Electrospray Device for Coupling Microscale Separations and Other

Jan 3, 2001 - Shown in Figure 1 is a low dead volume microsprayer system primarily .... or a well was accomplished with a Watec (Las Vegas, NV) model ...
0 downloads 0 Views 107KB Size
Download PDF
Anal. Chem. 2001, 73, 632-638

Electrospray Device for Coupling Microscale Separations and Other Miniaturized Devices with Electrospray Mass Spectrometry Timothy Wachs and Jack Henion*

Analytical Toxicology, Department of Population Medicine and Diagnostic Sciences, New York State College of Veterinary Medicine, Cornell University, 927 Warren Drive, Ithaca, New York 14850

A miniaturized ion sprayer device is described which is suitable for coupling with chip-based analytical separation devices, multiwell plates, or surfaces containing residues of prepared samples. Two versions of a similar device are described. A “microsprayer” device suitable for coupling to the terminal edge of a capillary electrophoresis (CE) chip is constructed from modified 1/16-in. HPLC fittings. This microsprayer employs a free-standing liquid junction formed via continuous delivery of a flow (2-6 µL/min) of suitable solvent which carries the CE effluent through a pneumatically assisted electrospray (ion spray) needle positioned in front of an atmospheric pressure ionization (API) mass spectrometer. A related but larger “minisprayer” device is also described which employs the same features as the microsprayer, but with an extended sampling capillary tube which can reach into the depths of 96-, 384-, and 1536-multiwell plates containing either sample solutions or dried sample residues. The minisprayer may be positioned in front of an API ion sampling orifice and the multiwell plate positioned stepwise from sample to sample for analysis of trace samples contained in the wells. The resulting infusion-ion spray mass spectrometric analyses can provide sequential analysis of previously prepared biological samples containing small drug compounds, proteins, and related compounds. This same device is also shown to be useful for sampling from a surface containing trace level compounds of biological interest. Results are shown that demonstrate microscale separations and selected ion monitoring (SIM) capillary electrophoresis/mass spectrometry (CE/MS) detection of berberine and palmatine using the microsprayer. SIM ion spray determination of a 2 ng/µL solution of berberine contained as a dry residue in the bottom of a 384-well plate as well as full-scan electrospray mass spectra for lowpicomole levels of cytochrome c contained in a 1536-well microtiter plate are shown. The respective micro- and minisprayer devices provide a simple yet effective means of transferring trace-level samples either from a microscale or chip-based separation device as well as samples contained in multiwell plates which are increasingly employed in high-throughput applications in the pharmaceutical industry. 632 Analytical Chemistry, Vol. 73, No. 3, February 1, 2001

Miniaturized devices promise to offer exciting opportunities for analytical chemistry in the future.1 The application of chipbased analytical devices for zone electrophoresis separation of analytes with conventional spectroscopic detectors was first demonstrated by Harrison and coauthors.2-4 In addition, ongoing developments have further demonstrated impressive analytical applications involving capillary electrophoresis performed on a chip-based substrate.5-13 Some recent reports have demonstrated capillary electrophoresis (CE) separations within time frames of seconds14 and even milliseconds.15 Each of these reports has employed various spectroscopic detectors and used direct detection of the analytes on the chip. Atmospheric pressure ionization (API) mass spectrometry (MS) techniques could provide an important, alternative detection system for chip-based devices. Consequently, efforts to couple this technology with chip-based sample handling systems has attracted considerable recent interest. Developments in miniaturized chipbased devices for analytical applications that have appeared over the past decade suggest these devices coupled to API MS systems could provide some useful analytical advances.16,17 * Corresponding author: (e-mail) [email protected]. (1) Marshall, S. R&D Mag. 1999, 18-27. (2) Harrison, D. J.; Manz, A.; Fan, Z.; Ludi, H.; Widmer, H. M. Anal. Chem. 1992, 64, 1926-1932. (3) Harrison, D. J.; Fluri, K.; Seiler, K.; Fan, Z.; Effenhauser, C. S.; Manz, A. Science 1993, 261, 895-897. (4) Harrison, D. J.; Fan, Z.; Seiler, K.; Manz, A.; Widmer, H. M. Anal. Chim. Acta 1993, 283, 361-366. (5) Manz, A.; Fettinger, J. C.; Verpoorte, E. M. J.; Ludi, H.; Widmer, H. M. Harrison, D. J. Trends Anal. Chem. 1991, 10, 144-149. (6) Manz, A.; Harrison, D. J.; Verpoorte, E. M. J.; Fettinger, J. C.; Ludi, H.; Widmer, H. M. J. Chromatogr. 1991, 593, 253-258. (7) Seiler, K.; Harrison, D. J.; Manz, A. Anal. Chem. 1993, 65, 1481-1488. (8) Fan, Z.; Harrison, D. J. Anal. Chem. 1994, 66, 177-184. (9) Colyer, C.; Mangru, S. D.; Harrison, D. J. J. Chromatogr., A 1997, 781, 271-276. (10) Jacobson, S. C.; Ramsey, J. M. Electrophoresis 1995, 16, 481-486. (11) Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1997, 69, 3212-3217. (12) Kutter, J. P.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1997, 69, 51655171. (13) Hutt, L. D.; Glavin, D. P.; Bada, J. L.; Mathies, R. A. Anal. Chem. 1999, 71, 4000-4006. (14) Bings, N. H.; Wang, C.; Skinner, C. D.; Coler, C. L.; Thibault, P.; Harrison, D. J. Anal. Chem. 1999, 71, 3292-3296. (15) Jacobson, S. C.; Culbertson, T.; Daler, J. D.; Ramsey, M. J. Anal. Chem. 1998, 70, 3476-3480. (16) Henry, C. Anal. Chem. 1997, 69, 359A-361A. (17) Bruins, A. P.; Covey, T. R.; Henion, J. D. Anal. Chem. 1987, 59, 26422646. 10.1021/ac000935y CCC: $20.00

