Gas Chromatography-Microchip Atmospheric Pressure Chemical

Anal. Chem. 2006, 78, 3027-3031

Gas Chromatography-Microchip Atmospheric Pressure Chemical Ionization-Mass Spectrometry Pekka O 2 stman,† Laura Luosuja 1 rvi,‡ Markus Haapala,† Kestas Grigoras,§ Raimo A. Ketola,† Tapio Kotiaho,‡ Sami Franssila,§ and Risto Kostiainen*,†

Department of Pharmaceutical Chemistry, Faculty of Pharmacy, P.O. Box 56, FIN-00014 University of Helsinki, Finland, Laboratory of Analytical Chemistry, Department of Chemistry, P.O. Box 55, FIN-00014 University of Helsinki, Finland, and Microelectronics Centre, Helsinki University of Technology, P.O. Box 3500, FIN-02015 HUT, Finland

An atmospheric pressure chemical ionization (APCI) microchip is presented for combining a gas chromatograph (GC) to a mass spectrometer (MS). The chip includes capillary insertion channel, stopper, vaporizer channel, nozzle and nebulizer gas inlet fabricated on the silicon wafer, and a platinum heater sputtered on a glass wafer. These two wafers are joined by anodic bonding creating a two-dimensional version of an APCI microchip. The sample from GC is directed via heated transfer line capillary to the vaporizer channel of the APCI chip. The etched nozzle forms narrow sample plume, which is ionized by an external corona discharge needle, and the ions are analyzed by a mass spectrometer. The GCmicrochip APCI-MS combination provides an efficient method for qualitative and quantitative analysis. The spectra produced by microchip APCI show intensive protonated molecule and some fragmentation products as in classical chemical ionization for structure elucidation. In quantitative analysis the GC-microchip APCI-MS showed good linearity (r2 ) 0.9989) and repeatability (relative standard deviation 4.4%). The limits of detection with signal-to-noise ratio of three were between 0.5 and 2 µmol/L with MS mode using selected ion monitoring and 0.05 µmol/L with MS/MS using multiple reaction monitoring. Gas chromatography combined with mass spectrometry (GC/ MS) is a powerful analytical technique for volatile and thermally stable analytes, as demonstrated by the great number of applications in, for example, environmental,1 drug,2 and food analysis.3 In GC/MS, the separated analytes are traditionally ionized in a vacuum prior to detection by MS. The most common ionization method is electron ionization (EI), which provides high ionization efficiency with reproducible and characteristic mass spectra. Positive and negative ion chemical ionization (CI) are also commonly used in GC/MS,4 particularly in determination of * Corresponding author: (e-mail) [email protected]; (phone) +3589-191 59 134; (fax) +358-9-191 59 556. † Department of Pharmaceutical Chemistry, University of Helsinki. ‡ Laboratory of Analytical Chemistry, University of Helsinki. § Microelectronics Centre, Helsinki University of Technology. (1) Bernd, S. R. T. Mass Spectrom. Rev. 2005, 24, 719-765. (2) Maurer, H. H. J. Chromatogr. 1992, 580, 3-41. (3) Lehotay, S. J.; Hajslova, J. Trends Anal. Chem. 2002, 21, 686-697. 10.1021/ac052260a CCC: $33.50 Published on Web 04/04/2006

