Sheathless Electrospray from Polymer Microchips - Analytical

Solomon Ssenyange, Justine Taylor, D. Jed Harrison, and Mark T. McDermott. Analytical Chemistry .... a glass microdevice. Terry Koerner , Richard D. O...
0 downloads 0 Views 262KB Size
Download PDF
Anal. Chem. 2003, 75, 3934-3940

Sheathless Electrospray from Polymer Microchips Malin Svedberg,† Andreas Pettersson,‡ Stefan Nilsson,‡ Jonas Bergquist,‡ Leif Nyholm,‡ Fredrik Nikolajeff,† and Karin Markides*,‡

The A° ngstro¨m Laboratory, Department of Materials Science, Uppsala University, P.O. Box 534, 751 21 Uppsala, Sweden, and Department of Analytical Chemistry, Uppsala University, P.O. Box 599, 751 24 Uppsala, Sweden

In this study, sheathless electrospray from polymer microchips with conducting layer on the emitter tip is described for the first time. The electrospray emitter tips were fabricated directly from the end of the microchips that were made of polycarbonate or poly(methyl methacrylate) with injection molding. A variety of tip shapes and conducting coatings were evaluated using an electrospray time-of-flight mass spectrometer run in the sheathless mode. Stable electrospray was obtained both from hand-polished and machine-milled three-dimensional tips coated with either polymer-embedded gold particles or graphite particles as the conducting layer. Sputtered gold, on the other hand, suffered from poor stability mainly due to bad adhesion to the polymer tip. The durability of the different coatings was confirmed with electrochemical experiments under simulated electrospray conditions. The relative standard deviations of the response received from the ion current of the MS analysis were in the range of 3.5-12%. The detection limit for a standard mixture containing five neuropeptides was lower than 0.5 fmol. The low detection limit makes the emitter tips highly attractive for the analysis of low-abundance biological species. The increased interest in proteomics and the analysis of biological samples present in low concentrations and often in limited amount of sample has increased the demands for better and more sensitive analysis systems.1 The lab-on-a-chip technology2 offers great advantages in separation applications with its fast, high-efficiency separation of small sample volumes and low dead volumes. Interest in polymer-based microchips for chemical analysis purposes has increased in recent years, since methods for accurate plastic replication that fulfill the analytical chemical requirements have become available. The use of polymers in microfabrication gives increased possibilities to vary the chemical and biochemical properties, either by choice of polymer or by surface treatment. Replicated microstructures in plastic also has the potential to enable low-cost biochemical devices, which leads to the possibility of producing disposable chips to be used, for example, clinical analysis.3 * Corresponding author. E-mail: [email protected]. Fax: +46(0)18 471 3692. † Department of Materials Science. ‡ Department of Analytical Chemistry. (1) Papac, D. I.; Shahrokh, Z. Pharmacol. Res. 2001, 18, 131-145. (2) Mello, A. J. Lab Chip 2001, 1, 7N-12N.

