Comparison of spectrum normalization techniques for univariate analysis of stainless steel by laser-induced breakdown spectroscopy

Analytical performance of six different spectrum normalization techniques, namely internal normalization, normalization with total light, normalization with background along with their three-point smoothing methods were studied using LIBS for quantification of Cr, Mn and Ni in stainless steel. Optimization of the number of laser shots per spectrum was carried out to obtain the best analytical results. Internal normalization technique model was used for selecting the best emission lines having sufficient intensity and spectral purity for Cr, Mn and Ni for comparison of different normalization techniques. For detailed evaluation of these normalization techniques, under optimized experimental conditions, three statistical parameters i.e., standard error of prediction, relative standard deviation and average bias, were compared for these techniques using the selected emission lines. Results show that the internal normalization technique produces the best analytical results followed by total light normalization. The smoothing of the raw spectra reduces the random error and produces better analytical results provided the peak under study has sufficient (a parts per thousand yen7) number of pixels.