© 2006 American Chemical Society

molecular weights, since fragmentation is considerably less with CI than with EI. Modern liquid chromatography-mass spectrometry (LC/MS) is based on atmospheric pressure ionization (API) techniques: electrospray ionization (ESI),5,6 atmospheric pressure chemical ionization (APCI),7,8 and recently introduced atmospheric pressure photoionization (APPI).9,10 These techniques provide efficient ionization for a wide variety of molecules. ESI is an excellent method for ionic and polar compounds and can be applied for small and large molecules such as peptides and proteins, whereas APCI and especially APPI are better ionization techniques for less polar and neutral small molecules, but they are not suitable for large biomolecules. These methods have become very popular since they are able to ionize molecules that are not amenable to GC/MS. GC/MS and LC/MS have been separate instruments, since in GC/MS compounds are ionized at vacuum (EI and CI) and in LC/MS at atmospheric pressure (ESI, APCI, APPI). However, Dzinic et al.11 showed already in 1976, that the interfacing of GC to MS using APCI is straightforward. After this GC-APCI-MS in negative ion mode was applied in analysis of environmental samples including tetrachlorodibenzo-p-dioxin,12-14 tetrachlorodibenzofuran,15 nitro-polycyclic aromatic hydrocarbons (nitroPAH),16-21 amino-PAH,22 and pesticides.23 Since then GC-APCI(4) Munson, M. S. B.; Field, F. H. J. Am. Chem. Soc. 1966, 88, 2621-2630. (5) Dole, M.; Mack, L. L.; Hines, R. L.; Mobley, R. C.; Ferguson, L. D.; Alice, M. B. J. Chem. Phys. 1968, 49, 2240-2249. (6) Yamashita, M.; Fenn, J. B. J. Phys. Chem. 1984, 88, 4611-4615. (7) Horning, E. C.; Carroll, D. I.; Dzidic, I.; Haegele, K. D.; Horning, M. G.; Stillwell, R. N. J. Chromatogr. Sci. 1974, 12, 725-729. (8) Carroll, D. I.; Dzidic, I.; Stillwell, R. N.; Haegele, K. D.; Horning, E. C. Anal. Chem. 1975, 47, 2369-2373. (9) Robb, D. B.; Covey T. R.; Bruins, A. P. Anal. Chem. 2000, 72, 3653-3659. (10) Syage, J. A.; Evans, M. D.; Hanold, K. A. Am. Lab. 2000, 32, 24-29. (11) Dzidic, I.; Carroll, D. I.; Stillwell, R. N.; Horning E. C. Anal. Chem. 1976, 48, 1763-1768. (12) Mitchum, R. K.; Moler, G. F.; Korfmacher, W. A. Anal. Chem. 1980, 52, 2278-2282. (13) Mitchum, R. K.; Korfmacher, W. A.; Moler, G. F. Anal. Chem. 1982, 54, 719-722. (14) Korfmacher, W. A.; Moler, G. F.; Delongchamp, R. R.; Mitchum, R. K.; Harless, R. L. Chemosphere 1984, 13, 669-685. (15) Korfmacher, W. A.; Mitchum, R. K.; Hileman, F. D.; Mazer, T. Chemosphere 1983, 12, 1243-1249. (16) Korfmacher, W. A.; Fu, P. P.; Chou, M.; Mitchum, R. K. HRC & CC 1984, 7, 41-42. (17) Korfmacher, W. A.; Miller, D. W. HRC & CC 1984, 7, 581-583. (18) Korfmacher, W. A.; Rushing, L. G. HRC & CC 1986, 9, 293-295.

