Anal. Chem. 2010, 82, 3764–3771
Ion Chemistry of VX Surrogates and Ion Energetics Properties of VX: New Suggestions for VX Chemical Ionization Mass Spectrometry Detection Anthony J. Midey,†,§,⊥ Thomas M. Miller,†,§ A. A. Viggiano,*,† Narayan C. Bera,‡ Satoshi Maeda,‡ and Keiji Morokuma‡ Air Force Research Laboratory, Space Vehicles Directorate, 29 Randolph Rd., Hanscom AFB, Massachusetts 01731-3010, and Cherry L. Emerson Center for Scientific Computation and Department of Chemistry, Emory University, Atlanta, Georgia Room temperature rate constants and product ion branching ratios have been measured for the reactions of numerous positive and negative ions with VX chemical warfare agent surrogates representing the amine (triethylamine) and organophosphonate (diethyl methythiomethylphosphonate (DEMTMP)) portions of VX. The measurements have been supplemented by theoretical calculations of the proton affinity, fluoride affinity, and ionization potential of VX and the simulants. The results show that many proton transfer reactions are rapid and that the proton affinity of VX is near the top of the scale. Many proton transfer agents should detect VX selectively and sensitively in chemical ionization mass spectrometers. Charge transfer with NO+ should also be sensitive and selective since the ionization potential of VX is small. The surrogate studies confirm these trends. Limits of detection for commercial and research grade CIMS instruments are estimated at 80 pptv and 5 ppqv, respectively. Trace detection of chemical warfare agents (CWA) is critical for applications in both homeland security and battlefield operations. The analytical requirements are rigorous, given that harm to humans is caused by minute quantities. According to a 2005 National Academy of Sciences (NAS) report1 on monitoring the air quality at CWA disposal facilities, real-time detection is needed in the plants where the live agents are routinely handled. Of the currently available technologies, the panel concluded that the only current method that demonstrated enough promise to pursue in the short term is chemical ionization mass spectrometry (CIMS). CIMS has already proven successful at selective and sensitive detection of trace concentrations of many atmospheric neutrals.2-9 Using the CIMS technique (as well as the related ion mobility spectrometer technique (IMS)) for CWA detection requires a * Corresponding author. E-mail:
[email protected]. † Air Force Research Laboratory. ‡ Emory University. § Under contract to the Institute for Scientific Research, Chestnut Hill, MA. ⊥ Current address: Excellims Corp., Acton, MA. (1) Kolb, C. E.; Steinfeld, J. I.; Drake, E. M.; Drury, C. G.; Gibson, J. r.; Koller, L. D.; Klugh, J. R.; Sides, G. D.; Viggiano, A. A.; Walt, D. R. In Monitoring at Chemical Agent Disposal Facilities, NAS Report, NRC; National Academy of Sciences: Washington, DC, 2005.
3764
Analytical Chemistry, Vol. 82, No. 9, May 1, 2010
search for readily generated primary ions that react rapidly with the agents, while yielding product ions with unique mass signatures (mobilities for IMS). The latter goal is typically achieved by employing ion-molecule reactions that keep the neutral reagent structure intact (or nearly so) after ionization.10,11 Because it is highly sensitive, pretreatment of the gas sample before analysis with CIMS detection of CWA should not be necessary. This allows real time detection. Commercial examples of this method, such as proton transfer reaction-mass spectrometry (PTR-MS)12-14 and selected ion flow tube-mass spectrometry (SIFT-MS),15-18 have been shown to have limits of detection (LOD) approaching tens of pptv.19 Custom built instruments for atmospheric trace gas detection have reached limits of quantitation (LOQ) of 5 ppqv.3,20 (2) Ballenthin, J. O.; Thorn, W. F.; Miller, T. M.; Viggiano, A. A.; Hunton, D. E.; Koike, M.; Kondo, Y.; Takegawa, N.; Irie, H.; Ikeda, H. J. Geophys. Res. 2003, 