Anal. Chem. 2007, 79, 6368-6375
Single-Particle Aerosol Mass Spectrometry for the Detection and Identification of Chemical Warfare Agent Simulants Audrey N. Martin,†,‡ George R. Farquar,*,† Matthias Frank,† Eric E. Gard,† and David P. Fergenson†
Lawrence Livermore National Laboratory, Livermore, California 94550, and Michigan State University, East Lansing, Michigan 48824
Single-particle aerosol mass spectrometry (SPAMS) was used for the real-time detection of liquid nerve agent simulants. A total of 1000 dual-polarity time-of-flight mass spectra were obtained for micrometer-sized single particles each of dimethyl methyl phosphonate, diethyl ethyl phosphonate, diethyl phosphoramidate, and diethyl phthalate using laser fluences between 0.58 and 7.83 nJ/µm2, and mass spectral variation with laser fluence was studied. The mass spectra obtained allowed identification of single particles of the chemical warfare agent (CWA) simulants at each laser fluence used although lower laser fluences allowed more facile identification. SPAMS is presented as a promising real-time detection system for the presence of CWAs. Detection of chemical warfare agents (CWAs) in a real-time manner is essential due to their toxicity and the necessity of immediate medical treatment after exposure.1 In addition to fast analysis, the ideal detection system would be sensitive, selective, and able to operate at a specified physical distance from the population it is designed to protect.2,3 This would allow the prevention or mitigation of casualties after a chemical attack by reducing the number of victims exposed and their level of exposure.1,4 Trace detection is critical due to the small amount of CWA that is sufficient to cause serious injury or death.3 CWAs present the additional challenge for detection in that the agent could be present and harmful in either the liquid or vapor phase.4 Several detection schemes have targeted one phase or the other.1-3 Surface acoustic wave chemical sensors have been used to detect small amounts of vapor-phase CWAs with the capability of preconcentrating the sample before analysis.5-7 Due to the nature of the sensor, only gaseous forms of the agents can * To whom correspondence should be addressed. E-mail:
[email protected], (925)424-4275. † Lawrence Livermore National Laboratory. ‡ Michigan State University. (1) Smith, W. D. Anal. Chem. 2002, 74, 462A-466A. (2) Hill, H. H., Jr.; Martin, S. J. Pure Appl. Chem. 2002, 74, 2281-2291. (3) Seto, Y.; Kanamori-Kataoka, M.; Tsuge, K.; Ohsawa, I.; Matsushita, K.; Sekiguchi, H.; Itoi, T.; Iura, K.; Sano, Y.; Yamashiro, S. Sens. Actuators, B: Chem. 2005, 108, 193-197. (4) Chemical Casualty Care Division (USAMRICD). Medical Management of Chemical Casualties Handbook. 3rd ed.; EAI Corp.: Abingdon, MD, 2000. (5) Hartmann-Thompson, C.; Hu, J.; Kaganove, S. N.; Keinath, S. E.; Keeley, D. L.; Dvornic, P. R. Chem. Mater. 2004, 16, 5357-5364. (6) Laljer, C. E. MITRE Tech. Pap. 2000, 1-9.
6368 Analytical Chemistry, Vol. 79, No. 16, August 15, 2007
be detected online. Other technologies, such as electrochemical biosensors,8,9 are capable of detecting liquid-phase organophosphate agents in a short time.10,11 Ion mobility spectrometry (IMS) has also been commonly used to detect CWAs.1,2,12 A fielddeployable IMS, the “Improved Chemical Agent Monitor”, is capable of detecting vapor-phase G-type and V-type CWAs after 1 min of exposure using a radioactive ionization source.2 A detection system capable of multiphase analysis would be advantageous; mass spectrometry has been used to independently detect each phase of agent although not simultaneously. Several