© 2001 American Chemical Society Published on Web 01/03/2001

Attempts to couple mass spectrometry with a chip-based fluid channel have to-date required that the latter deliver the analyte via an electrospray plume directed to the ion sampling orifice of an API mass spectrometer. An early report was provided by Karger and co-workers18 and was followed by subsequent reports by these same researchers19 and others.14,20-24 The strategy for producing a microelectrospray plume from the chip has evolved from the Xue et al. direct spraying from the flat edge of the chip18,19 to improved means of producing an electrospray plume. One of these was an integrated miniaturized pneumatic nebulizer coupled via a subatmospheric liquid junction electrospray interface,25 while another was a pulled glass capillary centered in a carefully drilled flat-bottomed hole centered with the exit of a microfabricated CE channel in a glass substrate.14 Additional reports have included a sleeve to support the sprayer capillary22 on the edge of the chip as well as disposable emitters for CE/MS.26 Other developments have included gluing a pulled capillary sprayer on the flat, larger surface of the chip24 aligned with the CE channel. An alternative approach that may be applicable to chip-based separations has also been reported that employs a microfabricated monolithic nozzle surrounded by an annular cavity on the surface of a silicon substrate.27,28 In this device, the fluid is delivered from the backside of the silicon chip substrate using fluid flows of a few hundred nanoliters per minute via a through-chip channel that terminates in a nanoelectrospray nozzle. Another alternative for spraying from a chip was recently reported.29,30 This microsprayer device is described in detail in this report and in itself does not embody microfabrication. However, it is a miniaturized ion spray device17 that incorporates a makeup flow of fluid that produces a micro liquid junction31,32 between the sprayer and substrates or miniaturized sample holders where fluid transfer and analysis by mass spectrometry is desired. Herein we describe the application of this coupled miniaturized ion sprayer device for analyzing a variety of samples including small drug molecules and a representative protein. (18) Xue, A.; Dunayevskiy, M.; Foret, F.; Karger, B. L. Rapid Commun. Mass Spectrom. 1997, 11, 1253-1256. (19) Xue, Q.; Foret, F.; Dunayevskiy, Y. M.; Zavracky, P. M.; McGruer, N. E.; Karger, B. L. Anal. Chem. 1997, 69, 426-430. (20) Li, J.; Thibault, P.; Bings, N. H.; Skinner, C. D.; Wang, C.; Colyer, C.; Harrison, D. J. Anal. Chem. 1999, 71, 3036-3045. (21) Figeys, D.; Ning, Y.; Aebersold, R. Anal. Chem. 1997, 69, 3153-3160. (22) Figeys, D.; Gygi, S. P.; McKinnon, G.; Aebersold, R. Anal. Chem. 1998, 70, 3728-3734. (23) Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 1997, 69, 1174-1178. (24) Lazar, I. M.; Ramsey, R. S.; Sundberg, S.; Ramsey, J. M. Anal. Chem. 1999, 71, 3627-3631. (25) Zhang, B.; Liu, H.; Karger, B. L.; Foret, F. Anal. Chem. 1999, 71, 32583264. (26) Bateman, K. P.; White, R. L.; Thibault, P. Rapid Commun. Mass Spectrom. 1997, 11, 307-315. (27) Schultz, G.; Corso, T.; Prosser, S. In Proceedings of the 47th ASMS Conference on Mass Spectrometry and Allied Topics; American Society for Mass Spectrometry: Orlando, FL, 1999. (28) Schultz, G.; Corso, T.; Prosser, S.; Zhang, S. Anal. Chem. 2000, 72, 40584063. (29) Wachs, T.; Henion, J. D. In Proceedings of the 47th ASMS Conference on Mass Spectrometry and Allied Topics; American Society for Mass Spectrometry: Orlando, FL, 1999. (30) Wachs, T.; Deng, Y.; Henion, J. D. In Proceedings of the 48th ASMS Conference on Mass Spectrometry and Allied Topics; American Society for Mass Spectrometry: Long Beach, CA, 2000. (31) Lee, E. D.; Muck, W.; Covey, T. R.; Henion, J. D. Biomed. Environ. Mass Spectrom. 1989, 18, 844-850. (32) Lee, E. D.; Covey, T. R.; Henion, J. D. In Cornell Research Foundation; Cornell Research Foundation: Ithaca, NY, 1991.