108, ACH7. (3) Chen, G.; Huey, L. G.; Trainer, M.; Nicks, D.; Corbett, J.; Ryerson, T.; Parrish, D.; Neuman, J. A.; Nowak, J.; Tanner, D. J.; Holloway, J.; Brock, C.; Crawford, J.; Olson, J. R.; Sullivan, A.; Weber, R.; Schauffler, S.; Donnelly, S.; Atlas, E.; Roberts, J.; Flocke, F.; Hu ¨ bler, G.; Fehsenfeld, F. J. Geophys. Res. 2005, 110, D10S90. (4) Eisele, F. L.; Tanner, D. J. J. Geophys. Res. 1993, 98, 9001. (5) Hunton, D. E.; Ballenthin, J. O.; Borghetti, J. F.; Federico, G. S.; Miller, T. M.; Thorn, W. F.; Viggiano, A. A.; Anderson, B. E.; Cofer, W. R.; McDougal, D. S.; Wey, C. C. J. Geophys. Res. 2000, 105, 26841. (6) Marcy, T. P.; Gao, R. S.; Northway, M. J.; Popp, P. J.; Stark, H.; Fahey, D. W. Int. J. Mass Spectrom. 2005, 243, 63. (7) Mauldin, R. L., III; Tanner, D. J.; Eisele, F. L. J. Geophys. Res. 1998, 103, 3361. (8) Schneider, J.; Burger, V.; Arnold, F. J. Geophys. Res. 1997, 102, 25. (9) Tremmel, H. G.; Schlager, H.; Konopka, P.; Schulte, P.; Arnold, F.; Klemm, M.; Droste-Franke, B. J. Geophys. Res. 1998, 103, 10803. (10) Midey, A. J.; Miller, T. M.; Viggiano, A. A. J. Phys. Chem. A 2008, 112, 10250. (11) Midey, A. J.; Miller, T. M.; Viggiano, A. A. J. Phys. Chem. A 2009, 113, 4982–4989. (12) Petersson, F.; Sulzer, P.; Mayhew, C. A.; Watts, P.; Jordan, A.; Mark, L.; Mark, T. D. Rapid Commun. Mass Spectrom. 2009, 23, 3875–3880. (13) Mayhew, C. A.; Sulzer, P.; Petersson, F.; Haidacher, S.; Jordan, A.; Mark, L.; Watts, P.; Mark, T. D. Int. J. Mass Spectrom. 2010, 289, 58–63. (14) Cordell, R. L.; Willis, K. A.; Wyche, K. P.; Blake, R. S.; Ellis, A. M.; Monks, P. S. Anal. Chem. 2007, 79, 8359. (15) Francis, G. J.; Milligan, D. B.; McEwan, M. J. Anal. Chem. 2009, 81, 8892– 8899. (16) Milligan, D. B.; Francis, G. J.; Prince, B. J.; McEwan, M. J. Anal. Chem. 2007, 79, 2537–2540. (17) Enderby, B.; Lenney, W.; Brady, M.; Emmett, C.; Spanel, P.; Smith, D. J. Breath Res. 2009, 3, 036001. (18) Smith, D.; Pysanenko, A.; Spanel, P. Rapid Commun. Mass Spectrom. 2009, 23. 10.1021/ac100176r 2010 American Chemical Society Published on Web 04/12/2010
Many current CWA field monitoring methods such as the Chemical Agent Monitor (CAM) and RAID systems use ion mobility spectrometry to detect the live agents after ionization based on their signature drift time through a buffer gas under the influence of the IMS drift field. These instruments typically use atmospheric pressure ionization (API) through reactions of the sampled gas with air ions generated with a 63Ni beta emitter.21 Knowing the kinetics, particularly the product ion distributions, provides an understanding of the ionization chemistry prior to analysis. This knowledge of the ion chemistry that is occurring within the system can guide the design of improved trace detection systems for applications such as CIMS, IMS, and API mass spectrometry (APIMS). As CWA’s cannot be readily studied in most laboratories, the use of surrogates and theoretical calculations are required to screen for the best reagent ion. This approach has recently been employed in our laboratory to find positive and negative ion CIMS reaction schemes for mustard (HD), and sarin (GA), and soman (GD) using the surrogates, 2-chloroethyl ethyl sulfide (2-CEES)10 and dimethyl methylphosphonate (DMMP),11 respectively, in a research-grade selected ion flow tube-mass spectrometer (SIFTMS). Francis et al. have similarly used a commercial SIFT-MS with a similar combination of experiment and theory to detect these simulants and several others with good sensitivity.15 A version of that instrument has been partially tested with some of the live agents.22 In addition, Cordell et al. have used a modified proton transfer reaction-mass spectrometer (PTR-MS) to make similar studies for blister and GX surrogates with some live agent testing for mustard (HD) and sarin (GB).14 The work presented here expands upon this approach to find new ion-molecule reaction schemes that can be used to detect VX, a schematic of whose structure is given in Figure 1a. VX has both alkylamine and organophosphorus functionalities as seen in Figure 1a. Thus, two separate surrogates have