Atmospheric Pressure Photoionization-Mass Spectrometry with a

Anal. Chem. 2004, 76, 6797-6801

Atmospheric Pressure Photoionization-Mass Spectrometry with a Microchip Heated Nebulizer Tiina J. Kauppila,† Pekka O 2 stman,† Seppo Marttila,‡ Raimo A. Ketola,† Tapio Kotiaho,§ Sami Franssila,‡ and Risto Kostiainen*,†,|

Viikki Drug Discovery Technology Center, Faculty of Pharmacy, Department of Chemistry, and Faculty of Pharmacy, Division of Pharmaceutical Chemistry, University of Helsinki, FIN-00014, Helsinki, Finland, and Microelectronics Centre, Helsinki University of Technology, P.O. Box 3500, FIN-02015, Espoo, Finland

A novel, microfabricated heated nebulizer chip for atmospheric pressure photoionization-mass spectrometry (APPIMS) is presented. The chip consists of fluidic and gas inlets, a mixer, and a nozzle etched onto silicon wafer that is anodically bonded to a Pyrex glass wafer, on which an aluminum heater is sputtered. A krypton discharge lamp is used as the source for 10-eV photons to initiate the photoionization process. Dopant, delivered as part of the sample solution, is used to achieve efficient ionization. The use of the microfabricated heated nebulizer with APPI in the analysis of four analytes is demonstrated, and the spectra are compared to those obtained with a conventional APPI source. Ionization in positive and negative ion modes was successfully achieved and the spectra were mainly similar to those obtained with conventional APPI, indicating that the ionization in microfabricated and conventional APPI sources takes place by the same mechanisms. The flow rates with conventional APPI are ∼100 µL/min, whereas the microchip heated nebulizer allows the use of flow rates 0.05-5 µL/min, thus being compatible with microfluidic separation systems or microand nano-LC. A stable signal was demonstrated throughout a 5-h measurement, which proved the excellent stability of the micro-APPI. The same heated nebulizer chip can be used for weeks. The newest trend in the world of analytical chemistry has been the miniaturization of conventional analysis techniques.1 The purpose of the miniaturization is the achievement of rapid analyses with very low solvent consumption with the capability of analyzing very small amounts of sample without sensitivity loss. Also, the possibility of automated analysis, with all analysis stepsssample preparation, separation, and detectionsintegrated on the same microchip, fascinates scientists. * To whom correspondence should be addressed. Tel: +358-9-191 59134. Fax: +358-9-191 59556. E-mail: [email protected]. † Viikki Drug Discovery Technology Center, Faculty of Pharmacy, University of Helsinki. ‡ Helsinki University of Technology. § Department of Chemistry, University of Helsinki. | Faculty of Pharmacy, Division of Pharmaceutical Chemistry, University of Helsinki. (1) Huikko, K.; Kostiainen, R.; Kotiaho, T. Eur. J. Pharm. Sci. 2003, 20 (2), 149-171. 10.1021/ac049058c CCC: $27.50 Published on Web 10/07/2004