3934 Analytical Chemistry, Vol. 75, No. 15, August 1, 2003

To benefit from highly efficient separations, a sensitive and reliable detection method is needed. One detection technique that has gained much interest is electrospray ionization (ESI) coupled to a time-of-flight mass spectrometer (TOF-MS) mainly because of its ability to detect large biomolecules with great sensitivity and accuracy. ESI also enables liquid separations to be coupled to mass spectrometry with high ionization efficiency. Several constructions of an interface between a chip and ESIMS have been reported in the literature. A few groups4,5 have described direct electrospray from microchips made in glass without any tip at the exit channel. The problem they encountered was to establish a well-defined Taylor cone at the exit of the microchannel. Extensive wetting of the emerging liquid over the flat edge of the microchip resulted in poor cone formation. Rohner and co-workers6 constructed a similar planar microchip in poly(ethylene terephthalate) with decreased wetability, but the problem of aligning the Taylor cone in front of the microchannel exit still existed. External electrospray emitters have been added to the chip by inserting a spray capillary at the microchannel exit,7-15 by constructing a microplate16 or by using a liquid junction between the microchip and a multiple electrospray emitter placed in front of the mass spectrometer.17 Although these techniques show high sensitivity and robustness, and allow multiple analysis without (3) Becker, H.; Locascio, L. E. Talanta 2002, 56, 267-287. (4) Ramsey, R. S.; Ramsey J. M. Anal. Chem. 1997, 69, 1174-1178. (5) Xue, Q.; Foret, F.; Dunayevskiy, Y. M.; Zavracky, P. M.; McGruer, N. E.; Karger, B. L. Anal. Chem. 1997, 69, 426-430. (6) Rohner, T. C.; Rossier, J. S.; Girault, H. H. Anal. Chem. 2001, 73, 53535357. (7) Xu, N.; Lin, Y.; Hofstadler, S. A.; Matson, D.; Call, C. J.; Smith, R. D. Anal. Chem. 1998, 70, 3553-3556. (8) Meng, Z.; Qi, S.; Soper, S. A.; Limbach, P. A. Anal. Chem. 2001, 73, 12861291. (9) Gao, J.; Xu, J.; Locascio, L. E.; Lee, C. S. Anal. Chem. 2001, 73, 26482655. (10) Jiang, Y.; Wang, P. C.; Locascio, L. E.; Lee, C. S. Anal. Chem. 2001, 73, 2048-2053. (11) Vrouwe, E. X.; Gysler, J.; Tjaden, U. R.; Van der Greef, J. Rapid Commun. Mass Spectrom. 2000, 14, 1682-1688. (12) Chartogne, A.; Tjaden, U. R.; Van der Greef, J. Rapid Commun. Mass Spectrom. 2000, 14, 1269-1274. (13) Liu, H.; Felten, C.; Xue, Q.; Zhang, B.; Jedrzejewski, P.; Karger, B. L.; Foret, F. Anal. Chem. 2000, 72, 3303-3310. (14) Figeys, D.; Aebersold, R.; Lock, C. J. Capillary Electrophor. Microchip Technol. 1999, 6, 1-6. (15) Chen, S. H.; Sung, W. C.; Lee, G. B.; Lin, Z. Y.; Chen, P. W.; Liao, P. C. Electrophoresis 2001, 22, 3972-3977. (16) Tang, K.; Lin, Y.; Matson, D. W.; Kim, T.; Smith, R. D. Anal. Chem. 2001, 73, 1658-1663. (17) Kameoka, J.; Craighead, H. G.; Zhang, H.; Henion, J. Anal. Chem. 2001, 73, 1935-1941. 10.1021/ac030045t CCC: $25.00