complementary techniques have been used for sample introduction in mass spectrometry.13-23 INFICON (http://www. inficon.com) has developed Hapsite,24 a semiportable, batteryoperated, GC/MS that is capable of sampling air, soil, or water. After separation in the GC, the sample components are introduced into the mass spectrometer, ionized using electron impact ionization, and analyzed in a quadrupole ion trap.24 Sekiguchi et al. used the Hapsite detection system for the analysis of vapor-phase CWAs, reporting detection limits of 0.2, 0.5, and 8 µg/m3, (7) McGill, R. A.; Nguyen, V. K.; Chung, R.; Shaffer, R. E.; DiLella, D.; Stepnowski, J. L.; Mlsna, T. E.; Venezky, D. L.; Dominguez, D. Sens. Actuators, B: Chem. 2000, 65, 10-13. (8) Simonian, A. L.; Grimsley, J. K.; Flounders, A. W.; Schoeniger, J. S.; Cheng, T.; DeFrank, J. J.; Wild, J. R. Anal. Chim. Acta 2001, 442, 15-23. (9) Lin, Y.; Lu, F.; Wang, J. Electroanalysis 2003, 16, 145-149. (10) Mulchandani, A.; Mulchandani, P.; Chen, W.; Wang, J.; Chen, L. Anal. Chem. 1999, 71, 2246-2249. (11) Mulchandani, P.; Chen, W.; Mulchandani, A. Environ. Sci. Technol. 2001, 35, 2562-2565. (12) Eiceman, G. A.; Stone, J. A. Anal. Chem. 2004, 76, 390A-397A. (13) Steiner, W. E.; Clowers, B. H.; Matz, L. M.; Siems, W. F.; Hill, H. H., Jr. Anal. Chem. 2002, 74, 4343-4352. (14) Steiner, W. E.; Clowers, B. H.; Haigh, P. E.; Hill, H. H., Jr. Anal. Chem. 2003, 75, 6068-6076. (15) Kientz, C. E. J. Chromatogr., A 1998, 814, 1-23. (16) Steiner, W. E.; Klopsch, S. J.; English, W. A.; Clowers, B. H.; Hill, H. H., Jr. Anal. Chem. 2005, 77, 4792-4799. (17) Tοrnes, J. A. Rapid Commun. Mass Spectrom. 1996, 10, 878-882. (18) Richardson, D. D.; Sadi, B. B. M.; Caruso, J. A. J. Anal. At. Spectrom. 2006, 21, 396-403. (19) Makas, A. L.; Troshkov, M. L.; Kudryavtsev, A. S.; Lunin, V. M. J. Chromatogr., B 2004, 800, 63-67. (20) Mulligan, C. C.; Justes, D. R.; Noll, R. J.; Sanders, N. L.; Laughlin, B. C.; Cooks, R. G. Analyst 2006, 131, 556-567. (21) Shu, Y.; Su, A.; Liu, J.; Lin, C. Anal. Chem. 2006, 78, 4697-4701. (22) Ellis-Steinborner, S.; Ramachandran, A.; Blanksby, S. J. Rapid Commun. Mass Spectrom. 2006, 20, 1939-1948. (23) Cotte-Rodriguez, I.; Justes, D. R.; Nanita, S. C.; Noll, R. J.; Mulligan, C. C.; Sanders, N. L.; Cooks, R. G. Analyst 2006, 131, 579-589. (24) Hapsite Smart Chemical Identification System; Inficon: East Syracuse, NY,l 2004. 10.1021/ac070704s CCC: $37.00
© 2007 American Chemical Society Published on Web 07/14/2007
respectively, for sarin, soman, and tabun at S/N ) 3.25 A preconcentrating step is required for vapor detection, which increases the overall analysis time, and samples with low volatility must be derivatized in order to increase their volatility and allow separation and detection. D’Agostino et al. used liquid chromatography-electrospray ionization-MS (LC-ESI-MS) for the identification of CWAs in aqueous extracts of office media (paper, carpet, drywall).26 Sarin, cyclosarin, soman, and triethyl phosphate were all recovered from the materials, with efficiencies between 20 and 100%, depending on the medium tested. Desorption electrospray ionization (DESI) was also used to analyze the spiked media.27 A solid-phase microextraction fiber was used to preconcentrate the sample from headspace and was then analyzed by DESI; sarin and tabun were detected at µg of CWA/g of media in this manner.27 IMS has recently been used as a sample separation and introduction device for mass spectrometry. Steiner et al. used electrospray ionization to introduce a liquid CWA into an IMS system for separation,13 followed by ion introduction into a TOFMS for further