The Study of Thermal Decomposition of RDX by Corona Discharge ...

0 downloads 0 Views 455KB Size Report
Keywords: corona discharge, negative ions, ion mobility spectrometry, mass ... storage and considered as one of the most brisant military explosives. Therefore ...
The Study of Thermal Decomposition of RDX by Corona DischargeIon Mobility Spectrometry-Mass Spectrometry Zuzana Lichvanováa, Martin Sabo a , and Štefan Matejčík a a

Comenius University in Bratislava, Faculty of Mathematics, Physics and

Informatics, Department of Experimental Physics, Mlynská dolina, F2 842 48, Bratislava, Slovakia, [email protected] In this study we have examined the formation of ions from hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) sample evaporated at 473 K in reactions with reactant ions (RI) produced in negative corona discharge (CD). The ions have been analysed and detected by ion mobility spectrometry coupled to orthogonal acceleration time of flight spectrometer (IMS-oaTOF). The chemical ionisation was carried out by O2-, O2- (H2O), N2O2-, N2O3- and N2O3(H2O) RI formed in CD operated in zero air in the reverse gas flow mode at elevated temperatures 438 K and by Cl-(H2O)n (n=0,1,2) at 332 K. In case of 332 K, the RI were prepared by admixture of carbon tetrachloride (CCl4 – dopant gas) in order to enhance sensitivity of RDX. Keywords: corona discharge, negative ions, ion mobility spectrometry, mass spectrometry, thermal decomposition, RDX

1. Introduction The detection of trace amounts of explosives is an important security issue. There exists a growing demand to deploy detection devices in controlled environments such as the airports, mass transit stations, military checkpoints and densely crowded places. Over the past three decades, progress has been made in research focused on understanding of chemical and physical properties of explosives, particularly well-known hexahydro1,3,5-trinitro-1,3,5-triazine (RDX) (Najarro et al. 2012; Gapeev et al. 2003). RDX is an explosive nitroamine compound which is widely used as an ingredient of solid propellants in military and in industrial applications. It has been also developed as a more powerful explosive than TNT. It is stable in storage and considered as one of the most brisant military explosives. Therefore, many studies have been dedicated to its ignition, decomposition and combustion behaviour (Liebman et al. 1987). Several techniques have been used for a research of RDX explosive focused on its thermal decomposition. RDX has been examined over selected thermal ranges using a pyrolysis technique interfaced to an atmospheric pressure chemical ionization tandem mass spectrometer MS-MS system by Liebman et al.(1987). Another research of RDX has been focused on using pyrolysis-gas chromatography (PGC) and pyrogenic products of this explosive were identified by pyrolysis capillary gas chromatography-mass spectrometry (PCGC-MS) (Huwei and Ruonong 1989). In addition to experimental studies of RDX, theoretical calculations dealing with the mechanism of the gas phase unimolecular decomposition of RDX using the principles of gradient-corrected density functional theory have been developed (Wu and Fried 1997). Among the methods and techniques reported for the detection of explosives so far, an IMS has proved to be one of the best methods for detection trace amounts of explosives. The IMS systems are based on separation of the ions according to their mobility as they drift through gas under the influence of an applied electric field (Eiceman et al. 2003). The conventional ionization source used in IMS devices is a radioactive 63Ni foil which emits high energy electrons (Crawford and Hill 2013). Several alternative ionization sources have been

1

discovered including UV light (Baim et al. 1983), corona discharge (Sabo and Matejčík 2012) and electrospray (Hilton et al. 2010). Explosives are commonly detected in IMS in the negative mode. The explosives or their fragments posses high electron affinity and thus efficiently form negative ions. For observation of thermal decomposition of different types of explosives and to demonstrate a reasonable explosives detection limit, positive corona discharge has been also applied as an ionization source in IMS by Tabrizchi (2010). In this paper, we discuss the reactions of RI produced by negative CD with RDX and the influence of high temperature evaporation using the Corona Discharge - Ion Mobility Spectrometry - orthogonal acceleration Time Of Flight (CD-IMS-oaTOF) technique. The IMS was operated in reverse gas flow mode at elevated temperatures of 438 K and 332 K (in zero air at atmospheric pressure. Additionally, the results of the IMS-oaTOF study on response to carbon tetrachloride (CCl4) admixture added to zero air in the negative CD and its influence on the composition of negative reactant ions produced in the discharge will be presented. The dopant CCl4 enhances the sensitivity of the IMS device for detection of explosives (Puton 2008; Borsdorf et al. 2011).