© 2003 American Chemical Society Published on Web 07/04/2003

cross-contamination, the drawback of these chips will always be the increased dead volume with subsequent hampered usefulness for real applications. An alternative approach has been to fabricate an electrospray tip from the bulk material at the outlet of the microchannel. This has been done in microchips in a variety of different bulk polymers resulting in enhanced spray performance and a more well-defined Taylor cone formation, obtained with a lower electrospray potential compared to the planar chip-exit ESI.18-23 Wen and co-workers constructed microchannels in a polycarbonate (PC) chip material to be used for isoelectric focusing.18 The tip was mechanically machined directly from the chip and successfully evaluated with different sheath flows of liquid and gas to aid the spray. Another microchip was fabricated in poly(dimethyl siloxane) material, and the electrospray potential was applied via a liquid junction that gave long durability of the spray in a large range of flow rates.19,20 Different outer tip angles have also been evaluated by mechanically constructing microchannel openings in poly(methyl methacrylate) (PMMA) with a knife. The width of the channels used was 375 µm, and an electrospray potential of 3800 V was needed to get stable electrospray.21 Gobry and co-workers constructed microchannels in a polyimide substrate by plasma etching where the emitter tip was sharpened with a knife. A limit of detection of 40 fmol/µL was obtained for a peptide solution.22 Recently, a microchip in Zeonor was presented where a triangular parylene piece sandwiched between two microchip pieces was used as the electrospray emitter. The construction allowed the attachment of a dense ESI array device without cross-contamination from adjacent emitter tips and with good performance for high-throughput analysis.23 Common to all chip-ESI devices described above is the fact that the electrospray potential has been applied upstream in the microchannel. This approach limits the ability to control the effective electrospray potential since a potential drop will be present in the solution. This may result in further band broadening between the electrode and the electrospray. One way to minimize this problem is to apply the potential on a conducting layer adhered to the outside of the electrospray tip, an approach that in this paper will be referred to as sheathless ESI. Here, no aiding liquids or gases are used to produce the spray, which means that there is no dilution of the sample. In addition, low flow rates can be used, which is beneficial from a sensitivity point of view since lower flow rates may provide smaller droplets and hence higher ionization efficiency.24 Since the ESI potential is applied on a conducting layer onto the tip, the need for a durable coating is an important issue. Barnidge and co-workers 25 used polyimide to glue 2-µm gold particles to form a conducting layer on a sharpened fused-silica capillary. The method, also known as “fairy (18) Wen, J.; Lin Y.; Xiang, F.; Matson, D. W.; Udseth, H. R.; Smith, R. D. Electrophoresis 2000, 21, 191-197. (19) Kim, J. S.; Knapp, D. R. J. Am. Soc. Mass. Spectrom. 2001, 12, 463-469. (20) Kim, J. S.; Knapp, D. R. J. Chromatogr., A 2001, 924, 137-145. (21) Yuan, C -H.; Shiea, J. Anal. Chem. 2001, 73, 1080-1083. (22) Gobry, V.; Van Oostrum, J.; Martinelli, M.; Rohner, T. C.; Reymond, F.; Rossier, J. S.; Girault, H. H. Proteomics 2002, 2, 405-412. (23) Kameoka, J.; Orth, R.; Ilic, B.; Czaplewski, D.; Wachs, T.; Craighead, H. G. Anal. Chem. 2002, 74, 5897-5901. (24) Cole, R. B. Electrospray ionization mass spectrometry: fundamentals, instrumentation and applications; John Wiley & Sons: New York, 1997. (25) Barnidge, D. R.; Nilsson, S.; Markides, K. E. Anal. Chem. 1999, 71, 41154118.

dust”, prolonged the emitter lifetime from hours in earlier described designs to months. This technique was later developed to also include the use of different gluing media 26 and also conducting graphite particles 27 instead of gold particles, without any loss in durability. In this work, sheathless electrospray from polymer microchips with conducting layer on the emitter tip is described for the first time. The aim has been to fabricate and evaluate possible designs of sheathless electrospray emitter tips made directly from plastic chips. The final goal was to find a fast and reliable replication approach as well as a sensitive and reproducible mass spectrometric method in order to analyze low-abundant endogenous compounds such as neuropeptides. The challenges in this study were to find a sufficiently stable conducting coating that adheres to the emitter tip and to construct tips with as small an outer radius as possible. The tips were fabricated directly from injection-molded microchips made in the thermoplastics PMMA and PC, and the tips were tested for sheathless electrospray with different conducting coatings. The stabilities of the conducting coatings were also further tested with electrochemical methods.28 PMMA has been demonstrated to be a suitable column material for separation with capillary electrophoresis (CE) 29-34 and could therefore be of high interest to utilize in a separation chip. PC, whose material properties and experience of injection molding parameters are available from the CD industry, has also been found to provide good properties for CE applications.35 EXPERIMENTAL SECTION Chemicals. The mixture of neuropeptides was purchased from Sigma Chemicals Co. (St. Louis, MO). Methanol (Lichrosolv), acetic acid, nitric acid (both pro analysi), and methylene blue (electrophoresis grade) were from Merck (Darmstadt, Germany). The deionized water used in all experiments was obtained from a Milli-Q+ system (Millipore Corp., Marlborough, MA). Pellets of Makrolon 1265 (Bayer) and Plexiglas DQ501 (Ro¨hm) were used for injection molding of the PC disks and PMMA disks, respectively. Fabrication of the Microfluidic Channels. The device contained a channel with two sections (see Figure 1a). The entrance channel, which was 370 µm wide with a depth of 180 µm and a length of 2 cm, was used to connect the chip with a fused-silica capillary (Polymicro Technologies, Phoenix, AZ) to (26) Nilsson, S.; Markides, K. E. Rapid Commun. Mass Spectrom. 2000, 14, 6-11. (27) Nilsson, S.; Wetterhall, M.; Bergquist, J.; Nyholm, L.; Markides, K. E.,Rapid Commun. Mass Spectrom. 2001, 15, 1997-2000. (28) Nilsson, S.; Svedberg, T. M.; Pettersson, J.; Bjo¨refors, F.; Markides, K.; Nyholm, L. Anal. Chem. 2001, 73, 4607-4616. (29) Shao, X.; Shen, Y.; O’Neill, K.; Lee, M. L. J. Microcolumn Sep. 1999, 11, 325-329. (30) Qi, S.; Lee, M. L. J. Microcolumn Sep. 1998, 10, 605-609. (31) Lee, G.-W.; Chen, S.-H.; Huang, G. R.; Sung, W.-C.; Lin, Y.-H. Sens. Actuators, B 2001, 75, 142-148. (32) Ford, S. M.; Davis, J.; Kar, B.; Qi, S. D.; McWhorter, S.; Soper, S. A.; Malek, C. K. J. Biomed. Eng. 1999, 121, 13-21. (33) Ford, S. M.; Kar, B.; McWhorter, S.; Davies, J.; Soper, S. A.; Klopf, M.; Calderon, G.; Saile, V. J. Microcolumn Sep. 1998, 10, 413-422. (34) Chen, S. H.; Lee, G. B.; Sung, W. C.; Huang, G. R.; Chen, Y. H.; Young, K. C.; Chang, T. T. Proceedings of the Micro Total Analysis Systems 2000 Symposium, Enschede, The Netherlands, May 14-18, 2000; pp 497-500. (35) Liu, Y.; Ganser, D.; Schneider, A.; Lin, R.; Grodzinski, P.; Kroutchinina, N. Anal. Chem. 2001, 73, 4196-4201.

