High-Pressure Ion Source Combined with an In ... - ACS Publications

Sep 9, 2000 - Low-pressure NICI experiments were performed with a home-built instrument made of an EI/CI high-pressure ion source (Nermag, France) ...
2 downloads 0 Views 209KB Size
Download PDF
Anal. Chem. 2000, 72, 5063-5069

High-Pressure Ion Source Combined with an In-Axis Ion Trap Mass Spectrometer. 2. Application of Selective Low-Pressure Negative Ion Chemical Ionization T. Faye, J. C. Mathurin, 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

Negative ion chemical ionization was carried out using a quadrupole ion trap mass spectrometer with selected reactant negative ions, primarily injected from a homemade dual EI/CI external ion source. Hence, selective ion/molecule reactions were provided according to the reaction time, which induce a greater control over bimolecular ionization mechanisms than in conventional a high-pressure ion source combined with beam instruments, where several competitive ionization processes take place mainly due to source conditions (e.g., temperature, pressure, and repeller). By selecting the reactant ions, ion/molecule reactions were specifically produced (i.e., charge exchange, proton transfer, nucleophilic substitution, and/or r-β elimination) with several organic target compounds. Gas-phase reactivity of phosphorusand nitrogen-containing compounds (such as phosphonates as representative for chemical warfare agents and phosphorothionates, phosphorodithionates, and triazines for pesticides) as well as dinitro aromatic compounds (for pesticides) has been explored, in the present work, to ensure further unambiguous detection. The detection of contaminant and energetic compounds has always been an active topic of interest in environmental and forensic science. Phosphorus-, nitrogen-, and nitro-containing compounds for pesticide material have to be detected at ultratrace level when monitoring complex matrixes.1-4 Mass spectrometry * Corresponding author: (phone) (33) 01 44 27 32 63; (fax) (33) 01 44 27 38 43; (e-mail) [email protected]. (1) Yinon, J. Mass Spectrom. Rev. 1991, 10, 179. (2) Modern Methods and Applications in Analysis of Explosives; Yinon, J., Zitrin, S., Eds.; Wiley: Chichester, U.K., 1993. (3) Schachterle, S.; Brittain, R. D.; Mills, J. D. J. Chromatogr., A 1994, 683, 185. (4) Drevenkar, V.; Stengl, B.; Frobe, Z. Anal. Chem. Acta 1994, 290, 277. 10.1021/ac000172e CCC: $19.00 Published on Web 09/09/2000

© 2000 American Chemical Society

as a powerful trace detector must first be specific and sensitive. Gas-phase ionization is commonly used for such analysis of volatile compounds. Over the wide range of existing mass spectrometers, the quadrupole ion trap mass spectrometer5 (QITMS) has the advantage because of its versatile modes of operation. Indeed, MSn experiments and selective ion/molecule reactions can be easily performed within the trapping volume, for structural elucidation of particular compounds. Nevertheless, the commercially available QITMS was limited to the study of analytes with relatively low molecular weight and to the positive ion modes (i.e., EI and PICI). Difficulties encountered with the in situ formation of negative ions6,7 via the electron capture process can be overcame by the use of an external ion source.8,9 Production and injection of a reactant negative ion seems to be an efficient way of operation, since the specificity of ion/ molecule interactions can be dramatically enhanced with the isolation of reactant ions before the reaction period.10 In this work, negative ion chemical ionization (NICI) was carried out using a homemade instrument where selected reagent anions were injected, from a dual EI/CI external high-pressure (5) (a) Quadrupole Storage Mass Spectrometry; Chemical Analysis Series; March, R. E., Hughes, R. J., Eds.; 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. (6) (a) 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. (b) Gre´goire, S.; Mathurin, J. C.; March, R. E.; Tabet, J. C. ICR/Ion Trap News Lett. 1996, 43, 12. (7) (a) Mathurin, J. C.; Gre´goire, S.; Brunot, A.; Tabet, J. C.; March, R. E.; Catinella, S.; Traldi, P. J. Mass Spectrom. 1997, 32, 829. (b) Gre´goire, S.; Mathurin, J. C.; March, R. E.; Tabet, J. C. Can. J. Chem.-Rev. Can. Chim. 1998, 76, 452. (8) Eckenrode, B. A.; Glish, G. L.; McLuckey, S. A. Int. J. Mass Spectrom. Ion Processes 1990, 99, 151. (9) McLuckey, S. A.; Goeringer, D. E.; Asano, K. G.; Vaidyanathan, G.; Stephenson, J. L. Rapid Commun. Mass Spectrom. 1996, 10, 287. (10) Mathurin, J. C.; Faye, T.; Brunot, A.; Tabet, J. C.; Wells, G.; Fuche´, C. Anal. Chem., preceding paper in this issue.

