Mobile phase influence on electrospray ionization for the analysis of smokeless powders by gradient reversed phase high-performance liquid chromatography-ESIMS

Mobile phase influence on electrospray ionization for the analysis of smokeless powders by gradient reversed phase high-performance liquid chromatography-ESIMS

Forensic Science International 154 (2005) 159–166 www.elsevier.com/locate/forsciint Mobile phase influence on electrospray ionization for the analysi...

315KB Sizes 0 Downloads 27 Views

Forensic Science International 154 (2005) 159–166 www.elsevier.com/locate/forsciint

Mobile phase influence on electrospray ionization for the analysis of smokeless powders by gradient reversed phase high-performance liquid chromatography-ESIMS$ John A. Mathis1, Bruce R. McCord* Ohio University, Department of Chemistry and Biochemistry, Clippinger Laboratories, Athens, OH 45701, USA Received 23 March 2004; received in revised form 4 October 2004; accepted 4 October 2004 Available online 9 December 2004

Abstract The solution parameters associated with gradient elution reversed phase high-performance liquid chromatography (HPLC) were characterized by evaluating the electrospray ionization mass spectrometry (ESIMS) response of selected smokeless powder additives. Using direct liquid infusion, the positive ion ESIMS responses were determined for three general classes of powder constituents: dialkylphthalate acid esters, N,N0 -dialkyl-N,N0 -diphenyl urea based stabilizers and nitroso-, nitro-, and dinitro- derivatives of diphenylamine. The relative ESIMS intensities of the powder components were investigated as a function of three solution parameters: ammonium acetate concentration, pH, and percent methanol. The effect of the ammonium acetate concentration demonstrates that the electrolyte is required for efficient ionization and the ESIMS intensity was optimal at a concentration of 2 mM for the selected compounds, except 2,40 -dinitrodiphenylamine. The aqueous solution pH, corresponding to the available protons in solution, did not have a significant effect on the ESIMS intensity of the analytes. The percent methanol was evaluated with both decreasing and constant electrolyte concentrations to demonstrate the effects of droplet stability and ion transfer into the gas phase. These findings were applied to the analysis of the selected smokeless powder additives using HPLC– ESIMS to illustrate increased sensitivity for the protonated molecules. # 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Forensic; Electrospray ionization mass spectrometry; Gradient; Smokeless powders

1. Introduction The use of electrospray ionization mass spectrometry (ESIMS) is becoming an important technique in forensic $ Research presented from Doctoral Dissertation at Ohio University, Department of Chemistry and Biochemistry, Clippinger Laboratories, Athens, OH 45701, USA. * Corresponding author. Present address: Department of Chemistry, Florida International University, Miami, FL 33199, USA. E-mail addresses: [email protected] (J.A. Mathis), [email protected] (B.R. McCord). 1 Present address: Noramco Inc. World Headquarters, 1440 Olympic Drive, Athens, GA 30601, USA.

applications involving the analysis of toxicological and explosive samples [1–12]. The selection of the optimal ESIMS detection parameters involves consideration of analyte characteristics, instrument operation, and in regard to reversed phase high-performance liquid chromatography (HPLC), the mobile phase composition [10,13–16]. In HPLC–ESIMS applications, certain solution parameters associated with the mobile phase influence the ESIMS signal. These include the organic modifier which facilitates chromatographic resolution and the electrolyte which promotes ionisation. The physical properties of the solution, such as volatility, surface tension, viscosity, conductivity, and dielectric constant are associated with the mechanisms by which ions are produced during electrospray ionization.

0379-0738/$ – see front matter # 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2004.10.008