2. Experiment The CD-IMS-oaTOF device used in present study was developed at Department of Experimental Physics. It was described in detail in previous work (Sabo et al. 2014). The experimental device is depicted in Figure 1. The IMS consists of 8stainless steel rings electrodes separated by Teflon rings. The homogeneous electric field in drift tube has been secured by nine 5 MΩ resistors. The applied intensity of the electric field along the IMS drift tube was 436 V.cm-1. The IMS device was operated in zero air in reverse gas flow mode (Sabo and Matúška 2011). In this mode the O2-, O2- (H2O), N2O2-, N2O3- and N2O3- (H2O) RI were formed and used for chemical ionization. In the case of doping the CD by vapours of carbon tetrachloride in order to enhance sensitivity the main RI were Cl-(H2O)n (n=0,1,2) (Puton et al. 2008). The IMS was operated at atmospheric pressure and gas temperature of 438 K and in case of CCl4 dopant at 332 K. It is noteworthy to mention that the gas temperatures within the ionization region and the drift tube were derived from the ion mobility of N2O2- ions, using known value of the reduced mobility of 2.37 cm2V-1s-1 of this ion in zero air (Sabo and Matúška 2011). We used zero air purified with molecular sieve traps (Agilent) with flow 1.2 l/min as a drift gas. The dopant gas flow rate was 40ml/hour. The IMS was equipped with negative polarity CD ionization source in point to plane geometry. The point electrode of CD was prepared from tungsten wire of 100 µm diameter axially located against the centre of cathode’s orifice. The CD was supplied by high voltage power supplies and the potential difference was -3.5 kV across the discharge gap (-11.8 kV at point electrode and -8.3 kV at plane electrode). The typical CD current was 28 µA. The Bradbury-Nielsen type of shutter grid (SG) was operated at frequency of 80 Hz with a pulse width of 20 µs. The interface of IMS to oaTOF was described in detail in previous work (Sabo et al. 2013). The IMS-oaTOF instrument can be operated in three different regimes: i) IMS regime, where only IMS spectra are recorded, ii) TOF regime, where SG of the IMS is fully open and the mass spectra are recorded and iii) in two-dimensional (2D) regime, where 2D IMS-MS spectra are recorded. The operation parameters of the experiment are gathered in Table 1.

2

2.1. Sample introduction The RDX samples (38 mg or 33 mg in case of doping) placed in a small aluminium pot were inserted into a heatable cylindrical brass container of 5 cm height and of 4 cm diameter located outside the IMS as depicted in Figure 1. The container was heated by two small heaters of 125 W. Using this arrangement we were able to reach the temperature of 473 K close to the melting point RDX and far above the decomposition temperature of RDX of 443 K (Najarro et al. 2012). The present experiments were carried out at temperatures far below the critical temperature of the RDX of 840 K (Maximov 1992) so that the auto ignition of RDX was not possible; moreover, the low amount of the explosive limited the amount of the energy potentially released to very low value. The temperature of sample has been measured by digital thermometer during entire experiment and was stable. The nitrogen gas with flow rates 45 and 25 ml/min was used as a sample gas which flew through the heating system of sample and carried sample vapours into the IMS. At this low N2flow rate we assumed that the sample gas temperature was equal to the container’s temperature. 2.2. Gases and chemicals The mixture of high-purity oxygen O2 and nitrogen N2 (99.999%) in ratio 1:4 was used (zero air) as the drift gas at 1.2 l/min gas flow rate. The vapours of carbon tetrachloride CCl4 of purity 99 % were used as the dopant gas. The compound RDX was obtained from the Slovak Ministry of Defence at purity up to 99%. The sample was placed inside the container in solid phase, heated and its vapours were carried by N2 sample gas to the IMS and subsequently studied by IMS.

