do-it-yourself microelectrophoresis chips with integrated sample recovery

Swomitra K. MohantyDongshin KimDavid J. Beebe

Department of BiomedicalEngineering,University of Wisconsin-Madison,Madison, WI, USA

Received April 17, 2006Revised June 22, 2006Accepted June 23, 2006

Research Article

Do-it-yourself microelectrophoresis chipswith integrated sample recovery

We present a microelectrophoresis chip that is simple to fabricate using the mfluidictectonics (mFT) platform (Beebe, D. J. et al., Proc. Natl. Acad. Sci. USA 2000, 97,13488–13493; Agarwal, A. K. et al., J. Micromech. Microeng. 2006, 16, 332–340). Thedevice contains a removable capillary insert (RCI) for easy sample collection afterseparation (Atencia, J. et al., Lab Chip 2006, DOI: 10. 1039/b514068d). Device con-struction is accomplished in less than 20 min without specialized equipment tradition-ally associated with microelectrophoresis chip construction. mFT was used to build aPAGE device utilizing two orthogonal microchannels. One channel performs standardseparations, while the second channel serves as an access point to remove bands ofinterest from the chip via the RCI. The RCI contains an integrated electrode that facil-itates the removal of bands using electrokinetic techniques. The device was char-acterized using prestained proteins (Pierce BlueRanger and TriChromRanger). Sam-ples were loaded into the microelectrophoresis device via a standard micropipette. Anelectrical field of 40 V/cm was used to separate and collect the proteins. The micro-PAGE device is simple to fabricate, benefits from microscale analysis, and includes anon-chip collection scheme that interfaces the macroworld with the microworld.

Keywords: Microelectrophoresis / mfluidic tectonics / Removable capillary insertDOI 10.1002/elps.200600238

1 Introduction

Miniaturization of electrophoresis analysis via microelec-trophoresis chips has been an area of active research inrecent years [1–3]. Advantages of miniaturization includereduced material usage, compact analysis space, andhigh speed analysis. However obstacles to microelec-trophoresis chips are cost and fabrication techniques.These chips require expensive cleanroom facilities orspecialized equipment not available to most biologicalresearchers. In addition, most chips do not contain aconvenient method for sample collection, which may fur-ther discourage researchers from using the technology,as sample analysis after separation is often needed.

Historically, electrophoresis is one of the most commonmethods to separate charged molecules such as DNAand proteins [4]. These molecules have unique physical

characteristics based on size, charge, and shape thatallow researchers to separate them in the presence of asieving matrix and electric field. Typical separations aredone in either slab gels or capillaries. Slab gel electro-phoresis is a well-established technique that has beenused for decades. It has proven to be economical andyield reliable analytical results. However, separationsdone using gel electrophoresis are very time consuming(often taking several hours to complete). In the 1990s CEwas introduced, which had several desirable attributesincluding reduced separation time, reduced sample/reagent use, and automated analysis. However, CErequires expensive complex equipment and is not alwaysappropriate for all separations.

As microelectrophoresis chips are not readily available,one alternative that some researchers have employed toimprove upon traditional methods is inexpensive micro-slab gel electrophoresis systems [5]. The microslab gelsmade it possible to achieve higher electric field strengthwith lower voltages, decreasing analysis time. However,fabrication of microslab gels is cumbersome and creat-ing small loading wells is difficult. Recently a simplemethod for making microslab gels was reported [6]. Thismethod casts gels (25 6 30 6 0.6 mm) between a

Correspondence: Mr. Swomitra Mohanty, Department of Biomedi-cal Engineering, Room 2139 ECB, 1550 Engineering Drive, Madison,WI 53706, USAE-mail: [email protected]: 1608-265-9239

Abbreviations: �FT, mfluidic tectonics; RCI, removable capillaryinsert; PIBA, poly iso-bornyl acrylate

3772 Electrophoresis 2006, 27, 3772–3778

© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

Electrophoresis 2006, 27, 3772–3778 Miniaturization 3773

microscope glass slide (24 6 30 mm) and transparencyfilm. By cutting injection holes into the film, definedinjection wells in the gel were created by taking advan-tage of polymerization inhibition properties of oxygen.The gel is polymerized at the film interface where oxygenis not present. However, gel at the injection holes wereexposed to oxygen which inhibits polymerization. Onedraw back to this method is that band collection afterseparation is difficult. The size of the gel and space be-tween the bands is very small making traditional bandcutting challenging.

