Low-pressure chemical ionization in ion trap mass spectrometry

S. Bouchonnet, D. Libong and M. Sablier, Eur. J. Mass Spectrom. 10, 509–521 (2004)

509

Low-pressure chemical ionization in ion trap mass spectrometry

Stéphane Bouchonnet,* Danielle Libong and Michel Sablier Département de Chimie des Mécanismes Réactionnels, Ecole Polytechnique, route de Saclay, 91128 Palaiseau Cedex, France. E-mail: [email protected]

This article presents the specificities of low-pressure chemical ionization in ion trap mass spectrometry. One main feature is the ability to perform chemical ionization with liquid reagents as readily as with “conventional” gases (methane, isobutane and ammonia). The reactivities and analytical applications of gas and liquid reagents are summarized from literature data and are compared when possible. Keywords: ion trap mass spectrometry, chemical ionization, reagents

Introduction Mass spectrometry is well recognized as a highly sensitive and selective tool for the analysis of target compounds in forensic sciences.1 Over the wide range of existing mass spectrometers, the ion trap mass spectrometer has the advantage of offering versatile modes of operation associated with a well-recognized flexibility of use which allows, in particular, both single stage (MS) and multiple stages of analysis (MS/MS, MSn).2–4 Compared with sector or linear quadrupole instruments, the ion trap mass spectrometer generally affords higher sensitivity due to its capability to trap and accumulate ions. Moreover, the easy coupling of gas chromatography (GC) to these benchtop-like mass spectrometers offers a third-dimensional separation feature, permitting the detection of analytes from complex matrices within ultra-trace detection levels. As a consequence, ion trap mass spectrometers have found an increasingly widespread use in analytical laboratories. Two kinds of ion trap can be differentiated between, those using external ionization and those using in situ ionization. The latter are widespread in coupling with gas chromatography, mainly because they allow switching between electron ionization (EI) and chemical ionization (CI) in a few seconds (see below). Although adequate for many purposes, EI is not always the ideal ionization method, but it is the more routinely used because it permits automatic identification of compounds via database search. Molecular weight determinations are often more appropriately based on CI, which has the advantage of allowing better control (through the choice of reagent gases) of the internal energies deposited during the ionization processes. This paper

DOI: 10.1255/ejms.649

recalls the main features of chemical ionization using an ion trap mass spectrometer using in situ ionization and compares them to those of “classical” external sources of linear quadrupole instruments. Reactants that have been used in ion traps are listed and compared in terms of reaction efficiency, selectivity, applications, etc. Reactants were divided into two categories: “usual” gas reagents (methane, isobutane and ammonia) and “unusual” liquid reagents. Discussions on reagent ion selection and on the optimization of conditions for performing CI in the ion trap as well as comparative studies of different types of mass spectrometers have taken place early on.2,5–9 Instrumental The aim of this paper is obviously not to present in detail the theory of the ion trap mass spectrometer; we briefly recall the operating principles of in situ chemical ionization in ion trap mass spectrometry and develop the specificities of this ionization mode. As presented on Figure 1, an ion trap consists of three electrodes separated by two spacers. All internal surfaces of recent apparatus are passivated with a silicon dioxide layer of 300 to 400 Å to reduce the adsorption of polar compounds.10 For ion traps using in situ ionization, the end of the GC capillary column enters the ion trap through a hole in one of the spacers; analytes are eluted inside the analyzer and ionized between the electrodes. Whatever the ionization mode (EI or CI), electrons are produced from a metallic filament and sequentially introduced (at 70 eV) into the ion trap

ISSN 1356-1049

© IM Publications 2004

510

Low-Pressure Chemical Ionization in Ion Trap Mass Spectrometry

Figure 2. Typical scan function for operation of the ion trap in the CI mode. A = reagent ionization; B = ejection of unwanted ions (i.e. non-reagent ions); C = reaction of reagent ions with sample molecules; D = ejection of ions with m/z ratios lower than the lowest m/z ration of the analytical scan; E = acquisition of the spectrum.

Figure 1. Ion trap mass spectrometer using in situ ionization.

through an electron gate. In the CI mode, electron ionization of the reagent lasts between 10 and 2500 µs. The reagent molecular ions rapidly fragment or react with neutrals according to the reactions presented in the second part of this article. The reagent undergoes electron ionization at the same time as eluted molecules. Unwanted sample ions formed by electron ionization are ejected using a low frequency waveform, which is applied between end-cap electrodes. The reagent ions finally react with analyte molecules via reactions of protonation, hydride or alkyl abstraction, charge transfer or bonding. The time imparted for these reactions is generally between 1 and 130 ms. Figure 2 displays a typical scan function for the operation of the ion trap in the CI mode. Operated in the so-called mass-selective instability mode, in which only a radio frequency (rf) voltage is applied to the ring electrode when the end-cap electrodes are grounded, the ion trap presents the characteristic that the operating voltage determines the range of mass-to-charge ratios of ions trapped in the quadrupolar field. The gradual increase in this rf operating voltage results in the setting of a mass window wherein ions of increasing mass escape from the device and are detected on an external detector. For best performances, the motion of the ions stored within the trap is damped by a background pressure of helium typically set at a value of 10–3 Torr. Specificity of in situ chemical ionization in an ion trap mass spectrometer

ensure that a sufficient number of collisions occur to generate abundant reagent ions. At a pressure of 1 Torr, the reagent gas is generally 1000-fold or more in excess of the sample, precluding significant electron ionization of the analyte. A pressure of 1 Torr cannot be set in an ion trap: too many collisions and space charge effects would greatly disrupt ion trajectories. In 1972, Todd et al. explained that “CI should be observable in trapping devices, but at pressure some 10,000 times lower than those normally employed in conventional CI ion sources”.8 In an ion trap, CI is usually performed by using only 10–6 to 10–4 Torr of reagent; ion storage allows the CI process to be carried out within reaction periods in the millisecond range. Both the sample and reagent pressures are low, so that ample conversion of sample molecules to ions requires long reaction times. Adduct ions, such as (M + C2H5)+ and (M + C3H5)+ in the case of methane CI, which are frequently observed in high pressure sources, are not observed in ion trap experiments.5,6,11 This is one important feature of the comparison between CI mass spectra recorded with an ion trap and mass spectra recorded with a high pressure ion source. This is interpreted in terms of thermalization of ions: in the ion trap, the pressure of reagent gas is too low to provide enough stabilizing collisions to remove internal energy from adduct ions so that the latter dissociate. In the same way, protonated molecules are less stable in ion traps than in high pressure sources. This excess of internal energy leads to a slightly, but nevertheless significantly, increased fragmentation in ion traps. The main disadvantage of ion traps using in situ ionization is their incapacity to allow negative CI. Electron capture reactions cannot be performed because thermal electron and heavy negative ions cannot be trapped at the same time.12 Ability to switch between EI and CI during chromatographic separation

Low reagent pressure

In “traditional” external sources fitted with quadrupole and magnetic sector analyzers, the residence time of ions is very short, about 10–5 s. Under such conditions, a high pressure of reagent gas, of the order of 1 Torr, is required to

With mass spectrometers using an external source, the ionization chamber volume must be changed when changing the ionization mode. In fact, the ionization chamber is more gas-tight in CI than in EI, in order to retain a reagent gas pressure of about 1 Torr.13 With the ion trap, CI analysis


is as routine as EI; there is no change in ion source geometry.14 When introducing the CI reagent into the ion trap, the reagent pressure is stabilized at about 10–5 Torr within a few seconds. In the same way, the reagent is quickly pumped out when closing the valve. This permits a switch between EI and CI modes in a few seconds. Thus, a specificity of ion trap analyzers using in situ ionization is their ability to change the ionization mode during the chromatographic separation. This can be very helpful in analyzing mixtures of compounds (such as benzodiazepines, for example) among which some are better detected under EI while some others are better detected under CI conditions.15 Use of liquid reagents

