paper | lab on a chipmmb.bme.wisc.edu/research/beebepubs/beebe pubs/2367745537kim… · pumps.1–4...

Hydrogel-based reconfigurable components for microfluidic devices{

Dongshin Kimab and David J. Beebe*ab

Received 11th September 2006, Accepted 20th November 2006

First published as an Advance Article on the web 7th December 2006

DOI: 10.1039/b612995a

In situ liquid-phase photopolymerization (LP3) has been applied to the field of microfluidics to

create components within integrated systems. As an extension of LP3 technology, we present

reconfigurable components that utilize the swelling of hydrogels. These components can be

conveniently used to enhance microfluidic functions and applications. In order to utilize the

swelling characteristic of hydrogels to the fullest, we demonstrate strategies to increase the

swelling performance temporally and spatially. To this end, two successful applications using the

reconfigurable components were tested: (1) active walls to block or divert flow at different steps in

the fabrication or assay process, and (2) delivery pistons to move objects to specific locations

within the microchannels after device fabrication.

Introduction

Hydrogels belong to a class of polymeric materials that swell

upon hydration. Recently, hydrogels have been used in

microfluidic systems to provide actuation for valves and

pumps.1–4 Also, hydrogels possess many characteristics that

resemble or mimic biological systems and microenvironments,

such as the cross-linked nature of the extracelluar matrix, the

swelling behavior of biological tissues, and tension or spring

forces generated by muscles.5–7 Specifically, the swelling

phenomenon of stimuli-responsive hydrogels may aid in

understanding biological functions like muscle movement8–10

or plant motion.11,12

Fabrication of reconfigurable components is another area

where hydrogels can prove useful. The volumetric swelling

behavior inherent to stimuli-responsive hydrogels can be

utilized as functional components to reconfigure elements in

a system. In general, once a microdevice is completed via

traditional microfabrication methods, it is difficult to reconfi-

gure the device by, for example, changing the microchannel

path, relocating certain components, or sealing off areas.

Using active components as part of the system assembly

process provides new options typically not available via

traditional passive assembly.

In this paper, several basic characteristics of hydrogel

swelling will be discussed along with corresponding applica-

tions that employ hydrogel-based reconfigurable components.

Theory

To realize the utility of hydrogels, it is necessary to understand

their swelling behavior. Thus, swelling characteristics of

hydrogels are discussed and two reconfigurable components–

active walls and a delivery piston–are subsequently proposed.

Radial swelling vs. longitudinal swelling

While the basic swelling kinetics of hydrogels have been well

studied,13–20 the swelling behavior of hydrogels in different

shapes has barely been examined.2 To date, most hydrogel

kinetics data are based on circular hydrogel disks or on a

macro scale. To realize diverse microfluidic functions that

utilize the swelling of various shapes of hydrogels, additional

information regarding the relationship between geometry and

swelling is important. The swelling behavior of the hydrogel is

expected to vary significantly from a circular hydrogel disk to

a long, rectangular hydrogel bar when both components are

constricted by the top and bottom of the channel. Fig. 1 shows

the radial and longitudinal swelling for the hydrogel disk and

bar, respectively, when the diameter of the hydrogel disk and

the length of the hydrogel bar are identical. Both structures are

constricted between a channel top and bottom, thus preventing

swelling towards the top or bottom. The rate of the ionic shell

formation indicating the diffusion of the stimuli is the same

with both the hydrogel disk and hydrogel bar. However, the

smallest dimension of the hydrogel bar (i.e., width) is much

smaller than that of the hydrogel disk (i.e., diameter), and

therefore the swelling time of the hydrogel bar, is expected to

be considerably faster than that of the hydrogel disk.21,22

Channel height can also affect the swelling behavior, as

shown in Fig. S1 (ESI).{ According to Hooke’s law, the

magnitude of the normal reacting force (Fn) onto the hydrogel

bar is proportional to the amount of constricted perpendicular

swelling (dy) if it is assumed that the swollen hydrogel behaves

like a Hookean material and has negligible weight.13

Additionally, the frictional force (Ff) on the top and bottom

is proportional to Fn (i.e. Ff = mFn, where m is the frictional

coefficient). If there is no friction present on the top and

bottom, the volume of swollen hydrogel bar (V) can be

represented as:

V = wlh = nV0 = nw0l0h0 (1)

aDepartment of Mechanical Engineering, University of Wisconsin,Madison, WI 53706, USAbDepartment of Biomedical Engineering, University of Wisconsin, 2142Engineering Centers Building, 1550 Engineering Drive, Madison,WI 53706, USA. E-mail: [email protected]; Tel: (608) 262-2260{ Electronic supplementary information (ESI) available:Supplementary figures. See DOI: 10.1039/b612995a

PAPER www.rsc.org/loc | Lab on a Chip

This journal is � The Royal Society of Chemistry 2007 Lab Chip, 2007, 7, 193–198 | 193

where w, l, h, and n are width, length, thickness, and the extent

of volume change, respectively. Variables with naught sub-

scripts represent the initial values. If we use a for aspect ratio

of width over length of the hydrogel bar, which is assumed to

be same for both the initial and swollen states, eqn (1) becomes

eqn (2):

V = al2h = nal20h0 (2)

Since h equals h0 when the hydrogel bar is constricted by the

channel top and bottom, the longitudinal swelling distance

(dx) is:

dx~l{l0~ffiffiffi

np

{1� �

l0 (3)

proving that dx is independent of h. Thus, dx will theoretically

be the same for both thick and thin hydrogel bars with the

same initial length when they reach equilibrium. In reality,

however, Ff exists on both the channel top and bottom. Ff for

the thick bar is larger because of a greater dy; as such, dx of the

thick hydrogel bar is expected to be smaller than that of the

thin hydrogel bar.

Buckling and active wall

Hydrophilicity of the channel surfaces increases the frictional

stress on a contacting material in the presence of water at the

interface due to the capillary force, which ultimately affects the

hydrogel swelling.23,24 If Ff is larger than the longitudinal

swelling force or if the concentration of the stimulus is not

uniform throughout the whole perimeter, the hydrogel bar

swells irregularly and therefore buckles as shown in Fig. 2. To

avoid buckling, the hydrogel bar and microfluidic chamber

need to be optimized for linear swelling so that swelling starts

and proceeds evenly throughout the whole perimeter of the

hydrogel bar. If one end of the hydrogel bar is constricted, the

hydrogel bar swells toward the free end. This unbuckled

hydrogel bar has been used as a micro-valve.24

Intentional buckling of hydrogel bars can be used for certain

microfluidic functions. If both sides of the hydrogel bar are

constricted, the reacting force (F) at the end of hydrogel bar,

generated by longitudinal swelling, induces buckling since the

bar acts like an elastic beam under axial compression (Fig. S2{).

The buckled hydrogel bar seals the channel and can therefore be

used as a blocking component within microfluidic systems.

The critical buckling load, Pcr, (i.e., any load increase

beyond this critical value will lead to buckling) of a hydrogel

bar which has both ends constricted, is given by Euler’s

formula:

Pcr~EIp2

l2(4)

where E, I, and l are Young’s modulus of the hydrogel bar,

area moment of inertia of the cross-section, and length of the

hydrogel bar, respectively.25 According to eqn (4), the critical

buckling load of longer hydrogels is smaller than that of the

shorter hydrogels when they are made of same material and

have same cross-sectional area. As a result, the longer

Fig. 1 Comparison between radial and longitudinal swelling for the hydrogel disk and bar when the diameter of the hydrogel disk and the length

of the hydrogel bar are identical. Both structures are constricted between a channel top and bottom, thus preventing swelling towards the top or

bottom. The rate of the ionic shell formation indicating the diffusion of the stimuli is the same with both the hydrogel disk and hydrogel bar.

However, the smallest dimension of the hydrogel bar (i.e., width) is much smaller than that of the hydrogel disk (i.e., diameter), and therefore the

swelling time of the hydrogel bar, is expected to be considerably faster than that of the hydrogel disk.

Fig. 2 Buckled hydrogel bar. The hydrogel bar is constricted by the

rigid post on left. Due to uneven swelling, part of the bar buckles.

194 | Lab Chip, 2007, 7, 193–198 This journal is � The Royal Society of Chemistry 2007

hydrogels will buckle faster than the shorter hydrogels under

the same swelling condition. Thus, it is possible to program a

buckling time by geometrically adjusting the length of

hydrogels.

