Non-Proximate Detection of Small and Large ... - ACS Publications

Ismael Cotte-Rodrıguez, Christopher C. Mulligan, and R. Graham Cooks*. Department of Chemistry, Purdue University, West Lafayette, Indiana 47907. Amb...
0 downloads 0 Views 335KB Size
Anal. Chem. 2007, 79, 7069-7077

Non-Proximate Detection of Small and Large Molecules by Desorption Electrospray Ionization and Desorption Atmospheric Pressure Chemical Ionization Mass Spectrometry: Instrumentation and Applications in Forensics, Chemistry, and Biology Ismael Cotte-Rodrı´guez, Christopher C. Mulligan, and R. Graham Cooks*

Department of Chemistry, Purdue University, West Lafayette, Indiana 47907

Ambient surfaces are examined by mass spectrometry at distances of up to 3 m from the instrument without any prior sample preparation. Non-proximate versions of the desorption electrospray ionization (DESI) and desorption atmospheric pressure chemical ionization experiments are shown to allow rapid, sensitive, and selective detection of trace amounts of active ingredients in pharmaceutical drug formulations, illicit drugs (methamphetamine, cocaine, and diacetylmorphine), organic salts, peptides, chemical warfare agent simulants, and other small organic compounds. Utilizing an ion transport tube to transport analyte ions to the mass spectrometer, nonproximate DESI allows one to collect high-quality, largely interference-free spectra with signal-to-noise (S/N) ratios of more than 100. High selectivity is achieved by tandem mass spectrometry and by reactive DESI, a variant experiment in which reagents added into the solvent spray allow bondforming reactions with the analyte. Ion/molecule reactions were found to selectively suppress the response of mixture components other than the analyte of interest in nonproximate-DESI. Flexible ion transport tubing is also investigated, allowing performance similar to stainless steel tubing in the transport of ions from the sample to the mass spectrometer. Transfer tube temperature effects are examined. A multiple sprayer DESI source capable of analyzing a larger sample area was evaluated to decrease the sampling time and increase sample throughput. Low nanogram detection limits were obtained for the compounds studied from a wide variety of surfaces, even those present in complex matrixes. Since the first experiments of Sir J. J. Thomson involving the use of the parabola mass spectrograph,1 mass spectrometry (MS) has evolved into one of the most powerful techniques for chemical * To whom correspondence should be addressed. Telephone: (765) 494-5262. Fax: (765) 494-9421. E-mail: [email protected]. (1) Thomson, J. J. Rays of Positive Electricity and Their Applications to Chemical Analysis; Longmans: London, 1913. 10.1021/ac0707939 CCC: $37.00 Published on Web 08/14/2007

© 2007 American Chemical Society

identification and characterization.2 Newer experimental methods, such as tandem mass spectrometry and ion/molecule reactions, allow the selective identification of analytes in complex matrixes, even at trace levels.3-7 Sample ionization in MS can be performed from the solution, gas, or condensed phase. Typically, ionization of analytes in the solution phase is performed by electrospray ionization (ESI),8 whereas gas-phase samples have been ionized by atmospheric pressure chemical ionization (APCI).9 Condensed phase samples, the most challenging group to ionize, have been successfully studied by several desorption ionization (DI) methods, especially secondary ion mass spectrometry;10 matrix-assisted laser desorption ionization (MALDI);11 and its atmospheric pressure analog, AP-MALDI.12 Recently, ambient ionization methods, such as desorption electrospray ionization (DESI),13 desorption atmospheric pressure chemical ionization (DAPCI),14 and direct analysis in real time (DART),15 have been introduced to allow ionization in the ambient environment without any sample preparation. Additional methods have recently been added to this group, including electrospray-assisted laser desorption ionization (ELDI)16 and atmospheric solid sampling probe (ASAP),17 which (2) McLuckey, S. A.; Wells, J. M. Chem. Rev. 2001, 101, 571-606. (3) Eberlin, M. N. J. Mass Spectrom. 2006, 41, 141-156. (4) Busch, K. L.; Glish, G. L.; McLuckey, S. A. Mass Spectrometry/Mass Spectrometry: Techniques and Applications of Tandem Mass Spectrometry; VCH Publishers, Inc.: New York, NY, 1988. (5) Cooks, R. G.; Glish, G. L. Chem. Eng. News 1981, 59, 40-52. (6) McLafferty, F. G. Tandem Mass Spectrometry; Wiley: New York, 1983. (7) Chen, H.; Chen, H. W.; Cooks, R. G. J. Am. Soc. Mass Spectr. 2004, 15, 998-1004. (8) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science 1989, 246, 64-71. (9) Lane, D. A. Environ. Sci. Technol. 1982, 16, A38-A46. (10) Pachuta, S. J.; Cooks, R. G. Chem. Rev. 1987, 87, 647-669. (11) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299-2301. (12) Laiko, V. V.; Baldwin, M. A.; Burlingame, A. L. Anal. Chem. 2000, 72, 652657. (13) Takats, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G. Science 2004, 306, 471-473. (14) Cotte-Rodriguez, I.; Taka´ts, Z.; Talaty, N.; Chen, H.; Cooks, R. G. Anal. Chem. 2005, 77, 6755-6764. (15) Cody, R. B.; Laramee, J. A.; Durst, H. D. Anal. Chem. 2005, 77, 22972302. (16) Shiea, J.; Huang, M. Z.; Hsu, H. J.; Lee, C. Y.; Yuan, C. H.; Beech, I.; Sunner, J. Rapid Commun. Mass Spectrom. 2005, 19, 3701-3704.