3. Results and discussion 3.1. Reverse gas flow mode The IMS device was operated in reverse gas flow mode during the entire experiment (Sabo and Matúška 2011). The gas flow through the CD is oriented in the opposite direction to the movement of ions in the CD ion source. In this flow regime the neutral products of the CD discharge are removed from the interaction region of the ions. The generation of negative RI in the CD in this mode at ambient temperature was examined and in detail described in our former study(Sabo and Matúška 2011). As it was stated there, the dominant reactant ions in reverse flow in zero air were O2-, O2- (H2O), N2O2-, N2O3- and N2O3- (H2O).These ions appeared as three peaks in IMS spectrum with reduced ion mobilities values (K0) of 2.37 cm2V-1s-1 (N2O2-), 2.27 cm2V-1s-1 (O2-, O2-(H2O)) and 2.2 cm2V-1s-1 (N2O3-, N2O3- (H2O)). In present study the temperature of the drift gas was set to 438 K in contrast to our earlier study performed at room temperature. Due to this difference in drift gas temperature the IMS spectra are slightly different. In IMS spectrum depicted in Figure 2a the RI in zero air consists of only two peaks at drift time (td) 7.0 and 7.58 ms and K0 of 2.37 cm2V-1s-1 (N2O2-) and 2.17 cm2V-1s-1, respectively. The assignment of these peaks to particular m/z can be made on the basis of Figure 2b (O2- with m/z 32 Da, O2-(H2O) with m/z 50 Da,N2O2- with m/z 60 Da, O2(H2O)2 with m/z 68 Da and N2O3-(H2O) with m/z 94 Da) (Sabo and Malásková 2014). The negative ions interact with the water molecules in IMS and one IMS peak thus consists of several ions and their water clusters. The composition of IMS peak of several ions and their water clusters can be explained on

3

the basis of the equilibrium between water cluster generation and dissociation of the clusters in drift tube (Sabo and Malásková 2014). 3.2. Interaction of RI with RDX at high temperature Thermal decomposition of RDX proceeds via several pathways, including N-NO2, C-H bond cleavage, dissociation and transfer of the O2 molecule from the NO2 to an adjacent CH2. Two most common pathways preferred during the thermal decomposition of RDX are N-NO2 bond rupture and symmetric concerted ring fission (Wu and Fried 1997). In the present work, the sample of RDX was heated to the 473 K and we expected that thermal decomposition of the RDX occurs already at temperature 443 K(Najarro et al. 2012). The RDX vapours were carried by N2 gas with temperature473 K into the IMS where reactions of the sample with RI took place. As a result several new peaks appeared in IMS spectrum (Figure 2a). The corresponding mass spectrum is depicted in Figure 2b and 2D map in Figure 2c. The first IMS peak (td=7.1ms and K0=2.32 cm2V-1s-1) was composed of five ions m/z 42, 46, 59, 62, 68 and 92 Da. The second one detected at 7.86 ms (K0=2.09 cm2V-1s-1) consisted according to 2D map of m/z 50, 60 and 75 Da, whereas m/z 50 and 60 Da represented RI. The third peak appeared at 8.34 ms (K0=1.97 cm2V-1s-1) and consisted of ions m/z 85 and 89 Da. Next peak detected at td=8.92 ms (K0=1.85 cm2V-1s-1) we assigned to m/z 105 Da. Two last peaks formed in td of 12.02 and 13.06 ms (K0 of 1.36 and 1.26 cm2V-1s-1) we ascribed to ions of m/z 221 Da and 264 Da. The ions m/z 42 Da we assigned to NCO-. As it was stated in recent studies (Oyumi and Brill 1985; Yinon 1982), the general tendency at higher temperatures for RDX is to undergo N-N bond homolysis (NO2, HONO, HCN production), whereas at lower temperatures the N-N bonds retain to a greater extent at the expense of C-N bond cleavage (CH2O, N2O production). Gongwer et al. detected hydrogen isocyanate HNCO as a volatile product of decomposition of RDX by IR spectroscopy (Gongwer et al. 1998). It has been proved (Gongwer and Brill 1998)that HNCO mainly comes from the pyrolysis of amides like C-hydroxyl-N-methyl formamide (HMFA) which represents another product of pyrolysis. We suggested that the formation of HNCO under the thermal decomposition is followed by ion-molecule reaction with RI to form NCO-. The ions m/z 46 Da have been observed in many previous studies and therefore we assigned it according to them to NO2-(Liebman et al. 1987; Ewing and Waltman 2009). An initial step for its generation is the cleavage of N-NO2 bond in RDX (Liebman et al. 1987). Thus, the thermal decomposition of the RDX leads to cleavage of NO2 fragment and NO2-isalso formed in reactions of RI with NO2 molecule: RI + RDX → NO2- + other fragment

(1a)

-

There exists also different pathway for NO2 formation: RI + NO2 → NO2- + other fragment

(1b)

