Characterization of Triacetone Triperoxide by Ion Mobility

ARTICLE pubs.acs.org/ac

Characterization of Triacetone Triperoxide by Ion Mobility Spectrometry and Mass Spectrometry Following Atmospheric Pressure Chemical Ionization Robert G. Ewing,* Melanie J. Waltman, and David A. Atkinson Pacific Northwest National Laboratory, Richland, Washington 99354, United States ABSTRACT: The atmospheric pressure chemical ionization of triacetone triperoxide (TATP) with subsequent separation and detection by ion mobility spectrometry has been studied. Positive ionization with hydronium reactant ions produced only fragments of the TATP molecule, with m/z 91 ion being the most predominant species. Ionization with ammonium reactant ions produced a molecular adduct at m/z 240. The reduced mobility value of this ion was constant at 1.36 cm2V1s1 across the temperature range from 60 to 140 °C. The stability of this ion was temperature dependent and did not exist at temperatures above 140 °C, where only fragment ions were observed. The introduction of ammonia vapors with TATP resulted in the formation of m/z 58 ion. As the concentration of ammonia increased, this smaller ion appeared to dominate the spectra and the TATPammonium adduct decreased in intensity. The ion at m/z 58 has been noted by several research groups upon using ammonia reagents in chemical ionization, but the identity was unknown. Evidence presented here supports the formation of protonated 2-propanimine. A proposed mechanism involves the addition of ammonia to the TATPammonium adduct followed by an elimination reaction. A similar mechanism involving the chemical ionization of acetone with excess ammonia also showed the formation of m/z 58 ion. TATP vapors from a solid sample were detected with a hand-held ion mobility spectrometer operated at room temperature. The TATPammonium molecular adduct was observed in the presence of ammonia and TATP vapors with this spectrometer.

R

ecent terrorist attempts that have involved the use of triacetone triperoxide (TATP) have elevated the interest in the detection of peroxide-based explosives and other homemade explosives. TATP is relatively unstable and thus does not have industrial or military applications. However, due to relative ease of synthesis and availability of starting materials, TATP has been discovered in many instances of illicit production, from either intended terrorist activity or seemingly simple curiosity. With the heightened awareness of TATP use, interest in improving detection of peroxide-based explosives has grown. A review of some these methods was published in 2006.1 In this review, it was noted that only one paper on TATP detection by ion mobility spectrometry (IMS) had been published. IMS has historically been one of the most important techniques for the detection of explosive compounds, with extensive deployment around the globe for security applications. Most common explosives such as nitro-organic species are ionized and subsequently detected as negative ions.2 TATP, however, appears to ionize by only forming positive ions with atmospheric pressure chemical ionization (APCI). A variety of ions has been reported in the literature resulting from chemical ionization and APCI of TATP with mass spectrometry (MS) analyses. Wilson et al.3 reported TATP fragment ions of m/z 59, 77, 91, and 117, along with the protonated molecular ion m/z 223 in a selected ion flow tube (SIFT)-MS experiment with H3Oþ reagent ions. Using O2þ as a reagent ion, m/z 43, 58, 77, 93, 133, and 222 (TATPþ) were reported. r 2011 American Chemical Society

With NOþ as the reagent ion, m/z 43, 88, 191, and 222 (TATPþ) were observed. These major shifts in fragmentation pattern in this low pressure, higher energy environment (compared to APCI) illustrate the sensitivity of the TATP fragmentation to specific changes in ionization chemistry. Sigman et al.4 analyzed TATP using gas chromatography with positive ion chemical ionization MS. With methane as the reagent gas, m/z 43 ion was the largest peak with other ions observed at m/z 59, 74, 75, and 91. When ammonia was used as the reagent gas, the m/z 240 (TATP 3 NH4þ) ion was observed along with m/z 58 and 223. An interesting highlight to this work was that TATP analysis was performed with both a quadrupole mass spectrometer and an ion trap-based system.4 The resulting differences will be explored later in this manuscript in the context of the experimental results presented here. Using APCI conditions with ammonia, a variety of ions have been reported: m/z 74, 89, 91, 223, and 240.57 From these results, the pseudomolecular ion at m/z 240 is readily formed under a variety of APCI conditions containing ammonia. Upon ms/ms analysis of the m/z 240 ion, Xu et al.6 showed fragments m/z 223, 132, 91, and 74, and Widmer et al.7 showed fragments m/z 91 and 74. Desorption electrospray ionization (DESI) was also shown to produce pseudomolecular ions of TATP with both Received: February 26, 2011 Accepted: April 28, 2011 Published: April 28, 2011 4838