Figure 1. Micro ion spray source designed for interfacing to chipbased separation devices incorporating a self-contained liquid junction. Key: (A) 1 cm × 31 gauge ss tubing, (B) 0.020 in. i.d. × 0.0625 in. o.d. ss tubing, (C) 0.3 cm × 23 gauge ss tubing, (D) makeup liquid entrance, (E) epoxy seal, (F) 0.0625-in. Teflon tubing, (G) 0.4 cm × 23 gauge ss tubing, (H) hole, and (I) gas inlet.

These applications include on-line CE/MS analyses employing short, fused-silica capillaries as well as representative microsample holders including 384- and 1536-well plates. A successful outcome from this effort could provide electrospray mass spectrometric detection for each of these application areas and thus the corresponding high sensitivity, selectivity, and accuracy often required in today’s challenging research environment. EXPERIMENTAL SECTION Apparatus. Two different pneumatically assisted electrosprayers were constructed. Shown in Figure 1 is a low dead volume microsprayer system primarily intended for coupling to chip-based separation systems. The metal sprayer tube (A) is a 1-cm length of 31-gauge stainless steel hypodermic tubing (Small Parts, Miami Lakes, FL). The body (B) of the device is the makeup liquid supply tube, which is a length of 0.020 in. i.d. × 0.0625 in. o.d. stainless steel tubing center-drilled with a small hole perpendicular to its length to accommodate the metal sprayer tube. The hole was enlarged for part of its depth to accommodate a 0.3-cm length of 23-gauge stainless steel tubing (C) which is the liquid delivery tube to the sprayer tube. This tube is held in place with Varian (Lexington, MA) TorrSeal epoxy resin. The metal sprayer tube is inserted into the liquid makeup tube, and the inlet end of the sprayer tube is allowed to protrude approximately 0.1-0.2 mm beyond the liquid makeup tube (see detail in Figure 1). It is held in place with epoxy on the other side of the liquid supply tube. The makeup liquid (D) is supplied to the main body tubing by way of a length of fused-silica tubing inserted into the end of the main body tubing and sealed with epoxy (not shown in Figure 1). The other end of the main body tubing is also sealed with epoxy (E). Makeup liquid flow to the device was via a Hamilton Gas-Tight syringe (Reno, NV) coupled with the appropriate Upchurch Scientific (Oak Harbor, WA) fittings and Harvard Apparatus (Holliston, MA) model 22 infusion pump. Pneumatic nebulization was incorporated by adding a gas supply line (F) of 0.0625-in. Teflon tubing (Upchurch Scientific) Analytical Chemistry, Vol. 73, No. 3, February 1, 2001

633

Figure 2. Mini ion spray source for sampling liquids and solids from wells and surfaces. Key: (A) 1/16-in. stainless steel tee, (B) 31-gauge ss tubing, (C) 0.0625-in.-o.d. PEEK tubing, (D) modified nut (see text), (E) 23-gauge ss tubing, (F) 0.0625-in.-o.d. Teflon tubing, (G) shortened nut, (H, I) see text for details, (J) 1/8-in.-diameter Teflon plug, (K) 23-gauge ss tubing, (L) gas inlet, and (M) liquid inlet.