In reaction (1a), NO2- is formed in direct reaction of RI with RDX accompanied with formation of neutral fragments (Farber 1979). Additional ions associated with thermal decomposition of RDX appear in MS at m/z 59 Da (Figure 2b). Snyder et al. (1989), who focused on characterization of RDX by pyrolysis-atmospheric pressure ionization - tandem mass spectrometry detected a protonated form of N-methylformamide (C2H5NO)H+ product. Another study by Gongwer et al. (1998) identified N-methylformamide (CH3NHCHO) and acetamide (CH3CONH2) as volatile, higher-molecular-weight products of RDX thermal decomposition. Both mentioned volatiles have the same

4

molecular weights, but they differ from each other by their structures. In present experiment we have detected for m/z 59 only one IMS peak. However, we are not sure about which of the ions (CH3NHCHO)- or (CH3CONH2)-is formed via charge transfer reaction or if both ions are formed and posses similar value of reduced ion mobility. The ions with m/z 62 Da drifted within the first IMS peak (Figure 2a) and we ascribed them to NO3- anions. These ions are usually formed as RI in negative CD in zero air under standard gas flow mode (Sabo et al. 2014).However, in reverse flow mode applied in present experiment, NO3- generation was suppressed and these ions did not appear among RI. For this reason we supposed that the detected NO3-ions originate from RDX. This is supported by the study of Asbury et al. (2000) who detected NO3- as fragment of thermal decomposition of RDX . The ions with m/z 92 Da (K0of 2.32 cm2V-1s-1) we considered to be NO2-(NO2) or N2O4- radical anions. These ions have also been detected in previous work (Sabo and Malásková 2014) in the mixture of synthetic air and atmospheric air in CD-IMS-oaTOF(Sabo and Malásková 2014). We assume that these ions are formed in association reactions of NO2- with neutral NO2 formed via thermal decomposition of RDX. The second IMS peak in Figure 2a (7.86 ms, K0=2.09 cm2V-1s-1) was composed of ions m/z 50 and 60 Da already known as RI and ions of m/z 75 Da. A key volatile with m/z 75 Da has been already mentioned in research of Liebman et al. (1987) who used analytical pyrolysis with triple quadrupole mass spectrometer for detection of complex products of RDX. The ion of m/z 75 Da is the strongest fragment detected in present study and it is unambiguously a product of RDX decomposition. We incline to the interpretation of Gongwer et al. (1998) who assigned it to C-hydroxyl-N-methylformamide C2H5NO2 (HMFA m/z 75 Da). As it was already mentioned above we assumed that the HMFA is formed by thermal decomposition and undergoes charge transfer reaction with RI. This substance is ionised in reaction with RI forming (C2H5NO2)-. Next fragment resulting from thermal decomposition of RDX observed by IMS-oaTOF m/z 85 Da appeared at td=8.32 ms (K0=1.97 cm2V-1s-1). It has been already mentioned by Liebman et al. (1987) and Snyder et al. (1989) in atmospheric pressure ionization (API) studies. This fragment has been postulated as triazine (C3H7N3). Yinon et al. (1982) studied structure of ions by collision induced dissociation (CID) of RDX ions ionized by chemical ionisation (CI) and triazine has appeared as one among other products of RDX. We assume that formation of this product in our experimental conditions is possible as well and (C3H7N3)- is formed in ion-molecule reactions of RI to RDX. The IMS spectrum of RDX exhibits peak observed at td of 8.88 ms (K0=1.85 cm2V-1s-1), which we related to m/z 105 Da. Since there are not any relevant references in literature, we tentatively assumed that these ions are formed in association reaction of earlier described neutral fragments ((CH3NHCHO) or (CH3CONH2)) with dominant product of thermal decomposition of RDX, NO2- ions, and generate complex ion (CH3NHCHO.NO2)or (CH3CONH2.NO2)-. IMS peak appearing in IMS spectrum at td of 8.32 ms and K0=1.97 cm2V-1s-1was composed of ions m/z 85 and 89 Da. Based on our best knowledge, the ions of m/z 89 have not yet been detected from RDX. We suspected that they originate from fragmentation of ions (CH3NHCHO.NO2)- or (CH3CONH2.NO2)- (m/z 105 Da) by cleavage of oxygen atom. The ions formed in IMS spectrum at td =12.02 ms (K0=1.36 cm2V-1s-1, m/z 221 Da) we assigned to (RDX-H)-. This is in agreement with Cheng et al. (2014) who applied O2- as RI in negative photoionization in IMS for