Researchers continue to focus on improving electropho-resis through continued miniaturization [7–13]. As men-tioned above, miniaturization improves electrophoreticanalysis by reducing material usage, analysis space, andanalysis time. Due to high surface to volume ratios at themicroscale, Joule heating and temperature gradients areless problematic allowing for higher electric fieldsstrengths (compared to microslab gels) in shorter dis-tances. In addition, the reduced space used for micro-electrophoresis allows for development of high-through-put systems that can be interfaced with commercialautomation devices. One caveat to increased miniatur-ization is that sample collection after on-chip analysisbecomes increasingly difficult. The details for off chipsample manipulations must be addressed for microelec-trophoresis chips to be fully useful. In cases whereseparation of an unknown DNA sample is conducted, aspecific band of interest is present that needs to beextracted for further analysis (for example Southern blot,PCR, DNA sequencing). The trend often observed inmicrofluidics is that the more miniaturized analysis de-vices become, the more inclined these devices are touse automated systems/custom instrumentation forsample loading and recovery. However, these automatedliquid handling systems add to the cost of microelec-trophoresis chips providing the same economical obsta-

cle to researchers that CE does. Another issue in thecost of microelectrophoresis chips is in the materialsused and fabrication methods. Using alternative materi-als and fabrication techniques can make microelec-trophoresis chips more economical and accessible to lifescience labs.

Various materials have been used to fabricate microelec-trophoresis chips including polydimethylsiloxane (PDMS),glass, and silicon [9, 14]. Each type of material providesadvantages and disadvantages with regard to ease offabrication, cost, and repeatable results (Table 1). PDMSis a class of elastomers that is made up of an inorganicsiloxane backbone and organic methyl groups [15]. Whenpolymerized, PDMS is optically clear and generally con-sidered non-toxic, making it a useful construction mate-rial for microfluidic devices involving biological material.Using a process known as soft lithography [16], it is pos-sible to make micro cell culture constructs [17], micro invitro fertilization devices [18], and electrophoresis devices[14]. Although PDMS is easy to handle, it is not optimal forelectrophoretic separations without surface modification[19]. For example the surface of PDMS attracts biomole-cules causing them to stick to the channel instead ofseparating out during electrophoresis. In addition, mold-ing PDMS requires using relief master molds. The mastermold fabrication process can take several hoursdepending on the thickness and feature size. Although amaster can be used many times to make a device, thematerials and equipment needed to make masters can becostly.

Glass and silicon are materials that work well for micro-electrophoresis devices utilizing acrylamide gel or capil-lary fluid because of their surface properties and precisefabrication processes. Glass is the material used in tradi-tional devices (PAGE electrophoresis systems), thus itschoice for microscale devices is logical. Acrylamide gelpolymerizes within glass, which helps create a uniform gel

Table 1. Comparison of microelectrophoresis fabrication methods

Fabrication technique Advantages Disadvantages

Silicon/glass micro-electromechanicalsystem (MEMS)

Sub-micrometer resolution machining,well-established techniques for repeatabledevices

Requires cleanroom and expensive specializedequipment, time- and labor-intensive,complicated device assembly

PDMS 10-mm resolution , simple fabricationmethod, easy device assembly, self-bonding material

Additional surface modification needed forelectrophoresis separations, requiresfabrication of a relief master mold that addsto material costs and construction time

mfluidic tectonics Fast/simple fabrication method, no mastersneeded, easy device assembly, nobonding step required