Since the introduction of chemical ionization by Munson and Field in the 1960s, most CI experiments have been performed using methane, isobutane and ammonia.16 The specificities of each one of these gases are detailed below. In an ion trap, the low reactant pressure allows the use of reactant molecules difficult to introduce into high-pressure sources because of their low volatility. Reagents, which are in the liquid state under standard conditions of temperature and pressure, can thus be used. Methanol and acetonitrile are, by far, the most routinely used; default parameters for performing CI with each one of the two reagents are preset, for instance, in the Varian Saturn ion trap software.17 Other liquid reagents such as water, acetone or dimethyl ether have been tested; the main results are reported below. Performing chemical ionization with a liquid reagent costs less and is more convenient than with “classical” CI gas; it does not require a gas tank and gauge. A few millilitres of methanol or acetonitrile enable CI to be performed for several weeks. Given that the vapor pressure of the reagent must be sufficient at the laboratory temperature to be introduced into the mass spectrometer by vacuum aspiration, the partial pressure can simply be adjusted with a manual needle valve. Ability to select reagent ions

Unlike high-pressure sources where all the reagent ions produced under electron ionization can react with sample molecules, ion traps offer the advantage of allowing ion– molecule reactions with a mass-selected reagent ion or a mass-selected range of ions. This mass-selective capability is an important feature of ion trap mass spectrometry that has been extensively used to study fundamental aspects of CI.18,19 In the ion trap, the electron gate is switched on for a short period at the start of the reaction sequence during which both reagent and sample molecules undergo electron ionization. The sample-to-reagent gas ratio may even be 1 : 10 because space charge effects limit the total ion population. Thus, it must be kept in mind that El fragments from the sample can contribute significantly to the resulting CI mass spectrum. These analyte ions can also react with the neutral molecules of the reagent gas, providing “adulterated” CI spectra; this phenomenon cannot be avoided with high-pressure sources.20 The mass-selective storage capabilities of the ion trap may be

511

used to eliminate the problem of El interference in the resulting CI spectra. Cairns and co-workers showed that modifications to the scan function of the radio frequency applied to the ring electrode make it possible to produce CI spectra that are no longer adulterated by EI-like ions.21 It has also been demonstrated that non-resonance excitation can be used as a low-pass mass filter in the CI mode to eject undesired fragment ions that result from direct electron ionization.22 This makes it possible to produce a “classical” CI spectrum rather than a hybrid EI / CI spectrum. Effect of the damping gas

In all ion trap devices, helium is used as a damping gas to stabilize ion trajectories close to the center of the potential well. The buffer gas partial pressure is usually about 10–3 Torr. A study by Brodbelt et al. showed that helium plays a significant role in the CI process.5 Helium stabilizes the protonated molecule by collisional deactivation, leading to fragmentation ratios lower than those observed in the absence of damping gas. This thermalization effect has been confirmed by more recent studies.23 Usual reagents: methane, isobutane and ammonia Methane, isobutane and ammonia are the reagent gases that have been the most widely used since the development of chemical ionization in mass spectrometry. With the exception of a few articles, essentially by fundamental research teams, most authors do not justify the choice of the reagent gas. While CI (with a given reagent) is sometimes compared with EI in terms of sensitivity, there is not much comparison with the performances of CI reagents in the literature involving in situ ionization ion trap mass spectrometry. A general comparison of sensitivities of EI and CI cannot be significant, since it obviously depends on the reagent and on the analyte considered. Anyway, for a given compound and whatever the reagent, EI is more sensitive than CI, in most cases, when analyses are performed in the full scan mode.9,24,25 In “single storage” and “tandem” experiments, detection thresholds are often of the same order of magnitude for EI and CI; CI is often preferred for the analysis of matrix extracts because it allows discrimination against some low-basicity compounds of the sample matrix.26 There are a few papers where the comparison between EI and CI is notably in favor of CI in terms of sensitivity.23 Cooks and co-workers showed that, as in high pressure sources, the degree of fragmentation is coherent with the respective proton affinities (PA) of the reagent gas: with most organic compounds, ammonia27 (PA = 201 kcal mol–1) CI does not provide any fragment ion and isobutane27 (PA = 193 kcal mol–1) causes less fragmentation than methane28 (PA = 125 kcal mol–1).29 Comparing methane and isobutane CI, they pointed out that methane produces more absolute ion abundances but that the MH+ peak in the isobutane CI spectra is of much greater relative abundance than that in the

512


Figure 3. Relative occurence of reagent gases in the literature.

methane CI spectra. Providing an ionic current that is mainly constituted of MH+ ions makes isobutane a reagent of choice for the analyst. It allows direct access to the molecular weight of the compound and also provides ideal conditions to perform sensitive MS/MS or MSn experiments. However, other studies show that charge exchange is the dominant process observed using isobutane with some compounds such as polychlorinated biphenyls.30 Fales et al. showed strong differences between the mass spectra resulting from ionization of sulfur S8 with methane, isobutane, ammonia and water.31 Methane produces an abundant MH+ ion and numerous fragment ions derived from the (M + C2H5)+ adduct ion. Using isobutane or ammonia, the main process is charge exchange to produce S8+•, while water CI mainly provides an abundant S8H+ ion. Figure 3 compares the relative occurrences of gas reagents in the literature. Methane

Among the three “conventional” reagent gases, methane is undoubtedly the one which has been the most studied from a fundamental point of view. Formation of reagent ions involves Reactions (1)–(7). CH4 + e– → CH4+• (m/z 16) + 2 e–

(1)

CH4+• → CH3+ (m/z 15) + H•

(2)

CH4+• → CH2+• (m/z 14) + H2

(3)

CH4+• + CH4 → CH5+ (m/z 17) + CH3•

(4)

CH2+• + CH4 → C2H3+ (m/z 27) + H2 + H•

(5)

CH3+ + CH4 → C2H5+ (m/z 29) + H2

(6)

C2H3+ + CH4 → C3H5+ (m/z 41) + H2

(7)

In the absence of a reliable pressure measurement, the methane pressure in the ion trap is usually estimated by the relative abundances of m/z 16, 17 and 29 ions. At a partial pressure of 1–2 × 10–5 Torr, the ratios of the signal intensities of m/z 17 to 16 and m/z 17 to 29 are about 10 : 1 and 1 : 1, respectively.32 The mass-selective capability of the ion trap allowed Tsuji et al. to study extensively the reactivity of

