Detection of Explosives and Related Compounds by Low

ARTICLE pubs.acs.org/ac

Detection of Explosives and Related Compounds by Low-Temperature Plasma Ambient Ionization Mass Spectrometry Juan F. Garcia-Reyes,†,‡ Jason D. Harper,†,§ Gary A. Salazar,† Nicholas A. Charipar,†,§ Zheng Ouyang,†,§ and R. Graham Cooks*,† †

Department of Chemistry and Center for Analytical Instrumentation Development and §Weldon School of Biomedical Engineering, Purdue University, West Lafayette, Indiana 47907, United States ‡ Analytical Chemistry Research Group, Department of Physical and Analytical Chemistry, University of Jaen, 23071 Jaen, Spain

bS Supporting Information ABSTRACT: Detection of explosives is important for public safety. A recently developed low-temperature plasma (LTP) probe for desorption and ionization of samples in the ambient environment (Anal. Chem. 2008, 80, 9097) is applied in a comprehensive evaluation of analytical performance for rapid detection of 13 explosives and explosives-related compounds. The selected chemicals [pentaerythritol tetranitrate (PETN), trinitrotoluene (TNT), cyclo-1,3,5-trimethylenetrinitramine (RDX), tetryl, cyclo-1,3,5,7-tetramethylenetetranitrate (HMX), hexamethylene triperoxide diamine (HMTD), 2,4-dinitrotoluene, 1, 3-dinitrobenzene, 1,3,5-trinitrobenzene, 2-amino-4,6-dinitrotoluene, 4-amino-2,6-dinitrotoluene, 2,6-dinitrotoluene, and 4-nitrotoluene) were tested at levels in the range 1 pg-10 ng. Most showed remarkable sensitivity in the negative-ion mode, yielding limits of detection in the low picogram range, particularly when analyzed from a glass substrate heated to 120 °C. Ions typically formed from these molecules (M) by LTP include [M þ NO2]-, [M]-, and [M - NO2]-. The LTP-mass spectrometry methodology displayed a linear signal response over three orders of magnitude of analyte amount for the studied explosives. In addition, the effects of synthetic matrices and different types of surfaces were evaluated. The data obtained demonstrate that LTP-MS allows detection of ultratrace amounts of explosives and confirmation of their identity. Tandem mass spectrometry (MS/MS) was used to confirm the presence of selected explosives at low levels; for example, TNT was confirmed at absolute levels as low as 0.6 pg. Linearity and intraand interday precision were also evaluated, yielding results that demonstrate the potential usefulness and ruggedness of LTP-MS for the detection of explosives of different classes. The use of ion/molecule reactions to form adducts with particular explosives such as RDX and HMX was shown to enhance the selectivity and specificity. This was accomplished by merging the discharge gas with an appropriate reagent headspace vapor (e.g., from a 0.2% trifluoracetic acid solution).

ocietal and scientific1 interest in the detection of explosives and explosives-related compounds (ERCs) continues to increase. High sensitivity, the specificity needed to minimize false positives or false negatives, applicability to nonvolatile and thermally unstable analytes, real-time response, and instrumentation portability and reliability are the main characteristics required in explosives detection technology. No existing detection technology fulfils all these requirements. Explosives of the nitro class degrade over time, for example, when buried in landmines, to yield a variety of chemicals including nitrobenzene (NB), dinitrobenzene (DNB), and nitrotoluene (NT).2 The development of a multiclass method capable of detecting accurately multiple explosives and ERCs is of great interest both for public safety reasons and for detecting landmines via the higher vapor pressure degradation products. A wide variety of techniques has been used to detect bulk and trace amounts of explosives.3 Trace detection methods include cavity ring-down spectroscopy (CRDS),4,5 Raman spectroscopy,6 laser-induced breakdown spectroscopy7,8 together with other

S

r 2010 American Chemical Society

laser-based detection techniques,9,10 immunoassays,11,12 chemiluminescence,13 and ion mobility spectrometry (IMS).14 Among these technologies, IMS is probably the most widely used.15 Detection of explosives by IMS is based on the gas-phase mobilities of their ions in a weak electric field. IMS is a sensitive, rugged, trace detection technology used extensively for rapid nitro-containing explosives detection, particularly as a screening tool at airports. They feature extraordinarily high sensitivity but less impressive chemical specificity. The technology is broadly but not universally applicable to the detection of all explosives classes. The standard sampling method for trace explosives detection with IMS is based on manual surface wiping with sampling swabs. These two factors have resulted in continued interest in evaluating alternatives for the detection of trace explosives in transportation security, among which mass spectrometry is one possibility. Received: November 5, 2010 Accepted: December 7, 2010 Published: December 21, 2010 1084

dx.doi.org/10.1021/ac1029117 | Anal. Chem. 2011, 83, 1084–1092

Analytical Chemistry

ARTICLE

Figure 1. Schematic showing configuration of the LTP probe for ambient ionization MS.