5

explosives exploration. In purified air, reaction of O2−reactant ions with RDX resulted in the formation of unstable adduct ions, RDX.O2-. Since the C-H bond in RDX sample is relatively weak, the hydrogen could be abstracted by O2-, forming product ions, (RDX –H)-. Therefore, we assumed a following reaction as a source of these ions O2- + RDX →RDX.O2- → (RDX-H)- + HO2

(2) 2

-1 -1

The last peak in IMS spectrum in Figure 2a drifting at td=13.06 ms (K0=1.26 cm V s ) we tentatively assigned to (RDX).NCO- with corresponding m/z of 264 Da. To our best knowledge, there do not exist any relevant information in literature about these ions as a response to RDX. We believe that the ion is formed via association reaction: RDX+ NCO- +M → RDX. NCO- + M

(3)

where M represents a neutral molecule acting as third particle (N2 or O2). 3.3 RDX detection with doping by CCl4 The ionization of RDX and explosives based on charge transfer reaction very often results in NO2- or NO3- which interact further with the explosives and their fragments. These cluster ions are usually instable and further disintegrate in IMS drift tube. It is responsible for deterioration of the selectivity and sensitivity of the IMS instrument. Better results can be achieved by ionization based on the associative ionization of the explosives by halogen negative ions. These ions are formed by doping the halogenated molecules (particularly with chloride) to ion source in order to provide Cl- reactant ions in the ionization region (Marotta et al. 2005). The ionization mechanism by Cl-reactant ions usually exhibits charge transfer, proton abstraction, adduct formation or nucleophilic displacement. Additionally, Cl- RI increase the detection sensitivity as more characteristic ionsare collected on IMS detector (Puton at al. 2008). The Cl- RI are in IMS present mainly in the form of Cl-.(H2O)n (n=0,1,2) (Puton et al. 2008; Marotta et al. 2005). In the case of Cl- RI the CD-IMS-oaTOF was working in reverse gas flow mode and the temperature of IMS drift gas was 332 K. The CD has been doped with 26 ppm CCl4 and this dopant passed together with nitrogen N2 directly through the CD gap. After doping the CD with CCl4 we have detected RI peaks with reduced ion mobilities of 2.51, 2.37, 2.28 and 2.18 cm2V-1s-1 (corresponding td of 8.46, 8.96, 9.32 and 9.76 ms) depicted in Figure 3a. The MS spectrum (Figure 3b) exhibits m/z 35, 37, 53, 55, 71, 73, 81, 95, 125 and 127 Da. According to our earlier studies on doping of the CD-IMS by CCl4(Sabo and Malásková 2014) and on basis of present 2D map (Figure 3c), we assigned the first RI peak with K0=2.51 cm2V-1s-1 (composed of ions m/z 35, 37, 53, 55,71 and 73 Da) to Cl-.(H2O)n (n=0,1,2). The second RI peak K0=2.37 cm2V-1s-1 in IMS spectrum consisted of m/z = 35,37 and 81, 83 Da, the third one K0=2.28 cm2V-1s-1 of m/z 35, 37 and 95, 97 Da and the last one K0=2.18 cm2V-1s-1 of m/z = 125 and 127 Da. The ions m/z 35, 37 Da at IMS peaks 2.37 and 2.28 cm2V-1s-1 presumably originate from fragmentation of larger RI ion on IMS-oaTOF interface. The assignment of m/z 35 and 37 Da is straightforward ascribed to

35

Cl- and

37

Cl- formed via dissociative

electron attachment to CCl4 in CD. The presence of 35Cl- and 37Cl- in mass spectrum is easy to be recognised by the typical isotopic ratio 3:1 of these isotopes and this ration is typical also for all ions exhibiting Cl-. The ions with m/z 53 and 55 Da we have assigned to

35

Cl-(H2O) and

37

Cl-(H2O) (n=1) and m/z 71 and 73 Da to

35

Cl-

(H2O)2 and 37Cl-(H2O)2 (n=2). In previous study (Sabo and Malásková 2014) we did not detect m/z 71 and 73 Da (within K0=2.51 cm2V-1s-1 peak). We believe that it was due to higher temperature of drift gas in that study and