and a concentric nebulizing tube (G) to the sprayer. The nebulizing tube is a 0.4-cm length of 23-gauge stainless steel tubing that had a small hole (H) ground through the wall of the tubing. The Teflon gas supply tubing was pierced with a sewing needle near the end of the tubing and perpendicular to its length (F), the nebulizing tube was slipped through the holes, the nebulizing tube slipped over the sprayer tube, and the whole assembly was attached to the main supply tube and sealed with epoxy resin. A supply of nitrogen gas controlled by a pressure regulator provided nebulizing gas (I) flow to the sprayer. Electrical connection to the sprayer was via the metal supply tube (A). The sprayer was fastened with electrically insulating Teflon hardware to an xyz model K33-485 positioner (Edmund Scientific, Barrington, NJ). During use, a stage was placed behind the xyz positioner holding the sprayer such that the terminus of the capillary electrophoresis separation channel could be aligned with the inlet end of the sprayer and at which point a liquid junction could be formed. The combination sprayer and separation device was then placed in front of the ion sampling orifice of a PE-Sciex API-III triple quadrupole mass spectrometer (Concord, ON, Canada). High voltage was supplied to the sprayer from a Bertan (Hicksville, NY) series 230, 5-kV high-voltage supply. The high-voltage supply was “decoupled” from the other voltages used on the chip for separation or injection by including a 25-MΩ resistor from the sprayer supply to ground. Figure 2 shows the construction of a related but larger “minisprayer” used for sampling solutions from multiwell plates, reconstituting and recovering samples in the wells of multiwell plates, or sampling materials on surfaces. The body of this larger minisprayer is a 1/16-in. stainless steel (ss) tee (A) (Upchurch Scientific) equipped with requisite fittings and ferrules. The metal sprayer tube (B) is a length of 31-gauge stainless steel tubing projecting through the Tee and held with a short piece of 0.0625-in.-o.d. PEEK tubing (C) of the appropriate inside diameter, a PEEK ferrule, and a specially modified nut (D) (see below). At 634 Analytical Chemistry, Vol. 73, No. 3, February 1, 2001

both the inlet end and the sprayer end of sprayer tube (B) were added concentric stainless steel capillaries. These provided a continuous flow of makeup liquid for the inlet end as well as pneumatic nebulization of the electrospray process at the outlet end. The inlet end employed a makeup liquid tube constructed of a length of 23-gauge stainless steel tubing (E) which was slipped over the sprayer tubing and held in the tee with a short piece of 0.0625-in.-o.d. Teflon tubing (F) and a PEEK ferrule. To minimize the overall length of the sprayer, the nut (G) used to swage this ferrule was shortened as much as possible while still retaining enough thread to allow tightening the ferrule. The 23-gauge makeup liquid tube (E) is positioned such that the sample inlet sprayer tube protrudes ∼0.1 mm when used for sampling from solid surfaces. While sampling solutions, the makeup tube is positioned such that the sprayer tube is flush or slightly inside the makeup tube (see details in Figure 2 (H, I)). To provide pneumatic nebulization and aspiration while maintaining a short path length, the pneumatic nebulizer portion of the sprayer was incorporated into the same nut assembly that holds the sprayer tubing in the tee. The head of a long-reach nut (D) was drilled to an inside diameter of 1/8 in. for approximately three-fourths of the head length (see Figure 2 (D) for details.) A 1/ -in.-diameter Teflon plug (J) equipped with a central hole to 8 accommodate the 23-gauge nebulizer tube (K) was press-fitted into the 1/8-in. drilled hole. The nebulizing tube (K) can be moved along the length of the sprayer tube to optimize the nebulization and aspiration process. A gas inlet (L) tube was epoxy-sealed into the side of the nut. The makeup liquid was pumped through the base of the tee using appropriate fittings and tubing (M) and the same infusion pump used with the microsprayer described above. Electrical contact was made to the metal tee. The whole assembly was mounted on an electrically insulating arm that can be rigidly positioned in front of a movable xyx stage. The stage and sprayer

Figure 3. (a) “Mock” CE chip and details of self-contained liquid junction. Key: (A) microsprayer, (B) 6 cm × 75 µm i.d. × 375 µm o.d. fused silica, (C) 1.5 mL polyethylene centrifuge tube, (D) ss wire electrode, (E) hole, and (F) detail of liquid junction. (b) Liquid junction used to sample from surface of a well.

were positioned in front of the ion sampling orifice of a PE-Sciex API-IIIplus mass spectrometer. A “mock” capillary electrophoresis chip, Figure 3a, with microsprayer (A) was constructed from a 6-cm length of 75 µm i.d. × 375 µm o.d. fused-silica tubing (B) attached to a glass microscope slide to demonstrate coupling the microsprayer to a pseudochip-based separation device. Buffer and sample reservoirs were constructed from 1.5-mL polyethylene centrifuge tubes (C) that were held with a separate clamp. A hole was drilled in the side of the centrifuge tube (E) below the liquid level and slightly larger in diameter than the fused-silica tubing (B). The surface tension of the buffer was high enough to keep the buffer from leaking out of the hole in the reservoir. High voltage was applied for the CE separation via an electrode (D). The ion spray voltage was applied to the sprayer as noted above and was decoupled from the CE voltage via the liquid junction31,32 and the resistor to ground. Multiwell plates used for sampling from wells with the above minisprayer were either 384- or 1536-well polystyrene plates (Greiner America, Lake Mary, FL). The 384-well plates were black, and the 1536-well plates were clear polystyrene. The clear plastic plates facilitated the positioning of the inlet of the sprayer during development stages. The plates containing samples of interest were mounted vertically on an xyz positioner to allow the inlet end of the minisprayer (Figure 3b) to reach into and near the bottom of a selected well. The current setup precludes horizontal mounting of the plates, although this could be accomplished with some instruments. Observation of the sprayer position with respect to either a surface or a well was accomplished with a Watec (Las Vegas, NV) model 502A CCD camera. Reagents. Makeup liquid used for forming the liquid junction or for sampling from a multiwell plate was 80% methanol (Fisher Scientific, Pittsburgh, PA)/20% water prepared in-house with a