160

J.A. Mathis, B.R. McCord / Forensic Science International 154 (2005) 159–166

The production of ions by electrospray ionization has been described by two conventional mechanisms: charged residue model and ion evaporation theory [17,19,20]. The charged residue model proposes that a droplet fissioning process called Coulomb fission, results from solvent evaporation and an increasing surface to volume ratio. The droplet charge density increases due to Coulomb fission, which continues until no further solvent evaporation occurs and the remaining particles contain only a single ionic species. The charged residue model has been shown to be responsible for the production of large multiply charged ions, such as proteins [17,21]. A second mechanism, ion evaporation theory, maintains that during the droplet desolvation process, the surface charge density increases. This process offsets the surface tension of the solvent until coulombic repulsion exceeds ion adhesion and charged particles are released from the droplet surface. Ion evaporation theory has been used to show the effect of the electrolyte level by comparing the response of several analytes over a large range of electrolyte concentrations [20]. The results demonstrated a practical electrolyte concentration range for analyte quantitation from 108 to 105 M. In addition, these studies illustrated that ionization suppression occurs at increased electrolyte concentrations [18,20,22]. A third proposed mechanism known as the equilibrium partitioning model, is similar to ion evaporation theory in that ions are generated directly from the droplet surface [23]. The equilibrium partitioning model differs from the ion evaporation theory in it’s assumption that ions partition between the charged exterior (surface) and neutral interior of the droplet, which exist as two distinct phases. According to the equilibrium partitioning model, the ions observed in the mass spectrum depend upon several solution parameters such as polarity of the solvent and solute ion. For example, the nonpolar character of the side chain groups in small peptides has been related to ESIMS response [24]. This relationship was extended to correlate the ESIMS response of small peptides to their retention characteristics in reversed-phase HPLC, where the intensity increased with the hydrophobicity of the molecules [25]. Additional physical properties of the mobile phase composition, such as pH have been shown to affect the electrospray process [14,21,26]. The pH influences the ESIMS signal with respect to both solution phase equilibria and gas phase processes. For example, the protonation of caffeine was used to demonstrate a process termed ‘‘wrong-way ‘round ionization’’ by Zhou and Cook, which involves the ionization of weakly basic neutral analytes resulting in the production of positive ions from strongly basic solutions [27]. Although the ESIMS mechanisms are well established, their relationship has not been correlated with the forensic application of gradient elution HPLC–ESIMS. Additionally, gradient elution HPLC is often applied without taking into account the inherent changes in the mobile phase composition and their effects on the ESIMS intensity. In the present

study, the effects of the solution parameters associated with mobile phase composition during gradient elution were investigated. The positive ion ESIMS responses of selected smokeless powder components were evaluated as function of the ammonium acetate electrolyte concentration, pH, and percent methanol using direct liquid infusion. To demonstrate the effects of solution parameters in this study, the results of the infusion experiments were applied to the reversed phase gradient elution HPLC–ESIMS method by comparing the responses using two different mobile phase compositions. The initial method was developed using 1 mM ammonium acetate in the aqueous fraction of the mobile phase, which results in decreasing levels of the electrolyte during the gradient elution method. In the second method, an ammonium acetate concentration of 2 mM was held constant during the gradient run. Additionally, to demonstrate an innovative application of the HPLC–ESIMS method, the analysis of different unburned smokeless powders was performed.

2. Materials 2.1. Chemicals Ammonium acetate (CH3COONH4), reagent grade acetic acid (CH3COOH), ammonium hydroxide (NH4OH), and HPLC grade methanol (CH3OH) were obtained from Fisher Scientific (Pittsburgh, PA). Milli-Q purified water was used throughout the experimental procedures. Analytical standards; 4,40 -dinitrodiphenylamine (4,40 DNDPA), 2,40 -dinitrodiphenylamine (2,40 DNDPA) (Cerilliant, Austin, TX), N-nitrosodiphenylamine (NsDPA) (Fluka, Milwaukee, WI), diphenylamine (DPA), 4-nitrosodiphenylamine (4sDPA), 2-nitrodiphenylamine (2NDPA), 4-nitrodiphenylamine (4NDPA), methyl centralite (MC), ethyl centralite (EC), dimethylphthalate (DMP), diethylphthalate (DEP), and dibutylphthalate (DBP) (Acros, Morris Plains, NJ) were used as received. 2.2. Smokeless powders The smokeless powder samples were used as received (unburned) and selected from various manufacturers including IMR (Plattsburgh, NY), Accurate Arms Company (McEwen, TN), Hodgden (Shawnee Mission, KS), and Alliant (Radford, VA). These samples were selected from different manufactures as well as different lots of the same smokeless powder.