Analytical Chemistry, Vol. 79, No. 18, September 15, 2007 7069

also allow the analysis of samples in the condensed phase. These ambient ionization methods are revolutionizing the way mass spectrometry is performed. We focus on DESI, which is becoming a widely used technique for the selective and sensitive detection of samples at trace levels from ambient surfaces. In a typical DESI experiment, a pneumatically assisted electrospray is directed onto a surface bearing a sample of interest. Analyte ions generated by the interaction of charged solvent microdroplets with the neutral molecules on the surface are desorbed and sampled by a mass spectrometer.18 DESI has shown potential value as a trace detection method, showing high selectivity and sensitivity in analyzing explosives and chemical warfare agent (CWA) simulants at trace quantities on contaminated surfaces, even in complex matrixes.14,18 Other applications include direct analysis of active ingredients in pharmaceutical tablets,19-28 metabolites in urine,29 polymer analysis,30,31 intact, untreated bacteria,32 lipids in tissue (tissue imaging),33 and biomolecules.34-37 Increased selectivity is obtained by DESI using tandem mass spectrometry and by a variant of DESI called reactive-DESI.14,18 In this experiment, solution-phase ion/molecule reactions are performed by doping the solvent spray with neutral reagents that can undergo bond-forming reactions with an analytebearing surface. In addition to DESI, the DAPCI, DART, ELDI, and ASAP methods are being applied to in situ analysis with little or no sample preparation, but so far, they have been restrained by the requirement of close spatial proximity of the sample and mass (17) McEwen, C. N.; McKay, R. G.; Larsen, B. S. Anal. Chem. 2005, 77, 78267831. (18) Cotte-Rodriguez, I.; Chen, H.; Cooks, R. G. Chem. Commun. 2006, 953955. (19) Rodriguez-Cruz, S. E. Rapid Commun. Mass Spectrom. 2006, 20, 53-60. (20) Leuthold, L. A.; Mandscheff, J. F.; Fathi, M.; Giroud, C.; Augsburger, M.; Varesio, E.; Hopfgartner, G. Rapid Commun. Mass Spectrom. 2006, 20, 103110. (21) Kauppila, T. J.; Wiseman, J. M.; Ketola, R. A.; Kotiaho, T.; Cooks, R. G.; Kostiainen, R. Rapid Commun. Mass Spectrom. 2006, 20, 387-392. (22) Williams, J. P.; Scrivens, J. H. Rapid Commun. Mass Spectrom. 2005, 19, 3643-3650. (23) Williams, J. P.; Nibbering, N. M. M.; Green, B. N.; Patel, V. J.; Scrivens, J. H. J. Mass Spectrom. 2006, 41, 1277-1286. (24) Williams, J. P.; Lock, R.; Patel, V. J.; Scrivens, J. H. Anal. Chem. 2006, 78, 7440-7445. (25) Williams, J. P.; Patel, V. J.; Holland, R.; Scrivens, J. H. Rapid Commun. Mass Spectrom. 2006, 20, 1447-1456. (26) Weston, D. J.; Bateman, R.; Wilson, I. D.; Wood, T. R.; Creaser, C. S. Anal. Chem. 2005, 77, 7572-7580. (27) Nyadong, L.; Green, M. D.; De Jesus, V. R.; Newton, P. N.; Fernandez, F. M. Anal. Chem. 2007, 79, 2150-2157. (28) Ricci, C.; Nyadong, L.; Fernandez, F. M.; Newton, P. N.; Kazarian, S. G. Anal. Bioanal. Chem. 2007, 387, 551-559. (29) Kauppila, T. J.; Talaty, N.; Salo, P. K.; Kotiaho, T.; Kostiainen, R.; Cooks, R. G. Rapid Commun. Mass Spectrom. 2006, 20, 2143-2150. (30) Jackson, A. T.; Williams, J. P.; Scrivens, J. H. Rapid Commun. Mass Spectrom. 2006, 20, 2717-2727. (31) Nefliu, M.; Venter, A.; Cooks, R. G. Chem. Commun. 2006, 888-890. (32) Song, Y.; Talaty, N.; Tao, W. A.; Pan, Z.; Cooks, R. G. Chem. Commun. 2007, 61-63. (33) Wiseman, J. M.; Ifa, D. R.; Song, Q. Y.; Cooks, R. G. Angew. Chem., Int. Ed. 2006, 45, 7188-7192. (34) Cooks, R. G.; Ouyang, Z.; Takats, Z.; Wiseman, J. M. Science 2006, 311, 1566-1570. (35) Takats, Z.; Wiseman, J. M.; Cooks, R. G. J. Mass Spectrom. 2005, 40, 12611275. (36) Bereman, M. S.; Nyadong, L.; Fernandez, F. M.; Muddiman, D. C. Rapid Commun. Mass Spectrom. 2006, 20, 3409-3411. (37) Shin, Y.-S.; Drolet, B.; Mayer, R.; Dolence, K.; Basile, F. Anal. Chem. 2007, 79, 3514-3518.