© 2004 American Chemical Society

In microsystems, separation of the sample components has mostly been achieved by methods based on capillary electrophoresis (CE) due to flow rates suitable to microchip scale.2 Some reports on the miniaturization of liquid chromatography (LC) have also been published, although on-chip LC has been more difficult to realize than on-chip CE.3 In analyte detection, fluoresence providing high sensitivity and selectivity has been the most popular.4 However, fluoresence detection mostly requires derivatization, which is time-consuming and may decrease quantitative repeatability. In the past few years, the use of mass spectrometry in microchip analysis has gained lots of interest and several reports on the subject have been published. Electrospray (ESI) is currently the method of choice, as the low flow rates used in ESI are ideal for microfluidic devices.5-23 However, no microchip realizations of ESI have been demonstrated for long-term routine (2) Manz, A.; Harrison, D. J.; Verpoorte, E. M. J.; Fettinger, J. C.; Paulus, A.; Luedi, H.; Widmer, H. M. J. Chromatogr. 1992, 593 (1-2), 253-8. (3) Manz, A.; Graber, N.; Widmer, H. M. Sens. Actuators 1990, B1 (1-6), 244-8. (4) Zhang, C.-X.; Manz, A. Anal. Chem. 2001, 73 (11), 2656-2662. (5) Figeys, D.; Gygi, S. P.; McKinnon, G.; Aebersold, R. Anal. Chem. 1998, 70 (18), 3728-3734. (6) Lazar, I. M.; Ramsey, R. S.; Sundberg, S.; Ramsey, J. M. Anal. Chem. 1999, 71 (17), 3627-3631. (7) Li, J.; Thibault, P.; Bings, N. H.; Skinner, C. D.; Wang, C.; Colyer, C.; Harrison, J. Anal. Chem. 1999, 71 (15), 3036-3045. (8) Zhang, B.; Liu, H.; Karger, B. L.; Foret, F. Anal. Chem. 1999, 71 (15), 3258-3264. (9) Liu, H.; Felten, C.; Xue, Q.; Zhang, B.; Jedrzejewski, P.; Karger, B. L.; Foret, F. Anal. Chem. 2000, 72 (14), 3303-3310. (10) Vrouwe, E. X.; Gysler, J.; Tjaden, U. R.; Van der Greef, J. Rapid Commun. Mass Spectrom. 2000, 14 (18), 1682-1688. (11) Deng, Y.; Henion, J.; Li, J.; Thibault, P.; Wang, C.; Harrison, D. J. Anal. Chem. 2001, 73 (3), 639-646. (12) Deng, Y.; Zhang, H.; Henion, J. Anal. Chem. 2001, 73 (7), 1432-9. (13) Kameoka, J.; Craighead, H. G.; Zhang, H.; Henion, J. Anal. Chem. 2001, 73 (9), 1935-1941. (14) Lazar, I. M.; Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 2001, 73 (8), 17339. (15) Xue, Q.; Foret, F.; Dunayevskiy, Y. M.; Zavracky, P. M.; McGruer, N. E.; Karger, B. L. Anal. Chem. 1997, 69 (3), 426-430. (16) Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 1997, 69 (6), 1174-1178. (17) Licklider, L.; Wang, X.-Q.; Desai, A.; Tai, Y.-C.; Lee, T. D. Anal. Chem. 2000, 72 (2), 367-375. (18) Schultz, G. A.; Corso, T. N.; Prosser, S. J.; Zhang, S. Anal. Chem. 2000, 72 (17), 4058-4063. (19) Yuan, C.-H.; Shiea, J. Anal. Chem. 2001, 73 (6), 1080-1083. (20) Kim, J. S.; Knapp, D. R. J. Am. Soc. Mass Spectrom. 2001, 12 (4), 463469. (21) Rohner, T. C.; Rossier, J. S.; Girault, H. H. Anal. Chem. 2001, 73 (22), 5353-5357.