CH5+, C2H5+ and C3H5+ ions with many organic molecules such as substituted benzenes, olefins, paraffins and C8–C18 alcohols.33 Since the proton affinity of CH4 (5.7 eV) is smaller than those of C2H4 (7.1 eV) and C3H4 (8.0 eV),34 more excess energies are released in the CH5+ reactions. In a general way, CH5+ ions exclusively induce proton transfer and dissociative proton transfer; this can be explained by the instability of the [M + CH5]+ adduct ion. The formation of [M + C2H5]+ and [M + C3H5]+ adduct ions is assumed to result from radiative association. The formation of such complexes depends on M. For example, while such adducts are observed with anilines, nitrobenzene and benzonitrile, they are not observed with benzene and toluene. It is to be noted that, in a high pressure methane CI source, benzene and toluene provide [M + C2H5]+ and [M + C3H5]+ adduct ions, as reported by Munson and Field.16 At a partial CH4 pressure of 10–5 Torr, secondary collisions with methane are not important enough to ensure collisional stabilization of complexes. The reactivities of CH5+, C2H5+ and C3H5+ ions obviously depend on the analyte considered and it is impossible to summarize here all the studies that have been performed. The case of olefins (CxH2x, x = 8 – 18) clearly illustrates the differences of reactivities between the three ions: the major reactions for the formation of CyH2y+1+ ions are proton transfer to a C=C bond and hydride abstraction from the alkyl portion in the CH5+ reactions, proton transfer to a C=C bond in the C2H5+ reactions and addition to a C=C bond in the C3H5+ reactions, while the major reactions for the formation of CyH2y–1+ are hydride ion abstraction in the CH5+ reactions, alkanide ion abstraction in the C2H5+ reactions and addition to a C=C bond in the C3H5+ reactions.33 Studying methane CI of pesticides in an ion trap, Cairns and co workers showed the co-existence of charge transfer and protonation reactions, resulting in an EI component superimposed on the CI spectrum.35 March and co-workers observed the same phenomenon with polychlorinated biphenyls and showed that methane CI with mass-selected C2H5+ ions provides a mass spectrum with a negligible amount of fragment ions and a negligible influence of charge exchange reactions.30,36 Dorey showed very important differences between alkane spectra recorded with a conventional source and with methane low pressure ionization. The increased fragmentation observed in ion trap spectra is assumed to be a consequence of the high kinetic energy of CI reagent ions (above 10 eV), as compared with conventional ion sources (less than 1 eV).37 From an analytical point of view, methane CI mass spectrometry is performed in all fields of analytical chemistry. It has been used in environmental analysis to detect pollutants such as polychlorinated biphenyls,30,36 N-nitrosodimethylamine,32 alkylbenzenesulfonates and their degradation products.23 In toxicology, it has been shown that the limits of detection of methamphetamines are of the same order for methane CI performed on underivatized compounds and for EI performed on HFBA (heptafluorobutyric anhydride) derivatives.38 Methane CI has also been used in pyrolysis-GC-MS


to establish the chemical composition of plants having antibacterial properties.39

Isobutane

Whereas a few fundamental articles about high pressure isobutane CI are available in the literature,40–42 to our knowledge fundamental aspects of formation and storage of isobutane reagent ions in ion traps have not been reported. Isobutane is not often used with traditional quadrupole analyzers because the presence of the m/z 57 reagent ion requires ion scanning above m/z 58. Moreover, this hydrocarbon tends to induce remanence effects in the mass spectrometer. The ion trap apparatus allows selective ejection of reagent ions before m/z scanning and the use of low partial pressure of C4H10 considerably reduces remanence effects. At a partial pressure 2 × 10–5 Torr, isobutane provides m/z 43 and m/z 57 ions in a 1 : 2 ratio.43 m/z 43 results from CH3• elimination from C4H+•10 while m/z 57, the t-butyl ion C4H9+, is assumed to result from hydride abstraction from C4H10 by the primary fragment ion.42 Reagent ions are formed according to Reactions (8)–(10). C4H10 + e– → C4H10+• (m/z 58) + 2 e–

(8)

C4H10+• → C3H7+ (m/z 43) + CH3•

(9)

C3H7+ + C4H10 → C3H8 + C4H9+ (m/z 57)

(10)

Schollenberger and co-workers showed that isobutane improves the stability of mass spectra compared with methanol for the detection of trichothecenes (secondary metabolites of mycotoxins).44 For butyltin compounds, it has been shown that CH4 and C4H10 provide very similar spectra, with a sharp decrease in the sensitivity of detection using isobutane.45 For pesticides, for which molecular ions are less fragmented with C4H10, the comparison of CH4 CI and C4H10 CI sensitivities is in favor of the latter.46 Studying methylenedioxymetamphetamine and its chiral metabolites, De Boer et al. demonstrated that isobutane CI simultaneously induces proton transfer and charge transfer reactions in the ion trap.47 The literature shows that isobutane is today the reagent that has been the most widely used in CI ion trap mass spectrometry. Although the choice of C4H10 is usually not rationalized by the authors, this success can be explained by the fact that isobutane is the reagent that generally provides the greatest amount of MH+ ions. In a recent study, we showed that this is due to the combination of two factors: the formation of reagent ions is energetically of low cost and the storage of reagent ions is more efficient than with other “classical” reagents.11 Isobutane CI has been used in environmental analysis for the detection of insecticides48,49 and pesticides.50–52 It has also been widely used in medical and veterinary research as well as in analytical toxicology for the detection of propafenone (antiarrhythmic agent)53 animal drug residues,54 cocaine,55 cannabidiol56 and their

513

metabolites. Isobutane CI has been successfully performed in biological studies to detect trichothecenes57 and juvenile insect hormones.43 It has also been performed on chemical warfare agent simulants such as diisopropylmethylsulfonate and chloroethylsulfide.26 Ammonia

A study by Creaser and co-workers illustrates the great interest of combining CI and ion trap mass spectrometry for petroleum fraction analysis.58 Nitrogen and sulfur heterocycles are selectively detected in petroleum fractions by NH4 – CI / ion trap MS. In ammonia CI, both NH3•+ and NH4+ reagent ions are formed. NH3•+ results from electro-ionization of NH3 and NH4+ results from the reaction between NH3•+ and NH3 to give NH2• and NH4+ [reactions (11)–(12)]. NH3 + e– → NH3+• (m/z 17) + 2 e–

(11)

NH3+• + NH3 → NH4+ (m/z 18) + NH2•

(12)

Whereas NH3•+ undergoes charge–exchange reactions with aromatic hydrocarbons (the ionization potential of NH3, 10.16 eV, is higher than that of most organic compounds), NH4+ is unreactive toward a wide range of hydrocarbons. The use of an ion trap mass spectrometer makes it possible to create a large population of NH4+ ions while minimizing the amount of NH3•+ ions, thus allowing the selective detection of nitrogen and sulfur compounds. In spite of the great selectivity that it provides, ammonia is definitely less used than methane and isobutane. This could be due to its corrosiveness which tends to reduce the filament life-time. NH3 CI was successfully used in studies about the toxicities of a widely encountered terpene, d-limonene, and one of its main metabolites, perillyl alcohol in rat and human plasmas.59,60 Liquid reagents Positive-ion CI has the advantage of offering a powerful tool in increasing selectivity for selected classes of compounds by using reagents of different proton affinities and reaction properties. Ion traps that enable the trapping of ions over a relatively long period enormously increase the effectiveness of potential CI reagent systems that are barely employed, if ever. As stated above, one advantage of CI is its ability to control the amount of energy deposited into the ion (and thereby the degree of fragmentation) by selection of the reagent gas. Different reagent compounds can be used in CI experiments to selectively ionize the analytes in correlation with their thermochemical criteria.42 Consequently, the choice of an appropriate reagent permits the analysis of certain classes of compounds in a mixture. In addition to the “classical” gaseous CI reagents currently used for chemical

514


CH3CN•+ → [CH2CN]+ (m/z 40) + H•

(14)

CH3CN•+ + CH3CN → [CH3CNH]+ (m/z 42) + H2CN]• (15) [CH2CN]+ + CH3CN → [C3H4N]+ (m/z 54) + HCN

Figure 4. Relative occurrences of liquid reagents in the literature.