Analytical Chemistry, Vol. 72, No. 20, October 15, 2000 5063

Table 1. Structures and Molecular Weights of the Organic Compounds Used in This Study

ion source into an ion trap, to specifically react with target pollutant compounds (Table 1). Further ion/molecule reactions are obtained that complement those available in positive ion chemical ionization (PICI) and that similarly may be selected to ensure specific information required in particular analysis. EXPERIMENTAL SECTION Low-pressure NICI experiments were performed with a homebuilt instrument made of an EI/CI high-pressure ion source (Nermag, France) combined with an ion trap mass spectrometer, a Saturn III QITMS (Varian, Palo Alto, CA). Details on experimental procedures and a synoptic view of this interface have been presented previously.10 The negative reactant ions were produced in the external ion source under electron impact and relative highpressure conditions. A time-controlled reaction period whereby the selected negative ions reacted with the neutral compounds was provided in order to realize in situ selective ion/molecule interactions in the ion trap. The detection of negative ions was achieved by a modified conversion dynode (venetian blind type, Thorn EMI) placed before the electron multiplier detector (Channeltron, Galileo). The ion trap manifold temperature was set to 170 °C and helium gas bath pressure was ∼3 × 10-5 Torr. An analytical scan function for all NICI-MS experiments was realized with QISMS software and was displayed elsewhere.10 Phosphonate compounds were obtained from the CEB, and pesticide compounds were purchased from Aldrich Chemical Co. The concentration of sample solutions varied typically from about 5064

Analytical Chemistry, Vol. 72, No. 20, October 15, 2000

10 to 30 ng/µL. Compounds were introduced in the QITMS by a 30 m × 0.25 mm i.d. × 0.25 µm thickness DB-5MS (J&W Scientific) capillary column. The temperature program employed was as follow: 50 °C for 1 min, then ramped at 40 °C/min to 140 °C, and finally 10 °C/min to 260 °C for 2 min. Injector temperature was 230 °C, and 1-µL volume was injected in the splitless mode. The transfer line between the gas chromatograph (Star 3400 CX, Varian) and the ion trap mass spectrometer was fixed at 230 °C. The injection of reagent ions was time-controlled by the AGC software and reaction times of about 50-100 ms. The total acquisition time was ∼300-400 ms with an analytical scan rate fixed at 5555 Th‚s-1. RESULTS AND DISCUSSION (1) Reactant Negative Ion Formation. The advantages of chemical ionization mass spectrometry, particularly in negative ion mode, for studying compounds that exhibit extensive fragmentations with no or very weak molecular ion in EI-MS have been long known.11 Although PICI mass spectra can give some fragmentations, the intact quasimolecular ion produced from bimolecular ionization occurring at the collision frequency (i.e., for exothermic proton transfer and reactant ion attachment) allows enhancement of the sensitivity and selectivity of mass spectrometry. Such a behavior improves mass spectral quality and makes (11) Harrison, A. G. Chemical Ionization Mass Spectrometry, 2nd ed.; CRC Press: Boca Raton, FL, 1992.

Alternatively, the CH3O- methoxide (m/z 31) was likely produced from the following reactions, under high pressure (PCH3OH ) 2.5 × 10-4 Torr).

CH3OH + e f (CH3OH•-)*

(1)

(CH3OH•-)* + CH3OH f CH3O• + CH3O- + H2 (2)

Figure 1. Gas-phase acidities12 (from AH forms) and electron affinities12 (from radical A• or molecule) of four reagent gases used in NICI experiments (CH2Cl2, CH3OH, H2O, N2O).

