Introduction

Triacetone triperoxide (TATP) can be easily synthesized, so it is often used in terrorist activities such as the July 2005 London transport bombings, the November 2015 Paris bombing, and the March 2016 Brussels airport and subway station bombings. Therefore, it is necessary to develop a real-time and online technology for detecting TATP.

In recent years, a series of methods for detecting TATP have been developed, such as gas chromatography mass spectrometry (GC-MS) [1], liquid chromatography mass spectrometry (LC-MS) [2], and liquid chromatography method with post-column ultraviolet irradiation and fluorescence detection [3]. Unfortunately, these methods need sample pretreatment and chromatographic separation, which usually takes dozens of minutes, so they are only suitable for laboratory analysis of TATP. Ion mobility spectrometry (IMS) [4,5,6] can be used for on-site detection, but it may misreport because of limited resolution. The electrochemical sensor method [7] needs sonic processing of tens of minutes for the electrodes in order to remove contaminants from electrode surfaces. The N,N-dimethyl-p-phenylenediamine (DMPD) sensor method [8] needs sample acid treatment. None of these methods can be used for reliable online detection of TATP.

Proton transfer reaction mass spectrometry (PTR-MS) is an online detection technology with fast response time and a good limit of detection (LOD) [9,10,11]. Detailed descriptions of the PTR-MS technology can be found elsewhere [12,13,14,15]. In recent years, PTR-MS has been widely applied in the field of public safety, such as the detection of explosives [16,17,18,19] and illicit and controlled prescription drugs [20]. Our research team had pioneered the study of detecting TATP using PTR-MS. The results showed that when the discharge gas in the ion source was changed from water vapor to ammonia under a low reduced-field condition, the sensitivity for detecting TATP could be improved effectively [21]. However, for a common user of PTR-MS, it is inconvenient to switch the discharge gas between water vapor and ammonia. Furthermore, the voltage conditions of ion source need to be changed accordingly. Nazarov and Satoshi Inomata respectively introduced methods of using dopants in IMS and PTR-MS systems to get some new reagent ions [22, 23], inspiring the idea that a dopant with ammonia in PTR-MS may be a good choice for TATP detection based on our previous study [21]. This work developed ammonia-assisted PTR-MS. Through an injection port bypass, ammonia was injected into a homemade PTR-MS device instead of changing the discharge gas. The drift tube voltage and ammonia flow were optimized respectively. The characteristic mass spectra for TATP detection were gained under optimized conditions. Further detection research was carried out on samples of TATP mixed with white sugar or salt.

Experimental

Experimental Device

The TATP detection was performed on a homemade PTR-MS device (Figure 1). Briefly, the PTR-MS instrument consists of an ion source, drift tube, and ion detection system. The reagent ions H3O+(H2O)n (n = 0, 1, 2) were generated in the ion source by glow discharge on the water vapor and then were injected into the drift tube. The flow of water vapor was maintained at 120 mL h−1 during the experiment. Driven by air through the injection port, the sample gas TATP vapor, including the assisted gas ammonia, was introduced into the drift tube. At this time, the ammonia underwent a proton transfer reaction with H3O+(H2O)n (n = 0, 1, 2) (Eq. (1)) [24] because the proton affinity (PA) of ammonia (PA = 850.6 kJ/Mol) is higher than that of H2O (PA = 691.0 kJ/mol) [25] and its cluster ions [26]. In addition, the TATP underwent a reaction with NH4+ (Eq. (2)).