dx.doi.org/10.1021/ac200466v | Anal. Chem. 2011, 83, 4838–4844

Analytical Chemistry ammonium and sodium ions.8 Also, it was noted that lower source temperatures for APCI worked best to produced protonated species and molecular adducts of organic peroxides.9 IMS, as it is typically deployed, uses 63Ni-based APCI to generate product ions. This ionization chemistry closely parallels that of typical APCI-MS and is very sensitive to certain types of compounds. Only two publications have discussed the detection of TATP by IMS, with conflicting results.10,11 Buttigieg et al. reported the best detection of TATP in the presence of toluene as a protonated molecular ion, m/z 223 with a reported Ko of 2.71 cm2V1s1.11 Marr and Groves reported a TATP ammonium adduct in the presence of ammonia vapor with a Ko = 1.36 cm2V1s1.10 Marr and Groves also noted that at temperatures above 130 °C the molecular adduct was no longer visible, but two other ions m/z 73 and 89 appeared. The investigation in this manuscript attempts to further understand the processes that occur for the ionization of TATP in the presence of H3Oþ and NH4þ reactant ions and to determine ion formation and fragmentation relating to temperature and other instrumental conditions.

’ EXPERIMENTAL SECTION Chemicals. TATP was obtained from AccuStandard (New Haven, CT) in hexane at a concentration of 1.0 mg mL1. Solutions of TATP were diluted further 1:10 in methanol and in hexane. Hexane and methanol (99.9% pure) were obtained from Fisher (Pittsburgh, PA). Ammonium hydroxide (2830% NH3) was obtained from Sigma-Aldrich (St. Louis, MO). Purified air was generated from compressed air and purified with a Pure Air Generator model CO2RP140 (Domnick Hunter, Charlotte, NC) with moisture levels less than 10 ppmv measured by a Moisture Image Series I Hygrometer (GE Panametrics, Billerica, MA). Instrumentation. The primary instrument used in this study was an ion mobility spectrometer/mass spectrometer model MMS 160 IMS/MS/MS (PCP Inc., West Palm Beach, FL) with a 63Ni ionization source. This was described in detail previously.12 Ion mobility spectrometer temperature was varied from 60 to 180 °C. Purified air was used for both the drift gas and carrier gas at flow rates of 600 mL min1 and 100 mL min1, respectively. The electric field was maintained at 200 V/cm throughout the experiment. Ion mobility spectrometer spectra were collected using a 200 μs gate width. Reduced mobility values were calculated from mobility spectra obtained from the faraday plate. Temperature was measured as the internal gas temperature from a thermocouple inserted near the ionization region. This temperature was 2 to 7 °C cooler than the housing temperature, with the larger differences occurring at the higher temperatures. A hand-held ion mobility spectrometer, an LCD ABBII (Smiths Detection, Watford, U.K.), was used for detecting TATP vapors from solid TATP. The design and operation has been described previously.13,14 Modifications included removing ammonia vapors from the instrument by replacing the ammonia doped molecular sieve with fresh undoped molecular sieve several times over an extended period to change the reactant ion chemistry from ammonium ions to hydronium ions. Procedures. Four types of spectra were obtained with this system including: normal mass spectra (MS); ion mobility spectra with the Faraday plate at the end of the drift region (IMS-F); ion mobility spectra with the electron multiplier (IMS-EM); and ion mobility spectra with mass spectrometer