3. Methods 3.1. Instrumentation 3.1.1. Electrospray ionization mass spectrometry A Bruker Esquire (Bruker Daltonics, Bremen, Germany) electrospray ionization interface with a quadrupole ion-trap

J.A. Mathis, B.R. McCord / Forensic Science International 154 (2005) 159–166

161

mass spectrometer was used with direct liquid infusion (Cole–Parmer Infusion Pump, Vernon Hills, IL). Mixtures of standards in methanol/aqueous ammonium acetate were infused at a flow rate of 20 mL/min. in the positive ion mode for the detection of protonated molecules [M + H]+. The optimized ESIMS settings were held constant throughout all experiments: ESI needle (4 kV), capillary temperature (250 8C), nebulizing gas (N2, 40 psi), drying gas (N2, 5 L/ min.) and ion optics: capillary offset voltage (35 V), skimmer 1 (25 V), skimmer 2 (6 V), octapole (2.4 V), lens 1 (5 V) and lens 2 (60 V).

3.1.4. Smokeless powders analysis The smokeless powder sample preparation method developed for GC–MS analysis was adapted for use with HPLCUV [11,28]. Individual smokeless powders samples were prepared by extracting 5 mg of the unburned powder with 250 mL methylene chloride overnight. A 20 mL aliquot was removed into a clean vial and evaporated under a stream of nitrogen gas. The samples were reconstituted with 40 mL methanol. An injection volume of 5 mL was used for the quantitative HPLC–ESIMS analysis.

3.1.2. Solution parameters Analytical standards containing DPA (5 mg/mL), NsDPA (1.5 mg/mL), 4NDPA (1.5 mg/mL), 2,40 DNDPA (5 mg/mL), MC (0.5 mg/mL), EC (0.5 mg/mL), DMP (1.5 mg/mL), and DBP (0.5 mg/mL) comprised a stock solution that was diluted 1:4 in each sample. The ESIMS responses of the selected compounds were evaluated as protonated molecules by direct liquid infusion of discreet samples at the selected factor levels. The solution parameters investigated include the ammonium acetate concentration, pH, and percent methanol. To evaluate different electrolyte concentrations, the samples were prepared in 50% methanol with varying levels of aqueous ammonium acetate. The effect of the available protons was evaluated by pH adjustment of the aqueous solution followed by dilution to 50% methanol. The pH of the aqueous fraction was adjusted by the addition of 1% (v/v) acetic acid or ammonium hydroxide with 2 mM ammonium acetate to maintain a constant electrolyte concentration without correction for dilution in methanol. The effect of the percent methanol was investigated by using variable and constant ammonium acetate concentrations.

4. Results and discussion

3.1.3. HPLC conditions A Hewlett-Packard 1100 HPLC system (Agilent, Palo Alto, CA) controlled by Chemstation software (LC–MS ver.A.06.03) was used with a Restek Pinnacle octyl column; ˚ (C-8), 2.1 mm  100 mm, 3 mm particle size, and 120 A average pore size at room temperature. The HPLC detection system consisted of a variable wavelength UV detector, 230 nm, coupled to the Bruker ESIMS operated in the positive ion mode. The mobile phase consisted of methanol (MeOH) and aqueous phases with two different electrolyte compositions. The gradient was initially developed using 1 mM ammonium acetate in the aqueous fraction, which decreased during the course of the gradient elution profile. Alternatively, an ammonium acetate concentration of 2 mM in both the organic and aqueous phases was used to provide a constant level of electrolyte. A linear gradient of 50–95% MeOH in 25 min was used at a flow rate of 0.25 mL/min. Standard mixtures with concentrations between 0.05–20 mg/ mL were prepared for the validation studies. An autosampler injection volume of 5 mL was used for the HPLC–ESIMS analysis.