7070

Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

spectrometer, which is inconvenient for applications in which access to the sample is limited. Previously, we reported the use of DESI for the detection of explosives and CWAs from ambient surfaces placed at distances of up to 3 m from the MS.38 In that system, ionization occurred in a stand-off manner, and the resulting ions were transported directly and continuously into the instrument for analysis. This stand-off (remote) sampling method is referred to as nonproximate DESI. In the present paper, we expand greatly on our previous report38 by performing ion/molecule reactions inside the ion transport tube (length up to 3 m) connecting the sample and the mass spectrometer. The high selectivity and sensitivity obtained by performing ion/molecule reactions in reactive DESI allows the nonproximate experiment to be used in ultratrace detection of compounds in complex matrixes. In some cases, reagent vapors were introduced into the ion transport tube to improve performance. Several systems were studied, including pharmaceutical drug formulations, organic salts, peptides, CWAs, and small organics. Illicit drugs were directly analyzed from skin with high selectivity and sensitivity and with no sample preparation. Controlled fragmentation and formation of stable adducts were achieved by changing the temperature of the transport tube up to 250 °C. It is shown that it is possible to perform high-throughput analysis by using a multiple DESI sprayer source capable of covering high surface areas. In addition, use of a flexible ion transport tube is shown to allow better control of the nonproximate-DESI device for applications where sample accessibility is limited. Nonproximate DAPCI is also evaluated as an alternative analysis from surfaces that are susceptible to damage by direct application of solvent. These experiments using a rudimentary nonproximate detection device lift the required spatial proximity of the sample and the mass spectrometer. Nonproximate DESI may expand the capability of mass spectrometry by increasing its flexibility in application areas including forensics, public safety, clinical diagnosis, and military applications (e.g., detection of improvised explosive devices). EXPERIMENTAL SECTION Single/Multiple Sprayer DESI Sources and Transport Tubing for Non-Proximate DESI. In nonproximate detection by DESI, the mass spectrometer and the DESI source are physically separated, but connected by a long stainless steel (SS) ion transport tube (1.8 mm i.d., 3.18 mm o.d., Swagelok, Solon, OH), which allows the transport of ions through air to the mass spectrometer after they leave the surface (Figure 1a). Flexible conductive silicone tubing, used for particle transport (4.83 mm i.d., 6.35 mm o.d., TSI Incorporated, Shoreview, MN) was also used as an ion transport tube. The length of the ion transport tube is normally in the range of 1-3 m. This transport tube is connected to the LTQ mass spectrometer by a precut transfer capillary with adapter (0.508 mm i.d., 1.59 mm o.d., 6 in. long, Small Parts. Inc, Miami Lakes, FL). The spray emitter is positioned at an angle of 35° to the normal and to the inlet of the ion transport tube (Figure 1a). Ion transport through air is assisted by the vacuum of the mass spectrometer. The transport tube has a control arm that allows rastering of the DESI source across large (38) Cotte-Rodriguez, I.; Cooks, R. G. Chem. Commun. 2006, 2968-2970.

Figure 1. Apparatus for DESI analysis up to 3 m away from the mass spectrometer. (a) The mass spectrometer and the DESI source are physically separated, but connected by a long stainless steel ion transport tube. A secondary inlet allows introduction of neutral reagent gas. (b) Multiple-sprayer setup used for high-surface-area screening. The high-surface-area multiple sprayer is composed of three DESI emitters that are welded together and separated from each other by a total distance of 12 mm. (c) DAPCI source used for nonproximate experiments.

surfaces. The N2 supply line and the solvent spray inner capillary for the DESI source are conveniently routed along the length of the transport tube. The solvent was sprayed at a flow rate of 3 µL/min with an applied high voltage of 5 kV. The area covered per sprayer at the optimized tip-to-surface distance of 2 mm was evaluated. Spot size experiments were performed by electrospraying methyl violet dye onto filter paper and measuring the spots sizes at tip-to-surface distances varying from 0.5 to 5 mm. A linear dependence between the spot size and tip-to-surface distance (R2 ) 0.9975) was observed, with a spot radius of 1.1 mm being attained at the optimum tip-to-surface distance of 2 mm, giving a total surface area covered per sprayer of 4 mm2. To overcome this limitation in total surface area coverage, a high surface area multiple sprayer was built. The high surface area multiple sprayer consists of three individual DESI emitters arranged linearly and mounted in three 1/16-in. Swagelok SS tees welded together and separated from each other by a total distance of 12 mm (Figure 1b). Desorbed analyte ions were gathered from the surface by an aluminum transfer funnel (Figure 1b), allowing the collection of ions generated from all three sprayers into the transport tube, assisted by the suction generated by the vacuum of the mass spectrometer. Temperature Effects, Ion/Molecule Reactions, and NonProximate DAPCI. Temperature studies were performed by equipping a 1-m SS ion transport tube with a controllable heating element and thermocouple (Omega Engineering Inc., Stamford, CT), allowing accurate temperature measurement and control up to 250 °C. Gas-phase ion/molecule reaction experiments were carried out by welding a small section of SS transport tube (1.8 mm i.d., 3.18 mm o.d., 2 in. length, Swagelok) to the main stainless steel ion transport tube (1.8 mm i.d., 3.18 mm o.d., 1 m in length, Swagelok), allowing introduction of a reagent gas as shown in Figure 1a. This secondary transport inlet was welded at an angle of 35° with respect to the main ion transport tube and