Figure 1
figure 1

Schematic diagram of PTR-MS apparatus. GD glow discharge, SD source drift region, IC intermediate chamber, EM electron multiplier

Full size image
$$ {\mathrm{NH}}_3+{\mathrm{H}}_3{\mathrm{O}}^{+}{\left({\mathrm{H}}_2\mathrm{O}\right)}_n\to {{\mathrm{NH}}_4}^{+}+{\left({\mathrm{H}}_2\mathrm{O}\right)}_{n+1}\ \left(n=0,1,2\right) $$
(1)
$$ \mathrm{TATP}+{{\mathrm{NH}}_4}^{+}\to {\left[\mathrm{TATP}+{\mathrm{NH}}_4\right]}^{+} $$
(2)

After these series of reactions in the drift tube, the product ions first passed through a differentially pumping intermediate chamber, in which they were focused by ion lens before entering the ion detection system. The ion detection system, which is mainly made up of a quadrupole mass spectrometer, detected the m/z and number of ions and finally obtained the mass spectrum. During this experiment, the data sampling rate was 1 Hz and averaged 120 times for one mass spectrum. The temperature of the drift tube was maintained at 60(± 1) °C and the pressure of the drift tube was 0.7 Torr.

Experimental Process

In the drift tube, the product ions generated from the reaction of reagent ions H3O+(H2O)n (n = 0, 1, 2) or NH4+ with the samples will be affected by the ionic collision energy. When the collision energy is too high, many fragment ion peaks will be formed. When the collision energy is too low, many cluster ion peaks will be formed. Both of these conditions make the analysis of mass spectra complicated. The ionic collision energy in the drift tube is closely related to the voltage of the drift tube. To test the proper drift tube voltage condition for TATP detection, the voltage was optimized. The assisted gas ammonia was used to change the reagent ion, so the flow of ammonia was also optimized. The ammonia came from the saturated steam of ammonia water, and an injection pump controlled its flow.

Under the optimized conditions, the characteristic mass spectra of TATP detection were obtained under the condition of conventional PTR-MS and ammonia-assisted PTR-MS. The volume specific concentration C of TATP was calculated by Eq. (3). The equation was derived from the ideal gas equation.

$$ C=\frac{P_S}{P}\frac{F_S}{F_T} $$
(3)

In this equation, Ps (5.25 × 10−2Torr) is the saturated vapor pressure of TATP at room temperature [27], and P (760 Torr) is the environmental pressure of the laboratory. Fs is the saturated vapor flow of TATP, which was controlled by an injection pump, and FT (244.2 mL h−1) is the total flow through the injection port, which was determined by the instrument.

To test the applicability of ammonia-assisted PTR-MS detecting TATP in real complex matrixes, the method was applied to detect TATP mixed with sugar or salt under the optimized conditions.

TATP Synthesis and Reagents

The small amount of TATP used in the experiment was synthesized according to a method described in the previous literature [28]. The purity of the TATP sample was analyzed using gas chromatography mass spectrometry (GC-MS) (Figure S1). The results showed that there was only a little residual acetone in the TATP sample. This is because, for safety reasons, the synthetic TATP sample was not dried. The ammonia solution (25~28% NH3) was purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China, http://www.sinoreagent.com). The sugar and salt used in the experiment were purchased from a nearby grocery store.

Results and Discussion

Optimization of the Drift Tube Voltage and Ammonia Flow

When the injection port bypass ammonia flow was closed in conventional PTR-MS, 7.58 ppm TATP steam was injected into the injection port. The ion [TATP+H]+ at m/z 223 was chosen as the characteristic ion. The drift tube voltage was regulated from 50 to 240 V with 10 V increments. The intensity variation of [TATP+H]+ with the drift tube voltage is shown in Figure 2a. With the increase of the drift tube voltage, the [TATP+H]+ ionic signal intensity first increased and then decreased, reaching its maximum at 140 V. The normal drift tube voltage was 358 V for this homemade PTR-MS apparatus. This phenomenon indicates that the drift tube voltage needed for TATP detection is lower than that of usual detection. In other words, the [TATP+H]+ ions fragment easily.

