Anal. Chem. 2002, 74, 5897-5901
An Electrospray Ionization Source for Integration with Microfluidics Jun Kameoka,† Reid Orth,† Bojan Ilic,† David Czaplewski,† Tim Wachs,‡ and H. G. Craighead*,†
School of Applied and Engineering Physics and Analytical Toxicology College of Veterinary Medicine, Cornell University, 927 Warren Drive, Ithaca, New York 14850
We have demonstrated a new electrospray ionization (ESI) device incorporating a tip made from a shaped thin film, bonded to a microfluidic channel, and interfaced to a timeof-flight mass spectrometer (TOFMS). A triangular-shaped thin polymer tip was formed by lithography and etching. A microfluidic channel, 20 µm wide and 10 µm deep, was embossed in a cyclo olefin substrate using a silicon master. The triangular tip was aligned with the channel and bonded between the channel plate and a flat plate to create a microfluidic channel with a wicking tip protruding from the end. This structure aided the formation of a stable Taylor cone at the apex of the tip, forming an electrospray ionization source. This source was tested by spraying several solutions for mass spectrometric analysis. Because the components are all made by lithographic approaches with high geometrical fidelity, an integrated array system with multiple channels can be formed with the same method and ease as a single channel. We tested a multichannel system in a multiplexed manner and showed reliable operation with no significant cross contamination between closely spaced channels. Micrototal analytical systems (µTAS) are being investigated for high-speed chemical analysis. For example, devices using capillary electrophoretic techniques for small molecule separation,1-5 DNA separation,6-9 two-dimensional peptide separation,10 and microchip immunoassay11 have been reported recently. Optical fluorescence or electrochemical detection systems in microfluidic channels were used with those devices. A growing interest in * E-mail:
[email protected]. † School of Applied and Engineering Physics. ‡ Analytical Toxicology College of Veterinary Medicine. (1) Harrison, D. J.; Manz, A.; Fan, Z.; Ludi, H.; Widmer, H. M. Anal. Chem. 1992, 64, 1908-1919. (2) Harrison, D. J.; Fluri, K.; Seiler, K.; Fan, Z.; Effenhauser, C. S.; Manz, A. Science 1993, 261, 895-897. (3) Harrison, D. J.; Fan, Z.; Seiler, K.; Manz, A.; Widmer, H. M. Anal. Chim. Acta 1993, 283, 361-366. (4) Jacobson, S. C.; Hergenroder, R.; Koutny, L. B.; Ramsey, J. M. Anal. Chem. 1994, 66, 2369-2373. (5) Duffy, C. D.; McDonald, C. J.; Schueller, O.; Whiteside, M. G. Anal. Chem. 1998, 70, 4974-4984. (6) Han, J.; Craighead, H. G. Science 2000, 288, 1026-1029. (7) Han, J.; Craighead, H. G. Anal. Chem, 2002, 74, 394-401. (8) Han, J.; Turner, S, W.; Craighead, H. G. Phys. Rev. Lett, 1999, 83, 16881691. (9) Han, J.; Craighead, H. G. J. Vac. Sci. Technol. 1999, A17, 2142-2147. (10) Rocklin, R. D.; Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 2000, 72, 52445249. (11) Koutny, L. B.; Schmalzing, D.; Taylor, T. A.; Fuchs, M. Anal. Chem. 1996, 68, 18-22. 10.1021/ac020396s CCC: $22.00 Published on Web 10/22/2002
© 2002 American Chemical Society
peptide and protein analysis demands instruments for rapid analysis of many compounds. Mass spectrometry plays a key role in such analysis, and there are many advantages in efficiency and sample consumption that can be attained by the use of microfluidics. Therefore, there is a need for systems that can effectively interface microfluidic systems with mass spectrometry. Leading sources for mass spectroscopic analysis of biomolecules include matrix-assisted laser desorption ionization (MALDI) and electrospray ionization (ESI). The MALDI technique incorporates a pulsed UV laser to desorb and ionize the sample matrix. Alternatively, the ESI technique employs