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004 6797

analysis thus far, partly due to spray instability as well as clogging of the small-diameter emitters.13,15,22,24 The analyte signal in ESI can also suffer from suppression due to high ion concentration, which limits the use of buffers and demands thorough sample preparation before analysis.25 In addition, because of the nature of electrospray, it is best suited for molecules that can already be ionized in the solution, which leaves out a large group of nonpolar analytes. Recently, a microfabricated device, based on the principle of atmospheric pressure chemical ionization (APCI), has been introduced.25 The APCI microchip consisted of channels for the introduction of the sample and nebulizing gas flows as well as an integrated heater. A corona discharge needle was placed in front of the microchip, which ionized the nebulized sample. The microchip APCI has been shown to provide a very stable signal, high sensitivity, and good quantitative repeatability. Microchip APCI can be applied to compounds that are slightly less polar than those that can be ionized by using ESI, but it is still inefficient in the analysis of completely nonpolar compounds. Atmospheric pressure photoionization (APPI) is a novel ionization technique for liquid chromatography-mass spectrometry (LC-MS), which was originally developed for the ionization of nonpolar compounds, which cannot be ionized by ESI or APCI. Two distinct APPI apparatuses have been described by Bruins et al.26 and Syage et al.,27 which are based on the same operational principle. The ionization in APPI is initiated by 10-eV photons emitted by a krypton discharge lamp. The ionization of analytes can take place either directly by photoionization or indirectly via gas-phase reactions with other species.28-30 APPI has shown its efficiency in the ionization of nonpolar analytes and has thus broadened the group of compounds that can be analyzed by atmospheric pressure ionization-mass spectrometry (API-MS) techniques. A review on APPI and applications using it has recently been published.31 The optimal flow rates in commercial APPI are ∼0.1 mL/min, excluding the possibility of combining micro-LC or microfluidic devices to MS with APPI. Therefore, in this study we present for the first time a microfabricated APPI device, which utilizes the microchip nebulizer designed for miniaturized APCI.25 The silicon-glass microchip nebulizer consists of microchannels for sample solution and nebulizing gas as well as a heater system. The corona discharge needle used in APCI is replaced by a krypton discharge lamp, so that the ionization is initiated by photons. The performance of the new micro-APPI source is (22) Huikko, K.; Oestman, P.; Grigoras, K.; Tuomikoski, S.; Tiainen, V. M.; Soininen, A.; Puolanne, K.; Manz, A.; Franssila, S.; Kostiainen, R.; Kotiaho, T. Lab Chip 2003, 3 (2), 67-72. (23) Dethy, J.-M.; Ackermann, B. L.; Delatour, C.; Henion, J. D.; Schultz, G. A. Anal. Chem. 2003, 75 (4), 805-811. (24) Smith, A. D.; Moini, M. Anal. Chem. 2001, 73 (2), 240-246. (25) O ¨ stman, P.; Marttila, S. J.; Kotiaho, T.; Franssila, S.; Kostiainen, R., Anal. Chem. in press. (26) Robb, D. B.; Covey, T. R.; Bruins, A. P. Anal. Chem. 2000, 72, 36533659. (27) Syage, J. A.; Evans, M. D.; Hanold, K. A. Am. Lab. 2000, 24-29. (28) Kauppila, T. J.; Kuuranne, T.; Meurer, E. C.; Eberlin, M. N.; Kotiaho, T.; Kostiainen, R. Anal. Chem. 2002, 74, 5470-5479. (29) Kauppila, T. J.; Kotiaho, T.; Bruins, A. P.; Kostiainen, R. J. Am. Soc. Mass Spectrom. 2004, 15 (2), 203-211. (30) Hanold, K. A.; Fischer, S. M.; Cormia, P. H.; Miller, C. E.; Syage, J. A. Anal. Chem. 2004, 76 (10), 2842-2851. (31) Raffaelli, A.; Saba, A. Mass Spectrom. Rev. 2003, 22, 318-331.

6798 Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

Figure 1. Photograph of the heated nebulizer microchip.

demonstrated and compared to that of the conventional APPI source. EXPERIMENTAL SECTION Reagents. Acridine, 2-naphthol, 1,4-naphthoquinone, 2-naphthoic acid, and anisole (99.9%) were purchased from Sigma-Aldrich (Steinheim, Germany) and methanol and toluene from J. T. Baker (Deventer, Holland). All solvents were of analytical or chromatographic grade. Sample Preparation. Stock solutions (10-2 M) of the analytes were prepared in toluene or anisole. For the analyses with the microfabricated APPI, the stock solutions were diluted in a mixture of methanol and toluene (9:1) or methanol and anisole (9:1) to final concentrations of 10-4 (positive ion mode samples) and 10-3 M (negative ion mode samples). For the analyses with conventional APPI, the stock solutions were diluted in methanol to final concentrations of 10-4 or 10-5 M (2-naphthol). Fabrication Process of the Heated Nebulizer Microchip. The heated nebulizer microchip contained fluid inlets, flow channels for both the liquid sample and the nebulizer gas, a mixing chamber, an integrated microheater, and an exit nozzle (Figures 1 and 2). Three photomasks were required for chip fabrication. The first mask defined the channel, mixer, and nozzle patterns on the silicon front side, the second defined the inlet holes on the wafer backside, and the third was used to define the microheater on the glass wafer. Patterns on the wafer front side were aligned to fluidic inlet holes on the back using double-side lithography equipment (Electronic Visions AL6-2). Anisotropic silicon etching was done in 25 wt % tetramethylammonium hydroxide solution at 80 °C. The widths of the nebulizer gas and liquid sample channels were 300 and 120 µm, respectively. The channel depth was 85-190 µm, depending on the channel width. A 300-nm aluminum layer was sputtered on the glass wafer. Aluminum was wet etched in phosphoric acid solution (PS 70-10