Nanopure water system (Barnstead, Dubuque, IA). Depending upon the analyte chemistry, the bulk makeup liquid contained either 0.1% formic acid (GFS Chemicals, Columbus, OH) or 3 mM ammonium acetate (Fisher Scientific.) This solution was infused into the makeup tube at a rate of 2-6 µL/min depending upon the experiment. The makeup flow, sprayer voltage, nebulizing gas flow, liquid junction gap, and makeup liquid composition are all interacting parameters affecting operation of the sprayer. Operation of the sprayer was optimized for both a stable junction and a stable spray for a given makeup liquid composition. Once determined, these settings were reproducible for a given mode of operation. A 1 ng/µL solution of both berberine and palmatine (Sigma Chemical, St. Louis, MO) in water prepared by dilution from a 1000 ng/µL stock solution in methanol was used for the CE separation. A 2 ng/µL solution of berberine in methanol was prepared by dilution from the 1000 ng/µL stock solution in methanol. Ten microliters of this solution were evaporated in a well to provide a dry residue to demonstrate the ability of the sprayer to recover a dried sample from a surface. A stock solution of cytochrome c was prepared by dissolving of bovine heart cytochrome c (Sigma Chemical) in 1:1 methanol/ water-0.1% formic acid. This solution was then diluted to a concentration of 30 ng/µL (2 pmol/µL) in 100% methanol-0.1% formic acid or 15 ng/µL (1 pmol/µL) in 1:1 methanol/water0.1% formic acid or 100% water-0.1% formic acid. For sampling a liquid from a multiwell plate, 10 µL of the 1:1 methanol/water solution or 100% aqueous solution was transferred to a well in a 1536-well plate. For the reconstitution experiment, 10 µL of the 100% methanolic solution of cytochrome c was transferred to a well on a 1536-well plate and the solvent allowed to evaporate, leaving a dry residue of the sample on the surface of the well. Procedure. For experiments requiring a low dead volume, such as coupling to a nanoseparation chip, the microsprayer device was employed (Figure 1). For static experiments, such as sampling from wells, the larger minisprayer was used (Figure 2). The mass spectral data were collected in either the single MS full-scan mode or by selected ion monitoring (SIM). Unit mass resolution (peak widths of 0.5-0.7 Da at half-height) was maintained in all experiments. The sprayer voltage was maintained at 5 kV for all experiments reported here. Makeup flow and nebulizer gas flow rates were optimized for the specific experiment that was performed. The position of the sprayer tube relative to the makeup liquid tube could be varied to optimize the performance of the larger minisprayer. The inlet of the sprayer was aligned with the separation channel of a “mock chip” or other chip device and makeup liquid flow initiated. High voltage was applied to the sprayer while nitrogen gas was introduced to the nebulizing tube. Once a spray was established the voltage, gas pressure, and makeup flow were adjusted for a stable spray while a narrow liquid bridge was maintained between the sprayer and the chip edge (Figure 3a (F)). Both high ion current signal intensity and low baseline noise were factors considered in the optimization process. “Mock Chip” CE/MS. A 6-cm length of 75-µm-i.d. fused-silica tubing was used to simulate coupling of the sprayer to an actual chip (Figure 3a). The capillary was filled with the optimized CE running buffer by sliding the capillary through the drilled hole in the side of a centrifuge tube (E) and below the surface of the buffer, which allowed capillary action to fill the capillary. The Analytical Chemistry, Vol. 73, No. 3, February 1, 2001

635

Figure 4. CE/MS with a “mock chip” (5-cm fused-silica capillary) coupled via liquid junction microsprayer.