Highest resolution 20–50 mm, limited tophoto-definable materials


3774 S. K. Mohanty et al. Electrophoresis 2006, 27, 3772–3778

matrix for separation. For capillary separations, the glasssurface allows charged layers to develop facilitating EOF.The charged layers are important for many types ofseparations involving positive and negative charged mole-cules. Silicon is a material traditionally used in integratedcircuit chip fabrication. As a result, micromachining tech-niques have already been developed for silicon, makingmicroelectrophoresis chip fabrication achievable with littleprocess adjustment. Fabrication in silicon also allows forintegration of electronic components such as detectors,heaters, and electrodes, making it very attractive for com-plete lab-on-a-chip applications. Since silicon fabricationinvolves batch processing, lab-on-a-chip devices can bemade simultaneously on a massive scale reducing the costper chip. However, such mass produced chips are notreadily available to life science researchers. Successfuldevices have been made using glass and silicon [20, 21].However, they are not as simple to build as PDMS devices.Traditional micromachining techniques are time consum-ing and require expensive cleanroom facilities not availableto most researchers. A typical glass electrophoresis devicetakes several hours to complete. Assembly of devices isalso an issue generally requiring anodic bonding to create acompletely sealed channel. Anodic bonding requires ex-tremely clean surfaces, relatively high temperatures, andhigh voltages to achieve. It is more labor intensive thanPDMS bonding. While silicon and glass make excellentmicroelectrophoresis chips, the particular fabricationmethodology that is traditionally employed makes itimpractical for most life sciences research laboratories.

We have addressed the issues of high cost microelec-trophoresis chips and sample collection using alternativematerials and fabrication methods than those previouslydescribed. We present a microelectrophoresis chip con-taining a removable sample collection component fabri-cated using the mfluidic tectonics (mFT) platform [22, 23].The device is simple to fabricate with minimal costs andequipment requirements. The mFT platform uses poly iso-bornyl acrylate (PIBA) and liquid-phase polymerization topattern microstructures such as microfluidic channels,mixers and valves [24]. PIBA is a high glass transitiontemperature (Tg 887C) polymer which is solid at roomtemperature. It has excellent adhesion properties to var-ious substrates, such as glass, plastic, and metal. PIBAcan be polymerized by liquid-phase photopolymerization,and shows minimal change (5–10%) in volume after po-lymerization. Photopolymerization of PIBA is done quicklywith typical polymerization times ranging from 5 to 40 s.After polymerization, PIBA is stiff and optically clear. De-vices are fabricated in a matter of minutes allowing forquick design improvements if needed. This contrasts withtraditional glass and PDMS fabrication where both meth-ods take hours to complete a device. While PDMS de-

vices require a master (inverse of the desired shape) forfabrication, mFT creates the channels via direct photo-polymerization of the final device material eliminating theneed for masters. One other attractive aspect of mFT is nobonding is required for creating microchannels since theyare automatically sealed during the channel formation.

2 Materials and methods

2.1 Device overview

A scheme for the microelectrophoresis device and remo-vable capillary insert (RCI) [25] is shown in Fig. 1. The de-vice contains two orthogonal channels with RCI. EachRCI contains integrated electrodes (platinum wire) forelectric field application and sample recovery. When aband of interest is identified, it is directed down theattached orthogonal channel by activating the electrodesat the end of the channel. This method of on-chip sampleextraction has been previously reported by Burns et al.[26], giving complete details modeling electric fields at theintersection and design criteria. The sample recoverymethods described here expand on this work by addingthe RCI for easy off-chip removal.

Figure 1. Scheme of microelectrophoresis device. Twoorthogonal channels are shown with an RCI at each end.Each RCI has an integrated electrode to work as ananode or cathode during electrophoretic separation andcollection. Proteins are initially sent down a separationchannel. When a band of interest is identified and reachesthe intersection, it is directed down the access channel byactivating the electrodes in that channel and shutting offelectric field in the separation channel.

2.2 Device fabrication

Microelectrophoresis chips were fabricated with threelayers of materials: glass-PIBA-glass (Fig. 2). Two adhe-sive spacers of 0.5 cm width, 5.5 cm length, and 125 mm



Figure 2. Fabrication processfor microelectrophoresis chip.Microelectrophoresis chips arefabricated of three layers ofmaterials, glass-PIBA-glass. (a)Two adhesive spacers areplaced on the edges of the bot-tom glass. The top slide (No. 1cover slip) is drilled with 1-mmholes to provide access to thepatterned channels and to cre-ate reservoirs for sample load-ing/collection. (b) The coverglass is placed on the adhesiveforming a chamber. (c) PIBA isintroduced using capillaryaction and a (d) photo-mask isvisually aligned over the sub-strate. The substrate is exposedto UV light. (e) Any un-poly-merized PIBA is vacuum-removed, leaving an emptychannel ready for acrylamidegel. (f) The RCI is placed at theoutput of the separation channeland ends of the access channel.Acrylamide gel solution isplaced into the channel throughthe RCI. This allows the gel tonot only fill the channel but toremain in the RCI itself for po-lymerization. Once filled, thedevice is exposed to UV lightagain. This results in micro-channels and RCI filled withacrylamide gel. (g) Platinum wireis placed in the RCI directly inthe gel and acts as an electrodefor the separation process. Theinput reservoir of the channelalso has a platinum electrode inelectrical contact to serve as thecathode for the separation.