Analytical Chemistry, Vol. 75, No. 15, August 1, 2003

3935

Figure 1. (a) Experimental setup with the electrospray device. The sample is introduced via a syringe pump, and the high voltage is applied to the conductive coating at the tip. (b) Schematic representation of the tips fabricated in this work: I, machine-milled 3D tip; II, machine-milled 2D tip; III, hand-polished 3D tip.

the chip. The exit channel was 100 µm wide, 70 µm deep, and 4 cm long and ended at the tip outlet. The pattern was first defined in a photosensitive resist layer spun onto silicon wafers, 525 µm thick with a diameter of 100 mm, using photolithography. Microchannels for the device were wet etched in the silicon wafers in a 440 g/L potassium hydroxide solution at 80 °C for 158 min. The wet etching continued until a depth of 180 µm was received in the wider channel. The narrower channel was at that time completely etched and the cross section was V-formed, while the wider channel had a trapezoidal cross section; see Figure 1a. The channel pattern was next transferred to a nickel mold by electroplating. To increase the conductivity of the silicon wafer during the electroplating, a seed layer consisting of 25-nm titanium and 25-nm nickel was first deposited using a magnetron sputter (MDX Advance Energy, Balzers, Liechtenstein). The electroplating was carried out using equipment for replication of compact disks (Microform 12 from Toolex Alpha, Sundbyberg, Sweden). Nickel was plated to a thickness of 300 µm in a bath of nickel sulfamate and boric acid at 55 °C. Afterward, the silicon wafer was etched away from the nickel using a 30 wt % potassium hydroxide solution at 80 °C. The whole replication procedure is described in Figure 2. The nickel master was punched to become the stamper and used for the mass production of plastic compact disks (CDs) in 3936 Analytical Chemistry, Vol. 75, No. 15, August 1, 2003

Figure 2. Schematic of the replication from silicon wafer to plastic chip.

PC and PMMA using a commercial injection-molding machine (Mould Express, Toolex Alpha). The nickel mother and the injection molded plastic disks were characterized with a stylus profilometer (Dektak 200-Si, Veeco Instruments) in order to check the quality of the replication.

