A New Approach for Material Classification Based on Quantitative Analysis

Accurate identification and classification of a material based on energy dispersive spectrometry (EDS) has always been a challenge, down to the variability between systems, acquisition conditions, spectrum processing methods and accuracy in quantitative analysis results. Traditional methods rely on template spectra from known chemical compositions to match against an unknown spectrum. The spectrum matching algorithm assumes that template spectra are acquired with the same system and under the same conditions, which poses difficulties and in some cases is impossible. In this paper, a different approach is proposed based on quantitative analysis results as the basis for matching. Quantification-based matching removes the ambiguity introduced by different systems and acquisition conditions. Databases can be created based on the nominal composition of compounds without acquiring a spectrum from known samples. This paper explains the details of the matching algorithm and presents three examples to demonstrate the performance.