ARTICLE

set for selected ion monitoring (IMS-SIM). The mass spectra were collected by operating the ion mobility spectrometer with the ion gates fully open. The IMS-F spectra were used to measure drift times and subsequent calculation of the reduced mobility values (Ko) reported in this paper. Calculated reduced mobility values measured throughout the experiments for a given ion and set of conditions remained constant and varied only slightly, on the order of 0.01 cm2 V1 s1. IMS-EM spectra, which used the electron multiplier from the mass spectrometer as the detector while the quadrupoles were operated to pass all ions, produced spectra similar to the IMS-F spectra with the exception of a shift of approximately 0.5 ms to longer drift times. To identify the drift time of an ion with a specific m/z, an IMS-SIM was collected by setting the quadrupoles to pass a specific m/z while collecting ions over the drift time spectral period. The drift time of an ion could then be compared to the drift times from the IMS-EM and from this correlated to the IMS-F to link m/z to a specific reduced mobility value. Various ammonia vapor concentrations were generated from ammonium hydroxide placed in a permeation tube as well as in two diffusion tubes with different cross sectional area and lengths. The permeation tube consisted of a 3 in. long piece of 1/4 in. Teflon tubing (1/16 in. wall thickness) capped on both ends with 1/4 in. stainless steel Swagelok caps. Two diffusion tubes were assembled from 2 mL amber glass vials (Fisher Scientific) containing ammonium hydroxide and having screw cap lids with Teflon septa, through which syringe needles were placed. The needles for the small and large diffusion tubes were 18 and 23 gauge, respectively (Becton, Dickinson and Company, 1 1/2 in. length). The tube in use was placed in a sealed stainless steel vessel (approximate volume of 33 in.3) with air flow in at 100 mL min1 and the outlet flow split variably between the ion mobility spectrometer and the exhaust. The concentration of ammonia was gravimetrically determined over several weeks. The permeation tube emitted 2.9  107 g min1; the small diffusion tube emitted 4.1  107 g min1, and the large diffusion tube emitted 7.1  107 g min1. For TATP introduction, a small aliquot of TATP solution (e2 μL) was deposited on the end of a 8 in.-long stainless steel wire, and the solvent was allowed to evaporate for 10 to 20 s prior to insertion into the inlet of the instrument. For the introduction of ammonia vapor, additional air containing various concentrations of ammonia was delivered through a Swagelok tee and passed over the sample wire and into the ionization region. The sample flow entering the ionization region was thus a combination of the carrier gas, the ammonia gas stream (when added), and TATP vapor desorbing from the end of the wire.

’ RESULTS AND DISCUSSION The introduction of TATP to the IMS/MS instrument at 60 °C with Hþ(H20)n reactant ions produced only fragment ions and not the protonated molecular ion at m/z 223. The IMS spectra of TATP with Hþ(H2O)n reactant ions at temperatures of 60 °C, 100 °C, 140 °C, and 180 °C are displayed in Figure 1. The fragment ions observed for TATP under these conditions were located close to the reactant ion peak and were mass identified to be m/z 59 and 91, with 91 being the predominant ion at all temperatures. The fragment ions appear to have some temperature dependence with the m/z 59 ion appearing above 120 °C. The fragment ions m/z 91 (C3H6O3Hþ) and m/z 59 4839

dx.doi.org/10.1021/ac200466v |Anal. Chem. 2011, 83, 4838–4844

Analytical Chemistry

ARTICLE

Table 1. Reduced Mobility Values (cm2V1s1) of Various Reactant Ions and TATP Product Ions over a Range of Temperatures

Figure 1. IMS-F spectra of positive ions in air with about 10 ng of TATP from 1 μL of a dilute solution in hexane at various temperatures. Only fragments ions of TATP were observed. The dominant ion at m/z 91 appears at all temperatures investigated with m/z 59 appearing at elevated temperatures. The ion masses indicated above each peak were determined by IMS-SIM.

Figure 2. Ion mobility spectra about 10 ng of TATP at 60 °C. (A) IMS spectra collected with the Faraday plate detector. *The intensity is not directly comparable to the other spectra collected with the electron multiplier. (B) IMS spectra collected with the electron multiplier of the mass spectrometer. (C) IMS spectra with SIM of m/z 59 and (D) IMS spectra with SIM of m/z 91.

(C3H6OHþ) are ones that have been consistently observed with chemical ionization mass spectrometry.3,4,15,16 The peaks in the ion mobility spectra shown in Figure 1 were mass identified by IMS-SIM as shown in Figure 2, which contains various spectra of TATP at 60 °C. Figure 2A is the IMS-F

temp.

