Fabrication of Porous Polymer Monoliths in Polymeric Microfluidic

Jan 14, 2006 - the clogging of the channels, which is a common problem ... (10) Lazar, I. M.; Ramsey, R. S.; Sundberg, S.; Ramsey, J. M. Anal. Chem. 1...
0 downloads 0 Views 502KB Size
Anal. Chem. 2006, 78, 1130-1138

Fabrication of Porous Polymer Monoliths in Polymeric Microfluidic Chips as an Electrospray Emitter for Direct Coupling to Mass Spectrometry Mohamed F. Bedair and Richard D. Oleschuk*

Department of Chemistry, Queen’s University, Kingston, Ontario, Canada, K7L 3N6

Coupling of polymeric microfluidic devices to mass spectrometry is reported using porous polymer monoliths (PPM) as nanoelectrospray emitters. Lauryl acrylate-coethylene dimethacrylate porous polymer monolith was photopatterned for 5 mm at the end of the channel of microfluidic devices fabricated from three different polymeric substrate materials, including the following: poly(dimethylsiloxane) (PDMS), poly(methyl methacrylate) (PMMA), and cyclic olefin copolymer (COC). Spraying directly from the end of the chip removes any dead volume associated with inserted emitters or transfer lines, and the presence of multiple pathways in the PPM prevents the clogging of the channels, which is a common problem in conventional nanospray emitters. Spraying from a microfluidic channel having a PPM emitter produced a substantial increase in TIC stability and increased sensitivity by as much as 70× compared to spraying from an open end chip with no PPM. The performance of PPM emitter in COC, PMMA, and PDMS chips was compared in terms of stability and reproducibility of the electrospray. COC chips showed the highest reproducibility in terms of chip-to-chip performance, which can be attributed to the ease and reproducibility of the PPM formation due to the favorable optical and chemical properties of COC. We have further tested the performance of the COC chips by constant infusion of poly(propylene glycol) solution at organic content ranging from 10 to 90% methanol and at flow rates ranging from 50 to 1000 nL/min, showing optimum spraying conditions (RSD < 5%) at 50-70% organic content and at flow rates from 100 to 500 nL/ min. The PPM sprayer was also used for protein preconcentration and desalting prior to mass spectrometric detection, and results were comparable with a chip spraying from an electrospray tip. The progress in proteomic research has increased the demand for high-throughput analysis of biological samples present at very low concentration or in very limited sample volumes. Microfluidic analytical devices coupled with mass spectrometry detection are possible solutions to the current analysis bottleneck. As a result, interfacing microfluidic chips to mass spectrometry detection has received considerable attention.1-4 Initial work on coupling microfluidic chips to electrospray mass spectrometry was demon* To whom correspondence should be addressed. E-mail: Oleschuk@ chem.queensu.ca.