thickness (Grace Bio-Labs Secure-Seal Adhesive SA-S-1L) were placed on the edges of the bottom glass (Corn-ing 50 mm 6 75 mm plain microslide). The top slide(Fisher No. 1 premium cover glass 35650 mm) was dril-led with 1-mm holes using a diamond tip drill bit (EurotoolDiamond Bur fine/med grit 1 mm, 1500 rpm) to provideaccess holes for the patterned channels and to createreservoirs for sample loading/collection. Since the coverglass was so thin, drilling took less than 10 s per hole withminimal breakage (typically one in eight slides cracked). Aphoto-definable PIBA solution was made as described byBeebe et al. [22]. The cover glass was placed on theadhesive forming a chamber. PIBA was introduced usingcapillary action and a photomask printed on an acetatesheet with a clear field pattern was visually aligned overthe substrate. Features patterned for this work rangedfrom 500 mm to 1000 mm in width by 3 cm in length; how-ever, features sizes down to 20–50 mm can be created.The substrate was exposed to UV light (Acticure EFOSA4000) at 6.7 mW/cm2 for 17 s. PIBA was cured by UVlight, therefore areas of the mask that blocked light formthe channel shape. The ends of the channel were alignedwith the holes on the cover glass. A vacuum pump wasplaced over the holes using a vacuum cup interface(Integrated Dispensing Solutions part no. 8880264) toremove any un-polymerized PIBA leaving an emptysealed channel ready for acrylamide gel. Channels werecleaned out using deionized (DI) water to remove anyPIBA residue on the surface of the glass.

2.3 Device interface and RCI construction

Connectors to interface with the device were fabricatedout of PDMS via a coring process (no masters were used).PDMS was poured over an adhesive (Grace Bio-LabsSecure-Seal Adhesive SA-S-1L) and cured. Small 1 cm6 1 cm squares were cut and holes were bored throughcreating an access port. A one-inch piece of capillarytubing (Cole-Palmer PTFE no. 06417–31) was insertedinto the connector forming the RCI. The RCI was placedat the output of the separation channel and accesschannel (Fig. 2f). UV curable acrylamide solution wasprepared using 500 mL 25% acrylamide/bisacrylamide(37.5:1), 250 mL 1.5 M Tris-HCl at pH 8.8, 2 mL riboflavin(0.2%), 50 ml TEMED (10%) and 448 mL of DI water toyield a final concentration of 12.5%. Acrylamide gel solu-tion was placed into the channel though the RCI (Fig. 2f).Once filled, the device was exposed to UV light for 10 minat 25 mW/cm2. This resulted in microchannels and RCIswith polymerized acrylamide gel. A platinum wire wasplaced in the RCI and acted as an electrode for theseparation process (Fig. 2g). The input reservoir of theseparation channel was in contact with a platinum wire

wich served as the cathode for the separation. The elec-trode was kept 2 cm away from the loading zone to mini-mize effects on the protein markers (bubbles generatedfrom electrolysis). A drop of 0.56 Tris-borate-EDTA buffer(TBE) was placed over the reservoir once the separationprocess started to insure electric field continuity.

2.4 Device operation

Characterization of the device was done performingseparations on prestained protein markers (Pierce Blue-Ranger and TriChromRanger). TriChromRanger/Blue-Ranger consists of a stabilized and lyophilized formula-tion of seven proteins, ranging from 16.5 K to 210 K and iseasily visualized in normal light. The marker solution (1 mL)was loaded into the channel reservoir. An electric fieldstrength of 40 V/cm was applied (Thermo EC 6000Pelectrophoresis power supply) for 15 min. Typically inother microfluidic systems electrokinetic injection is usedto inject samples into a channel [27]. This setup is com-plex and requires extra work and equipment (extra powersupplies, controllers, programing) to manipulate electricfields. To simplify sample loading and make it moreaccessible to life science labs, we used standard meth-ods via a micropipette for the experiments describedhere. Experiments were visualized on a standard stereo-scope (Olympus SZX12) using Leica DFC300 FX camerafor imaging.

