 2. Once the HRP sample filled up
the reaction chamber by the flow from inlet 1, it made a contact
with the stop valve 2 and put the stopped flow in motion. Then,
Fig. 8. The simulation result for the self-fluid replacement. Images indicate (a) the initial situation that the red liquid filled the reaction chamber and started to flow
into the waste chamber, (b) the situation when the sample replacement is being proceeded, (c) the situation when the red liquid in the reaction chamber is mostly
replaced by the blue liquid, and (d) the completion of liquid replacement. (For interpretation of the references to color in this figure legend, the reader is referred to
the web version of the article.)

Page 9

K.H. Chung et al. / Analytica Chimica Acta 585 (2007) 1–10
9
Fig. 9. A sequential view of the enzyme (HRP)-catalyzed precipitation reaction
in the reaction chamber of the microfluidic chip (a) and the time course change
in color intensity registered in the reaction chamber (b).
both solutions flowed through the side retardation channel. It
was a concern that unwanted precipitation reaction could take
place in the side retarding channel and block the flow, which
was not the case in this test. The flow in the side retarding chan-
nel is known to be laminar, resulting in a plug flow of liquid
samples. Once the samples reached the stop valve 1 and broke
the valve, the liquid moved into the waste chamber. This step
brought about the flow of 4-CN in the reaction chamber, and the
HRP-catalyzedprecipitationreactiontookplaceinthechamber.
Thecolorintensityfromtheprecipitateformationincreasedand
leveledoffataround20min,exhibitingatypicalpatternofenzy-
maticcatalysis(datanotshown).Thereactioninitiatedwhenthe
two samples met, but additional time was required to observe
an amplified signal change. As the task of the chip was to self-
replacethesampleinthereactionchamberwithanothersample,
the plug flow in the reaction chamber was not enough to create
sufficient mixing and diffusion for the precipitation reaction to
occur. Therefore, a method of reducing the time requirement for
signal development was necessary.
Toachievethisgoal,HRP-taggedmicrobeadswereloadedin
the reaction chamber to increase the reaction area and generate
faster and amplified reaction results. Polystyrene microbeads,
conjugated with HRP, were smeared over the reaction cham-
ber prior to the PET film sealing. The substrate solution was
applied in the inlet port 1, and the PBS solution was deposited
in another inlet port 2. The color intensity changed instantly as
the substrate solution filled the reaction chamber. Fig. 9 shows
the test results and the change in color intensity in the reaction
chamber. Compared to 20min in the prior experiment without
microbeadutilization,thetimerequirementforthesignalingwas
dramaticallyreducedto3s(Fig.9b),supportingtheusageofthe
microfluidic chip for point-of-care-testing.
4. Conclusions
We have developed a polymer microfluidic chip for self-
sample exchange. The design, the fabrication process and the
results of experiment for the self-fluid replacement chip were
presented. All the fluidic actions in the chip were pre-designed
and accomplished by only capillary force. The presented
microfluidic method could be applicable to biochemical lab-on-
a-chips having washing steps or fluid replacements by simple
modifications of the design. Also, to meet the desired purposes
of fluidic actions, flow time, stopping pressure and sample vol-
ume would be adjusted by modifying the chip geometry and
channel configuration. The test of the microfluidic device was
successfully carried out by enzymatic precipitation reaction.
The performance of the bioanalytical reaction within the chip
wasenhancedbyemployingbio-functionalizedmicrobeads.An
extension of this principle by using HRP-tagged antibody for
immunoanalysis is currently underway.
Acknowledgements
This research was supported by the Korea Research Foun-
dation grant (KRF2005-041-D00258), funded by the Korean
government (MOEHRD) and also by the ERC (KOSEF) for
Advanced Bioseparation Technology.
References
[1] D.D. Cunningham, Anal. Chim. Acta 429 (2001) 1–18.
[2] D.R. Reyes, D. Iossifidis, P.A. Auroux, A. Manz, Anal. Chem. 74 (2002)
2623–2636.
[3] P.A. Auroux, D. Iossifidis, D.R. Reyes, A. Manz, Anal. Chem. 74 (2002)
2637–2652.
[4] S. Shoji, Top. Curr. Chem. 194 (1998) 163–188.
[5] C.M. Ho, Y.C. Tai, Annu. Rev. Fluid Mech. 30 (1998) 579–612.
[6] P. Mitchell, Nat. Biotechnol. 19 (2001) 717–721.
[7] D. Erickson, D. Li, Anal. Chim. Acta 507 (2004) 11–26.
[8] T. Bayraktar, S.B. Pidugu, Int. J. Heat Mass Transfer 49 (2006) 815–824.
[9] M. Gad-el-Hak, J. Fluids Eng. 121 (1999) 5–33.
[10] P.G. Wapner, W.P. Hoffman, Sens. Actuators B 71 (2000) 60–67.
[11] P.G. de Gennes, Rev. Modern Phys. 57 (1985) 827–862.
[12] J. Melin, N. Roxhed, G. Gimenez, P. Griss, W. Wijngaart, G. Stemme,
Proceedings of Transducers ’03, 2003, pp. 1562–1565.
[13] P.F.Man,C.H.Mastrangelo,M.A.Burns,D.T.Burke,ProceedingsofIEEE
MEMS ’98, 1998, pp. 45–50.
[14] D. Banerjee, J. Fragala, T. Duenas, R. Shile, B. Rosner, Proceedings of
microTAS, 2003, pp. 57–60.
[15] D.S. Kim, K.C. Lee, T.H. Kwon, S.S. Lee, J. Micromech. Microeng. 12
(2002) 236–246.
[16] F. Goldschmidtboing, R. Schlosser, S. Schonhardt, P. Woias, Proceedings
of Transducers ’03, 2003, pp. 1883–1886.
[17] E. Kim, G.M. Whitesides, J. Phys. Chem. B 101 (1997) 855–863.
[18] B.V.Zhmud,F.Tiberg,K.Hallstensson,J.ColloidInterfaceSci.228(2000)
263–269.