1130 Analytical Chemistry, Vol. 78, No. 4, February 15, 2006

strated by electrospraying directly from the flat end of a glass chip5,6 but suffered from large Taylor cone formation. To increase the efficiency and stability of the electrospray, a fused-silica nanospray emitter7-10 or a transfer line to the electrospray tip11-15 was inserted at the end of the separation channel of the microfluidic device. Although this approach has proven successful, the drilling and critical alignment of the nanospray emitter to the end of the channel and the dead volume created present a significant challenge. As the interest in polymeric microfluidic devices has grown in recent years, many research groups have reported the microfabrication of the electrospray emitter directly from bulk material of the microchip. The fabrication of a three-dimensional tip in thermoplastic chips (poly(methyl methacrylate) (PMMA) and PC) produced a stable electrospray in the sheathless mode,16 while the planar sharp tip required the assistance of gas and liquid sheath flow.17 Recently, a nozzle that was micromilled in a PMMA foil used as cover for the microchip was evaluated as a possible nanospray emitter.18 The nozzle had a diameter of 30 µm with an apex of 60° that produced a stable electrospray with solutions having organic content as low as 5% methanol. (1) Oleschuk, R. D.; Harrison, D. J. TrAC, Trends Anal. Chem. 2000, 19, 379388. (2) De Mello, A. J. Lab Chip 2001, 1, 7N-12N. (3) Limbach Patrick, A.; Meng, Z. Analyst 2002, 127, 693-700. (4) Sung, W.-C.; Makamba, H.; Chen, S.-H. Electrophoresis 2005, 26, 17831791. (5) Xue, Q.; Foret, F.; Dunayevskiy, Y. M.; Zavracky, P. M.; McGruer, N. E.; Karger, B. L. Anal. Chem. 1997, 69, 426-430. (6) Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 1997, 69, 1174-1178. (7) Xiang, F.; Lin, Y.; Wen, J.; Matson, D. W.; Smith, R. D. Anal. Chem. 1999, 71, 1485-1490. (8) Li, J.; Thibault, P.; Bings, N. H.; Skinner, C. D.; Wang, C.; Colyer, C.; Harrison, J. Anal. Chem. 1999, 71, 3036-3045. (9) Bings, N. H.; Wang, C.; Skinner, C. D.; Colyer, C. L.; Thibault, P.; Harrison, D. J. Anal. Chem. 1999, 71, 3292-3296. (10) Lazar, I. M.; Ramsey, R. S.; Sundberg, S.; Ramsey, J. M. Anal. Chem. 1999, 71, 3627-3631. (11) Figeys, D.; Ning, Y.; Aebersold, R. Anal. Chem. 1997, 69, 3153-3160. (12) Zhang, B.; Liu, H.; Karger, B. L.; Foret, F. Anal. Chem. 1999, 71, 32583264. (13) Zhang, B.; Foret, F.; Karger, B. L. Anal. Chem. 2000, 72, 1015-1022. (14) Deng, Y.; Henion, J.; Li, J.; Thibault, P.; Wang, C.; Harrison, D. J. Anal. Chem. 2001, 73, 639-646. (15) Meng, Z.; Qi, S.; Soper, S. A.; Limbach, P. A. Anal. Chem. 2001, 73, 12861291. (16) Svedberg, M.; Pettersson, A.; Nilsson, S.; Bergquist, J.; Nyholm, L.; Nikolajeff, F.; Markides, K. Anal. Chem. 2003, 75, 3934-3940. (17) Wen, J.; Lin, Y.; Xiang, F.; Matson, D. W.; Udseth, H. R.; Smith, R. D. Electrophoresis 2000, 21, 191-197. (18) Schilling, M.; Nigge, W.; Rudzinski, A.; Neyer, A.; Hergenroeder, R. Lab Chip 2004, 4, 220-224. 10.1021/ac0514570 CCC: $33.50

