Anal. Chem. 2000, 72, 5055-5062
High-Pressure Ion Source Combined with an In-Axis Ion Trap Mass Spectrometer. 1. Instrumentation and Applications J. C. Mathurin, T. Faye, A. Brunot, and J. C. Tabet*
Laboratoire de Chimie Structurale Organique et Biologique, CNRS UMR 7613, Universite´ Pierre et Marie Curie, 4 place Jussieu Boite 45, 75252 Paris Cedex 05, France G. Wells
Varian Chromatography Systems, 2700 Mitchell Drive, Walnut Creek, California 94598 C. Fuche´
CREL, Ministe` re de l’Inte´ rieur, 168 route de Versailles, 78150 Le Chesnay, France
A new combination of a dual EI/CI ion source with a quadrupole ion trap mass spectrometer has been realized in order to efficiently produce negative ions in the reaction cell. Analysis of volatile compounds was performed under negative ion chemical ionization (NICI) during a reaction period where selected reactant negative ions, previously produced in the external ion source, were allowed to interact with molecules, introduced by hyphenated techniques such as gas chromatography. The O2•-, CH3O-, and Cl- reactant ions were used in this study to ensure specific ion/molecule interactions such as proton transfer, nucleophilic displacement, or charge exchange processes, respectively leading to even-electron species, i.e., deprotonated [M - H]- molecules, diagnostic [M - R]- ions, or odd-electron M-• molecular species. The reaction orientation depends on the thermochemistry of reactions within kinetic controls. First analytical results are presented here for the trace-level detection of several contaminants under NICI/Cl- conditions. Phosphoruscontaining compounds (malathion, ethyl parathion, and methyl parathion as representative for pesticides) and nitro-containing compounds (2,4,6-trinitrotoluene for explosive material) have been chosen in order to explore the analytical ability of this promising instrumental coupling. The use of mass spectrometry in negative ion mode is well recognized as a highly sensitive and selective tool for the analysis of volatile organic compounds, especially in the environmental and forensic sciences.1-3 Nevertheless, low-pressure negative ion * Corresponding author: (phone) (33) 01 44 27 32 63; (fax) (33) 01 44 27 38 43; (e-mail)
[email protected]. (1) Budzikiewicz, H. Mass Spectrom. Rev. 1986, 5, 345. (2) Advances in Physical Organic Chemistry; Nibbering, N. N., Ed.; Academic Press: New York, 1988; Vol. 24, p 1. (3) Harrison, A. G. Chemical Ionization Mass Spectrometry, 2nd ed.; CRC Press: Boca Raton, FL, 1992. 10.1021/ac000171m CCC: $19.00 Published on Web 09/09/2000
© 2000 American Chemical Society
mass spectrometer has never been really successfully used compared to its positive ion mode.4-6 Indeed, the production of low-energy electrons, necessary for the negative ion formation via resonance electron capture (EC), is mainly impeded by the lack of thermalizing gas. Much of the progress in the use of mass spectrometry in negative ion mode has taken place in the past three decades, due to the development of new techniques that enhanced the performance of this mode. Indeed, high-pressure chemical ionization (CI) conditions, typically ranging from 0.1 to 1 Torr (1 Torr ) 133.3 Pa) allowed the improvement of the situation markedly. A resurgence of interest in the negative ion mode began with the advent of a simple and convenient means of generating, in the ion source, a large number of electrons with near thermal kinetic energy. Adopting the method of CI, introduced by Munson and Field,7 this was achieved by Dougherty and Weisenberger8 and later by Hunt et al.9 Under high-pressure conditions, a buffer gas was employed as a collisional moderator for both the initially energetic electrons, as a means of kinetic energy cooling, and the odd-electron negative ions formed by EC to remove the amount of excess internal energy via a three-body attachment. Even with appropriate instrumental modifications, further limitations to the analytical application of negative ion mode were observed, particularly for EC processes. The reproducibility of negative ion chemical ionization (NICI) mass spectra was greatly affected by the experimental conditions, i.e., the ion source (4) Melton, C. E. Mass Spectrometry of Organic Ions; McLafferty, F. W., Ed.; Academic Press: New York, 1963. (5) Bowie, J. H.; Williams, B. D. MTP International Review of Science; Maccoll, A., Ed.; Physical Chemistry Series Two; Butterworths: London, 1975; Vol. 5, p 89. (6) Jennings, K. R. Gas-Phase Ion Chemistry; Bowers, M. T., Ed.; Academic Press: New York, 1979. (7) Munson, M. S. B.; Field, F. H. J. Am. Chem. Soc. 1966, 88, 2621. (8) Dougherty, R. C.; Weisenberger, C. R. J. Am. Chem. Soc. 1968, 90, 6570. (9) Hunt, D. F.; Stafford, G. C., Jr.; Crow, F. W.; Russell, J. W. Anal. Chem. 1976, 48, 2098.
