Dissociation Enthalpies of Chloride Adducts of Nitrate and Nitrite

Subscriber access provided by BRIGHAM YOUNG UNIV

Article

The Dissociation Enthalpies of Chloride Adducts of Nitrate and Nitrite Explosives Determined by Ion Mobility Spectrometry Maneeshin Yasassri Rajapakse, Peter E Fowler, Gary A Eiceman, and John A. Stone J. Phys. Chem. A, Just Accepted Manuscript • DOI: 10.1021/acs.jpca.5b10765 • Publication Date (Web): 18 Jan 2016 Downloaded from http://pubs.acs.org on January 26, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry A is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

The Dissociation Enthalpies of Chloride Adducts of Nitrate and Nitrite Explosives Determined by Ion Mobility Spectrometry Maneeshin Y. Rajapakse†, Peter E. Fowler†, Gary A. Eiceman*† and John A. Stone‡. †

Department of Chemistry and Biochemistry, New Mexico State University, Las Cruces, New Mexico, 88003, United States ‡

Department of Chemistry, Queen’s University, Kingston, Ontario K7L 3N6, Canada

ABSTRACT The kinetics for thermal dissociations of the chloride adducts of the nitrate explosives 1,3-dinitroglycerin (1,3-NG), 1,2-dinitroglycerin (1,2-NG), the nitrite explosive 3,4dinitrotoluene (3,4-DNT), and the explosive taggant 2,3-dimethyl-2,3-dinitrobutane (DMNB) have been studied by atmospheric pressure ion mobility spectrometry. Both 1,3-NG·Cl‾ and1,2NG·Cl‾ decompose in a gas-phase SN2 reaction in which Cl‾ displaces NO3‾ while 3,4-DNT·Cl‾ and DMNB·Cl‾ decompose by loss of Cl‾. The determined activation energies (kJ mol-1) and pre-exponential factors (s-1) values for the dissociations respectively are: 1,3-NG·Cl‾, 86 ± 2, 2.2 x 1012; 1,2-NG·Cl‾, 97 ±2, 3.5 x 1012; 3,4-DNT·Cl‾, 81 ± 2, 4.8 x 1013; and DMNB·Cl‾, 68 ± 2, 9.7 x 1011. Calculations by density functional theory show the structures of the nitrate ester adducts involve three hydrogen bonds, one from the hydroxyl group and the other two from the two nitrated carbons. The relative Cl‾ dissociation energies of the nitrates together with the previously reported smaller value for glycerol trinitrate and the calculated highest value for glycerol,1-mononitrate are explicable in terms of the number of hydroxyl hydrogen bond participants. The theoretical enthalpy changes for the nitrate ester displacement reactions are in

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

agreement with those derived from the experimental activation energies but considerably higher for the nitro-compounds. Keywords: ion mobility, nitro- and nitrate-explosives, kinetics, enthalpy, rate constants, chloride adducts

INTRODUCTION Ion mobility spectrometry (IMS) at atmospheric pressure is a fast, sensitive, inexpensive and portable explosive detection technique.1 It is used worldwide in both military and civilian venues such as on the battlefield, at airports, and at crime scenes.2, 3 The basis of the technique is gas-phase ion-molecule chemistry that provides for the detection of molecules desorbed from surfaces or already present in the atmosphere at sub parts per million by volume (ppmv) concentrations.1, 4 In IMS, a measurement is dependent on the reaction of an analyte of interest with a gas-phase reactant ion, most often generated in deployed instruments by electrons from a 63

Ni source. The ion product(s) is characterized for mobility as swarms in a weak electric field.

Explosives are determined with an ion mobility spectrometer operating in the negative mode because unique product ions can usually be produced with high selectively.5, 6 The chloride ion is a much-used, easily-formed, reactant ion for such reactions because it is the conjugate base of HCl, a strong gas-phase acid, and is therefore not prone to proton abstraction from many analytes. Chloride adducts are readily formed with gas-phase acids weaker than HCl and the adducts are useful for the detection of both nitrate- and nitro-containing compounds.5-11 Hand-held, field-operable, mobility spectrometers of necessity operate at ambient temperatures because of power constraints but static instruments without such constraints may operate at a higher temperature where interfering reactions usually decrease. Upper limits on temperature exist because the association of an ion with a molecule to form an ionic adduct is


Page 2 of 36

Page 3 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


always an exothermic process and adduct ions will not be observable above certain temperatures. Adduct ions will be most stable at low temperature though once formed will dissociate at an increasing rate as the temperature is raised. The upper temperature limit for observation and the rate of dissociation at temperatures below this limit will be different for different analytes. Ion mobility spectrometry is well-suited for observing the stability of ion-molecule adducts as a function of temperature and can, under favorable circumstances, be used to obtain rate constants and activation energies for adduct formation 12, 13 and adduct dissociation.14-18 The advantage of producing and observing ions drifting in low electrostatic fields at atmospheric pressure is that the ions, present in extremely low concentration, are essentially in thermal equilibrium with the surrounding gas molecules and are at the well-defined gas temperature. The increase in translational kinetic energy and hence effective temperature due to a drift field of ca. 250 V cm-1, calculated using the two temperature theory of Viehland and Mason, is negligible compared to its thermal energy.19 For example, an ion traveling in air with a thermal temperature of 373 K and in a field of 250 V cm-1 has an effective temperature of 373.1 K. In addition, the collision frequency of an ion in a gas at one atmosphere pressure is greater than 1010 s-1, resulting in very efficient stabilization of hyper-energetic nascent adducts. The chloride adducts (M·Cl‾) of some nitrate- and nitro-explosives M have rate constants for dissociation in the temperature range from 70 to 200 oC that is available for experimentation in an ion mobility spectrometer. 5,20,16 In this paper, we report the results of a study of the dissociation reactions of M·Cl‾ for 1,2- and 1,3-dinitroglycerin, that are both explosive propellant components, and the decomposition products of the important trinitroglycerin. They are present as metabolites of microbial action at sites where trinitroglycerin has been used as an explosive and in blood plasma and sewage when it is used as a vasodilator.21-26 Comparison is



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

made with the kinetic data obtained for the thermal dissociation of the chloride adduct of the nitrite ester explosives 3,4-dinitro toluene and 2,3,6-trinitrotoluene and that of 2,3-dimethyl-2,3dinitrobutane. The latter is not an explosive but is a detection taggant for explosives because its favorable vapor pressure allows IMS detection at ppbv levels. Rate constants for the dissociation of each adduct were measureable over a narrow, unique temperature range, allowing the activation energy to be determined. Computed structures and reaction enthalpies for competitive dissociation products are compared with those derived from the experimental activation energies and show the importance of the type and positions of the hydrogen bonds in the relative stability of the chloride adducts.