© 2006 American Chemical Society Published on Web 01/14/2006

Spraying from the end tip of a poly(dimethylsiloxane) (PDMS) chip with a channel width and depth of 10 × 10,19 30 × 50,20 and 50 × 50 µm open groove21 has also been reported. The electrospray voltage was applied at the eluent reservoir. As expected, smaller channel dimensions showed higher spray performance but exhibited problems with device clogging.19 A PDMS chip with a three-dimensional emitter tip was micromolded around a 50µm metallic wire. The shape of the emitter tip was drilled into the mold and the chip tip was coated by dusting with graphite, which allowed the electrospray voltage to be applied directly to the tip,22 thus improving the stability of the electrospray. Other polymeric materials have also been used to fabricate the chip emitter. A planer niblike emitter containing a sample reservoir and a 8-µm capillary slot fabricated from negative photoresist SU-8 required only 0.8 kV to produce stable electrospray23 with solution flowing by capillary action. Capillary action was also used to drive the solution in a polyimide microchip with a tip of 20-µm diameter fabricated by plasma etching.24 Another application using polyimide chips was reported by Yin et al., from Agilent Technology, where the channels in the chip and a conically shaped tip are formed by laser ablation fabrication methods. The chip has both an integrated enrichment column and a separation column.25 Coupling the chip to MS with microfabricated nozzles was also reported using silicon etching26 or parylene tip micromachined on a silicon chip.27 A parylene nanospray emitter was embossed in a cyclic olefin (Zeonor) chip; the triangular-shaped thin parylene tip was formed by lithography and etching.28 Recently our laboratory reported a nanospray emitter utilizing a porous polymer monolith (PPM) fabricated at the end of a capillary.29,30 A stable electrospray was produced at different flow rates where a single Taylor cone was generated at flow rates higher than 100 nL/min, while multiple Taylor cones generating a mist were developed at flow rates lower than 100 nL/min. The multiple paths of the PPM minimize clogging problems usually associated with nanobore emitters, and the hydrophobic nature of the PPM limits the problems of the droplet spreading at the emitter. The hydrophobic effect has also been demonstrated by spraying from a polycarbonate chip having a porous poly(tetrafluoroethylene) membrane thermally bonded to the channel exit.31 (19) Huikko, K.; Oestman, P.; Grigoras, K.; Tuomikoski, S.; Tiainen, V. M.; Soininen, A.; Puolanne, K.; Manz, A.; Franssila, S.; Kostiainen, R.; Kotiaho, T. Lab Chip 2003, 3, 67-72. (20) Kim, J.-S.; Knapp, D. R. Electrophoresis 2001, 22, 3993-3999. (21) Svedberg, M.; Veszelei, M.; Axelsson, J.; Vangbo, M.; Nikolajeff, F. Lab Chip 2004, 4, 322-327. (22) Dahlin, A. P.; Wetterhall, M.; Liljegren, G.; Bergstroem, S. K.; Andren, P.; Nyholm, L.; Markides, K. E.; Bergquist, J. Analyst 2005, 130, 193-199. (23) Le Gac, S.; Arscott, S.; Rolando, C. Electrophoresis 2003, 24, 3640-3647. (24) Gobry, V.; Van Oostrum, J.; Martinelli, M.; Rohner, T. C.; Reymond, F.; Rossier, J. S.; Girault, H. H. Proteomics 2002, 2, 405-412. (25) Yin, H.; Killeen, K.; Brennen, R.; Sobek, D.; Werlich, M.; Van de Goor, T. Anal. Chem. 2005, 77, 527-533. (26) Schultz, G. A.; Corso, T. N.; Prosser, S. J.; Zhang, S. Anal. Chem. 2000, 72, 4058-4063. (27) Licklider, L.; Wang, X.-Q.; Desai, A.; Tai, Y.-C.; Lee, T. D. Anal. Chem. 2000, 72, 367-375. (28) Kameoka, J.; Orth, R.; Ilic, B.; Czaplewski, D.; Wachs, T.; Craighead, H. G. Anal. Chem. 2002, 74, 5897-5901. (29) Koerner, T.; Turck, K.; Brown, L.; Oleschuk, R. D. Anal. Chem. 2004, 76, 6456-6460. (30) Lee, S. S. H.; Douma, M.; Koerner, T.; Oleschuk, R. D. Rapid Commun. Mass Spectrom. 2005, 19, 2671-2680.

In this report, we investigate the utilization of a porous polymer monolith as a nanospray emitter for coupling different polymeric microfluidic devices to a mass spectrometer. A PPM was photopatterned at the end of the channel of microfluidic devices fabricated from three different polymeric substrates including PMMA, cyclic olefin copolymer (COC), and PDMS. The stability and reproducibility of the electrospray are compared for the different substrate materials. It follows that in addition to using the PPM to effect a robust microfluidic-MS coupling, the PPM material may be used for separation and sample preparation. Few research groups have reported the photopatterning of PPM in microfluidic devices fabricated from glass32,33 or from cyclic olefin copolymer34,35 where the PPM was used for chromatography or SPE. To our knowledge, this is the first report of the formation of organic-based porous polymer monoliths in PDMS. EXPERIMENTAL SECTION Materials and Reagents. All solutions for polymerization and mass spectrometry were prepared in >18 MΩ Milli-Q water (Millipore, Bedford, MA). Lauryl acrylate (LA), ethylene glycol dimethacrylate (EDMA), 2-acrylamido-2-methyl-1-propanesulfonic acid (AMPS), methyl acrylamide (MA), poly(ethylene glycol) diacrylate (PEG diacrylate), benzyl alcohol (BA), sodium periodate, benzoin methyl ether, benzophenone, angiotensin I, myoglobin (from horse skeletal muscle), and O-(2-aminopropyl)-O′(2-methoxyethyl)polypropylene glycol 500 (PPG) were purchased from Sigma-Aldrich (Oakville, ON, Canada). Glacial acetic acid, methanol, and 2-propanol (HPLC grade) were obtained from Fisher Scientific (Ottawa, ON, Canada). Slygard 184 silicone elastomer and curing agent were purchased from Dow Corning Corp. (Midland, MI), PMMA sheets were acquired from Warehoused Plastic Sales Inc. (Toronto, ON, Canada), and cyclic olefin plastic plates (ZEONOR750R) were obtained from Zeon Chemicals (Louisville, KY). All chemicals were used without purification. Fabrication of the Microchips. PDMS microchips with channels 50 µm wide and 20 µm deep were fabricated as described previously in detail.36 Briefly, Slygard 184 PDMS prepolymer was mixed thoroughly in a 10:1 mass ratio of silicone elastomer to curing agent and degassed for 1 h. A PDMS master was prepared by pouring the prepolymer mixture onto a glass substrate containing an array of six etched devices purchased from Micralyne (Edmonton, AB, Canada) and was cured at 65 °C for 4 h. The PDMS substrate was prepared by molding the prepolymer mixture against the PDMS master. The PDMS substrate and an unpatterned PDMS cover were then oxidized using a Tesla coil (Fisher Scientific) for 5 min, laid on top of each other, and sealed at 95 °C for 5 h to form an irreversible bond. PMMA and COC substrates were hot embossed over a nickel master manufactured by Tecan Ltd. (Dorset, U.K.) using a HEX (31) Wang, Y.-X.; Cooper, J. W.; Lee, C. S.; DeVoe, D. L. Lab Chip 2004, 4, 363-367. (32) Yu, C.; Davey, M. H.; Svec, F.; Frechet, J. M. J. Anal. Chem. 2001, 73, 5088-5096. (33) Ngola, S. M.; Fintschenko, Y.; Choi, W.-Y.; Shepodd, T. J. Anal. Chem. 2001, 73, 849-856. (34) Stachowiak, T. B.; Rohr, T.; Hilder, E. F.; Peterson, D. S.; Yi, M.; Svec, F.; Frechet, J. M. J. Electrophoresis 2003, 24, 3689-3693. (35) Benetton, S.; Kameoka, J.; Tan, A.; Wachs, T.; Craighead, H.; Henion, J. D. Anal. Chem. 2003, 75, 6430-6436. (36) Wang, B.; Abdulali-Kanji, Z.; Dodwell, E.; Horton, J. H.; Oleschuk, R. D. Electrophoresis 2003, 24, 1442-1450.