4.1. Electrospray ionization mass spectrometry response The influence of the HPLC mobile phase on the ESIMS response involves a variety of factors. To investigate the effects of the solution parameters associated with the gradient elution mobile phase, the ESIMS responses of the selected analytes were evaluated as a function ammonium acetate concentration, aqueous solution pH, and percent methanol. The influence of the ammonium acetate concentration was evaluated at different methanol levels with both decreasing and constant electrolyte concentrations in order to investigate the interaction of these two factors. The direct liquid infusion experiments were designed with the sample preparation and run order randomized to evaluate the ESIMS peak intensities as a function of each solution parameter. 4.2. Effect of ammonium acetate concentration Initial studies on the analysis of smokeless powder components by HPLC–ESIMS and ESIMS/MS in the positive ion mode illustrated the detection of certain compounds but did not report the use of electrolytes to promote ionization [8,9,29]. Other investigations reporting the use of HPLC–ESIMS for smokeless powder analysis, including negative ion ESIMS studies, utilized ammonium acetate at arbitrarily chosen concentrations [2,8,11,29]. It is expected that ammonium acetate is required and promotes ionization at an optimal concentration. Furthermore, both the ion evaporation theory and equilibrium partitioning model describe mechanisms that rationalize the use of electrolytes for efficient ion production by electrospray ionization [18,22]. In order to evaluate the reversed phase HPLC–ESIMS method, the ESIMS peak intensities of selected smokeless powder additives were determined as a function of the methanol gradient by infusion experiments with different levels of methanol and 1 mM ammonium acetate in the aqueous fraction of the mobile phase. This approach yields increasing percent methanol and decreasing electrolyte concentrations during the analysis. The decreasing ammonium acetate concentration had a significant effect on the response of the selected smokeless powder components as shown in

162

J.A. Mathis, B.R. McCord / Forensic Science International 154 (2005) 159–166

Fig. 1. ESIMS response of smokeless powder additives as a function of the percent methanol (lower scale) and decreasing ammonium acetate concentrations (upper scale).

Fig. 1. The ESIMS signal of the smokeless powder additives decreased with decreasing ammonium acetate concentrations. The decreasing ESIMS response is attributed to the fact that the electrolyte is below the optimal concentration at increasing methanol levels and decreasing ammonium acetate concentrations. Therefore, the effect of the ammonium acetate concentration alone was investigated. Fig. 2 shows the ESIMS response of the selected analytes at different ammonium acetate concentrations in 50% methanol/water. Ammonium acetate was required for efficient ionization as shown by the increased intensity compared to 0 mM and was optimal for most compounds at 2 mM, with the exception of 2,40 DNDPA. Additionally, at higher ammonium acetate concentrations, the intensity decreases due to ionization suppression [18,22].

As shown in Figs. 1 and 2, changes in the percent methanol and ammonium acetate concentration affected the intensity of 2,40 DNDPA differently compared to the other analytes. The effect of the solution parameters on the ESIMS response of 2,40 DNDPA can be explained by the electrospray mechanisms involved in ion production. Fig. 1 shows that the ammonium acetate concentration from 1.5 to 0.5 mM did not have a significant effect on the ESIMS response of 2,40 DNDPA. This result follows the ion evaporation theory in that the droplet stability and thus 2,40 DNDPA surface activity is reduced with ammonium acetate concentrations below an optimal level. In addition, the formation of an intramolecular hydrogen bond between the oxygen of the 2-nitro group and the amine hydrogen would cause an apparent decrease in the polarity of 2,40 DNDPA within

Fig. 2. ESIMS response of smokeless powder additives at different ammonium acetate concentrations in 50% methanol/water.

J.A. Mathis, B.R. McCord / Forensic Science International 154 (2005) 159–166

163

Fig. 3. ESIMS intensity of the selected analytes illustrating the effect of the aqueous solution pH in 50% methanol/2 mM aqueous ammonium acetate.

the electrosprayed droplet. This hypothesis is consistent with the equilibrium partitioning model which has demonstrated increased ESIMS intensity for more hydrophobic molecules [30]. 4.3. Effect of solution pH An alternative explanation of the decreased ESIMS response shown in Fig. 1 is that the signal decreases with the water content or the available protons in the infused sample. Thus, the effect of the aqueous solution pH was investigated in 50% methanol/2 mM aqueous ammonium acetate. As shown in Fig. 3, the pH of the aqueous solution did not have a significant effect on the ESIMS intensity of the selected analytes as indicated using two-way ANOVA (p = 0.91). The negligible effect of pH on the ESIMS intensity illustrates that the protonation of the selected analytes may be more dependent upon the gas-phase proton affinity than the level of available protons in solution. For example, the responses of 4NDPA and 2,40 DNDPA slightly increased with increasing pH. Because these compounds are weakly basic in solution, the increased ESIMS intensity with pH can be described as following the ‘‘wrong-way ‘round ionization’’ process illustrated for the production of protonated species from strongly basic solutions [17,27,31]. 4.4. Effect of percent methanol As shown in Fig. 1, the ESIMS intensity decreased with increasing percent methanol and was correlated with the decreasing ammonium acetate concentration in the aqueous fraction of the mobile phase. Thus, the influence of the two factors was confounded and the effect of the percent metha-