ionization, varieties of unusual reagents have been developed for specific applications in CI mass spectrometry and, more recently, in ion trap mass spectrometry.61 The criteria for choosing a CI reagent lies in its availability as a pure compound and its relative inertness toward the analytes to preclude the appearance of secondary ions, making it difficult to interpret the resulting mass spectra. While proton transfer CI reagents are obviously the most commonly used for the identification of volatile analytes, an alternative method to provide direct information on the molecular weight of the analytes under soft CI conditions results in the binding of a relatively inert reagent ion to the sample. Complexes, consequently, are formed with a correlated representative mass peak in the resulting mass spectrum. Trimethylsilyl adduct cations have been used in this way as selective reagents.61 The use of specific charge exchange reagents illustrates fairly well the possibility offered by performing selective ionization of certain compounds in a mixture while not ionizing other components under chemical exchange processes. Benzene and carbon disulfide have been used as ionizing reagents for the analysis of hydrocarbons in petroleum fractions under low-energy ionization conditions.62 An increasing number of studies have focused on the investigation of alternative CI reagents and, obviously, this growing interest in characterizing new reagents is correlated to the increasing availability and developing capabilities of ion trap mass spectrometers. The section below reports the fundamental studies and analytical applications of unusual CI reagents in ion trap mass spectrometry. Acetonitrile and, to a smaller extent, methanol are preferable in routine analysis while other liquid reagents are preferable in fundamental research. Figure 4 compares the relative occurrences of liquid reagents in the literature. Acetonitrile

The proton affinity of acetonitrile is 788.3 kJ mol–1.42 The ionization of acetonitrile in ion-trap mass spectrometry provides four principal ionic species m/z 40 [C2H2N]+, m/z 41 [C2H3N]+•, m/z 42 [C2H4N]+ and m/z 54 [C3H4N]+, according to Reactions (13)–(16). CH3CN + e– → CH3CN•+ (m/z 41) + 2 e–

(13)

(16)

At 10–5 Torr, the relative abundance of ions as follows: 100% for m/z 42, 16% for m/z 54 and m/z 41 and 8% for m/z 40. Acetonitrile CI provides a useful tool for the detection of alkanes in mass spectrometry, which is difficult for many reasons. Electron ionization in the case of long-chain alkanes and alkenes has severe limitations, due to the low abundance of molecular ions and to the formation of fragment ions that are not always easily related to the original structure. The abundance of the molecular ion is usually lower than 1% with an EI ion source at high temperature but it increases when the temperature decreases. In chemical ionization, hydrocarbons are not easily protonated due to their low proton affinity. Moneti and Pieraccini identified long-chain hydrocarbons by acetonitrile CI in an ion trap.63 They showed that [M + C2H2N]+ ions are easily produced and generate adducts with long-chain saturated hydrocarbons which are much more stable than MH+ ions. Other reagent ions do not complex with alkanes. Long-chain saturated alcohols were also studied by acetonitrile chemical ionization. The competition between the protonation of alcohols and the formation of [M + C2H2N]+ ions is in favor of protonation. This has been confirmed by isotopic labeling.64–66 Plzak et al. compared methane, isobutane, acetonitrile and methanol for routine trace analysis of organotins (wood preservatives and agrochemicals stabilizers).45 Methane CI and isobutane CI provide spectra similar to those obtained by EI. Methanol CI provides spectra characterized by [M – H]+ ions. Acetonitrile CI provides simple mass spectra, characterized by the [M + C2H2N]+ ions, which can be used for complementary spectral information to EI. More, acetonitrile provides the best sensitivity. Although m/z 54 is of small abundance (16%) in the ionic plasma, many studies demonstrate its use in locating the double-bond position in unsaturated hydrocarbons.67 Because of their low proton affinities, chemical ionization of alkenes does not provide [MH]+ ions, only two ionic species: [M + C2H2N]+ and [M + C3H4N]+. Subsequent fragmentation by collision-induced dissociation of the [M + 54]+ leads to two fragment ions which allow one to establish the localization of the double bond in the original molecule. Functionalized long-chain of unsaturated alcohols and esters were studied in the same manner.67–69 In contrast to hydrocarbons, a competition between adduct formation and protonation is observed in favor of the latter with m/z 41 and m/z 42 ions. Raising the radio frequency value above m/z 50 allows the [C3H4N]+ reagent ions to be trapped selectively and permits the double bond localisation to be focused. [C3H4N]+ ion reagents are used routinely for the localisation of double bond of natural fatty acid methyl esters and some MS/MS libraries of unsaturated frames are available.70 This technique has some


limitations; it does not provide information on the cis/trans geometry and has not been evaluated for the identification of branches in the fatty acid chains. Acetonitrile CI is routinely used in environmental71,72 and biochemical73,74 trace analysis. It is also widely used in toxicology, for the detection of drugs such as benzodiazepines15 and illicit drugs such as cocaine,75 methadone76 and lysergic acid diethylamide (“LSD”).77 Comparing methane and acetonitrile for the detection of arecoline in the treatment of Alzheimer disease, Shetty et al. showed that acetonitrile chemical ionization permits hydride abstraction from arecoline that is observed with methane to be limited significantly.73 Methanol

The proton affinity of methanol is 761.1 kJ mol–1.42 Under EI, methanol provides two principal ionic species: the molecular ion m/z 32 [CH3OH]+• and m/z 33, [CH3OH2]+, which results from H• abstraction from CH3OH by m/z 32 [Reactions (17)–(18)]. CH3OH + e– → CH3OH+• (m/z 32) + 2 e–

(17)

CH3OH+• + CH3OH → [CH3OH2]+ (m/z 33) + [CH2OH]• (18) At 10–5 Torr, the relative abundances of m/z 33 and m/z 32 are about 10 : 1.17 The proton affinity of methanol is close to that of acetonitrile and both reagents are generally used in the same fields of application. Methanol CI provides simple spectra with a very abundant [MH]+ ion and no adduct ions. Methanol is often used for the detection of polar molecules and some applications are found, for example, with organotins.45 Dimethyl ether, ethers

Dimethyl ether is certainly the most versatile unusual CI reagent that has been applied in ion trap mass spectrometry. Two main reagent ions are generated under CI conditions: protonated dimethyl ether (m/z 47) and methoxymethyl ether ion [CH3O = CH2]+ (m/z 45). This latter ion induces predominantly methylene addition with compounds possessing electron-releasing substituents, whereas compounds bearing electron-withdrawing functions lead to adduct formation via a methyl addition reaction. A detailed investigation of ion–molecule reactions of dimethyl ether with simple aromatic compounds (hydroxyacetophenone, methoxyacetophenone, methoxyphenol, hydroxybenzaldehyde, anisaldehyde and vanillin) has been conducted by Brodbelt et al. and has furnished useful insights on the effective adduct formation yields compared with a conventional quadrupole ion source.78 As a CI reagent, dimethyl ether has demonstrated both functional group and positional selectivity in the ion trap through the reaction of the methoxymethyl ether cation, m/z 45, which induces selective formation of [M + 13]+ and [M + 15]+ adducts with the aromatic compounds investigated and offering a structural selectivity specific to low pressure

