Explosives in fingerprints on packaging by LIBS

Scientists in Spain have improved the accuracy and selectivity of laser-induced breakdown spectroscopy (LIBS) for distinguishing between different explosives left in fingerprints on packaging. This technique has been proposed in the past but suffers from poor selectivity due to an inability to assign the correct structures to the spectral emissions. It has been improved using neural networks to develop a training set which can then be applied to unknown samples as reported in Talanta.

Six real explosives (DNT, TNT, RDX, PETN, chloratite and ammonal) were tested along with 17 harmless products that share some characteristics with the explosives and might be found as contaminants on packaging surfaces. These included hand cream, butter, soap, sugar, sweetener, fuel oil and motor oil, as well as some common chemicals like potassium and sodium chloride and potassium and sodium perchlorate. They were deposited as fingerprints from lab personnel onto three types of packaging materials: a cardboard box, a padded brown envelope and a white envelope. LIBS was carried out with a Nd:YAG laser at 1064 nm and the emitted light was monitored.

The spectra contain patterns that were used to classify the substances via decision trees based on three internal nodes. A substance was sorted according to its hazardous nature, ie whether it is chlorite, ammonal, an organic explosive, or a harmful substance. The system is not completely accurate, but with false positives and false negatives of up to 10%, it represents a major improvement in accuracy.

The research team, from the University of Malaga, is currently extending the scope of the study to take in more harmless materials and substances found encountered in everyday life.