An active wall is an example of a microfluidic component

using the buckling phenomenon of a hydrogel bar. An active

wall can function as a blocking component that prevents or

reconfigures flow in certain channels or parts of a microfluidic

device. The active wall is comprised of two rigid posts, a

responsive hydrogel bar, and straight channel walls as shown

in Fig. S2.{ The channel is open during normal operation

(Fig. S2a) and is closed by a buckled hydrogel (Fig. S2b) upon

introduction of a swelling stimulus such as pH or temperature.

The active wall can be used to enhance fabrication quality or

improve device performance in microfluidic systems. One

example is a detour channel—created by the active wall—for

protecting certain sensitive components (e.g. filter or readout

disk) from chemicals (used during other fabrication steps) that

could degrade the component and, in turn, adversely affect

component performance.

Flipping and delivery piston

As discussed previously, the hydrogel bar can be designed to

swell longer distances as compared to the hydrogel disk.

However, the longer the bar, the easier it can buckle or flip

over, resulting in improper function. To make the hydrogel bar

swell longer distances without any buckling or flipping over in

irregular, non-repeatable directions, the hydrogel bar can be

formed to have a slanted cross-section, as depicted in Fig. S3.{As the hydrogel bar swells, the off-center reacting forces (Fn)

act on the top and bottom of the hydrogel bar; once the

moment (M) of the hydrogel bar becomes large enough to

overcome the negative moment (Mf) caused by Ff, the hydrogel

bar flips over. By exploiting this phenomenon, it is possible to

control the direction of flipping. Once the hydrogel bar has

flipped over, it does not touch the top or bottom channel

surface; the hydrogel bar can therefore fully swell and contract

with minimal hindrance (i.e., friction).

Long distance swelling can function similar to a micro-

actuator and be used as a delivery tool in microfluidic systems.

Current technologies, such as optical tweezers and dielectro-

phoretic traps which facilitate entrapment or movement of

particles, are not suitable for particles ranging beyond 10 mm

in diameter.26–29 A delivery piston that capitalizes on the

physical reorientation (90u rotation about the x-axis) of a

rectangular hydrogel bar can be employed to physically move

or reposition relatively large objects inside a microchannel

(Fig. S4{). When triggered, the hydrogel piston can push a disk

to a designated location for subsequent analysis or treatment.

The ability to relocate particles, similar to the function of a

conveyor belt in a macro-system, grants a certain amount of

autonomy to microdevices.

Results and discussion

Radial swelling vs. longitudinal swelling

Theoretically, longitudinal swelling distance of the hydrogel

bar should be equal to the radial swelling distance of the

hydrogel disk due to the same ratio of volumetric change.

Thus, the maximum swelling distance would be same for the

disk and bar if enough swelling time (i.e., hours) was allowed.

The swelling rate and swelling distance of the hydrogel bar was

faster and longer, respectively, than that of the hydrogel disk

during the 50 min swelling experiment, as anticipated (Fig. 3

and Table 1), for both 125 mm and 250 mm thick devices. As a

result, the hydrogel bar is a superior design when compared to

the circular hydrogel disk because of its smaller volume and

faster swelling. The thickness of the hydrogel also affected

swelling behavior. The swelling of the thin hydrogels (125 mm)

was greater in magnitude than that of the thick hydrogels

(250 mm), presumably due to the lower frictional force of the

thin bar owing to the smaller perpendicular swelling (dy)

(Fig. S1{). Thus, thinner hydrogels are better to achieve longer

swelling.

Buckling and active wall

Fig. S5{ shows the result of preliminary tests to investigate the

buckling time and mode of the active walls in different length.

The time before the buckling started to close the active walls

took 4 min, 10 min, and 20 min for 4.8, 3.8, and 2.8 mm

hydrogels, respectively. The length of time to seal the channel

was greater for the shorter bars, as predicted by eqn (4).

Fig. 4 shows the result of the active wall using 2.8 mm long

hydrogel bars to protect a filter during other manufacturing

Fig. 3 Comparison of radial (disk) and longitudinal (bar) swelling.

The swelling speeds of the hydrogel bars are significantly faster than

those of the hydrogel disks.

Table 1 Comparison of measured swelling distance and time betweenhydrogel bar and disk.

