Development of SRM 2907 Trace Terrorist Explosives Simulants for

ARTICLE pubs.acs.org/ac

Development of SRM 2907 Trace Terrorist Explosives Simulants for the Detection of Semtex and Triacetone Triperoxide William MacCrehan,* Stephanie Moore, and Diane Hancock Analytical Chemistry Division, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, United States

bS Supporting Information ABSTRACT: Effective and accurate detection of trace explosives is crucial in the effort to thwart terrorist explosives attacks. A National Institute of Standards and Technology (NIST) standard reference material (SRM) has been developed for the evaluation of trace explosives detectors that sample by collection of residue particles using swiping or air filtration. SRM 2907 Trace Terrorist Explosives Simulants consists of two materials individually simulating the residues of the plastic explosive Semtex [for pentaerytritol tetranitrate (PETN)] and the improvised explosive triacetone triperoxide (TATP). Unique challenges were encountered in the development of these materials, including the selection of suitable inert substrates, material preparation, thermal stability testing, and analytical method development. Two independent analytical methods based on liquid chromatography with ultraviolet absorbance and mass spectrometric detection, LC-UV and LC/MS, respectively, were developed and used to certify the mass fractions of PETN and TATP. These materials were further evaluated for their suitability on a field swipe-sampled trace explosives detectors based on ion mobility spectrometry (IMS).

T

errorist explosives attacks, particularly in passenger transportation, have created the necessity to reliably detect trace explosives residues in field settings such as airports, border crossings, cargo transfer docks, and building entries. Although the exact nature of terrorist explosive devices is closely guarded information, the news media reporting on the so-called Lockarbie, shoe bomber, London subway, and underwear bomber attacks have implicated the use of Semtex containing pentaerytritol tetranitrate (PETN) and peroxide explosives such as triacetone triperoxide (TATP). As most explosives have a very low vapor pressure, the goal of trace detectors in screening applications is to collect and detect the particulate residues that remain following the fabrication of the improvised explosives devices. Typically, articles for inspection are swiped with a fabric mounted on a wand holder. Air-sampling systems, jet and thermal plume portals, and shoe vacuum systems are also under development. Following collection, the residue particles are heated to liberate the explosives for vapor-phase detection. In order to evaluate the effectiveness of these trace explosives detectors, both the residue particle collection and explosives vapor detection aspects of the measurement should be determined. The goal of developing standard reference material (SRM) 2907 Trace Terrorist Explosives Simulants is to provide a reasonable simulation of explosive particulate residues possessing a well-characterized amount of explosive in a nonexplosive form that may be used to optimize and validate the performance of trace explosives detectors. Previous study of the particle distribution of fingerprint residues from the plastic explosives Composition C-4 and Semtex 1A found a large distribution of sizes, with many particles below 10 μm.1 However, only particles of 10 μm and above possess significant This article not subject to U.S. Copyright. Published 2011 by the American Chemical Society

enough mass for effective detection using current technology. A “sweet spot” for residue detection was projected to be in the 10
30 μm size range based on frequency of occurrence and mass. This is the size range that was targeted in the development of the previous SRM 2905 Trace Particulate Explosives Simulants (certified for RDX, TNT, and HMX content)2 as well as the new SRM 2907. As with SRM 2905, in preparing SRM 2907, real explosive formulations are dissolved in suitable solvents and coated onto the inert simulant matrix. The preparation of the candidate materials, suitability testing for stability, performance using a field trace explosives detector, and the analytical methods for certification of this SRM are detailed.

’ EXPERIMENTAL SECTION Materials. Hydrogen peroxide 50 wt % (Sigma-Aldrich, St. Louis, MO), ACS reagent-grade sulfuric acid, and HPLCgrade acetone were use to synthesize TATP (see Supporting Information). Semtex 1A (Explosia, Semtin, Czeck Republic) was a gift from the Bureau of Alcohol, Tobacco, Firearms, and Explosives in Ammendale, MD. Substrate powders were based on a 20
30 μm diameter C18 silica (Protein and Peptide C18, Grace Discovery Sciences, no. 218TPB2030, Deerfield, IL) and 12
20 μm diameter polystyrene
divinylbenzene porous polymer beads PRP-1 (Hamilton, no. 79581, Reno, NV). Received: July 29, 2011 Accepted: October 17, 2011 Published: October 17, 2011 9054