515

CI.78 These selective ion–molecule reactions revealed how the reagent cations can be used to derivatize a molecule and alter its dissociative behavior in order to obtain structural information through collisionally-activated dissociation experiments. The ion–molecule reactions of dimethyl ether ions with model hydroxylated and non-substituted 1,4 benzodiazepines have shown their analytical applications in probing structural differences in the 1,4 benzodiazepine series by the detection of the characteristic [M + 13]+ products with the hydroxylated molecules whereas [M + 15]+ adducts appeared with the non-hydroxylated benzodiazepines.79 On the other hand, ion–molecule reactions of 1,4-benzodiazepines with protonated ethylene oxide cations generated from electron ionization of ethylene oxide did not offer competitive product formation pathways or reaction site selectivity through the dissociation of the observed [M + 45]+ adduct ions.79 Similarly, examination of ortho effects in the collisionallyactivated dissociation processes of closed-shell aromatic ions confirmed the selectivity of ion–molecule reactions of dimethyl ether ions.80 The reactions of dimethyl ether ions with amino alcohols were compared with both an ion trap mass spectrometer and a quadrupole mass spectrometer.81 Ion–molecule reactions of the amino alcohols with dimethyl ether ions yielded two predominant products [M + 1]+ and [M + 13]+ resulting from protonation and from methylene substitution to neutrals, respectively. Collisionally-activated dissociation studies on the [M + 45]+ adduct observed under high pressure stabilizing conditions in the quadrupole ion source confirmed that this adduct leads to the only observed [M + 13]+ product under low pressure conditions in the ion trap where the intermediate [M + 45]+ adduct was not detected. Investigations of the competition between reactive sites in simple monofunctional and bifunctional molecules revealed a net preponderance for the protonation reaction with simple alcohols, whereas monofunctional amines, diamines81 and diols82 formed significant amounts of [M + 13]+ ions and related dissociation products. On the contrary, the intermediate [M + 45]+ adducts were detected during ion–molecule reactions of dimethyl ether ions with lactams in an ion trap, confirming its precursor role in the formation of methylene substitution and methyl cation transfer products, [M + 13]+ and [M + 15]+, respectively.83 Dependence on ring size and substituent positions were observed for the formation of [M + 13]+ and [M + 16]+ ions, contrary to the protonation and methyl cation transfer reaction. The methylated lactone ion [M + 16]+ was reported for the first time and was attributed to a reaction product of the [M + 45]+ adduct through a ringopening, methylene imine elimination and subsequent cyclisation sequence.83 Comparison of reactions under chemical ionization conditions in an ion trap and a tandem quadrupole mass spectrometer showed the formation of the same adduct ions except for an [M + 47]+ ion in the quadrupole ion source attributed to a loosely proton-bound complex between the lactam and the dimethyl ether that could not be observed under the low collisional stabilization conditions of the ion trap.83 Extended to the example of lactone compounds, com-

516


parison between the types of adducts observed with both apparatus revealed that one could not expect a differentiation between lactones and lactams using high-pressure CI compared with the low-pressure conditions in use within the ion trap. The analysis of hydantoin and succinimide based anticonvulsants was carried out by Shen and Brodbelt with the aim of characterizing both derived compounds by the mean of protonation reaction, [M + 1]+ adduct, or methylene substitution, [M + 13]+ adduct.84 Phenyl substitution on the skeletal structure appeared to enhance the intrinsic basicities of the nitrogen and carbonyl oxygen atoms in the hydantoin and succinimide rings, while the phenyl ring might deactivate those positions for methylene substitution. Collisionallyactivated dissociation conducted on both adducts showed similar dissociation pathways uncorrelated to the position of the attached proton or methylene group.84 Reactions of dimethyl ether ions with polycyclic aromatic hydrocarbons (PAH) were elucidated by isolating the major reagent ions generated during ionization of dimethyl ether in the ion trap and reacting them with neutral PAH.85 Nevertheless, the authors concluded that the use of dimethyl ether as a CI reagent does not provide useful information for the differentiation of the investigated PAH isomers. In the course of their studies on ion–molecule reactions in an ion trap for the development of improved CI processes, Alvarez and Brodbelt investigated a series of homologous ether reagent cations generated from dimethyl ether, di-n-butyl ether and 2-methoxyethanol as potential CI reagents.86 The reagent ion, CH3OCH2CH2OCH2+, generated from ionization of the methoxymethanol precursor, showed the greatest promise in the absence of any observed selectivity during the analysis of nucleoside antibiotics with dimethyl ether and di-n-butyl ether. The reactive species, CH3OCH2CH2OCH2+ (m/z 89), presented several advantages as a reagent cation due to its structural similarity with the methoxymethylene ion [CH3O = CH2]+ previously investigated and its polarizability related to its larger size inducing a greater interaction energy on the formation of ion–molecule complexes with substrates.86 The reaction of the CH3OCH2CH2OCH2+ ion resulted in the formation of [M + 13]+ and [M + 89]+ adducts covalently bound where the nucleosidic moiety was shown to dominate both reactive and dissociative behaviors of the investigated nucleoside antibiotics.

the field of ion trap mass spectrometry and only two studies related to the use of NO+ as a CI reagent have been reported in the scope of this review. Nitric oxide CI in a glow discharge ionization source coupled to quadrupole ion trap has been evaluated for the characterization of automotive exhausts and used as a diagnostic method for the determination of hydrocarbons at the parts-per-billion levels.87 This technique was found to yield diagnostic ions for alkanes, alkenes and arenes in a standard gas mixture, and its ability to measure a variety of hydrocarbons present at 15 ppb was demonstrated making the combination of NO+ CI with ion trap attractive for tailpipe exhaust measurements. Nitrosonium ion CI has been implemented in an ion trap for the characterization of epoxides through the collisionally-activated dissociation of the hydride abstraction [M – H]+ product ion generated under chemical ionization conditions.88

Nitric oxide

Nitric oxide has a long history as a reagent gas for analytical mass spectrometry and, in particular, is well-known to be successful for the differentiation of primary, secondary and tertiary alcohols of aldehydes from ketones and of cycloalkanes from alkanes.61 One inherent advantage in using nitric oxide as a CI reagent with an ion trap lies in the fact that unavoidable filament consumption due to the contact of NO gas, a strong metal oxidizing agent, in the ion source, is eliminated in such a device, inducing a more practical use of NO, pure or diluted, for routine analysis. Nevertheless, analytical applications of nitrosonium ion CI are very scarce in

Vinyl methyl ether

Vinyl methyl ether is known to react through bimolecular processes leading to four-membered cyclic intermediate ions which can be used for the location of olefinic double bonds in mono- and poly-unsaturated compounds.61,88 However, the use of pure vinyl methyl ether as reagent gas under high-pressure CI conditions allows the formation of charged protonated bound dimers and polymeric ions which complicate the interpretation of mass spectra.61 Vinyl methyl ether plasmas, resulting from CI conditions, are known to be complicated by the diversity of chemical processes involved.88 As a consequence, isolation of the vinyl methyl ether radical cation, through a scanning sequence of the ion trap, simplifies the resulting CI spectra. Vinyl methyl ether was shown to yield molecular ions, resulting from charge transfer process, and [M + 15]+ adduct ions, characteristic of a methyl transfer, with α,β-unsaturated epoxides providing, in combination with the MSn capabilities of the ion trap, an efficient selectivity for this class of compounds.88 Vinyl methyl ether has been proposed as an alkene-selective CI reagent for the implementation of a fast-response detection method for the measurement of atmospheric isoprene with a membrane introduction system coupled to an ion trap.89 Ionized vinyl methyl ether molecular ion was proposed to react with the isoprene through a Diels–Alder cycloaddition process to form an unstable methoxycyclohexene radical cation which further fragments to lose methanol to generate an adduct ion [M + 58 – 32]+•.89 Quantitation of isoprene and related compounds was based on the intensity measurement of the resulting methyl loss fragment ion (m/z 79) upon collisionally-activated dissociation. For further applications, it was proposed that the vinyl methyl ether CI technique be applied to other atmospheric biogenic compounds such as methyl vinyl ketone, methacrolein and terpenes. Acetone