3 Results and discussion

3.1 Band separation/collection

All experiments were conducted using 12.5% acrylamidegel looking at band separation and post separation sam-ple collection using the RCI (Fig. 3). Figure 3a shows thetop view of the device during operation. Figure 3b showsseven bands of TriChromRanger as it separates. Thebands shown are myosin (210 K), phosphorylase B(110 K), BSA (80 K), ovalbumin (47 K), carbonic anhy-drase (32 K), trypsin inhibitor (25 K), and lysozyme(16.5 K). Figure 3c shows an intersection with theseparation and access channel. As the band approachedthe intersection, the electrodes in the collection channelwere activated (electrodes in the separation channel wereshut off) driving the band into the access channel towardsthe RCI for collection.

3.2 Band-broadening and leakage

When the band reaches the intersection, some broad-ening in the shape occurs as the band is no longer con-fined within the separation channel. In addition, band



leakage occurs into channels at the cross-section when itis not collected (Fig. 3d). These results are similar to thatreported by Lin et al. [26] and more detail regarding thiscan be found in their publication. If a band of interestcomes after a band that is not collected, most of theleaked material was removed from the path by activatingthe electrodes in the access channel and driving the

material in the direction of a “waste” RCI directly oppositeof the RCI used for sample extraction. This is done beforethe band of interest reaches the intersection. Occasionalbubbles could show up in the gel during fabrication asseen in Fig. 3c. The bubbles did not have any majorimpact on the performance of the microelectrophoresischip.

Figure 3. Separation and collection of proteins with microelectrophoresis device. (a) Top view of adevice during use. (b) Image of TriChromRanger proteins as separation is occurring. (c) Orthogonalintersection of separation channel and access channel. The band of interest approaches the inter-section. Once it is in range of the access channel, the electrodes at the end of the access channel areactivated directing the protein to the RCI where it can be collected. (d) Band leakage into the channelwas observed if a band was not collected. (e) The RCI with integrated electrode inserted into a PDMSconnector to interface with the device. (f) Acrylamide gel removed from the RCI after collecting theprotein. The protein is stored in a cylindrical shaped gel. Dark portion of the gel is BlueRanger protein.



3.3 Sample removal off-chip

Figure 3e shows an image of the RCI with the electrodecoming out of the capillary tubing. The RCI is inserted intoa PDMS connector to interface with the device. As men-tioned above, acrylamide gel was polymerized in the RCIand a positive potential was applied at the electrode todrive the protein bands towards the RCI. After the samplewas collected inside the tubing, it was removed from thechip by pulling the capillary out. The RCI provided a con-venient package to transport the protein. The gel wasremoved from the capillary by applying pressure at oneend leaving a cylindrical shaped gel. Figure 3f shows aprotein (dark color) filled gel from a capillary lying on amicroelectrophoresis device.

4 Concluding remarks

The work shown here illustrates a microelectrophoresischip with an integrated sample collection scheme that isconstructed using a simple, inexpensive fabricationmethodology. Expensive micromachining equipmentassociated with microtechnology is not needed. The onlymaterials required are glass slides, a photo-mask trans-parency, PIBA, and a UV source. The cost of these mate-rials is minimal and will allow researchers to take advan-tage of the benefits of microscale analysis. Devices arebuilt in minutes and modifications to a design are madeeasily by changing a photo-mask transparency. The costof such a mask is typically $20 US (as compared tostandard chrome masks that cost $400–$500 US).

One important point this device addresses is the issue ofpost separation sample collection without the use ofexpensive liquid handling instruments. The RCI presentedhere provides a sample recovery method that can inter-face with the macro world, allowing researchers to furtherprocess samples with traditional molecular biology tech-niques. For demonstration purposes, we have shown asimple device for 1-D single lane separation of biomole-cules. It should be noted that parallel lanes can be easilyintegrated into this platform allowing for multiple sampleanalysis. As mentioned above, various microfluidic toolshave been demonstrated using mFT. Electrophoresis andthe RCI are additional tools that can be added to the mFTtoolbox to create a versatile platform for electrophoreticand other biological analysis methods.