EXPERIMENTAL Kinetic ion mobility spectrometer A ion mobility spectrometer equipped with dual ion shutters and a gas chromatograph inlet for pre-separation has been described previously in detail.16 The 19.10 cm long drift tube is divided into three sections by two sequential ion shutters of Tyndall-Powell design, Shutter 1 and Shutter 2, giving a combined source and reaction region of 6.50 cm, a 12.60 cm drift length when only Shutter 1 is in operation and a 6.60 cm drift length when only Shutter 2 is in operation. Ions are formed with a 10 mCi 63Ni source. The whole mobility spectrometer is housed in a temperature regulated oven to ensure a steady-state temperature throughout the instrument with less than 1% variation of the drift gas temperature from drift gas inlet to drift gas outlet of the drift tube. The 0.10 mm id DB-5 gas chromatograph column is 30 m long with nitrogen (>99% purity by Airgas Inc., Radnor, PA) as carrier gas. The injector port was held at 120 oC and the


Page 4 of 36

Page 5 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


column at 100 oC. The sample transfer line between the chromatograph and the ion source was at 100 oC, as was the line carrying a stream of air containing the CCl4 dopant. The 280 mL min-1 drift gas, purified air (generated by an Aadco 737 series pure air generator, Cleves, OH), was dried by passing through a 1.5 m molecular sieve column. Ion mobility-mass spectrometer (IMS-MS) The identities of the ions present in the kinetic ion mobility spectrometer were investigated in a specially constructed mobility drift tube interfaced to a mass spectrometer. A single shutter mobility spectrometer was assembled by stacking 0.30 cm thick, 1.50 cm inner diameter and 3.80 cm outer diameter stainless steel electrodes with 1.22 cm thick, 1.96 inner diameter and 5.00 cm outer diameter Teflon spacers. Two aluminum end flanges hold the rings in compression. The ion source is a 10 mCi 63Ni foil in a brass holder fitted into a Teflon ring. The distance from the source end flange to the detector end flange is 25.0 cm. The drift tube is divided into two sections by an ion shutter, a combined source and reaction region of 14.9 cm and a drift region of 9.7 cm. The drift tube is heated to a desired temperature by a circumferential rope heater (Omega, Stamford, CT). The mobility spectrometer is interfaced to a gas chromatograph (Hewlett-Packard 5890 series ii) by passing the end of a 5 m long capillary column (DB-5, J&W Scientific Inc., Folsom, CA) via a heated transfer line into the reaction region. Headspace vapor over liquid CCl4 was introduced using the same transfer line, which in all experiments was kept at 100 oC. The chromatographic conditions used were as follows: injector 150oC, initial column temperature 70oC, final column temperature150 oC, ramp 20 oC min-1, carrier gas N2 (>99% purity, Airgas Inc., Radnor, PA), split ratio 20:1. Drift gas was air purified by passage through a molecular sieve tower to maintain a moisture level of ~ 1 ppmv. A 1.58 mm ID stainless steel tube extending from the 1.58 mm diameter hole of the Faraday



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

detector of the drift tube is attached to the 1.58 mm diameter heated capillary inlet of the mass spectrometer (Shimadzu-2010 LCMS, Shimadzu Cooperation, Kyoto, Japan) via a stainless steel Swagelok union (Swagelok Company, Solon, OH) to form the GC-IMS-MS system. Chemicals and Reagents: Standard solutions for trinitroglycerine (NG), 1,2-dinitroglycerine (1,2-NG), 1,3dinitroglycerine (1,3-NG), 3,4-dinitrotoluene (3,4-DNT), 1-mono nitroglycerine (MN), dimethyl dinitro butane (DMNB), and pentaerythrityl tetranitrate, (PETN), were obtained at a concentration of 100 ng µL−1 in methanol from AccuStandard Inc. (New Haven, CT). Reagent grade carbon tetrachloride was from Sigma Aldrich Chemical Co. (Milwaukee, WI). Procedures: The flow of CCl4 dopant in the kinetic ion mobility spectrometer was adjusted to obtain the highest intensity for the chloride peak in the mobility spectrum in the absence of analyte. One µL of standard was injected into the gas chromatograph and mobility spectra were recorded for 8 minutes using Linear IMS_1.0.vi software and a PCI 6035E interface card (National Instrument Corporation, Austin, TX) with 50 digitally averaged scans recorded for each stored spectrum, 163 spectra in total. The mobility spectrometer temperature was varied from 60 to ~ 195 oC for each sample to observe the evolution of the spectrum and determine the range over which ion decomposition was observable. The range was narrow, of the order of 10 to 20 oC, and was different for each of the analytes. The instrument was stabilized overnight at a set temperature with all gases flowing to ensure a steady-state temperature throughout the system before any kinetic measurement. Mobility spectra were obtained in three modes: with only Shutter 1 operating; with only Shutter 2 operating; and with both shutters operating but with a time delay ∆t between the


Page 6 of 36

Page 7 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


opening of Shutter 1 and the later opening of Shutter 2 where ∆t was the time of the peak maximum for an ion of interest, injected through Shutter 1 and isolated by Shutter 2 for injection into the 6.60 cm drift region. Both shutters were operated with 0.300 ms pulse widths to isolate an ion of interest and obtain the spectra for the thermal decomposition experiments. Multiple IMS spectra were recorded over the gas chromatographic runtime. In measurements by GC IMS MS, 1 µL of sample was injected into the chromatograph and the ion mobility spectrometer was operated with chloride as the reactant ion, a drift tube electrostatic field of 400 V cm-1, and the ion shutter open to pass all ions to the mass spectrometer. The capillary tube inlet of the mass spectrometer was maintained at 60 oC to optimize the ion transfer efficiency from the drift tube to the mass spectrometer. Experiments were made with drift tube temperatures between 70 to144oC both to identify the product ions and determine their thermal stabilities. Mass spectra were recorded using a 2010 MS and LCMSsolution software (Shimadzu Cooperation, Kyoto, Japan) over the m/z range from 30 to 600Da. Ab initio Calculations SPARTAN 10 (Wave function, Inc., Irvine, CA) software with density functional theory (DFT) and the 6-311+G(d,p) basis set was used to provide energy-minimized chemical structures for the reactant ions, product ions, and ion-molecule intermediates.

RESULTS AND DISCUSSION Mass spectrometric investigations Since the sensitivity of the GC-IMS-MS instrument was insufficient for investigations with the mobility spectrometer operating in the pulsed mode, the ion shutter was disabled and the 7 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 36

full spectrum of ions emanating from the ion source were determined with ion drift conditions characteristic of the kinetic ion mobility spectrometer. The low sensitivity was almost certainly due to a very inefficient transmission of ions through the mobility spectrometer – mass spectrometer interface. At 70 oC, each of the nitroglycerin explosives, NG, 1,2-NG and 1,3-NG, and also the non-explosive 1-NG exhibited M·Cl‾ as the base peak and only product ion peak in the mass spectrum. Adduct formation (Eq. 1) was the only reaction between the chloride reactant ion and the glycerin nitrates. M + Cl‾ → M·Cl‾

(1)

When a high concentration of sample was present in the ion source then a chloride bound dimer M2Cl‾ also appeared (Eq. 2). M·Cl‾ + M → M2·Cl‾

(2)