CH3O• has again a relatively medium EA value [EA(CH3O•) ) 146 kJ‚mol-1]12 and therefore is unable to undergo an exothermic charge exchange reaction with most of organic analytes. Nevertheless, the weak Bro¨nsted CH3OH acid [∆H°acid (CH3OH) ) 1590 kJ‚mol-1]12 yields a frequently used CH3O- reactant which efficiently leads to the formation of deprotonated [M - H]molecules through an exothermic proton-transfer ion/molecule reaction. Although the less acidic commonly used reactant is water, through the OH- negative ion, it was not employed here due to the reasons mentioned previously,10 i.e., the low trapping efficiency of the produced negative ions characterized by higher m/z ratios. A mixture of argon and nitrogenous oxide, in a 1:10 ratio, was used to generate dioxygen O2•- negative ion (m/z 32) within total pressure of 5 × 10-4 Torr. The presence of the oxygen negative ion at m/z 16 (eq 3) must be emphasized and may contribute to dioxygen negative ion formation according to eq 4:

easier the analysis of target compounds in complex matrixes, especially in the presence of large chemical background noise. NICI can be considered as complementary to PICI, and the absence of fragmentation can be turned out by using MSn experiments, with the advantage of charge location to limit the fragmentation pathways. Reactant gas were carefully chosen in order to produce negative reactant ions and allow specific and regioselective ion/ molecule reactions to be performed in the ion trap mass spectrometer. Gas-phase properties of the reactant species used in this study are illustrated in Figure 1 as a function of the gas-phase acidity and electron affinity values.12 One should note that these ions are widely spread over the graphical representation and therefore will react differently toward the studied neutral substrates. For instance, chloride negative ions (m/z 35) were readily formed in the external ion source by a dissociative electron capture process from dichloromethane (as shown by Budzikiewicz13) and introduced via a metered needle valve at a constant uncorrected pressure of 2 × 10-4 Torr. The HCl reactant neutral could be considered as a strong Bro¨nsted acid [∆H°acid (HCl) ) 1395 kJ‚mol-1],12 and its corresponding negative reactant ion Cl- will abstract a proton only from stronger acidic compounds. On the other hand, Cl• is also characterized by a relative high electron affinity value [EA(Cl•) ) 345 kJ‚mol-1],12 which indeed prevents any charge exchange reaction with most of organic compounds. Furthermore, for compounds such as esters, for instance, characterized by relative low values of both electron affinity and gasphase acidity, only Cl- solvatation by analyte is observed, to give [M + Cl]- adduct ions. These ions, which cannot dissociate into [M - H]-, chose another less endothermic fragmentation pathway. Indeed, Cl-, as a powerful nucleophilic reactant, undergoes (i) nucleophilic substitution and gives rise to the formation of [M + Cl - RCl]- ions (R being an alkyl radical) and/or (ii) elimination to produce [M + Cl - HCl - (R - H)]- ion. In fact, these ions are similar and are noted as [M - R]-.

O•- was formed via direct dissociative electron capture (eq 3) while O2•- was produced by ion/molecule reactions between the produced ion O•- and the neutral reactant N2O (eq 4). The neutral O2 has a low electron affinity [EA(O2) ) 42.2 kJ‚mol-1],12 and the O2•- ions react generally via an exothermic charge exchange reaction with molecules possessing suitabily higher EA14,15 and therefore enhance a direct confirmation of the analyte molecular weight. The gas-phase acidity of the formal HO2• radical [∆H°acid -1 12 (HO2•) ) 1476 kJ‚mol ] is intermediate between that of HCl and CH3OH, and thus, its reactant O2•- ion must be considered as a stronger base to remove proton from many analytes, yielding [M - H]- ions.16 In view of these considerations, the gas-phase behavior of various substrates toward the injected and selected reactant ions has been compared. Particularly, the different ion/molecule reactions have been scrutinized according to the reactant negative ion used in ion trap. (2) Specific Negative Ion/Molecule Reactions. (a) Behavior of Pesticide Compounds under Selective NICI Conditions. Phosphorus-containing compounds are commonly used as insecticides in agricultural or in the domestic field, and nitrogencontaining compounds constitute a wide class of chemical herbicides. Several compounds representative of both the families were

(12) (a) Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W. G. J. Phys. Chem. Ref. Data 1988, 17 (Suppl. 1). (b) NIST Standard Reference Database; Mallard, W. G., Ed.; 1998, 69; NIST Chemistry WebBook, http://webbook.nist.gov/chemistry/. (13) Budzikiewicz, H. Mass Spectrom. Rev. 1986, 5, 345.