90° to the DESI source. This setup allowed a stream of DESIgenerated ions and a neutral reagent gas to interact throughout the length of the transfer tube. For nonproximate DAPCI experiments, the DESI source was simply replaced by a DAPCI source14 (Figure 1c). The gaseous reagent molecules (methanol, water, etc.) used in DAPCI are delivered by a stream of nitrogen gas and ionized by a corona discharge formed by applying a high voltage, typically 5 kV, to a thin, SS needle. Reagent Chemicals and Analytes Studied. Trifluoroacetic acid (TFA), 2,4,6-triphenylpyridine, nicotine, bradykinin, loratadine, naproxen, aspirin, ibuprofen, acetaminophen, caffeine, benzoic acid, and benzophenone were purchased from Sigma-Aldrich (Milwaukee, WI). Methanol (HPLC grade) and sodium chloride were purchased from Mallinckrodt Baker, Inc. (Phillipsburg, NJ). Deionized water was obtained from a Barnstead/Thermolyne deionizer unit (Barnstead Mega-Pure System, Dubuque, IA). Dimethyl methylphosphonate (DMMP) was purchased from Lancaster Synthesis (Pelham, NH). Over-the-counter tablets (Claritin, Excedrin, Alleve, acetaminophen, aspirin, and ibuprofen) were obtained from local pharmacies. Cocaine and mixtures of illicit drugs containing D-methamphetamine, cocaine, and diacetylmorphine were purchased as 1 mg/mL solutions (in methanol) and 250 µg/mL solutions (in acetonitrile), respectively, from Alltech-Applied Science Labs (State College, PA). Ammonia gas was purchased from Scott Specialty gases (474 ppm in nitrogen, Plumsteadville, PA). Ammonium acetate (NH4AOC) was purchased from Fisher Scientific (Fair Lawn, NJ). The organic salt 1,2,4,6-tetraphenylpyridinium perchlorate was obtained from the Florida Center for Heterocyclic Compounds (Gainesville, FL). Standard solutions were prepared in methanol at the desired concentration. From each solution, 10 µL was pipetted onto the desired surface with 1 cm2 total surface area, unless otherwise specified, and then allowed to dry prior analysis. Pharmaceutical drug formulations and tablets were directly analyzed without any sample preparation. Mass spectra for all compounds were generAnalytical Chemistry, Vol. 79, No. 18, September 15, 2007

7071

Figure 2. Positive ion nonproximate DESI mass spectrum of Claritin tablets placed at a distance of (a) 2 mm, (b) 1 m (c), and 3 m from the mass spectrometer. Methanol/water (50:50) was used as the spray solvent. The active ingredient loratidine gives the dominant peak at all sample distances. Absolute ion intensities for m/z 383 were 5.01 × 104 (2 mm), 2.75 × 102 (1 m), and 2.03 × 101 intensity units (3 m).

ated using methanol/water (50:50) as the spray solvent, sometimes doped with 10 mM NaCl or NH4AOC to enhance the selectivity and sensitivity by stable adduct formation. For DAPCI experiments, a mixture of methanol/water (50:50) vapor in nitrogen was used as reagent gas. Compound structures, molecular weights, suppliers, commercial names, and active ingredients of commercial drug formulations can be found as Supporting Information (Scheme S-1). Mass analysis was performed with a ThermoFinnigan LTQ mass spectrometer. RESULTS AND DISCUSSION Non-Proximate DESI of Active Ingredients in Pharmaceutical Drug Formulations. Loratadine, the active ingredient in Claritin tablets, is a long-acting tricyclic antihistamine that blocks the H1-receptor in cells.39,40 Figure 2 shows the positive ion DESI spectrum of a Claritin tablet when the sample is placed at (a) 2 mm (b) 1 m and (c) 3 m from the mass spectrometer inlet. The experiment was performed using methanol/water (50: 50) as the spray solvent. The DESI mass spectrum at 2 mm for Claritin is dominated by the protonated molecule of loratadine, observed as a peak doublet at m/z 383 and 385 due to chlorine (Figure 2a). Tandem MS was used to confirm the identity of the ion at m/z 383: it showed a characteristic fragment at m/z 337 (100% relative abundance) corresponding to the loss of ethanol from the ethyl ester side chain.41,42 Remarkably, as the sample/mass spectrometer distance is increased from 2 mm (Figure 2a) to 1 m (Figure 2b), the signal(39) Piwinski, J. J.; Wong, J. K.; Chan, T. M.; Green, M. J.; Ganguly, A. K. J. Org. Chem. 1990, 55, 3341-3350. (40) Yang, L. Y.; Wu, N.; Rudewicz, P. J. J. Chromatogr., A 2001, 926, 43-55. (41) Ramanathan, R.; Su, A. D.; Alvarez, N.; Blumenkrantz, N.; Chowdhury, S. K.; Alton, K.; Patrick, J. Anal. Chem. 2000, 72, 1352-1359. (42) Mueller, C. A.; Weinmann, W.; Dresen, S.; Schreiber, A.; Gergov, M. Rapid Commun. Mass Spectrom. 2005, 19, 1332-1338.