Dicing and Sealing. The CDs were diced into smaller pieces so that the short edges of each piece had open microchannels at both ends. PC pieces were cut using a pair of paper scissors. PMMA was broken by making a notch with a knife and then applying force. A cover lid of the same material and size was diced from a blank disk. Contact with the surfaces that were to be bonded was minimized during dicing. The channels and lid were rinsed with deionized water and dried with pressurized air. The cover lid was thermally bonded by hand onto the channel piece using hand pressure and heating the pieces on a hot plate at 160 and 185 °C for PMMA and PC, respectively. The channels were tested for leakage using a concentrated solution of methylene blue. After testing, the chips were rinsed with ethanol to remove traces of the dye. Fabrication of the ESI Emitters. After dicing and bonding, tips were machined directly from the chips. This was done in two ways. Some tips were made by manual polishing using abrasive SiC paper (Struers) with a mesh size of 300, 1200, and 4000 on a rotating polishing machine (Buehler, Metaserv 2000 grinder/ polisher). Cooling water was added continuously during polishing. Other tips were machined using a high-precision milling machine (Kern). Pressurized air was used as the cooling medium. Tips in both the two-dimensional (2D) and three-dimensional (3D) format were fabricated; see Figure 1b. Fused-silica capillaries (o.d. 150 µm, i.d. 70 µm) were inserted and glued into the larger end of the channel with a two-component epoxy glue (Bostik, Middleton, MO), which was allowed to cure at room temperature overnight. Deposition of the Conducting Layer. Three different coatings were applied on the fabricated electrospray chip devices. On some tips, gold was sputtered on both sides of the chips in a coater (E 5000 M Hercules, Bio-Rad) for scanning electron microscopy (SEM), using a plasma current of 10 mA, a voltage of 1.2 kV, and an argon pressure of 0.1 mbar for 3 min on each side. The thickness of the obtained coating was estimated to 35 nm. Other tips were covered with gold particles (Goodfellow, Huntingdon, U.K.) with a diameter of 2 µm according to the fairy dust procedure described previously.26 Some tips were also coated with 1-2-µm graphite particles (Aldrich Chemicals Inc., Milwaukee, WI).27 A thin layer of silicon glue was applied on the tip and a few centimeters upstream of the chip. Particles were then dusted onto the silicon glue (Bostik). The channel was purged with a low flow of nitrogen gas during the whole procedure in order not to let particles block the exit of the channel. The silicone were allowed to cure overnight. Mass Spectrometry. The performance of the ESI tips was investigated using a standard mixture of five neuropeptides: angiotensin II (MW 1046.2), [Arg8]-vasopressin (MW 1083.4), leucine-enkephalin (MW 555.3), methionine-enkephalin (MW 573.2), and luteinizing hormone releasing hormone, LHRH (MW 1181.6). The peptide standard was made up in 50:50 methanol/ Milli-Q water, with a 0.1% (v/v) acetic acid content and delivered to the MS with a syringe pump (Harvard PHD 2000, Holiston, MA) at flow rates in the range of 0.7-2.0 µL/min. All MS experiments were performed on a Jaguar orthogonal time-of-flight instrument (Leco, St. Joseph, MI). The original ESI interface was replaced by the microchip, which was mounted on