been studied individually to determine the reactivity at the amine and oxyphosphorus ends of the molecule. Triethylamine, (C2H5)3N, has been studied to represent the amine chemistry and diethyl methylthiomethylphosphonate (DEMTMP), (C2H5)2P(O)CH2SCH3, has been studied to represent the organophosphorus component. Structures of the simulants are shown in Figure 1b,c, respectively. Experimental rate constants and product ion branching ratios have been measured at the Air Force Research Laboratory (AFRL) using a SIFT at 298 K for a variety of ions reacting with the VX surrogates triethylamine and DEMTMP. The reactions of various protonated neutrals as well as the typical SIFT-MS cations H3O+, NO+, and O2+ with triethylamine are available in the literature.23-28 (19) Jordan, A.; Haidacher, S.; Hanel, G.; Hartungen, E.; Mark, L.; Seehauser, H.; Schottkowsky, R.; Sulzer, P.; Mark, T. D. Int. J. Mass Spectrom. 2009, 286, 122. (20) Mauldin, R. L., III; Frost, G. J.; Chen, G.; Tanner, D. J.; Prevot, A. S. H.; Davis, D. D.; Eisele, F. L. J. Geophys. Res. 1998, 103, 16713–16729. (21) Eiceman, G. A.; Karpas, Z. Ion Mobility Spectrometry, 2nd ed.; Taylor and Francis: London, 2005. (22) Francis, G. J.; Langford, V. S.; Milligan, D. B.; McEwan, M. J. Real time detection and quantification of dangerous substances in air without sample preparation using SIFT-MS; 56th ASMS Conference on Mass Spectrometry and Allied Topics, Denver, 2008. (23) Spanel, P.; Smith, D. Int. J. Mass Spectrom. 1998, 176, 203–211. (24) Feng, W. Y.; Lifshitz, C. Int. J. Mass Spectrom. Ion Proc. 1996, 152, 157– 168.
Figure 1. Schematic drawings of the structures of (a) nerve agent VX (b) triethylamine, and (c) diethyl methylthiomethylphosphonate (DEMTMP).
Therefore, only negative ion chemistry has been studied for this surrogate. In addition, theoretical calculations of the proton affinities (PA), ionization potentials (IP), and fluoride affinities (FA) of DEMTMP and VX have been performed to augment the studies involving the surrogate. The calculations also provide an understanding of the observed product ion distributions. The combination of theory and experiment allows us to predict which of the ions that sensitively and selectively detect the surrogates can also be employed to detect VX. The limit-of-detection (LOD) and limit-of-quantitation (LOQ) for a SIFT-type reactor are projected based on the kinetics measured, in the manner of Francis et al.15 for their commercial SIFT-MS. Similar limits are also estimated for the best custom built instrument. The viability of the combined approach for mustard (HD)10 and GX agents11 has been confirmed by recent simulant, mustard (HD), and sarin (GB) data.14,22 EXPERIMENTAL SECTION Kinetics Measurements. The experiments were performed at room temperature using the selected ion flow tube (SIFT) at AFRL shown in Figure 2. The instrument was described in detail elsewhere.29 Therefore, only a brief description is provided with the details pertinent to the current measurements. Ions were generated in a medium pressure source (∼0.1-1 Torr) using electron impact ionization of an appropriate neutral or combination of neutral molecules. A reactant ion with the desired mass-to-charge ratio (m/z) was selected with a quadrupole mass filter and then injected into a stainless steel flow tube at 0.5 Torr using a Venturi inlet with a fast flow of helium buffer gas. Neutral reagents were injected into the flow tube through finger inlets. For the triethylamine studies, a 5% mixture of the amine in helium was used as the neutral reagent. Methylthiomethylphos(25) Feng, W. Y.; Lifshitz, C. Int. J. Mass Spectrom. Ion Proc. 1995, 149-150, 13–25. (26) Feng, W. Y.; Goldenberg, M.; Lifshitz, C. J. Am. Soc. Mass Spectrom. 1994, 5, 695–703. (27) Feng, W. Y.; Ling, Y.; Lifshitz, C. J. Phys. Chem. 1996, 100, 35–39. (28) Feng, W. Y.; Iraqi, M.; Lifshitz, C. J. Phys. Chem. 1993, 97, 3510–3514. (29) Viggiano, A. A.; Morris, R. A.; Dale, F.; Paulson, J. F.; Giles, K.; Smith, D.; Su, T. J. Chem. Phys. 1990, 1149.