7072 Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

to-noise (S/N) ratio improves because of the greater reduction in the background ion signal relative to that for the analyte during transport through the tube to the mass spectrometer. Two orders of magnitude in signal intensity are lost for the ion at m/z 383 at 1 m (2.75 × 102 intensity units), as compared to that at 2 mm (5.01 × 104 intensity units). At 3 m (Figure 2c), the only species that survives transport through the tube to the mass spectrometer is protonated loratadine. At this distance, no interfering or matrix ions were observed. A further order of magnitude decrease in signal intensity (2.03 × 101 intensity units) was observed at 3 m for the ion at m/z 383 [loratadine + H]+ as compared to the absolute intensity recorded at 1 m. The high proton affinity of loratadine decreases the probability of proton loss by the ion of interest during transport to the mass spectrometer, increasing its abundance as compared to that of the background ions. Limits of detection at sample-to-mass spectrometer distances of 1 and 3 m for Claritin and other pharmaceuticals are included in Table S-1 in the Supporting Information. A summary of CID data for active ingredients in pharmaceutical drug formulations studied is given in Table S-2. To test the precision of the nonproximate DESI technique, five separate mass spectra were generated for the detection of loratadine from a single Claritin tablet at 1 m from the mass spectrometer inlet. The average peak height for the protonated loratadine molecule was calculated to be 2.83 × 102 intensity units, with a relative standard deviation of 7.5%. Further discussion of precision and the time dependence of signal intensity is provided in the Supporting Information. Ibuprofen, a nonsteroidal antiinflammatory drug (NSAID) with an acidic functional group (ibuprofen pKa ) 4.41),43 readily forms the deprotonated molecular ion that can be easily detected in the negative ion mode. This is advantageous because negative ionization is highly selective for compounds that have either high electron affinities or acidic groups. Negative ion DESI mass spectra of an Advil tablet containing 200 mg of the active ingredient ibuprofen were evaluated at sample/mass spectrometer distances of 2 mm, and 1 m, 3 m. Deprotonated ibuprofen, as well as the deprotonated dimer and sodiated dimer species, were observed at m/z 205 [ibuprofen H]-, m/z 411 [2ibuprofen - H]-, and m/z 433 [2ibuprofen - 2H + Na]-. Tandem MS of the ion at m/z 205 showed an abundant product ion at m/z 161, which corresponds to the loss of CO2, the primary fragmentation route of dissociation for ibuprofen.22,23 The dimer of ibuprofen is the base peak in the spectrum at m/z 411, which can be attributed to the low-energy deposition characteristic of DESI and the relatively large amount of analyte in this sample. The sodiated dimer at m/z 433 was previously observed by LC-MS-ESI using methanol/water/acetonitrile as the mobile phase.43 Possible sources of sodium are the water used in the spray solvent and sodium salts in the inert matrix. Contrasting the spectrum recorded at 2 mm with that obtained at 1 m, the greater distance resulted in a decrease in ion abundance by 2 orders of magnitude, but also a significant reduction in chemical noise. At a distance of 3 m, analyte ions were the only species surviving transport through the tube into the mass spectrometer, yielding a spectrum that lacks all signal due to background ions. At this distance, an additional order of (43) Schug, K.; McNair, H. M. J. Sep. Sci. 2002, 25, 760-766.

magnitude decrease in signal was observed, but S/N ratios of over 100 were observed. For the experimental data, refer to Figure S-1 in the Supporting Information. Alleve tablets were also evaluated in the negative ion mode by nonproximate-DESI. At a distance of 1 m, the deprotonated form of the active ingredient naproxen at m/z 229 [naproxen H]- and the deprotonated dimer at m/z 459 (data not shown) were seen. The main fragmentation route for the ion at m/z 229 was by loss of 44 mass units, corresponding to the loss of CO2, giving a product ion at m/z 185.23 The dimer fragmented by losing one naproxen molecule as a neutral, giving back deprotonated naproxen at m/z 229 (100% relative abundance) and two other fragments at m/z 185 (90% relative abundance) and m/z 170 (55% relative abundance).23 The capabilities of nonproximate DESI in the detection of multiple active ingredients in an over-the-counter drug formulation were evaluated by analyzing an Excedrin Migraine tablet, which contains 250 mg each of acetaminophen and aspirin, as well as 65 mg of the stimulant caffeine. As previously observed for loratadine and ibuprofen, the DESI mass spectrum taken from this tablet showed a marked reduction in background/formulation matrix signals as the sample-to-mass spectrometer distance was increased from 2 mm up to 3 m (See Figure S-2 in Supporting Information). Protonated aspirin, acetaminophen, caffeine, and monomeric and dimeric adducts formed with sodium and ammonium were observed when the tablet was placed 2 mm from the mass spectrometer inlet. At 1 m, the nonproximate DESI mass spectrum was dominated by protonated aspirin and acetaminophen and corresponding adducts, whereas at 3 m, only sodium adducts were observed. In separate experiments, Excedrin tablets were analyzed by nonproximate DESI to selectively detect aspirin, acetaminophen, and caffeine by doping the solvent spray with sodium chloride (NaCl), ammonium acetate (NH4OAC), or mixtures of both (NaCl/NH4OAC) to increase the selectivity and instrument response. When the solvent spray was doped with NaCl or NH4OAC, only sodium or ammonium adducts and protonated species were observed in the mass spectrum, showing an increased response for all species, as compared to the results obtained without using any dopants (Figure S-3 in Supporting Information). In addition to detecting active ingredients in pharmaceutical drug formulations, we also evaluated the performance of nonproximate DESI in analysis of other small organic compounds in both positive and negative ion detection modes. Figure 3 shows DESI spectra for 10 ng of the organic salt 1,2,4,6-tetraphenylpyridinium perchlorate deposited on paper (1 cm2) with the sample 2 mm and 1 m from the mass spectrometer. In the positive ion mode, this organic salt showed an abundant ion at m/z 384, corresponding to the intact tetraphenylpyridinium [C29H22N]+ cation (Figure 3a). As seen in Figure 3b, no signal corresponding to background was detected at 1 m. The perchlorate (ClO4)- anion of the organic salt was detected in the negative ion detection mode at m/z 99 at both 2 mm and 1 m distances. Benzoic acid was also studied by nonproximate DESI. Like ibuprofen, benzoic acid has an acidic functional group (benzoic acid pKa ) 4.20), allowing its detection in the negative ion mode as the deprotonated molecule. The negative ion nonproximate DESI spectrum for 10 ng of benzoic acid deposited on paper

Figure 3. Positive ion nonproximate DESI mass spectrum of 10 ng of 1,2,4,6-tetraphenylpyridinium perchlorate (total amount deposited) on 1 cm2 of paper located (a) 2 mm and (b) 1 m from the mass spectrometer. Methanol/water (50:50) was used as the spray solvent. Absolute ion intensity for m/z 384 (2 mm) was 1.21 × 104 intensity units, with a loss of 2 orders of magnitude at the 1-m length.