nol alone was not obvious. Following the determination of the optimal ammonium acetate concentration, the effect of the percent methanol was investigated with a constant electrolyte level of 2 mM. Fig. 4 shows the ESIMS intensity as a function of the methanol level with 2 mM ammonium acetate. The ESIMS signal generally increases with the percent methanol. Previous reports regarding the ESIMS response as a function of organic solvent levels attributed variations in the intensity to the surface activity and ion mobility of the analyte within the electrosprayed droplet [25,32]. In the present study, the electrolyte level is the prominent factor in both cases of increasing percent methanol. At decreasing ammonium acetate concentrations in Fig. 1, the electrolyte level was below the optimal level, which reduced droplet stability and analyte transfer into the gas phase. Furthermore, the intensity increase shown in Fig. 4 is credited to the reduced surface tension of the solution whereas the ammonium acetate concentration of 2 mM maintained droplet stability. 4.5. Optimized conditions for HPLC–ESIMS The influence of the ammonium acetate concentration, aqueous solution pH, and percent methanol have been evaluated as a function of the solution parameters associated with the methanol gradient. The decreasing electrolyte concentration had a significant effect on the ESIMS intensity of the selected smokeless powder components as shown in Fig. 1. The ESIMS response was enhanced with the ammonium acetate concentration at 2 mM, the pH of the aqueous solution had a negligible effect on the intensity, and the response increased for the selected analytes with increasing methanol levels at a constant electrolyte concentration as shown in Figs. 2–4, respectively. These results demonstrate

164

J.A. Mathis, B.R. McCord / Forensic Science International 154 (2005) 159–166

Fig. 4. The ESIMS intensity of smokeless powder additives as a function of percent methanol with the concentration of ammonium acetate held constant at 2 mM.

that pH control is not required while maintaining the ammonium acetate concentration at 2 mM. The effects of these solution parameters were tested with the validation of the gradient HPLC–ESIMS method using two different mobile phase compositions. Initially 1 mM ammonium acetate in the aqueous fraction was used in the gradient elution method resulting in decreasing levels of the electrolyte. To evaluate the effects of solution parameters in this study, an ammonium acetate concentration of 2 mM in both organic and aqueous portions of the mobile phase was used with no pH control. While the decreasing electrolyte concentrations during gradient elution represents the addition of ammonium acetate to the aqueous fraction of the mobile phase, the constant ammonium acetate concentration corresponds adding the electrolyte to both organic and aqueous phases.

The analysis of standard mixtures of the selected analytes was used for HPLC–ESIMS method validation. Linear regression was performed using the extracted ion chromatogram peak areas to establish the sensitivity of the HPLC– ESIMS method from the slope of the linear equations for each smokeless powder additive. The sensitivities of the regression lines using the two different mobile phase compositions are compared in Fig. 5. The method sensitivity using 2 mM total ammonium acetate concentration in the mobile phase was superior for the selected analytes, which was verified using two-way ANOVA (p = 0.02). The inset in Fig. 5 illustrates the increased HPLC–ESIMS response using 2 mM total ammonium acetate concentration for the extracted ion chromatograms for 2,40 DNDPA at 10 mg/ mL. The increased sensitivity confirms the inferences drawn

Fig. 5. Comparison of method sensitivities (slope of the regression line) using 1 mM ammonium acetate in the aqueous fraction and 2 mM total ammonium acetate concentration in both organic and aqueous portions of the mobile phase. The inset illustrates the increased ESIMS response using 2 mM total ammonium acetate concentration for the extracted ion chromatograms for 2,40 DNDPA.