sprayer was then aligned with the fused-silica tubing and positioned within ∼50 µm of the end of the tubing (F). The position was monitored with a video camera in the x and z directions. The vertical position, y, was set by visual inspection. Sample loading was accomplished by interchanging the buffer reservoir with a sample reservoir. A metal electrode in the solution (D) provided the electrical contact for electrokinetic injection. The sprayer (A) coupled via the liquid junction provides the counter electrode during electrokinetic injection. A difference of 2000 V between the sample reservoir and the sprayer for 10 s provided a sufficient quantity of sample loaded into the inlet of the CE capillary by electrokinetic injection. After the applied voltage was removed, the sample reservoir was interchanged with the buffer reservoir equipped with a metal electrode. Electrophoretic separation (Figure 4) was achieved by supplying a higher voltage than the sprayer voltage to the electrode. A separate infusion experiment at 200 nL/min of a 1 ng/µL solution of reserpine in methanol was performed with PEEK tubing coupled with the sprayer to demonstrate the feasibility of employing the sprayer with substrates other than fused silica or glass (data not shown here.) Sampling from Surfaces and Solutions. Three different modes of operation were used for the minisprayer for sampling from surfaces and solutions. The first mode involved sampling of an analyte from a surface by allowing a liquid junction to form between the sprayer inlet and the surface that held the analyte (Figure 3b.) This was demonstrated by evaporating to dryness 10 µL of a 2 ng/µL methanolic solution of berberine in a well of a 384-well plate and then inserting the sprayer into the well to form a liquid junction between the surface of the well and the sprayer (Figure 5.) For a demonstration of sampling from solution, the sprayer tube is positioned just inside the makeup tube (Figure 2 (I)). With the sprayer operating, the inlet is inserted into a solution held in a well. When the analyte solution is significantly more aqueous than the makeup liquid, there will often be an initial dilution of the solution after which the solution will be picked up by the sprayer. The makeup flow rate will effect how rapidly the solution is picked up by the inlet. To demonstrate the feasibility of this technique, solutions of cytochrome c in both 1:1 MeOH/water0.1% HCOOH and in 100% water-0.1% HCOOH were employed 636 Analytical Chemistry, Vol. 73, No. 3, February 1, 2001

Figure 5. Twenty nanograms of berberine sampled from the bottom of a well in a 384-well plate. SIM of m/z 336.2

Figure 6. Summed spectra of cytochrome c. Makeup liquid 0.1% HCOOH in 80:20 MeOH/water: (A) 10 pmol in solution of 1:1 MeOH/ H2O-0.1% HCOOH; (B) 10 pmol in solution of water-0.1% HCOOH; (C) 20 pmol deposited in well and then reconstituted with sprayer and makeup liquid.

at a concentration of 1 pmol/µL (Figure 6A,B). Approximately 10 µL of these solutions was sampled from the wells of a 1536-well plate. Makeup liquid was 80:20 MeOH/water-0.1% HCOOH. Another mode of operation involves reconstituting a dried sample of 20 pmol of cytochrome c from a well in a 1536-well plate (Figure 6C). In this mode, the sprayer is inserted into the well and then nebulizing gas and sprayer voltage are turned off. The makeup flow is allowed to partially fill the well, at which time the nebulizing gas and sprayer voltage are reapplied whereupon spraying commences. In either the solution sample mode or the reconstitution mode, the makeup liquid flow is adjusted to be slightly less than optimum for spray stability. This enables the sprayer to aspirate a portion of the bulk liquid while mixing it

with the makeup liquid that is flowing into the sprayer. Additionally, the sprayer tube is kept just inside of the makeup tube rather than outside when a liquid junction formation is desired. Safety Precautions and Considerations. Care must be taken when working with prototype systems such as those described above. Electrical power supplies should be current-limiting to the extent that accidental contact with an exposed portion of a highvoltage circuit cannot supply lethal currents. If this is not possible, then appropriate shields and/or interlocks should be used. Since potentially toxic vapors are produced during the electrospray process, appropriate ventilation should be provided. Normal care should be exercised in handling any potentially hazardous or flammable chemicals. RESULTS AND DISCUSSION The goal of coupling an API mass spectrometer to microseparation devices can be accomplished in several different ways. Ultimately, whatever device or approach is selected, it should not cause significant deterioration of the separation or alteration of the target analytes. Even though the addition of a makeup liquid can have a dilution effect, normal CE/MS is most often performed with either a sheath flow or a liquid junction interface.31-34 Because of the microscale of most chip-based separation systems, dead volume in the interface must be minimized. Although drawn fused-silica capillary sprayers may provide higher sensitivity than the described sprayer, the design described herein provides other advantages. Since the sprayer is a straight piece of metal tubing, it is less prone to becoming plugged by particles since there is no change in the inside diameter along its length. Additionally, bubbles in the sprayer tube, which may interrupt the spray from fused-silica systems by isolating the potential from the tip of the liquid, are not a problem with the all-metal spray tube since the potential on the makeup liquid is constant along the length of the sprayer and cannot be interrupted by a bubble in the sprayer tube. These features coupled with the pneumatic nebulization provide for a stable electrospray process. The microsprayer was designed exclusively to couple a chipbased separation device to an API mass spectrometer via a freestanding liquid junction between the sprayer tube and the separation device. The free-standing liquid junction detail depicted in Figure 3a (F) is basically a dynamic, flowing “liquid bridge” between the channel exit of the separation device and the inlet of the minisprayer. The bridge is maintained by the surface tension of the liquid and is also affected by the distance between the inlet of the sprayer and the surface of the chip. Additionally, the makeup flow rate must balance the rate at which the liquid is being drawn into the sprayer by the combined effects of the electrospray process and pneumatic nebulization. The conditions for maintaining both a stable spray and a stable junction were determined empirically. Little, if any, spreading of the liquid along the surface beyond the sprayer inlet was observed for both glass and polymer surfaces. Fluidic interactions between the sprayer and the separation channel are minimized with the described system. The liquid junction’s stability depends on a balanced flow between the (33) Smith, R. D.; Barinaga, C. J.; Udseth, H. R. Anal. Chem. 1988, 60, 19481952. (34) Lee, E. D.; Muck, W.; Henion, J. D.; Covey, T. R. J Chromatogr 1988, 458, 313-21.