NH4þ

Hþ(H2O)n

m/z 58

60 °C

2.28

2.13

80 °C

2.37

2.18

100 °C

2.48

2.23

2.29

120 °C

2.58

2.29

140 °C 160 °C

2.67 2.74

2.36 2.40

180 °C

2.86

m/z 59

m/z 91

m/z 240

2.17

2.00

1.36

2.22

2.04

1.36

2.15

2.08

1.37

2.29

2.24

2.14

1.36

2.32 2.33

2.29 2.31

2.18 2.20

1.36

2.34

2.30

2.20

spectrum which is used to calculate mobility values. Since the spectrum was collected using a different detector than the other three spectra in Figure 2 (Faraday plate vs electron multiplier), the intensity could not be compared and was scaled to approximate the other spectra in this figure. Figure 2B is the IMS spectrum collected with the electron multiplier, which shows the slight shift of about 0.5 ms to longer drift times due to the ions traversing the intermediate pressure region of the IMS/MS instrument interface and the mass spectrometer. Figure 2C,D represents IMS-SIM spectra for ions m/z 59 and m/z 91. The measured reduced mobility values at 60 °C for the TATP product ions m/z 59 and 91 were 2.07 and 2.00 cm2V1s1, respectively, with the reactant ions having a reduced mobility of 2.13 cm2V1s1. The reduced mobility values of these ions at all temperatures are listed in Table 1.The reported reduced mobility values varied by less than 0.5% within a given day. However, a parametric error analysis of the reduced mobility value was performed by Spangler17 and indicated an expected uncertainty of 1.8%; thus, the values were reported to 2 decimal places in Table 1 to reflect this uncertainty and display appropriate significant figures. The slight shift to higher mobility values with increases in temperature for a given ion are due to a shift in the equilibrium of these ions clustered with water molecules toward less solvation and thus smaller collisional cross sections of the ions. This behavior is consistent with ion mobility theory.18 The peak to the far left, labeled as H3Oþ, was previously identified as the hydrated proton reactant ion with observed masses of m/z 37 and 55, indicating 2 and 3 water moieties as observed by the mass spectrometer. These ions exist in equilibrium within the drift region of the ion mobility spectrometer and are not observable as discrete peaks in the mobility spectra. Each individual spectrum was a separate experiment resulting from discrete sample introductions of TATP. The variations of the relative intensities of the reactant ions and the product ions are due mainly to the variations in sample introduction. Increased sample concentration will result in an increase in the product ions and a subsequent decrease in the reactant ion intensity. Although attempts were made to introduce the same quantity of TATP, the relatively high vapor pressure, which has a range of reported values of 43, 59, 87, and 245 ppmv,1922 made this a challenge. Typically 12 μL aliquots of the TATP solution were added to the end of a wire, and the solvent was allowed to evaporate prior to insertion of the wire into the heated inlet of the ion mobility spectrometer. Solvents of methanol and hexane were selected because of their high vapor pressure and relatively low proton affinity, which was similar to or lower than that of water. The goal was to remove the solvent and let the TATP residue remain so as not to complicate the ionization process. Initially, 4840

dx.doi.org/10.1021/ac200466v |Anal. Chem. 2011, 83, 4838–4844

Analytical Chemistry

Figure 3. IMS-F spectra of 10 ng of TATP with ammonium reactant ions at a variety of temperatures.

analysis was performed by visually observing the disappearance of the droplet prior to analysis. However, it was discovered that the TATP signal varied significantly in a matter of seconds. To understand the significance of the evaporation of TATP, aliquots of 0.2 and 0.5 μL of the stock TATP solution were placed on the sample wire, thus depositing 200 and 500 ng. These were allowed to evaporate for 10 to 40 s before the wire was inserted into the ion mobility spectrometer. The intensity of the predominant ion was plotted versus evaporation time. The signal intensity was observed to be reduced by about 75% within about 10 s. Higher quantities shifted the time that the sample could be detected, but the signal decayed rapidly once the concentration fell within the range between saturated signal and the detection limit. As a result of the rapid sample evaporation, TATP analysis was conducted to keep the time between sample deposition and insertion constant for a set of experiments. This investigation was aimed at identifying TATP product ions formed in the chemical ionization process and not necessarily to probe detection limits. The amounts used in this study represent calculated amounts initially deposited onto the wire. Actual amounts delivered to the ion mobility spectrometer were significantly lower. In an attempt to produce larger ions that are better separated from the reactant ion species, ammonia vapor from the permeation tube was added to the ionization region to modify the reactant ion species from Hþ(H2O)n to NH4þ(H2O)m in order to produce the ammoniaTATP adduct observed by others.47,10 Such ions will have significantly lower mobilities and will appear in the IMS spectra well separated from the reactant ion species. This should greatly increase the confidence in identification of TATP by IMS. Ion mobility spectra with ammonium reactant ions at various temperatures are displayed in Figure 3. At temperatures below 120 °C, the main ion peak observed for TATP was TATP 3 NH4þ at m/z 240. As the temperature was increased, fragment ions appeared at m/z 91 and m/z 58. The correlation between mass and mobility of these peaks was verified by IMS-SIM. The reduced mobilities of these ions at various temperatures are listed in Table 1.

