Surface-Modified Poly(methyl methacrylate)Capillary Electrophoresis Microchips for Proteinand Peptide Analysis

Jikun Liu, Tao Pan, Adam T. Woolley, and Milton L. Lee*

Department of Chemistry and Biochemistry, Brigham Young University, Provo, Utah 84602-5700

Polymeric materials have emerged as appealing alterna-tives to conventional inorganic substrates for the fabrica-tion of microscale analytical systems; however, nativepolymeric surfaces typically require covalent modificationto ensure optimum biocompatibility. 2-Bromoisobutyrylbromide was immobilized on poly(methyl methacrylate)(PMMA) substrates activated using an oxygen plasma.Atom-transfer radical polymerization was then performedto graft poly(ethylene glycol) (PEG) on the PMMA surface.PMMA microcapillary electrophoresis (µCE) devices madewith the covalently modified surfaces exhibited substan-tially reduced electroosmotic flow and nonspecific adsorp-tion of proteins on microchannel surfaces. Experimentsusing fluorescein isothiocyanate-conjugated bovine serumalbumin indicated that both column efficiency and migra-tion time reproducibility were 1 order of magnitude betterwith derivatized compared to untreated PMMA µCE chips.Fast, reproducible, and efficient separations of proteinsand peptides were demonstrated using the PEG-graftedPMMA µCE chips. All analyses were completed in lessthan 60 s, and separation efficiencies as high as 5.2 ×104 plates for a 3.5-cm-long separation channel wereobtained. These results demonstrate the general ap-plicability of surface-grafted PMMA microdevices for abroad range of protein analyses.

The micro total analysis system approach to chemical analysishas generated significant interest during the last two decadesbecause of low sample consumption, high speed, capability toperform complex analyses in a single device, and potential to befabricated into high-density arrays to achieve high throughput.1,2

To date, the fabrication of microdevices has largely depended onthe use of rigid inorganic substrates such as glass, silicon, andquartz. Commodity polymeric materials are less expensive forlarge-scale manufacturing of microdevices, and plastic machiningtechnologies, including injection molding, casting, and embossing,are more versatile and accessible. In fact, it is feasible to considerdisposable microfluidic devices using plastic materials, which maybenefit applications in which cross contamination can generate

unacceptable results.3,4 Unfortunately, the surfaces of manypolymeric and inorganic materials are not compatible with biologi-cal samples. Hydrophobic, electrostatic, or other interactionsattract some analytes to the surface, leading to sample adsorptionand loss and analytical irreproducibility. Therefore, it is oftennecessary to modify the microdevice surface as part of thefabrication protocol.

Dynamic coating is a simple and rapid surface modificationtechnique that has been used widely in manipulation of electro-osmotic flow and separation of proteins, DNA, or oligosaccharideswith polymeric microdevices.5-14 Using this approach, surface-active compounds or surface modifiers are introduced into theseparation buffer, and the coating materials are then physicallyadsorbed to the channel surface. Although dynamic coating isconvenient, surface modifiers can be detrimental in applicationsthat require coupling to mass spectrometry or to miniaturizedchemical reactors. Therefore, microfluidic devices with chemicallymodified surfaces are preferred.

Poly(methyl methacrylate) (PMMA) is one of the mostcommonly used polymeric substrates in microfabrication. Amongmany approaches available for chemical modification of polymersurfaces, such as graft copolymerization,15 laser activation,16 ionbombardment,17 vacuum UV irradiation,18 plasma treatment,19-25

* To whom correspondence should be addressed. E-mail: milton_lee@byu.edu.(1) Reyes, D. R.; Iossifidis, D.; Auroux, P.-A.; Manz, A. Anal. Chem. 2002, 74,

2623-2636.(2) Auroux, P.-A.; Iossifidis, D.; Reyes, D. R.; Manz, A. Anal. Chem. 2002, 74,

2637-2652.

(3) Boone, T. D.; Ricco, A. J.; Fan, Z. H.; Tan, H.; Hooper, H. H.; Williams, S.J. Anal. Chem. 2002, 74, 78A-86A.

(4) Soper, S. A.; Ford, S. M.; Qi, S.; McCarley, R. L.; Kelly, K.; Murphy, M. C.Anal. Chem. 2000, 72, 642A-651A.

(5) Doherty, E. A. S.; Meagher, R. J.; Albarghouthi, M. N.; Barron, A. E.Electrophoresis 2003, 24, 34-54.