5 References

[1] Ugaz, V. M., Elms, R. D., Lo, R. C., Shaikh, F. A. et al., Philos.Trans. R. Soc. Lond. A 2004, 362, 1105–1129.

[2] Kan, C. W., Fredlake, C. P., Doherty, E. A. S., Barron, A. E.,Electrophoresis 2004, 25, 3564–3588.

[3] Lagally, E. T., Mathies, R. A., J. Phys. D 2004, 37, R245–R261.

[4] Vesterberg, O., Electrophoresis 1993, 14, 1243–1249.[5] Guttman, A., Ronai, Z., Electrophoresis 2000, 21, 3952–

3964.[6] Chen, X. J., Ugaz, V. M., Electrophoresis 2006, 27, 387–393.[7] Manz, A., Effenhauser, C. S., Burggraf, N., Verpoorte, E. et

al., Analusis 1994, 22, M25–M28.[8] Woolley, A. T., Mathies, R. A., Anal. Chem. 1995, 67, 3676–

3680.[9] Webster, J. R., Burns, M. A., Burke, D. T., Mastrangelo, C.

H., Micro Electro Mechanical Systems, 2000, The ThirteenthAnnual International Conference, Jan 23–27, 2000, Miya-zaki.

[10] Ugaz, V. M., Brahmasandra, S. N., Burke, D. T., Burns, M. A.,Electrophoresis 2002, 23, 1450–1459.

[11] Paegel, B. M., Blazej, R. G., Mathies, R. A., Curr. Opin. Bio-technol. 2003, 14, 42–50.

[12] Brahmasandra, S. N., Ugaz, V. M., Burke, D. T., Mas-trangelo, C. H. et al., Electrophoresis 2001, 22, 300–311.

[13] Buch, J. S., Kimball, C., Rosenberger, F., Highsmith, W. E. etal., Anal. Chem. 2004, 76, 874–881.

[14] Chen, G., Bao, H. M., Li, J. H., Chen, D., Microchim. Acta2006, 153, 151–158.

[15] Clarson, S. J., Semlyen, J. A., Siloxane Polymers, PrenticeHall, Englewood Cliffs 1993, p. 673.

[16] Xia, Y. N., Whitesides, G. M., Annu. Rev. Mater. Sci. 1998,28, 153–184.

[17] Yu, H. M., Meyvantsson, I., Shkel, I. A., Beebe, D. J., LabChip 2005, 5, 1089–1095.

[18] Clark, S. G., Walters, E. M., Beebe, D. J., Wheeler, M. B.,Biol. Reprod. 2002, 66, 312–312.

[19] Hu, S. W., Ren, X. Q., Bachman, M., Sims, C. E. et al., Elec-trophoresis 2003, 24, 3679–3688.

[20] Webster, J. R., Burns, M. A., Burke, D. T., Mastrangelo, C.H., Anal. Chem. 2001, 73, 1622–1626.

[21] Pal, R., Yang, M., Lin, R., Johnson, B. N. et al., Lab Chip2005, 5, 1024–1032.

[22] Beebe, D. J., Moore, J. S, Yu, Q., Liu, R. H. et al., Proc. Natl.Acad. Sci. USA 2000, 97, 13488–13493.

[23] Agarwal, A. K., Beebe, D. J., Jiang, H. R., J. Micromech.Microeng. 2006, 16, 332–340.

[24] Moorthy, J., Mensing, G. A., Kim, D., Mohanty, S. et al.,Electrophoresis 2004, 25, 1705–1713.

[25] Atencia, J., Beebe, D. J., Lab Chip 2006, 6, 575–577.[26] Lin, R. S., Burke, D. T., Burns, M. A., J. Chromatogr. A 2003,

1010, 255–268.[27] Xu, X., Luo, G. A., Chu, X. H., Lin, B. C., Chin. Sci. Bull. 1995,

40, 482–487.


do-it-yourself microelectrophoresis chips with integrated sample recovery

Documents