ARTICLE

Over the temperature range of 60 to 120 °C, the Ko for the TATPammonia adduct measured here remained constant at 1.36 cm2V1s1. This was in excellent agreement with Marr and Groves10 who also measured a Ko of 1.36 at temperatures e130 °C for this ion. This is in contrast to data presented by Buttigieg et al.,11 who reported the presence of a protonated molecular ion of TATP at m/z 223 with a Ko of 2.71 cm2V1s1 over a temperature range from 30 to 120 °C. In the experiments described in this manuscript, protonated molecular ions were never observed, nor were any fragments with m/z > 100. Although IMS does not provide specific m/z information like that of a mass spectrometer, strong correlations between mass and mobility have been reported.23,24 The mobility value reported by Buttigieg et al.11 of 2.71 cm2V1s1 correlates to ions with masses close to those for reactant ion species such as H3Oþ and NH4þ. An interesting observation was that the reported Ko of 2.71 cm2V1s1 for TATP in toluene was higher than their reported Kos for nitrogen, peroxide, and acetone, which are in the range expected for these smaller ions. In Buttigieg et al.’s results, the mass spectrometer that was used to identify the protonated molecular ion at m/z 223 was a different instrument than the ion mobility spectrometer used to produce the Ko value. It is possible that protonation of TATP could have occurred near the orifice to the mass spectrometer, thus minimizing the time available for the ion to fragment into smaller species. It is also possible that the ion observed in the ion mobility spectrometer may have been a fragment of the molecular species. Many mass spectrometry techniques, including SIFT-MS,3 PTR-MS,15 and CI-MS,4 have produced the protonated species of TATP, but it has always been a minor ion compared to other product ions resulting from either fragmentation or adduct formation with ammonia. However, the protonated species has not been observed elsewhere with IMS, and attempts to reproduce it in this work were not successful. With only protonated water as a reactant ion, the predominant fragment ion of m/z 91 and a smaller ion of m/z 59 was observed in the experiments described in this manuscript. These ions have been observed by others,3,4,15 and m/z 91 appears to be common throughout. When ammonium is used as the chemical reagent ion, the ammoniumTATP adduct is observed at temperatures below 140 °C in the IMS/MS system. Calculated structures by Sigman et al.4 show that the proton interaction with the TATP molecule lies on one of the oxygen atoms in the peroxide bond, forcing the carbonoxygen bond to lengthen, indicating a weaker bond. When the ammonium ion was used, it was shown coordinated near the center of TATP, allowing the charge to be shared across multiple peroxide bonds. This could explain why the TATPammonia adduct remains intact while the protonated molecular ion fragments during the milliseconds-long separation time available in IMS. Shen et al. also noted the instability of the protonated molecular ion of TATP and were only able to observe the ion using a reduced-field proton transfer reaction mass spectrometer (PTR-MS).15 In the presence of the ammoniumTATP adduct, a smaller ion with m/z 58 has also been observed in IMS/MS investigations.10 Marr and Groves noted the presence of the peak m/z 58 and attributed it to the acetone molecular ion.10 The reported mobility of this peak was 2.32 cm2V1s1, which is close to the mobility values measured in this study for m/z 58 at about 140 °C. Marr and Groves recognized the relationship between the presence of ammonia and the appearance of the 58 peak but did not propose a mechanism, only a tentative identity as a molecular ion of acetone. The authors noted that the 58 peak was dominant 4841