(6) Belder, D.; Ludwig, M. Electrophoresis 2003, 24, 3595-3606.(7) Dang, F.; Zhang, L.; Hagiwara, H.; Mishina, Y.; Baba, Y. Electrophoresis

2003, 24, 714-721.(8) Xu, F.; Jabasini, M.; Baba, Y. Electrophoresis 2002, 23, 3608-3614.(9) Zuborova, M.; Demianova, Z.; Kaniansky, D.; Masar, M.; Stanislawski, B. J.

Chromatogr., A 2003, 990, 179-188.(10) Song, L. G.; Fang, D. F.; Kobos, R. K.; Pace, S. J.; Chu, B. Electrophoresis

1999, 20, 2847-2855.(11) Ocvirk, G.; Munroe, M.; Tang, T.; Oleschuk, R.; Westra, K.; Harrison, D. J.

Electrophoresis 2000, 21, 107-115.(12) Dou, Y. H.; Bao, N.; Xu, J. J.; Chen, H. Y. Electrophoresis 2002, 23, 3558-

3566.(13) Liu, Y.; Fanguy, J. C.; Bledsoe, J. M.; Henry, C. S. Anal. Chem. 2000, 72,

5939-5944.(14) Barker, S. L. R.; Tarlov, M. J.; Canavan, H.; Hickman, J. J.; Locascio, L. E.

Anal. Chem. 2000, 72, 4899-4903.(15) Ichijima, H.; Okada, T.; Uyama, Y.; Ikada, Y. Makromol. Chem. 1991, 192,

1213-21.(16) Lawrence, J.; Li, L. Mater. Sci. Eng. A 2001, 303, 142-149.(17) Feurer, B.; Harel, J. P.; Lacabanne, C. Thin Solid Films 1983, 100, 249-

255.

Anal. Chem. 2004, 76, 6948-6955

6948 Analytical Chemistry, Vol. 76, No. 23, December 1, 2004 10.1021/ac040094l CCC: $27.50 © 2004 American Chemical SocietyPublished on Web 11/04/2004

and aminolysis,26 some have been adapted for fabrication ofPMMA microdevices. Johnson et al. used a pulsed UV excimerlaser below the ablation threshold to produce carboxyl groupson the surface of PMMA microchannels.27 Zangmeister and Tarlovmodified the surface of PMMA microdevices with UV/ozone.28

Henry et al. applied lithiated diamine to PMMA substrates to yieldamine-terminated PMMA surfaces. The same group furtherderivatized the aminolyzed PMMA microdevices with n-octa-decane-1-isocyanate.29 Although considerable effort has beendevoted to chemical modification of the PMMA surface, reliablederivatization techniques appropriate for PMMA bioanalyticalmicrodevices, especially microfabricated separation devices forproteins and peptides, still need to be explored and developed.

Atom-transfer radical polymerization (ATRP) is a novel tech-nique that can prepare polymers with controlled length in solutionor on a surface,30-33 making ATRP very attractive for surfacemodification of microcapillary electrophoresis (µCE) systems.Recently, Xiao and co-workers utilized ATRP to graft polyacryl-amide on the surface of PDMS,34 and they applied this techniquein the fabrication of PDMS µCE devices for protein separation.35

In electrophoretic separation of proteins and peptides, uniformpolymer coatings that have little interactions with the analytesusually improve the separation. In addition to polyacrylamide,neutral polymers such as hydroxypropyl cellulose, poly(2-hydroxy-propyl methacrylate), poly(2-hydroxyethyl methacrylate), poly-(ethylene glycol) (PEG), and poly(ethylene-propylene glycol)have been utilized to coat fused-silica capillary columns for highlyefficient protein separations. The best results were obtained usingPEG or PEG-like polymers.36-40 In the present work, we describea technique using ATRP to graft PEG on the surface of PMMAµCE devices for electrophoretic separations of proteins andpeptides.