acquired over the scan range m/z 50-500 using a step size of 0.1 Da and a dwell time of 0.3 ms. MS/MS measurements were carried out in the multiple reaction monitoring (MRM) mode. The data were collected by using a MacIntosh 8500/180 computer and a PE Sciex MassChrom software (version 1.1.1).

Figure 2. Schematic diagram of the microfabricated APPI.

from Merck). The resistance of the aluminum heater at room temperature was ∼90 Ω. Before bonding the surfaces were hydrophilized in ammonia/peroxide solution (RCA-1) to make alignment easier and to get higher bond strength. After alignment, the wafer stack was heated to 350 °C, and voltage of 300 V was applied for 15 min. Fluidic connectors (Nanoport, Upchurch Scientific) were glued with epoxy after wafer dicing. PEEK tubing 130-µm i.d., 150-µm o.d. (Upchurch Scientific) was used to access the chip. Chip size (18 × 29 mm2) was mainly determined by the size of the fluidic connectors. Sample and Solvent Delivery. In the measurements done with the micro-APPI, a microsyringe pump (Harvard Apparatus Inc., Holliston, MA) at flow rates of 0.05-5 µL/min was used for the sample delivery. In the measurements done with the conventional APPI, the samples were injected into a continuous solvent stream by using a 50-µL loop. An HP 1050 series pump was used to deliver the solvent at a flow rate of 200 µL/min. The dopant (toluene or anisole) was delivered by using a microsyringe pump (Harvard Apparatus Inc.) at a flow rate of 20 µL/min. Mass Spectrometry. The mass spectrometer was a PE Sciex API 300 triple quadrupole mass spectrometer (Sciex, Concord, Canada). In the micro-APPI measurements, the mass spectrometer was fitted with an x,y,z stage (Proxeon, Odense, Denmark) equipped with a Teflon folder (The Finnish School of Watchmaking, Espoo, Finland). Reference data were collected by using the same apparatus fitted with a conventional APPI source (Machine Shop, University of Groningen, The Netherlands). The photoionization lamp used with both ion sources was a 10-eV model PKS 100 krypton discharge lamp (Cathodeon Ltd., Cambridge, England). External power supply (GW GPS-3030, Good Will Instruments Co. Ltd.) was used to adjust the temperature of the nebulizer microchip. High-purity nitrogen (99.999%, Oy Woikoski Ab, Voikoski, Finland) was used as nebulizer gas and nitrogen delivered from a nitrogen generator (Whatman, Haverhill, MA) as curtain and collision gases. In addition, the conventional APPI source used compressed air filtered by an Atlas Copco air-dryer (Wilrijk, Belgium) as auxiliary gas and high-purity nitrogen as lamp gas. The instrument was operated in positive and negative ion modes. The mass spectrometer voltages were optimized to give a maximum signal of each analyte. Mass spectra were