Analytical Chemistry, Vol. 78, No. 4, February 15, 2006

1131

Figure 1. (a) Schematic diagram of (i) patterned substrate, (ii) cover plate with embedded electrode, and (iii) assembled chip with magnified image of PPM at the tip of a PMMA chip. Scanning electron micrographs of PPM in different substrate materials, (b) PDMS, (c) PMMA, and (d) delaminated COC. (e-g) show the further magnified SEMs of the porous structure of the PPM formed with the PDMS, PMMA, and COC microfluidic chips, respectively.

01 hot embossing system (Jenoptik Microtechnik, Jena, Germany) for 10 min at 10 kN and 70 °C for COC and 20 kN and 115 °C for PMMA. The substrates were bonded to the cover plates using thermal bonding to produce channels 50 µm wide and 20 µm deep. To fabricate a chip with an embedded electrode, a platinum wire (127 µm) was hot embossed onto the COC cover plate prior to thermal bonding with the COC substrate. Chips were compared to commercial chips purchased from Microfluidic ChipShop GmbH (Jena, Germany) with channel widths of 60 (top) and 32 µm (bottom) and a depth of 20 µm for PMMA and 70 (top) and 42 µm (bottom) and a depth of 20 µm for COC. An electrospray tip (150-µm o.d., 30-µm i.d.) was aligned at the end of the microchannel in a COC substrate and bonded to the cover plate, to produce a COC chip with a nanospray emitter for comparison with the PPM emitter. Nanoports (Upchurch Scientific, Oak Harber, WA) were attached to the reservoir holes of the COC and PMMA chips with epoxy glue PolyBond33 (Nbond Adhesives Int., Littleton, CO) and cured for 24 h at room temperature. Nanoports were attached to the PDMS chips by clamping the nanoport to the PDMS and relying on self-sealing coupled with applied pressure. 1132

