Abstract

Laser-induced breakdown spectroscopy (LIBS) produces line spectra from solid samples without any pre-treatment. The spectra were reduced to a matrix of intensities for the most intense lines and a number of different methods of identification were applied to determine if metal alloys could be distinguished by their spectra. The methods included principal component analysis, cluster analysis, multiple discriminant analysis, and spectral matching with a similarity index. The unsupervised methods, principal component analysis and cluster analysis, showed that the samples can be divided into groups based on their LIBS spectra. Discriminant analysis achieved an overall 97% correct classification into metal groups. Finally, a spectral matching approach was used to compare 234 individual spectra to a library of 39 average spectra. The accuracy of the matches varied from 97.4% correct prediction of the class of material to 79.9% correct identification of the specific alloy.