Acetone does not commonly appear as a usual CI reagent but has been reported for the characterization of mon-


osaccharides with a high pressure CI source61 or the characterization of codeine and methadone metabolites using GC/CI ion trap mass spectrometry.90–92 In ion trap mass spectrometry, acetone has been reported as an unusual CI reagent for the characterization of vincamine and has shown similarities in the resulting CI spectra with vinyl methyl ether CI since they both led to an [M + 43]+ adduct ion attributed to acetyl ion addition on the nitrogen position of the pyrole ring skeletal.93 This [M + 43]+ adduct was proposed to dissociate into protonated vincamine through a ketene loss, or into an [M + 43 – 18]+ ion through a water loss. However, dimethyl ether has proved to be a more efficient reagent for the effective detection of vincamine at the parts-per trillion level under CI conditions.93 Trimethyl borate

Trimethyl borate is well-known as a selective CI reagent for the characterization of stereochemical isomers of vicinal diols.61 The Lewis acid character of the dimethoxyborinium ion, easily generated by electron ionization of trimethylborate, has naturally induced the investigation on trimethylborate as a potential selective CI reagent for biologically-active molecules presenting functional groups of Lewis basicity.94 The borinium ion adducts included an [M + 73]+ corresponding to the binding of one dimethoxyborinium ion to the neutral

517

and an [M + 41]+ ion resulting from a loss of methanol from the preceding adduct. Collisionally-activated dissociation of these ions provided a large number of fragments in a limited yield that may, however, serve as effective fingerprints for compound identification during routine procedures.94 Pentafluorobenzyl alcohol

In recent years, there has been a growing interest in characterizing carboxylic acid products arising from the photo oxidation of isoprene and aromatics due to their potential influence on the natural acidity of rains and cloud waters. However, the identification of these unknown carboxylic acids can be difficult due to the lack of sensitive and selective techniques. Original methods for the detection of atmospheric chemistry derived products are still under investigation as proven, for example, by the proposed detection of isoprene by the use of vinyl methyl ether as a CI reagent.89 Recently, analysis of airborne carboxylic acids and phenols has been developed through the detection of their pentafluorobenzyl derivatives using gas chromatography/ion trap mass spectrometry with the pentafluorobenzyl alcohol as a CI reagent.95 The analytes are firstly derivatized with pentafluorobenzyl bromide to generate the corresponding esters and ethers which are subsequently separated by gas chromatography and analyzed in an ion trap. Pentafluorobenzyl

Table 1. Articles dealing with the analysis of compounds of interest in medicine, toxicology or biochemistry as a function of the reagent used

Acetaminophen

Methane

Isobutane

Ammonia

Acetonitrile

Methanol

11

11

11

11

11

Acetone

Antibiotics

83, 86

Anticonvulsants

84

Arecoline Benzodiazepines

73 11

Cannabidiol

11

11

11, 15

79

75

Codeine

90 5 59

Limonene 43

Insect juvenile hormones LSD Methamphetamines

11

56

Cocaine

Ephedrine

Dimethyl ether

77 38

47

518


Table 2. Articles dealing with the analysis of compounds of environmental interest as a function of the reagent used.

Methane Isobutane Ammonia Acetonitrile Methanol Acrilonitrile Dimethyl Pentafluoro- Vinylether benzyl methyl alcohol ether Alkylbenzenes

97

Alkylbenzenesulfonates

23

Butyltin compounds

45

45

45

45 95, 96, 98–101

Carbonyl compounds Chloroethylethylsulfide

26

Diisopropylmethyl phosphonate

26

Dioxines and furans

4

Fuels, petroleum fractions

57

Nitrosodimethylamine

32

Pesticides / insecticides

21, 35, 46

Plant extract Polychlorinated biphenyls

70

46 48–52

46, 71

102

48

39 30, 36

Vincamine

93

alcohol CI mass spectra provided a direct identification of the molecular masses of these products by the presence of an intense [M + 181]+ adduct ion resulting from the addition

of the pentafluorobenzyl cation (m/z 181) which represents the main ion obtained through CI conditions in the ion trap.95 The use of pentafluorobenzyl alcohol as a CI rea-

Table 3. Articles dealing with some organic compounds as a function of the reagent used.

Methane Acetic anhydride

5

Alcohols

5

Alkenes

33(c) 33(g)

Amines

5

Butyraldehyde

5

Cyclic aromatic compounds, PAH Cyclohexanone and derivatives

Isobutane

Acetonitrile Dimethyl ether

68 5

81, 82

5

85

5 69

Hydrocarbons and parafins

33(f)

65

Long chain alcohols

33(i)

66, 67

31

Vinylmethyl ether

87

88, 89

Halogenated reagents

5

Fatty acid methyl esters

Sulfur

Nitric oxide

81, 82

5, 18, 20, 33(a– d), 33(h) 5

Ammonia

31

31

102, 104


gent was initiated by the success encountered in the use of pentafluorobenzylhydroxylamine to derivatize airborne carbonyls subsequently detected by traditional GC/methane CI mass spectrometry.96,97 Hydroxy carbonyls were similarly measured in air by sampling aqueous solutions of trapped carbonyls derivatized with pentafluorobenzylhydroxylamine and reacted with bis(trimethylsilyl)-trifluoroacetamide using ion trap mass spectrometry.98,99 Ultra-trace levels of detection were achieved for the hydroxy carbonyls by the use of pentafluorobenzyl alcohol CI reagent which has shown an increase in molecular and pseudo-molecular ion signals in the resulting mass spectra when compare with methane CI results. Combination of EI, methane CI and pentafuorobenzyl alcohol CI mass spectra has permitted the first report for the detection of hydroxyacetone in an ambient atmosphere sample. The adaptability of pentafluorobenzyl derivatization coupled to GC/ion trap mass spectrometry has been demonstrated for the identification of oxygenated polar organics.100,101 The presence of [M – H]+, [M]+•, [M + H]+ and [M – 181]+ in the pentafluorobenzyl alcohol CI mass spectra was shown to provide a unique pattern which facilitates molecular weight determinations. Acrylonitrile, dichloromethane and halogenated reagents

Acrylonitrile and dichloromethane have been investigated for their properties as specific reagents for the characterization of isomers of chlordane compounds.102 These CI reagents reacted principally through a retro-Diels–Alder reaction, followed by HCl elimination in the case of dichloromethane which allows the differentiation of cis/trans isomers. Limits of detection at the pg level were achieved with heptachlor with both CI reagents in the selected ion monitoring mode, while limits of detection proved to be more favorable with acetonitrile compared with acrylonitrile and dichloromethane in the case of chlordane and nonachlor.102 Ion–molecule reactions between PAHs and selected positive reactant ions of dichloromethane, chloroform, carbon tetrachloride, 1,1-dichloroethane, difluoromethane103 and 1,1-difluoroethane103,104 have been employed to differentiate series of structural isomers of PAHs. Main product ions resulted from the formation of an adduct ion with the selected ions from the halocarbon precursors and from the elimination of an HX molecule (X = Cl, F) from these adducts. 1,1difluoroethane was the haloethane reactant which provided the largest number of different adduct-forming reactive ions, CH3CHF+ and CH3CF2+, respectively.104 Table 1 provides direct access to references of articles dealing with the analysis of compounds of interest in medicine, toxicology and biochemistry as a function of the reagent used. Table 2 provides direct access to references of articles dealing with the analysis of compounds of environmental interest as a function of the reagent used. Table 3 provides direct access to references of articles dealing with the analysis of some organic compounds as a function of the reagent used.

