Analysis and classification of smokeless powders by GC–MS and DART-TOFMS. (November 2018)

• Record Type: Journal Article
• Authors: Lennert, Emily; Bridge, Candice
• Is Part Of: Forensic science international. Volume 292(2018)
• Page Start: 11
• Page End: 22
• Publication Date: 2018-11
• Subjects: Smokeless powders -- DART-TOFMS -- GC–MS -- Chemometrics; Medical jurisprudence -- Periodicals; Chemistry, Forensic -- Periodicals; Forensic Medicine -- Periodicals; Médecine légale -- Périodiques; Chimie légale -- Périodiques; Gerechtelijke geneeskunde; Gerechtelijke chemie; Gerechtelijke psychiatrie; Chemistry, Forensic; Medical jurisprudence; Electronic journals; Periodicals; Electronic journals; 614.1
• DOI: 10.1016/j.forsciint.2018.09.003
• Languages: English
• ISSNs: 0379-0738

Abstract

Highlights: Smokeless powders analyzed via GC–MS and DART-TOFMS to compare instrumental methods. Compounds observed were comparable between instrumental methods. Robust classification models were produced from both instruments. Indicates each instrument has a high probability of classifying new samples. DART-TOFMS is a suitable screening method for organic components in powders. Abstract: Partially burned smokeless powder particles may be present as a form of evidence following a shooting or explosive event, such as the explosion of a pipe bomb. The characterization and classification of residual smokeless powers may allow for a known sample, i.e. sample collected from a suspect, to be connected to an unknown sample, i.e. sample obtained from a crime scene. In this study, thirty-four (34) smokeless powders were analyzed using GC–MS and DART-TOFMS to determine how comparable the discriminatory power of each instrument was based on the smokeless powder constituents identified within each sample. Analysis of smokeless powders by DART-TOFMS generated comparable results to GC–MS in a fraction of the time (∼30 seconds). Most peaks observed between the instruments were the same; however, N-nitroso-DPA was only observed in the DART-TOFMS spectra but was not a significant contributor. Samples were naturally grouped together using hierarchical cluster analysis and principal component analysis, based on underlying features in the resulting spectra. Samples were placed into groupings based on significant peaks observed and relative intensities. Classification models were presented for both GC–MS and DART-TOFMS and subsequently tested and compared. The classification models used in this study were linear discriminant analysis, k-nearest neighbors, and random forest modeling. The groups observed were similar between the two instruments, indicating that DART-TOFMS provides comparable data to GC–MS and could be used as a rapid screening technique.