EXPERIMENTAL SECTIONMaterials. 2-Bromoisobutyryl bromide (98%), poly(ethylene

glycol) methyl ether methacrylate (PEGMEMA, MW ∼475), 2,2′-dipyridyl (99+%), copper(I) chloride (98+%), and copper(II)bromide (99%) were purchased from Aldrich (Milwaukee, WI) andused without further purification. Heptane (reagent grade), tetra-hydrofuran (THF, reagent grade), absolute methanol (reagentgrade), pyridine (reagent grade), urea (reagent grade), anddithiothreitol (molecular biology grade) were obtained from FisherScientific (Fair Lawn, NJ). Iodoacetamide was purchased fromAmersham Biosciences (Piscataway, NJ). Ethylenediaminetetra-acetic acid disodium salt dihydrate (Na2EDTA‚2H2O, 99+%) wasobtained from Invitrogen Life Technologies (Carlsbad, CA).Fluorescein isothiocyanate (FITC), bovine serum albumin (BSA),trypsin, FITC-conjugated BSA (FITC-BSA), and FITC-conju-gated insulin (FITC-insulin) were purchased from Sigma (St.Louis, MO). R-Phycoerythrin (R-PE) was obtained from Poly-sciences (Warrington, PA). Recombinant, enhanced green fluo-rescent protein (GFP) was purchased from Clontech (Palo Alto,CA). The 18.2 MΩ‚cm deionized water used was from a Milli-QUF Plus water purification system (Millipore, Billerica, MA), andthe buffer solution used throughout the experiments was 10 mMTrisma hydrochloride (Tris-HCl) at pH 8.8, which was filteredusing 0.2-µm syringe filters (Pall Gelman Laboratory, Ann Arbor,MI).

Microchip Fabrication. Acrylite FF sheets (Cyro, WestPaterson, NJ) were used as substrates for PMMA µCE chips. Thefabrication protocol was adapted from Kelly and Woolley.41 First,an 800-nm etch mask layer of silicon dioxide was grown on a 4-in.silicon wafer (Encompass Distribution Services, Pleasanton, CA)at 1110 °C in an atmosphere of oxygen and water. Next, Shipley812 (Shipley, Marlborough, MA) positive photoresist was spincoated on the silicon wafer at 3500 rpm for 120 s. To increaseadhesion of the photoresist and drive off residual solvent, the waferwas baked at 90 °C for 2 min. The photoresist was then exposedto UV radiation for 40 s through a photomask using a PLA-501Fmask aligner (Canon, Tokyo, Japan). The photomask was de-signed using CAD software (CleWin, WieWeb Software) andprinted onto transparency film using a 3600 dpi printer. Followingexposure, the photoresist was developed with 20% aqueousMicroposit 351 developer (Shipley) for 30 s. After being placedin an oven for 30 min at 150 °C, which helped to harden thephotoresist, the wafer was immersed in buffered hydrofluoric acidfor 10 min to remove silicon dioxide from the areas not coveredby the photoresist. Finally, the silicon wafer was etched using 40%aqueous KOH solution at 80 °C for 30 min. The resulting silicontemplate with protruding features was used to imprint the channelpattern (Figure 1) into 1.5-mm-thick PMMA substrates at 120 °C.The fabrication of µCE chips was finished after surface derivati-zation (see below) by thermally bonding the patterned surface toa 3.0-mm-thick PMMA substrate with access holes at 95 °C. Theimprinting and bonding processes were performed in an HP 5890gas chromatography oven. The access holes were created usinga C-200 CO2 laser engraving system (Universal Laser Systems,Scottsdale, AZ). The dimensions of the channel features weremeasured using an Alpha-step 200 profilometer (KLA-Tencor, SanJose, CA).

(18) Hozumi, A.; Masuda, T.; Hayashi, K.; Sugimura, H.; Takai, O.; Kameyama,T. Langmuir 2002, 18, 9022-9027.

(19) Shenton, M. J.; Stevens, G. C. J. Phys. D: Appl. Phys. 2001, 34, 2761-2768.

(20) Lim, H.; Lee, Y.; Han, S.; Cho, J.; Kim, K.-J. J. Vac. Sci. Technol., A 2001,19, 1490-1496.

(21) Weikart, C. M.; Yasuda, H. K. J. Polym. Sci. A: Polym. Chem. 2000, 38,3028-3042.

(22) Groning, P.; Collaud, M.; Dietler, G.; Schlapbach, L. J. Appl. Phys. 1994,76, 887-892.

(23) Kang, I. K.; Kwon, B. K.; Lee, J. H.; Lee, H. B. Biomaterials 1993, 14, 787-792.

(24) Schulz, U.; Munzert, P.; Kaiser, N. Surf. Coat. Technol. 2001, 142-144,507-511.

(25) Vargo, T. G.; Gardella, J. A. J. Polym. Sci. A: Polym. Chem. 1989, 27, 1267-1286.