Analytical Chemistry, Vol. 78, No. 9, May 1, 2006 3027

MS has been used only occasionally.24 Recentlym McEwen and McKay modified a commercial LC/MS instrument to allow both atmospheric pressure LC/MS and GC/MS.25 Revelsky et al. presented APPI-MS in analysis of gaseous samples and later they showed that GC-APPI-MS provides an efficient method for the analysis of wide variety of analytes including n-alkanes, alcohols, ketones, esters, amines, amino acids, and aromatic hydrocarbons.26,27 Although the potential of APCI and APPI in GC/MS have been demonstrated, the methods have not become popular. This is partly because APCI and APPI interfaces optimized for GC/ MS are not yet commercially available. The current trend in analytical chemistry is miniaturization using microchip technology in order to improve sensitivity and speed of analysis. Coupling of MS to microfluidic systems has been intensively studied. These developments have been focused nearly exclusively on microchip ESI techniques, which are not suitable for gaseous samples. Recently, a microchip APCI-MS system was presented for liquid samples,28 which provided flow rates from tens of nanoliters to several microliters per minute, efficient ionization, excellent sensitivity, and good reproducibility. Furthermore, Kauppila et al. introduced a microchip APPI also for liquid samples.29 Both of these microchips could offer a potential ionization technique for GC/MS. In this work, we present a modified APCI microchip using an external corona discharge needle for coupling GC to MS. The sample from GC is directed via heated transfer line inserted from the rear edge into a stopper inside the vaporizer channel of the APCI chip. The microfabricated stopper in the vaporizer channel ensures correct positioning of the capillary, which leads to reproducible insertion and minimization of the dead volumes inside the microchip APCI. The feasibility of new microchip APCI for quantitative GC/MS work was investigated with a set of volatile organic compounds and with testosterone representing semivolatile compounds. EXPERIMENTAL SECTION Fabrication Process of the APCI Microchip. The fabrication process is presented in more detailed by Franssila et al.30 Shortly, double-side polished 100-oriented silicon wafers (380 µm thick, high resistivity >500 Ω cm) were used as substrates, and 500(19) Korfmacher, W. A.; Rushing, L. G.; Engelbach, R. J.; Freeman, J. P.; Djuric, Z.; Fifer, E. K.; Beland, F. A. HRC & CC 1987, 10, 43-45. (20) Korfmacher, W. A.; Rushing, L. G.; Arey, J.; Zielinska, B.; Pitts, J. N., Jr. HRC & CC 1987, 10, 641-646. (21) Engelbach, R. J.; Korfmacher, W. A.; Rushing, L. G. HRC & CC 1988, 11, 661-663. (22) Kinouchi, T.; Lopez M.; Arleen T.; Rushing, L. G.; Beland, F. A.; Korfmacher, W. A. High Resolut. Chromatogr. 1990, 13, 281-284. (23) Korfmacher, W. A.; Rushing, L. G.; Siitonen, P. H.; Branscomb, C. J.; Holder, C. L. HRC & CC 1987, 10, 332-336. (24) Guimbaud, C.; Bartels-Rausch, T.; Ammann M. Int. J. Mass Spectrom. 2003, 226, 279-290. (25) McEwen, C. N.; McKay R. G. J. Am. Soc. Mass Spectrom. 2005, 16, 17301738. (26) Revelsky, I. A.; Yashin, Y. S.; Voznesensky, V. N.; Kurochkin, V. K.; Kostyanovsky, R. G. Izv. Akad. Nauk SSSR, Ser. Khim. 1987, 9, 19871992. (Article in Russian). (27) Revelsky, I. A.; Yashin, Y. S.; Sobolevsky, T. G.; Revelsky, A. I.; Miller, B.; Oriedo, V. Eur. J. Mass Spectrom. 2003, 9, 497-507. (28) O ¨ stman, P.; Marttila, S. J.; Kotiaho, T.; Franssila, S.; Kostiainen, R. Anal. Chem. 2004, 76, 6659-6664. (29) Kauppila, T. J.; O ¨ stman, P.; Marttila, S.; Ketola, R. A.; Kotiaho, T.; Franssila, S.; Kostiainen, R. Anal. Chem. 2004, 76, 6797-6801. (30) Franssila, S.; Marttila, S.; Kolari, K.; O ¨ stman, P.; Kotiaho, T.; Kostiainen, R.; Lehtiniemi, R.; Fager, C.-M.; Manninen, J. Submitted to J. MEMS.