519

References 1. J. Yinon, Forensic Applications of Mass Spectrometry. CRC Press, Boca Raton, FL, USA (1995). 2. R.E. March and J.F.J. Todd, Practical Aspects of Ion-Trap Mass Spectrometry Vols 1–3. CRC Press, Boca Raton, FL, USA (1995). 3. J.F.J. Todd and A.D. Penman, Int. J. Mass Spectrom. Ion Processes 106, 1 (1991). 4. R.E. March, J. Mass Spectrom. 32, 351 (1997). 5. J.B. Brodbelt, J.N. Louris and R.G. Cooks, Anal. Chem. 59, 1278 (1987). 6. S.M. Boswell, R.E. Mather and J.F.J. Todd, Int. J. Mass Spectrom. Ion Processes 99, 139 (1990). 7. J.V. Johnson, R.A. Yost, P.E. Kelley and D.C. Bradford, Anal. Chem. 62, 2162 (1990). 8. R.F. Bonner, G. Lawson and J.F.J. Todd, J. Chem. Soc. Chem. Commun. 11, 79 (1972). 9. B.A. Eckenrode, S.A. McLuckey and G.L. Glish, Int. J. Mass Spectrom. Ion Processes 106, 137 (1991). 10. D. Libong, S. Pirnay, C. Bruneau, F. Rogalewicz, I. Ricordel and S. Bouchonnet, J. Chromatogr. A 1010, 123 (2003). 11. D. Libong, M. Sablier and S. Bouchonnet, submitted for publication. 12. J.R. Chapman, Practical Organic Mass Spectrometry, 2nd Edn. John Wiley & Sons Ltd, Chichester, UK (1993). 13. A.A. Mosi and G.K. Eigendorf, Curr. Org. Chem. 2, 145 (1998). 14. B. Brittain and C. Feigel, Am. Lab. 26, 44 (1994). 15. S. Pirnay, I. Ricordel, D. Libong and S. Bouchonnet, J. Chromatogr. A 954, 235 (2002). 16. M.S.B. Munson and F.H. Field, J. Am. Chem. Soc. 88, 2621 (1966). 17. Varian, Advanced MS Techniques SATURN 2000 GC/MS (1996). 18. D.W. Berberich, M.E. Hail, J.V. Johnson and R.A. Yost, Int. J. Mass Spectrom. Ion Processes 94, 115 (1989). 19. R.E. March and R.J. Hughes (Eds), Quadrupole Storage Mass Spectrometry. John Wiley & Sons, New York, USA (1989). 20. K.J. Hart, S.A. McLuckey and G.L. Glish, J. Am. Soc. Mass Spectrom. 3, 549 (1992). 21. T. Cairns, K.S. Chiu, E. Siegmund and M. Weber, Rapid Commun. Mass Spectrom. 6, 449 (1992). 22. M. Wang, S. Schachterle and G. Wells, J. Am Soc. Mass Spectrom. 7, 668 (1996). 23. W.H. Ding, J.H. Lo and S.H. Tzing, J. Chromatogr. A 818, 270 (1998). 24. P.G. Vanysacker, H.G.M. Janssen, H.M.J. Snijders, P.A. Leclercq, C.A. Cramers and H.J.M. Vancruchten, J. Microcolumn Sep. 5, 413 (1993). 25. R.L. Fitzgerald, C.L. O’Neal, B.J. Hart, A. Poklis and D.A. Herold, J. Anal. Toxicol. 21, 445 (1997). 26. M.V. Buchanan, R.L. Hettich, J.H. Xu, L.C. Waters and A. Watson, J. Hazard. Mater. 42, 49 (1995). 27. W.J. Richter and H. Schwarz, Angew Chem., Int. Ed. Engl. 17, 424 (1979).

520


28. R. Walder and J.L. Franklin, Int. J. Mass Spectrom. Ion Phys. 85, 636 (1980). 29. J.N. Louris, R.G. Cooks, J.E.P Syka, P.E. Kelley, G.C. Stafford and J.F.J. Todd, Anal. Chem. 59, 1677 (1987). 30. M. Lausevic, J.B. Plomley, X. Jiang, R.E. March and C.D. Metcalfe, Eur. Mass Spectrom. 1, 149 (1995). 31. H.M. Fales, Q.L Pu and R.T. Mason, Int. J. Mass Spectrom. Ion Processes 106, 273 (1991). 32. J.B. Plomley, C.J. Koester and R.E. March, Anal. Chem. 66, 4437 (1994). 33. (a) M. Tsuji, E. Oda and Y. Nishimura, Chem. Lett. 781 (1997); (b) M. Tsuji, T. Arikawa and Y. Nishimura, Bull. Chem. Soc. Jpn 72, 293 (1999); (c) M. Tsuji, Y. Tanaka, T. Arikawa and Y. Nishimura, Chem. Lett. 892 (2000); (d) M. Tsuji, T. Arikawa and Y. Nishimura, Bull. Chem. Soc. Jpn 73, 131 (2000); (e) M. Tsuji, T. Arikawa, Y. Tanaka and Y. Nishimura, Bull. Chem. Soc. Jpn 73, 1673 (2000); (f) Y. Tanaka, M. Tsuji and Y. Nishimura, Bull. Chem. Soc. Jpn 73, 2703 (2000); (g) Y. Tanaka and M. Tsuji, Bull. Chem. Soc. Jpn 74, 839 (2001); (h) Y. Tanaka and M. Tsuji, Bull. Chem. Soc. Jpn 75, 241 (2002). (i) Y. Tanaka and M. Tsuji, Bull. Chem. Soc. Jpn. 75, 1729 (2002). 34. S.G. Lias, J.E. Bartmess, J.F. Liebman, J.L. Holmes, R.D. Levin and W.G. Mallard, J. Phys. Chem. Ref Data 17, Suppl. 1 (1988). 35. T. Cairns, K.S. Chin, E. Siegmund and M. Weber, Rapid Commun. Mass Spectrom. 6, 331 (1992). 36. T.M. Vasiljevic, M.D. Lausevic and R.E. March, J. Serb. Chem. Soc. 65, 431 (2000). 37. R.C. Dorey. Org. Mass Spectrom. 24, 973 (1989). 38. A.H.B. Wu, T.A. Onigbinde, S.S. Wong and K.G. Johnson, J. Anal. Toxicol. 16, 137 (1992). 39. G.C. Galletti, G. Chiavari and Y.D. Kahie, Rapid Commun. Mass Spectrom. 7, 651 (1993). 40. M.S.B. Munson, J.L. Franklin and F.H. Field, J. Phys. Chem. 68, 3098 (1964). 41. F.H. Field, J. Am. Chem. Soc. 91, 2829 (1969). 42. A.G. Harrison, Chemical Ionization Mass Spectrometry. CRC Press, Boca Raton, Florida, USA (1992). 43. P.E.A. Teal, A.T. Proveaux and R.R. Heath, Anal. Biochem. 277, 206 (2000). 44. M. Schollenberger, U. Lauber, H. Terry Jara, S. Suchy, W. Drochner and H.M. Müller, J. Chromatogr. A 815, 123 (1998). 45. Z. Plzak, M. Polanska and M. Suchanek, J. Chromatogr. A 699, 241 (1995). 46. G.C. Mattern, G.M. Singer, J.B. Louis, M. Robson and J.D. Rosen, J. Agric. Food Chem. 38, 402 (1990). 47. D. De Boer, L.P. Tan, P. Gorier, R.M.A. Van de Wal, J.J. Kettenes-Van den Bosch, E.A. De Bruijn and R.A.A. Maes, J. Mass Spectrom. 32, 1236 (1997). 48. G.F. Bauerle, Jr, K.L. Ray and J.S. Brodbelt, Anal. Chim. Acta 317, 137 (1995). 49. Y. Hirahara, Y. Tsumura, Y. Nakamura, Y. Tonogai and T. Shibata, J. Food Protect. 60, 305 (1997). 50. K.S. Roinestad, J.B. Louis and J.D. Rosen, J. Assoc. Off. Anal. Chem. Int. 76, 1121 (1993).