Figure 2
figure 2

Dependence of the characteristic ionic intensity on the drift tube voltage using (a) conventional PTR-MS for detecting TATP and (b) ammonia-assisted PTR-MS for detecting TATP

Full size image

When the injection port bypass ammonia flow was 1 mL h−1 in ammonia-assisted PTR-MS, 7.58 ppm TATP steam was injected into the injection port. The ion [TATP+NH4]+ at m/z 240 was chosen as the characteristic ion. The drift tube voltage was regulated from 50 to 330 V with 10 V increments. The intensity variation of [TATP+NH4]+ with the drift tube voltage is shown in Figure 2b. With the increase of drift tube voltage, the [TATP+NH4]+ ionic signal intensity first increased and then decreased, reaching its maximum at 230 V. Compared with Figure 2a, the optimized drift tube voltage is higher, which shows that [TATP+NH4]+ is more stable than [TATP+H]+. Also, the [TATP+NH4]+ ionic signal intensity was higher than that of [TATP+H]+ for the same concentration of TATP steam. This phenomenon indicates that ammonia-assisted PTR-MS is more sensitive to TATP than conventional PTR-MS.

When the drift tube voltage was set to 230 V in ammonia-assisted PTR-MS, 7.58 ppm TATP steam was injected into the injection port. The ammonia flow optimization was carried out by changing the ammonia flow. The result is shown in Figure 3. One can see that the intensity of the characteristic ions [TATP+NH4]+ gradually leveled off after the initial ramp-up, and the optimum ammonia flow is 9 mL h−1. The reason for this trend should be as follows. With the increase of ammonia flow, more reagent ions H3O+(H2O)n (n = 0, 1, 2) should be converted to NH4+, so the ion [TATP+NH4]+ intensity will increase. When the ammonia flow is large enough, H3O+(H2O)n (n = 0, 1, 2) should be completely converted to NH4+. At this time, the increase of ammonia flow will not increase the intensity of NH4+, so the intensity of [TATP+NH4]+ will tend to level off.

Figure 3
figure 3

Dependence of the [TATP+NH4]+ intensity on the ammonia flow F

Full size image

To test the above speculation, the mass spectra for the laboratory air were measured at 230 V drift tube voltage; the results are shown in Figure 4. Figure 4a is the mass spectrum for the conventional PTR-MS. The peaks of the reagent ions appear at m/z 19, 37, and 55, and they can be attributed to H3O+(H2O)n (n = 0, 1, 2). Figure 4b is the mass spectrum for ammonia-assisted PTR-MS with an ammonia flow of 9 mL h−1. The peaks of the reagent ions appear at m/z 18, 35, and 36 and can be attributed respectively to NH4+, NH4+NH3, and NH4+H2O. One can see that almost all the H3O+ and its cluster ions can be transformed to NH4+ and its cluster ions. And this phenomenon confirms the above speculation.

Figure 4
figure 4

Mass spectra of laboratory air measured by (a) conventional PTR-MS and (b) ammonia-assisted PTR-MS

Full size image

Contrast of Sensitivity and LOD of TATP Detection Using Conventional PTR-MS and Ammonia-Assisted PTR-MS

With the drift tube voltage set to 140 V in conventional PTR-MS, the concentration of TATP steam was prepared as 7.58, 17.5, 26.2, 40.8, and 53.6 ppm. The dependence of the intensity of [TATP+H]+ at m/z 223 and the concentration of TATP steam are shown in Figure 5a. When the drift tube voltage was set to 230 V and the ammonia flow was 9 mL h−1 in ammonia-assisted PTR-MS, the concentration of TATP steam was the same as above. The dependence of the intensity of [TATP+NH4]+ at m/z 240 and the concentration of TATP steam is shown in Figure 5a. The ratio of characteristic ion [TATP+NH4]+ intensity to [TATP+H]+ intensity is shown by numbers on arrows in Figure 5a. The background signal intensity at m/z 223 and 240 was less than 5 cps (counts per second), which is far less than the ion intensity of TATP seen in Figure 5a. Therefore, the ratio of characteristic ion intensity is approximately equal to the ratio of sensitivity. From Figure 5a, one can see that the sensitivity of TATP detection is enhanced more than 37 times by ammonia-assisted PTR-MS.