Thickness/mm

Swollen distance/mm aSwelling time/min

Bar Disk Bar Disk

125 205 133 10.4 38.5250 159 107 10.3 40.8a Time taken to reach 95% of the maximum swelling distance.


steps. The dark area between the top and bottom channels is

an in situ fabricated filter. The active walls blocked auxiliary

channels used during preliminary steps of the filter fabrication

process, successfully diverting flow through the filter. The

buckling of the hydrogel bars closed the auxiliary channel and

remained closed under pressures up to 207 kPa, which was the

maximum measurable pressure of the current measurement

system.

The active wall made of the temperature-sensitive hydrogel

tended to swell without buckling (Fig. S6{). The swelling and

shrinking of the temperature-sensitive hydrogel was rapid and

was complete in less than two minutes. Although it requires

additional materials for operation, such as heating and/or

cooling devices, the temperature-sensitive gels can function as

active walls or active micro-valves that are chemically resistant

to swelling or shrinking upon interaction with sample fluid.

An active wall may also function to reconfigure channel

networks or isolate designated parts of a device. In addition, if

the swelling of the hydrogel bar is adjusted to have a certain

degree of allowable pressure, the active wall may serve as a

safety valve or regulator to prevent excessive pressure build-up

inside a microfluidic chamber.

Flipping and delivery piston

As a result of flipping, a delivery piston was able to move a

disk to a designated location (Fig. 5). The hydrogel piston was

photopolymerized inside the channel to have a slanted angle;

due to the unidirectional net moment developed during

swelling, this allowed the slanted piston to flip in one direction.

The flipping occurred at the tip (Fig. 5b) and propagated

toward the base of the hydrogel piston (Fig. 5c). After

10 minutes of swelling, the 500 mm disk was moved to its

designated location by the hydrogel piston. The resultant

swollen distance of the piston was 1.8 mm, which could be

adjusted by using buffer solutions of varying pH.2 Once the

hydrogel piston had flipped over, it would not flip back to its

initial orientation even in the contracted state because there

was no reaction force to realize the negative rotation moment.

This delivery piston provides new functionality for moving

big objects such as readout disks or possibly embryos30 inside

microfluidic devices. For example, the delivery piston could be

used to reposition a functionalized disk for enzyme-linked

Fig. 4 Active walls for filter fabrication. Once the filter was

fabricated and washed (a), buckled hydrogels, functioning as active

walls, diverted flow through the filter (b). The dyed color indicates the

rinsing and sample flow.

Fig. 5 Time-lapse images of the hydrogel delivery piston. The 500 mm diameter disk was moved into the designated location by flipping of the

hydrogel piston (200 mm wide; 3.3 mm long) after 10 min swelling. The resultant swollen distance of the piston was 1.8 mm. The white dotted circles

in (b) and (c) represent the point where the bar has flipped.


immunosorbent assay (ELISA), a detection procedure used in

diagnostics that utilizes antibody specificity and enzymatic

reactions.31 One of the problems with ELISA within a

microfluidic device is nonspecific background signal from the

surface of the treatment channel, which lowers the signal to

noise ratio. One method used to address this problem has been

the use of blocking agents such as albumin. The delivery piston

could be an alternative way to enhance the signal by

repositioning the readout disk to a clean channel that has

not been exposed to the fluorophore, ultimately avoiding the

use of albumin or other blocking agents typically employed for

ELISA.

Conclusions

Swelling characteristics of hydrogels were studied to better

understand their behavior and provide insight for exploring

new microfluidic functions and components. The results

suggest that the hydrogel bar is the preferred geometry for

actuator functions than the circular design of the hydrogel disk

due to the bar’s faster swelling time. Next, methods utilizing

the elastic energy of swelling hydrogels were proposed and

demonstrated with novel reconfigurable microfluidic compo-

nents, such as active walls and delivery pistons: active walls

either reconfigured the channel network for protection of

system components (e.g. the in situ fabricated filter) or simply

redirected flow; additionally, a delivery piston repositioned a

relatively large object (500 mm disk). The reconfigurable

components demonstrated in this paper were accomplished

by combining a passive element (channels formed during LP3)

and an active element (stimuli-responsive hydrogel). These

components may be expanded upon to create other compo-

nents, such as safety or release valves, degradable blocking

components, and microfluidic actuators. The use of stimuli-

responsive materials to reconfigure existing components

further increases the options available to microfluidic

designers.