dx.doi.org/10.1021/ac201967m | Anal. Chem. 2011, 83, 9054–9059

Analytical Chemistry 4-Nitrobenzonitrile (4-NBN, Avocado Research Chemicals, Heysham, Lancashire, U.K.) was used as the internal standard for liquid chromatography with ultraviolet absorbance detection (LC-UV) determination of TATP and PETN. It was twice recrystallized from 50:50 (volume fraction) pentane/acetone. 1,2,4-Butanetriol trinitrate (BTTN, no. B-002, Cerilliant, Round Rock, TX) was used as internal standard for the liquid chromatography/mass spectrometry (LC/MS) measurements of PETN. Ammonium acetate pH 5.5 (≈1 mol/L) was prepared by neutralization of 56 mL of electronic-grade glacial acetic acid dissolved in 500 mL of water with Ultrex ammonia (Merck, Darmstat, Germany) to pH 5.5 and final dilution to 1 L. A volume of 10 mL of this buffer was added to each liter of the two LC solvents (10 mmol/L) for the LC/MS determinations of TATP. Analytical reagent-grade ammonium nitrate (Mallinckrodt, Paris, KY) was used in the preparation of the aqueous mobile phase (0.2 mmol/L) for the PETN LC/MS determination. 2-Methylbutane (2-MB) was Chromasolv-grade (Sigma-Aldrich). All additional solvents used were HPLC-grade or equivalent. Preparation of TATP and 13C3-TATP. Synthesis of the TATP materials followed the procedure by Oxley et al.3 with notable modifications, see the Supporting Information. Preparation of Prototype and Candidate Powders. Simulated residues were prepared by coating explosives from solution onto inert substrate powders using rotary evaporation. Safety Note: For the rotary evaporative coating of these simulant materials, an explosion-resistant flask was used, and the apparatus was placed behind an explosion shield in a fume hood with the windows closed. The following personal protective equipment was worn: a buttoned cotton lab coat, safety glasses with side shields, a full face shield, 98 dB over-the-ear hearing protection, and nitrile laboratory gloves. Prototype Materials. Prototype materials containing Semtex and TATP were prepared for stability testing as described in the Supporting Information. One gram samples were stored in plastic squeeze or glass bottles at
20, 5, 23, 35, and 50 °C and withdrawn at predetermined times for evaluation by LC-UV. Candidate Material 0.04% Semtex on C18 Silica. An amount of 307.91 g of the C18 substrate was washed twice with diethyl ether and dried with N2 and subsequent rotovap vacuum at 35 °C for 3 h and 60 °C for 2 h. Then fresh diethyl ether was added to the powder in the rotovap flask in sufficient volume to “wet” the powder. An amount of 115.2 mg of Semtex 1A was dissolved in about 125 mL of diethyl ether and ultrasonically agitated for 30 min (adding small amounts of ice to the bath to prevent excessive heating). Since traces of residual Semtex particles were noted, the solution was decanted into the rotovap flask and an additional 50 mL volume of diethyl ether was added and ultrasonically agitated for 30 min. Care was taken not to transfer any particulate matter from this solution to the rotovap flask. The powder mixture was dried in the rotovap under N2, then under maximum vacuum for 2 h at 50 °C, followed by 0.5 h at 60 °C. However, since the powder still smelled of ether, a second washing was performed with 350 mL of 2-MB. The powder was dried under N2, and rotary vacuum was applied for 5 h at 35 °C and 2 h at 50 °C. The gravimetric value was determined from the mass of Semtex and substrate taken for preparation of this material and was ≈0.37 mg/g. Allowing for the ≈80% PETN content of this Semtex provided an expected value of ≈0.30 mg/g PETN. The bulk candidate material was stored at
20 °C. For the SRM packaging, 1.0 g samples were placed in the plastic squeeze bottles and stored at
20 °C.