At 122oC, the relative intensity of NG·Cl‾ decreased as two new peaks appeared in the spectrum and these were NO3‾ at 62 Da and the nitrate adduct, NG·NO3‾, at 289 Da. The thermal decomposition of NG·Cl‾ proceeds by the displacement of NO3‾ by Cl‾ (Eq. 3) with the free NO3‾ forming the nitrate adduct by association with the ambient NG of the source region (Eq. 4).16 NG·Cl‾ → [NG − NO3 + Cl] + NO3‾

(3)

NG + NO3‾ → NG·NO3‾

(4)

When the temperature was raised to 144oC, NG·NO3‾ was the only ion in the spectrum. At this temperature, NO3‾ appeared in the 1,2-NG and 1,3-NG spectra as did adducts of M·NO3‾ at 244 Da, although the intensity of 1,2-NG·NO3‾ was very weak and the chloride adducts were still present with good intensity. In contrast to the more highly nitrated glycerins, the spectrum of 1-


Page 9 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


NG at both 122 oC and 144 oC contained only the peaks due to NG·Cl‾ at 172 Da and 174 Da for the two chlorine isotopes. These results show that the thermal stability of a Cl‾ adduct of a glycerin nitrate ester is dependent both on the number of nitrate groups and their positions. NG has the least stable adduct while comparison between the adducts of 1,2-NG and 1,3-NG shows that the adjacent NO3 groups in 1,2-NG·Cl‾ leads to greater thermal stability compared with the non-adjacent ones in 1,3-NG·Cl‾. Surprisingly, 1-NG has the most stable chloride adduct. As described later, this order of stabilities of is confirmed by both experimentally determined reaction activation energies for the dissociation of the adducts and by theoretical calculations. Tetranitrate ester pentaerithrytol tetranitrate (PETN) was also investigated. At 122 oC the PETN spectrum was considerably weaker than that of the other compounds and showed evidence of decomposition of the molecule before entering the ion source as has been previously reported.27 The products were: PETN35Cl‾, 351Da; PETN37Cl‾, 353Da; PETN·NO3‾, 378Da; [PETN-H]‾, 315Da; and PETN·NO2‾, 362Da. The ion intensities with this compound were low and decreased with increasing temperature precluding further investigation of PETN·Cl‾. TNT had only one peak in its mass spectrum at any temperature, [TNT-H] ‾ at 226Da, while 3,4-DNT, with only two NO2 substituents compared with the three of TNT, showed only the two 3,4-DNT·Cl‾ chlorine isotopic peaks at 217 Da and 219Da at 70 oC. At higher temperatures, [3,4-DNT-H]‾ at 183Da appeared with increasing intensity relative to the intensity of the carrier gas adduct. DMNB formed a very stable chloride adduct at 70 oC and also a chloride bound dimer (DMNB)2Cl‾. The ion intensity was poor and decreased at higher temperatures so that a definitive statement regarding adduct decomposition could not be made.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 36

Thermal dissociation of chloride adducts in the kinetic ion mobility spectrometer The chloride reactant ion was the only ion present in the mobility spectrum in the absence of sample in the chromatographic eluent. An eluting sample that reacted with chloride was evidenced by a decrease in the reactant ion peak and the appearance of one or more new product ions. Examples of the mobility spectra are shown in Figure 1 for DMNB at 80 oC and Figure 2 for 1,3-NG at 155 oC. Ion identification was by correlation of the effect of temperature on the mass spectra obtained in the GC-IMS-MS experiments with the peaks observed in the ion mobility spectra here. In addition, further confirmation was obtained by comparison of calculated reduced mobilities Ko (Eq. 5) with those available in the literature. In the equation the ion mobility K = L/Et where L (cm) is the drift length, E (V cm-1) the electrostatic drift field and t (s) the drift time is normalized to standard temperature 273 K and pressure 760 mm Hg-1. The equation treats air as an ideal gas which is justified because in the range of temperatures and pressures employed the number density of air deviates considerably less than 1% of that of an ideal gas.28 Ko = K

273 P L 273 P = T 760 Et T 760

(5)

Comparison of the experimental Ko values with available literature values is presented in Table 1. No values are available for the glycerin dinitrate ions or for 3,4-DNT·Cl‾ but excellent agreement with literature values is obtained for NO3‾ and [TNT-H]‾ and there is reasonable agreement for DMNB·Cl‾. The introduction of DMNB at 80 oC resulted in only one new peak in the spectrum, identified in Figure 1a as the adduct DMNB·Cl‾. This product has previously been reported from an IMS-MS experiment.5 There may be a second product present that appears as an inflection on the shoulder of the Cl‾ peak but insufficient resolution prevented confirmation. Ion intensity 10 ACS Paragon Plus Environment

Page 11 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


between the Cl‾ peak and the DMNB·Cl‾ peak denotes that decomposition of DMNB·Cl‾ has occurred in the drift region.14 A decomposing ion results in the formation of a product ion of lower mass whose mobility is higher and drift time lower than its precursor. Decomposition occurring in the drift region results in signal appearing in the spectrum between the product and precursor peaks because the charge is carried part of the way by the slower-moving precursor and the rest of the way by the faster product. Operation of dual ion shutters allows the gating of the precursor adduct into the drift region with no interference from other ions, especially product ions formed before the second ion shutter. A detailed description of adduct ion isolation by the dual shutter method and kinetic parameter calculations are presented in prior publications.14-17 Allowing only DMNB·Cl‾ into the 6.60 cm drift region by setting a suitable delay between the opening Shutter 2 and that of Shutter 1 produced the spectrum in Figure 1b. The time scales are different for the figures because of the different drift lengths but peaks may be correlated because drift times are proportional to drift lengths, 12.60 cm for Figure 1a and 6.60 cm for Figure 1b.The only ion product of the dissociation of DMNB·Cl‾ is clearly Cl‾. The spectrum obtained with 1,3-NG and chloride reactant at 155 oC in Figure 2a contains three products, NO3‾, 1,3NG·Cl‾ and 1,3-NG·NO3‾ (Figure 2a). The raised baseline shows that NO3‾ is the product of the dissociation of either or both of 1,3-NG·Cl‾ and 1,3-NG·NO3‾. revious IMS experiments found that the thermal decomposition of NG·Cl‾ produced exclusively NO3‾ and this was ascertained to be true for 1,3-NG·NO3‾ by using the dual-shutter capabilities of the instrument as shown in the spectrum of Figure 2b16 where NO3‾ is the only product ion. By comparison, isolation of 1,2-NG·NO3‾ at 155 oC showed the ion to be stable with no sign of its decomposition. Isolation of this ion at higher temperatures showed that decomposition produced NO3‾. Surprisingly, the adduct of the mononitrate, 1-NG·Cl‾, was the most stable of all



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 36

the chloride adducts showing no evidence of decomposition even at the highest available experimental temperature of 196 oC. The isolated chloride adduct of 3,4 DNT like that of DMNB was found to thermally decompose by loss of Cl‾, with no evidence for competing formation of NO2‾. Determination of the activation energies for the thermal decomposition of chloride adducts The decomposition of a chloride adduct in its passage through the drift region is the first order reaction of Eq. 6 in which the product X‾ was determined to be Cl‾ for 3,4-DNT·Cl‾ and DMNB·Cl‾ and NO3‾ for 1,2-NG·Cl‾ and 1,3-NG·Cl‾. N was M for the nitrite esters and [M − NO3 + Cl] for the nitrate esters. M·Cl‾ → N + X‾