strong electric fields to create ions from solution. ESI couples well to on-line liquid phase separation systems, such as liquid chromatography (LC) or capillary electrophoresis (CE). This has motivated substantial research on effective sources for ESI that can be efficiently integrated with microfluidic systems. Several groups have investigated approaches for integrating microfluidic channels with electrospray ionization sources for mass spectrometry. Some previous approaches involved connecting a tapered capillary nanoelectrospray emitter to an outlet of a microfluidic channel at an edge of a chip on substrates, such as glass and polymer. In some cases, a manually controlled drilling process was performed to make a hole at the channel exit for a tubing insertion and sealed using an adhesive.12-18 This method is not useful for fabricating highly dense electrospray array devices because of the complexity of the manual fabrication processing steps. Direct electrospray from planar microfluidic channel exits has been described in several papers.19-22 Difficulty in establishing a stable Taylor cone due to liquid spreading at the flat edge has been reported. The flat edge of a glass chip was coated by a (12) Liu, H.; Felten, C.; Xue, Q.; Zhang, B.; Jedrzejewski, P.; Karger, B. L.; Foret, F. Anal. Chem. 2000, 72, 3303-3310. (13) Lazar, I. M.; Ramsey, R. S.; Sundberg, S.; Ramsey, J. M. Anal. Chem. 1999, 71, 3627-3631. (14) Chan, J. H.; Timporman, A. T.; Qin, D.; Aebersold. R. Anal. Chem. 1999, 71, 4437-4444. (15) Bings, N. H.; Wang, C.; Skinner, C. D.; Colyer, C. L.; Thibault, P.; Harrison, D. J. Anal. Chem. 1999, 71, 3292-3296. (16) Li, J. H.; Kelly, J. F.; Chernushevivh, I.; Harrison, D. J.; Thibault, P. Anal. Chem. 2000, 72, 599-609. (17) Jiang, y.; Wang, P.; Locascio, L. E.; Lee, C. S. Anal. Chem. 2001, 73, 20482053. (18) Figeys, D.; Aebersold. R. Anal. Chem. 1998, 70, 3721-3727. (19) Lazar, I. M.; Ramsey, R. S.; Jacobson, S. C.; Foote, R. S.; Ramsey, J. M. J. Chromatogr., A 2000, 892, 195-201. (20) Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 1997, 69, 1174-1178. (21) Xue, Q.; Foret, F.; Dunayevskiy, Y. M.; Zavracky, P. M.; McGruer, N. E.; Karger, B, L. Anal. Chem. 1997, 69, 426-430. (22) Rohner, T. C.; Rossier, J. S.; Girault, H. H. Anal. Chem. 2001, 73, 53535357.
Analytical Chemistry, Vol. 74, No. 22, November 15, 2002 5897
Figure 1. Fabrication process of the single triangular tip electrosprayer integrated with the microfluidic channel. The parylene C film was sandwiched between two plastic chips and thermally bonded.
Figure 2. Scanning electron micrograph of the triangular tip taken at a 45° tilt angle. The scale bar is 20 µm. The base of this triangle is 100 µm.
hydrophobic chemical21 or a hydrophobic substrate was used to prevent liquid spreading at the channel exit.22 However, the position of the Taylor cone was not stable in these studies. This liquid-spreading phenomenon at the exit of the microfluidic channel creates difficulty not only for establishing a Taylor cone but also for making a high-density array of devices as a result of cross-channel contamination at channel exits. An external ESI device coupled to microfluidic devices via a liquid junction has been investigated.23-26 These devices can be coupled to on-line separation systems and interfaced to mass spectrometry. However, because the ESI device was not fully integrated with the microfluidic channels, there was limited advantage to this level of ESI integration. (23) Zhang, B.; Foret, F.; Karger, B, L. Anal. Chem. 2000, 72, 1015-1022. (24) Wachs, T.; Henion, J. Anal. Chem. 2001, 73, 632-638. (25) Deng, Y.; Henion, J.; Thibault, P.; Harrison, D. J. Anal. Chem. 2000, 73, 639-646. (26) Kameoka, J.; Craighead, H. G.; Zhang, H.; Henion, J. Anal. Chem. 2001, 73, 1935-1941.