ARTICLE

Candidate Material 0.4% TATP on PRP-1. An amount of 287.70 g of the PRP-1 was washed with two excess volumes of pentane. The powder was then dried under N2 and then full vacuum in the rotovap for 5 h at 50 °C and 2 h at 70 °C. After this drying, the powder still smelled of pentane. So the powder was further washed twice with 2-MB, dried with N2 and then full vacuum for 2 h each at 30 , 50, and 60 °C. Since the powder still had some odor, the powder was allowed to dry overnight under full vacuum at 30 °C. To this dried powder, 1.0216 g of refined TATP (described in the Supporting Information) was dissolved in ≈300 mL of 2-MB and added to the rotovap flask. The slurry was mixed for 15 min in the rotovap, bulk 2-MB was removed with N2, and then the powder was dried under full vacuum for 2 h at 35 °C and 3 h at 50 °C. The gravimetric value for this material was ≈3.8 mg/g. The bulk candidate material was stored at
20 °C. For the SRM packaging, 1.0 g samples were placed in the plastic squeeze bottles and stored at
20 °C. Primary Standard Purity Determination. 1H NMR Evaluation of TATP Purity. The identity and purity of the primary standard TATP was assessed with proton nuclear magnetic resonance spectroscopy (1H NMR) using the (100%
x) method. The NMR experiments were acquired at 283 K on a Bruker Avance 600 MHz spectrometer, operating under Topspin (version 2.1, PLl) software, using a 5 mm broad-band inverse detection probe. Data was collected using a one-dimensional pulse program with inverse-gated composite decoupling pulse (GARP) applied to 13C signals during acquisition. A 30° excitation pulse with a 45 s relaxation delay ensured reliable quantitative results. NMR experiments were performed in 2H3acetonitrile (Cambridge Isotope Laboratories, Andover, MA, DLM-53, 99.96% 2H). The water impurity in the acetonitrile solvent (≈700 μL) was determined before the addition of 9.96 mg of the TATP so that water content of the TATP could be established by difference. Typically spectra were acquired for 16 h 20 min. Raw data was line-broadened to 0.3 Hz by exponential multiplication and zeroed-filled, providing a spectral resolution of 0.114 Hz/point. Chemical shifts were referenced to the residual acetonitrile peak (1.941 ppm) and spectral peak assignments for the TATP and impurities were made using literature data.4,5 The presence of the rinse solvent, 2-MB, was confirmed by comparison with a spectrum of this compound dissolved in 2H3-acetonitrile. Additional Purity Evaluations. The purity of the primary standard PETN was determined by LC-UV, differential scanning calorimetry, and the certificate of analysis provided by the manufacturer (Cerilliant, Round Rock, TX). For the LC-UV purity determinations, columns with different selectivities, a C18 and a pentafluorphenyl propyl column, were used with UV detection at 210 nm. The purity of the TATP primary standard was determined by LC-UV with a C18 column (detector at 210 nm), gas chromatography with flame ionization detection (GC-FID), and gas chromatography/mass spectrometry (GC/ MS). Both GC separations were obtained using DB-5MS columns [30 m, 0.25 mm i.d., 0.25 μm film thickness (Agilent Technologies, Santa Clara, CA)]. The final purity assignments were a synthesis of all chromatographic and NMR measurements. Value Assignment Measurements. Sample Extractions. For the LC-UV determinations of TATP, ≈50 mg of the 0.4% TATP powder was transferred into a glass tube with stopper and the exact mass was determined. Then 1 mL of acetonitrile with 4-NBN internal standard was added, and the mass was determined. After 30 min of extraction with ultrasonic agitation, the 9055