RESULTS AND DISCUSSION The positions of the nebulizer chip and the krypton discharge lamp were found to be very critical for efficient ionization and were optimized together by monitoring the intensity of the toluene radical cation signal. The best signal was obtained when the lamp was positioned as close to the MS orifice as possible with the nebulizer microchip in its immediate proximity. Because of the large area demanded by the lamp, the place of the chip had to be compromised and the chip could not be fitted as close to the orifice as when the nebulizer chip was used with the corona discharge needle.25 Therefore, a modified design with a smaller lamp could permit the position of the nebulizer chip closer to the mass spectrometer orifice, which could increase the sensitivity. A repeller was needed to create a suitable electric field to transfer the ions from the ion source into MS. This was achieved by attaching a metal plate in front of the Teflon holder, which was connected to the high-voltage supply (1.3-2 kV) (Figure 2). The performances of the micro-APPI and the conventional APPI were compared in the analysis of four compounds. Toluene and anisole, which were introduced as part of the solvent, were used as dopants. The dopant is a substance that has ionization energy (IE) below the energy of the photons and that thus can be used to enhance the ionization efficiency, as the photons easily lose their energy in collisions with surfaces and gas-phase particles. The dopant was found necessary for efficient ionization in all measurements. Acridine and 2-naphthol were chosen as analytes in the positive ion mode, because they both have low ionization energies but different proton affinities (PAs).32 Acridine has high PA and can readily accept protons from protonated solvent molecules or other species that have PAs below that of acridine. A protonated molecule (MH+) of acridine, formed by proton transfer, was observed in the acridine spectra with the micro-APPI as well as with the conventional APPI source (Figure 3a and b). 2-Naphthol was chosen as an example of low-PA compounds that cannot be ionized by proton transfer.28 The 2-naphthol spectra showed an intense radical cation (M+.) with both sources (Figure 3c and d), indicating that efficient ionization of nonpolar analytes can be achieved using the micro-APPI. This has also been shown to be the case with conventional APPI.28,33 Anisole was used as the dopant when 2-naphthol was analyzed, because anisole has higher PA than toluene and therefore the anisole radical cation does not react as easily with solvent molecules. As the anisole radical cations stay in the system, the charge exchange reaction can take place more efficiently.33 In negative ion mode, 1,4-naphthoquinone and 2-naphthoic acid were chosen as model compounds. 1,4-Naphthoquinone has positive electron affinity, and therefore, it can be ionized by electron capture or charge exchange reaction.29 Oxidation reac(32) Linstrom, P. J., Mallard, W. G., Eds. NIST Chemistry WebBook, NIST Standard Reference Database 69; National Institute of Standards and Technology, Gaithersburg, MD 20899, July 2001 (http://webbook.nist.gov). (33) Kauppila, T. J.; Kostiainen, R.; Bruins, A. P. Rapid Commun. Mass Spectrom. 2004, 18 (7), 808-815.

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

6799

Figure 3. Spectra of the analytes using micro-APPI and conventional APPI. (a) Acridine (100 µM), (b) acridine (100 µM), (c) 2-naphthol (100 µM), (d) 2-naphthol (10 µM), (e) 1,4-naphthoquinone (1 mM), (f) 1,4-Naphthoquinone (100 µM), (g) 2-naphthoic acid (1 mM), and (h) 2-naphthoic acid (100 µM). Micro-APPI (a, c, e, and g) solvent methanol/toluene (9:1); except (c) solvent methanol/anisole (9:1), flow rate 2.5 µL/min; except (e) flow rate 5 µL/min. Conventional APPI (b, d, f, h) 50-µL loop injection, solvent methanol (200 µL/min), dopant toluene (20 µL/min); except (d) dopant anisole (20 µL/min).

tions with atmospheric oxygen have also been reported.29 In this study, the 1,4-naphthoquinone spectra showed both a negative molecular ion M-•, formed by electron capture or charge exchange and oxidation products (Figure 3e and f). However, the proportion of the oxidation products was much higher with the conventional APPI than with the micro-APPI, which indicates that 6800

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

either the oxidation reactions are catalyzed by the metal surfaces inside the conventional APPI ion source or the formation of oxidation products is dependent on the residence time, which may be shorter with the micro-APPI. 2-Naphthoic acid was chosen because of its high gas-phase acidity and thus the tendency to form a deprotonated molecule ([M - H]-) by proton transfer. As

Figure 4. Signal of the three main product ions of the MH+ of acridine (100 µM) in methanol/toluene (9:1) at flow rates 0.05-5 µL/ min with the micro-APPI (measured in MRM mode).