(26) Xu, G.-F.; Bergbreiter, D. E.; Letton, A. Chem. Mater. 1992, 4, 1240-1245.(27) Johnson, T. J.; Ross, D.; Gaitan, M.; Locascio, L. E. Anal. Chem. 2001, 73,

3656-3661.(28) Zangmeister, R. A.; Tarlov, M. J. Langmuir 2003, 19, 6901-6904.(29) Henry, A. C.; Tutt, T. J.; Galloway, M.; Davidson, Y. Y.; McWhorter, C. S.;

Soper, S. A.; McCarley, R. L. Anal. Chem. 2000, 72, 5331-5337.(30) Kato, M.; Kamigaito, M.; Sawamoto, M.; Higashimura, T. Macromolecules

1995, 28, 1721-1723.(31) Wang, J.-S.; Matyjaszewski, K. J. Am. Chem. Soc. 1995, 117, 5614-5615.(32) Matyjaszewski, K.; Xia, J. Chem. Rev. 2001, 101, 2921-2990.(33) Kamigaito, M.; Ando, T.; Sawamoto, M. Chem. Rev. 2001, 101, 3689-3745.(34) Xiao, D.; Zhang, H.; Wirth, M. J. Langmuir 2002, 18, 9971-9976.(35) Xiao, D.; Le, T. V.; Wirth, M. J. Anal. Chem. 2004, 76, 2055-2061.(36) Huang, M.; Vorkink, W. P.; Lee, M. L. J. Microcolumn Sep. 1992, 4, 233-

238.(37) Zhao, Z.; Malik, A.; Lee, M. L. J. Microcolumn Sep. 1992, 4, 411-417.(38) Huang, M.; Lee, M. L. J. Microcolumn Sep. 1992, 4, 491-496.(39) Malik, A.; Zhao, Z.; Lee, M. L. J. Microcolumn Sep. 1993, 5, 119-125.(40) Zhao, Z.; Malik, A.; Lee, M. L. Anal. Chem. 1993, 65, 2747-2752. (41) Kelly, R. T.; Woolley, A. T. Anal. Chem. 2003, 75, 1941-1945.

Analytical Chemistry, Vol. 76, No. 23, December 1, 2004 6949

Oxygen Plasma Activation. An oxygen plasma generatedusing a DEM-451 reactive ion etcher (Anelva, Tokyo, Japan) wasused to oxidize the surface of PMMA before grafting. The plasmawas excited at 13.56 MHz, and the cathode/sample holder wascooled with running water. Prior to plasma oxidation, the patternedPMMA substrates and cover plates were annealed at 100 °C for1 h. Next, the substrates were rinsed with methanol and waterand dried with nitrogen gas. The PMMA substrates were loadedon the cathode in the etching chamber, and the chamber pressurewas reduced below 10 mTorr using a two-stage vacuum pump.Oxygen was introduced into the chamber at a flow rate of 3.10standard cubic centimeters per minute using a mass flow control-ler. The plasma was started when the chamber pressure reached100 mTorr, and the power of the plasma was maintained at 80 Wusing a manual impedance matching network during the oxidationprocess, which took ∼3 min.

During the plasma oxidation, ionized oxygen or oxygenradicals attacked the PMMA backbone or side chains. As a result,oxygen-containing functionalities such as hydroxyl and carboxylgroups were generated on the PMMA surface, which providedhandles for subsequent chemical reactions.

Immobilization of Initiator. Immobilization of a typical ATRPinitiator (Scheme 1), 2-bromoisobutyryl bromide, was performedaccording to the procedure of Carlmark and Malmstrom.42

Immediately after oxygen plasma activation, the PMMA substratewas immersed in a heptane/THF solution containing 50 mM2-bromoisobutyryl bromide and 55 mM pyridine. After 24 h, thesubstrate was removed and washed thoroughly with methanol anddeionized water. Nitrogen gas was then used to dry the substrate.

Grafting of Poly(ethylene glycol) on the PMMA Surface.Grafting of PEG on the PMMA surface (Scheme 2) was imple-mented using a protocol adapted from the literature.43-45 First,

40 mL of PEGMEMA and 60 mL of deionized water were addedto a 250-mL round-bottom flask. The flask was then sealed with asleeved rubber stopper and subjected to laboratory vacuumfollowed by 20 psi nitrogen gas for 30 min. This purging procedurewas repeated three times to remove oxygen. Afterward, CuCl(0.424 g, 1.8 mmol), CuBr2 (0.287 g, 0.54 mmol), 2,2′-dipyridyl(1.74 g, 4.68 mmol), and 40 wt % PEGMEMA aqueous solutionwere mixed inside a glovebox. To start the ATRP reaction, theresulting dark brown mixture was transferred to a crystallizationdish containing the initiator-immobilized PMMA substrate undera nitrogen atmosphere. The reaction was allowed to proceed atroom temperature for 24 h. To quench the reaction, the PMMAsubstrate was taken out of the glovebox, immersed in saturatedaqueous Na2EDTA solution to remove residual copper(II) ions,and rinsed with a copious volume of deionized water. Before chipassembly, nitrogen gas was used to dry the PMMA substrates.

