Discrimination and classification of high explosives and other organic materials based on laser-induced plasma spectroscopy

Abstract

A substantial number of works have focused on the identification and classification of high explosives and organic compounds based on nanosecond laser-induced plasma spectroscopy (LIPS), but the major spectral differences between different molecular solids are still not well understood. The objective of this work is to determine the spectral regions that are most discriminatory of each material using chemometrics. Five high explosives and five organics containing benzene rings are measured in an air environment using nanosecond LIPS. Six different regions of the spectrum are subsequently obtained using a six-channel spectrometer, and then k-means clustering analysis in two-dimensional principal component space is used for each wavelength band to reveal the greatest spectral disparities. It is speculated that more carbon atoms in the functional groups tend to participate in the formation of the C₂ emission, while the carbon in the carbon ring tends to contribute more to the formation of the CN emission. Additionally, the quantitative correlations between C₂ emission intensity, oxygen balance (OB), and aroma index are also clarified for the first time. Finally, well-performing classification models are established using well-chosen spectral features with a supervised quadratic discriminant classifier (QDC) and linear support vector machine (SVM) with both specificity and accuracy of above 99%. This work proposes that nanosecond LIPS can be used as a chemical probe to detect structural discrepancies among organic samples and thus distinguish them well.