J.A. Mathis, B.R. McCord / Forensic Science International 154 (2005) 159–166

165

Table 1 Composition (%) of unburned smokeless powder samples determined using methylene chloride extraction and analysis by the HPLC–ESIMS method with the Pinnacle octyl column and methanol gradient with 2 mM in the mobile phase Powder Lot no. Compound

m/z

4-40 -Dinitrodiphenylamine N-Nitrosodiphenylamine Methyl centralite 4-Nitrodiphenylamine Diphenylamine 2-Nitrodiphenylamine Ehtyl centralite Dibutylphthalate

260 199 241 215 170 215 269 279

WIN 296

WIN 296

H380

H380

H-S 700-X

H-S 700-X

H-S 700-X

H-S 700-X

IMR 4895

IMR 4895

IMR 4895

D61A %

B5A %

D601 %

D901 %

869 %

818 %

850 %

916 %

933 %

929 %

926 %

5

5

5

5

0.02 0.2

2

3

4

0.1

0.2

0.1 0.07 4 0.02

0.06 5 0.02

0.07 4 0.03

0.09 5 0.1 0.08 2

0.09 4 0.08 0.05 3

0.1 4 0.09 3 0.3

from the characterization of the solution parameters associated with the methanol gradient. The ESIMS intensity and thus method sensitivity is dependant on a number of factors regarding instrumental parameters and the mobile phase solution. The instrumental settings are analyte dependent and cannot be as readily modified as the mobile phase composition. Because the instrumental parameters in the present study were held constant in the course of evaluating the different solution parameters, the results allowed for the characterization of the ESIMS response as a function of the methanol gradient. While these results were specific to the particular constituents used in smokeless powders, the experimental approach can be used to evaluate the ESIMS response of other analytes. Additional studies must consider the important instrumental parameters such as ionization potential, interface temperature, and nebulizing gas flow rate. In addition, the solution flow rate, conductivity, and surface tension affect the generation of ions by electrospray at specific instrumental settings. 4.6. Smokeless powders analysis While the fundamental effects of gradient elution on the electrospray ionization mass spectrometric response of smokeless powder additives were investigated, the practical application of this technique provides a unique approach toward characterizing smokeless powder samples. The quantitative analysis of several unburned smokeless powders was performed using the newly developed HPLC–ESIMS method. The percent composition of the additives in twelve unburned smokeless powders is given in Table 1. As described previously, these results allow for the differentiation of various unburned powders from different manufacturers by comparing the composition of each powder [11]. In the present study, unburned powders from the same manufacturer but different lots were also analyzed. The smokeless powders from the different manufactures can be distinguished by the presence or absence of particular compounds. For example, Winchester 296 (WIN 296) was

0.09 4 0.09 2 0.3

1

4

1

2

the only powder of the eleven in which dibutylphthalate was detected. Furthermore, while ethyl centralite was detected in the Hodgden 380 (H380) and Hi-Skor 700-X (H-S 700-X) powders, the two lots of H380 also contain 44DNDPA. The varying composition of certain additives in the same powder and different lots demonstrates the advantage of the HPLC– ESIMS analysis method to differentiate smokeless powders. For example, the different levels of ethyl centralite in the H-S 700-X and N-nitrosodiphenylamine and diphenylamine in IMR 4895 can be used to show the minor compositional variation among the different lots of the same smokeless powder.

5. Conclusions The ionization of neutral analytes in positive ion ESIMS depend on a combination of factors. In the present study, the ESIMS responses of smokeless powder additives have been characterized as a function of solution parameters related to the gradient elution profile. The effects of the ammonium acetate concentration, pH, and percent methanol were investigated by direct liquid infusion of selected smokeless powder components with varying solution parameters. The results demonstrate that the ESIMS intensities of the smokeless powder additives were dependent upon the ammonium acetate concentration with varying and constant levels of methanol, not affected by pH, and increased with the percent methanol with the electrolyte concentration held constant. The HPLC–ESIMS responses of the non-polar components were enhanced by maintaining the ammonium acetate concentration constant throughout the linear gradient.

Acknowledgments The authors acknowledge the Ohio University Research Incentive Fund for financial support and Janet Doyle of the Federal Bureau of Investigation Forensic Science Research and Training Center for assistance with this project.