6

thus cluster distribution was shifted to smaller clusters. The ions m/z 81 and 83 Da we ascribed to Cl-(NO2) and m/z 95 and 97 Da ions to Cl-(NO)2 clusters (Sabo and Malásková 2014). The slowest RI peak (td=9.76 ms and K0=2.18 cm2V-1s-1) was composed of Cl-(NO)3 (m/z =125 and 127 Da) ions. The comparison of RI in present study and former study (Sabo and Malásková 2014) shows very similar composition, but the values of reduced ion mobilities of RI ions are slightly different. We believe that this is mainly due to differences

in the temperature of the drift gas (Borsdorf and Mayer 2012;Tabrizchi and

Rouholahnejad 2005). The gas temperature also affects the relative abundance of formed product ions, the gas density, the mean free path of ions, the collision rate (Borsdorf and Mayer 2012).The differences in reduced mobilities also arise from differences in cluster size distribution due to temperature effect (Mayer and Borsdorf 2014).The most pronounced difference in value of reduced ion mobility to former study was detected for the first IMS peak (td=8.46 ms) and we believe that it is due to different water clusters sizes. The lower temperature in present study resulted in larger distribution of cluster sizes and this affected the value of the reduced ion mobility. The introduction of RDX sample into IMS resulted in appearance of the IMS peak with reduced mobility 2.37 cm2V-1s-1 (m/z 46 Da). These ions have been detected also in operation without CCl4 dopant (the K0 was slightly different) and have been ascribed to NO2- ions. We supposed that NO2- result from ion-molecule reaction of RI with NO2 and with RDX (Sabo and Malásková 2014). Within this IMS peak we have detected also ions m/z 60 Da, which we assigned to N2O2-. We assumed that this ion originates from following process at atmospheric pressure (4) (Sabo et al. 2010) O- + N2O + M → N2O2- + M

(4)

where M is a third particle (N2, or O2). The O- ions are formed via dissociative electron attachment to O2 in the CD in the vicinity of the tip. These ions have very short lifetime in the high pressure environment, due to rapid conversion via ion molecule reactions with molecules and radicals. We suppose that N2O molecules are formed by thermal decomposition of RDX (Dong et al. 2012; J. Stals 1970; Liebman et al. 1986; Oyumi and Brill 1985) and subsequently converted into the ions N2O2- in the reaction (4). It is noteworthy, we have been able to detect the ions N2O2- despite of the fact they are preferably formed at room temperature (Sabo and Malásková 2014). The ions with the most significant response in IMS spectrum at td=9.32 ms (K0=2.28 cm2V-1s-1 m/z 62 Da) (Sabo et al. 2013) we attributed to the NO3- ions. The NO3 molecules represent one of the major fragments of thermal decomposition of RDX. Since the electron affinity of NO3 molecules (3.93 eV) (Weaver et al. 1991) is higher than electron affinity of Cl- ions (3.61 eV) (Sabo and Malásková 2014), the ionization occurs via charge transfer reaction. Another IMS peak formed (td=9.76 ms, K0=2.18 cm2V-1s-1 and m/z 88 Da) we attributed according to Yinon et al. to CH2N3O2-(Yinon et al. 1982) and the ions detected (td=10.54 ms, K0=2.01 cm2V-1s-1 and m/z 98 Da) we also associate to RDX fragment in IMS. Stals et al. (1970), Snyder et al. (1989)and Yinon et al. (1982) identified it as C3H4N3O- and their appearance is result of pyrolysis, negative ion chemical ionization or collisional induced dissociation. Finally, we have detected in the IMS spectrum peak at td=14.1 ms (K0=1.51 cm2V-1s-1and m/z 257 Da), please look at Figure 3a and 3c. We associated this ion with (RDX).Cl-, which is formed via reaction: RDX + Cl- → (RDX).Cl-

(5)

7

These adduct ions (RDX).Cl- have been reported as a dominant peak in recent work of Asbury with reduced mobilities 1.44 and 1.4 cm2V-1s-1, respectively (Asbury 2000). Due to thermal decomposition of large part of RDX in high temperature evaporation source, in present experiment we have seen this peak only at lower intensity.

4. Conclusion We have studied the formation of ions from RDX and its neutral fragments formed after thermal decomposition process reached by evaporation of RDX at high temperature (473 K). The negative CD was operated in zero air and zero air doped by CCl4and the drift gas temperature was set to 437 K or 332K (in case of dopant). The negative reactant ions formed in the CD were analyzed by the CD-IMS-oaTOF method. The negative ions formed in interaction of RI with RDX and the thermal decomposition products were detected and identified. The application of carbon tetrachloride CCl4 as dopant was in order to prepare different RI and to achieve higher stability and sensitivity of formed products.