Analytical Chemistry, Vol. 72, No. 20, October 15, 2000 5055
temperature,10,11 the moderator gas pressure,12 and the analyte concentration.13-15 Lens potential and repeller conditions were also incriminated and were instrument dependent.16 Furthermore, competitive reactions may be generated in the ion source under these high-pressure NICI conditions such as EC and ion/molecule reactions. Nevertheless, the ability to form even-electron negative ions (e.g., [M - H]-) has proved to be useful in many analytical applications, since quasi-molecular species often show greater stability than the corresponding positive ions.3 Especially, its internal energy related to the reaction exothermicity is dissipated in the counterpart neutral. Thus, the weak internal energy of these quasi-molecular species allows enhancement of the isomeric and diastereoisomeric structural differentiation.17 Charge location avoids the competitive decompositions promoted by the deprotonation site, as available from MH+ ion. Furthermore, if oddelectron negative ions are also more stable than the corresponding positive species, they present many fragmentations, as well as electron detachment, leading to decrease in sensitivity. Mass spectrometry analysis of volatile compounds mixture is widely carried out through hyphenated techniques, such as gas chromatographic mass spectrometry (GC/MS). Among the commercial benchtop mass spectrometers, obviously the most flexible instrument offering high sensitivity is the quadrupole ion trap mass spectrometer (QITMS).18 With its low cost and ruggedness, the QITMS system can be easily operated alternatively between electron ionization (EI) and low-pressure CI in positive ion mode, even during a single injection step. Additional features related to the “in time” analytical scanning, allowing sequential MSn experiments, are very useful for structural elucidation, particularly for distinguishing isomers, for instance.19 Moreover, GC/MS/MS offers a third dimensional separation feature, permitting the detection of target analytes from complex matrixes that may be identified and quantified within ultr trace levels.20 To combine these versatile operation modes of QITMS with the analytical performance of the negative ion mode, several studies have been performed on the in situ formation of such ions.21-24 During ionization processes available under conventional EI conditions (noted as in situ ionization related to internal (10) Miwa, B. J.; Garland, W. A.; Blumenthal, P. Anal. Chem. 1981, 53, 793. (11) Brumley, W. C.; Nesheiu`, S.; Trucksess, M. W.; Dreifuss, P. A.; Roach, J. A. G.; Andrezejewski, D.; Eppley, R. M.; Pohland, A. E.; Thorpe, C. W.; Sphon, J. A. Anal. Chem. 1981, 53, 2003. (12) Szulegko, J. E.; Howe, I.; Beynon, J. H.; Schlunegger, U. P. Org. Mass Spectrom. 1980, 15, 5. (13) Stemmler, E. A.; Hites, R. A. Anal. Chem. 1985, 57, 684. (14) Betowski, L. D.; Webb, H. M.; Sauter, A. D. Biomed. Mass Spectrom. 1983, 10, 369. (15) Hunt, D. F.; Crow, F. W. Anal. Chem. 1978, 50, 1781. (16) Ong, V. S.; Hites, R. A. J. Am. Soc. Mass Spectrom. 1993, 4, 270. (17) (a) Winkler, F. J.; Stahl, D. J. Am. Chem. Soc. 1978, 100, 6779. (b) Winkler, F. J.; Stahl, D. J. Am. Chem. Soc. 1979, 101, 3685. (18) (a) Quadrupole Storage Mass Spectrometry; March, R. E., Hughes, R. J., Eds.; Chemical Analysis Series; Wiley: New York, 1989; Vol. 102. (b) Practical Aspects of Ion Trap Mass Spectrometry; March, R. E., Todd, J. F. J., Eds.; CRC Press: Boca Raton, FL, 1995; Vols. 1-3. (19) (a) Evans, C.; Catinella, S.; Traldi, P.; Vettori, U.; Allegri, G. Rapid Commun. Mass Spectrom. 1990, 4, 335. (b) Favretto, D.; Guidugli, F.; Seraglia, R.; Traldi, P.; Ursini, F.; Sevanian, A. Rapid Commun. Mass Spectrom. 1991, 5, 240. (20) Van Pelt, C. K.; Haggarty, P.; Brenna, J. T. Anal. Chem. 1998, 70, 4369. (21) Matter, R. E.; Todd, J. F. J. Int. J. Mass Spectrom. Ion Phys. 1980, 33, 159. (22) Berberich, D. W.; Yost, R. A. J. Am. Soc. Mass Spectrom. 1994, 5, 757. (23) Catinella, S.; Traldi, P.; Jiang, X.; Londry, F. A.; Morrison, R. J. S.; March, R. E.; Gre´goire, S.; Mathurin, J. C.; Tabet, J. C. Rapid Commun. Mass Spectrom. 1995, 9, 1302.