analysis. The orthogonal nature of the data obtained from the IM(TOF)MS allowed the separation and identification of numerous CWA degradation products both neat and in water samples with LODs less than 100 ppb (µg/L) for most compounds.13 The instrument was also adapted for vapor-phase analysis using preconcentration through a heated capillary or membrane,14 and tested using dimethyl methyl phosphonate (DMMP) aerosolized in various solvents producing the same drift times and masses of ions as the DMMP sampled in the aqueous phase.16 Single-particle aerosol mass spectrometry (SPAMS) has been developed over the past decades into a real-time, online detection system capable of sizing and obtaining positive and negative ion mass spectra of single particles.28-31 SPAMS is capable of analysis in a variety of environments,28,29 and has been applied at LLNL to the detection of single and composite explosives,32 as well as various biological samples: Bacillus spores,33-39 viruses,33 toxin simulants,33 fungi,35 and Mycobacteria,39 under the name of (25) Sekiguchi, H.; Matsushita, K.; Yamashiro, S.; Sano, Y.; Seto, Y.; Okuda, T.; Sato, A. Forensic Toxicol. 2006, 24, 17-22. (26) D’Agostino, P. A.; Hancock, J. R.; Chenier, C. L.; Lepage, C. R. J. J. Chromatogr., A 2006, 1110, 86-94. (27) Wood, M.; Laloup, M.; Samyn, N.; del Mar Ramirez Fernandez, M.; de Bruijn, E. A.; Maes, R. A. A.; De Boeck, G. J. Chromatogr., A 2006, 1130, 3-15. (28) Gard, E.; Mayer, J. E.; Morrical, B. D.; Dienes, T.; Fergenson, D. P.; Prather, K. A. Anal. Chem. 1997, 69, 4083-4091. (29) Noble, C. A.; Prather, K. A. Environ. Sci. Technol. 1996, 30, 2667-2680. (30) Noble, C. A.; Prather, K. A. Mass Spectrom. Rev. 2000, 19, 248-274. (31) Sinha, M. P. Rev. Sci. Instrum. 1984, 55, 886-891. (32) Martin, A. N.; Farquar, G. R.; Gard, E. E.; Frank, M.; Fergenson, D. P. Anal. Chem. 2007, 79, 1918-1925. (33) Steele, P. T.; Srivastava, A.; Pitesky, M. E.; Fergenson, D. P.; Tobias, H. J.; Gard, E. E.; Frank, M. Anal. Chem. 2005, 77, 7448-7454. (34) Srivastava, A.; Pitesky, M. E.; Steele, P. T.; Tobias, H. J.; Fergenson, D. P.; Horn, J. M.; Russell, S. C.; Czerwieniec, G. A.; Lebrilla, C. B.; Gard, E. E.; Frank, M. Anal. Chem. 2005, 77, 3315-3323. (35) Fergenson, D. P.; Pitesky, M. E.; Tobias, H. J.; Steele, P. T.; Czerwieniec, G. A.; Russell, S. C.; Lebrilla, C. B.; Horn, J. M.; Coffee, K. R.; Srivastava, A.; Pillai, S. P.; Shih, M.-T. P.; Hall, H. L.; Ramponi, A. J.; Chang, J. T.; Langlois, R. G.; Estacio, P. L.; Hadley, R. T.; Frank, M.; Gard, E. E. Anal. Chem. 2004, 76, 373-378. (36) Steele, P. T.; Tobias, H. J.; Fergenson, D. P.; Pitesky, M. E.; Horn, J. M.; Czerwieniec, G. A.; Russell, S. C.; Lebrilla, C. B.; Gard, E. E.; Frank, M. Anal. Chem. 2003, 75, 5480-5487. (37) Czerwieniec, G. A.; Russell, S. C.; Tobias, H. J.; Pitesky, M. E.; Fergenson, D. P.; Steele, P.; Srivastava, A.; Horn, J. M.; Frank, M.; Gard, E. E.; Lebrilla, C. B. Anal. Chem. 2005, 77, 1081-1087. (38) Tobias, H. J.; Pitesky, M. E.; Fergenson, D. P.; Steele, P. T.; Horn, J.; Frank, M.; Gard, E. E. J. Microbiol. Methods 2006, 67, 56-63.
Figure 1. Schematic of a SPAMS used to obtain positive and negative ion time-of-flight mass spectra of chemical warfare agent simulants. Aerosol samples are formed in a Collison nebulizer and sampled into the SPAMS where they are aerodynamically focused, tracked, and desorbed and ionized by a Nd:YAG laser (266 nm). Positive and negative ion spectra of material ionized from individual particles are simultaneously collected by a dual-polarity reflectron time-of-flight mass spectrometer and identified in embedded hardware and real-time software.