Methods

Ethylene glycol dimethacrylate (EGDMA) was purchased

from Polysciences Inc. (Warrington, PA). Isobornyl acrylate

(IBA) was obtained from Surface Specialties, (Smyma, GA).

N-isopropylacrylamide (NIPAAm), N,N9-methylenebisacryl-

amide (NMBA), and dimethylsulfoxide (DMSO) were from

Acros Organics (Geel, Belgium). 2-hydroxyethyl methacrylate

(HEMA), 2,2-dimethoxy-2-phenylacetophenone (DMPA), tet-

raethyleneglycol dimethacrylate (TEGDMA), and acrylic acid

(AA) were purchased from Sigma-Aldrich Corp. (St. Louis,

MO). IBA and HEMA were vacuum distilled before use, while

all the other materials were used as obtained.

Hydrogel bar and disk for swelling comparison

Hydrogel bars (180 mm wide, 1 mm long; 125 mm and 250 mm

thick) and disks (1 mm od; 125 mm and 250 mm thick) were

fabricated from a HEMA-based hydrogel prepolymer contain-

ing HEMA, DMPA, and EGDMA (1 : 0.03 : 0.03 weight

ratio). Channel networks and chambers to enclose the

hydrogels were fabricated from an IBA-based prepolymer

mixture consisting of IBA, TEGDMA, and DMPA (1.9 : 0.1 :

0.06 weight ratio) using the tectonics platform.32

A polycarbonate cartridge with an adhesive gasket at the

edges (125 mm or 250 mm thick for corresponding hydrogel

thickness; Grace Bio-Labs, Bend, OR) was attached to a

polycarbonate sheet fixed on a microscope glass slide and filled

with the IBA-based prepolymer. A film photomask detailing

the channels and chambers was placed on top of the cartridge

and exposed to UV light (365 nm, 12 mW cm22, 12 s) (EXFO

Acticure 4000, Mississauga, Ontario, Canada). After removing

the unpolymerized monomer material with a vacuum pump,

the channel was flushed deionized (DI) water. The channel was

then filled with DI water and exposed again to UV light

(12 mW cm22, 12 s) to obtain well-defined channel walls and

chambers. Next, the chamber was filled with the HEMA-based

hydrogel prepolymer. The photomask representing the shape

of a hydrogel bar or disk was placed on top of the device and

exposed to UV light (24 mW cm22, 60 s) with similar post-

polymerization treatment.

Radial and longitudinal swelling were compared using

hydrogel bars and disks. The prepared device was placed on

a stereoscope and swelling was initiated by introducing DI

water in the chamber. Time-lapse images were taken by a

digital camera (DFC300FX, Leica, Cambridge, UK) attached

to the stereoscope. The images were calibrated and analyzed to

quantify swelling of the hydrogel bars and disks using an

imaging software (IM50, Leica, Cambridge, UK).

Filter device employing active wall

Fig. 4 shows the design of a microfluidic filter that is protected

by active walls during other manufacturing steps. Channel

networks, chambers, and hydrogel bars were fabricated using

the technique outlined in the previous section. Polycarbonate

cartridges of 250 mm thick were used for the filter devices.

Three different lengths of hydrogel bars (2.8, 3.8 and 4.8 mm

with 200 mm width) were tested to investigate the time range of

buckling by DI water. 2.8 mm long hydrogel bars were chosen

for the filter device to have 20 minutes before the active wall

closed so that there was enough time for the porous filter

photopolymerization and 5–10 minutes of rinsing. Next, the

chamber was filled with a filter mixture containing HEMA,

DMPA, EGDMA, and water (0.22 : 0.002 : 0.008 : 0.78 weight

ratio). A porous filter was photopolymerized in situ33 by

placing a photomask on top of the device and exposing to UV

light (25 mW cm22, 4 min). DI water was introduced to rinse

unpolymerized filter material from the area. The rinsing fluid

passed alongside (not through) the fabricated filter during the

cleaning process (Fig. 4a). Twenty minutes after polymeriza-

tion, the active walls buckled and closed the auxiliary channels

(Fig. 4b). Closed active walls were investigated separately to

characterize the blocking function with a flow rate measure-

ment method developed previously.34 As an alternate

approach, a temperature-sensitive hydrogel prepolymer con-

sisting of NIPAAm, NMBA, DMPA, DMSO, and water

(0.545 : 0.015 : 0.0385 : 0.75 : 0.25 weight ratio)35,36 was used to

make active walls that were not stimulated by sample fluid.

