Quantitative estimation of elemental composition employing a synthetic generated spectrum

Abstract

Composition of multielement samples is estimated by using a synthetic generated spectrum utilizing a nonlinear fitting routine. By fitting simultaneously a large number of emission lines, the error in the estimation is minimized. The procedure for synthetic spectrum generation includes self-absorption of emission lines by taking into account the number density of different species in the plasma. The Stark width of different emission lines is iteratively calculated based on the observed width and the degree of self-absorption in the individual lines. This procedure is found to be successful for achieving convergence of the retrieval algorithm even for dense spectra as well as for resolving merged lines with accuracy. This scheme was applied on a stainless steel and brass sample, and the laser-induced breakdown spectroscopy results match well with the reference value obtained from the electron probe microanalyzer measurement.