(14) Hunt, D. F.; McEven, C. N.; Harvey, Z. T. M. Anal. Chem. 1975, 47, 1730. (15) Hunt, D. F.; Stafford, G. C.; Crow, F. W.; Russell, J. W. Anal. Chem. 1976, 48, 2098. (16) Jennings, K. R. High Performance Mass Spectrometry; Gross, M. L., Ed.; American Chemical Society: Washington, DC, 1978.

N2O + e f [N2O]•-* f O•- + N2

(3)

O•- + N2O f [N2O,O]-•* f O2•- + N2

(4)

Analytical Chemistry, Vol. 72, No. 20, October 15, 2000

5065

Table 2. Main Ions (m/z) Produced from Selective NICI-MS Reactions According to the Reagent Aniona compounds

NICI/Cl-

methyl parathion (Mw ) 263 u)

248 [M - CH3]- (100)

262 [M - H]- (100), 248 [M - CH3]- (17)

ethyl parathion (Mw ) 291 u)

262 [M - C2H5]- (100)

290 [M - H]- (100), 262 [M - C2H5]- (27)

malathion (Mw ) 330 u)

315 [M - CH3]- (100)

329 [M - H]- (14), 315 [M - CH3]- (49), 157 [PS2[OCH3)2]- (100) 335 M•- (60), 334 [M - H]- (100) 214 [M - H]- (100) 239 [M - H]- (100)

trifluraline (Mw ) 335 u) atrazine (Mw ) 215 u) dinoterbe (Mw ) 240 u) aRelative

293 [M - H]- (100)

NICI/O2263 M•- (52), 248 [M - CH3]- (100), 154 [NO2C6H4S]- (76) 291 M•- (100), 262 [M - C2H5]- (56), 154 [NO2C6H4S]- (53) 329 [M - H]- (71), 315 [M - CH3]- (53), 157 [PS2[OCH3)2]- (100) 335 M•- (100) 214 [M - H]- (100) 240 M•- (100)

abundances of ions are displayed in Parentheses.

submitted to NICI experiments with three reactant negative ions: Cl-, CH3O-, and O2•-. Table 2 shows the main diagnostic ions produced by selective ion/molecule reactions induced by these reactant negative ions. Organophosphorus Compounds. Malathion (MLT) containing several acidic sites and various leaving groups may be considered as an acidic and electrophilic reactant. Among the possible ion/ molecule reactions, i.e., proton abstraction, R-β elimination, and/ or nucleophilic substitution, one may be more favored, according to the chemical and thermochemical properties of the selected reactant ions. Methyl parathion (MTP) and ethyl parathion (ETP) both include acidic (i.e., CH3-P) and electrophilic sites (P-O-C and PdS) and also contain an aromatic nitro group which enhances the electron affinity, allowing a charge exchange reaction with a suitable reactant ion characterized by a weaker electron affinity. The general reaction patterns of the three studied organophoshorus compounds under NICI/Cl- conditions shows a similar behavior. The results are summarized in Table 2. Under the experimental conditions, the diagnostic ion [M - R]- is observed (where R corresponds to the alkoxy group). The NICI/Cl- mass spectra of the MLT, MTP, and ETP samples exhibit a base peak corresponding to the [MLT - CH3](m/z 315), [MTP - CH3]- (m/z 248), and [ETP - C2H5]- (m/z 262) ions, respectively. For these three compounds, the SN2-C displacement occurs and is oriented at the activated carbon atom of the alkoxy group. Similar diagnostic ions were previously reported by Hogdes and al.17 for trimethyl phosphate ester. Note that SN2-C at the carboethoxy site of MLT is not a favorable process since ethyl loss is not observed here. Actually, these results enlighten the effect of the larger polarizability of the PdS group compared with the CdO group. This factor definitively orients the SN2-C process. Since no other ion or fragmentation was observed during these ion/molecule reactions, this demonstrates that the Cl- ion specifically and regioselectively reacts toward organophosphorus compounds. It appears to be a powerful reactant ion for analytical purposes, and it can be easily prepared from CH2Cl2 solvent as gas. When an acid weaker than HCl, such as CH3OH, was used, MTP and ETP react mainly via exothermic proton transfer to (17) Hodges, R. V.; Sullivan, S. A.; Beauchamp, J. L. J. Am. Chem. Soc. 1980, 102, 935.