The temperature-responsive active walls were closed by cool-

ing to less than 20 uC.


Delivery piston

Channel networks and chambers (250 mm thick) were

fabricated according to the above protocols. A base-responsive

hydrogel prepolymer consisting of AA, HEMA, EGDMA,

and DMPA (4.05 : 29.3 : 0.334 : 1 weight ratio) filled the

channel network and the hydrogel piston (200 mm wide, 3.3 mm

long) was polymerized. The piston was exposed to UV light

that was slanted at approximately 10 degrees in order to

produce a slanted shape; such a slant allowed for controlled

flipping in one direction (Fig. S6{). A flat disk (500 mm

diameter, 125 mm thick) was inserted into its initial location by

blowing with nitrogen gas. The hydrogel piston was swelled

using a 10 mM buffer, pH 10.

Acknowledgements

This research was partly supported by the US Department of

Homeland Security (DHS) (grant number N-00014-04-1-

0659), through a grant awarded to the National Center for

Food Protection and Defense (NCFPD) at the University of

Minnesota and by a grant from DARPA-MTO (F30602-00-1-

0570).

References

1 A. K. Agarwal, S. S. Sridharamurthy, D. J. Beebe and H. Jiang,Programmable Autonomous Micromixers and Micropumps,J. Microelectromech. Syst., 2005, 14(6), 1409–1421.

2 D. J. Beebe, J. S. Moore, J. M. Bauer, Q. Yu, R. H. Liu,C. Devadoss and B. H. Jo, Functional hydrogel structures forautonomous flow control inside microfluidic channels, Nature,2000, 404(6778), 588–590.

3 R. H. Liu, Q. Yu and D. J. Beebe, Fabrication and characteriza-tion of hydrogel-based microvalves, J. Microelectromech. Syst.,2002, 11(1), 45–53.

4 A. Richter, D. Kuckling, S. Howitz, T. Gehring and K. F. Arndt,Electronically controllable microvalves based on smart hydrogels:Magnitudes and potential applications, J. Microelectromech. Syst.,2003, 12(5), 748–753.

5 A. Batorsky, J. H. Liao, A. W. Lund, G. E. Plopper andJ. P. Stegemann, Encapsulation of adult human mesenchymalstem cells within collagen-agarose microenvironments, Biotechnol.Bioengin., 2005, 92(4), 492–500.

6 H. Kurihara and T. Nagamune, Cell adhesion ability of artificialextracellular matrix proteins containing a long repetitive Arg-Gly-Asp sequence, J. Biosci. Bioeng., 2005, 100(1), 82–87.

7 Y. N. Yan, X. H. Wang, Z. Xiong, H. X. Liu, F. Liu, F. Lin,R. D. Wu, R. J. Zhang and Q. P. Lu, Direct construction of athree-dimensional structure with cells and hydrogel, J. Bioact.Compat. Polym., 2005, 20(3), 259–269.

8 T. Kaneko, Development of liquid crystalline hydrogel soft-actuator working under body temperature range, KobunshiRonbunshu, 2005, 62(8), 373–379.

9 R. Yoshida, Design of functional polymer gels and theirapplication to biomimetic materials, Curr. Org. Chem., 2005,9(16), 1617–1641.

10 E. A. Moschou, S. F. Peteu, L. G. Bachas, M. J. Madou andS. Daunert, Artificial muscle material with fast electroactuationunder neutral pH conditions, Chem. Mater., 2004, 16(12),2499–2502.

11 Y. Forterre, J. M. Skotheim, J. Dumais and L. Mahadevan, Howthe Venus flytrap snaps, Nature, 2005, 433(7024), 421–425.

12 J. M. Skotheim and L. Mahadevan, Physical limits and designprinciples for plant and fungal movements, Science, 2005,308(5726), 1308–1310.