(6)

The rate constant for the reaction is determinable using the ion isolation method from the time at which dissociation occurs and the concentration of ions present at that time. If td is the time for non-decomposed M·Cl¯ to travel from Shutter 2 to the detector plate and tm is the drift time for X¯ formed at or before the shutter to travel the same distance, then the presence of ion intensity at time t between tm and td denotes M·Cl‾ decomposition during its travel between ion shutter and detector plate. The actual time tx at which decomposition occurred is given in terms of tm, td and t by Eq. 7 in which the term in brackets represents the fraction of the time t spent as dimer.14 t − tm tx = td td − tm

(7)

The value of tx varies from 0 when t = tm to td when t = td and the number of non-dissociated M·Cl¯ present at time tx is proportional to the numerically integrated ion intensity from t to the end of the mobility spectrum. The first order rate constant k for the thermal decomposition of M·Cl¯, Eq. 8, is obtained from its integrated form, Eq. 9. A plot of the natural logarithm of


Page 13 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


remaining M·Cl¯ ion intensity versus tx gives a linear relationship with a negative slope equal to the rate constant k for the thermal decomposition of M·Cl¯. −

dMCl‾ = kMCl‾ (8) dt

ln MCl‾ = ln MCl‾% − kt (9) The measurement of k over a range of temperatures allows the determination of the activation energy Ea from a plot of ln k versus 1/T(K) (Eq. 10). In this Arrhenius equation A (s-1) is the preexponential factor, T(K) is the drift gas temperature and R is the gas constant (J K-1 mol-1). lnk = lnA −

E( (10) RT

Arrhenius plots for the thermal decomposition of the chloride adducts of 1,2-NG·Cl¯,1,3NG·Cl¯, DMNB·Cl¯, 3,4-DNT·Cl¯ together with that reported previously for NG·Cl¯ 16 are shown in Figure 3. The slope and intercept of each regression line was used to obtain, respectively, Ea and A for each chloride adduct decomposition. Each point on the figure represents the average of three separate determinations and each point was obtained on a different day following overnight temperature stabilization of the instrument. With the assumption that Ea is the change in internal energy from reactants to products, the standard enthalpy change for each dissociation reaction is ∆Ho = Ea + RT. Table 2 summarizes the decomposition reactions identified, the temperature range over which decomposition was measureable, the experimental Ea and A values and the derived enthalpy changes for the reactions. The temperature for calculation of each enthalpy value is that at the mid-point of the respective temperature range. Included in the table is the result previously reported for NG·Cl‾. The temperature range over which dissociation was observable for any of the chloride adducts was very limited, from 10 oC for 3,4-DNT to 16 oC for 1,2-NG·Cl‾, due to the usual



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 36

fixed length of the drift region. The absolute temperature ranges for the glycerin nitrates increase in the same order of increasing stability determined from temperature studies in the GCIMS-MS experiments, NG·Cl‾ < 1,3-NG·Cl‾ < 1,2-NG·Cl‾ < 1-NG·Cl‾. Indeed 1-NG·Cl‾, which was found to be very stable in the GC-IMS-MS experiments, showed no tendency to decompose even at 196 oC,, the upper temperature limit of the mobility spectrometer. Unfortunately, the isolated PETN·Cl‾ signal was so weak that it was impossible to characterize its mode of decomposition. All the chloride adducts of the nitrate ester explosives decompose thermally by a reaction in which Cl‾ displaced NO3‾ in the gas-phase SN2 reaction of Eq. 3 that was previously shown to occur for both the chloride adduct of ethylene glycol dinitrate and NG·Cl‾.16,17 The path from M + Cl‾ to the products, the nitrate ion and a molecule in which Cl has replaced NO3 has two potential wells, due to two ion-molecule complexes, separated by a transition state (Eq. 11) and Figure 4. M + Cl‾ → M·Cl‾ → TS → [M-NO3+Cl]·NO3‾ → [M−NO3+Cl] + NO3‾

(11)

The nitro-containing compound adducts, DMNB·Cl‾ and 3,4-DNT·Cl‾ decomposed by loss of Cl‾. The temperature range for observation of the decomposition of DMNB·Cl‾ was the lowest of all the chloride adducts. This low temperature range is consistent with two previous IMS experiments that found the adduct was not observable above 60 oC and above 80 oC.5,29 There was no evidence for the formation of [M-H]‾ from 3,4-DNT·Cl‾ that was observed in the GC-IMS-MS experiments. The only peak in the TNT GC-IMS spectrum was thermally stable and was identified in the GC-IMS-MS experiments as [TNT-H]‾ , presumably formed by loss of HCl from the hyperenergetic TNT-Cl¯ encounter complex (Eq. 12). The adduct TNT·Cl‾ has been reported to exist


Page 15 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


with an intensity 3.2 % of that of [TNT-H]‾ in a corona discharge ion mobility spectrometer at 51 o

C and its relative intensity increased to 6 % at 90 oC.30 The increase with temperature was

ascribed to a high and unregulated concentration of TNT in the drift region due to TNT desorption from the tube walls. Despite a thorough search we found no evidence for the existence of TNT·Cl‾ in either the GC-IMS or GS-IMS-MS experiments. The low Arrhenius pre-exponential factors for the displacement reactions are not unexpected because there is a loss in entropy in forming the transition state when the carbon at which displacement occurs attains a pentacoordinate geometry.17 The pre-exponential factor for the dissociation of 3,4-DNT·Cl‾, 4.8 x 1013 s-1, is larger as expected for a simple dissociation, although not as large as the ca.1015 s-1 found for proton-bound dimers of dimethyl methylphosphonate, 2,4-dimethylpyridine and several ketones.14, 31 The low pre-exponential value of 4.8 x 1013 for 3,4-DNT·Cl‾ decomposition and the extremely low value of 9.7 x 1011 s-1 for DMNB·Cl‾ decomposition are unexpected for the simple departure of an atomic ion that involves the simple elongation of a hydrogen bond. C6 H5 CH3 + Cl− = C6 H5 CH3 · Cl−∗ → C6 H5 CH− 2 + HCl

(12)

Density functional theory (DFT) calculations Calculations of both structures and associated enthalpies were completed by DFT with the 6-311+G(d,p) basis set for the lowest conformers determined for each of the neutral molecules of 1,2-NG, 1,3-NG, 1-NG, DNMB, 3,4-DNT and their relevant adduct ions. Each structure was at a local minimum on the potential energy surface with no imaginary frequencies. The structures of the neutral molecules and their Cl‾ adducts are shown in Figure 5. Included for comparison are the previously reported structures for NG and NG·Cl¯.16