Analytical Chemistry, Vol. 78, No. 4, February 15, 2006

Photopatterning of the Porous Polymer Monolith. PDMS chips were photografted according to Hu et al.,37 the assembled microchip was rinsed with benzophenone solution (0.25 M in acetone); the channel was then rinsed with water and filled with the grafting solution (8% AMPS or MA, 2% PEG diacrylate, and 0.5% BA in aqueous solution of 0.5 mM sodium periodate). The chip was irradiated with 8 W of UV light at 254 nm (Spectroline, model ENF-280c) for 10 min. The chip was then rinsed with water and dried with a nitrogen stream. The PPM monomer solution (24% LA, 16% EDMA, 30% methanol, 30% 2-propanol, and 0.4% benzoin methyl ether) was degassed by purging with nitrogen for 10 min and then introduced to the channels of grafted PDMS, PMMA, and COC chips. The PPM was patterned in the channel by masking the undesired parts with electrical tape and irradiating the chip with 8-W UV light at 365 nm for 15 min at a distance of 4.5 cm. Instrumentation. MS experiments were performed on an API 3000 triple quadrupole mass spectrometer (MDS-Sciex, Concord, Canada) fitted with a nanoelectrospray source (Proxeon, Odense, Denmark) consisting of an x-y-z stage and two CCD camera (37) Hu, S.; Ren, X.; Bachman, M.; Sims, C. E.; Li, G. P.; Allbritton, N. L. Anal. Chem. 2004, 76, 1865-1870.

Figure 2. Total ion current for constant infusion of angiotensin I (1 µM, 50% methanol, 0.5% AcOH) for COC chip with a PPM sprayer or open channel with representative mass spectrum extracted from each of the TICs. Flow rate 500 nL/min; applied voltage of 4.5 kV. Mass spectra were obtained from averaging 10 s of the TIC and represent ∼80 fmol. (We emphasize that the PPM filled channel produced a 1 order of magnitude increase in counts per second compared to open channeled device.)

Figure 3. Total ion current for constant infusion of angiotensin I (1 µM, 50% methanol, 0.5% AcOH) sprayed from a PDMS chip with a PPM or open channel. Flow rate of 500 nL/min; applied voltage 4.5 kV. Mass spectra were obtained from averaging 10 s of the TIC and represent ∼80 fmol.

kits to aid in the positioning of the chip ∼5 mm from the orifice of the MS. Solution was driven by an Eksigent NanoLC pump (Livermore, CA) to the chip nanoport through a 50-µm-i.d. fusedsilica capillary. The ES voltage was supplied by connecting the MS power supply either to the embedded platinum electrode or to a stainless steal union in the transfer line. All spectra were acquired in positive ion mode with a scan range of 300-1200 m/z at 1 Hz and nitrogen curtain gas flow of 2 L/min. Stability of the TIC was calculated by dividing the standard deviation of the TIC data points on the average of the data points and represented as percentage (RSD).

RESULTS AND DISCUSSION Polymeric µTAS devices can be mass-produced rapidly using technologies such as casting, injection molding, or hot embossing. The cost-effectiveness of polymeric devices permits their use in “single-use” applications, thus eliminating any cross-contamination or sample carryover problems known for continuous use devices. Interfacing these devices to a universal detection tool such as mass spectrometry has advantages over other detection techniques that require the analyte to be fluorescent. We describe a robust method for coupling polymeric microfluidic devices to mass spectrometry, utilizing porous polymer monolith as an electrospray emitter Analytical Chemistry, Vol. 78, No. 4, February 15, 2006

1133

Figure 4. Total ion current for constant infusion of angiotensin I (1 µM, 50% methanol, 0.5% AcOH) sprayed from a PMMA chip with a PPM sprayer or open channel. Flow rate of 500 nL/min; applied voltage 4.5 kV. Mass spectra were obtained from averaging 10 s of the TIC and represent ∼ 80 fmol.

Figure 5. Optical microscope images of stable Taylor cone established at a microchannel exit of a COC chip with PPM patterned at the exit of the channel for 10 mm, applied voltage of 4.0-4.5 kV at flow rate of 100 (a), 500 (b), and 1000 (c) nL/min at a counter electrode spacing of 5 mm.

patterned at the end of the microchannel. Spraying directly from the end of the microchannel removes any dead volume associated with inserted emitters or transfer lines, and the presence of multiple pathways in the PPM prevents the clogging of the channels, which is a common problem in conventional nanospray emitters. The PPM multiple pathways also divide the flow in the channel into smaller streams to create the potential of multiple electrospray emitters, thus producing smaller droplets from individual Taylor cones that require less desolvation producing improved ionization efficiency. Formation of the PPM Emitter in the Polymeric Chips. Lauryl acrylate-co-ethylene dimethacrylate PPM was photopatterned for 5 mm at the end of the channel of microfluidic devices fabricated from three different polymeric substrates including COC, PMMA, and PDMS, Figure 1. The polymerization conditions were chosen in the light of the work published by Yu et al.,38 where it was shown that use of low molecular weight alcohols as the porogenic solvent provided a compatible solvent system with the polymeric chips. 1134