5066

NICI/CH3O-

Analytical Chemistry, Vol. 72, No. 20, October 15, 2000

CH3O- methoxide, yielding the deprotonated [MTP - H]- and [ETP - H]- molecules observed at m/z 262 and 290, respectively (Table 2). Note that other ions are also competitively produced, e.g., [MTP - CH3]- (m/z 248) and [ETP - C2H5]- (m/z 262) ions, by SN2-C but with less efficiency. Although exothermic, these displacement reactions are not favored compared to the exothermic and efficient proton-transfer reactions. This is due to the presence of a higher intrinsic energy barrier in the displacement reactions. Considering the fact that CH3OH is a weaker acid than HCl, three competitive reactions with MLT may obviously occur, such as SN2-C displacement yielding [MLT - CH3]- ion (m/z 315), proton-transfer reaction (i.e., [MLT - H]-, m/z 329), and R-β elimination. This latter mechanism, described in Scheme 1, yields an abundant signal at m/z 157 corresponding to the [PS2(OCH3)2]ion, necessarily resulting from the primary proton abstraction at the R position of the carboethoxy group and regioselective to the β position of the sulfur atom concomitant with the C-S bond cleavage. This contrasts very likely with the direct proton abstraction from the R position of sulfur atom, yielding the [MLT - H]- ion. The [MLT - CH3]- ions, corresponding to SN2-C herein, are also produced but in less abundance than in NICI/ Cl-. This other competitive process is very likely hindered by R-β elimination, a more efficient gas-phase process which is strengthened by the acidic character of the R-carboethoxy position and the labile property of the C-S bond compared to the C-O bond. As expected, O2•- reacts with both the parathion analogues via a charge exchange reaction to produce the odd-electron MTP•and ETP•- molecular ions at m/z 263 and 291 (Table 2). The EA of the O2 molecule being very weak [EA(O2) ) 42.2 kJ‚mol-1],12 the corresponding reactant ion O2•- can react with most of the organic compounds characterized by higher EA values. Formation of MTP•- and ETP•- ions is substantially in agreement with the presence of nitro group in the parathion molecules, which increases their respective electron affinities and therefore enhances an exothermic charge exchange reaction with O2•negative ion. The relative stability of the molecular negative ion species is also strengthened by the delocalization of the negative charge within the nitro group, a well-known electron-withdrawing group.18 An additional diagnostic [NO2C6H4S]- ion corresponding to the p-nitrothiophenoxide appears at m/z 154. This ion has a high relative abundance similar to that displayed in the NICI mass (18) Stemmler, H.; Hites, R. A. Biomed. Environ. Mass Spectrom. 1987, 14, 417.

Scheme 1. Formation of [PS2(OCH3)2]- Ion (m/z 157) via an r-β Elimination Mechanism (a) and a Deprotonation (b) Proposed for the Reaction between CH3O- Anion and Malathion (Mw ) 330)