3028

Analytical Chemistry, Vol. 78, No. 9, May 1, 2006

µm-thick Pyrex glass wafers (Corning 7740) as channel cover plates. Three photomasks were required for chip fabrication. The first mask defines the capillary insertion channel, vaporizer channel, and nozzle (on silicon top side), mask two is for throughwafer nebulizer gas inlet holes (on the silicon backside), and mask three defines the integrated platinum heater (on glass wafer). Silicon wafers were cleaned in ammonia-peroxide and hydrogen chloride-peroxide mixtures (RCA-1 and RCA-2) followed by thermal oxidization at 1000 °C, resulting in 1000-nm-thick oxide. The top side feature shapes were patterned first, followed by the backside nebulizer gas inlet definition using double-sided alignment. Oxide was partly etched by a reactive ion etching technique first and then completed by buffered hydrofluoric acid (BHF) etching. Photoresist was stripped in acetone and 2-propanol. Silicon anisotropic etching was performed in potassium hydroxide (20 wt %) and 2-propanol (10% bw) mixture at 82 °C. Capillary insertion channel width and depth were 520 and 230 µm, respectively. Vaporizer channel width was 300 µm and depth 240 µm. Pyrex glass wafers were cleaned in acetone and 2-propanol. The heater patterning was done by a liftoff method. First, a 5-µmthick layer of AZ 4562 photoresist was spin-coated, exposed through a photomask, and developed. Chromium adhesion promotion layer (10 nm) was sputtered first, followed by a 100nm-thick platinum layer. Platinum was sputtered in two steps to reduce thermal cracking of the photoresist. Finally, photoresist was removed in acetone and 2-propanol baths with additional ultrasound agitation. The resistance of the heater was ∼250 Ω. Before bonding, the silicon oxide was removed completely from the silicon wafer by etching in BHF solution and the surface was hydrophilized in RCA-1 solution. After alignment, the silicon and glass wafers were joined together by anodic bonding. The wafer stack was heated to 320 °C, and a voltage of 500 V was applied for 15 min, with maximum current reaching 15 mA. Fluidic connector (Nanoport, Upchurch Scientific) for the nebulizer gas inlet and sample inlet capillary (i.d. 0.15 mm, o.d. 0.22 mm, deactivated fused-silica capillary, SGE) was glued with high-temperature epoxy (Cotronics 4703, Cotronics Corp., New York) after wafer dicing into the separate chips. Chemicals and Samples. Benzaldehyde and 2-acetylnaphthalene were obtained from Sigma-Aldrich Chemie GmbH (Steinheim, Germany). Anisole was purchased from Fluka Chemie GmbH (Buchs, Switzerland) and acetoacetone from E. Merck AG (Darmstadt, Germany). Stock solutions of benzaldehyde, 2-acetylnaphthalene, and anisole were prepared by dissolving the compounds in hexane (Merck KGaA, Darmstadt, Germany) to a concentration of 1 mmol/L. Stock solution of acetoacetone was prepared by dissolving the compound in ethanol (AA, Primalco, Rajama¨ki, Finland) to a concentration of 10 mmol/L. Final working solutions of the analytes were prepared by diluting stock solutions further with hexane. Human urine sample prepared for analysis of steroids was obtained from United Laboratories Ltd., Finland. Sample pretreatment consisted of enzymatic hydrolysis of steroid glucuronites followed by liquid extraction at basic pH.31 The testosterone concentration of the sample was 1.12 µmol/L (corresponding to 48.66 ng/mL). Prior to GC-microchip APCI-MS measurement, the ethanol solvent was evaporated to dryness and changed to hexane. (31) Thevis, M.; Geyer, H.; Mareck, U.; Scha¨nzer, W. J. Mass Spectrom. 2005, 40, 955-962.

Figure 1. (A) Schematic presentation of the positioning of the microchip APCI and the corona discharge needle in front of the mass spectrum. (B) Schematic presentation of the microchip APCI. Sample inlet capillary is inserted from the rear edge of the chip to the narrowing. The nebulizer gas flows from behind and along the capillary to the flow channel.