166

J.A. Mathis, B.R. McCord / Forensic Science International 154 (2005) 159–166

References [1] M.J. Bogusz, J. Chromatogr. B 733 (1999) 65–91. [2] F. Garofolo, A. Longo, V. Migliozzi, C. Tallarico, Rapid Commun. Mass Spectrom. 10 (1996) 1273–1277. [3] F. Saint-Marcoux, G. Lachatre, P. Marquet, J. Am. Soc. Mass Spectrom. 14 (2003) 14–22. [4] K. Srinivasan, P. Wang, A.T. Eley, C.A. White, M.G. Bartlett, J. Chromatogr. B 745 (2000) 287–303. [5] L. Geiser, S. Cherkaoui, J.L. Veuthey, J. Chromatogr. A 895 (2000) 111–121. [6] W. Ahrer, E. Scherwenk, W. Buchberger, J. Chromatogr. A 910 (2001) 69–78. [7] L.M. Wu, J.R. Almirall, K.G. Furton, HRC-J. High Resolut. Chromatogr. 22 (1999) 279–282. [8] Z.P. Wu, Y. Tong, J.Y. Yu, X.R. Zhang, C.D. Yang, C.X. Pan, X.Y. Deng, Y.X. Wen, Y.C. Xu, J. Forensic Sci. 46 (2001) 495– 501. [9] Y. Tong, Z.P. Wei, C.D. Yang, J.Y. Yu, X.R. Zhang, S.J. Yang, X.Y. Deng, Y.C. Xu, Y.X. Wen, Analyst 126 (2001) 480–484. [10] R. Dams, T. Benijts, W. Gunther, W. Lambert, A. De Leenheer, Rapid Commun. Mass Spectrom. 16 (2002) 1072–1077. [11] J.A. Mathis, B.R. McCord, J. Chromatogr. A 988 (2003) 107– 116. [12] S.A. White, A.S. Kidd, K.S. Webb, J. Forensic Sci. 44 (1999) 375–379. [13] A. Asperger, R. Efer, T. Koal, W. Engewald, J. Chromatogr. A 937 (2001) 65–72. [14] J.J. Zhao, A.Y. Yang, J.D. Rogers, J. Mass Spectrom. 37 (2002) 421–433. [15] G.D. Wang, R.B. Cole, J. Am. Soc. Mass Spectrom. 7 (1996) 1050–1058.

[16] W.H. Schaefer, F. Dixon, J. Am. Soc. Mass Spectrom. 7 (1996) 1059–1069. [17] P. Kebarle, M. Peschke, Anal. Chim. Acta 406 (2000) 11–35. [18] T.L. Constantopoulos, G.S. Jackson, C.G. Enke, J. Am. Soc. Mass Spectrom. 10 (1999) 625–634. [19] M.S. Wilm, M. Mann, Int. J. Mass Spectom. Ion Process. 136 (1994) 167–180. [20] P. Kebarle, L. Tang, Anal. Chem. 65 (1993) A972–A986. [21] R.B. Cole, J. Mass Spectrom. 35 (2000) 763–772. [22] R. King, R. Bonfiglio, C. Fernandez-Metzler, C. MillerStein, T. Olah, J. Am. Soc. Mass Spectrom. 11 (2000) 942– 950. [23] C.G. Enke, Anal. Chem. 69 (1997) 4885–4893. [24] N.B. Cech, C.G. Enke, Anal. Chem. 72 (2000) 2717–2723. [25] N.B. Cech, C.G. Enke, Mass Spectrom. Rev. 20 (2001) 362– 387. [26] S.L. Zhou, B.S. Prebyl, K.D. Cook, Anal. Chem. 74 (2002) 4885–4888. [27] S.L. Zhou, K.D. Cook, J. Am. Soc. Mass Spectrom. 11 (2000) 961–966. [28] R.M. Martz, L.D. Lasswell, in: J. Yinon (Ed.), Proceedings of the International Symposium on the Analysis and Detection of Explosives, US Government Printing Office: Washington, DC, FBI Academy, Quantico VA, 1983. pp. 245–254. [29] Z.P. Wu, Y. Tong, J.Y. Yu, X.R. Zhang, C.X. Pan, X.Y. Deng, Y.C. Xu, Y.X. Wen, Analyst 124 (1999) 1563–1567. [30] N.B. Cech, J.R. Krone, C.G. Enke, Anal. Chem. 73 (2001) 208–213. [31] J.F. Gal, P.C. Maria, E.D. Raczynska, J. Mass Spectrom. 36 (2001) 699–716. [32] P.J.R. Sjoberg, C.F. Bokman, D. Bylund, K.E. Markides, J. Am. Soc. Mass Spectrom. 12 (2001) 1002–1010.