makeup input and the spray rate. This can be established by placing the sprayer inlet on a portion of the chip edge where a channel outlet is not present. Once proper conditions are established, the channel outlet can then be aligned with the sprayer inlet. Even though the gap is narrow, the free-standing nature of the liquid junction precludes any significant fluidic forces on the channel. The SIM CE/MS electropherogram shown in Figure 4 demonstrates the viability of the sprayer system as an interface between a microscale separation system and an API MS system. Although this experiment was performed on a “mock chip”, which consisted of a short length of fused-silica capillary used to simulate an actual chip, the dimensions and conditions are similar to those experienced when using an actual chip-based device. This SIM CE/MS determination of berberine and palmatine provided separation and detection of the respective quaternary cations of these compounds at m/z 336.2 and 352.2 (Figure 4). The total selected ion current profile for the SIM CE/MS determination of these two analytical standards is shown in Figure 4A. Although the two compounds are observed to be only partially separated in Figure 4A, their respective electrophoretic profiles are clearly evident in Figure 4B and Figure 4C where their corresponding selected ion current profiles at m/z 336.2 and 352.2 are shown. The symmetry of the electrophoretic peaks indicates low dead volume conditions within both the sprayer and the liquid junction. The liquid junction is continuously renewed as the makeup liquid is swept from the outer makeup tube into the inner sprayer tube (Figure 1 (C)). The sprayer tube has a calculated volume of ∼125 nL. At a flow rate of 4 µL/min for the makeup liquid, the transit time through the sprayer tube would be ∼1.9 s. Although it is not possible for us to determine the exact nature of the flow through the junction and into the sprayer, the actual junction volume is on the order of 1-2 nL; hence, the transit time of a target analyte through the junction would be on the order of 20 ms. It is also not possible to determine how well the analyte is dispersed in the makeup liquid once it is picked up at the exit of the separation channel. This is likely a function of the electrophoretic flow produced in the separation channel. Of course, with any system using a makeup liquid, a reverse electrophoretic flow would result in makeup liquid being drawn into the separation channel with possible deleterious effects. The analytical ruggedness of a system is crucial if a large number of samples are to be analyzed. Inspection of the baseline noise in Figure 4 indicates that the sprayer operates in a stable mode. The same sprayer was used for almost one year before it had to be rebuilt due to electrochemical etching of the sprayer tip. Once conditions were determined for stable operation, the separation device could be removed and replaced repeatedly without having to readjust anything other than the liquid junction gap. For an automated system, this would work well either by removing and replacing chips or by indexing the chip to a new separation channel on a multichannel chip. There was also minimal spreading of the makeup liquid outside the diameter of the makeup tube, so that channels could be at closely spaced intervals on a chip. Since the sprayer is essentially electrically and fluidly decoupled from the separation device, electrokinetic (or pressure) injection can be carried out while the sprayer is actually spraying. Once again this is an advantage for an automated system in which Analytical Chemistry, Vol. 73, No. 3, February 1, 2001