spectrum recorded under conventional high-pressure ion source conditions.19 In earlier experiments, these species have also been observed under electron capture in situ ionization in a QITMS.20 These findings suggest that this thioxide results from the dissociation of the excited molecular MTP•-* and ETP•-* ions. The mechanism of this ion formation involves a rearrangement process already observed in the case of phosphorothioxide compounds.21 On the other hand, the NICI/O2•- mass spectrum of MLT is quite similar (except in ion abundance) to the previous NICI/ CH3O- mass spectrum. No molecular negative ion has been observed because of the absence of a stronger electron-withdrawing group in MLT, conferring it a lower electron affinity than the neutral reactant O2. The base peak is the [PS2(OCH3)2]- ion at m/z 157 produced via R-β elimination process (as shown in Scheme 1) in competition with the formation of the deprotonated [MLT - H]- molecule (m/z 329). Another competitive reaction is observed with the substituted product [M - CH3]- ion (m/z 315) in similar yields. Nitrogen-Containing Compounds. Trifluraline (Table 1), with two different electron-withdrawing sites such as CF3 and NO2 groups, must mainly give rise to the formation of odd-electron molecular M•- ions (Table 2) through charge exchange ion/ molecule reaction with the CH3O- and O2•- reactants. Dinoterbe also includes two NO2 groups, but a phenolic site provides relatively high gas-phase acidity allowing proton-transfer reaction with a weaker acid. Finally, atrazine has one chlorine atom and one acid site like a conjugated secondary amine. Thus, an ion/ molecule reaction between a weak CH3OH acid and atrazine will mainly lead to deprotonated molecule. Considering the ion/molecule reactions between Cl- ion and these analytes, only dinoterbe reacts to form a deprotonated [M - H]- molecule (m/z 293), resulting from exothermic proton (19) Stemmler, H.; Hites, R. A. Biomed. Environ. Mass Spectrom. 1988, 17, 311. (20) (a) Gre´goire, S. Thesis, Universite´ Pierre et Marie Curie, 1997. (b) Mathurin, J. C. Thesis, Universite´ Pierre et Marie Curie, 1997. (21) (a) Busch, K. L.; Bursey, M. M.; Hass, J. R.; Sovocool, G. Appl. Spectrosc. 1978, 32, 273. (b) Stan, H.-J.; Kellner, G. Biomed. Mass Spectrom. 1985, 12, 695.

transfer (Table 2). This trend evidences the larger gas-phase acidity of dinoterbe (containing dinitrophenol moiety) than that of HCl and even greater than that of atrazine or trifluraline, since no reaction takes place with Cl- on these compounds. Moreover the presence of this specific [M - H]- ion in the NICI/Cl- mass spectrum is obviously helpful, when selective ion monitoring or MSn experiments are needed, due to the large sensitivity obtained with this single ion and without neighboring m/z ions. Otherwise, under selected NICI/CH3O- conditions, deprotonated [M - H]- molecules are the main ions displayed in all the mass spectra (Table 2). Very likely trifluraline and atrazine, being more acidic than methanol, give rise to the formation of the unambiguous diagnostic ions at m/z 334 and 214, respectively. Similarly, dinoterbe reacts to form [M - H]- at m/z 293. Another important specific ion is produced by the reaction of CH3O- with trifluraline, molecular M•- negative ion, is provided via charge exchange reaction due to the presence of nitro and fluorine groups enhancing the EA of these dinitroaromatic compounds compared to that of CH3O•. On the other hand, trifluraline and dinoterbe NICI/O2•- mass spectra demonstrate that these compounds specifically react to generate the formation of M•- ions (Table 2) because both have electron-withdrawing groups. One should note, as discussed before, that only the reaction of dinoterbe with Cl- leads to the [M - H]- ion without competition. Thus, very likely dinoterbe is not only a stronger acid than HCl but also HO2•. The results show no [M - H]- ion in the NICI/O2•- mass spectrum although HO2• is a very weaker acid than HCl [∆H°acid (HO2•) ) 1476 kJ‚mol-1 > ∆H°acid (HCl) ) 1395 kJ‚mol-1].12 In these experiments, the charge exchange reaction seems to be a more efficient process than for proton transfer, accounting for the absence of deprotonated molecule. Atrazine exhibits a significant signal at m/z 214 corresponding to the deprotonated [M - H]- molecule as observed during reaction with CH3O-. Atrazine is then more acidic than HO2• and CH3OH but does not contain any strong withdrawing electron group inducing any charge exchange reaction; thus, the proton-transfer process is the only observable pathway. Analytical Chemistry, Vol. 72, No. 20, October 15, 2000

5067

Scheme 2. r-β Elimination and SN2-C Mechanisms Proposed for the Reaction Induced by Methoxide and Alkyl Phosphonate Esters

Table 3. Major Ions (m/z) Observed in NICI-MS Reactions between Organophosphonate Compounds and CH3O- Methoxide organophosphonate diestersa DEMPa (Mw ) 152 u) MiPrMPb (Mw ) 152 u)