Figure 5
figure 5

(a) Dependence of characteristic ions intensity on TATP concentration using conventional PTR-MS and ammonia-assisted PTR-MS; monitor graphs of characteristic ions using (b) conventional PTR-MS and (c) ammonia-assisted PTR-MS

Full size image

To investigate the LOD of TATP detection, 7.58 ppm TATP steam was detected using conventional PTR-MS and ammonia-assisted PTR-MS. The monitor graphs of characteristic ions are shown in Figure 5b and c. After a signal-to-noise ratio of 2:1 is considered, the LOD can be calculated respectively using the results in Figure 5b, c [29]. The LOD of TATP detection was 33.6 ppb for conventional PTR-MS and was 1.3 ppb for ammonia-assisted PTR-MS. The LOD of TATP for ammonia-assisted PTR-MS is better even than that of ammonia discharge PTR-MS due to the higher reagent ions in this PTR-MS instrument [21].

Analysis of Typical Mass Spectra for Detecting TATP

When the drift tube voltage was set to 140 V in conventional PTR-MS, the typical mass spectrum was measured for the air inside the laboratory; the result is shown in Figure 6a. To reduce the damage of the electronic multiplier, the water cluster ions at m/z 55 and 73 were skipped in this scan. In Figure 6a, some obvious peaks at m/z 36, 37, 57, 75, and 91 can be seen. These peaks can be attributed respectively to NH4+H2O, H3O+H2O, and O18 isotopic peaks of H3O+(H2O)2, and to O18 isotopic peaks of H3O+(H2O)3 and H3O+(H2O)4. The reagent ions are all in the form of water clusters because of the low drift tube voltage.

Figure 6
figure 6

Conventional PTR-MS for detecting (a) air and (b) TATP; ammonia-assisted PTR-MS for detecting (c) air and (d) TATP

Full size image

When the drift tube voltage was set to 140 V in conventional PTR-MS and the concentration of TATP steam was 40.8 ppm, the mass spectrum of TATP was measured; Figure 6b shows the results. In contrast to Figure 6a, the new rising ionic peaks appeared at m/z 43, 59, 74, 75, 77, 79, 91, 92, 93, 97, 109, 117, and 223. The ion at m/z 223 is the characteristic ion [TATP+H]+. Methane positive ion chemical ionization mass spectrometry (PICI-MS) [30] and PTR-MS [21] used for detecting TATP all detected the ions at m/z 43, 59, 74, 75, and 91; these are fragment ions from [TATP+H]+ and have been assigned as CH3CO+, C3H7O+, C3H6O2+, C3H7O2+, and C3H7O3+ respectively [30]. The ions at m/z 75 and 91 also have a minor contribution from water clusters. To investigate the influence of residual acetone, acetone was added to the injection port at this voltage. The rising ion peaks were found at m/z 59, 77, and 117, which should be protonated acetone, a clustering ion of m/z 59 with a water molecule, and dimer of acetone (Figure S2a). This means that the ions at m/z 59 have partial contributions from residual acetone and the ions at m/z 77 and 117 are also the result of residual acetone. The ions at m/z 79 and 97 may be the clustering ions of m/z 43 with two water molecules C2H3O+(H2O)2 and of m/z 43 with three water molecules C2H3O+(H2O)3. As for the ions at m/z 92, 93, and 109, they should be the clustering ions of m/z 74, 75, and 91 with a water molecular and are assigned as C3H6O2+ H2O, C3H7O2+ H2O, and C3H7O3+ H2O.

When the drift tube voltage was set to 230 V and ammonia flow was 9 mL h−1 in ammonia-assisted PTR-MS, the typical mass spectrum measured for the air inside the laboratory is as shown in Figure 6c. The reagent ion is mainly NH4+ at m/z 18 (see Figure 4b), which was skipped in this scan because of its too strong intensity. Furthermore, some other obvious peaks at m/z 30, 35, and 36 can be seen; they may be NO+, NH4+NH3, and NH4+H2O.