Analytical Chemistry, Vol. 78, No. 4, February 15, 2006

Among the polymeric substrate materials tested, COC was found to be the most reliable for patterning a PPM within a microchannel. COC exhibits good optical transparency above 230 nm and shows good chemical resistance to both the polymerization solution and the organic solvents commonly used in electrospray (e.g., methanol and acetonitrile). Compared to COC, PMMA has a lower UV transparency at 254 nm, and we have found the chemical resistance of PMMA strongly depended on the source of the PMMA sheets used. For example, the in-laboratory embossed PMMA chips showed higher chemical resistance to the polymerization solution compared to the commercially obtained chips. However, channels within chips prepared with both PMMA materials were distorted when high content acetonitrile (>70%) solutions were used. PDMS is both chemically resistant to most solvents and highly optically transparent; however, no PPM could be formed in native or oxidized PDMS channels and, consequently, required a grafting (38) Yu, C.; Xu, M.; Svec, F.; Frechet, J. M. J. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, 755-769.

Figure 6. Extracted ion current (432.8-433.2) for constant infusion of angiotensin I (1 µM, 50% methanol, 0.5% AcOH) sprayed from a COC chip having a PPM emitter at flow rate of 50 (a) and 100 nL/min (b). TIC where spraying voltage 4.0 kV was turned on and off to regenerate the electrospray process at a flow rate of 50 nL/min (c).

step to facilitate the PPM formation. The difficulty in producing PPM within native PDMS devices is probably due to the adsorption of the monomers/porogenic solvent or the permeability of PDMS to oxygen. Oxygen quenches free radical polymerization, which either results in no polymerization or in the formation of a nonporous gel in the channels. To minimize such adsorption, PDMS channels were surface modified. Several methods have been described in the literature to modify the channel wall surfaces of PDMS devices.37,39-42 Three grafting procedures were examined including the following: oxidation of PDMS surface followed by covalent modification with AMPS using cerium(IV) catalyzed polymerization,42 UV grafting of native PDMS with acrylamide,41 and UV grafting of PDMS with preadsorbed benzophenone with methyl acrylate or AMPS/poly(ethylene glycol) diacrylate.37 UV grafting with preadsorbed benzophenone proved to be the most efficient of the three methods for minimizing the adsorption of the polymerization solution and the formation of the PPM. The adsorption of the benzophenone to the inner surface of the PDMS channels prior to the introduction of the grafting solution increases the rate of surface grafting to polymerization within the microchannel. (39) Makamba, H.; Hsieh, Y.-Y.; Sung, W.-C.; Chen, S.-H. Anal. Chem. 2005, 77, 3971-3978. (40) Kang, J.; Wistuba, D.; Schurig, V. Electrophoresis 2002, 23, 1116-1120. (41) Rohr, T.; Ogletree, D. F.; Svec, F.; Frechet, J. M. J. Adv. Funct. Mater. 2003, 13, 264-270. (42) Slentz, B. E.; Penner, N. A.; Regnier, F. E. J. Chromatogr., A 2002, 948, 225-233.

Electrospray Performance. Spraying from a chip having a PPM emitter produces a substantial increase in the TIC stability and sensitivity compared to spraying from the corresponding chip with an open channel, Figures 2-4. The stability (RSD < 5.0%) of the TIC produced by constant infusion of 1 µM solution of angiotensin I increased by 1 order of magnitude for COC and PDMS chips having a PPM emitter compared to spraying from an open end chip with no PPM (Figures 2 and 3, respectively). The multiple flow paths in the PPM emitter generate a potential of multiple electrospray emitters that may create smaller Taylor cones, which produce smaller sized droplets with a large surface area-to-volume ratio. The reduced droplet size increases the analyte concentration at the droplet surface, thus requiring less desolvation. In the case of PMMA chips, Figure 4, there was no increase in sensitivity with the patterned PPM, which may be attributed to the smaller dimensions of the PMMA channel compared to that of the COC and the PDMS, Figure 1. The shallower channel produced smaller droplets resulting in better TIC stability. The low signal stability obtained by spraying from an open channel is contributed to spreading of the large Taylor cone formed at the exit of the channel. The Taylor cone formation for the COC chip having a PPM emitter was characterized at different flow rates (100, 500, and 1000 nL/min), Figure 5. A 4.0-4.5-kV potential was applied to generate the Taylor cone with a counter electrode spacing of ∼5 mm. At a flow rate below 200 nL/min, no Taylor cone was observed; only the liquid jet mist of the electrospray was seen. Analytical Chemistry, Vol. 78, No. 4, February 15, 2006