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 36

Table 3 summarizes the computed enthalpy changes for the observed displacement reactions of the glycerin nitrate adducts and the observed Cl‾ loss from 3,4-DNT·Cl‾ and DMNB·Cl‾. Also included for comparison are the enthalpies for the non-observed loss of Cl‾ from the M·Cl‾ ions of NO3‾ from the M·NO3‾ adducts that were observed by mass spectrometry. The previously reported information for the analogous reactions involving NG is also given.16 Table 3 clearly shows that the binding enthalpy of Cl‾ to each of the glycerin nitrate esters is always greater than that of NO3‾ but the difference is not large, 18, 9, 7 and 15 kJ mol-1 respectively for NG, 1,3-NG, 1,2-NG and 1-NG. In the IMS/MS spectra, however, the Cl‾ adduct disappears at a lower temperature than does the NO3‾ adduct because it can participate in the displacement reaction. A displacement reaction involving M·NO3‾ would involve NO3‾ displacing NO3‾, which is a thermoneutral reaction that does not produce an identifiable product. The enthalpy changes for the overall displacement reactions of Eq. 13 that may be computed from the data of Table 3 are essentially identical viz. 37, 37, 36 and 40 kJ mol-1 respectively for NG, 1,3-NG, 1,2-NG and 1-NG. Therefore the order of the reaction enthalpies for the M + Cl‾ displacements, NG < 1,3-NG < 1,2-NG are almost entirely due to the Cl‾ binding enthalpies that become more negative in the order -81 < -89 < -96 kJ mol-1. Scheme 1 shows a thermodynamic cycle for the displacement reaction in terms of bond dissociation energies (BDEs) and electron affinities (EAs). M + Cl‾

[M−NO +Cl] + NO ‾ 3

BDE(C−ONO2)

3

−BDE(C−Cl)

EA(Cl) - EA(NO ) 3

[M-NO ] + NO ‾ + Cl

[M-NO ] + NO + Cl‾ 3

3

3


3

Page 17 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Scheme 1 EA(Cl) - EA(NO3) = 3.6131 eV – 3.937 eV = -31.25 kJ mol-1 which can account for much of the -37 kJ mol-1 overall reaction enthalpy for the three glycerin nitrates. This implies that the driving force for the overall displacement reaction (Eq. 13) is the higher electron affinity of the departing NO3 that is 31 kJ mol-1 larger than that of Cl. M + Cl‾ = [M – NO3 + Cl] + NO3‾

(13)

The computed enthalpy changes for the displacement reaction, Eq. 13 are in excellent agreement with the experimental values as shown in Table 3, which confirms that the lowest energy process for the decomposition of NG·Cl‾, 1,3-NG·Cl‾ and 1,2-NG·Cl‾ is the displacement reaction Eq. 5. 1-NG·Cl‾ is anomalous in that it has a computed enthalpy change for the displacement reaction of 85 kJ mol-1 and the much higher 125 kJ mol-1 for loss of Cl‾ but the displacement reaction was not observable. The probable reason for this anomalous behavior is discussed below. The nitroglycerin molecules have lowest energy structures in which the NO3 groups as shown in Figure 4 are positioned as far apart as possible from each other within the rotational constraints of the carbon chain. In forming the M·Cl‾ adduct with each molecule there is a conformational change that presents three acidic hydrogens to form hydrogen bonds with Cl‾ in a tridentate manner. The hydrogens gain their acidity either by being on an NO3-bearing carbon or on the oxygen of the hydroxyl group. A hydrogen atom on the latter site has the greater acidity and, molecular constraints aside, will form the strongest hydrogen bond. Neutral NG has the terminal NO3 groups below and the central NO3 above the molecular plane. 1,3-NG has the H of the Hydroxyl group oriented to a terminal ester oxygen atom of the nitrate at a distance of 2.446 Å, a distance small enough to imply a hydrogen bond. Supporting this is the elongation of the



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

O−NO2 bond to 1.442 Å compared with the 1.429 Å length of the other terminal NO3. The hydroxyl H of 1,2NG does not participate in a hydrogen bond the two O−NO2 being essentially the same length, 1.431 Å and 1.428 Å and the O−H bond length is 0.962 Å compared with the 0.964 Å for 1,3-NG. The Cl in NG·Cl¯ is hydrogen bonded to two terminal Hs at distances of 2.469 Å and 2.383 Å while the H on C2 is at the greater distance of 3.362 Å . The two terminal C− H bonds increase in length by 0.010 Å in the presence of Cl¯ but the central C−H increases by only 0.001 Å, consistent with a much weaker hydrogen bond strength. The presence of Cl¯ in 1,3-NG·Cl¯ causes the re-orientation of the hydroxyl group so that it together with two terminal Hs form three hydrogen bonds. The OH····Cl distance is 2.103 Å and the O−H bond has lengthened by 0.031 Å from its 0.964 Å in 1,3NG and the two CH····Cl bonds are 2.748 Å and 2.662 Å in length. The associated C−H bonds increase by less than 0.001 Å. There are three hydrogens oriented toward Cl in 1,2-NG·Cl¯, one from each C with no participation from hydroxyl. However, the O−H bond slightly increases by 0.001 Å as it weakly hydrogen bonds with the ester O of an NO3 at a distance of 2.143 Å. In 1,3-NG·Cl‾ there are three hydrogen bonds to Cl but the one weak OH····ONO2 hydrogen bond in the neutral molecule is broken, leading to 1,3-NG having a lower Cl‾ affinity than1,2-NG whose OH is non-hydrogen bonded in the neutral molecule but is bonding in the adduct. NG has no hydroxyl to form a strong hydrogen bond with Cl‾ and as a consequence it has the weakest Cl‾ affinity. The above argument is consistent with the computed order of binding enthalpies for NG·Cl‾, 1,3-NG·Cl‾ and 1,2-NG·Cl‾ i.e. 118 kJ mol-1 < 126 kJ mol-1 < 132 kJ mol-1.


Page 18 of 36

Page 19 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The reaction in which Cl‾ displaces NO3‾, Equation 11, has a lower energy requirement than the competing reversion to M + Cl‾ for the dinitrate esters as was previously experimentally demonstrated for NG. The transition state for NG·Cl‾ was found to be at essentially the same enthalpy level as that of the final products. The transition state, between 1,2-NG·Cl‾ and [1,2-NG - NO3+Cl]·NO3‾ was determined at the 6-311+G(d,p) level by finding the maximum energy as the C····N distance for the departing NO3‾ was gradually increased. The structure has one imaginary frequency as required for a transition state and with the carbon atom showing pentavalent character for an SN2 displacement reaction. The enthalpy of the transition state is 89 kJ mol-1 above the level of 1,2-NG·Cl‾. This is slightly less than the 96 kJ mol-1 computed for the displacement reaction and confirms that the measured activation energy and resulting enthalpy change for the dissociation ar for the complete reaction of Eq. 11. A seen in Table 3, the computed enthalpy changes for the displacement reactions are in excellent agreement with the order of and with the absolute values of the experimental enthalpies. The exception to this statement is 1-NG·Cl¯, calculated to have a displacement reaction enthalpy of 85 kJ mol-1 that is next lowest of the four glycerin nitrates but found to be non-dissociative at a temperature of 196 o