Gas Chromatography. A HP5890A gas chromatograph (Hewlett-Packard, Waldbronn, West Germany) was used. The column used was a 15 m × 0.25 mm i.d. (5% phenyl-95% dimethylpolysiloxane, 0.25-µm film thickness; FactorFour VF-5ms, Varian Inc.). Helium (99.9995%, Oy Woikoski Ab, Voikoski, Finland) was used as a carrier gas. Samples (1 µL) were injected manually with 1-min splitless injection. Injector temperature was 280 °C. The temperature program was from 34 (4 min) to 230 °C (20 °C/min) in the analysis of volatile organic compounds. In the analysis of testosterone sample, the program was from 60 (1.5 min) to 265 °C (35 °C/min) and from 265 to 300 °C (3 °C/min) (4-min hold). Deactivated fused-silica capillary used as a transfer line between GC and microchip APCI was connected to the GC column with a capillary column butt connector (Supelco). The other end of the transfer line was pushed into the microchip via the insertion channel until the stopper and the capillary was fixed by gluing it at the rear end of the microchip with epoxy standing at high temperature. The transfer line was heated by self-made resistor wire heater. The temperature was adjusted to 280 °C (300 °C for the testosterone sample) with external power supply (GW GPS-3030, Good Will Instruments Co. Ltd., Taiwan) and monitored with a Fluke 54 Series II thermometer (Fluke Corp.) with a K-type thermoelement. The microchip APCI was positioned 1 cm away from the mass spectrometer orifice in orthogonal position (Figure 1A). Mass Spectrometry. The mass spectrometer was a PE Sciex API-300 triple quadrupole (Perkin-Elmer Sciex, Concord, Canada). Nitrogen produced by a Whatman 75-720 nitrogen generator (Whatman Inc., Haverhill, MA) was used as curtain and nebulizer gas. Nebulizer gas was introduced to the microchip APCI via a 510-µm-i.d. PEEK tubing (Upchurch Scientific) through the Nanoport with a back pressure of 0.15 bar. The external corona darning needle was a knitting needle (John James size 3/9, Entaco Limited, Warwickshire, England). The needle current was set to 1 µA. The corona discharge needle distance was 0.5 cm from the MS and 1 cm in front of the microchip APCI. External power supply (EPS EP-6515, EPS Stromversorgung GmbH, Augsburg, Germany) was used to adjust the temperature of the microchip APCI in the range of 70-170 °C (chip outside temperature, measured with Fluke 54 Series II thermometer). The mass spectrometer was operated in a positive ion mode. In the measurements of mass spectra, the scan range was m/z 30-230 (1.0 s/scan with a step size of 0.1 Da). Limits of detection

Figure 2. Selected ion chromatograms of the protonated molecules of the sample compounds. The concentrations were 100 µmol/L.

(LODs) with signal-to-noise ratio (S/N) of 3 were measured by MS using selected ion monitoring (SIM) of the protonated molecule and one fragment ion with dwell time of 124 ms. LODs (S/N ) 3) were measured also by MS/MS using multiple reaction monitoring (MRM) of two product ions with dwell tine of 495 ms. The same MRM was used in determination of linearities and repeatabilities. Data were acquired with a MassCrom 1.1.1 software (PE Sciex). RESULTS AND DISCUSSION The fused-silica transfer line capillary is inserted through the insert channel until the capillary meets the stopper, which is a conical narrowing fabricated inside the APCI chip (Figure 1B). The stopper provides reproducible inserting of the capillary to the right position and minimization of the dead volumes. The nebulizer gas, introduced behind the stopper to the insertion channel, flushes the compounds eluated from the transfer line efficiently toward the vaporizer and the exit nozzle minimizing memory effects and peak tailing. The dead volume of microchip APCI is ∼1.5 µL, which does not cause peak broadening, since the nebulizer gas flow rate is ∼2.5 mL/s and the whole vaporizer channel is flushed in less than 1 ms. Platinum as a chemically inert heater material provides stable and long-term operation even at high temperatures. With platinum heater, the temperatures were tested up to the glass softening temperature of ∼600 °C.30 Glass cracking was observed at this temperature. In practice, the maximum operational surface temperature limit of ∼300 °C was set by the epoxy glue used in connecting the Nanoport assemblies. The etched nozzle, as shown also in our earlier work,30 produces a narrow plume toward the corona needle. Because the initial ionization takes place in a very small volume near the needle tip,32 it can be expected that a confined plume results in a larger fraction of ionized analytes, leading to improved sensitivity. Separation efficiency of the system is excellent with narrow peaks, as shown in Figure 2. The peak half-widths and peak asymmetry factors of the test compounds were 0.02-0.04 min and (32) Chen, J.; Davidson J. H. Plasma Chem. Plasma Process. 2002, 22, 495522.