5898
Analytical Chemistry, Vol. 74, No. 22, November 15, 2002
Sacrificial layer methods have been used to fabricate a microfluidic channel and a sprayer nozzle on silicon by depositing silicon nitride27 or parylene,28 using photo resist as the sacrificial layer. A monolithic sprayer nozzle fabrication on a planar surface of a silicon wafer was investigated, and it showed better stability than the tapered capillary nanoelectrospray.29 An electrospray nozzle array fabrication on a planar surface of a polymer substrate was performed using a laser ablation technique.30 It was difficult for these devices to be integrated with microfluidic networks because of the planar nozzle design. Improvements in the spraying performance from a flat channel exit were obtained by sharpening the channel exit.31-33 This improved the sprayer stability and decreased the voltage required to establish a Taylor cone. A polymer substrate such as poly(dimethylsiloxane) (PDMS) or polycarbonate (PC) was used for these investigations. In other work, a sharp tip at an exit of an open channel was manually fabricated on the surface of a polymer substrate. An electrospray ionization source interfaced to a mass spectrometer, with this device, was used for analysis of large biomolecules.34 This device required a large-scale fluidic channel to minimize the evaporation effect from the open channel. In this paper, we describe a triangle electrospray emitter tip integrated with a microfluidic device and interfaced to a mass spectrometer. The triangular tip acts like a nozzle or wick that (27) Desai, A.; Tai, Y.; Davis, M. T.; Lee, T. D. Transducers 97 2001, 2, 927930. (28) Licklider, L.; Wang, X.; Desai, A.; Tai, Y.; Davis, M. T.; Lee, T. D. Anal. Chem. 2000, 72, 367-375. (29) Schultz, G. A.; Corso, T. N.; Prosser, S. J.; Zhang, S. Anal. Chem. 2000, 72, 4058-4063. (30) Tang, K.; Lin, Y.; Matson, D. W.; Kim, T.; Smith, R. D. Anal. Chem. 2000, 73, 639-646. (31) Lin, Y.; Wen, J.; Fan, X.; Matson, D. W.; Smith, R.D. SPIE 1999, 3877, 28-35. (32) Kim, J.; Knapp, D. R. J. Am. Soc. Mass Spectrom. 2001, 12, 463-469. (33) Kim, J.; Knapp, D. R. Electrophoresis 2001, 22, 3993-3999. (34) Yuan, C.; Shiea, J. Anal. Chem. 2001, 73, 1080-1083.
Figure 3. Configuration of ESI device interfacing to mass spectrometer. The device was mounted on the X, Y, Z stage positioner and adjusted to obtain the maximum total ion current. A 600 V current was applied to the MS orifice. The chip reservoir was connected to the syringe pump via silica capillary tubing. The gold wire was located at the reservoir for electrospraying.
prevents liquid from spreading laterally at the exit of the microfluidic channel. The shape of the tip helps to form and fix the position of the Taylor cone. A liquid droplet with a critical curvature for establishing a Taylor cone is formed at the apex of the triangle. An electrospray ionization array device with four independent triangle electrospray tips was also fabricated and tested for spraying multiple compounds in a multiplexed format. The fabrication process for these devices was straightforward and lends itself to integration and mass production. EXPERIMENTAL SECTION Device Fabrication. The fabrication process for the single electrospray tip device is shown in Figure 1. A 2.5-cm-long, 20µm-wide, and 10-µm-deep microfluidic channel was embossed in a polymer chip using a silicon master followed by a drilling of a reservoir hole. The fabrication process for the ESI array device was the same as a single tip device except that it has four microfluidic channels, four reservoirs, and four electrospray tips. The length, width, and depth of the microfluidic channels for the array device were the same as the single tip device. The embossing process of the polymeric substrate and the fabrication process for the silicon master have been described in a previous paper.26 Our tip material was a 5-µm-thick film of parylene C deposited onto a silicon surface using a PDS 2010 deposition machine (Specialty Coating Systems Inc.) and pattered by optical lithography using an HTG contact aligner (Hybrid Technology Group) followed by oxygen plasma etching using a PlasmaTherm 72 (UNAXIS, Zurich, Switzerland).35 The polymeric films patterned on the silicon wafer were peeled from the silicon surface in an isopropyl alcohol solution. Films formed in this way are hydrophilic. This film was then sandwiched between two pieces of plastic and thermally bonded. A scanning electron microscope (SEM) image of the channel exit and two-dimensional triangle emitter tip at the edge of the device is shown in Figure 2. It was taken at a 45°-tilt angle. The apex angle of the triangular tip (angle a in Figure 2) is 90°. Angles b and c in Figure 2 are 45°. The base of this triangular tip is 100 µm, and the height is 50 µm. (35) Ilic, B.; Craighead, H. G. Biomed. Microdevices 2000, 2, 317-322.