Mass spectrometry- (MS-) based methods have several advantageous features for the rapid and specific trace detection of particular organic compounds, including high specificity, selectivity, and broad applicability. However, until the development of the ambient ionization methods,16-21 MS methods required significant sample manipulation. Another barrier to the use of mass spectrometry is that some of the explosives are nonvolatile compounds that are not easily ionized by traditional methods. Although a wide variety of desorption ionization methods are available for the MS analysis of compounds on surfaces, they require operation under vacuum. Included in this group are various laser-based ionization techniques combined with mass spectrometry, among them resonance-enhanced multiphoton ionization (REMPI) time-of-flight mass spectrometry (TOFMS),22-24 single-photon laser ionization time-of-flight mass spectrometry (SPI-TOFMS),25 and ultrafast multiphoton ionization (MPI) TOFMS.26 Electrospray ionization mass spectrometry in the negative-ion mode,27-29 atmospheric-pressure chemical ionization,30 and liquid chromatography-mass spectrometry31,32 have been widely used for the detection of explosives in solution. Gas chromatographybased methods in combination with electron capture or mass spectrometric detection33,34 have been applied for the vaporphase analysis of explosives.35 Despite the availability of all these reliable methods, the need remains for rapid measurements on solid residues on surfaces with little or no sample preparation. In this context, ambient ionization mass spectrometry16-21 is a novel approach the features of which map well onto the requirements for trace explosive detection, mainly because these techniques remove or minimize the need for sample preparation and allow fast in situ desorption/ionization of sample surfaces. The original two ambient desorption ionization methods, desorption electrospray ionization (DESI)36-39 and direct analysis in real time (DART),40 have both been applied to trace detection of explosives. Secondary electrospray ionization mass spectrometry41 and thermal desorption/ambient chemical ionization42 have also been employed for direct sampling of explosives in the gas phase or sampled/accumulated onto cotton swabs, respectively. In recent work, advanced laser methods of ambient ionization17,43-45 have been used, as has another plasma-based ambient desorption ionization source, low-temperature plasma (LTP).46 The LTP source uses a dielectric-barrier discharge to generate a plasma (using helium or ambient air as discharge gas) that is capable of desorbing and ionizing species in the condensed

phase. Zhang et al.47 reported an initial study characterizing explosives using the LTP probe in a study of TNT, RDX, and PETN. Here we report a feasibility study on the application of LTPMS for the trace detection of multiclass explosives. The LTP probe was coupled to an ion trap mass spectrometer and its sensitivity and specificity were assayed for several different multiclass explosives and explosives-related compounds including pentaerythritol tetranitrate (PETN), trinitrotoluene (TNT), cyclo1,3,5-trimethylenetrinitramine (RDX), tetryl, cyclo-1,3,5,7-tetramethylenetetranitrate (HMX), hexamethylene triperoxide diamine (HMTD) 2,4-dinitrotoluene, 1,3-dinitrobenzene, 1,3,5-trinitrobenzene, 2-amino-4,6-dinitrotoluene, 4-amino-2,6-dinitrotoluene, 2,6-dinitrotoluene, and 4-nitrotoluene. The effects of synthetic matrices and the use of different types of surfaces on the analytical performance of the method were also evaluated. Experiments aimed at enhancing selectivity and specificity were conducted by adding to the plasma gas chemical reagents that produce ion-molecule reactions with selected explosives such as RDX and HMX.

’ EXPERIMENTAL SECTION Chemicals and Reagents. Trifluoroacetic acid (TFA) was purchased from Sigma-Aldrich (Milkwaukee, WI) and methanol and acetonitrile (HPLC-grade) were from Mallinckrodt Baker Inc. (Phillipsburg, NJ). Deionized/distilled water was obtained from a Barnstead/Thermolyne deionizer unit (Barnstead Mega-Pure System, Dubuque, IA). All standards were used without further purification. Commercially available mixtures of explosives standards solutions EPA 8330A and EPA8330B were acquired from Supelco Inc. (Bellefonte, PA). HMTD, HMX, PETN, RDX, and TNT standards were purchased as 0.1 or 1 mg/ mL solutions in methanol/acetonitrile (1:1) from AccuStandard Inc. (New Haven, CT). Individual standards and/or standards mixtures were prepared by diluting the stock solutions with methanol to a final concentration of 1-1000 μg/L of the selected explosives. Low-Temperature Plasma Mass Spectrometry. Experiments were performed on a Thermo LTQ linear ion trap mass spectrometer (Thermo Finnigan, San Jos
e, CA). Data were acquired via the Xcalibur software. LTP-MS analysis was performed in the positive and negative-ion mode for all compounds studied. The instrument was set to collect spectra under automatic gain 1085

dx.doi.org/10.1021/ac1029117 |Anal. Chem. 2011, 83, 1084–1092

Analytical Chemistry

ARTICLE

Table 1. LTP-MS/MS Analysis of Selected Explosivesa

1086

dx.doi.org/10.1021/ac1029117 |Anal. Chem. 2011, 83, 1084–1092

Analytical Chemistry

ARTICLE

Table 1. Continued

a Compounds studied include those from the EPA 8330 method. Experiments were done with glass as substrate. b Molecular mass (Mr) was calculated from isotope-averaged atomic masses for the constituent elements. c LOD was calculated from solvent standards with glass as substrate. All experiments were performed with 3 μL of solvent standards deposited on glass slides, at room temperature. d LOD was calculated from solvent standards with glass as substrate, heated at 120 °C. All experiments were performed with 3 μL of solvent standards deposited on glass slides.

