Energy dispersive X-ray diffraction to identify explosive substances : spectra analysis procedure optimization

To detect the presence of explosives in packages, automated systems are required. Energy dispersive X-ray diffraction (EDXRD) represents a powerful non invasive tool providing information on the atomic structure of samples. In this paper, EDXRD is investigated as a suitable technique for explosive detection and identification. To this end, a database has been constructed, containing measured X-ray diffraction spectra of several explosives and common materials. In order to quantify spectral resolution influence, this procedure is repeated with two different detectors which have different spectral resolution. Using our database, some standard spectrum analysis procedures generally used for this application have been implemented. Regarding to the results, it's possible to conclude on the robustness and the limits of each analysis procedure. The aim of this work is to define a robust and efficient sequence of EDXRD spectra analysis to discriminate explosive substances. Since our explosive substances are crystalline, the first step consists in using characteristic of the spectrum to estimate a crystallinity criterion which allows to remove a large part of common materials. The second step is a more detailed analysis, it consists in using similarity criterion and major peaks location to differentiate explosive from crystalline common materials. The influence of the spectral resolution on the detection is also examined.