1135

Figure 7. Total ion currents for constant infusion of 1 µM PPG in different percent organic sprayed from a COC chip having a PPM emitter at flow rate of 500 nL/min (a), mass spectrum obtained by averaging 5 s of the TIC and represent ∼40 fmol from the TIC generated with 50% organic in (b).

The mist presumably results from a number of extremely small Taylor cones that emanate from one of the many pores at the exit aperture of the device. The lack of a large-volume Taylor cone should reduce band broadening associated with spraying directly from a chip.6 It follows that porous polymer assisted electrospray should be compatible with chips that utilize electroosmotic pumping that typically produce flow rates of ∼25-100 nL/min. At a higher flow rate of 500 nL/min, a single confined smallvolume Taylor cone was produced at the exit of the channel and no droplet spreading was apparent. Similar to the low flow rates, band broadening associated with Taylor cone production should be minimal at this flow rate. As the flow rate increased to 1000 1136

Analytical Chemistry, Vol. 78, No. 4, February 15, 2006

nL/min, the volume was too large for the sprayer and the Taylor cone wetted the triangular tip of the chip exit. The optimal performance of the PPM sprayer at low flow rates is in agreement with our recently published results obtained by spraying from a glass chip with a PPM sprayer.43 The interchip reproducibility was tested by comparing the MS data obtained from three different chips using constant infusion of angiotensin I. COC proved to produce the highest chip-to-chip reproducibility (RSD < 20%) due to the favorable conditions of forming the PPM emitter in the COC compared to the PDMS or (43) Koerner, T.; Oleschuk Richard, D. Rapid Commun. Mass Spectrom. 2005, 19, 3279-3286.

Figure 8. Total ion current for constant infusion of 1 µM PPG (400 nL/min, 3.5 kV applied voltage) sprayed from a COC with 100- (a) and 50-µm (b) channel width. Both chips had PPM emitters (8 mm) and embedded platinum electrodes positioned at 10 mm from the exit of the chip.

the PMMA. The intrachip reproducibility was found to be less than 5% RSD for MS data collected for a period over 9 h. The chip PPM emitter was found to be very durable; chips were stored after testing for several weeks in a dry state and reused without significant loss of performance. The performance of the electrospray from a COC chip with PPM emitter with constant infusion of angiotensin I was examined at flow rates of 50 and 100 nL/min. This flow rate was chosen because it is within the limits normally generated by electroosmotic flow. Figure 6a and b shows comparable performances in terms of stability and sensitivity of the chip at both flow rates. The ability to initiate and stop the electrospray from a chip would be advantageous during certain sample preparation schemes (i.e., preconcentration). To test the ability of the PPM chip, we toggled the spray on and off. Figure 6c shows the TIC generated by the PPM sprayer at 50 nL/min by turning off the electrospray voltage for 10 s and back on to 4.0 kV, the electrospray was immediately regenerated when the voltage was reapplied. The effect of the organic modifier in the electrospray solution on the surface wetting of the chip exit and the electrospraying process was also studied. The constant infusion of 1 µM poly(propylene glycol) solutions with different organic contents ranging from 10 to 90% methanol were electrosprayed from a COC chip with a PPM emitter, Figure 7. An electrospray could be established with an organic modifier as low as 10% due to the hydrophobic nature of the PPM as well as the cyclic olefin chip. The stability of the electrospray was, however, best (RSD of the TIC < 4.0%) at 50% methanol, similar to that observed with conventional nanospray. The electrospray voltage was applied to the solution in the microchannels by connecting the MS power supply either to the embedded platinum electrode or to a stainless steal union in the transfer line. To fabricate a COC chip with embedded electrode, a platinum wire was first embossed on the cover plate for 5 min at 70 °C and 5 kN before the subsequent thermal bonding with the patterned plate. Three electrode diameters (25, 127, 300 µm) were tested for robustness of fabrication and possible blockage