bioaerosol mass spectrometry (BAMS). The SPAMS system is capable of detecting a low concentration of threat particles in a high background environment. The automated data analysis system developed at LLNL, which is coupled to SPAMS, can be used for immediate analysis providing early warning if a threat particle is detected. These characteristics make SPAMS an ideal candidate for a real-time point detection system to monitor an environment for a wide variety of threat agents simultaneously. The current work builds upon our recent application of SPAMS to the detection of explosives32 and biological agents by expanding the tested detection capability of the instrument to also include CWA detection in the liquid phase. Studies of several simulants ionized over a range of laser energies are presented, and the optimal conditions are discussed. This work discusses only liquidphase agents; as CWAs can be harmful in both the liquid phase as well as the vapor phase, vapor studies were also pursued and will be presented in a future work. MATERIALS AND METHODS Sample Preparation. CWA simulants, DMMP (>97.0% purity, a sarin simulant), diethyl ethyl phosphonate (DEEP; >98.0%, a sarin simulant), diethyl phosphoramidate (DEPA; >98%, a tabun simulant), and diethyl phthalate (DEP; >99.5%, a VX simulant) were obtained from Sigma-Aldrich (St. Louis, MO). DMMP, DEEP, and DEP were obtained in the liquid phase; DEPA was obtained as a solid and dissolved in methanol (10 mg/10 mL) before analysis. A 10-mL aliquot of the CWA simulant was placed in a glass low-volume Collison nebulizer (BGI Inc., CN-25, 27) connected to a house nitrogen line for aerosol generation. The aerosolized particles were passed through a diffusion drier and transported to the SPAMS via ∼8 m of 3/8-in. conductive tubing. Instrument Parameters. SPAMS instrumentation has been previously described35 and is shown in Figure 1. The dried aerosol (39) Tobias, H. J.; Schafer, M. P.; Pitesky, M.; Fergenson, D. P.; Horn, J.; Frank, M.; Gard, E. E. Appl. Environ. Microbiol. 2005, 71, 6086-6095.
Analytical Chemistry, Vol. 79, No. 16, August 15, 2007
6369
Table 1. Chemical Name and Structure of Several Chemical Warfare Agents and the Simulants Commonly Used To Study Their Detection
enters the instrument through an interface, which includes a converging nozzle, pressure flow reducer, and aerodynamic lens stack, resulting in a tightly focused aerosol beam. The particles then pass through a tracking region of up to six pairs of lasers (CW, 660 nm) and photomultiplier tubes, three of which were operated in this study. The motion of a particle as it passes through the lasers is used to determine its velocity and aerodynamic diameter by recording the time between scattering events
of each laser. The particle velocity is used to trigger the firing of the desorption/ionization (D/I) laser (Q-switched, frequencyquadrupled Nd:YAG, Ultra CFR, Big Sky Laser Technologies, Inc., 266 nm, 7-ns pulses) at the time when the particle is centered in the source region of the mass spectrometer. The D/I laser output is optically modified to produce a relatively flat-topped beam profile,33 the energy of which is controlled by a manually rotatable half-wave plate and vertical polarizer prior to reaching the ion
Figure 2. Single-particle mass spectra of aerosolized DMMP, DEEP, DEPA, and DEP obtained with a laser fluence of 1.11 nJ/µm2 (note differences in y-axes). Each spectrum is from an individual particle of the agent and demonstrates the ability to differentiate the agents with one particle. 6370 Analytical Chemistry, Vol. 79, No. 16, August 15, 2007
Figure 3. Single-particle mass spectra of aerosolized DMMP at laser fluences of 0.58, 1.11, 2.57, 5.05, and 7.66 nJ/µm2 (note also differences in y-axes). Each spectrum is the average of 1000 single-particle spectra. The average particle was 1.15 ( 0.09 µm in diameter. Note the [DMMP + H]- peak at m/z -125 is present at all laser fluences.
source region. The laser light is focused to an ∼330 µm diameter spot at the point where it intersects the particle beam. The energy of each laser pulse is measured at the exit of the ion source chamber with a laser power meter (Coherent, Inc., J25LP-MUV) with digital readout. Before the experiments described here, the laser power meter was calibrated with and without the laser exit optical window to account for the window’s absorbance. Power measurements are reported as an average of 50 pulses and corrected for the measured energy loss in the exit window. Fluence was calculated by dividing the net laser energy by the cross-sectional area of the laser beam (330-µm diameter). The positive and negative ions generated in the source region were simultaneously extracted in opposite directions to two reflectronTOF mass analyzers. Ions were detected via microchannel plates (Burle Technologies, Inc.), and the signal was digitized (Signatec, Inc., PDA1000) at 8 bits dynamic range on a 333-mV scale. Data were analyzed using software developed in-house. RESULTS AND DISCUSSION The CWA simulants studied in this work are commonly used to simulate detection of the toxic CWAs.12,14,16,20,40 They were selected based on similarities in structure, volatility, or both to the nerve agents. Table 1 shows the chemical structures of several CWAs along with those of the simulant used for analysis.