control for a maximum ion trap injection time of 200 ms with 2 microscans per spectrum. An average spectral acquisition time of 3 s was used for each mass spectrum. The main experimental parameters used were as follows: m/z range 150-600; capillary temperature 200 °C; tube lens -65 V; capillary voltage -15 V. Tandem mass spectrometry experiments were performed with collision-induced dissociation (CID) experiments in order to confirm the presence of the targeted species. These experiments were performed with an isolation window of 1.5 (m/z units) and 25-35% collision energy (manufacturer's unit). The LTP probe (Figure 1) is described in detail elsewhere.46 It consists of a glass tube (o.d. 6.35 mm and i.d. 3.75 mm) with an internal grounded electrode (stainless steel, diameter 1.57 mm) centered axially and an outer electrode (copper tape) surrounding the outside of the tube. The wall of the glass tube serves as the dielectric barrier. An alternating high voltage, 5-10 kV at a frequency of ca. 2.5 kHz, is applied to the outer electrode with the center electrode grounded to generate the dielectric barrier discharge. The discharge AC voltage was provided by a custombuilt, variable-frequency, variable-voltage power supply with total power consumption below 3 W. Helium was used as discharge gas at a flow rate of 0.4 L/min. The samples were placed on the sample holder, typically 0.5 cm away from the LTQ inlet. The LTP probe was placed with its end 4 mm away from the surface with an angle of ca. 30° from the sample surface. The samples were placed on the sample holder, typically 0.5-1 cm away from the LTQ inlet, and the LTP probe was placed with its end 4 mm away from the surface and at an angle of ca. 30° to the sample surface. For those experiments that employed a heated substrate, heating was achieved by directing a heat gun (NTE Electronics, Bloomfield, NJ) under the sample holder to increase the temperature of the substrate (glass slide) to ∼120 °C. Direct LTP-MS Analysis of Explosives Standards. Aliquots (2-3 μL of each solution) were deposited with a micropipet on a microscope glass slide (beveled micro slides, size 75  25 mm, thickness 1 mm, Gold Seal , Becton and Dickinson Co., Franklin Lakes, NJ).

’ RESULTS AND DISCUSSION Low-Temperature Plasma Mass Spectral Features of Selected Explosives and Explosives-Related Compounds. Ex-

periments were performed in both positive- and negative-ion

modes with pure samples in order to characterize LTP-MS(MS) performance. In most cases, high-quality spectra were obtained only in the negative-ion mode, as shown in Table 1. Note the difference in LOD values for heated and unheated substrates. The characteristic ions are molecular radical ions and nitro adducts. As an example, the full-scan LTP-MS spectrum of a mixture of explosives is shown in Figure 2a. The abundance of nitro-group adducts can be associated with the high concentration of these species in the plasma, as revealed by the background ions observed (Figure S1, Supporting Information). The fullscan negative-ion mode LTP-MS of the studied compounds provided limited fragmentation information for confirmation. Therefore, CID MS/MS experiments were performed to provide a higher degree of confidence as well as to decrease (improve) detection limits. As shown in Figure 2, in LTP-MS, TNT undergoes electron capture ionization yielding the negative ion m/z 227 as the main ion, although the deprotonated molecule, m/z 226, [M - H]-, is also observed (Figure 2a). Collision-induced dissociation of the 2,4,6-TNT radical anion, m/z 227, produced major fragments seen in Figure 2b at m/z 210 (loss of OH due to an ortho effect) and m/z 197 (loss of NO). These fragmentations are consistent with both APCI and ESI mass spectra data available in the literature.48,49 The production of 2,4,6-trinitrotoluene (TNT) involves toluene formation from methanol and benzene as reagents, followed by three steps of nitration. Typical byproducts include the various isomers of trinitrotoluene, dinitrotoluene, trinitrobenzene, and dinitrobenzene, in proportions that depend on details of the manufacturing process.49 The LTP-MS mass spectra of this family of compounds are characterized by the presence of molecular anions in the negative-ion mode, while MS/MS fragmentation typically involves the loss of NO. The features of the LTP-MS (and MS/MS) spectra of these compounds are included in Table 1. As an example, the CID MS/MS spectra of TNT, 1,3-dinitrobenzene, 1,3,5-trinitrobenzene, and 2-amino4,6-dinitrobenzene are illustrated in Figure 2b-e. Interestingly, a curious example of isomer differentiation occurs in the case of 2,4- and 2,6-dinitrotoluene (Figure 3). Under several experimental conditions, the 2,4- and 2,6-dinitrotoluene isomers gave different mass spectra. 2,6-DNT is detected as a molecular ion at m/z 182 but 2,4-DNT is detected at m/z 1087