in the microchannel after thermal bonding with the patterned plate. The 127-µm-diameter platinum wire was found to be optimum for electrode construction, while the 25 µm was too fragile for the hot embossing and the 300 µm tends to block the microchannel. The application of the electrospray voltage 1 cm from the chip exit decreased the required ES voltage to 3.5 kV compared to the 4.5 kV needed when the voltage applied in the solution transfer line, as can be seen from the total ion current for constant infusion of 1 µM poly(propylene glycol) in Figure 8. The ability to embed an electrode at the exit of the chip both enables facile control of the electrospray voltage and facilitates efforts to connect microfluidic chips to MS with electroosmotic flow. The length and relative hydrophobicity/hydrophilicity of the PPM can be controlled to either maximize or minimize analyte retention. As a result, the patterned PPM can be used to function as either a separation column and sprayer or sprayer alone. The potential of the PPM sprayer was demonstrated by its application as a solid-phase extraction bed for the preconcentration and desalting of protein samples prior to the mass spectrometric detection as shown in Figure 9. A long sample plug pf 225 nL of myoglobin (2 × 10-6 M in 50 mM ammonium acetate) was injected in a chip having a 5-cm PPM bed; the sample was stacked on the top of a PPM bed and desalted by rinsing with water for 10 min; the myoglobin was eluted with a step gradient to 80% acetonitrile having 0.5% acetic acid. Spraying from the end of the PPM-filled channel was compared to that from a chip having an embedded electrospray tip at the end the PPM-filled channel. Results in terms of sensitivity were comparable in both the XIC (monitoring [M + 21H]21+ ion) and the resulting MS spectrum. Although the sensitivity of the two coupling methodologies is similar, the PPM sprayer is significantly easier to fabricate and provides a zero dead volume connection. CONCLUSIONS PPM-assisted electrospray provides a facile methodology for coupling microfluidic devices to mass spectrometry that is sheathAnalytical Chemistry, Vol. 78, No. 4, February 15, 2006

1137

Figure 9. Desalting and preconcentration of myoglobin (2 × 10-6 M in 50 mM ammonium acetate) using COC chips having 5-cm PPM column. Injection of 225 nL, followed by rinsing with water for 10 min and eluting with 80% acetonitrile containing 0.5% acetic acid. (a) The extracted ion chromatograms of myoglobin [M + 21H]21+, 893.5, sprayed from the end of the separation channel (b) and from a COC chip with embedded electrospray tip. XIC chromatograms are offset by 20% in the both the x and y axes. The MS representing each XIC was averaged from a 20-s window around the peak maximum.

less and has zero dead volume. This study shows that a PPM patterned at the exit orifice of a microfluidic device significantly enhances the stability of the ESI process. The improvement presumably stems from improved ionization efficiency resulting from the formation of multiple Taylor cones at flow rates lower than 100 nL/min. The appearance of extremely small Taylor cones at low flow rates should minimize any dead volume associated with the Taylor cone, preventing a loss in separation resolution. The PPM emitter can be prepared at various lengths and in any position within the microfluidic architecture through appropriate photomasking. PPM emitters constructed in three low-cost substrate materials (i.e., PDMS, COC, and PMMA) all produced stable electrosprays at a flow rate of 500 nL/min. However, COC showed the highest reproducibility as a result of the enhanced robustness of the PPM fabrication method. The performance of the PPM-assisted electrospray at low flow rates (50-100 nL/min) compares favorably with electroosmotic flow pumping velocities in microfluidic channels. Samples containing different amounts of organic modifier (10-90%) could be successfully electrosprayed from a COC PPM patterned microfluidic device with maximum 1138 Analytical Chemistry, Vol. 78, No. 4, February 15, 2006

sensitivity and stability with 50% CH3OH. Choosing different PPM monomers enables the incorporation of surface functionalities within the PPM. It follows that in addition to acting to enhance electrospray performance when directly coupling microfluidic devices with mass spectrometry the PPM can be customized to act as a separation medium. Although PPM-assisted electrospray is demonstrated to provide enhanced performance as a microfluidic MS interface, the method should find utility in other lowflow separation and MS coupling applications. ACKNOWLEDGMENT We thank Genome Prairie (Enabling Technologies Project), the Natural Sciences and Engineering Research Council of Canada, Canadian Foundation for Innovation, Ontario Innovation Trust, and Queen’s University for financial support.

Received for review August 12, 2005. Accepted December 8, 2005. AC0514570