an xyz translation stage (Protana) and placed in front of the interface plate. The distance between the microchip tip and the orifice entrance was 3-5 mm. The microchip was connected to the ESI voltage of the mass spectrometer via the conducting layer on the microchip tip by a simple crocodile clip connection. No nebulizing gas or other spray-forming aids was used with the exception of a nitrogen curtain gas flow that was set to 600 mL/ min and heated to 100 °C in order to enhance the solvent evaporation from the spray. All spectra were obtained over the m/z range 0-6000 with a pulsing frequency of 5 kHz. The software used for collecting and evaluating the mass spectrometry data were Sensar Corp. v 1.61 and ChromaTOF. 1.00 Jaguar driver 3.12 Beta. (Leco). Electrochemical Examinations. Unstructured PC and PMMA pieces with the different conductive coatings were investigated with cyclic voltammetry (CV) and chronoamperometry (CA) in order to study the electrochemical behavior of conducting layers on the plastic materials under simulated electrospray conditions. The experiments were performed using a PAR 273 potentiostat (EG&G Princeton Applied Research, Princeton, NJ). The coated plastic piece acted as the working electrode in the voltammetric measurement setup and was prepared as follows. Flat pieces, 1 × 2 cm2, of PC and PMMA were diced from blank CDs and covered with sputtered gold, fairy dust, or graphite particles on one side in the same manner as described for the electrospray devices. Prior to measurement, the pieces were rinsed with ethanol and Milli-Q water. A graphite rod was used as the counter electrode, and an Ag/ AgCl electrode served as the reference electrode. The CV and CA experiments were performed in a water solution containing 0.12 mM HNO3. In the CV, the potential was scanned between 0 and +2V with a scan rate of 50 mV/s. In the CA experiments, a constant potential was applied, for 600 s, while the current was monitored. Both cathodic (application of a constant potential of -2.5 V versus Ag/AgCl) and anodic (application of a constant potential of +2.5 V versus Ag/AgCl) CA experiments were performed. Data from the PAR 273 potentiostat were collected using the PAR Electrochemistry software (EG&G Princeton Applied Research), and the data were processed using Matlab 6.1 (MathWorks Inc.). Microscope Examinations. Some of the electrospray devices and electrochemically evaluated pieces were further examined using SEM with a LaB6 filament (Leo 440, Thornwood, NY) and optical microscopy. Safety Considerations. All acids and bases should be handled with great care. Precautions should be taken to avoid exposure of the eyes, especially when handling concentrated potassium hydroxide solution. Nickel solutions are allergenic. To avoid electrical shock, the high-voltage power supplies should be handled with extreme care. RESULTS AND DISCUSSION Fabrication of the Electrospray Device. Stylus profiles show that the replication fidelity of PC is almost 100% while the degree of mold filling for PMMA may be a little bit lower. The fidelity for PMMA is a question of optimization of the injection molding parameters and should not be considered as a limitation when PMMA is used for fabrication of injection-molded microfluidic structures. In Figure 3, SEM pictures of two ESI emitters in Analytical Chemistry, Vol. 75, No. 15, August 1, 2003

3937

Figure 3. SEM pictures of 3D tips: (a) hand-polished PMMA tip; (b) outlet from a machined-milled electrospray tip in PMMA.

Figure 4. SEM pictures of gold sputtered on PMMA, where the gold has flaked off during reduction at -2.5 V for 600 s. Dark regions are areas where the gold has flaked off while the gold film is still intact in the light regions. Arrows point at some regions where the rounded shapes of the flaked-off areas can be seen.

PMMA are shown. Of all the examined fabricated chips, the outlets of the channels can only be observed on one chip, in one of the machine-milled tips shown in Figure 3b. On all other microchips, the outlets are covered with a thin layer of polymer, which means that the opening through which the sample solution flowed was very small. The reason for this is probably that the polymer was smeared out over the outlet during the milling procedure. Even in the machine-milled cases, this effect was significant and will require further studies. Despite these problems, the tips were clearly functioning well as electrospray spray devices. Inspection with optical microscopy on some hand-polished tips in PC indicates that the hand polishing resulted in shrinkage of the channel near the outlet at the tip (this means that the applied force during the milling may also influence the size of the outlets). The connection between the inlet infusion capillary and the microchip had in the best case a 10-nL dead volume, which 3938

Analytical Chemistry, Vol. 75, No. 15, August 1, 2003

depends on how the capillary was glued into the microchip. Since these experiments were done by direct infusion, the dead volume was not our focus and was therefore not optimized in this work. In the fabrication process, the bonding procedure was the yieldlimiting step. The bonding yield was ∼20% for both plastics. It may be increased significantly by oxygen plasma pretreatment and by using a pressurized heater instead of a hot plate for the thermal bonding. The polishing and milling of the tips never resulted in a stopping of flow through the microchannels. The tips could sometimes require repolishing if the channel failed to be centered in the tip. All centered polished 3D tips were shown to be functional as electrospray devices, although the goldsputtered tips failed after a short time due to flaking of the gold film. The process time was 15 s for injection molding of a plastic disk, followed by a few minutes for dicing and bonding, less than 1 h for hand polishing, 20 min for machine milling, and a few