dx.doi.org/10.1021/ac1029117 |Anal. Chem. 2011, 83, 1084–1092

Analytical Chemistry

ARTICLE

Figure 2. (a) Mixture of explosives (30 ng each) examined from a heated glass slide. Full-scan negative-ion mode LTP-MS spectrum of 2,4-dinitrotoluene (2,4-DNT) (m/z 181 and 183); 1,3-dinitrobenzene (1,3-DNB) (m/z 168); 1,3,5-trinitrobenzene (1,3,5-TNB) (m/z 213 and 259); 2-amino-4,6-dinitrotoluene (2A-4,6-DNT) (m/z 197 and 243); RDX (m/z 268 and 284); and TNT (m/z 226 and 227). See Table 1 for assignments of the ions. Data were obtained via LTP-MS with sample substrate (glass slide) heated at 120 °C, as described in the Experimental Section. (b-e) LTPMS/MS product ion negative-ion mode spectra of the M-. radical anions of selected explosives: (b) m/z 227 (TNT); (c) m/z 168 (1,3-dinitrobenzene); (d) m/z 213 (1,3,5-trinitrobenzene); and (e) m/z 197 (2-amino-4,6-dinitrobenzene). Data were acquired with the sample substrate (glass slide) at room temperature.

181 and 183 (see Figure 3), with the relative abundances of these two ions depending on the geometry and positioning of the inner electrode with respect to the outer electrode. The MS/MS data for 2,6-DNT (Figure 3a) show loss of NO to give an ion at m/z 152 [2,6-DNT - NO]-, a result that is consistent with those for the other nitro compounds included in the study. The MS/MS spectrum (Figure 3b) for m/z 181 from 2,4-DNT yields a product ion at m/z 116. This is consistent with a previous work by Lubman and co-workers,50 who used laser-based ionization with

laser-REMPI under atmospheric conditions. These authors also observed large differences in the mass spectra of the two dinitrotoluene isomers. However, to our knowledge this characteristic fragmentation behavior of each isomer has not been noted in either electrospray or atmospheric-pressure chemical ionization mass spectrometry. Tetryl is known to undergo hydrolysis, forming N-methylpicramide (replacement of a nitro group with a hydrogen atom).27,51 The LTP-MS spectrum of tetryl (Figure S2, Supporting 1088

dx.doi.org/10.1021/ac1029117 |Anal. Chem. 2011, 83, 1084–1092

Analytical Chemistry

ARTICLE

Figure 3. LTP-MS and MS/MS negative-ion mode spectra of 30 ng of (a) 2,6-DNT [(a.1) full scan; (a.2) MS/MS of m/z 182] and (b) 2,4-DNT [(b.1) full scan; (b.2) MS/MS of m/z 181]. Data were acquired with the sample substrate (glass slide) at room temperature.

Information) in the negative-ion mode shows m/z 241, which corresponds to [M - NO2]- but could be due to hydrolysis. Tetryl standards were then prepared while the use of water was avoided, in order to avoid hydrolysis. The resulting LTP-MS spectrum showed additional adduct ions at m/z 257 and 286, corresponding to an oxidized form of tetryl and to the intact [M - H]-. The ion [M - NO2]-, m/z 241, was examined by collision-induced dissociation and its MS/MS product ion spectrum is shown in Figure S3 (Supporting Information). The main fragment ions detected from the cleavage of the precursor ion ([M - NO2]- (m/z 241) are m/z 181, 213, and 194 (in order of relative abundance), which correspond tentatively to the loss of CH2NO2, NCH2, and HNO2, respectively. These fragments are the same to those described by Yinon et al.,27 who used electrospray mass spectrometry. Many organic high explosives, particularly those found in plastic explosives, do not produce a sufficient vapor pressure to allow headspace vapor detection. This is the case for both RDX and HMX. The powerful high explosive RDX (Table S1, Supporting Information) has a very low volatility and is difficult to analyze by traditional ionization methods. The LTP-MS spectra of both RDX and HMX are shown in Figure S3 (Supporting Information). The MS/MS CID spectra of the stable nitro

adducts yield LOD data that are not as good as those for other explosives. The identification of the target explosives in this case was accomplished by use of two characteristic ions for each compound: m/z 268 ([M þ NO2]-) and m/z 286 ([M þ NO3]-) for RDX, and m/z 342 ([M þ NO2]-) and m/z 358 ([M þ NO3]-) for HMX. Additional experiments with these two explosives were performed via a modified ionization assembly with vapor-phase reagents passing through the LTP probe, as discussed below. As a representative of the nitrate ester explosives class, PETN (pentaerythritol tetranitrate) was studied by LTP-MS. Good signal was detected in the negative- but not in the positive-ion mode. The LTP-MS mass spectrum is characterized by typical adducts m/z 362 ([M þ NO2]-) and m/z 378 ([M þ NO3]-). The LOD of a PETN standard in the MS/MS experiment was 60 pg. This presages the potential application of LTP-MS/MS to related compounds such as glycerol nitrate (nitroglycerin) and ethylene glycol dinitrate, which are more volatile than PETN. Because of safety considerations, the peroxide explosives were not tested extensively, although their relatively high volatility and excellent DESI performance suggest that LTP should be successful. The one member of this group that was examined, HMTD (hexamethylene triperoxide diamine), could be detected only in 1089

dx.doi.org/10.1021/ac1029117 |Anal. Chem. 2011, 83, 1084–1092

Analytical Chemistry

ARTICLE

Table 2. Analytical Performance of Different Substrates in LTP-MS Analysis of Selected Explosives