C at which all the other Cl‾ adducts are not viable. This temperature cannot be high enough for

either the occurrence of either the displacement reaction or the simple dissociation to 1-NG + Cl¯. 1-NG has only two hydrogens, both on C1, whose acidity can be affected by the presence of the NO3 group. The C2 hydroxyl is oriented towards the oxygen of the C3 hydroxyl giving an O −H····O angle of 109o with the O−H bond length 0.005 Å greater than that of the terminal hydroxyl. 1-NG·Cl¯ has three hydrogen bonds to Cl, one with a C1 hydrogen and the other two from the hydroxyl groups. The C1 hydrogen is 3.076 Å distant from Cl and can therefore be only



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

weakly hydrogen bonding with Cl. The two more acidic hydroxyl hydrogens are almost equidistant from Cl, the one on C2 at 2.103 Å and the one on C3 at 2.127 Å. The H—Cl—H angle is 69.4o and the OHHO dihedral angle is 5.7o so there is almost planar geometry in the hydrogen bonding and the chloride adduct is showing diol rather than nitrate character. For comparison, the computed H—Cl—H angle in ethanediol·Cl‾ was reported to be 69o and the chloride binding energy was 106 kJ mol-1, determined by the pulsed electron beam high-pressure mass spectrometry equilibrium method.32 This compares with the computed value for 1-NG·Cl‾ of 125 kJ mol-1. 1-NG is behaving differently towards Cl‾ than the other glycerin nitrates in that the NO3 group has little influence on the binding enthalpy or on the geometry of the biding site. The enthalpy of the transition state between NG·Cl‾ and [1-NG – NO3 + Cl]·NO3‾, obtained in the same manner as for that of 1,2-DNT·Cl‾, is 112 kJ mol-1, which unlike for the other glycerin nitrates, is considerably higher than that of the final displacement reaction products. Although this reaction is as enthalpically favored over loss of Cl‾ as for the other glycerin nitrate-chloride adducts the higher activation energy requires a higher temperature for its observation than is attainable in the mobility spectrometer. The formation of NO3‾ in the decomposition of the chloride adducts of the nitroglycerines explains why M·NO3‾ is often a major product in the analysis of nitrate esters when chloride is the reactant ion. 3 If the nitrate ester is at a sufficiently high concentration, the NO3‾ released in the displacement reaction associates with the ester to form M·NO3‾ that, as Table 3 illustrates, always has a higher enthalpy for dissociation than M·Cl‾. The computed structure of DMNB·Cl¯ has Cl‾ hydrogen bonded to the two terminal methyl groups at distances of 2.444 Å and 2.477 Å with the two nitro groups conferring acidity on the hydrogens situated, as usual, on the opposite side of the molecule to chloride. The two


Page 20 of 36

Page 21 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


next-nearest hydrogens are on the other two methyls at distances of 3.540 Å and 3.376 Å. The NCCN dihedral angle is 162o compared with 40o in the neutral molecule. The computed value for the dissociation of DMNB·Cl¯ of 79 kJ mol-1 is higher than the experimental 70 kJ mol-1 and as Table 3 shows, more towards the value reported previously that was obtained by an equilibrium method. The Cl¯ in 3,4-DNT·Cl‾ is hydrogen bonded to one H of the methyl group at a distance of 2.613 Å and the C6 hydrogen at 2.397 Å. The greater acidity of the ring hydrogen can be attributed to the resonance effect of the nitro group on C3 being stronger than the resonance effect of the C4 nitro group on the methyl hydrogen. In forming these hydrogen bonds the methyl C−H has lengthened by 0.011 Å and the ring C−H by 0.010 Å. As is common to all the chloride adducts the C−H bonds of the neutral molecule lengthens when hydrogen bonding occurs. The methyl C−H bond increases by 0.011 Å while the ring C−H increases by slightly less, 0.009 Å. The calculated enthalpy for loss of Cl¯ from 3,4-DNT·Cl‾ of 91 kJ mol-1 is in reasonable agreement with the experimental value of 99 kJ mol-1 (Table 3). For comparison, the experimental binding energy of Cl¯ with nitrobenzene is 69 kJ mol-1, with Br¯ it is as expected for the larger ion slightly less, 63 kJ mol-1, and for Br¯ with 1,2-dinitrobenzene is 87.9 ± 7.5 kJ mol-1.33 No value for the enthalpy of association of Cl‾ with 1,2-dinitrobenzene is available but a value around 95 kJ mol-1 would be expected, consistent with the value of 91kJ mol-1 for 3,4DNT·Cl‾. The longer hydrogen bond length from CH3 compared with that from the ring shows that the main binding strength comes from resonance.

CONCLUSIONS Chloride adducts of glycerin nitrate esters are readily formed in ion mobility spectrometers with chloride as reactant ion. The observable thermal decomposition of the 21 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

adducts occurs over a narrow temperature range that is unique to each ester, the fully nitrated NG ester over the lowest range and the mono nitrate at a range above 196 oC that was not attainable experimentally. The order of thermal stability of nitrate explosives increase in the reverse order of the number and position of the nitrate groups viz. NG·Cl‾ < 1,3-NG·Cl‾ < 1,2-NG·Cl‾ < 1NG·Cl‾, where the dissociation of 1-NG·Cl‾ was unable to be determined due to its stability in the drift tube operating temperature range. The Cl‾ is multiply hydrogen bonded to the ester through either or both H attached to the carbon chain and hydroxyl with the latter giving the shorter, stronger bond. The sole product of the decompositions is NO3‾ and calculation shows that the reaction is the displacement of NO3‾ by Cl‾ in a gas-phase SN2 reaction with a transition state that is approximately at the same enthalpy level as the final products. The enthalpy requirement for the displacement reaction of the mononitrate is lower than that for the dinitrates but the transition state enthalpy level is 27 kJ mol-1 higher than the final products, consistent with the required, but experimentally unattainable, higher reaction temperature. Loss of Cl‾ is the only thermal dissociation reaction observed for the Cl‾ adducts of the two nitro compounds DMNB and 3,4-DNT. Loss of HCl to give [3,4-DNT-H]‾ was observed in the GC-IMS-MS experiments but not in the thermal decomposition experiments. The reason for the discrepancy may be that competitive thermal HCl formation is insignificant compared to loss of Cl‾ or that [3,4-DNT-H]‾ was formed by collisional activated dissociation in the mass spectrometer interface region. In either case, the dominant thermal reaction is Cl‾ loss. A driving force in the SN2 displacement reaction of Cl for NO3 with the glycerol nitrates is the significant difference in the electron affinities, 3.61 eV and 3.94 eV respectively. This difference of 32 kJ mol-1 is almost the same as the calculated, essentially constant ~37 kJ mol-1 for the overall reaction (Eq. 13). The absence of a displacement reaction with the nitro22 ACS Paragon Plus Environment

Page 22 of 36

Page 23 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


containing DMNB·Cl‾, in which Cl‾ is hydrogen-bonded in a similar manner to NG·Cl¯ is mainly attributable to the prohibitively low electron affinity of NO2, 2.37 eV.