minutes for gluing the capillary (curing time excluded). Therefore, our fabrication method is highly suitable for small-scale production of ESI tips. Spray Performance. 3D tips, hand polished and machine milled in both PC and PMMA were evaluated in ESI experiments. With these tips, an electric field sufficient for electrospray could be obtained with an applied electrospray potential in the range of 2800 -3300 V. Stable electrospray was obtained with all 3D tips after they were coated with fairy dust gold or graphite particles. No significant difference in spray performance could be seen between PC chips and PMMA chips. The relative standard deviation of the ion current for m/z between 510 and 600 for runs lasting 5 min was in the range of 3.5-12%. Since the most timeconsuming part of the microchip construction was the milling of the tip, a legitimate question was whether electrospray chips could be fabricated with less advanced structures at the channel exit. However, since this work was done in sheathless electrospray mode where the high voltage was applied on a conducting layer on the tip, a chip without any defined tip was omitted due to the large droplet spreading on the conducting coating along the flat edge of the microchip. 2D tips were therefore tested since these can more easily be fabricated simultaneously with the channels during the replication, compared with the corresponding 3D structures. However, due to the requirement of an unreasonably high electrospray potential (more than 5000 V), stabilization of the spray by further increasing the applied electrospray potential was not possible and only resulted in discharge from the corners of the emitter tip. Previous work5 has shown the possibility of coating the tip end with hydrophobic polymers in order to limit the spreading of droplets from the microchannel. This approach would be possible to apply to the 2D tips, but the durability of the hydrophobic coating on the tip must then be further evaluated. Durability of the Conductive Coating. One challenge in sheathless electrospray is to find a conductive coating that is stable enough to endure the harsh conditions present in the electrospray process.24 Deposition of gold by sputtering was the most straightforward way of applying a conductive layer. This gold film was found to form a smooth and even surface. Stable electrospray was obtained, but after less than 30 s, the signal intensity dropped dramatically resulting in discharges and finally a discontinued spray. Polymer-embedded gold (fairy dust) and graphite particles have shown to give sustainable coatings for ESI-MS tips.25-27 When these coatings were applied onto microchip tips, a stable spray was obtained for more than 10 consecutive 5-min runs without any signs of decrease in signal intensity and sensitivity. The observations from the electrospray were modeled with electrochemical experiments for sputtered gold- and polymerembedded graphite to evaluate the electrochemical processes of the electrospray interface and to better understand what determines the stability of the coatings.36 The potentials that were applied in the CA experiments were chosen, so that oxidation or reduction of water occurred. In the CA experiments on the sputtered gold coatings, the anodic current decreased to a level of less than 20% of the initial current and the cathodic current decreased to ∼25% in less than 30 s. The behavior was the same for both PC and PMMA. In the (36) Soper, S. A.; Henry, A. C.; Vaidya, B.; Galloway, M.; Wabuyele, M.; McCarley, R. L. Anal. Chim. Acta 2002, 470, 87-99.

Figure 5. Mass spectra of peptide standards obtained under different conditions. (a) Single spectrum obtained during 160 ms of a 1 µM peptide standard corresponding to 2.6 fmol of each peptide, (b) Expansion of Figure 3a between m/z 510 and 600. (c) A 10 nM peptide sample spectrum obtained during 2.71 s corresponding to 0.5 fmol of each peptide. A, angiotensin II (MW 1046.2); B, [Arg8]vasopressin (MW 1083.4); C, leucine-enkephalin (MW 555.3); D, methionine-enkephalin (MW 573.2); E, luteinizing hormone releasing hormone, LHRH (MW 1181.6).

case of the graphite coating, the anodic current stabilized at a level of 80% of the initial current and the cathodic at a corresponding 95% level, which is compatible with the results from the ESI experiments with our plastic microchips. The samples with fairy dust coatings could not be evaluated electrochemically since their resistance was too high (>20 MΩ) for the simulated conditions. The fact that the gold flakes off very quickly when the samples act as both anodes and cathodes indicates that the adhesion between the gold and the polymer is poor and that the gold is flaking off due to gas bubbles forming during the electrospray process.36 A detailed inspection of the regions where the gold has flaked off also points to bubble formation as a reason the gold film is flaking off. Figure 4 shows a SEM picture of a PMMA Analytical Chemistry, Vol. 75, No. 15, August 1, 2003