Table 3. Detection of Multiclass Explosives in Synthetic Matrices a

limits of detection (LOD)a compd

glass

paper

PTFE

latex

limits of detection (LOD)b fabric compd

surface concn

surface concn

(pg 3 mm2) solventc

(pg 3 mm2) matrixd

RDXb

30 pg

1 ng

750 pg

5 ng

30 ng

tetryl

15 pg

400 pg

20 pg

150 pg

600 ng

TNT

0.1

0.3

TNT

0.6 pg

40 pg

15 pg

60 pg

300 pg

RDX tetryl

1.5 0.07

15 3.5

a

Data shown were obtained by LTP-MS with sample substrate (glass slide) heated at 120 °C, as described in the Experimental Section. b RDX was detected in full-scan negative ionization mode, and two ions were used for identification purposes: m/z 268 ([M þ NO2]-) and m/z 284 ([M þ NO3]-). Tetryl and TNT were detected via the MS/MS transitions shown in Table 1.

the positive-ion mode ([M þ H]þ) and gave a LOD of 90 pg when a heated substrate was used. Analytical Performance of LTP-MS/MS in the Trace Detection of Explosives. Selected explosives were tested as neat samples at levels in the range from 1 pg to 10 ng. The detection limit was calculated as the minimum amount of analyte deposited on a glass slide that gave a signal (for the product ion in the selected MS/MS transition used for detection of the target analyte) that could be distinguished from background. The criterion S/N ratio g3, for averaged signals (for both analyte and background) of 3 s, was used to define detection limits. Most of the compounds showed remarkable sensitivity in the negativeion mode, yielding LODs in the low picogram range, particularly when a heated glass substrate at 120 °C was used (Table 1). The data confirm that LTP-MS allows the detection and confirmation of ultratrace amounts of explosives. MS/MS confirmation of selected explosives such as TNT and 2,6-dinitrotoluene were obtained at absolute levels as low as 0.6 pg. The use of heated substrate improved the detection of less volatile compounds (e.g., RDX and HMX). We noticed that the compounds with higher vapor pressures were less affected by the use of heated substrate, and low picogram detection was still achieved with the substrate at room temperature for compounds such as trinitrobenzene, dinitrotoluene, or even TNT. This fact, together with the small impact on analyte signal associated with the angle and positioning of the probe in relation to the sample, suggests that thermal desorption is a necessary step in ionization by LTP. The lack of a strong geometry dependence suggests the ionization occurs only with gas-phase sample molecules, unlike, for instance, DESI, where charged droplets directed to the surface in a particular direction give secondary droplets that leave the surface in a particular direction. It seems likely (but is still speculative) that in LTP the plasma generates “hot spots” on the surface causing (thermal) desorption of analytes and allowing them to undergo atmospheric-pressure chemical ionization reactions with ions formed from excited species generated in the plasma. Linearity and intra- and interday precision were also evaluated, yielding satisfactory results that prove the ruggedness of LTP-MS for the detection of explosives of various types. Interday and intraday RSD values were calculated for a typical case, that of 2, 6-DNT. Intraday precision was satisfactory. The relative standard deviation (RSD) for 5-10 s averaged signals was ca. 5-10%, while the RSD values for the absolute highest signals were lower than 15-20%. Note that part of this variation is due to manual sample spotting/pipetting. Interday RSD was evaluated over a

1,3-dinitrobenzene

0.1

0.6

1,3,5-trinitrobenzene

0.15

0.6

2-amino-4,6-dinitrotoluene

0.1

4

a

The synthetic matrix was prepared with 1 mL of soap (Windex cleaner) in 5 mL of EtOH. An aliquot of 10 μL was spread onto the area on which the analyte had been deposited. b Data shown were obtained by LTP-MS with sample substrate (glass slide) heated at 120 °C, as described in the Experimental Section. c The surface concentration of the experiment in pure solvent was estimated by assuming a circular droplet of 3 mm diameter (area of ca. 7 mm2). d The surface concentration of the experiment with synthetic matrix was estimated; area of spot ca. 300 mm2.