AUTHOR INFORMATION *(G.A.E.) E-mail: [email protected]. Fax: 575-646-6094. Tel: 575-646-2146. Notes The authors declare no competing financial interest. “This material is based upon work supported by the U.S. Department of Homeland Security, Science and Technology Directorate, Office of University Programs, under Grant Award 2013ST-061-ED0001. The views and conclusions contained in this document are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of the U.S. Department of Homeland Security.”

ACKNOWLEDGMENTS Financial support from the U.S. Department of Homeland Security, ALERT - Awareness and Localization of Explosives-Related Threats, through the Northeastern University is gratefully acknowledged.

REFERENCES (1) Eiceman, G.A.; Karpas, Z.; Hill, H. H. Ion Mobility Spectrometry, 3rd Ed.; CRC Press: Boca Raton, FL, 2013 (2) Makinen, M.; Nousiainen, M.; Sillanpaa, M. Ion Spectrometric Detection Technologies for Ultra-Traces of Explosives: A Review. Mass Spectrom. Rev. 2011, 30, 940-973. (3) Ewing, R.G.; Atkinson, D.A.; Eiceman, G.A.; Ewing, G.J. A Critical Review of Ion Mobility Spectrometry for the Detection of Explosives and Explosive Related Compounds.Talanta 2001, 54, 515-529. 23 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(4) Creaser, C.S.; Griffiths, J.R.; Bramwell, C.J.; Noreen, S.; Hill, C.A.; Thomas, C.L.P. Mobility Spectrometry: A Review. Part 1. Structural Analysis by Mobility Measurements. Analyst 2004, 129, 984-994. (5) Lawrence, A.H.; Neudorfl, P.; Stone, J.A. The Formation of Chloride Adducts in the Detection of Dinitrocompounds by Ion Mobility Spectrometry. Int. J. Mass Spec. 2001, 209, 185-195. (6) Proctor, C.J.; Todd, J.F.J. Alternative Reagent Ions for Plasma Chromatography. Anal. Chem. 1984, 56, 1794-1797. (7) Matz, L.M.; Tornatore, P.S.; Hill, H.H. Evaluation of Suspected Interferents for TNT Detection by Ion Mobility Spectrometry.Talanta 2001, 54, 171-179. (8) Daum, K.A.; Atkinson, D.A.; Ewing, R.G. Formation of Halide Reactant Ions and Effects of Excess Reagent Chemical on the Ionization of TNT in Ion Mobility Spectrometry. Talanta 2001, 55, 491-500. (9) Mathis, J.A.; McCord, B.R. The Analysis of High Explosives by Liquid Chromatography/Electrospray Ionization Mass Spectrometry: Multiplexed Detection of Negative Ion Adducts. Rapid Commun. Mass Spectrom. 2005, 19, 99-104. (10) Daum, K.A.; Atkinson, D.A.; Ewing, R.G.; Knighton, W.B.; Grimsrud, E.P. Resolving Interferences in Negative Mode Ion Mobility Spectrometry Using Selective Reactant Ion Chemistry.Talanta 2001, 54, 299-306. (11) Vigneau, O.; Mandard, X.M. A LCMS Method Allowing the Analysis of HMX and RDX Present at the Picogram Level in Natural Aqueous Samples Without a Concentration Step.Talanta 2009, 77, 1609-1613. (12) Jazan, E.; Tabrizchi, M. Kinetic Study of Proton-Bound Dimer Formation Using Ion Mobility Spectrometry. Chem. Phys. 2009, 355, 37-42. (13) Valadbeigi, Y.; Farrokhpour, H.; Tabrizchi, M. Effect of Hydration on the Kinetics of Proton-Bound Dimer Formation: Experimental and Theoretical Study. J. Phys. Chem. A 2014, 118, 7663-7671. (14) Ewing, R.G.; Eiceman, G.A.; Harden, C.S.; Stone, J.A. The Kinetics of the Decompositions of the Proton Bound Dimers of 1,4-Dimethylpyridine and Dimethyl Methylphosphonate from Atmospheric Pressure Ion Mobility Spectra. Int. J. Mass Spectrom. 2006, 255-256, 76-85. (15) An, X.; Eiceman, G.A.; Rasanen, R.M.; Rodriguez, J.E.; Stone, J.A. Dissociation of Proton Bound Ketone Dimers in Asymmetric Electric Fields with Differential Mobility Spectrometry and in Uniform Electric Fields with Linear Ion Mobility Spectrometry. J. Phys. Chem. A 2013, 117, 6389-6401. 24 ACS Paragon Plus Environment

Page 24 of 36

Page 25 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(16) Rajapakse, M.Y.; Stone, J. A.; Eiceman, G.A. Decomposition Kinetics of Nitroglycerine·Cl¯(g) in Air at Ambient Pressure with a Tandem Ion Mobility Spectrometer. J. Phys. Chem, A 2014, 118, 2683-2692. (17) Rajapakse, R.M.M.Y; Stone, J.A.; Eiceman, G.A. An Ion Mobility and Theoretical Study of the Thermal Decomposition of the Adduct Formed Between Ethylene Glycol Dinitrate and Chloride. Int. J. Mass Spectrom. 2014, 371, 28-35. (18) An, X. Kinetics of Decomposition of Gas-Phase Ions of Esters and Ketones in Air at Ambient Pressure Using Differential Mobility Spectrometry and Ion Mobility Spectrometry. Ph.D. Dissertation, New Mexico State University, Las Cruces, NM, January 2010. (19) Viehland, L.A.; Mason, E.A. Gaseous Ion Mobility in Electric Fields of Arbitrary Strength. Ann. Phys. 1978, 110, 287-328. (20) Khayamian, T.; Tabrizchi, M.; Jafari, M.T. Analysis of 2,4,6-Trinitrotoluene, Pentaerythritol Tetranitrate and Cyclo-1,3,5-Trimethylene-2,4,6-Trinitramine Using Negative Corona Discharge Ion Mobility Spectrometry. Talanta 2003, 59, 327-33. (21) Miyayama, T.; Tsou, P.S.; Fung, S.M.; Fung, H.L. Simultaneous Determination of Nitroglycerin and Dinitrate Metabolites in Metabolism Studies Using Liquid Chromatography Mass Spectrometry with Electrospray Ionization. J. Chromatogr. B 2006, 835, 21-26. (22) Aburawi, S.; Curry, S.H.; Whelpton, R. Chemical Denitration of Nitroglycerin, and Conversion of 1,2-dinitroglycerin to 1,3-dinitroglycerin. Int. J. Pharm. 1984, 22, 327-336. (23) Akrill, P.; Guiver, R.; Cocker, J. Biological Monitoring of Nitroglycerin Exposure by Urine Analysis. Toxicol. Lett. 2002, 134, 271-276. (24) Martel, R.; Godin, B.; Levesque, R.; Cote, S. Determination of Nitroglycerin and Its Degradation Products by Solid-Phase Extraction and LCUV. Chromatographia. 2010, 71, 285289. (25) DeTata, D.; Collins, P.; McKinley, A. A Fast Liquid Chromatography Quadrupole Time-ofFlight Mass Spectrometry (LC-QToF-MS) Method for the Identification of Organic Explosives and Propellants. Forensic Sci. Int. 2013, 233, 63-74. (26) Husserl, J.; Hughes, J.B. Biodegradation of Nitroglycerin in Porous Media and Potential for Bioaugmentation with Arthrobacter Sp. Strain JBH1. Chemosphere 2013, 92, 721-724. (27) Eiceman, G.A.; Preston, D.; Tiano, G.; Rodriguez, J.E.; Parmeter, J.E. Quantitative Calibration of Vapor Levels of TNT, RDX, and PETN Using a Diffusion Generator with Gravimetry and Ion Mobility Spectrometry. Talanta 1997, 45, 57-74. (28) Perry, R.H.; Green, D.W. (Eds). Perry’s Chemical Engineers’ Handbook, 8th Ed. McGraw25 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Hill, New York, NY, 2008. (29) Munro, W.A.; Thomas, C.L.P.; Langford, M.L. Characterisation of the Ion Mobility Spectrometric Behaviour of a Dinitrobutane Under Varying Conditions of Temperature and Concentration. Analytica Chimica Acta 1998, 375, 49-63. (30) Sabo, M.; Michalczuk, B.; Lichvanová, Z.; Kavicky, V.; Radjenovic, B.; Matejcík, S. Interactions of Multiple Reactant Ions with 2,4,6-Trinitrotoluene Studied by Corona Discharge Ion Mobility-Mass Spectrometry.Int. J. Mass Spectrom. 2015, 380, 12-20. (31) An, X.; Eiceman, G.A.; Stone, J.A. A Determination of the Effective Temperatures for the Dissociation of the Proton Bound Dimer of Dimethyl Methylphosphonate in a Planar Differential Mobility Spectrometer. Int. J. Ion Mobiityl. Spectrom. 2010, 13, 25-36. (32) Zhang, W.; Beglinger, C.; Stone, J.A. High-Pressure Mass Spectrometric Study of the GasPhase Association of Cl‾ with α-ω Diols. J. Phys. Chem. 1995, 99, 11673-11679. (33) NIST Chemistry Webbook//webbook.nist.gov/chemistry