Contact Angle Measurements. Immediately after 4 µL ofdeionized water was placed on the PMMA surface using a syringe,an NRL-100 goniometer (Rame-hart, Mountain Lakes, NJ) wasused to measure the contact angle. The contact angle wascalculated as the mean of the left and right contact angles of thewater drop.

Electroosmotic Flow Measurements. The electroosmoticflow (EOF) in the PMMA microchannel was measured using thecurrent monitoring method.46 In a typical measurement, 1 mL ofdeionized water was pumped through the microchannel at a flowrate of 50 µL/min using a syringe pump (11-Plus, HarvardApparatus, Holliston, MA). Following this rinsing step, the channelwas conditioned with 1 mL of 50 mM Tris-HCl buffer at a pH of8.8 using the syringe pump (flow rate, 50 µL/min). The reservoirsused to provide electrical contact to the channel were emptied,10 µL of 50 mM Tris-HCl buffer was added to one reservoir, andthe same volume of 1 mM Tris-HCl buffer was introduced intothe other reservoir. A PS-350 high-voltage supply unit (StanfordResearch Systems, Sunnyvale, CA) was employed to provide highvoltage during the measurement, and current variation wasrecorded using a PCI-1200 data acquisition board (NationalInstruments, Austin, TX) and an in-house-written LabView 6isoftware program (National Instruments).

Tryptic Digest Preparation. A digestion protocol describedin the literature was followed.47 Briefly, before tryptic digestion,6 M urea-denatured BSA (1 mg) was reduced and alkylated usingdithiothreitol and iodoacetamide, respectively. The pretreated BSAwas then digested with 20 µg of trypsin at 37 °C at pH 8.0. Thetryptic digest was desalted using a Spectra/Por cellulose esterdialysis membrane (MWCO 100, Spectrum Medical Industries,Houston, TX) for 24 h, and 6 mM FITC in acetone was added tothe product at a 10:1 FITC/BSA molar ratio. The derivatizationreaction was allowed to proceed in the dark at room temperaturefor 24 h. Before CE separation, the digest was diluted 2.5-foldwith 10 mM Tris-HCl buffer at a pH of 8.8.

Detection. The laser-induced fluorescence detection systemand the setup for data acquisition have been reported elsewhere.41

The sampling rate for data collection was 20 Hz.

(42) Carlmark, A.; Malmstrom, E. E. Biomacromolecules 2003, 4, 1740-1745.(43) Huang, W.; Kim, J.-B.; Bruening, M. L.; Baker, G. L. Macromolecules 2002,

35, 1175-1179.(44) Huang, W.; Baker, G. L.; Bruening, M. L. Angew. Chem., Int. Ed. 2001,

40, 1510-1512.

(45) Wang, X.-S.; Armes, S. P. Macromolecules 2000, 33, 6640-6647.(46) Huang, X.; Gordon, M. J.; Zare, R. N. Anal. Chem. 1988, 60, 1837-1838.(47) Kinter, M.; Sherman, N. E. Protein Sequencing and Identification Using

Tandem Mass Spectrometry; Wiley-Interscience: New York, 2000.

Figure 1. (A) Schematic diagram of the µCE chips used in this work(1, sample reservoir; 2, sample waste reservoir; 3, buffer reservoir;4, buffer waste reservoir). (B) Cross-sectional dimensions of themicrochannels.

6950 Analytical Chemistry, Vol. 76, No. 23, December 1, 2004

Chip Operation. Channels were filled with 10 mM Tris-HClbuffer from reservoir 4 (Figure 1) using a syringe pump. Prior toseparation, 30 µL of protein sample was introduced into reservoir1 and a platinum electrode was inserted into each reservoir toprovide electrical contact. Voltages were applied to the reservoirsusing PS-300 and PS-350 high-voltage supply units (StanfordResearch Systems). The two voltage supplies were connectedusing a home-built switching circuit board. “Pinched” injection 48

was used to introduce the sample into the channel with anestimated injection volume of 230 pL. During injection, reservoirs1, 3, and 4 were grounded, and reservoir 2 was maintained at +0.6kV. During separation, reservoirs 1 and 2 were set at +0.6 kV,reservoir 3 was grounded, and reservoir 4 was set at +2.0 kV.These conditions were used for all of the µCE experiments.