3939

surface with sputtered gold, after reduction at -2.5 V for 600 s. The areas where the gold has flaked off and the naked polymer surface is visible (the darker areas in Figure 4) are sometimes bubble-shaped. Three such regions are demonstrated with arrows in Figure 4. Limit of Detection (LOD). The major advantage of sheathless electrospray compared to liquid junction/sheath flow mode is the ability to obtain data with high sensitivity and with a minimal loss of sample. Stable electrospray were obtained in the flow rate range of 0.7-2 µL/min without any decrease in performance. Lower flow rates resulted in unstable spray while flow rates higher than 2 µL/min resulted in high noise levels. Panels a and b of Figure 5 show a spectrum for 1 µM peptide standard dissolved in 50:50 methanol/water (v/v) + 0.1% acetic acid collected during 160 ms with a 3D hand-polished PMMA tip coated with fairy dust gold. The flow rate in this experiment was 1 µL/min, and the spectrum itself represented 2.6 fmol of each neuropeptide. The peak at m/z 413 is a phthalate, which probably originates from plastisicers present in the laboratory, since this peak occurs even when a fused-silica capillary was used as the ESI emitter. Since 1 µM is a relatively high concentration of peptide, seldom present in biological samples, spectra were also collected for a 10 nM peptide sample solution. Figure 5c shows such a spectrum collected during 2.72 s with a flow rate of 1 µL/min. The spectrum, which represents 0.5 fmol of peptide standard, denotes the LOD, three times signal-to-noise ratio, for [Arg8]-vasopressin, leucine-enkephalin, and methionine-enkephalin. It should be noted that these experiments were done by direct infusion, which means that a signal suppression effect may be present that gives lower sensitivities than if only one compound is analyzed at the time.24 If a separation step, e.g., CE, is introduced on the chip before the electrospray, an increased sensitivity would be expected. This study shows the possibility to achieve stable and durable sheathless electrospray from injection-molded PMMA and PC microchips with a low detection limit. Both these plastics have also shown promising properties for chemical modifications that would further extend their usefulness in many real applications.36 These features together with the ability to replicate disposable plastic chips are highly attractive for the chemical analysis of

3940

Analytical Chemistry, Vol. 75, No. 15, August 1, 2003

complex biological samples. The main objective for future work on on-chip ESI is to improve the replication procedure of the tip and also to incorporate a sample cleanup and separation step into the chip before the electrospray tip. CONCLUSIONS On-chip sheathless electrospray have been demonstrated with microchips made in PC and PMMA. Stable spray in sheathless mode was achieved for both hand-polished and machine-milled 3D tips of both materials with fairy dust or graphite particles as the conducting layers. The stability of the coatings was confirmed by electrochemical investigations. No difference was observed in electrospray performance between the PC and PMMA chips. The fabrication methods used in this work are suitable for the preparation of tips for on-chip applications upstream in the microchannel, such as separation and sample pretreatment. The low detection limits make the emitter tips highly attractive for the analysis of low-abundant species in complex biological samples. Although the plastic on-chip tips reported on in this paper will produce somewhat uncontrolled chip tip channel outlet sizes, they gave a surprisingly good performance and strong encouragement to search for a repeatable and controlled method. ACKNOWLEDGMENT The authors thank the following people: Carola Strandman for support in the clean room, Lars Lundbladh for running the injection molding machine, Leif Ha¨ggbom for help with the milling machine, Anna-Lisa Tiensuu and Ove O ¨ hman for support with the electroplating. A° mic AB is greatly acknowledged for the kind support of the special equipment such as the electroplating and injection molding machine. Financial support from the Swedish Foundation for Strategic Research (SSF) and Swedish Research Council (13123, 621-2002-5261, 5104-706 and K-AA/KV 09368-320) is greatly acknowledged.

Received for review January 29, 2003. Accepted May 12, 2003. AC030045T