5056
Analytical Chemistry, Vol. 72, No. 20, October 15, 2000
ionization), negative as well as positive charged ions are simultaneously produced with a very low ratio of ∼1:1000. Moreover, it should be emphasized that longer ionization times and relatively high analyte pressure into the ion trap are required to effectively observe the negative ions produced via EC process. Ion-ion recombinations rapidly raise to a great extent and further decrease the total number of negative ions. These extreme limiting conditions lead to space charge24 perturbations due to the repulsive Coulombic forces strength, from the highly abundant positive ions, favor the observation of the higher mass/charge ratio negative ions by increasing their trapping potential well.24 However, electron/molecule interactions leading to negative ions require electrons of near thermal kinetic energy, which are limited by the accelerating property of the rf storage drive potential which will compel the electrons to undergo unstable trajectories with excessive kinetic energy. Note that the relatively low pressure conditions, ∼1 mTorr (0.133 Pa) within ion trap, cannot either efficiently thermalize these high kinetically energetic electrons or collisionally stabilize (or quench) molecular negative ions. In conclusion, the low yield of negative ions in the QITMS does not provide a convenient tool for analytical applications that impose low-level detection for environmental contaminants.22 An alternative solution for producing negative species in the QITMS was provided by combining it with an external highpressure ion source. Mainly, two different approaches have been proposed. The first approach used a quadrupole ion trap as a massselective detector.25,26 Analytes were introduced in the external ion source (called external ionization) using gas chromatography or by direct probe insertion. They were mostly ionized via resonant electron capture under high-pressure conditions (typically 0.1-1 Torr). Unfortunately with such a high-pressure source, some ion/ molecule reactions compete with the previous EC process. The produced negative ions were then injected into QITMS and were conventionally analyzed by scanning the rf drive potential. The second approach relied on the production and the injection of reactant negative ions27 (odd- or even-electron species) from an external high-pressure ion source. Analyte compounds were separately introduced into the QITMS via GC mode, and NICI under low-pressure conditions was carried out after selection of injected reactant negative ions. This in situ analyte ionization via specific ion/molecule reactions was preferred for two reasons. The first was even clearly depicted and involved limitations occurring in negative ion formation into the high-pressure ion source, which was greatly affected by the experimental conditions. Obviously, low-pressure NICI with a reaction time period is less dependent than high-pressure NICI upon noncontrolled operating factors.28 Second, the injection drawbacks are highly reduced. The loss of sensitivity due to the low injection efficiency (from 1 to (24) (a) Gre´goire, S.; Mathurin, J. C.; March, R. E.; Tabet, J. C. ICR/Ion Trap News Lett. 1996, 43, 12. (b) Mathurin, J. C.; Gre´goire, S.; Brunot, A.; Tabet, J. C.; March, R. E.; Catinella, S.; Traldi, P. J. Mass Spectrom. 1997, 32, 829. (c) Gre´goire, S.; Mathurin, J. C.; March, R. E.; Tabet, J. C. Can. J. Chem.Rev. Can. Chim. 1998, 76, 452. (25) McLuckey, S. A.; Glish, G. L.; Asano, K. G. Anal. Chim. Acta 1989, 225, 25. (26) McLuckey, S. A. Goeringer, D. E.; Asano, K. G.; Vaidyanathan, G.; Stephenson, J. L. Rapid Commun. Mass Spectrom. 1996, 10, 287. (27) Eckenrode, B. A.; Glish, G. L.; McLuckey, S. A. Int. J. Mass Spectrom. Ion Processes 1990, 99, 151. (28) Siegel, M. W. Int. J. Mass Spectrom. Ion Phys. 1983, 46, 325.