dx.doi.org/10.1021/ac201967m |Anal. Chem. 2011, 83, 9054–9059

Analytical Chemistry sample tube was centrifuged and the supernatant withdrawn and placed in a 3 mL polypropylene 0.45 μm centrifugal filter that had been precleaned with acetonitrile. A fresh ≈1 mL portion of acetonitrile was added to the sample, extracted, and recovered. This process was repeated, resulting in three extractions with fresh solvent. All extracts were combined and filtered. The filtered solution was vortex-mixed for 10 s and placed in an LC sampling vial. LC/MS determination used the internal standard 13C3-TATP in isopropyl alcohol as extractant. A single ultrasonic extraction of ≈50 mg with 1 mL of solvent for 99 min and centrifugal filtration provided samples for determination. For the LC-UV measurements of PETN, samples were extracted as for TATP using 99 min of ultrasonic agitation using 4-NBN as internal standard in acetonitrile. For the LC/MS measurements, isopropyl alcohol was used as the extraction solvent with a BTTN internal standard. For all determinations, sample extracts were maintained at 5 °C in the LC autosampler tray. LC-UV Determinations. All LC separations used a two-pump LC system with UV absorbance detection at 210 nm. The separation used for value assignment of TATP used a reversed-phase ACE3 column (MACMOD Analytical, Chadds Ford, PA, 15 cm  0.3 cm, 3 μm particles) with a mobile phase flowing at 0.7 mL/min using the following solvent program: 15% acetonitrile/water mobile phase for 13 min then ramped to 43% over 7 min, hold for at 43% for 8 min, then stepped to 90% for 2 min before return to initial conditions. A 4 μL sample injection was used. For the value assignment of PETN, the same column, injection size, and flow rate were used. The mobile phase composition was 37.5% acetonitrile/water for 25 min then stepped to 100% for 1 min before return to initial conditions. LC/MS Determinations. The LC/MS value assignment measurements for TATP used a single-quadrupole mass spectrometer with atmospheric chemical ionization in the positive ion mode (APCI+). To enhance the sensitivity of TATP detection, an ammonium ion adduct (M + NH4)+ was formed6 by the addition of 10 mmol/L ammonium acetate to both LC solvents. An Eclipse XDB-C18 (15 cm  0.46 cm, 5 μm particles, Agilent Technologies) was used with an isocratic mobile phase that was 65% solvent B, where solvent A was 10 mmol/L NH4OAc and B was 99.5% methanol/water with 10 mmol/L NH4OAc. Flow rate was 1.0 mL/min with an injection size of 6 μL. Detection parameters were drying gas 5 L/min, nebulizer pressure 0.41 MPa (60 psi), drying gas temperature 325 °C, vaporizer temperature 325 °C, capillary voltage 3.5 kV, and corona current 4 μA. The (M + NH4)+ adduct ions monitored were at m/z 240.1 for TATP and m/z 243.1 for the internal standard 13C3-TATP. Methanol provided much larger MS signals for TATP than did analogous solvent compositions based on acetonitrile. For the LC/MS value assignment of PETN, atmospheric pressure chemical ionization in the negative mode (APCI
) was used. To enhance the sensitivity and improve the reproducibility of the measurements, a nitrate ion adduct (M + NO3)
was formed by the addition of 0.2 mmol/L of NH4NO3 to the aqueous eluent component. The same separation conditions were used as for TATP with a 65% methanol/water (0.2 mmol/L NO3
) eluent. MSD parameters were APCI
, gas 5 L/min, nebulizer pressure 0.41 MPa (60 psi), drying gas temperature 280 °C, vaporizer temperature 350 °C, capillary voltage 3 kV, and corona current 15 μA. The (M + NO3)
ions monitored were m/z 378.0 for PETN and m/z 308.0 for the internal standard BTTN. Although the operating conditions were above the reported decomposition temperature for ammonium nitrate, the inlet of the MS was cleaned with isopropyl alcohol/water daily after use.

ARTICLE

Ion Mobility Spectrometry Measurements. The instrument response for both the uncoated C18 and PRP-1 substrates and explosives-coated prototypes was evaluated on an Itemiser3 ion mobility spectrometry (IMS) detector (GE Ion Track, Wilmington, MA). Calibration experiments were done with the Itemiser3, which uses diffusion through a poly(dimethylsiloxane) membrane to select vapor-phase explosive molecules for detection. “Multipurpose” sample traps (Morpho Detection, Wilmington, MA) were used for swipe introduction of test materials. No alarm response was noted for the uncoated substrates. Prototype materials were “dusted” onto the swipe pretared on an ultramicrobalance, and the mass was determined to (0.5 μg. For detection of TATP, the positive ion channel was used with the ammonia dopant gas and the drift times were calibrated with solution-deposited TATP with subsequent evaporation. The detection of TATP used temperatures for the desorber of 163 °C and detector 162 °C. The desorber sampling time for TATP was 7 s. For detection of PETN, the negative ion channel with the dichloromethane dopant gas was used. The drift times were calibrated using PETN and assigning the drift time of “PETN 1”. Temperatures for the calibration were desorber 163 °C and detector 122 °C. The “PETN 1” peak response (≈9.1 ms drift time) was used without consideration of the “PETN 2” fragmentation peak at ≈4.4 ms. The desorber sampling time for PETN was set to 10 s.