When the drift tube voltage was set to 230 V and ammonia flow was 9 mL h−1 in ammonia-assisted PTR-MS, 40.8 ppm TATP steam was injected into the injection port. The measured mass spectrum is shown in Figure 6d. In contrast to Figure 6c, the rising ionic peaks appear at m/z 43, 58, 73, 74, 75, 76, 89, 91, and 240. The ion at m/z 240 is the characteristic ion [TATP+NH4]+. The experimental results of TATP detection by PTR-MS [21] showed that when ammonia was used as the discharge gas in the ion source, the reagent ions changed from H3O+ to NH4+ and NH4+NH3, and the signal peak of characteristic ion [TATP+NH4]+ could be obtained. Meanwhile, there were some fragment ion peaks in the mass spectrum, such as ions at m/z 43, 58, 74, 75, and 91. As for ions at m/z 58, Ewing et al. suggested that this peak was due to protonated 2-propanimine C3NH8+ generated by the interaction between [TATP+NH4]+ and NH3 [5]. In addition to the ions reported in the literature [21], new ion peaks appeared at m/z 73, 76, and 89 in this paper. To investigate the influence of residual acetone, acetone was added to the injection port under the same conditions of drift tube voltage and ammonia flow, and arising ion peak was found at m/z 76, which should be adduct ions [C3H6O+NH4]+ (Figure S2b). Therefore, the ion peak at m/z 76 is the result of residual acetone. As for ion peaks at m/z 73 and 89, Marr et al. [31] used IMS to detect TATP; they found that when the temperature of the migration tube was higher than 130 °C, the [TATP+NH4]+ peak disappeared and new ion peaks appeared at m/z 73 and 89. Therefore, the ions at m/z 73 and 89 may be thermal decomposition fragment ions of [TATP+NH4]+.

Analysis of Mass Spectra of TATP Mixed with Real Complex Matrixes

TATP is like white sugar or salt in appearance. To test the applicability of ammonia-assisted PTR-MS for detecting TATP in real complex matrixes, 45 mg solid TATP, 45 mg solid TATP mixed with 450 mg white sugar, and 45 mg solid TATP mixed with 450 mg salt were respectively placed in the injection port. When the drift tube voltage was set to 230 V and ammonia flow was 9 mL h−1 in ammonia-assisted PTR-MS, the measured mass spectra are as shown in Figure 7. One can see that the peak positions and relative abundances of the product ions are basically the same as the result in Figure 6d except for ions at m/z 76. The ion peak at m/z 76 is the result of residual acetone (Figure S2b). A possible reason is as follows. The TATP measured in Figure 6d was saturated vapor, which was obtained from hours of sublimation of solid TATP in a sealed glass bottle. In this situation, acetone steam originating from residual acetone will accumulate in the TATP saturated vapor, which increased the concentration of acetone in that vapor. The TATP measured in Figure 7 was solid TATP, so there was no cumulative process, which reduced the ion intensity at m/z 76. Figure 7 shows that even if TATP is mixed with sugar or salt at a mass ratio of 1:10, the characteristic ion [TATP+NH4]+ can still be detected. This result shows that ammonia-assisted PTR-MS can be used to detect TATP in real complex matrixes.

Figure 7
figure 7

Mass spectra of (a) only TATP, (b) mixture of TATP and white sugar, and (c) mixture of TATP and salt, all detected using ammonia-assisted PTR-MS

Full size image

Conclusion

A novel method, ammonia-assisted PTR-MS, was developed to detect TATP explosive through changing the reagent ions. The drift tube voltage and the ammonia water headspace steam flow were optimized. The optimal drift tube voltage was 230 V, and the optimal ammonia water headspace steam flow was 9 mL h−1 for ammonia-assisted PTR-MS. The sensitivity was enhanced more than 37 times compared with TATP detection using conventional PTR-MS, and the LOD reached 1.3 ppb. Ammonia-assisted PTR-MS can also be applied for detecting samples of TATP mixed with sugar or salt at a mass ratio of 1:10. Compared to the conventional PTR-MS, ammonia-assisted PTR-MS has better sensitivity and LOD for TATP detection. Ammonia-assisted PTR-MS provides common users of PTR-MS with a convenient and quick method to change reagent ions. In the same vein, the users of PTR-MS could easily obtain other reagent ions by injecting different assisted gases into the injection port to meet different detection needs.