Identification of Solid Materials by Correlation Analysis Using a Microscopic Laser-Induced Plasma Spectrometer

The primary goal of this work was the instant identification of solid materials on the basis of spectral libraries stored in a computer using laser-induced plasma spectroscopy. The libraries were obtained prior to analysis and consisted of representative spectra from different groups of materials to be analyzed. Special attention was paid to identification of samples with very similar chemical composition, such as certain series of stainless steel and cast iron standards. Both linear and rank correlation methods were applied. Rank correlation proved to be more reliable, yielding probability of correct identification close to unity for almost all studied samples. This technique should have applications in the metallurgical, mineralogical, and semiconductor industries and in medical and forensic sciences.