EiPrMPc (Mw ) 166 u)

DiPrMPd (Mw ) 179 u)

NICI/CH3O-b 151 [M - H]- (100) 137 [M - R + 14]- (14) 123 [M - R]- (58) 151 [M - H]- or/and [M - R2 + 14]- (100) 137 [M - R2]- (9) 123 [M - R1 + 14]- (5) 109 [M - R1]- (49) 165 [M - H]- (100) 151 [M - R2 + 14]- (18) 137 [M - R1 + 14]- (45) or/and [M - R2]- (45) 123 [M - R1]- (46) 179 [M - H]- (100) 151 [M - R + 14]- (22) 137 [M - R]- (95)

aRelative abundances of ions related to the base peak are noted in parentheses. b (a)RdC2H5,(b)R1)CH(CH3)2,R2)CH3,(c)R1)CH(CH3)2, R2)C2H5, and (d) R1)CH(CH3)2, R2)CH(CH3)2.

(b) Selective NICI Mass Spectra of Dialkyl Methyl Organophosphonate Diesters. Dialkyl methyl phosphonates are considered as representative weak toxic organophosphonate compounds (OPC) and could, therefore, be used as models for studying the gas-phase behavior of such substrata. Indeed, they contain acid-base sites (i.e., CH3 group versus PdO group) and present ambident electrophilic and nucleophilic characters, respectively. The acidic and electrophilic reactivities of methyl phosphonates toward the selected CH3O- negative ion, listed in Table 1, have been carefully scrutinized in order to investigate the driving forces, which guide the ion/molecule reactions yielding diagnostic ions. These latter are displayed in the NICI mass spectra as major ions (Table 3) and are regioselectively produced via bimolecular pathways. Their relative abundances depend on the substitutent group. For these methyl phosphonates, the formation of deprotonated [M - H]- molecules is the main favored channel (corresponding to the base peak), which involves the exothermic proton-transfer reaction due to the higher gas-phase acidity of phosphonates (∆H°acid values lower than 1500 kJ‚mol-1)22 than that of methanol 5068 Analytical Chemistry, Vol. 72, No. 20, October 15, 2000

[∆H°acid (CH3OH) ) 1590 kJ‚mol-1].12 Such chemical properties provide highly efficient proton transfers (kexp/kcoll ) 1)23 which allow formation of deprotonated [M - H]- molecules, useful for unambiguous molecular weight determination. Low-abundance diagnostic ions such as [M - R1]- {or/and [M-R2]-} (Scheme 2) and [M - R1 + 14]- (or/and [M - R2 + 14]-) (Scheme 3) are displayed in the NICI mass spectra of the studied OPC. Their formation cannot be attributed to the dissociation pathways of the excited [M - H]-* ions due to exothermicity of the proton-transfer reaction. Actually, the exothermicity is carried out by the lost neutral, i.e., CH3OH, where the chemical bond is created. Thus, the [M - H]- ions must be characterized by a weak internal energy which is not sufficient to promote fragmentations. Consequently, the formation of these [M - R]- and [M - R + 14]- ions must find their origins from bimolecular reactions rather than via dissociation processes. Note that recently, Steiner et al.24 described similar diagnostic ions which were generated by ion/molecule reactions occurring between DEMP and isopropyl alkoxide [(CH3)2CO]- reactant. The proposed explanation is based upon two possible bimolecular processes for the formation of the [M - R]- ions, either an R-β elimination25 or a SN2-C displacement as displayed in Scheme 2. Note that these two feasible pathways lead to diagnostic ions with minor abundance in these NICI mass spectra. Their reaction efficiencies do not correlate with the reaction thermochemistry (requiring exothermicity) but rather depend on the height of the intrinsic barrier (as tight transition state) between both minimums of the potential well for the two ion-neutral complexes as described by Brauman26,27 using a potential energy surface representation. It could be therefore assumed that E2 elimination and SN2-C displacement are depicted by a nonnegligible intrinsic (22) (a) Leroy, A.; Fournier, F.; Tabet, J. C.; Dissard, J.; Daoust-Maleval, I.; Tambute´, A. Proc. 43rd Conference on Mass Spectrometry and Allied Topics, Atlanta, GA, 1995; p 406. (b) Leroy, A. Thesis, Universite´ Pierre et Marie Curie, 1996. (23) (a) De Koning, L. J.; Nibbering, N. N. J. Am. Chem. Soc. 1988, 110, 2066. (b) Occhiucci, G.; Speranza, M.; De Koning, L. J.; Nibbering, N. N. J. Am. Chem. Soc. 1989, 111, 7387. (24) Steiner, V.; Daoust-Maleval, I.; Tabet, J. C. Int. J. Mass Spectrom., in press. (25) Su, T.; Bowers, M. T. Gas-Phase Ion Chemistry; Bowers, M. T., Ed.; Academic Press: New York, 1979; Vol 1. (26) Olmstead, W. N.; Brauman, J. I. J. Am. Chem. Soc. 1977, 99, 4219. (27) Brauman, J. I.; Kinetics of Ion Molecule Reactions; Ausloos, P., Ed.; Plenum Press: New York, 1979.