13 S. K. De, N. R. Aluru, B. Johnson, W. C. Crone, D. J. Beebe andJ. Moore, Equilibrium swelling and kinetics of pH-responsive

hydro g e l s : mod e l s , expe r im ent s , a nd s imula t i ons ,J. Microelectromech. Syst., 2002, 11(5), 544–555.

14 Q. Yu, J. M. Bauer, J. S. Moore and D. J. Beebe, Responsivebiomimetic hydrogel valve for microfluidics, Appl. Phys. Lett.,2001, 78(17), 2589–2591.

15 M. G. Olsen, J. M. Bauer and D. J. Beebe, Particle imagingtechnique for measuring the deformation rate of hydrogelmicrostructures, Appl. Phys. Lett., 2000, 76(22), 3310–3312.

16 J. Valencia and I. F. Pierola, Swell ing kinetics ofpoly(N-vinylimidazole-co-sodium styrenesulfonate) hydrogels,J. Appl. Polym. Sci., 2002, 83(1), 191–200.

17 S. K. Bajpai, Swelling–deswelling behavior of poly(acrylamide-co-maleic acid) hydrogels, J. Appl. Polym. Sci., 2001, 80(14),2782–2789.

18 X. Qu, A. Wirsen and A. C. Albertsson, Structural change andswelling mechanism of pH-sensitive hydrogels based on chitosanand D,L-lactic acid, J. Appl. Polym. Sci., 1999, 74(13), 3186–3192.

19 S. J. Kim, H. I. I. Kim, S. J. Park and S. I. Kim, Shape changecharacteristics of polymer hydrogel based on polyacrylic acid/poly(vinyl sulfonic acid) in electric fields, Sens. Actuators, A, 2004,115(1), 146–150.

20 C. Khoury, T. Adalsteinsson, B. Johnson, W. C. Crone andD. J. Beebe, Tunable microfabricated hydrogels—A study inprotein interaction and diffusion, Biomed. Microdevices, 2003, 5(1),35–45.

21 T. Tanaka, E. Sato, Y. Hirokawa, S. Hirotsu and J. Peetermans,Critical kinetics of volume phase-transition of gels, Phys. Rev.Lett., 1985, 55(22), 2455–2458.

22 E. S. Matsuo and T. Tanaka, Kinetics of discontinuous volumephase-transition of gels, J. Chem. Phys., 1988, 89(3), 1695–1703.

23 D.-I. Kim, H.-S. Ahn and D.-H. Choi, Effect of surfacehydrophilicity and water vapor pressure on the interfacial shearstrength of adsorbed water layer, Appl. Phys. Lett., 2004, 84(11),1919–1921.

24 D. Kim and D. J. Beebe, A bi-polymer micro one-way valve, Sens.Actuators, A, 2006, DOI: 10.1016/j.sna.2006.11.004.

25 R. D. Cook and W. C. Young, Advanced Mechanics of Materials,2nd edn, 1999, Prentice-Hall, Inc., Upper Saddle River, NJ, p. 481.

26 D. G. Grier, A revolution in optical manipulation, Nature, 2003,424(6950), 810–816.

27 J. Voldman, R. A. Braff, M. Toner, M. L. Gray and M. A. Schmidt,Holding forces of single-particle dielectrophoretic traps, Biophys.J., 2001, 80(1), 531–541.

28 S. Rajaraman, H. Noh, P. J. Hesketh and D. S. Gottfried, Rapid,low cost microfabrication technologies toward realization ofdevices for dielectrophoretic manipulation of particles andnanowires, Sens. Actuators, B, 2006, 114(1), 392–401.

29 D. R. Albrecht, G. H. Underhill, T. B. Wassermann, R. L. Sah andS. N. Bhatia, Probing the role of multicellular organization inthree-dimensional microenvironments, Nat. Methods, 2006, 3(5),369–375.

30 I. K. Glasgow, H. C. Zeringue, D. J. Beebe, S. J. Choi, J. T. Lyman,N. G. Chan and M. B. Wheeler, Handling individual mammalianembryos using microfluidics, IEEE Trans. Biomed. Eng., 2001,48(5), 570–578.

31 J. Moorthy, G. A. Mensing, D. Kim, S. Mohanty, D. T. Eddington,W. H. Tepp, E. A. Johnson and D. J. Beebe, Microfluidic tectonicsplatform: A colorimetric, disposable botulinum toxin enzyme-linked immunosorbent assay system, Electrophoresis, 2004, 25,1705–1713.