dx.doi.org/10.1021/ac200466v |Anal. Chem. 2011, 83, 4838–4844

Analytical Chemistry when they had ammonia present, but when the ammonia concentration was “minimized”, the ion at m/z 240 became dominant.10 The authors did not mention the actual ammonia concentration levels present in their experiments. As part of this investigation, the ammonia concentrations were controlled to understand the effects on the ionization of TATP. Ammonia concentrations were estimated on the basis of gas flow rates and diffusion rates determined gravimetrically. Displayed in Figure 4 are two mobility spectra of TATP collected at 80 °C under the same conditions with the exception of varying the ammonia concentration. Figure 4A,B shows ion mobility spectra of TATP with ammonia concentrations in the ionization region of about 4.7 ppmv and 8.1 ppmv, respectively. The dramatic shift in the relative intensities of the 240 peak to the 58 occurs with a modest (less than double) increase to the ammonia vapor concentration. As previously noted, others have observed the appearance of the 58 peak in ammonia chemical ionization studies.4,15 Shen et al. noted the appearance of the m/z 58 ion and refuted the idea that it was acetone since the introduction of acetone to their PTR-MS with ammonium reagent ions produced m/z 76, which is the ammoniumacetone adduct. Furthermore, they did not observe m/z 58.15 Clearly, ammonia concentration plays a significant role in the chemical ionization product ions from TATP. First, in the presence of small levels of ammonia, the reactant ion species changes from H30þ to NH4þ and, upon the introduction of TATP, produces the TATPammonium adduct at m/z 240. The subsequent addition

Figure 4. IMS of a 0.2 μL stock TATP at 80° (A) with 4.7 ppmv ammonia (B) with 8.1 ppmv ammonia.

ARTICLE

of higher concentrations of ammonia then leads to the formation of m/z 58. With protonated organic species that have even number masses, it is likely that the ion contains an odd number of nitrogen atoms. With the relatively low mass of 58, the ion likely contains only one nitrogen atom. This is a reasonable assumption since the production is related to the presence of high concentrations of ammonia. Furthermore, since the initial ion formed is the TATPammonium adduct and the addition of more ammonia causes the production of the ion at m/z 58, a proposed mechanism likely involves the interactions of the adduct ion with an ammonia molecule. A proposed mechanism is shown in Figure 5, which leads to the production of protonated 2-propanimine. The proton affinity of 2-propanimine is fairly high with a value of 932.3 kJ mol1, which is significantly higher than ammonia with a value of 853.6 kJ mol1 and many other species, thus explaining predominance in the spectra.25 The premise for this mechanism was based upon common reactions found in organic chemistry texts for the addition of derivatives of ammonia to aldehydes and ketones. The addition of ammonia to the carbonyl group followed by the elimination of water produces an imine. In the case for TATP, the carbon is bonded to two separate oxygen atoms instead of the carbonoxygen double bound found in a carbonyl group and is facilitated by the protonated ammonia already attached to TATP. To further test the proposed mechanism, it was postulated that a similar mechanism should also work for an aldehyde or ketone. To test this theory, the gas phase chemical ionization of acetone in the presence of ammonia was investigated. First, 10 μL from the headspace of acetone was injected into the ion mobility spectrometer with ammonia concentrations of about 8.1 ppmv as shown in Figure 6A. The three peaks in this figure are ammonium ion, the proton bound dimer of ammonia, and the ammoniumacetone adduct at m/z 76. Only a small peak at m/z 58 was observed. To increase the ammonia concentration, the experiment was repeated by adding 15 μL of headspace from a concentrated solution of ammonium hydroxide to the 10 μL headspace of acetone vapor, both in a 25 μL gastight syringe. This mixture was then injected into the ion mobility spectrometer, producing Figure 6B. The ion at m/z 58 clearly dominates the spectra and is evidence of the production of protonated 2-propanimine. Although this was not tested further, it is likely to occur for other aldehydes and ketones. In reviewing the data by Sigman et al. for the ammonium chemical ionization of TATP, it was noted that the spectra from both a quadrupole and an ion trap mass spectrometer were presented.4 The ammonium adduct was observed in both instances, but the ion at m/z 58 was only observed in the ion trap system. The authors noted the presence of m/z 58 but did not postulate on its identity or how it was formed. It was likely formed in the ion trap via the same mechanism proposed here due to the increased time available for the adduct to interact with

Figure 5. Mechanism of formation of m/z 58 from m/z 240 in the presence of excess ammonia. 4842

dx.doi.org/10.1021/ac200466v |Anal. Chem. 2011, 83, 4838–4844

Analytical Chemistry

Figure 6. Mass spectra of 10 μL of headspace of acetone at 80° with (A) large ammonia diffusion tube at 50 mL min1 (B) and additional 15 μL of headspace of ammonia vapor plus large ammonia diffusion tube at 50 mL min1.

