Collecting data under field conditions for forensic investigations of chemical warfare agents calls for the use of portable instruments. In this study, a set of aged, crude preparations of sulfur mustard were characterized spectroscopically without any sample preparation using handheld Raman and portable IR instruments. The spectral data was used to construct Random Forest multivariate models for the attribution of test set samples to the synthetic method used for their production. Colored and fluorescent samples were included in the study, which made Raman spectroscopy challenging although fluorescence was diminished by using an excitation wavelength of 1064 nm. The predictive power of models constructed with IR or Raman data alone, as well as with combined data was investigated. Both techniques gave useful data for attribution. Model performance was enhanced when Raman and IR spectra were combined, allowing correct classification of 19/23 (83%) of test set spectra. The results demonstrate that data obtained with spectroscopy instruments amenable for field deployment can be useful in forensic studies of chemical warfare agents.

在野外条件下收集数据以进行化学战剂的法医调查需要使用便携式仪器。在这项研究中，使用手持式拉曼光谱仪和便携式红外仪器对一组老化的粗制硫芥子油进行了光谱表征，而无需任何样品制备。光谱数据用于构建随机森林多元模型，以将测试集样本归因于用于其生产的合成方法。有色和荧光样品包括在研究中，这使拉曼光谱具有挑战性，尽管通过使用1064 nm的激发波长可以减少荧光。研究了仅使用IR或拉曼数据以及组合数据构建的模型的预测能力。两种技术都为归因提供了有用的数据。结合拉曼光谱和红外光谱可以提高模型性能，从而可以对19/23（83％）的测试集光谱进行正确分类。结果表明，适用于现场部署的光谱仪器获得的数据可用于化学战剂的法医研究。