(1 cm2) at separation distances of (a) 2 mm and (b) 1 m can be observed in Figure S-5 in the Supporting Information. The base peak in the mass spectrum at 1 m corresponds to the deprotonated molecular ion at m/z 121, [C7H6O2 - H]-, and there is a marked reduction in background/matrix signal as compared to the spectrum obtained at 2 mm. Non-Proximate DESI of Illicit Drugs on Skin. The rapid and accurate detection of illicit drugs is a topic of increasing concern, and detection of these species from unconventional surfaces, such as skin, could prove useful to the fields of law enforcement and forensics. Nonproximate DESI of drugs of abuse was demonstrated from skin, conducted under a human subject protocol approved by the Institutional Review Board of Purdue University. Figure 4 shows the mass spectrum for 1 ng of cocaine deposited on the index finger of a test subject, recorded with the person (a) 2 mm and (b) 1 m from the mass spectrometer inlet. Using a mixture of methanol/water (50:50) as the spray solvent, protonated cocaine at m/z 304 appears as the base peak in the mass spectrum. Tandem MS confirmed the identity of this ion, with the product ion spectrum showing a characteristic fragment at m/z 182 formed by loss of benzoic acid.44-46 A similar experiment was performed on a mixture of three drugs of abuse (Dmethamphetamine, cocaine, and diacetylmorphine). The mixture (1 ng of each component) was deposited on skin (index finger) in a total area of 1 cm2, and the only ions observed at 1 m from the mass spectrometer corresponded to protonated methamphetamine, cocaine, and diacetylmorphine at m/z 150 (35% relative abundance), m/z 304 (100% relative abundance), and m/z 370 (10% relative abundance), respectively (data not shown). Tandem mass spectrometry corroborated the identities of these ions. (44) Hows, M. E. P.; Lacroix, L.; Heidbreder, C.; Organ, A. J.; Shah, A. J. J. Neurosci. Methods 2004, 138, 123-132. (45) Needham, S. R.; Jeanville, P. M.; Brown, P. R.; Estape, E. S. J. Chromatogr., B 2000, 748, 77-87. (46) Jeanville, P. M.; Estape, E. S.; Needham, S. R.; Cole, M. J. J. Am. Soc. Mass Spectrom. 2000, 11, 257-263.

Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

7073

Figure 4. Positive ion nonproximate DESI mass spectrum from skin containing 1 ng of cocaine deposited over a total area of 1 cm2. Finger was placed at a distance of (a) 2 mm and (b) 1 m from the mass spectrometer. Methanol/water (50:50) was used as the spray solvent. Absolute ion intensity for m/z 304 (2 mm) was 1.64 × 104 intensity units, with a loss of ∼2 orders of magnitude at the 1-m length.

Figure 5. Nonproximate DESI mass spectrum for 10 ng of bradykinin deposited on Teflon in an area of 1 cm2 using nitrogen or air as nebulizing gas, with the sample placed 2 mm (a and c) and at 1 m (b and d) from the mass spectrometer. Absolute ion intensities for m/z 531 at 2 mm with air and nitrogen were 1.22 × 104 and 1.96 × 104 intensity units, respectively, with a loss of ∼2 orders of magnitude at the 1-m length.

Peptide Analysis by Non-Proximate DESI Using Air for Nebulization. To investigate the detection of larger molecules and to achieve a better understanding of their nonproximate DESI behavior during atmospheric pressure ion transport, a simple experiment was performed in which the peptide bradykinin was analyzed using air as a substitute for nitrogen for nebulization. Figure 5 shows the nonproximate DESI mass spectrum for 10 ng of bradykinin deposited on Teflon in an area of 1 cm2 and examined using nitrogen (Figure 5a and 5b) and air (Figure 5c and 5d) as nebulizing gas. For the doubly charged molecule at m/z 531, the absolute ion intensities at 2 mm for nitrogen and air were 1.22 × 104 and 1.96 × 104 ion intensity units, respectively. 7074

Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

As was the case for other systems in this study, the background was drastically reduced in the 1-m experiment, which showed only protonated bradykinin and the doubly charged molecule (Figure 5b and 5d), both with S/N ratios of more than 100. At 1 m, the singly charged species is favored over the doubly charged molecule, possibly due to partial deprotonation of the latter ions as they travel back through the transfer tube to the mass spectrometer. No significant differences were observed between using nitrogen or air as the nebulizing gas, which is of interest for field applications, allowing the use of a simple air compressor instead of the compressed nitrogen gas cylinders typically used in DESI. Non-Proximate DESI of Chemical Warfare Agent Simulants via Adduct Formation. The chemical warfare agent simulant DMMP was detected in the positive ion mode by its sodiated adducts at m/z 147 [DMMP + Na]+ and m/z 271 [2DMMP + Na]+. In an experiment in which 0.5 ng of DMMP was deposited in an area of 1 cm2 on paper and the sample placed at a distance of 2 mm from the mass spectrometer, a high-quality spectrum was observed when using an aqueous NaCl spray solution to enhance the selectivity and instrumental response for these adduct ions (See Supporting Information, Figure S-6). The above ions were also the most significant species in the mass spectrum recorded at a distance of 1 m, whereas at 3 m, where even more stabilizing and reactive ion/molecule collisions occur, the sodiated dimer at m/z 271 was the only species observed. A marked reduction in the background ion population is observed at 1 m, as compared to the mass spectrum taken at 2 mm from the mass spectrometer, providing a high quality spectrum in spite of the decrease of 2 orders of magnitude in absolute abundance for m/z 271, as compared to the absolute abundance obtained at 2 mm. Neither background nor fragment ions were observed in the nonproximate-DESI mass spectrum of DMMP taken at 3 m. Nonproximate DESI with Ion Transport via a Flexible Transport Tube. In a typical nonproximate DESI experiment, the ion transport tube used to collect and transport ions from the surface of the sample to the mass spectrometer is a straight, SS tube. This tube allows effective ion transport even when bent by up to 45 degrees. A straight tube is not appropriate for some applications, such as in the analysis of large, geometrically complex objects. For this reason, flexible ion transport tubes were evaluated. Figure 6 shows the positive nonproximate DESI mass spectrum of 15 ng of loratadine deposited on paper (total area of 1 cm2) placed at both 2 mm and 1 m from the mass spectrometer inlet. The performance of the flexible conductive tubing was similar to that obtained with the SS tubing. The absolute signal intensity at 2 mm (typical DESI experiment) was 2.5 × 104, decreasing approximately 2 orders of magnitude when the distance was increased to 1 m using either the SS ion transfer tube (5.1 × 102) or conductive tubing (4.3 × 102). Temperature Effects on Non-Proximate DESI. The nonproximate DESI responses for Excedrin and DMMP were evaluated at temperatures ranging from ambient up to 250 °C. At 100250 °C, it is possible to remove signals due to molecules other than the analytes of interest, possibly due to enhanced thermal dissociation of background ions. At an operating distance of 1 m and temperatures above 100 °C, aspirin and acetaminophen showed the formation of fragment ions via water and ketene loss,