Apparatus. The experimental configuration for characterization of a Taylor cone was based on a previous publication.36 The electrospray device was placed under an Olympus AX-70 microscope objective (Tokyo, Japan) with an aluminum counterelectrode placed 1.0 cm away from the polymeric tip. The aluminum counterelectrode was connected to a Keithley 486 picoammeter (Cleveland, OH) for measuring total ion current. The Harvard Apparatus syringe pump was connected to the reservoir, creating a 300 nL/min flow of buffer solution (50% methanol, 50% DI water with 0.1% formic acid). The experimental configuration of the device interfaced to a Mariner time-of-flight mass spectrometer (PerSeptive Biosystems, Inc. Framingham, MA) used for the infusion experiment is shown in Figure 3. The microchip was mounted on an X, Y, Z stage (Newport, Irvine, CA) for position adjustment between the emitter and the orifice of the mass spectrometer to maximize the total ion current. The silica capillary tubing was connected to the microfluidic channel via a pipet tip glued to the reservoir hole. A Harvard Apparatus syringe pump (Natick, MA) was used to create the desired liquid flow rate through the silica capillary tubing during the infusion experiment. There was no leakage observed under this connection. A gold wire was also glued to the reservoir hole under the pipet tip for the high voltage power supply. A 600 V potential was applied to the MS orifice, and the heated capillary of the MS was maintained at 80 °C. The syringe pump created a stable flow of 300 nL/min to supply liquid to the channel outlet. A voltage of 2.5-3.0 kV was applied between the gold wire and the MS orifice that was 8 to 12 mm away from the ESI triangular tip. MATERIALS AND CHEMICALS A cyclo-olefin plastic plate (ZEONOR 1020R) with dimensions of 2 × 100 × 150 mm was purchased from Zeon Chemicals (Louisville, KY). The cyclo olefin polymer has strong chemical (36) Kameoka, J.; Orth, R.; Ilic, B.; Czaplewski, D.; Craighead, H. G. IEEE-EMBS Society Special Topic Conference on Microtechnologies in Medicine & Biology 2002, 62-65.
Analytical Chemistry, Vol. 74, No. 22, November 15, 2002
5899
Figure 6. ESI array device with four triangular tips. The no. 2 tip is spraying buffer solution.
Figure 4. (a) Optical image of a triangular tip. The syringe pump is not activated. (b) A Taylor cone is established on the triangular tip after the syringe pump and the power supply were activated; the liquid jet is also seen.
Figure 5. (a) Mass spectrum of chicken cytochrome c. (b) Total ion current of chicken cytochrome c for 10 min. RSD is 1.3%.
resistance to alcohols, ketones, and acids.37 In this study, because methanol and formic acid were employed with the device, it was important to have high resistance to these chemicals. Parylene C dimer was purchased from Specialty Coating Systems (Indianapolis, Indiana). Chicken cytochrome c (MW, 12 222), desipramine (MW, 266), and imipramine (MW, 280) were purchased from (37) Lamonte, R.; McNally, D. Plastic Eng. 2000, 51-55.
5900 Analytical Chemistry, Vol. 74, No. 22, November 15, 2002
Sigma Chemicals (St. Louis, MO). All samples were diluted in 50%methanol, 50% DI water with 0.1% formic acid. Methanol and isopropyl alcohol were purchased from J. T. Baker (Phillisburg, NJ). The formic acid and the acetic acid were obtained from Fisher Scientific (Fair Lawn, NJ). Deionized 18 MΩ/cm water was generated in-house with a Barnstead Nanopure II Filtration System (Barnstead Thermolyne, Dubuque, IA). Safety Precautions and Considerations. When using any high-voltage power supplies that are capable of lethal currents, a current-limiting resistor should be placed in series with the connection to the device. The resistance should be chosen to limit the current to the low microampere range when the power supply is at its maximum voltage. Additionally, shielding and proper insulation should be used along with normal precautions. Since potentially toxic or flammable materials are being sprayed with a device such as this, proper ventilation and vapor removal should be provided. Normal practices for handling laboratory chemicals should be followed. RESULTS AND DISCUSSION Taylor Cone Formation. The liquid motion on the triangular tip was characterized before interfacing with the mass spectrometer. The liquid from the channel outlet wets the two-dimensional triangle tip, forming a triangularly shaped droplet. There was no observed liquid expansion in lateral directions beyond the triangular tip. A Taylor cone was established at the apex of the triangle tip. A 2.0-2.8 kV potential applied to the gold wire immediately established a Taylor cone, and electrospray ionization activity was confirmed by monitoring the total ion current. A stable total ion current of 30-40 nA was measured by a Keithley 486 picoammeter. This confinement of liquid in the particular small area at the edge of the device enables high-density ESI tip array fabrication without cross-channel contamination. An optical image of the triangle prior to activation is shown in Figure 4a. The image of the Taylor cone and the liquid jet established on the triangular tip after pumping and the application of high voltage are shown in Figure 4b. The Taylor cone volume in Figure 4b was estimated to be 0.06 nL. It appeared that the two angles of the triangular tip at the channel exit (angles b and c in Figure 2) need to be smaller than