637

a background mass peak could be continuously monitored and used as an error monitor in case of the system failure. Another advantage of this system is the flexibility of the sprayer system. Not only can it be coupled to chips constructed of various materials, e.g., glass, plastics, etc., it is also not limited to coupling separation systems to an API MS. Since the sprayer is not part of a closed fluid system, the liquid junction can be formed with many types of surfaces. The junction can thus be used to sample a surface or sample a targeted analyte that may be on a surface. To demonstrate this, a modified minisprayer was constructed that incorporated a longer sampling tube that could reach, for example, to the bottom of a well on a multiwell plate. Since separations are not involved in the following examples, the increased dead volume of this sprayer is less important. Figure 5 shows the ion current profile for m/z 336.2 of berberine sampled from the bottom of a well on a 384-well plate. The ion current profile shown in Figure 5 results from the reconstitution and uptake of the berberine dry residue. Thus, the liquid junction formed between the inlet of the minisprayer (Figure 3b) and the dry residue of sample effectively redissolves or “reconstitutes” the sample and transfers it through the minisprayer for ion spray mass spectrometric detection. Figure 5 shows that the sample is introduced to the mass spectrometer within ∼40 s, increases to a maximum ion current signal in ∼1 min and is consumed by the end of ∼2 min. The small secondary peak was the effect of moving the sprayer position along the bottom of the well. This process could be performed repeatedly if one required the sequential analysis of samples contained in their respective wells of these multiwell plates. If the described system were to be operated in a robotic, automated mode, it is possible that the high-throughput analysis of combinatorial library synthetic products could be achieved. These analyses could provide rapid determination of product integrity as well as some information on the sample purity. If the relative position of the makeup liquid tube is changed such that the sprayer inlet lies at the surface (flush) or withdrawn slightly inside the makeup liquid tube (Figure 2 (I)), the sprayer can readily combine the makeup liquid with a sample solution. If the makeup flow is adjusted such that it is less than the sprayer aspiration flow rate, then the sprayer can aspirate the surrounding sample solution in addition to the makeup liquid. This technique can be used to advantage when an aqueous sample solution is involved since aqueous solutions are often more difficult to electrospray. The rate at which the solution is actually sampled is not only a function of the makeup flow but also of the relative compositions of the makeup liquid and the sample solution. If the composition of the sample solution is significantly more aqueous than the makeup liquid, it has been observed that there is actually a decrease in the spray flow with a resulting increase in solution volume until the composition of the liquid in the vicinity of the sprayer inlet becomes more similar to the makeup liquid. Panels A and B of Figure 6 show the electrospray mass spectra obtained from solutions of cytochrome c contained in their

638

Analytical Chemistry, Vol. 73, No. 3, February 1, 2001

respective wells of a 1536-well plate. These summed spectra represent 10 pmol of the protein in 1:1 methanol/aqueous-0.1% HCOOH (Figure 6A) and also in 100% water containing 0.1% HCOOH (Figure 6B). The ion current intensity for the protein dissolved in 100% aqueous solution is less than that in the 1:1 methanol/water-0.1% HCOOH solution as might be expected, but still a reasonable signal was obtained from what was originally a 100% aqueous system. This same sprayer can be used in another mode to reconstitute dried samples (residues) such as peptides and proteins contained in wells. The makeup liquid is used first to redissolve or reconstitute the dry residue sample with the makeup, whereupon the resultant solution is transferred to the sprayer for electrospray ionization as described in the Procedure section. The electrospray mass spectral data for 20 pmol of cytochrome c in the well are presented in Figure 6C. Although there is the possibility for adsorptive losses of the protein to the walls of the well, an acceptable mass spectrum was obtained from the deposited 20 pmol of this protein. This method also reduces the potential losses associated with multiple transfers of “sticky” samples. These three data sets of electrospray mass spectra for cytochrome c reveal the same general behavior and demonstrate that the described minisprayer is capable of sampling a protein from either a solution or a dry residue contained in the wells of a 1536-well plate. With automated transitional stages and proper control systems, an entire 1536-well plate could be sampled robotically. Although no investigations were performed to determine the extent of cross contamination from material clinging to the sprayer, a separate wash step could be implemented to rinse both the insides and outsides of the tubes that contact the sample. CONCLUSIONS This work demonstrates the feasibility of coupling an easily constructed microelectrospray device to a miniaturized separation device or other fluid delivery substrate for the electrospray determination of organic compounds. The concept of the liquid junction originally reported by this laboratory31,32 was employed to provide facile coupling of miniaturized fluid delivery system for mass spectrometric analysis of samples. It is suggested that the concepts described here are also applicable to other substrates including multiwell plates, SDS-PAGE plates, TLC plates, and related surface-available sample sources. ACKNOWLEDGMENT The authors thank PE Sciex for instrumentation support and Ms. Laura Lauman for her continued generous support of our research. We also thank Dr. N. Spooner and SmithKline Beecham Pharmaceuticals for the gift of the API IIIplus tandem triple quadrupole mass spectrometer system used in this report. Received for review August 8, 2000. Accepted November 16, 2000, AC000935Y