Table 1. Structures and Molecular Weights of the Organic Compounds Used in This Study
5%25,29) of externally generated ions, the dissociation,30-32 and the mass discrimination were avoided with respect to the ions of interest. Moreover, the limitations encountered with injected ions, which are related to nonlinear resonance effects leading to ion ejection and dissociation, are bypassed with the use of internal ionization. In the present work, NICI-MS was examined with several reactant negative ions in order to perform reactive and selective collisions into the ion trap with several priority compounds. Nitroand phosphorus-containing compounds have been extensively investigated in negative ion mode, with respect to their electrophilic properties as well as their high electron affinities. Detection of such relevant species at trace level is of much concern for civil and environmental protection.33,34 The characterization of these substances by mass spectrometry is often complicated by the variety of fragmentation pathways available in the positive EI-MS conditions, resulting in complex mass spectra as shown in a commercial mass spectra library. Table 1 shows the compounds that have been subjected to NICI-MS analysis in order to study their gas-phase reactivities and to demonstrate the analytical feasibility of the combination of a high-pressure ion source and a QITMS coupled with GC. EXPERIMENTAL SECTION Instrumentation and Material. Sequential MS and MS2 experiments were performed with a homemade instrument composed of a modified dual EI/CI high-pressure ion source (Nermag, France) externally fitted with a Saturn III QITMS (29) Morand, K. L.; Horning, S. R.; Cooks, R. G. Int. J. Mass Spectrom. Ion Processes 1991, 105, 13. Williams, J. D.; Reiser, H.; Kaiser, R. E.; Cooks, R. G. Int. J. Mass Spectrom. Ion Processes 1991, 108, 199. (30) Van Berkel, G. J.; McLuckey, S. A.; Glish, G. L. Anal. Chem. 1991, 63, 1098. (31) Quarmby, S. T.; Yost, R. A.; Schwartz, J. C.; Syka, J. E. P. Proc. 44th ASMS Conference on Mass Spectrometry and Allied Topics, Portland, OR, 1996; p 765. (32) Steiner, V.; Beaugrand, C.; Liere, P.; Tabet, J. C. J. Mass Spectrom. 1999, 34, 511. (33) Convention on the Prohibition of the Development, Production, Stockpilling and Use of Chemical Weapons and on their Destruction. United States Control and Disarmament Agency, Washington, DC, 1993. (34) (a) Kolla, P.; Angew. Chem., Int. Ed. Engl. 1997, 36, 800. (b) Yinon, J. Forensic Applications of Mass Spectrometry; CRC Press: Boca Boton, FL, 1994. (c) Yinon, J.; Zitrin, S. Modern Methods and Applications in Detection of Explosives; Wiley: Chichester, U.K., 1993.
Figure 1. Schematic view of the homemade instrument composed of a modified high-pressure EI/CI ion source combined with an ion trap mass spectrometer.
(Varian, Palo Alto, CA). A synoptic view of the instrument is shown in Figure 1. The original manifold of the commercial GC/QITMS has been preserved. Note first that the ion source vacuum pumping was achieved by a 250 L/s turbomolecular pump (Varian V250 ISO) and by a 280 L/s diffusion pump (Edwards 100/300 P) for the QITMS manifold (Figure 1). A differential pumping was achieved through the hole (inner diameter 1.8 mm) of the last focalization lens located between both stages of this homemade instrument. Consequently, reactant gas diffusion from the ion source into the ion trap was highly decreased. This allowed us to reduce or avoid secondary ion/molecule reactions which take place at low collision energy (e.g., self-ionization) with ions produced from substrate. Dissociations with high rate constants occurring from the higher kinetic energy product ions are somewhat limited when these ions are formed in the ion trap mass spectrometer. Thus, in contrast to high-pressure NICI, where several processes compete, here selective ion/molecule reactions are realized. The injection of reactant ions was time controlled by a gate lens, pulsed at (150 V and located between the extraction lens and the first focalization lens joined to the ion source housing. The repeller lens was also modified in order to accept a second filament, on-axis to the injection beam. Then, three analysis modes of operation were available: the EI, PICI, and NICI conditions. Two selective detection modes were easily Analytical Chemistry, Vol. 72, No. 20, October 15, 2000
5057
Table 2. External Ion Source Conditions for Preparation of Reactant Negative Ionsa precursor neutral
pressure (× 10-4 Torr)
reactant ions
main processa
CH2Cl2 CH3OH N2O
2 2.5 5
ClCH3OO2•-
a b c
a (a) Dissociative electron capture, (b) consecutive ion/molecule reaction, and (c) dissociative electron capture followed by ion/molecule reaction.
