Part 2: Forensic attribution profiling of Russian VX in food using liquid chromatography-mass spectrometry

Jansson, Daniel; Lindström, Susanne Wiklund; Norlin, Rikard; Hok, Saphon; Valdez, Carlos A.; Williams, Audrey M.; Alcaraz, Armando; Nilsson, Calle; Åstot, Crister

doi:10.1016/j.talanta.2018.02.103

Title: Part 2: Forensic attribution profiling of Russian VX in food using liquid chromatography-mass spectrometry

Journal Article · 03 March 2018 · Talanta

DOI: https://doi.org/10.1016/j.talanta.2018.02.103 · OSTI ID:1513146

Jansson, Daniel ^[1]; Lindström, Susanne Wiklund ^[1]; Norlin, Rikard ^[1]; Hok, Saphon ^[2]; Valdez, Carlos A. ^[2]; Williams, Audrey M. ^[2]; Alcaraz, Armando ^[2]; Nilsson, Calle ^[1]; Åstot, Crister ^[1]

Swedish Defence Research Agency, Umeå (Sweden)
Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States). Forensic Science Center

This work is part two of a three-part series in this issue of a Sweden-United States collaborative effort towards the understanding of the chemical attribution signatures of Russian VX (VR) in synthesized samples and complex food matrices. We describe the sourcing of VR present in food based on chemical analysis of attribution signatures by liquid chromatography-tandem mass spectrometry (LC-MS/MS) combined with multivariate data analysis. Analytical data was acquired from seven different foods spiked with VR batches that were synthesized via six different routes in two separate laboratories. The synthesis products were spiked at a lethal dose into seven food matrices: water, orange juice, apple purée, baby food, pea purée, liquid eggs and hot dog. After acetonitrile sample extraction, the samples were analyzed by LC-MS/MS operated in MRM mode. A multivariate statistical calibration model was built on the chemical attribution profiles from 118 VR spiked food samples. Using the model, an external test-set of the six synthesis routes employed for VR production was correctly identified with no observable major impact of the food matrices to the classification. The overall performance of the statistical models was found to be exceptional (94%) for the test set samples retrospectively classified to their synthesis routes.

View Accepted Manuscript (DOE)

View Accepted Manuscript (Publisher)

Cite

Export

Save

Research Organization:: Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States); Swedish Defence Research Agency, Umeå (Sweden)

Sponsoring Organization:: USDOE National Nuclear Security Administration (NNSA); Dept. of Homeland Security (DHS) (United States); Swedish Civil Contingencies Agency