Analytical Chemistry, Vol. 78, No. 9, May 1, 2006

3029

Table 1. Retention Times (tR), Peak Widths at Half-Height (Wh), Peak Base Widths (Wb), Theoretical Plates (N), Plates per Length (N/L), and Peak Asymmetry (As)a compound

Rt

Wh (min)

Wb (min)

N

N/L

As

acetoacetone anisole benzaldehyde 2-acetylnaphthalene

4.45 6.66 7.24 12.60

0.030 0.022 0.039 0.030

0.284 0.251 0.134 0.504

118709 512354 190923 964355

7914 34157 12728 64290

1.09 1.20 1.20 0.97

a Calculated from selected ion chromatograms of the protonated molecules of the test compounds. The concentrations were 100 µmol/ L, and injection volume was 1 µL.

Figure 4. Typical background mass spectrum of the GC-microchip APCI-MS interface.

Figure 3. Total ion chromatogram of testosterone measured in MRM mode (precursor ion m/z 289.0, product ions m/z 109.1 and 97.1) from human urine sample.

1.09-1.20, respectively (Table 1), typical of GC separations. These results show that the dead volumes of the APCI chip are minimal. The usability of the GC-microchip APCI-MS in analysis of semivolatile compounds was tested with a testosterone urine sample using maximal heater temperature. Testosterone is known to eluate as a broad peak at high temperatures in GC, and it is therefore normally derivatized prior to injection. The successful detection of underivatized testosterone shows that the platinum heater provides adequate temperature for the analysis of semivolatile compounds (Figure 3). In addition, no memory effects were observed in analysis of volatile organic compounds or testosterone. The small size of the APCI chip and the high thermal conductivity of silicon make it possible to heat up the chip from room temperature to 300 °C in less than 1 min.30 This means that the temperature can be changed during the GC run and optimized for each analyte. This is an advantage when the sample includes both relatively labile volatile organic compounds and semivolatile compounds. The thermal fragmentation of the labile volatile compounds can be minimized by using lower temperatures at the beginning of the run, and after their elution, the temperature can be raised to optimize elution of semivolatile compounds. Since the main ionization mechanism in APCI is a protontransfer reaction, the reagent ion composition has a great effect on the ionization efficiency. The main background ions (Figure 3030 Analytical Chemistry, Vol. 78, No. 9, May 1, 2006

4) were protonated water clusters [(H2O)nH]+ (n ) 2, at m/z 37; n ) 3, at m/z 55; n ) 4, at m/z 73) originating from surrounding air and ions N3+ (at m/z 42) and [(H2O)N3]+ (at m/z 60) originating from the nebulizer gas (nitrogen). Protonated water clusters dominate the spectrum contributing ∼80% of the signal. This is in good agreement with earlier studies.33-35 Raising the flow rate of nebulizer gas increased the relative abundances of N3+ and [(H2O)N3]+, whereas the increase of chip temperature had no significant effect on the relative abundances of the reagent ions. The proton affinity (PA) of H2O is relatively low (691 kJ/mol),36 and therefore, the most of the organic molecules can be protonated using H3O+ as a reagent ion. However, the PA of water clusters is higher than that of a single molecule. For example, Wentwold and Goebbert showed that the PA of the water dimer is on average 20 kJ/mol higher than that of a single water molecule.37 Also, the impurities with high proton affinity in laboratory air may have a significant effect on reagent ion composition. However, all the test compounds produced abundant protonated molecules, indicating that the proton affinities of the test compounds are high enough for efficient proton-transfer reaction. The PAs of acetylacetone, benzaldehyde, anisole, and 2-acetylnaphthalene are 874, 834, 840, and 909 kJ/mol, respectively.36,38 The mass spectra of all compounds also showed some abundant fragment ions. Acetoacetone (Figure 5A) showed a cleavage of acetone [M + H - C3H6O]+ (at m/z 43). Anisole (Figure 5B) showed fragment ions [M + H - CH3]+ (at m/z 94) and [M + H - CH3OH]+ (at m/z 77). Benzaldehyde (Figure 5C) produced fragment ions [M + H - CH2O]+ at m/z 77 and [M + H - CO]+ at m/z 79. Both anisole and benzaldehyde formed hydronium ion adducts at m/z 127 and 125, respectively. 2-Acetylnaphthalene (Figure 5D) showed only one weak fragment ion [COCH3]+ at m/z 43. The extent of fragmentation was slightly increased when the chip temperature was raised. Also the (33) Shahin, M. M. J. Chem. Phys. 1965, 45, 2600-2605. (34) Good, A., Durden D. A.; Kebarle P. J. Chem. Phys. 1970, 52, 222-229. (35) Kambara, H.; Mitsui, Y.; Kanomata, I. Anal. Chem. 1979, 51, 1447-1452. (36) NIST Chemistry WebBook (http:// webbook.nist.gov/chemistry/). (37) Goebbert, D. J.; Wenthold P. G. Eur. J. Mass Spectrom. 2004, 10, 837845. (38) Kauppila, T. J.; Kuuranne, T.; Meurer, E. C.; Eberlin, M. N.; Kotiaho, T.; Kostiainen, R. Anal. Chem. 2002, 74, 5470-5479.