Figure 2. Analytical scan for sequential NICI-MS/O2•- experiments.
supplied by polarizing between (5000 V a modified conversion dynode (venetian blind type, Thorn EMI) located between the QITMS exit end cap and the detector (Channeltron, Galileo). Our approach to detect negative ions was to convert them to secondary positive ions by striking a conversion dynode maintained at a high positive potential.35 These secondary ions were then detected by impacting a standard continuous dynode electron multiplier. However, low conversion of negative to positive ions did not yield an efficient detection of negative ions in our homemade system. The ion trap manifold temperature was set to 170 °C. Helium gas bath pressure was ∼3 × 10-5 Torr. Analytical scan function for all NICI-MS experiments was realized with QISMS software. The NICI-MS/O2-• sequence, displayed in Figure 2 is made up of four steps. During the injection period (A), an optimal rf storage voltage amplitude was employed to maximize the number of injected reactant ions. The low-mass cutoff value (LMCO) was adjusted for each selected reactant ion such that the qz trapping parameter was maintained close to 0.4. Injection time was then controlled by the automatic gain control program (AGC)36 in order to prevent any space charge effects. In all experiments, reactant ions were extracted from the external ion source with a kinetic energy close to 5 eV. The reactant ion isolation period (B) was accomplished with the mass-selective instability scan combined to the use of broad-band waveforms.37 This ensured selective ion/ molecule reactions to be performed without production of any other interfering ions during the reaction period (C). Finally, a mass-selective ejection scan38 with axial modulation voltage (Vp-p ) 4 V) permitted ion detection from 50 to 650 Th. The analytical scan rate was fixed at 5555 Th/s. For MS2 experiments, an additional period was incorporated before period D, where a supplementary ac voltage is applied across the end-cap electrodes. The various reactant ions (Table 2) were produced by either dissociative electron capture or consecutive ion/molecule reaction. Reactant negative ions formation and properties were discussed in detail elsewhere.39 Pesticide compounds were purchased from Aldrich Chemical Co. They were separately introduced in the QITMS by capillary (35) Stafford, G. C. Environ. Health Perspect. 1980, 36, 85. (36) Stafford, G. C.; Taylor, D. M.; Bradshaw, S. C.; Syka, J. E. P.; Uhrich, M. Proc. 35th ASMS Conference on Mass Spectrometry and Allied Topics, Denver, CO, 1987; p 775. (37) Bolton, B.; Wells, G.; Wang, M. Proc. 41st ASMS Conference on Mass Spectrometry and Allied Topics, San Francisco, 1993; p 474a. (38) Stafford, G. C.; Kelley, P. E.; Syka, J. E. P.; Reynolds, W. E.; Todd, J. F. J. Int. J. Mass Spectrom. Ion Processes 1984, 60, 85. (39) Faye, T.; Mathurin, J. C.; Brunot, A.; Tabet, J. C.; Wells, G.; Fuche´, C. Anal. Chem., following paper in this issue.
5058
Analytical Chemistry, Vol. 72, No. 20, October 15, 2000
gas chromatography with Star 3400 CX (Varian). The column was 30 m × 0.25 mm i.d. × 0.25 µm thickness DB-5MS (J&W Scientific). The following temperature program was used: 50 °C for 1 min, ramped to 140 °C at 40 °C/min, and finally 10 °C/min to 260 °C for 2 min. Explosive material was obtained from the CREL and was introduced by a short chromatographic capillary column, i.e., 10 m × 0.25 mm × 0.25 µm thickness BP-5 (SGE). These conditions ensure a minimum thermal degradation, of the thermally labile explosive compounds, observed under standard analysis conditions. The temperature program was as follows: 60 °C for 1 min and then ramped at 10 °C/min to 250 °C. Liquid injections were performed at 230 °C in splitless mode with 1-µL volume. The interface to the mass spectrometer was maintained at 230 °C for all experiments. RESULTS AND DISCUSSION To evaluate the detection discrimination between negative and positive ions, signal intensities were compared for identical concentration of opposite charged ions within the ion trap. The Ar•+ (m/z 40) and I- (m/z 127) ions were used for this study. These ions were produced from argon and methyl iodide in the external ion source at the respective pressures of 3 × 10-4 and 2 × 10-4 Torr. The rf storage drive amplitude for the injection period was adjusted such that the qz trapping parameter was 0.3 for both these selected ions. The conversion dynode was typically polarized at (4000 V for negative and positive ions detection, respectively. Multiplier voltage was negatively biased at -1150 V to prevent any saturation of the detector. The maximum nAmax ion density in the QITMS could be estimated40 using the following eq 1,
nAmax )
( )
Ωqz 3 m 16 0 A zAe
2
(1)
where Ω ) 2π × 1.05 × 106 rad/s is the angular frequency of the rf drive potential, mAr ) 6.68 × 10-26 kg, and mI ) 2.12 × 10-25 kg. The corresponding calculated values of the maximum number of ions stored within the ion trap were nAr+ max ) 4.3 × 1011 ion/ m3 and nI- max ) 1.4 × 1012 ion/m3. Signal intensities of I- and Ar•+ were experimentally investigated as a function of the injection time. As shown Figure 3, it can be assumed that the ion trap was filled at 20 ms for both the positive and negative ion species. As regards the relative ion intensities and considering similar ionic (40) Schwebel, C.; Mo ¨ller, P. A.; Manh, P. T. J. Phys. Appl. 1975, 10, 227.