Acknowledgments This research was supported by the Slovak research and development agency project Nr. APVV-0259-12 and project VEGA 1/0417/15.

References Asbury G R, Klasmeier J, and Hill H H (2000) Analysis of explosives using electrospray ionization/ion mobility spectrometry (ESI/IMS). Talanta 50: 1291–98 Baim M A, Eatherton R L, Hill H H (1983) Ion mobility detector for gas chromatography with a direct photoionization source. Anal Chem 55:1761–66 Borsdorf H, Mayer T (2011) Response of halogenated compounds in ion mobility spectrometry depending on their structural features. Talanta 83: 815–22 Borsdorf H, Mayer T (2012) Temperature dependence of ion mobility signals of halogenated compounds. Talanta 101: 17–23 Crawford C L L and Hill H H (2013) Comparison of reactant and analyte ions for 63nickel, corona discharge, and secondary electrospray ionization sources with ion mobility-mass spectrometry. Talanta 107: 225–32 Dong X F, Yan Q L, Zhang X H, Cao D L, Xuan Ch L (2012) Effect of potassium chlorate on thermal decomposition of cyclotrimethylenetrinitramine (RDX). Journal of Analytical and Applied Pyrolysis 93: 160–64 Eiceman G A, Nazarov E G and Stone J A (2003) Chemical standards in ion mobility spectrometry. Analytica Chimica Acta 493: 185–94 Ewing R G and Waltman M J (2009) Mechanisms for negative reactant ion formation in an atmospheric pressure corona discharge. International Journal for Ion Mobility Spectrometry 12: 65–72 Farber M, Srivastava R D (1979) Mass spectrometric investigation of the thermal decomposition of RDX. Chemical Physics Letters 64: 307–10 Gapeev A, Sigman M, Yinon J (2003) Liquid chromatography/mass spectrometric analysis of explosives: rdx adduct ions. Rapid communications in mass spectrometry : RCM 17: 943–48

8

Gongwer P E, Brill T B (1998) Thermal decomposition of energetic materials 73: the identity and temperature dependence of ‘minor’ products from flash-heated RDX. Combustion and Flame 115: 417–23 Hilton Ch K, Krueger C A, Midey A J, Osgood M, Wu J, Wu Ch (2010) Improved analysis of explosives samples with electrospray ionization-high resolution ion mobility spectrometry (ESI-HRIMS). International journal of mass spectrometry 298: 64–71 Huwei L, Ruonong F (1989) Investigation of thermal decomposition of HMX and RDX pyrolysis-gas chromatography 138: 167–71 Cheng S, Wang W, Zhou Q, Chen Q, Peng L, Hua L (2014) Fast switching of CO3-(H2O)n and O2-(H2O)n reactant ions in dopant-assisted negative photoionization ion mobility spectrometry for explosives detection. Analytical chemistry 86: 2687–93 Stals J (1970) Chemistry of aliphatic unconjugated nitramines. Trans. Faraday Soc. 67: 1768–75 Liebman S A, Duff P J, Fickie K D, Schroeder M A, Fifer R A (1986) Degradation profile of propellant systems with analytical pyrolysis/concentrator/gc technology. Journal of Hazardous Materials 13: 51–56 Liebman S A (1987) Time-resolved analytical pyrolysis studies of nitramine decomposition with a triple quadrupole mass spectrometer system. Journal of Analytical and Applied Pyrolysis 12: 83–95 Marotta E, Scorrano G, Paradisi C (2005) Ionic reactions of chlorinated volatile organic compounds in air plasma at atmospheric pressure. Plasma Processes and Polymers 2: 209–17 Maksimov, Y Y (1992) Boiling point and enthalpy of evaporation of liquid hexogen and octagon. Russ. J. Phys. Chem. 66:280-281 Mayer T, Borsdorf H (2014) Accuracy of ion mobility measurement dependent on the influence of humidity. Anal Chem 86:5069-5076 Najarro M, Morris M E D, Staymates M E, Fletcher R, Gillen G (2012) Optimized thermal desorption for improved sensitivity in trace explosives detection by ion mobility spectrometry. The Analyst 137: 2614–22 Oyumi Y, Brill T B (1985) Thermal decomposition of energetic materials 3. A high-rate, in situ, ftir study of the thermolysis of rdx and hmx with pressure and heating rate as variables. Combustion and Flame 62: 213–24 Puton J, Nousiainen M, Sillanpää M (2008) ion mobility spectrometers with doped gases. Talanta 76: 978–87 Sabo M, et al. (2010) Atmospheric pressure corona discharge ionisation and ion mobility spectrometry / mass spectrometry study of the negative corona discharge in high purity oxygen and oxygen / nitrogen mixtures. International Journal of Mass Spectrometry 293: 23–27 Sabo M, Lichvanová Z (2014) Experimental simulation of negative ion chemistry in martian atmosphere using ion mobility spectrometry-mass spectrometry. The European Physical Journal D 68: 216 Sabo M, Malásková M, Matejčík Š (2014) Ion mobility spectrometry–mass spectrometry studies of ion processes in air at atmospheric pressure and their application to thermal desorption of 2,4,6-trinitrotoluene. plasma sources science and technology 23: 15025 Sabo M, Matejčík (2012) Corona discharge ion mobility spectrometry with orthogonal acceleration time of flight mass spectrometry for monitoring of volatile organic compounds. Anal Chem 84: 5327–34 Sabo M, Matúška J (2011) Specific O2 generation in corona discharge for ion mobility spectrometry. Talanta 85: 400–405