Conventional Capillary Electrophoresis. A Crystal CEmodel 300 capillary electrophoresis system (UNICAM, Madison,WI) was used to analyze unlabeled insulin. An 80-cm-long fused-silica capillary column (75 µm i.d. × 360 µm o.d., PolymicroTechnologies, Phoenix, AZ) was installed in this instrument, andthe effective length (the distance between the injection end ofthe capillary and the detection window) was 65 cm.

An insulin sample was prepared in 10 mM Tris-HCl at pH 8.8;the same buffer was also used as separation buffer. Hydrodynamicinjection was performed at 100 mbar for 6 s to introduce 1 mg/mL insulin into the capillary column. A voltage of 20 kV was thenapplied to the column to start the electrophoresis. The proteinwas detected using UV absorbance at 214 nm.

RESULTS AND DISCUSSIONSubstrate Fabrication. Unlike some commodity polymers,

the resistance of PMMA to many organic chemicals is low.49 Acids,amines, ketones, esters, cycloethers, aromatic hydrocarbons,nitriles, and halogenated hydrocarbons can dissolve PMMA.N,N-Dimethylacetamide, N,N-dimethyl formamide, and dimethylsulfoxide also damage PMMA. Xu and colleagues employed aTHF-heptane mixture (4:7 v/v) as a solvent to modify the PMMAsurface.26 In this work, we used THF-heptane to immobilize

2-bromoisobutyryl bromide on the PMMA surface. The ratio ofTHF to heptane was decreased to 1:4 (v/v) to minimize any effectson PMMA. Indeed, we found that this solvent mixture at roomtemperature did not damage the PMMA microchannels or causeloss of polymer clarity. Moreover, the background fluorescenceof the PMMA substrates did not increase after surface treatmentin this solvent mixture.

Cutting and drilling are often used in the fabrication ofpolymeric microdevices; however, the structures can be stressedduring conventional machining. We observed that after drilledPMMA substrates were treated with organic solvents, cracksappeared and extended rapidly into the stressed structure,presumably because the solvents entered and swelled the polymermatrix. To alleviate this problem, the PMMA components wereannealed in an oven before surface modification. Annealing wascarried out at the glass transition temperature of PMMA (100 °C)for 1 h, which relieved the stress. Figure 2 shows the effect ofannealing. Holes with diameters of 0.2 cm were drilled through 1cm × 1 cm, 1.5-mm-thick PMMA substrates using the CO2 lasersystem. Two drilled PMMA substrates, one annealed and one thatwas not annealed, were immersed in absolute methanol for 20min. Whereas cracks were found along the edge of the hole inthe untreated PMMA substrate (Figure 2A), cracks did not appearin the annealed PMMA substrate (Figure 2B).

Contact Angle and Electroosmotic Flow Measurements.Water contact angles for 1.5- and 3.0-mm-thick PMMA substrates(Table 1) were obtained, since both were utilized in the fabricationof microdevices. The contact angle for PEG-grafted PMMAsubstrates agrees well with results obtained from PEG monolayersgrafted on a silicon surface (∼41°),50 while the data for untreatedPMMA substrates are close to those measured by Henry and co-workers (∼66°).29 The comparison between contact angles beforeand after surface modification indicates that the wettability andsurface chemistry of PMMA were significantly changed after PEGgrafting.

The EOF mobility for an untreated PMMA microchannel was(1.6 ( 0.2) × 10-4 cm2‚V-1‚s-1 (CL% ) 95%; average of 4measurements), and the direction of the EOF was from the anodeto the cathode. In contrast, the variation in current during EOF(48) Jacobson, S. C.; Hergenroder, R.; Koutny, L. B.; Warmack, R. J.; Ramsey, J.

M. Anal. Chem. 1994, 66, 1107-1113.(49) Brandup, J.; Immergut, E. H.; Grulke, E. A.; Abe, A.; Bloch, D. R. Polymer

Handbook, 4th ed.; John Wiley: New York, 1999. (50) Sharma, S.; Johnson, R. W.; Desai, T. A. Langmuir 2004, 20, 348-356.