Figure 3. Signal intensity dependence upon the injection duration step of the selected and injected (a) Ar•+ (m/z 40) and (b) I- (m/z 127) ions.
densities, it can be seen that the yield of detected negative ions is low compared to that of positive ions, taking into account a constant conversion efficiency along the mass range for both ion species. A 30-fold decrease in sensitivity was observed (97% in absolute intensity). This behavior is evidence for the low-yield conversion of primary negative ions to secondary positive ions, which is responsible for the poor detection of negative ions. Nevertheless, the implantation of a conversion dynode permitted us to reach the main objective of detecting negative ions characterized by low m/z ratios. Reactant Negative Ion Selection. The wide choice of NICI gas allows production of a large panel of reactant ions (G- or G•-) in the external high-pressure source for further injection which can selectively react with neutral analyte in the QITMS. In this way, a series of reactant ions were tested for their ability to produce either an odd-electron M•- molecular ion or a deprotonated [M - H]- molecule (or adduct [M + R]- ions) via efficient ion/molecule reactions, permitting the analysis of compounds present in trace level for environmental (e.g., pesticides) or forensic applications (e.g., explosives and chemical warfare agents). Another major criterion for the choice of reactant negative ions results from the trapping properties inherent to QITMS. Indeed, a low rf storage voltage amplitude must be used in order to efficiently store the low m/z ratio reactant ions. Under these
conditions, ion/molecule products of high m/z ratios have low qz trapping parameters. For instance, a LMCO of 13 Th is imposed for NICI experiments with OH- (17 Th). Hence, a m/z 400 ion will be stored at qz ) 0.028. This low-qz parameter will ensure nonideal trapping conditions, i.e., a weak trapping pseudopotential well depth which could therefore involve mass discrimination during the analytical scan. Consequently to avoid such limitation, all the reactant ions used had relatively higher m/z ratios, i.e., m/z > 30. After preparation of reactant negative ions (Table 2), NICI experiments on TNT explosive compound were performed with three reactant ions (Cl-, CH3O-, and O2•-). For the pesticides, only results with Cl- were presented and those provided from both the CH3O- and O2•- reactants were discussed elsewhere.39 The main ions produced in the ion trap cell are reported in Table 3. (1) NICI Mass Spectra with Explosive Material. Nitroaromatic compounds are widely used as pesticides as well as explosives. Many of these compounds and their transformation products are of significant toxicological and forensic concern. Nitroaromatics are obviously an energetic material, frequently associated with crime as well as ubiquitous environmental pollutants. 2,4,6-Trinitrotoluene has been studied in this work since it is a major constituent of most explosive mixtures. This organic component contains three NO2 groups, conjugated with the aromatic moiety, which confers on it a large EA value. Therefore, the most sensitive detection method should take advantage of this chemical property by promoting the use of the negative ion mode. Indeed, positive ion production via electron ionization or chemical ionization usually leading to extensive fragmentation41 could be supplanted by the negative mode.26,27,34c,41,42 Lack of specific information (molecular weight for instance) was the main drawback for both positive ion modes. The EI mass spectrum of 2,4,6TNT displayed in Figure 4a, recorded with a commercial Saturn III QITMS under conventional in situ ionization and trapping conditions, reveals the absence of a stable molecular M•+ ion. Typically, a high degree of fragmentations is observed, such as the loss of the OH• radical to yield ions at m/z 210 and consecutive loss of NO• (or direct loss of HNO2) to give ions at m/z 180 and, as well as, numerous lower m/z ratio product ions. Moreover, formation of the protonated [M + H]+ molecule at m/z 228 is competitively produced by self-ionization. Some fragment ions, formed under EI of 2,4,6-TNT, transfer a proton to neutral 2,4,6TNT to yield the protonated molecule. Such a process, in the QITMS, is induced by the EI fragment ions often considered as protonated smaller systems. Their corresponding neutrals have lower proton affinity than that of the parent molecule. This phenomenon has been previously shown by Mechin et al.43 with DEMP under EI-MS conditions and significantly raises with relatively large residence times in the ion trap. On the other hand, the NICI/O2•- mass spectrum of 2,4,6-TNT displays specifically a molecular M•- negative ion at m/z 227 (Figure 4b) after reacting with the selected O2•- reactant ion in QITMS. This stable molecular negative ion M•- is formed via an exothermic charge exchange reaction, O2 possessing a low (41) Yinon, J. Mass Spectrom. Rev. 1991, 10, 179. (42) Yinon, J. J. Forensic Sci. 1980, 25, 401. (43) Me´chin, N.; Plombey, J.; March, R. E.; Blasco, T.; Tabet, J. C. Rapid Commun. Mass Spectrom. 1995, 9, 5.