Figure 5. Stability of the micro-APPI. Acridine (100 µM) in methanol/ toluene (9:1), MRM of three main fragments of MH+ of acridine.

assumed, the 2-naphthoic acid spectra showed a deprotonated molecule ([M - H]-) with both ion sources (Figure 3g and h). In addition, a breakdown product m/z 127 [M - H - CO2]- was shown in the spectrum with conventional APPI. The reason this fragment was missing from the spectrum obtained with microAPPI may be the lower temperature25 or the possibly shorter resident time with micro-APPI. The suitability of different flow rates was tested by monitoring the signal of the three main fragments of the MH+ of acridine (measured in MRM mode) at flow rates of 0.05-5 µL/min (Figure 4). Ionization was achieved at all flow rates, and the relative abundances of the three product ions were observed to stay the same throughout the measurement. However, the absolute abundance of the signal was observed to rise as the flow rate was increased, being most intense when the flow rate was >1 µL/ (34) Kauppila, T. J.; Kostiainen, R.; Bruins, A. P. Effect of the Flow Rate on the Ionization Efficiency in Atmospheric Pressure Photoionization Mass Spectrometry. Proceedings of the 51st ASMS Conference on Mass Spectrometry and Allied Topics, Montreal, PQ, Canada, June 8-12, 2003. (35) Kauppila, T. J.; Kostiainen, R.; Bruins, A. P., submitted.

min. This was thought to be due to the mass flow dependency of the system at these flow rates or the insufficient amount of the dopant (10% of solvent), as it was shown earlier that the amount of the dopant is a crucial factor in APPI and may limit the sensitivity.34,35 The performance of the micro-APPI was evaluated by introducing 100 µM acridine in methanol/toluene (9:1) with a flow rate of 2.5 µL/min and by monitoring the signal of three main product ions of acridine MH+. The micro-APPI produced a highly stable ion current as demonstrated in a 5-h continuous measurement (Figure 5). The stable performance of the micro-APPI is an important step in the development of a robust quantitative interface for the connection of microfluidic devices to mass spectrometer. CONCLUSIONS The use of a micro-APPI as an ion source for mass spectrometry was demonstrated for the first time. Ionization of analytes was successfully achieved in positive and negative ion modes. The results showed that the main ionization reactions observed with the conventional APPI sourcesproton transfer and charge exchange in positive ion mode and proton transfer and electron capture/charge exchange in negative ion modesalso take place with the micro-APPI. The low flow rates (0.05-5 µL/min) used in micro-APPI make possible the combination of micro/nano-LC or microfluidic devices to MS, which is not possible with commercial APPI. Furthermore, the manufacturing costs of microAPPI are much lower than those of conventional APPI. The longterm stability of the micro-APPI source, and the capability of analyzing polar and nonpolar analytes from very small sample volumes, are advantages over the present ESI chips and make the micro-APPI a considerable alternative for interfacing low flow rate separation systems with MS. In future, the performance of the micro-APPI can be improved by optimizing the voltage, the position, and the shape of the repeller. Also, a smaller lamp design would admit the nebulizer chip to be placed closer to the MS orifice, which may also increase the ionization efficiency. A new design of the chip with a special dopant inlet is now in preparation, which will make the source more practical as it will allow the use of solvents that are not soluble to typical dopants. ACKNOWLEDGMENT This study was financially supported by the Finnish Cultural Foundation, National Technology Agency of Finland, Orion Pharma, Juvantia Pharma Ltd., United Laboratories Ltd., and Danisco Sugar and Sweeteners. Received for review June 28, 2004. Accepted August 24, 2004. AC049058C

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

6801