the mass spectrum with an acquisition time of 1 s. The scan range of the mass-to-charge ratio (m/z) was from 600 to 1300. Figure 5b illustrates the stability of the total ion current for a 10-min continuous-infusion MS run with an acquisition speed of 1 spectrum/s. The spraying was stable, with a relative standard deviation (RSD) of 1.3% in the ion current. Cross-Channel Contamination Test of the Electrospray Ionization Array Device. Taylor cone formation on the ESI array device was observed with the same optical microscope and at the same conditions as the single tip device. Figure 6 shows the fourchannel tip array device. The distance between triangular tips is 80 µm. All four channels established Taylor cones with the application of 2.4-2.6 kV to the reservoirs using the 300 nL/min buffer solution flow rate. A stable total ion current of 28-35 nA was measured independently for each triangular tip. There was no sign of liquid spreading at the edge of the device. In Figure 6, the no. 2 electrospray was actively spraying buffer solution while the other three channels were off. All four channels were filled with the sample solution during the single channel activation. Two different compounds, 1 µM desipramine and 1 µM imipramine, were sprayed from four tips to examine the cross-channel contamination. The acquisition speed for this examination was 1 spectrum/s, scanned from m/z 200 to m/z 400. First, imipramine was sprayed from the no. 1 triangle tip with 2.8-3.2 kV applied to the reservoirs at a flow rate of 200 nL/min for 20 s. The mass spectrum of this electrospray is shown in Figure 7a. Second, desipramine was sprayed for 20 s from the no. 2 tip under the same spraying conditions as the no. 1 tip. The mass spectrum is shown in Figure 7b. Then the no. 3 triangular tip sprayed imipramine, and the no. 4 triangular tip sprayed desipramine under the same spraying conditions as the no. 1 tip, each for 20 s. These mass spectra of the two samples sprayed from the no. 3 and no. 4 tips are shown in Figure 7c,d. There was no sign of contamination from neighboring channels.
Figure 7. (a) Mass spectrum of imipramine sprayed from the no. 1 triangular tip. (b) Mass spectrum of desipramine sprayed from the no. 2 triangular tip. (c) Mass spectrum of imipramine sprayed from the no. 3 triangular tip. (d) Mass spectrum of desipramine sprayed from the no. 4 triangular tip.
that of the contact angle of the buffer solution on the cyclo olefin substrate. Thus, the liquid will be completely confined on the triangular tip. The thickness of the tip did not influence the spraying performance. We compared spraying performances of three different tips (3, 5, and 7 µm tip thickness) whose sizes were the same as the triangular tip in Figure 2. Total ion currents monitored by the picoammeter for these three tips were 35.5 ( 1.2 nA for the 3-µm-thickness tip, 35.3 ( 1.5 nA for the 5-µmthickness tip, and 35.8 ( 1.4 nA for the 7-µm tip. Liquid droplets for each case were completely confined on the tips. This indicates clearly that the ESI activity was only dependent on the top surface of the triangular tip. Stability of the Electrospray Ionization Device. The spraying stability performance was characterized by infusing the buffer solution containing 5 µM chicken cytochrome c. Figure 5a shows
CONCLUSION In this paper, we have described a triangular tip integrated in a microfluidic channel and interfaced to an ESI-MS that displayed a stable electrospray performance. The fabrication process for this device was straightforward and can be used to fabricate arrays of devices. A liquid droplet was confined at the channel outlet on the triangular tip without spreading to adjacent channels. Therefore, dense ESI array devices for high-throughput applications can be fabricated and used without cross-channel contamination. The precise characterization of the optimum tip design, such as tip angle and tip size effect, on the stability of the ion current is the focus of our ongoing research. Additionally, we plan to combine electrophoretic separation with this electrospray device in the future. ACKNOWLEDGMENT This work was supported in part by the National Science Foundation Cooperative Agreement ECS-9876771 through the Nanobiotechnology Center of Cornell University. The authors appreciate access and use of the Cornell Nanofabrication Facility located at Cornell University. Finally, we thank Prof. Jack Henion for valuable discussions and the use of the mass spectrometers. Received for review June 18, 2002. Accepted September 12, 2002. AC020396S Analytical Chemistry, Vol. 74, No. 22, November 15, 2002
5901