Analytical Chemistry, Vol. 72, No. 20, October 15, 2000
5059
Table 3. Main Ions (m/z) Produced from NICI-MS Reactions According to the Selected Reagent Aniona
a
compounds
NICI/Cl-
NICI/CH3O-
NICI/O2•-
2,4,6-TNT (Mw ) 227 u) methyl parathion (Mw ) 263 u) ethyl parathion (Mw ) 291 u) malathion (Mw ) 330 u)
226 [M - H]- (100) 248 [M - CH3]- (100) 262 [M - C2H5]- (100) 315 [M - CH3]- (100)
227 M•- (70), 226 [M - H]- (30) b b b
227 M•- (100) b b b
Relative ion intensities are displayed in parentheses. bThese produced ions are developed elsewhere.39
Figure 4. EI mass spectrum (a) of 2,4,6-TNT produced under in situ ionization in the ion trap. The [M + H]+ at m/z 228 and dominant fragment ion [M - OH]+ at m/z 210. The NICI mass spectrum after O2-• injection (b) of 2,4,6-TNT displays molecular M•- negative ion at m/z 227 without fragment ion.
electron affinity [EA(O2) ) 42.2 kJ‚mol-1].44 To unambiguously identify the 2,4,6-TNT analyte, a CID experiment was performed under resonant excitation (Figure 5) on the selected molecular M•- ion (m/z 227) to produce several diagnostic fragment ions. [M - OH]- ion at m/z 210 (related to the ortho effect), [M NO]- ion at m/z 197, [M - OH - NO]•- ion at m/z 180, [M 5060 Analytical Chemistry, Vol. 72, No. 20, October 15, 2000
2NO]•- ion at m/z 167, and [M - 3NO]- at m/z 137, as well as [M - OH - NO - CO]- at m/z 152, are the main product observed ions.45,46 Alternatively, ion/molecule reactions induced by CH3O- on the 2,4,6-TNT analyte yields the even-electron [M - H]- ion (m/z 226), which is produced in competition with the odd-electron M•-
Figure 5. CID induced by resonant excitation (qz ) 0.3, Vexc p-p• ) 0.2 V) of the selected M•- ion (m/z 227) produced by NICI/O2-• experiments from 2,4,6-TNT. The observed m/z 228 ion is due to the survivor natural isotopic 13C ion, not excited by the applied tickle.
molecular ion (Table 3). This particular reactivity enlightens the exothermic proton transfer related to the higher gas-phase acidity of 2,4,6-TNT [∆H°acid (2,4,6-TNT) ) 1320 kJ‚mol-1]44 than that of methanol [∆H°acid (CH3OH) ) 1590 kJ‚mol-1].44 Furthermore, a charge exchange reaction is also the most efficient process since 2,4,6-TNT has a larger EA than that of CH3O•, [EA(CH3O•) )146 kJ‚mol-1].44 Finally, Cl- negative ion exclusively reacts with 2,4,6-TNT to form the [M - H]- ion (m/z 226) because of the weaker acidity of HCl [∆H°acid (HCl) ) 1395 kJ‚mol-1]44 compared with that of the analyte. Here, in contrast to ion/molecule reactions with CH3O-, no molecular M•- ion is produced due to the very large EA value of Cl• [EA(Cl•) ) 345 kJ‚mol-1],44 which prevents any charge exchange reaction and suggests that EA(2,4,6-TNT) is lower than EA(Cl•) but larger than EA (CH3O•). (2) Application to Pesticide Compound Detection in a Complex Matrix. Organophosphorus compounds are used as insecticides in agricultural or in domestic application and also as chemical warfare agents. Malathion (MLT), methyl parathion (MTP), and ethyl parathion (ETP), Table 1, as one class of representative compounds, have to be detected at trace level in environmental field to monitor remediation of these pollutants from contaminated sites. Among the three reactant ions used for NICI-MS experiments as shown elsewhere39 (Cl-, O2•-, and CH3O-), chloride reactant provides an efficient tool for selective and sensitive ionization of target compounds present at trace level in complex matrixes. Cl- ion regiospecifically reacts toward organophosphorus compounds. It can be easily prepared from CH2Cl2 solvent as gas. The formation of [MLT - CH3]- (m/z 315), [MTP - CH3]- (m/z 248), and [ETP - C2H5]- (m/z 262) ions occurs via an SN2-C displacement at the activated carbon atom of (44) (a) Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W. G. J. Phys. Chem. Ref. Data 17, Supp. 1 1988. (b) NIST Standard Reference Database; Mallard, W. G., Ed.; 1998, 69;NIST Chemistry WebBook, http://webbook.nist.gov/chemistry/. (45) Yinon, J.; Fraisse, D.; Dagley, I. J. Org. Mass Spectrom. 1991, 26, 867. (46) Langford, M. L.; Todd, J. F. Org. Mass Spectrom. 1993, 28, 773.