period of 10 days, measuring 10 ng of 2,6-DNT, after tuning of the signal with a known compound (20 pg of herbicide ametryn) to set reproducible instrument conditions. The results were also satisfactory with interday RSD percentage lower than 30%. These results show that the proposed LTP-MS is reliable enough to provide (semi)quantitative data without the need for internal (deuterated) standards. Linearity of response was evaluated with different compounds, including TNT and its metabolite 2,6-DNT. Figure S4 (Supporting Information) shows the calibration curve obtained for 2,6-DNT spotted manually on a glass substrate and examined without heating [calibration curve equation y = 1115.9x þ 3394.7, R2 = 0.9932, where y is the intensity of m/z 182 and x is the amount (nanograms) of 2,6-DNT]. The linearity of the analytical response for LTP-MS approaches three orders of magnitude of concentration (above this range, the ion trapping capacity of the ion trap mass analyzer is the limiting factor). Figure S4 (Supporting Information) also shows the calibration plot for TNT [calibration curve equation y = 33 928x þ 21 117, R2 = 0.9842, where y is the intensity of m/z 227 and x is the amount (nanograms) of TNT]. In each determination, 3 μL of TNT was spotted manually on a glass substrate (a heated substrate in this case). Ruggedness of Low-Temperature Plasma Mass Spectrometry for Trace Detection of Explosives under Simulated Conditions . Effect of Different Surfaces on the Detection of Explosives by Low-Temperature Plasma Mass Spectrometry. Different surfaces were tested: glass, paper, poly(tetrafluoroethylene) (PTFE), latex, and fabric. The results in terms of limits of detection are shown in Table 2. From the data, it is clear that the performance on common substrates is not as good as with glass slides, but subnanogram or nanogram detection still was obtainable depending upon the explosive. Matrix Effects on LTP Detection. The effect of synthetic matrices on the performance of LTP-MS was also evaluated. A common commercial cleaner (Windex) was diluted 5 times in ethanol and used as the matrix. The experiment was performed on glass slides, where 3 μL of a neat solution of the target explosive was distributed evenly over an area of 3 cm2. Thereafter, 10 μL of the synthetic matrix was deposited onto the same surface. Once the surface was completely dry, it was interrogated 1090

dx.doi.org/10.1021/ac1029117 |Anal. Chem. 2011, 83, 1084–1092

Analytical Chemistry

ARTICLE

Figure 4. Reactive LTP-MS/MS using headspace of 1% TFA gas as reagent to create ions that react with target explosives (RDX and HMX). (a) Headspace gas is merged with the helium carrier gas (0.4 L 3 min-1 flow rate). (b) LTP-MS with 0.2% TFA headspace gas reagent. (c) Reactive LTPMS of 200 ng 3 mL-1 RDX (600 pg absolute); m/z 335 corresponds to the [TFA-RDX]- adduct The signal is less intense than that at m/z 268 and 284 due to the NO2 and NO3 adducts. (d) Reactive LTP-MS of 20 ng 3 mL-1 RDX (60 pg). Data were obtained via LTP-MS with sample substrate (glass slide) heated to 120 °C, as described in the Experimental Section.

by LTP-MS. Table 3 shows the effect of the addition of a synthetic matrix on the response to six selected explosives and related compounds (TNT, RDX, tetryl, 1,3-dinitrobenzene, 1,3,5-trinitrobenzene, and 2-amino-4,6-dinitrotoluene). The LODs obtained are shown in Table 3, expressed as analyte amount per surface area (picograms per square millimeter). The detection limits increase in the presence of matrix. However, the differences are not very large and therefore we can conclude that the ionization method is relatively tolerant to interfering species that may be present when real samples are interrogated. Enhancement of Selectivity by Use of Reactive LTP Mass Spectrometry. The use of additional chemical reagents in order to increase selectivity in the detection of selected explosives was studied. In the case of RDX and HMX, only poor-quality MS/MS spectra could be obtained due to the stability of the dominant [M þ NO2]- ion in the negative-ion mode. Therefore, only a single MS stage is used for these explosives. A modified version of the LTP experiment employs a reagent that yields a derivative (here a stable adduct) which improves ionization efficiency of these explosives. Trifluoroacetic acid (TFA) is known to form very stable adducts with RDX (m/z 335) and HMX (m/z 409) as described elsewhere.27,52,53 We used this feature in order to provide additional tools to enhance selectivity for explosives detection. The proposed assembly used is described in Figure 4a. The setup requires a small volume of a dilute aqueous solution of TFA (0.2%). The inherent vapor pressure of TFA in the headspace of the vial provides a concentration of TFA in the gas phase that allows the reactions to be carried out. TFA is ionized while

passing through the plasma once it is transported by the venturi effect, yielding m/z 113 together with its dimer (m/z 227) and trimer species (m/z 341), as shown in Figure 4b. An adduct ion with RDX occurs at m/z 335 and it fragments under CID conditions, giving rise to the TFA anion at m/z 113 by loss of a RDX molecule. Figure 4c,d shows the analysis of 600 pg and 60 pg of RDX . The analytical signal obtained is not as great as that for the nitro adduct from RDX or HMX, but it provides a useful alternative method for the unambiguous confirmation of the studied chemicals.

’ CONCLUSION A feasibility study on the potential application of LTP-MS for the trace detection of explosives has been completed. Analysis of the explosives is performed under ambient conditions from any surface very quickly, including the time for confirmatory MS/MS experiments. One of the main advantages of plasma methods is the fact that no solvents are required. LTP allowed the detection of explosives at levels as low as 0.6 pg for TNT and as much as 0.6 ng for HMX. Increased selectivity is obtained by MS/MS or, alternatively, by performing reactive LTP experiments in which reagents are added to the discharge gas. ’ ASSOCIATED CONTENT

bS

Supporting Information. Additional information (one table and four figures) as described in the text. This material is available free of charge via the Internet at http://pubs.acs.org.