’ RESULTS AND DISCUSSION The primary goal of this SRM development was to provide materials that were reasonable simulants of explosives residues and would provide accurately determined amounts of the target explosives in a nonexplosive format. Important secondary characteristics include homogeneity, stability, appropriate packaging, and obtaining the expected response in field explosives detectors. Our approach was to prepare prototype materials, test their properties and performance, and select candidate materials for value assignment. TATP Materials. As there are no vendors for solid TATP, it was necessary to arrange its preparation. Synthesis was done is a series of small batches. Following the procedure of Oxley et al.3 without modification resulted in a crude product that had many impurities. In addition to the cyclic side products (DADP) diperoxide and tetraacetone tetraperoxide (TrATrP), a series of linear oligomeric peroxides7 was also formed. Unanticipated products including a series of condensed ketones and keto acids were formed via the reaction of concentrated sulfuric acid with the acetone of the premixed reagent. These products were identified by GC/MS against the NIST Spectral Library and the literature,7,8 see Supplemental Figure 1 in the Supporting Information. To reduce these impurities, the crude product was dissolved in 2-MB and extracted three times with fresh volumes of 50% (volume fraction) methanol/water. This largely removed the polar impurities. Double recrystallizations successively from methanol and 2-MB provided the “primary standard TATP.” Additional recrystallizations from the “mother liquor” using sequentially methanol and then 2-MB resulted in a less pure but “refined TATP”. The NMR spectrum of the primary standard TATP prepared was found to be in good agreement with the literature spectra for TATP,4,5 securing its identity (Supplemental Figures 2 and 3 in the Supporting Information). The overall purity of the primary standard, evaluated by LC-UV (two columns), was estimated to be 2.2% (mass fraction) by GC/MS and 9056

dx.doi.org/10.1021/ac201967m |Anal. Chem. 2011, 83, 9054–9059

Analytical Chemistry 2.4% by GC-FID. The refined TATP was the material used for the preparation of the SRM candidate material. In the preparation of prototype and candidate materials for TATP, a low boiling solvent 2-MB (bp 27 °C) was used for dissolution for the rotary evaporative coating process. The goal was to enable the use of low temperatures in the removal of solvent. This would minimize losses of TATP by sublimation2 and/or decomposition.9 An important consideration in the preparation of the TATPcoated substrate material for SRM 2907 is the stability as a function of temperature. Upon preparation, prototype materials were stored at
20 °C and test samples were stored at
20, 5, 23, 35, and 50 °C. The first prototype material, 1% TATP on C18 silica, was stored in 1 g amounts in 4 mL plastic dropper bottles. After 18 days the samples were removed from storage and the concentration of TATP was determined by LC-UV, shown in Supplemental Figure 4 in the Supporting Information. The observed stability was poor, with ≈50% of the TATP content lost at room temperature. It is possible that TATP was being lost via decomposition10 or sublimation3 followed by transpiration through or adsorption on the plastic container. A new set of samples were stored in 2 mL glass containers and stored at the various temperatures, also shown in Supplemental Figure 4 in the Supporting Information. Evaluation of these glass-stored samples showed a very similar trend to the plastic bottles, with perhaps a little less loss. However, it was evident that the stability of the TATP on the C18 silica substrate was poor. It is possible that uncapped silanols provide reactive acidic sites for catalysis of TATP decomposition.10 It was clear that a TATP substrate with better stability was required. A polymeric particulate substrate would be free of residual silanols. PRP-1 is a porous polystyrene
divinylbenzene polymer with a nominal particle diameter of 12
20 μm, providing particles that are consistent in size with explosives residues.1 Once prepared via rotary evaporative coating, the 0.4% TATP on PRP-1 prototype material was stored at
20 °C and subsamples were stored at the various temperatures used for stability determination. The results are shown in Figure 1 where the PRP-1 substrate material is compared to the C18 after 18 days of storage. Since the concentrations of the TATP in the two prototypes are not equal, the plots are normalized to the value found at
20 °C. Much better stability, including surprising stability at 50 °C, was found for the TATP on the polymeric support. On the basis of this result for the prototype, the candidate material for TATP was prepared with the PRP-1. PETN Materials. For the preparation of Semtex component of SRM 2907, Semtex 1A was chosen over other Semtex formulations which contain a blend of PETN with significant amounts of RDX (hexahydro-1,3,5-trinitro-1,3,5-triazine). Extracts of this sample of Semtex 1A appeared to be free of RDX as determined by LC-UV and LC/MS. Materials representing Semtex residues were prepared on a C18 silica substrate. Dissolution of the polymer binder matrix (styrene butadiene) proved to be somewhat difficult and required the use of aggressive solvents. Most of the Semtex could be dissolved with either tetrahydrofuran (prototype material) or diethyl ether (candidate material) with ultrasonic agitation. Although the diethyl ether was highly volatile (bp 35 °C) and should have been easily removed by the rotary evaporative process, it proved to have significant retention on the C18 silica. It was largely removed by a second “wash” of the material with 2-MB and subsequent rotary evaporation. A study of the stability of PETN as a function of storage temperature over the range of
20 to 50 °C was performed for