neutral ammonia molecules within the ion trap (compared to the quadrupole), forming the protonated 2-propanimine. Shen et al. also noticed the appearance of the m/z 58 ion in the presence of ammonia in their PTR-MS system, which also allows additional reaction time in the drift tube.15 The relatively high vapor pressure of TATP makes it a good candidate for direct vapor detection. From the above discoveries, the ideal conditions for vapor detection would be an ion mobility spectrometer operated below 100 °C with appropriate ammonia concentrations in the reaction region. To test this, a hand-held ion mobility spectrometer, LCD ABBII, was used to sample the headspace vapors from TATP. Although this instrument was designed specifically for detecting chemical agents and toxic industrial chemicals, it is also capable of detecting a wide variety of organic vapors that are amenable to atmospheric pressure chemical ionization. The instrument has been shown to detect ethylene glycol dinitrate (EGDN) and 2,3-dimethyl-2,3-dinitrobutane (DMNB) vapors emitted from explosives.14 The LCD ABBII was operated without the ammonia dopant, using the hydronium ions as the main reactant ion species, as shown in the background spectra in Figure 7A. The reduced mobility value of the background peak was 2.06 cm2V1s1. Sampling the headspace from an ammonia permeation tube produced the spectra shown in Figure 7B with two peaks appearing to the left of the reactant ion with mobility values of 2.39 and 2.24 cm2V1s1. These peaks are likely protonated ammonia and the proton bound dimer of ammonia. When TATP headspace vapors were sampled in the absence of ammonia, a peak was visible just to the right of the reactant ion peak with a reduced mobility of 1.96 cm2V1s1 as shown in Figure 7C. When both TATP headspace vapors and ammonia vapors were sampled simultaneously as shown in Figure 7D, the TATPammonia adduct ion was observed at 8.47 ms with a reduced mobility value of 1.32 cm2V1s1. The mobility values determined here were calculated internally by the instrument firmware. The values for the ammonia, water, and TATP

ARTICLE

Figure 7. Ion mobility spectra of positive ions from a hand-held IMS, LCD-ABBII (Smiths Detection, Watford, UK), (A) Background spectra, H3Oþ(H2O)n; (B) ammonia vapor; (C) headspace from TATP crystals; (D) both ammonia and headspace from TATP crystals.

fragment ions were slightly lower than those measured with the ion mobility spectrometer/mass spectrometer, but this could be due to a shift in the equilibrium position of water clustering, since this instrument was operated at an ambient temperature of 21.7 °C as measured internally in the LCD ABBII. The lowest temperature run on the IMS/MS instrument was 60 °C. Although not directly comparable, these Ko values follow the trend of decreasing mobility as temperature decreases, as shown in Table 1. The reduced mobility value of 1.36 for the TATPammonia adduct measured with the ion mobility spectrometer/mass spectrometer appeared not to vary with temperature and compares well with the value of 1.32 cm2V1s1 determined with the LCD ABBII.

’ CONCLUSIONS Ion identities and mobility values for IMS analysis of TATP have not been well characterized previously. Using standard protonated water reactant ions, TATP fragments are observed at m/z 91 and m/z 59 and are postulated to be C3H6O3Hþ and C3H6OHþ, respectively. The mobility values of these peaks put them in a location near the reactant ions where the attribution of the peaks, specifically to TATP in an environmental sample, can be problematic. With the addition of an ammomia dopant, the TATPammonia adduct m/z 240 is present for temperatures below 140 °C. This adduct has a mobility value of 1.36 cm2V1s1, which is much lower than those resulting from the fragments and facilitates a higher confidence of TATP identification during an IMS measurement. At IMS temperatures below 120 °C, the ammoniaTATP adduct is the optimal product ion for detection of TATP by IMS, as it is well separated from the reactant ions, exhibits high sensitivity, and is reproducible if ammonia dopant concentrations are reasonably controlled. Excess ammonia alters the m/z 240 peak to m/z 58, believed to be C3NH8þ. The peak at m/z 58 has been reported previously but was left unidentified or attributed to an acetone molecular species. This work proposes the identity of 4843

dx.doi.org/10.1021/ac200466v |Anal. Chem. 2011, 83, 4838–4844

Analytical Chemistry the species as protonated 2-propanimine and postulates the mechanism of formation, which results from an ammonia driven reaction.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

ARTICLE

(21) Oxley, J. C.; Smith, J. L.; Luo, W.; Brady, J. Propellants, Explos., Pyrotech. 2009, 34, 539–543. (22) Oxley, J. C.; Smith, J. L.; Brady, J.; Naik, S. Propellants, Explos., Pyrotech. 2010, 35, 278–283. (23) Berant, Z.; Karpas, Z. J. Am. Chem. Soc. 1989, 111, 3819–3824. (24) Karasek, F. W.; Kim, S. H.; Rokushika, S. Anal. Chem. 1978, 50, 2013–2016. (25) Hunter, E. P.; Lias, S. G. J. Phys. Chem. Ref. Data 1998, 27 (3), 413–656.