Table 2. Limit of Detection (S/N ) 3) of Test Compounds in MS (SIM Mode) and MS/MS (MRM Mode) compound

LOD (MS) (µmol/L)

LOD (MS/MS) (nmol/L)

anisole benzaldehyde 2-acetylnaphthalene

2 0.5 1

50 50 50

ion. In our experiments, the main reagent ions were protonated water clusters having relatively low PAs compared to the analytes. This means that the proton-transfer reaction is relatively exothermic, which may cause increased fragmentation. The suitability of the GC-microchip APCI-MS interface for quantitative work was tested by determining the LODs using SIM and MRM modes for anisole, benzaldehyde, and 2-acetylnaphthalene. For acetoacetone, the background was too high for determining LOD. LODs with a signal-to-noise ratio of 3 were between 0.5 and 2 µmol/L with SIM mode and 0.05 µmol/L with MRM mode (Table 2). The linearity of response for anisole was measured using MRM mode within the range of 0.2-100 µmol/ L. Good linearity with regression coefficient (r2) of area 0.9989 was observed. The repeatability was evaluated with six anisole samples (2 µmol/L). The relative standard deviation (RSD) of the peak area was found to be 4.4%, indicating good repeatability even with manual injections. The RSD of retention times was 0.1%. These results show that GC-microchip APCI-MS is suitable for quantitative analysis. CONCLUSIONS The APCI microchip provides a convenient and easy method to combine GC to any API-MS equipped with an APCI source. The same chip can be used as an interface also in micro LC/MS or with microfuidic MS systems as shown in our earlier work.28 This means that the usability of MS systems designed for LC/ MS can be enlarged and separate systems for GC/MS and LC/ MS are not necessarily needed. The GC-microchip APCI-MS provides an efficient method for qualitative and quantitative analysis with detection limits down to the nanomole per liter range. The limits of detection demonstrated here can be easily improved by using a more sensitive mass spectrometer than used in this study. The spectra produced by microchip APCI showed protonated molecule and some fragmentation providing the very same type of information as classical chemical ionization for structure elucidation. The good repeatability and linearity of the GCmicrochip APCI-MS show the potential of the method in quantitative analysis.

Figure 5. Mass spectrum of (A) acetoacetone, (B) anisole, (C) benzaldehyde, and (D) 2-acetylnaphthalene. The concentrations were 100 µmol/L.

fragmentation is dependent on the enthalpy of the proton-transfer reaction, i.e., the PA difference between an analyte and a reagent

ACKNOWLEDGMENT We gratefully acknowledge The Academy of Finland, The Finnish National Technology Agency TEKES, VTT Technical Research Centre of Finland, Environics Oy, Silecs Oy, Braggone Oy, Labmaster Oy, and CHEMSEM graduate school for financial support and United Laboratories Ltd. for providing the urine sample. Received for review December 21, 2005. Accepted March 13, 2006. AC052260A Analytical Chemistry, Vol. 78, No. 9, May 1, 2006

3031