1091

dx.doi.org/10.1021/ac1029117 |Anal. Chem. 2011, 83, 1084–1092

Analytical Chemistry

’ AUTHOR INFORMATION Corresponding Author

*Tel: (765) 494-5262. Fax: (765) 494-9421. E-mail: cooks@ purdue.edu.

’ ACKNOWLEDGMENT We acknowledge funding from U.S. Department of Homeland Security (DHS), HSHQDC-09-9-00008 and 2007-ST-069-TSL001. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the U.S. Department of Homeland Security. J.F.G-R. acknowledges a Postdoctoral Fulbright Scholarship from the “Secretaría de Estado de Universidades e Investigaci
on” of the Spanish Ministerio de Educaci
on y Ciencia. ’ REFERENCES (1) Hallowell, S. F. Talanta 2001, 54, 447–458. (2) Jenkins, T. F.; Leggett, D. C.; Miyares, P. H.; Walsh, M. E.; Ranney, T. A.; Cragin, J. H.; George, V. Talanta 2001, 54, 501–513. (3) Pettersson, A.; Wallin, S.; Brandner, B.; Elds€ater, C.; Holmgren, E. . FOI, 2006. (4) Spicer, J. B.; Dagdigian, P. J.; Osiander, R.; Miragliotta, J. A.; Zhang, X.-C.; Kersting, R.; Crosley, D. R.; Hanson, R. K.; Jeffries, J. Proc. SPIE;Int. Soc. Opt. Eng. 2003, 5089, 1088. (5) Steinfeld, J. I.; Field, R. W.; Gardner, M.; Canagarantna, M.; Yang, S.; Gonz
alez-Casielles, A.; Wiltonsky, S.; Bhatia, P.; Gibbs, B.; Wilkie, B.; Loy, S. L.; Kachanov, A. Proc. SPIE;Int. Soc. Opt. Eng. 1999, 3853, 28. (6) Gupta, M.; Dahmani, R. Spectrochim. Acta, Part A. 2000, 56, 1453–1456. (7) Gonz
alez, R.; Lucena, P.; Tobaria, L. M.; Laserna, J. J. J. Anal. At. Spectrom. 2009, 24, 1123–1126. (8) L
opez-Moreno, C.; Polanco, S.; Laserna, J. J.; DeLucia, F., Jr.; Miziolek, A. W.; Rose, J.; Walters, R. A.; Whitehouse, A. I. J. Anal. At. Spectrom. 2006, 21, 55–60. € (9) Wallin, S.; Pettersson, A.; Ostmark, H.; Hobro, A. Anal. Bioanal. Chem. 2009, 395, 259–274. (10) Munson, C. A.; Gottfried, J. L.; De Lucia Jr., F. C.; McNesby, K. L.; Miziolek, A. W. . In Counterterrorist Detection Techniques of Explosives; Yinon, Y., Ed.; Elsevier: Amsterdam, 2007; Chapt. 10. (11) Bart, J. C.; Judd, L. L.; Kusterbeck, A. W. Sens. Actuators, B 1997, 28-29, 411–418. (12) Rabbany, S. Y.; Lane, W. J.; Marganski, W. A.; Kusterbeck, A. W.; Ligler, F. S. J. Immunol. Methods 2000, 246, 69–77. (13) Jim
enez, A. M.; Navas, M. J. J. Hazard. Mater. 2004, 106, 1–8. (14) Ewing, R. G.; Atkinson, D. A.; Eiceman, G. A.; Ewing, G. J. Talanta 2001, 54, 515–529. (15) Tom, M.; Hill, H. H., Jr. Anal. Chem. 2004, 76, 2741–2747. (16) Weston, D. J. Analyst 2010, 135, 661–668. (17) Huang, M.-Z.; Yuan, C.-H.; Cheng, S.-C.; Cho, Y.-T.; Shiea, J. Annu. Rev. Anal. Chem. 2010, 3, 43–65. (18) Cooks, R. G.; Ouyang, Z.; Takats, Z.; Wiseman, J. M. Science 2006, 311, 1566–1570. (19) Van Berkel, G. J.; Pasilis, S. P.; Ovchinnikova, O. J. Mass Spectrom. 2008, 43, 1161–1180. (20) Harris, G. A.; Nyadong, L.; Fernandez, F. M. Analyst 2008, 133, 1297–1301. (21) Venter, A.; Nefliu, M.; Cooks, R. G. TrAC, Trends Anal. Chem. 2008, 27, 284–290. (22) Kosmidis, C.; Marshall, A.; Clark, A.; Deas, R. M.; Ledingham, K. W. D.; Singhal, R. P. Rapid Commun. Mass Spectrom. 1994, 8, 521–526.