ARTICLE

Figure 1. Comparison of the stability of TATP on two substrates as a function of storage temperature for 18 days.

the prototype 0.1% Semtex on C18 Silica. Supplemental Figure 5 in the Supporting Information provides the results for testing after 6 weeks of storage. Very good stability was noted for storage at room temperature and below. About 2% of the PETN was lost after 6 weeks of storage at 50 °C. The stability of PETN on the C18 silica was judged to be sufficient for the preparation of the 0.04% Semtex candidate material. Certification Measurements. Certified value assignment of SRM 2907 required two independent measurement techniques. Although both the TATP and PETN are only weakly chromophoric, at the levels present in the candidate materials, it was possible to determine these analytes using LC with UV absorbance detection at 210 nm. Measurements using LC/MS with APCI provided the second independent determination. Both approaches used the internal standard method, added to the powders before solvent extraction. The LC separation of TATP has a unique feature. The two prominent conformations have sufficient stability that they may be easily separated by reversed-phase LC.4,11 The relative amounts of the two conformers were found to be dependent on the time since dissolution of the crystalline material, the solvent used, temperature, and even on the separation conditions. It is reasonable to assume that the absorbance and mass spectral responses of the two conformers are very nearly equal. Thus, in order to determine the total amount of TATP, the peak areas of both conformers were summed. LC-UV measurements used 4-NBN as internal standard with acetonitrile as the extraction solvent. Development of a method for the TATP determination on this substrate required complete separation of the internal standard and the two conformers (referred to as TATP1 and TATP2) from the UV-absorbing matrix constituents extracted from the polystyrene
divinylbenzene polymer. Analysis of PRP-1 solvent extracts by GC/MS and comparison to the NIST Spectral Library identified a series of alkyl benzenes and phenols as extractable matrix constituents. Since we could prepare extracts of the uncoated PRP-1 substrate, it was possible to optimize the chromatographic LC separation to provide “clean” baselines at the retention times of the three target compounds in the candidate material. Figure 2 shows the separation used for the LC-UV value assignment of TATP in the SRM material. This column/solvent combination provided an unusually high resolution of the TATP1 and TATP2 conformers compared to other LC separations evaluated and provided the required resolution to accurately determine 9057