Analytical Chemistry, Vol. 82, No. 9, May 1, 2010
3765
Figure 2. Diagram of the Selected Ion Flow Tube-Mass Spectrometer (SIFT-MS) at the Air Force Research Laboratory.
phonate (DEMTMP) was introduced into the reaction region by bubbling helium through the room temperature vapor. The reaction time was previously measured. The remaining reactant ions and the product ions were sampled through a 0.2 mm diameter aperture in a blunt nose cone, then analyzed with a second quadrupole mass filter, and detected with a conversion dynode detector. The bimolecular rate constant for reaction of an ion, A+, with DEMTMP was measured under pseudofirst order conditions according to eq 1 below,30
ln
IAt + IA0 +
) -k[DEMTMP]t
(1)
where I was the intensity of A+, k was the rate constant in cm3 molecules-1 s-1, and t was the previously measured reaction time in seconds. DEMTMP was introduced by passing a helium carrier through a stainless steel bubbler containing fiberglass wool saturated with DEMTMP. The helium flow rate was determined by a MKS mass flow controller. The DEMTMP flow rate added was then determined by measuring the total pressure in the bubbler, Ptotal, and using eq 2, where QHe was the flow rate of the helium carrier gas in sccm and vp was the room temperature vapor pressure of DEMTMP. QDEMTMP ) vp × QHe/(Ptotal-vp)
(2)
DEMTMP had a low vapor pressure at room temperature that was not well-known. It could, however, be calibrated by studying a reaction for which the rate constant was known or calculable. We used a highly exothermic proton transfer reaction. This class of reaction had been shown repeatedly15 to proceed at the SuChesnavich collision rate constant.31,32 The DEMTMP PA calculated here (see below) was 890 kJ mol-1, i.e., much higher than that of H2O, so this assumption should be valid for the reaction with H3O+. We used this assumption to derive the vapor (30) Midey, A. J.; Viggiano, A. A. J. Chem. Phys. 1998, 109, 5257. (31) Su, T.; Chesnavich, W. J. J. Chem. Phys. 1982, 76, 5183. (32) Su, T. J. Chem. Phys. 1988, 89, 5355.
3766
Analytical Chemistry, Vol. 82, No. 9, May 1, 2010
pressure which was used to determine all other rate constants. Over the course of a given experimental period, this calibration reaction was frequently monitored and no variation was observed. The monitoring was done in part to confirm that the sample of DEMTMP had not diminished. The DEMTMP concentration in the flow tube was given by the buffer concentration times the ratio of the DEMTMP flow to the buffer flow. Accounting for the uncertainty in measuring the pressures, gas flows, and the vapor pressure of the simulant, the rate constants had relative uncertainties of ±30% and absolute uncertainties of ±40%. The product ion branching ratios were determined by extrapolating the fraction of total ion counts in each channel plotted against the neutral concentration to zero concentration, a method referred to as “relative with dilutions.”15 This extrapolation corrected for any secondary reactions of the product ions with the remaining DEMTMP in the flow tube. The uncertainties in the product ion branching ratios were ±10%. Energetics Calculations. In order to extrapolate from the experimental results for the DEMTMP surrogates to VX, theoretical calculations of minimum energy structures and energetics were performed at the G3(MP2) level of theory33 using Gaussian 0334 for neutral DEMTMP. The corresponding ionic products for proton, fluoride, oxide, and electron transfer reaction are summarized in Table 1. Calculations of both the oxide affinity (OA) and electron affinity (EA) for attaching an O- and e-, respectively, gave negative values for DEMTMP; thus, these reaction mechanisms are not discussed. An average absolute deviation of ±5.4 kJ mol-1 was found for energies calculated using G3(MP2) methods.33 This approach was analogous to our recent work with mustard surrogate 2-CEES and GX surrogate DMMP.10,11 Given the size and complexity of the VX molecule, higher order theoretical methods were required to determine the ion energetics properties. On the basis of the ion chemistry seen in the previous studies, the most important properties were the PA, IP, and FA. (33) Curtiss, L. A.; Redfern, P. C.; Raghavachari, K.; Rassolov, V.; Pople, J. A. J. Chem. Phys. 1999, 110, 4703. (34) Frisch, M. J.; In Gaussian 03, Revision C.02 ed.; Gaussian, Inc.: Pittsburgh, PA, 2003.