ARTICLE

(23) Ledingham, K. W. D.; Lilic, H. S.; Kosmidis, C.; Deas, R. M.; Marshall, A.; McCanny, T.; Singhal, R. P.; Langley, A. J.; Shaikh, W. Rapid Commun. Mass Spectrom. 1995, 9, 1522–1527. (24) Kosmidis, C.; Ledingham, K. W. D.; Kilic, H. S.; McCanny, T.; Singhal, R. P.; Langley, A. J.; Shaikh, W. J. J. Phys. Chem. A 1997, 101, 2264–2270. (25) Mullen, C.; Irwin, A.; Pond, B. V.; Huestis, D. L.; Coggiola, M. J.; Oser, H. Anal. Chem. 2006, 78, 3807–3814. (26) Weickhardt, C.; Tonnies, K. Rapid Commun. Mass Spectrom. 2002, 16, 442–446. (27) Yinon, Y.; McClellan, J. E.; Yost, R. A. Rapid Commun. Mass Spectrom. 1997, 11, 1961–1970. (28) Wu, Z. G.; Hendrickson, C. L.; Rodgers, R. P.; Marshall, A. G. Anal. Chem. 2002, 74, 1879–1883. (29) Wu, Z.; Hendrickson, C. L.; Rodgers, R. R.; Marshall, A. G. Anal. Chem. 2005, 77, 2297–2302. (30) Song, Y. H.; Chen, H.; Cooks, R. G. Rapid Commun. Mass Spectrom. 2005, 19, 3493–3499. (31) Wood, M.; Laloup, M.; Samyn, N.; Ramírez-Fern
andez, M. M.; de Bruijn, E. A.; Maes, R. A. A.; De Boeck., G. J. Chromatogr. A 2006, 1130, 3–15. (32) Zhao, X. M.; Yinon, J. J. Chromatogr. A. 2002, 977, 59–68. (33) Halasz, A.; Groom, C.; Zhou, E.; Paquet, L.; Beaulieu, C.; Deschamps, S.; Corriveau, A.; Thiboutot, S.; Ampleman, G.; Dubois, C.; Hawari, J. J. Chromatogr. A 2002, 963, 411–418. (34) Walsh, M. E. Talanta 2001, 54, 427–438. (35) Barshick, S. A.; Griest, W. H. Anal. Chem. 1998, 70, 3015–3020. (36) Takats, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G. Science 2004, 306, 471–473. (37) Takats, Z.; Wiseman, J. M.; Cooks, R. G. J. Mass Spectrom. 2005, 40, 1261–1275. (38) Takats, Z.; Cotte-Rodriguez, I.; Talaty, N.; Chen, H.; Cooks, R. G. Chem Comm. 2005, 1950–1952. (39) Cotte-Rodríguez, I.; Takats, Z.; Talaty, N.; Chen, H.; Cooks, R. G. Anal. Chem. 2005, 77, 6755–6764. (40) Cody, R. B.; Laramee, J. A.; Durst, H. D. Anal. Chem. 2005, 77, 2297–2302. (41) Martínez-Lozano, P.; Rus, J.; Fern
andez de la Mora, G.; Hern
andez, M.; Fern
andez de la Mora, J. J. Am. Soc. Mass Spectrom. 2009, 20, 287–294. (42) Popov, I. A.; Chen, H.; Kharybin, O. N.; Nikolaev, E. N.; Cooks, R. G. Chem. Commun. 2005, 1953–1955. (43) Judge, E. J.; Brady, J. J.; Dalton, D.; Levis, R. J. Anal. Chem. 2010, 82, 3231–3238. (44) Brady, J. J.; Judge, E. J.; Levis, R. J. Rapid Commun. Mass Spectrom. 2009, 23, 3151–3157. (45) Nemes, P.; Barton, A. A.; Li, Y.; Vertes, A. Anal. Chem. 2008, 80, 4575–4582. (46) Harper, J. D.; Charipar, N. A.; Mulligan, C. C.; Zhang, X.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2008, 80, 9097–9104. (47) Zhang, Y.; Ma, X. X.; Zhang, S. C.; Yang, C. D.; Ouyang, Z.; Zhang, X. R. Analyst 2009, 134, 176–181. (48) Yinon, J.; McClellan, J. E.; Yost, R. A. Rapid Commun. Mass Spectrom. 1997, 11, 1961–1970. (49) Zhao, X.; Yinon, J. J. Chromatogr. A 2002, 946, 125–132. (50) Zhu, J.; Lustig, D.; Sofer, I.; Lubman, D. M. Anal. Chem. 1990, 62, 2225–2232. (51) Yinon, Y.; Zitrin, S.; Tamiri, T. Rapid Commun. Mass Spectrom. 1993, 7, 1051. (52) Evans, C. S.; Sleeman, R.; Luke, J.; Keely, B. J. Rapid Commun. Mass Spectrom. 2002, 16, 1883–1891. (53) Mathis, J. A.; McCord, B. R. Rapid Commun. Mass Spectrom. 2005, 19, 99–104.

1092

dx.doi.org/10.1021/ac1029117 |Anal. Chem. 2011, 83, 1084–1092