Figure 2 shows mass spectra obtained with a laser fluence of 1.16 nJ/µm2 (RSD 2.2%) from a single particle of each chemical agent. Positive ion and negative ion mass spectra were simultaneously obtained and are concatenated in this figure as well as those following. The top trace shows a randomly selected single particle of DMMP. The mass spectrum from this particle, along with nine subsequent particles of DMMP, were obtained in less than 1 s. The aerodynamic diameter of the particle shown was 1.33 µm, which correlates to a 1.2-pg particle by assuming the droplet is spherical. The remainder of the figure shows mass spectra from single particles of DEEP, DEPA, and DEP. Each spectrum contains peaks indicative of the agent that are discussed below. The information obtained from the analysis of a single particle is sufficient to differentiate these agents. Although each single particle contains the peaks indicative of each agent, averages of 1000 particles are used in the following experiments to ensure the repeatability of the measurement. The effect of laser energy on the desorption/ionization of neat DMMP was investigated. Average mass spectra of 1000 individual DMMP particles obtained with laser fluences covering the output range of the D/I laser, 0.58, 1.11, 2.57, 5.05, and 7.66 nJ/µm2, are shown in Figure 3. The particles’ average aerodynamic diameter (40) Novak, J. P.; Snow, E. S.; Houser, E. J.; Park, D.; Stepnowski, J. L.; McGill, R. A. Appl. Phys. Lett. 2003, 83, 4026-4028.
Analytical Chemistry, Vol. 79, No. 16, August 15, 2007
6371
Figure 4. Single-particle mass spectra of aerosolized DEEP at laser fluences of 1.18, 2.77, 5.13, and 7.59 nJ/µm2 (note differences in ion signal). Each spectrum is the average of 1000 single-particle spectra. The average particle was 1.11 ( 0.10 µm in diameter. Note the [DEEP]peak at m/z -166 is present at all laser fluences.
was 1.15 µm (RSD 7.8%, n ) 1000). The negative molecular ion [DMMP + H]- (m/z -125) was present in all the spectra and was most prevalent when a D/I laser fluence of 2.57 nJ/µm2 was used. As the laser fluence was increased, increased fragmentation occurred and less molecular ion and other characteristic negative peaks (e.g., m/z -277, -141) were present. At the lower D/I laser fluences (2.57 nJ/µm2 and below), several fragments of DMMP were identified including [P(OH)OCH3]-, [DMMP - CH3 CH2]-, and [DMMP - CH3]-, at m/z -79, -95, and -109, respectively, many of which have been seen in previous studies.17,19,20,26 The maximum signal from each of these peaks was also seen using a laser fluence of 2.57 nJ/µm2. The negative ion spectra also contain several peaks above the mass of the parent ion. The hydroxyl adduct of DMMP is seen at m/z -141 and again showed the highest signal when obtained using the 2.57 nJ/µm2 laser fluence. The peak at m/z -166, is one of the most prominent peaks in the spectra obtained at low laser fluences, with its maximum signal at 1.11 nJ/µm2. This peak could be due to an unknown adduct of DMMP or fragmentation of a DMMP dimer. Since this peak is above the mass of DMMP, it is not surprising that its maximum signal occurs at a lower laser fluence than the fragments described above; at fluences above 1.11 nJ/µm2, this compound may itself be fragmented. Several other peaks related to DMMP dimers are also seen at the low laser fluences: [2DMMP + OCH3 - 2H]- and [2DMMP + 51]- are present at m/z -277 and -299, respectively. These compounds are only present when the sample is ionized at 0.58 and 1.11 nJ/µm2; at higher fluences, the energy incident on the particles may be too great to allow dimer formation. 6372 Analytical Chemistry, Vol. 79, No. 16, August 15, 2007
The positive ion spectra also contain identifiable peaks due to DMMP fragments and dimers. Peaks due to the dimerization of DMMP include [2DMMP - O]+ and [2DMMP + H]+, at m/z +232 and +249, respectively; the [2DMMP + H]+ peak has been previously reported.41 Several unknown peaks at m/z +278 and +290, as well as a trimer peak at m/z +343 ([3DMMP - 2O + 3H]+) are seen at low laser fluences (