Page 26 of 36

Page 27 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Captions to figures Figure 1. (a) IMS spectrum of DMNB·Cl- using Shutter 1. (b) IMS spectrum of isolated DMNB·Cl¯ using dual shutter operation. Drift gas temperature 80oC and 250 V cm-1 field strength. Figure 2. (a) IMS spectrum of 1,3-NG.Cl¯ using Shutter 1. (b) IMS spectrum of isolated 1,3NG.Cl ¯ using dual shutter operation. Drift gas temperature 155oC and 250 V cm-1 field strength. Figure 3. Arrhenius plots for the thermal decomposition of (1) 1,2-NG·Cl¯, (2) 1,3-NG·Cl¯, (3) NG·Cl¯ (from Ref. 16), (4) 3,4-DNT·Cl¯, (5) DMNB·Cl¯. Figure 4. Representative energy levels for the chloride displacement reaction of Eq. 11. Figure 5. Structures of neutral molecules and their Cl¯ adducts.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 1. (a) IMS spectrum of DMNB·Cl- using Shutter 1. (b) IMS spectrum of isolated DMNB·Cl¯ using dual shutter operation. Drift gas temperature 80 oC and 250 V cm-1 field strength.


Page 28 of 36

Page 29 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2. (a) IMS spectrum of 1,3-NG.Cl¯ using Shutter 1. (b) IMS spectrum of isolated 1,3-NG.Cl ¯ using dual shutter operation. Drift gas temperature 155 oC and 250 V cm-1 field strength.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 3. Arrhenius plots for the thermal decomposition of (1) 1,2-NG·Cl¯, (2) 1,3NG·Cl¯, (3) NG·Cl¯ (from Ref. 16), (4) 3,4-DNT·Cl¯, (5) DMNB·Cl¯.


Page 30 of 36

Page 31 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4. Representative energy levels for the chloride displacement reaction of Eq. 11.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 5. Structures of neutral molecules and their Cl¯ adducts. 32 ACS Paragon Plus Environment

Page 32 of 36

Page 33 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 1. Reduced mobilities of ions (cm2 V-1 s-1) Ion Ko Ko Literature Cl¯

2.80 (123 oC)

2.93a (125 oC)

2.91 (144 oC) NO3¯

2.43

2.46b

1,3-NG·Cl¯

1.54

NA

1,3-NG·NO3¯

1.46

NA

1,2-NG·Cl¯

1.55

NA

1,2-NG·NO3¯

1.47

NA

DMNB· Cl¯

1.51

1.48-1.54c 1.46-1.57a

3,4-DNT·Cl¯

1.52

NA

[TNT-H]¯

1.57

1.55b

NA = not available. a value at 125 oC, reference 5; b value at 144 oC, reference 16; c reference 29



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 36

Table 2. Experimental results for chloride adduct dissociations. T range (oC), Arrhenius parameters Ea (kJ mol-1), A (s-1), and derived reaction enthalpies (kJ mol-1) Observed reaction

T range

Ea

A

∆Ho

NG.Cl¯ [NG − NO3 + Cl] + NO3¯ (Data from Ref. 16) 111-122 80±3 3.9 x 1012 83±3 1,3-NG·Cl¯ [1,3-NG − NO3 + Cl] + NO3¯

144-158 86±2 2.2 x 1012 88±2

1,2-NG·Cl¯ [1,2-NG − NO3 + Cl] + NO3¯

180-196 97±1 3.5 x 1012 99±2

DMNB·Cl¯ DMNB + Cl¯

69-84

68±2 9.7 x 1011 70±1

3,4-DNT·Cl¯ 3,4-DNT + Cl¯

80-90

81±2 4.8 x 1013 83±2


Page 35 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 3. Computed enthalpy changes for adduct ion decompositions Reaction NG·Cl‾ = NG + Cl‾ 1,3-NG·Cl‾ = 1,3-NG + Cl‾ 1,2-NG·Cl‾ = 1,2-NG + Cl‾ 1-NG·Cl‾ = 1-NG + Cl‾ NG·NO3‾ = NG + NO3‾ 1,3-NG·NO3‾ = 1,3-NG + NO3‾ 1,2-NG·NO3‾ = 1,2-NG + NO3‾ 1-NG·NO3‾ = 1-NG + NO3‾ NG·Cl‾ = (NG – NO3‾ + Cl) + NO3‾ 1,3-NG·Cl‾ = (1,3-NG – NO3‾ + Cl) + NO3‾ 1,2-NG·Cl‾ = (1,2-NG – NO3‾ + Cl) + NO3‾ 1-NG·Cl‾ = (1-NG – NO3‾ + Cl) + NO3‾ DMNB·Cl‾ = DMNB + Cl‾ 3,4-DNT·Cl‾ = 3,4-DNT + Cl‾

∆H1% (kJ mol-1) 118 126 132 125 100 117 125 110 81 (83)a 89 (88)b 96 (99)b 85 79 (70b,92c) 91 (99)b

Experimental results: a Reference 16, b Experimental, this work, c Reference 5.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

TOC Graphic


Page 36 of 36

Dissociation Enthalpies of Chloride Adducts of Nitrate and Nitrite

Recommend Documents