51. T. Miyoshi, Y. Hirahara, M. Narita, K. Okamoto, M. Miyata, S. Koiguchi, M. Hasegawa, K. Kamakura, T. Yamana and Y. Tonogai, J. Food Hyg. Soc. Japan 36, 274 (1995). 52. G.C. Mattern, J.B. Louis and J.D. Rosen, J. Assoc. Off Anal. Chem. 74, 982 (1991). 53. M.S. Leloux and R.A.A Maes, Biol. Mass Spectrom. 20, 382 (1991). 54. S.A. Barshick and M.V. Buchanan, J. Assoc. Off. Anal. Chem. 77, 1428 (1994). 55. M.R. Harkley, G.L. Henderson and C.H. Zhou, J. Anal. Toxicol. 15, 260 (1991). 56. P. Consroe, K. Kennedy and K. Schram, Pharmacol. Biochem. Behaviour 40, 517 (1991). 57. K.S. Roinestad, T.J. Montville and J.D. Rosen, J. Agric. Food Chem. 41, 2344 (1993). 58. C.S. Creaser, F. Krokos, K.E. O’Neil, M.J.C. Smith and P.G. McDowell, J. Am. Soc. Mass Spectrorn. 4, 322 (1993). 59. H. Chen, K.K. Chan and T. Budd, J. Pharm. Biomed. Anal. 17, 631 (1998). 60. Z. Zhang, H. Chen, K. Chan, T. Budd and R. Ganapathi, J. Chromatogr. B 728, 85 (1999). 61. M. Vairamani, U.A. Mirza and R. Srivinas, Mass Spectrom. Rev. 9, 235 (1990). 62. S.G. Roussis, W.P. Fitzgerald and A.S. Cameron, Rapid Commun. Mass Spectrom. 12, 373 (1998). 63. H. Wincel, R.H. Fokkens and N.M.M. Nibbering, Int. J. Mass Spectrom Ion Processes 96, 321 (1990). 64. H. Wincel, Int. J. Mass Spectrom. Ion Processes 175, 283 (1998). 65. G. Moneti and G. Pieraccini, Rapid Commun. Mass Spectrom. 10, 167 (1996). 66. G. Moneti, G. Pieraccini, D. Favretto and P. Traldi, J. Mass Spectrom. 34, 1354 (1999). 67. G. Moneti, G. Pieraccini, F. Dam, S. Turillazi, D. Favretto, P. Traldi, J. Mass Spectrom. 32, 1371 (1997). 68. N.J. Oldham and A. Svatos, Rapid. Commun. Mass. Spectrom. 13, 331 (1999). 69. A.L. Michaud, G.Y. Diau, R. Azbril and J.T. Brenna, Anal. Biochem. 307, 348 (2002). 70. A.J. Schubert and C.G. Enke, Anal. Chim. Acta 211, 83 (1988). 71. A.J.H. Louter, J. Van Doornmaien, J.J. Vreuls and U.A.T. Brinkman, J. High Resol. Chromatogr. 19, 679 (1996). 72. K. Schwadorf and H.M. Muller, Chromatographia 32, 137 (1991). 73. H.U. Shetty, E.M. Daly, N.H. Greig, S.I. Rapoport and T.T. Soncrant, J. Am. Soc. Mass Spectrom. 2, 168 (1991). 74. H.U. Shetty, H.W. Holloway and S.I. Rapoport, Anal. Biochem. 224, 279 (1995). 75. C. Girod and C. Staub, Forensic Sci. Int. 107, 261 (2000). 76. C. Girod and C. Staub, Forensic Sci. Int. 117, 175 (2001). 77. D. Libong, S. Bouchonnet and Y. Ricordel, J. Anal. Toxicol. 27, 24 (2003). 78. J. Brodbelt, C.C. Liou and T. Donovan, Anal. Chem. 63, 1205 (1991).


79. T.D. McCarley and J.S. Brodbelt, Anal. Chem. 65, 2380 (1993). 80. T.D. McCarley and J.S. Brodbelt, Org. Mass Spectrom. 27, 9 (1992). 81. E.S. Eichmann and J.S. Brodbelt, J. Am. Chem. Soc. Mass Spectrom. 4, 230 (1993). 82. E.S. Eichmann, E. Alvarez and J.S. Brodbelt, J. Am. Chem. Soc. Mass Spectrom. 3, 535 (1992). 83. T.D. McCarley and J.S. Brodbelt, J. Am. Chem. Soc. Mass Spectrom. 4, 352 (1993). 84. J.X. Shen and J. Brodbelt, J. Mass Spectrom. 31, 1389 (1996). 85. A.A. Mosi, R.H. Skelton and G.K. Eigendorf, J. Mass Spectrom. 34, 1274 (1999). 86. E.J. Alvarez and J.S. Brodbelt, J. Mass Spectrom. 30, 625 (1995). 87. M.A. Dearth, K.G. Asano, K.J. Hart, M.V. Buchanan, D.E. Goeringer and S.A. McLuckey, Anal. Chem. 69, 5121 (1997). 88. M. Sharif and J. Einhorn, Rapid Commun. Mass Spectrom. 11, 1185 (1997). 89. A. Colorado, D.J. Barket, J.M. Hurst and P.B. Shepson, Anal. Chem. 70, 5129 (1998). 90. S.P. Gygi, R.E. Joseph, E.J. Cone, D.G. Wilkins and D.E. Rollins, Drug Metab. Dispos. 24, 495 (1996). 91. D.G. Wilkins, P.R. Nagasawa, S.P. Gygi, R.L. Foltz and D.E. Rollins, J. Anal. Toxicol. 20, 355 (1996). 92. D.E. Rollins, D.G. Wilkins and G.G. Krueger, Eur. J. Clin. Pharmacol. 50, 391 (1996). 93. G.F. Bauerle, B.J. Hall, N.V. Tran and J.S. Brodbelt, J. Am. Soc. Mass Spectrom. 7, 250 (1996).

521

94. E.C. Kempen and J.S. Brodbelt, J. Mass Spectrom. 32, 846 (1997). 95. C.J. Chien, M.J. Charles, K.G. Sexton and H.E. Jeffries, Environ. Sci. Technol. 32, 299 (1998). 96. J. Yu, H.E. Jeffries and R. Le Lacheur, Environ. Sci. Technol. 29, 1923 (1993). 97. J. Yu, H.E. Jeffries and K.G. Sexton, Atm. Environ. 31, 2261 (1997). 98. R.G. Spaulding, X. Rao and M.J. Charles, Anal. Chem. 71, 3420 (1999). 99. R.S. Spaulding, R.W. Talbot and M.J. Charles, Environ. Sci. Technol. 36, 1798 (2002). 100. P. Frasey, X. Rao, R. Spaulding, B. Beld and M.J. Charles, Int. J. Mass Spectrom. 190/191, 343 (1999). 101. H. Destaillats, R.S. Spaulding and M.J. Charles, Environ. Sci. Technol. 36, 2227 (2002). 102. Z. Zencak, M. Oehme and S. Skopp, Rapid Commun. Mass Spectrom. 15, 1719 (2001). 103. A.A. Mosi, W.R. Cullen and G.K. Eigendorf, J. Mass Spectrom. 32, 864 (1997). 104. A.A. Mosi, W.R. Cullen and G.K. Eigendorf, J. Mass Spectrom. 33, 250 (1998).

Received: 19 January 2004 Revised: 15 February 2004 Accepted: 18 February 2004 Web Publication: 26 March 2004