Page 10

10
K.H. Chung et al. / Analytica Chimica Acta 585 (2007) 1–10
[19] J. Kohnle, G. Waibel, R. Cernosa, M. Storz, H. Ernst, H. Sandmaier, T.
Strobelt, R. Zengerle, Proceedings of IEEE MEMS ’02, 2002, pp. 77–80.
[20] F.G. Tseng, I.D. Yang, K.H. Lin, K.T. Ma, Sens. Actuators A 97 (2002)
131–138.
[21] P. Paik, V.K. Pamula, R.B. Fair, Lab. Chip 3 (2003) 253–259.
[22] K. Handique, D.T. Burke, C.H. Mastrangelo, M.A. Burns, Anal. Chem. 72
(2000) 4100–4109.
[23] A. Puntambekar, J.W. Choi, C.H. Ahn, S. Kim, V. Makhijani, Lab. Chip 2
(2002) 213–218.
[24] E.Shigematsu,M.Yamada,M.Yasuda,M.Seki,ProceedingsofTAS2005,
2005, pp. 1452–1554.
[25] K.S. Yun, I.J. Cho, J.U. Bu, G.H. Kim, Y.S. Jeon, C.H. Kim, E. Yoon,
Proceedings of IEEE MEMS, 2001, pp. 487–490.
[26] J.H. Lee, C.J. Kim, Proceedings of IEEE MEMS, 1998, pp. 538–543.
[27] H.J. Lee, D.J. Fermin, R.M. Corn, H.H. Girault, Electrochem. Commun. 1
(1999) 190–193.
[28] J.S. Ko, H.C. Yoon, H. Yang, H.B. Pyo, K.H. Chung, S.J. Kim, Y.T. Kim,
Lab. Chip 3 (2003) 106–113.
[29] F. Patolsky, M. Zayats, E. Katz, I. Willner, Anal. Chem. 71 (1999)
3171–3180.
[30] H.B. Pyo, Y.B. Shin, M.G. Kim, H.C. Yoon, Langmuir 21 (2005) 166–
171.