Scheme 3. Formation of Two Ion-Dipole Complexes c and d. Evolution toward [M - R + 14]- Ions

pentacovalent intermediate transition state29 a followed by the elimination of RO- alkoxide (addition-elimination process); a pentacoordinate transition state29 b with simultaneous P-OR bond cleavage and P-OCH3 bond formation (substitution process). This ion-dipole c can evolve into d by an internal protontransfer reaction prior to dissociation into several specific ions. The main observed pathway takes place through an exothermic internal proton abstraction since ROH alcohol is less acidic than (CH3O)PO(OR1)CH3: [∆H°acid (ROH) > ∆H°acid (methyl phosphonate)]. This stepwise process yielding ion-dipole complexes c and d will also lead to the formation of [M - R + 14]- ions. Such a nucleophilic substitution must be characterized by a nonnegligible intrinsic barrier related to the formation of the transition state a or b, leading to an inefficient ion/molecule reaction. CONCLUSION The formation, injection, and selection of reactant negative ions from a high-pressure ion source into a QITMS yield highly specific ion/molecule interactions via selective NICI-MS experiments and, thus, permit a greater control over the gas-phase competitive bimolecular mechanisms than occurs directly in a conventional high-pressure ion source. Upon the choice of the reactant negative ions, a low degree of fragmentation and direct confirmation of analyte molecular weights were achieved according to their chemical properties, i.e., gas-phase acidity, nucleophilicity, or/ and electron affinity. Phosphonate compounds mainly reacted via a proton-transfer process with CH3O- reactant. Other competitive reaction pathways have been revealed, like SN2-P, SN2-C, or R-β elimination processes. Nitrogen pesticides, depending upon the nature of the reactant ion (CH3O-, Cl-, or O2•-), led to two stable species, even-electron deprotonated [M - H]- molecules or oddelectron molecular M•- anion. Finally, organophosphorus pesticides showed the largest reactivity that easily permits identification for environmental applications. Indeed, several processes were observed, i.e., charge exchange, proton transfer, SN2-C, and R-β elimination.

barrier which is the bottleneck of the stepwise reactions, confering on them a relative low efficiency (kexp/kcoll , 1). On the other hand, formation of [M - R + 14]- ions occurs via a SN2-P displacement through the formation of ion-dipole complexes c28 and d, yielding an alkoxide RO- and (CH3O)PO(OR1)CH3 neutral (Scheme 3). Two probable mechanisms are proposed for the formation of both complexes c and d: a (28) Lum, R. C.; Grabowski, J. J. J. Am. Chem. Soc. 1993, 115, 7823.

ACKNOWLEDGMENT The authors acknowledge the Centre d′Etudes du Bouchet, the Centre de Recherches et d′Etudes de la Logistique de la Police Nationale, and Varian Associates for their financial and technical support. We also thank the CNRS and the University Pierre et Marie Curie. Received for review February 10, 2000. Accepted July 19, 2000. AC000172E (29) Lum, R. C.; Grabowski, J. J. J. Am. Chem. Soc. 1992, 114, 8619.

Analytical Chemistry, Vol. 72, No. 20, October 15, 2000

5069