Scheme 1

Scheme 2

Analytical Chemistry, Vol. 76, No. 23, December 1, 2004 6951

testing for the PEG-grafted PMMA microchannel was very small,and the current-time profile resembled a plateau over at least400 s. Thus, under the conditions used for separation of proteinsand peptides in this study, EOF was less than 1 × 10-5 cm2‚V-1‚s-1

in the surface-modified PMMA microchannels.Electrophoresis of Proteins and Peptides. Bovine serum

albumin is well known for its nonspecific adsorption on surfaces,which can negatively affect its analysis in microdevices. Thus,FITC-BSA was selected as a test compound to evaluate the effectof the ATRP-grafted PEG layer on the performance of PMMA µCEdevices.

To examine the adsorption of FITC-BSA on both untreatedand PEG-grafted PMMA microchannels, FITC-BSA was intro-duced to the µCE chips. After 30 min, the microchannels wereflushed with deionized water for 2 h at a flow rate of 30 µL/min.The microchips were then placed on the microscope stage, theinjection cross region was illuminated with a 488-nm laser line,and fluorescence was observed. As shown in Figure 3A, FITC-BSA adsorbed strongly on the surface of the untreated PMMAmicrochannels and bright fluorescence could be seen in all

channels. In contrast, no fluorescence was observed in the PEG-grafted microchannel (Figure 3B), indicating significantly lessprotein adsorption.

Figure 4 compares the µCE analysis of FITC-BSA using PEG-grafted and unmodified PMMA microdevices. As shown in Figure4A, three components were resolved in the electropherogramsobtained from the surface-grafted device. The two small peaks(1a and 1b) are suspected to be fragments of BSA or other bovineserum proteins. We also spiked the FITC-BSA sample with FITC,and a peak corresponding to the free label (peak 2) appeared

Figure 2. Photographs showing the effect of solvent on laser-drilled holes in PMMA. The diameter of the CO2 laser-machined holes was 0.2mm, and the PMMA substrates were immersed in absolute methanol for 20 min. (A) Untreated substrate; (B) annealed surface.

Figure 3. Fluorescence micrographs of PMMA µCE channels after FITC-BSA exposure. (A) Untreated PMMA microchip; (B) PEG-graftedPMMA microchip.

Table 1. Contact Angles for PEG-Grafted andUntreated PMMA Substrates

1.5-mm-thicksubstratea,b

3.0-mm-thicksubstratea,b

PEG-grafted PMMA 38.6 ( 1.2 41.2 ( 0.8untreated PMMA 65.7 ( 2.8 69.3 ( 2.0

a Each value was averaged from contact angle measurements offour separate water drops on the same PMMA substrate. b 95% CL.

6952 Analytical Chemistry, Vol. 76, No. 23, December 1, 2004

between peaks 1b and 1c in the electropherogram (Figure 5),confirming that peaks 1a, 1b, and 1c were not the free fluorescentlabel. The same protein sample was also analyzed using anuntreated PMMA µCE chip (Figure 4B). The migration time ofpeak 1c (Table 2) was 2.6-fold greater on the untreated PMMAµCE chip than on the PEG-grafted chip. The slower migrationwas due to a combination of interactions between FITC-BSA andthe PMMA surface, and the electroosmotic flow that opposedmigration of the protein in the untreated device. In addition, strongprotein-surface interactions in the untreated PMMA micro-channel resulted in severe band broadening, and therefore, thecolumn efficiency or total plate number for the untreated chipwas 1 order of magnitude lower compared to the PEG-grafted chip(Table 2). Moreover, adsorption of the negatively charged BSA

should change the local ú potential on the unmodified PMMAsurface, which may give rise to unstable EOF and, thus, poorerreproducibility of FITC-BSA migration times. Unstable EOF mayhave also caused baseline drift and poor injection reproducibilityin unmodified PMMA systems. However, these phenomena werenot found for the PEG-grafted microchips.

FITC-insulin was also used to test the PEG-grafted micro-chips. Seven peaks, three major ones and four minor ones,appeared in the electropherogram (Figure 6). In comparison, onlyone peak was observed in conventional capillary electrophoresisof unlabeled insulin (electropherogram not shown). We believethat the multiple peaks may have resulted from insulin derivatizedwith different numbers of FITC tags or from fluorescent labelingof peptide chains resulting from insulin decomposition. Polyacryl-amide gel electrophoresis of FITC-insulin confirms that thereare three major peptide components in the sample, which

Figure 4. µCE of FITC-BSA. (A) PEG-grafted chip; (B) untreated chip. Electropherograms were recorded for three consecutive runs. Peaks1a, 1b, and 1c are the three main components of the BSA sample.