the alkoxy group.39 During these selective reactions, no other competitive ion/molecule process (or dissociation) was observed. These results prove that the Cl- reactant is a powerful reactant ion for analytical purposes since the ionic current is only concentrated on one diagnostic ion. A mixture of pesticide compounds obtained from the extraction of a paint sample has been doped with 10 various pesticides at to two different dilutions (2 ng/µL and 200 pg/µL). The EI mass spectrum of the sample that contains 2 ng/µL parathion samples was performed on a nonmodified Saturn III QITMS under conventional analysis parameters. Pesticide ions are severely hidden due to a high level of interference from the chemical matrix (Figure 6a). Only the positive molecular m/z 263 and 291 ions could be used for identification of both pesticides. However, in the corresponding NICI/Cl- mass spectrum (Figure 6b), these background ions disappear. Thus, this latter technique has tremendous detection advantages due to the high selectivity offered by NICI/Cl- since only the parathion analogues exhibit a high-intensity signal involving [MTP - CH3]- (m/z 248) and [ETP - C2H5]- (m/z 262) ions. The limit of detection (S/N ) 20) achieved was 50 pg with a reaction time delay fixed at 50 ms using selective ion monitoring mode for m/z 248 corresponding to [M - CH3]- specific ions for methyl parathion. For the other pesticide compounds, a limit of detection from 0.2 to 1 ng was currently observed. CONCLUSION Owing to the difficulties encountered with in situ negative ionization in an ion trap by electron/molecule interactions, the use of an external ion source is obviously a convenient solution to produce negative ions in QITMS. Selective accumulation of reactant negative ions in the ion trap (filled in ∼20 ms) to perform in situ low-pressure chemical ionization is helpful since it enhances the NICI mass spectrum reproducibility compared with those available under high-pressure conditions. Moreover, negative ions produced near the center of the ion trap are then better trapped and less subject to nonlinear phenomena. In addition, the use of Analytical Chemistry, Vol. 72, No. 20, October 15, 2000
5061
Figure 6. Chromatograms of both the studied parathion pesticides in a paint sample spiked with 2 ng/µL under (a) EI-MS and (b) NICI-MS/Clconditions.
an efficient differential pumping, restricting reactant gas diffusion into the ion trap permits us to considerably reduce unwanted reactive and dissociative collisions. QITMS as an efficient (in time) tandem mass spectrometer provides one supplementary degree of specificity for analysis of chemical compounds. For instance, the MS/MS experiment on 2,4,6-TNT molecular negative ion described herein affords unambiguous methodology for the structural elucidation and the identification of this analyte. Preliminary analytical results show a detection limit of 50 pg for methyl parathion under NICI/Clconditions. To improve detection levels for explosive compounds, sampling will be further performed with headspace solid-phase (47) Arthur, C. L.; Pawliszyn, J. Anal. Chem. 1990, 62, 2145. (48) Amirav, A.; Dagan, S. Method and Device for the Introduction of a Sample into a Gas Chromatograph. U.S. patent 5686656, 1997: Japan application 43248/1997, 1997; and Chinese application 97102579.7, 1997. (49) Amirav, A.; Dagan, S. Eur. Mass. Spectrom. 1997, 3, 105.
5062
Analytical Chemistry, Vol. 72, No. 20, October 15, 2000
microextraction (SPME),47 which should ensure vapor concentration within short response times. Additional experiments using a direct insertion probe48,49 are also performed in our laboratory for studying other classes of explosive material with respect to their thermally labile properties, e.g., hexogen (RDX), octogen (HMX) and pentrite (PETN). ACKNOWLEDGMENT We acknowledge the Centre d’Etudes du Bouchet, the Centre de Recherches et d’Etudes de la Logistique de la Police Nationale, the CNRS, the University Pierre et Marie Curie. We also thank Varian Associates for their financial and technical support. Received for review February 10, 2000. Accepted July 19, 2000. AC000171M