32 D. J. Beebe, J. S. Moore, Q. Yu, R. H. Liu, M. L. Kraft, B.-H. Joand C. Devadoss, Microfluidic tectonics: A comprehensiveconstruction platform for microfluidic systems, Proc. Natl. Acad.Sci. U. S. A., 2000, 97(25), 13488–13493.

33 J. Moorthy and D. J. Beebe, In situ fabricated porous filters formicrosystems, Lab Chip, 2003, 3(2), 62–66.

34 D. Kim, N. C. Chesler and D. J. Beebe, A method for dynamicsystem characterization using hydraulic series resistance, Lab Chip,2006, 6(5), 639–644.

35 L. Dong, A. K. Agarwal, D. J. Beebe and H. R. Jiang, Adaptiveliquid microlenses activated by stimuli-responsive hydrogels,Nature, 2006, 442(7102), 551–554.

36 H. van der Linden, W. Olthuis and P. Bergveld, An efficientmethod for the fabrication of temperature-sensitive hydrogelmicroactuators, Lab Chip, 2004, 4(6), 619–624.


Supplementary Material (ESI) for Lab on a Chip This journal is © The Royal Society of Chemistry 2006

<Supplemental figures>

Figure S1 Force diagram of a cross-sectioned hydrogel bar or disk. The normal reacting force (Fn) onto the hydrogel bar or disk induced by swelling is proportional to the amount of constricted perpendicular swelling (δy) according to Hooke’s law when the swollen hydrogel is assumed to be a Hookean material and to have negligible weight. F and Ff indicate the longitudinal swelling force and frictional force, respectively.

(a) Top view of a hydrogel bar before buckling

(b) Top view of a hydrogel bar after buckling

Figure S2 Schematic illustration of a hydrogel bar undergoing buckling. The reacting force (F) at the end of hydrogel bar - induced by swelling of the hydrogel - causes buckling since the hydrogel bar acts like an elastic beam under axial compression.

Fn

Fn F f Channel top

Hydrogel bar or disk

δy

δx

F

Channel bottom Center line

F f

Swollen size

Channel wall Hydrogel bar

Post

FF

Post


Figure S3 A free body diagram of the cross-section of the hydrogel bar. As the hydrogel bar swells, the off-centered reacting forces (Fn) apply onto the top and bottom of the hydrogel bar and eventually the hydrogel bar flips over as the moment (M) of the hydrogel bar, induced by Fn, becomes large enough to overcome the negative moment (Mf) caused by the frictional force (Ff). Because of this, it is possible to control the direction of flipping and ultimately avoid the irregular flipping and buckling.

a) Before delivery b) After delivery

Figure S4 Conceptual diagram of the delivery piston. After inserting the disk through the inlet channel, the hydrogel piston is triggered to push the disk to the designated position.

Fn

M

Fn

Ff

Channel top

Channel bottom

Ff Mf

Hydrogel bar

Hydrogel piston

Inlet channel

Destination channel

Disk


(a) Buckling of 4.8 mm long hydrogel bar

(b) Buckling of 3.8 mm long hydrogel bar

(c) Buckling of 2.8 mm long hydrogel bar

Figure S5 Buckling of hydrogel bars polymerized to have a different length. The time before the buckling started to close the active walls took 4 min, 10 min, and 20 min for 4.8, 3.8, and 2.8 mm hydrogel bars, respectively.


(a) Contracted state of temperature sensitive hydrogel at 50°C

(b) Swollen state of temperature sensitive hydrogel below 20°C

Figure S6 Image of an active wall using a temperature-sensitive hydrogel. The temperature-sensitive hydrogel tended to swell without buckling.


Figure S7 Cross-sectional view of flipping. Once the channel was fabricated, the hydrogel pre-polymer filled the channel and the hydrogel bar was polymerized in a slanted shape due to slanting of UV light with respect to position of the device.

Photo mask

Hydrogel prepolymer Polymerized Hydrogel bar

UV

paper | lab on a chipmmb.bme.wisc.edu/research/beebepubs/beebe pubs/2367745537kim… · pumps.1–4...

Documents