dx.doi.org/10.1021/ac201967m |Anal. Chem. 2011, 83, 9054–9059

Analytical Chemistry

ARTICLE

the analyte and internal standard peak areas from substrate matrix constituents. LC/MS measurements used the synthesized 13C3-TATP as internal standard and isopropyl alcohol as extraction solvent. Both the unlabeled and stable isotope labeled TATP gave prominent (M + NH4)+ ions at m/z 240.1 and m/z 243.2 (see Supplemental Figure 6 in the Supporting Information), respectively, in the APCI+ mode with addition of the ammonium acetate mobile phase additive. The complete absence of detectable unlabeled TATP in the 13C-TATP spectrum confirms the successful synthesis of the labeled analogue. The molecular adduct ions were monitored in the selected ion (SIM) for quantitation. With only a difference of 3 Da and a natural abundance exact mass of 222, the labeled and unlabeled TATP conformers had nearly exactly the same LC retention time. For the LC/MS determination of PETN, we were unable to obtain a stable isotope labeled analogue of PETN. BTTN was chosen as a close analogue, with the structure of PETN substituting a
H for one of the four
CH2
O
NO2 groups. PETN and BTTN have similar UV spectra (decreasing absorbance at wavelengths greater than 200 nm). Even without the addition of mobile phase ionization additives, BTTN and PETN both form prominent molecular adduct ions at (M + NO3)
in the APCI
mass spectral detection mode (Supplemental Figure 7 in the Supporting Information). SIM monitoring was used at m/z 303.0 for BTTN and m/z 378.0 for PETN. These adducts presumably form via the decomposition of the alkyl nitrates to provide NO3
ions for attachment to another neutral molecule. Δ

alkyl nitrate f xNO3
þ decomposition products

ð1Þ

NO3
þ alkyl nitrate f ðM þ NO3 Þ


ð2Þ

The signals for BTTN (retention time ≈4.0 min) and PETN (retention time ≈5.5 min) did not overlap, which may be of concern if matrix components coelute and give rise to ionization suppression or enhancement of the MS signal for the target compounds. An extract of the C18 silica matrix was found to be

Figure 2. Gradient elution LC-UV separation of extract of SRM 2907 for determination of TATP.

free of extractable impurities (as determined by the UV at 210 nm) at the retention times of the target compounds. Thus, measurement interference via ionization suppression/enhancement should be negligible for this simple matrix. However, the uncertainty of the APCI
measurements without the addition of modifier was quite poor (11.4% RSD, n = 24). This may be the result of the sequential alkyl nitrate thermal decomposition and ionization reactions required for the formation of detectable ions. As noted previously with HMX and RDX, the addition of an adducting ion to the LC mobile phase can enhance the sensitivity of APCI detection and also improve the reproducibility of the measurements [ref 2 and references therein]. Supplying excess nitrate as a reactant ion can favor reaction 2 in the ionization process providing behavior resembling a “pseudo-first-order reaction” in the alkyl nitrate. Addition of 0.2 mmol/L ammonium nitrate enhanced the signal for BTTN by a factor of ≈8 and PETN by a factor of ≈22. More importantly, the uncertainty of the PETN determination using the internal standard method was reduced to 2.8% RSD, n = 24. For the certification measurements, binned randomized samplings from the sequential filling of the ≈275 bottles were used. Two samples were taken from each bottle. Twelve bottles were used for the LC-UV determinations and 6 bottles used for the LC/MS measurements. Table 1 provides the values obtained by each method (corrected for the purity of the primary standards), the certified values, and the expanded uncertainty. The certified value is a weighted mean of the results from two analytical methods.12 The uncertainty listed with each value is an expanded uncertainty about the mean, with coverage factor 2 (approximately 95% confidence) calculated by combining a between-method variance with a pooled, within-method variance13
15 following the ISO Guide.16 Evaluation of IMS Detector Response. Since the ultimate goal of this effort was to provide a reference material for the trace explosive detection, it was important to determine if the SRM could be used to calibrate a currently available IMS detector.

Figure 3. Calibration of the IMS detector with SRM 2907 and solutiondeposited TATP: diamonds, SRM 2907; circles, solution-deposited TATP. Inset: combination of particle and solution deposition of TATP in small amounts.

Table 1. Certification Results for SRM 2907 Trace Terrorist Explosives Simulants LC-UV (mg/g)

n

LC/MS (mg/g)

n

certified value (mg/g)

0.4% TATP (TATP)

3.60 (0.10)

22

3.67 (0.03)

12

3.63 ( 0.07

0.04% Semtex (PETN)

0.289 (0.008)

30

0.310 (0.007)

12

0.291 ( 0.014

9058

dx.doi.org/10.1021/ac201967m |Anal. Chem. 2011, 83, 9054–9059

Analytical Chemistry For this evaluation, the explosives were deposited from a dilute isopropyl alcohol solution onto the sampling swipe and the solvent was allowed to evaporate in the hood. The disappearance of visible liquid was monitored. The swipe was then quickly inserted (