Figure 5. µCE of a mixture of FITC and FITC-BSA.

Table 2. Column Efficiencies and Migration TimeReproducibilities for PEG-Grafted and UntreatedPMMA µCE Chips

migrationtime (s)a,b

RSD(%)

totalplatesa,b

PEG-grafted µCE chip 30.89 0.27 2.7 × 104

untreated µCE chip 80.91 3.1 2.5 × 103

a Data were measured/calculated for peak 1c in Figure 4. b Datawere calculated from five consecutive runs.

Analytical Chemistry, Vol. 76, No. 23, December 1, 2004 6953

correspond to peaks 1-3 in Figure 6. However, µCE of FITCunder the same conditions shows that the migration time of FITCis very close to peak 2. We spiked the FITC-insulin sample withFITC and found that the height of peak 2 increased. Whether ornot FITC is contained in peak 2 cannot be verified without usingother analytical techniques.

Peaks 1 and 3 were used to evaluate chip performance (Table3). As many as 5.2 × 104 plates for a 3.5-cm-long separationchannel were obtained in these analyses. The high plate numbersare a direct consequence of reducing both EOF and nonspecificadsorption on the PMMA surface with the grafted PEG layer.

A mixture of acidic proteins containing R-PE, FITC-BSA, andGFP was separated with a PEG-grafted PMMA µCE chip, and atypical result is shown in Figure 7. Three peaks (peak 1a anddoublet 1b) belonging to FITC-BSA appear in the electrophero-gram. R-PE (peak 2) appears between peak 1a and doublet 1b.GFP splits into two peaks (peaks 3a and 3b) during the separation.The total plates for (1a) of FITC-BSA, (3b) of GFP, and R-PEwere 3.9 × 104, 3.9 × 104, and 7.5 × 103, respectively.

A 40-s separation of a tryptic digest of BSA was also conductedusing a PEG-grafted PMMA µCE chip (Figure 8). The separationwas reproducible with relative standard deviations (RSD) for peaks1, 2, and 3 of 0.47, 0.32, and 0.60%, respectively (data were obtainedfor 4 consecutive runs). These results demonstrate the applicabilityof surface-grafted µCE systems in peptide digest analysis, animportant area of proteomics research.

CONCLUSIONSPoly(ethylene glycol) was grafted to the surface of PMMA

substrates using atom-transfer radical polymerization, whichsubstantially reduced electroosmotic flow and nonspecific adsorp-tion of proteins on the PMMA surface. As a result, fast, highlyefficient, and reproducible electrophoretic separations of proteinsand peptides were achieved using PEG-grafted PMMA µCE chips.We believe that this surface modification technique can not only

Figure 6. µCE of FITC-insulin. Electropherograms were obtained from four consecutive runs. Peaks 1 and 3 were used in the evaluation ofchip performance.

Table 3. Column Efficiencies and Migration TimeReproducibilities for Insulin Fragments

peakmigrationtime (s)a

RSD(%)

totalplatesa

1 19.36 0.57 5.2 × 104

3 32.28 1.0 4.2 × 104

a Data were calculated from four consecutive runs.

Figure 7. µCE of model acidic proteins. The sample contained R-PE(peak 2), FITC-BSA (peak 1a and doublet 1b), and GFP (peaks 3aand 3b).

Figure 8. µCE of a BSA tryptic digest.

6954 Analytical Chemistry, Vol. 76, No. 23, December 1, 2004

be utilized in the fabrication of µCE chips as shown here but canalso be applied in a much broader sense to other disposablepolymeric microfluidic devices for proteomics studies.

ACKNOWLEDGMENTThe work was funded by the National Institutes of Health (R01

GM064547-01A1). J.L. was supported by a R.K. Robins Fellowshipfrom the Department of Chemistry and Biochemistry at BrighamYoung University (BYU). All microfabrication work and PMMAsurface activation were performed in the Integrated Microelec-

tronics Laboratory at BYU. The authors acknowledge Dr. GeraldWatt for the use of a nitrogen glovebox and Dr. Matthew Linfordfor helpful instructions and the use of a goniometer. We aregrateful to Jenny Armenta for the analysis of the unlabeled insulinsample. We also thank Ziniu Zhou, Zhifeng Ye, and Jing Liu forhelpful discussions on chemical treatment of PMMA.

Received for review May 14, 2004. Accepted September10, 2004.

AC040094L

Analytical Chemistry, Vol. 76, No. 23, December 1, 2004 6955