Figure 6. Positive ion nonproximate DESI mass spectrum of loratadine (15 ng deposited on paper, total surface area of 1 cm2) examined (a) 2 mm and (b) 1 m from the mass spectrometer. Absolute ion intensities for m/z 383 were 2.5 × 104 (2 mm) and 4.3 × 102 intensity units (1 m).

Figure 7. Positive ion nonproximate DESI mass spectra for background ions and 10 ng of 2,4,6-triphenylpyridine deposited on 1 cm2 of paper. Sample placed 1 m from the mass spectrometer with either (a and c) ambient air or (b and d) ammonia gas being introduced into the ion transport tube. Methanol/water (50:50) was used as the spray solvent for DESI.

showing that heating must be carried out with some care. The data, proposed fragmentation mechanisms, and discussion appear as Supporting Information. Gas-Phase Ion-Molecule Reactions with Non-Proximate DESI. The nonproximate-DESI setup not only allows the study of temperature effects and the behavior of ions traveling long distances at near atmospheric pressure in the ion transport tube, but it also allows the study of these ions when neutral reagents are added into the ion transport tube to react via ion/molecule reactions. To explore this possibility, the nonproximate DESI setup was modified by adding an additional inlet to the ion transport tube (see the Experimental Section for details). Figure 7 shows the DESI spectrum for background ions when no analyte was present (Figure 7a and 7b) and for 10 ng of 2,4,6-triphenylpyridine (Figure 7c and 7d) at a distance of 1 m from the mass

spectrometer (sample deposited on paper in an area of 1 cm2), before (Figure 7a and 7c) and after adding neutral ammonia (Figure 7b and 7d) to the ion transport tube. As expected, species with lower proton affinities than ammonia (PA ) 853.6 kJ/mol)47 are removed by proton transfer to ammonia, resulting in a simpler background spectrum. This can be observed by comparing Figure 7a (before NH3 addition) and 7b (after NH3 addition). After NH3 addition to the ion transport tube, only two main ions contribute significantly to the background spectrum, those at m/z 256 and 102, the latter being the base peak with an absolute intensity of 4.70 × 101 instrument counts. When the same experiment was performed using 10 ng of 2,4,6triphenylpyridine deposited on paper in an area of 1 cm2, the mass spectrum before ammonia addition consisted of multiple peaks corresponding to several background ions and protonated 2,4,6triphenylpyridine at m/z 308 (Figure 7c). By adding ammonia in the ion transport tube (Figure 7d), the chemical noise was dramatically reduced, and the spectrum showed only protonated 2,4,6-triphenylpyridine at m/z 308 and a few background ions that contributed 5% to the total abundance relative to the sample ion at m/z 308. As expected, the absolute intensity of the ion at m/z 308 remained constant before (3.74 × 102 instrument counts) and after (5.27 × 102 instrument counts) addition of ammonia gas because of its high proton affinity relative to ammonia. Mixtures having two constituents varying in proton affinities were also examined by this method (see Figure S-9 in the Supporting Information). Multiple Sprayer DESI Source for Non-Proximate Analysis of Large Surface Areas. As observed in previous sections, DESI sources consisting of a single spray emitter have been used for remote detection of compounds at trace levels on ambient surfaces. When individual samples of interest have small surface areas, it is possible to complete analyses in a high-throughput fashion, but if large objects are to be examined, the total analysis time required to cover larger surface areas can be substantial. To decrease this time, a high surface area, multiple DESI sprayer source was built. Preliminary experiments were performed to optimize the area covered per spray emitter (spot size) and the ion transfer efficiency. In a typical DESI experiment, the area covered per sprayer is 4 mm2. By using the multiple sprayer, it was possible to cover an area of 72 mm2 by rastering linearly six times at intervals of 2 mm (36 mm length × 2 mm width ) 72 mm2), seen in Figure 1b. Larger areas can be covered by increasing the spray-to-surface distance (distances greater than 2 mm), but this is at the expense of sensitivity, which is decreased by the spray divergence. The coverage area by the high surface area multiple sprayer compared to that of a single sprayer was evaluated. Simulating an analyte surface of increased size, 10 ng of nicotine (total amount on paper) was deposited along a line covering a total area of 72 mm2. Figure 8 shows the response obtained using single (Figure 8a and 8b) and multiple (Figure 8c and 8d) sprayers with the sample 2 mm from the inlet to the mass spectrometer (Figure 8a and 8c) and 1 m (8b and 8d) from the instrument. The base peak in all four cases was the protonated nicotine molecule at m/z 163, but analysis using the multiple sprayer source yielded (47) Syracuse Research Corporation Physical Property Database, S. R. C., Syracuse, NY, 2003; http://www.syrres.com/esc/physdemo.htm.

Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

7075

Figure 8. Nonproximate DESI spectra using a single sprayer or multiple sprayers to examine 10 ng of nicotine deposited on paper over an area of 72 mm2. The single and multiple sprayers tips were placed at (a and c) 2 mm or (b and d) 1 m from the mass spectrometer. Absolute ion intensities for m/z 163 were 2.59 × 104 (single sprayer) and 2.99 × 104 intensity units (multiple sprayers), with a loss of ∼2 orders of magnitude for each at the 1-m length.

Figure 9. Nonproximate positive ion DAPCI spectrum of 10 ng of DMMP deposited on 1 cm2 of paper and examined at a distance of (a) 2 mm and (b) 1 m from the mass spectrometer. A mixture of methanol/water vapors was used as the reagent gas for DAPCI. Absolute ion intensities for m/z 125 were 1.73 × 104 intensity units, with a loss of ∼2 orders of magnitude at the 1-m length.

not only a higher signal intensity but also a considerably shorter total analysis time. Tandem MS of this ion showed that the main fragmentation under CID conditions is loss of methylamine, giving a product ion at m/z 132 that later loses H2 to form an ion at m/z 130.23 By using the single and multiple sprayers at a distance of 1 m (Figure 8b and 8d) from the mass spectrometer, the only ion observed corresponds to protonated nicotine. Non-Proximate DAPCI. Desorption atmospheric pressure chemical ionization (DAPCI) is an alternative ambient ionization method in which solvent does not come into contact with the sample surface. Instead, gaseous reagent ions (methanol, water, 7076 Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

etc.) are employed.14 Ion formation in DESI has been suggested to follow three different mechanisms: charge transfer, dropletpickup, and gas-phase ionization through proton/electron transfer or other ion/molecule reactions at atmospheric pressure, depending on the conditions used and the particular analyte.13,14 In DAPCI, the droplet-pickup mechanism is precluded, which leaves charge transfer (at the surface) and gas-phase ionization as the only two possible ionization mechanisms. Nonproximate DAPCI mass spectra for 10 ng of DMMP deposited on 1 cm2 of paper at 2 mm and 1 m from the mass spectrometer are shown in Figure 9a and 9b, respectively. A mixture of methanol/water (50:50) was introduced into a nitrogen flow and used as the reagent gas. Only the protonated molecule [DMMP + H]+ was observed for DAPCI analysis. Using DAPCI, a better response was obtained for the ion at m/z 125 (1.73 × 104 instrument counts, Figure 9a) than with the DESI method (1.10 × 104 instrument counts, data not shown) under similar experimental conditions. Both charge transfer and gas-phase ionization mechanisms are possible in the case of DMMP (DMMP vapor pressure ) 0.962 mmHg).47 CONCLUSIONS This set of experiments establishes that nonproximate DESI, analyses in which there is spatial separation between the sample and the mass spectrometer, is a promising technique. Data presented suggests that certain analyte ions survive passage through the ion transport system better than background ions, sample matrix, and fragment ions that are not thermodynamically favored. The distance-dependent phenomenon observed here is analogous to the time-dependent behavior typically observed in atmospheric pressure chemical ionization, in which thermodynamically stable ions increasingly come to dominate the mass spectra at long ionization times. In APCI, as here, the spectra containadductsoftheanalyteformedviaion/moleculeassociation.48-50 One way to further increase the selectivity of DESI is to add into the spray solvent a reagent that has a high affinity for the analyte of interest. An alternative is to achieve a similar result by allowing ion/molecule reactions with a reagent gas introduced into the ion transport tube. Analytes of high proton affinity will compete effectively for the charge in these experiments, allowing their discrimination from background and matrix species. This gives high selectivity and high S/N ratios in experiments performed at atmospheric pressure and ambient temperature; S/N ratios of more than 100 were observed at distances of 3 m from the mass spectrometer in virtually all cases studied. Alternatively, increasing the temperature in the transport tube can also result in a reduction in interferences, although care must be taken to avoid thermal fragmentation. The present approach may have significant applications in security and related areas. A portable51 or miniature52 mass spectrometer coupled to nonproximate DESI and operated at (48) Harrison, A. G. Chemical Ionization Mass Spectrometry, 2nd ed.; CRC Press: Boca Raton, FL, 1992. (49) Dougherty, R. C. Anal. Chem. 1981, 53, A625-A636. (50) Tannenbaum, H. P.; Roberts, J. D.; Dougherty, R. C. Anal. Chem. 1975, 47, 49-54. (51) Mulligan, C. C.; Talaty, N.; Cooks. R. G. Chem. Commun. 2006, 17091711. (52) Gao, L.; Song, Q.; Patterson, G. E.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2006, 17, 5994-6002.

distance could be used for in situ analysis. The possibility of using air instead of nitrogen as the nebulizing gas also points toward in situ applications. High-throughput analysis can also be achieved by using a high-surface-area multiple sprayer, allowing for possible applications in pharmaceutics, such as cleaning validation and content uniformity. Clinical diagnosis and drug screening are other areas that can be benefit from this method through in vivo analysis of specific biomarkers, proteins related to diseases, or specific drugs and their metabolites.

ACKNOWLEDGMENT This work was funded by the Office of Naval Research, Grant No. N00014-05-1-0454. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review April 19, 2007. Accepted July 10, 2007. AC0707939

Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

7077