9

Sabo M, Okuyama Y, Kučera M, Matejčík Š (2013) transport and stability of negative ions generated by negative corona discharge in air studied using ion mobility-oatof spectrometry. International Journal of Mass Spectrometry 334: 19–26 Snyder A P (1989) Characterization of cyclotrimethylenetrinitramine (RDX) by N, H Isotope Analyses with pyrolysis-atmospheric pressure ionization tandem mass spectrometry. Organic Mass Spectrometry 24: 15– 21 Tabrizchi M, Ilbeigi V (2010) Detection of explosives by positive corona discharge ion mobility spectrometry. Journal of hazardous materials 176: 692–96 Tabrizchi M, Rouholahnejad F(2005) Comparing the effect of pressure and temperature on ion mobilities. Journal of Physics D: Applied Physics 38: 857–62 Tam M, Hill H H (2004) Secondary electrospray ionization-ion mobility spectrometry for explosive vapor detection ( SESI ) as a nonradioactive ionization source to detect. Anal Chem 76: 2741–47 Weaver A, Arnold D W, Bradforth S E, Neumark D M (1991) Examination of the 2A’2 and 2E‘ States of NO3 by Ultraviolet Photoelectron Spectroscopy of NO−3. The Journal of Chemical Physics 94: 1740 Wu Ch, Fried L E (1997) Ab initio study of rdx decomposition mechanisms. J Phys Chem A 101: 8675–79 Yinon D H, Hass (1982) Mass spectral fragmentation pathways in rdx and hmx . a mass analyzed ion kinetic energy spectrometric/ collisional induced dissociation study.Organic Mass Spectrometry 17: 321–26 Yinon J (1982) Mass Spectrometry of Explosives : Nitro Compounds , Nitrate Esters , and Nitramines. Mass Spectrometry Reviews 1:257–307

10

Figures caption Fig. 1 The schematic view of the CD-IMS-oaTOF experimental apparatus operated in reverse gas flow mode. The arrows indicate the direction of the gas flow

Fig. 2 Response to RDX sample evaporated at 473 K at1.66 ppb concentration. IMS was working in reverse flow andnegative CD in zero air a) in IMS spectrum at 438 K b) corresponding MS spectrum c) two dimensional IMS-MS spectrum

Fig. 3 Response to RDX sample evaporated at 473 K at 5.16 ppb concentration. IMS was working in reverse flow and negative CD in zero air doped with 26 ppm of CCl4a) in IMS spectrum at 332 K b) corresponding MS spectrum c) two dimensional IMS-MS spectrum

Table caption Table 1Operation parameters of the CD-IMS-oaTOF device

11

Figure 1.

12

Figure 2.

13

Figure 3.

14

Table 1. IMS drift tube length Drift field intensity IMS operating temperatures Drift gas flow rate Sample gas flow rate Dopant vapour gas flow

11.44 cm 436 V.cm-1 332 K, 438 K 1000 ml/min 25, 45 ml/min 40 ml/ hour

CD current SG pulse width/frequency Average IMS resolving power (t/Δt) TOF pulse width TOF pulse frequency

28 µA 20µs/80Hz 50 10 µs 40 kHz

TOF acceleration voltage Mass accuracy Average TOF resolving power

2185 V 0.1 Da 800

15