’ ACKNOWLEDGMENT The authors would like to acknowledge the Laboratory Directed Research and Development program for funding this research through the Initiative for Explosives Detection at Pacific Northwest National Laboratory. The Pacific Northwest National Laboratory is a multiprogram national laboratory operated by Battelle Memorial Institute for the U.S. Department of Energy under Contract DE-AC05-76RL01830. The authors would also like to thank the U.S. Department of Homeland Security, Science and Technology, Transportation Security Laboratory, specifically Inho Cho, Joseph Kozole, and Richard Lareau, for assistance with TATP samples for vapor testing. ’ REFERENCES (1) Schulte-Ladbeck, R.; Vogel, M.; Karst, U. Anal. Bioanal. Chem. 2006, 386, 559–565. (2) Ewing, R. G.; Atkinson, D. A.; Eiceman, G. A.; Ewing, G. J. Talanta 2001, 54, 515–529. (3) Wilson, P. F.; Prince, B. J.; McEwan, M. J. Anal. Chem. 2006, 78, 575–579. (4) Sigman, M. E.; Clark, C. D.; Fidler, R.; Geiger, C. L.; Clausen, C. A. Rapid Commun. Mass Spectrom. 2006, 20, 2851–2857. (5) Cody, R. B.; Laramee, J. A.; Nilles, J. M.; Durst, H. D. JOEL News 2005, 40, 8–12. (6) Xu, X.; van de Craats, A. M.; Kok, E. M.; de Bruyn, P. C. A. M. J. Forensic Sci. 2004, 49, 1–7. (7) Widmer, L.; Watson, S.; Schlatter, K.; Crowson, A. Analyst 2002, 127, 1627–1632. (8) Cotte-Rodriquez, I.; Chen, H.; Cooks, R. G. Chem. Commun. 2006, 953–955. (9) Rondeau, D.; Vogel, R.; Tabet, J.-C. J. Mass Spectrom. 2003, 38, 931–940. (10) Marr, A. J.; Groves, D. M. Int. J. Ion Mobility Spectrom. 2003, 6, 59–62. (11) Buttigieg, G. A.; Knight, A. K.; Denson, S.; Pommier, C.; Denton, M. B. Forensic Sci. Int. 2003, 135, 53–59. (12) Ewing, R. G.; Eiceman, G. A.; Harden, C. S.; Stone, J. A. Int. J. Mass Spectrom. 2006, 255256, 76–85. (13) Taylor, S. J.; Piper, L. J.; Connor, J. A.; FitzGerald, J.; Adams, J. H.; Harden, C. S.; Shoff, D. B.; Davis, D. M.; Ewing, R. G. Int. J. Ion Mobility Spectrom. 1998, 1, 58–63. (14) Ewing, R. G.; Miller, C. J. Field Anal. Chem. Technol. 2001, 5, 215–221. (15) Shen, C.; Li, J.; Han, H.; Wang, H.; Jiang, H.; Chu, Y. Int. J. Mass Spectrom. 2009, 285, 100–103. (16) Zitrin, S.; Kraus, S.; Glattstein, B. Identification of Two Rare Explosives, Proceedings of the International Symposium on the Analysis and Detection of Explosives, FBI Academy, Quantico, VA, March 2931, 1983. (17) Spangler, G. E. Anal. Chem. 1993, 65, 3010–3014. (18) Mason, E. A.; Schamp, H. W., Jr. Ann. Phys. 1958, 4, 233–270. (19) Oxley, J. C.; Smith, J. L.; Shind, K.; Moran, J. Propellants, Explos., Pyrotech. 2005, 30, 127–130. (20) Damour, P. L.; Freddman, A.; Wormhoudt, J. Propellants, Explos., Pyrotech. 2010, 35, 514–520. 4844

dx.doi.org/10.1021/ac200466v |Anal. Chem. 2011, 83, 4838–4844