Table 1. Ion Energetic Properties for Dimethyl Methylthiomethylphosphonate (DEMTMP), Triethylamine ((C2H5)3N), and VXa
Table 2. Rate Constants for the Reaction of Various Negative Ions with Triethylamine ((C2H5)3N) at 298 K Measured in a Selected Ion Flow Tube (SIFT)a
DEMTMP
(C2H5)3N
VX
source ionization potential (IP) (eV) fluoride affinity (FA) (kJ mol-1)
G3(MP2) 8.26 157b 142c
ref 42 7.53
proton affinity (PA) (kJ mol-1)
910g 870h
982
DFT 7.3 111b 105d 88.6e 109f 943g 859h 1039i
a The proton affinities for VX represent the median value for each protonation location determined in the high level calculations. See text for details on the calculations. b F- on the P. c F- on the S. d F- complex at terminal CH3CH2O group. e F- complex at (i-C3H7)2N group. f Fcomplex between CH3CH2O and (i-C3H7)2N. g H+ on PdO oxygen. h H+ on the S. i H+ on N.
The initial structure search was performed using density functional theory (DFT)35-38 at the B3LYP/3-21G* level of theory to find the best starting geometries which were then optimized at the B3LYP/6-31G* level. To determine the IP, PA, and FA, single point energies were calculated for these geometries at the higher RIMP2/cc-pVTZ level.39 The goal of the experiments was to find fingerprint product ions for the reaction of various types of ions with the surrogate in order to extrapolate to reactions with VX. Thus, structure calculations were performed to find energetically allowed pathways for the observed fragment ions in order to gain some insight into the possible reaction pathways. As with the ion-molecule chemistry of 2-CEES and DMMP, barriers on the potential surfaces would influence the observed product ion branching ratios.10,11 However, a complete understanding of the reaction dynamics was beyond the scope of the current work. More detailed explorations of the potential surface will be examined in a separate publication. Special Material Handling. The following materials were used in the experiments: diethyl methylthiomethylphosphonate (Alfa Aesar, 98%), helium (AGA, 99.997%), oxygen (AGA, 99.999%), nitric oxide (Matheson, 99.5%), nitrogen dioxide (AGA, 99.5%), carbon dioxide (Middlesex Gases, 99.999%), sulfur hexafluoride (Matheson, 99.8%), ammonia (Matheson, 99.99%), acetone (Baker, HPLC grade), monomethylamine (Matheson, 99.5%), triethylamine (Aldrich, g99%), methanol (Baker, 99.8%), ethanol (Aldrich, anhydrous g99.5%), and distilled water. All of the reactant ions listed in Tables 1-3 were produced using the pure source gases. The materials could be handled using standard safety precautions for working with hazardous materials. DEMTMP was added to the bubbler in a fume hood via pipet, and then the bubbler was sealed with valves at the input and output prior to removal and installation on the instrument. The gaseous samples were introduced through vacuum tight connections that had been leak checked before use. The liquids were added by pipet to glass Becke, A. D. J. Chem. Phys. 1993, 98, 5648. Kim, K.; Jordan, K. D. J. Phys. Chem. 1994, 98, 10089–10094. Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98, 11623–11627. (39) Weigend, F.; Haser, M.; Patzelt, H.; Ahlrichs, R. Chem. Phys. Lett. 1998, 294, 143–152. (35) (36) (37) (38)
ion -
F SF5SF6CO3NO2NO3OO2-
products, (C2H5)3N
rate constant, 10-9 cm3 s-1
collision rate constant, 10-9 cm3 s-1
[F• (C2H5)3N]no reaction no reaction no reaction no reaction no reaction OH- + (C2H5)2N(C2H4) no reaction
0.094
