Laser-induced breakdown spectroscopy; Generalized extreme value distribution; shot-to-shot statistics ;
The purpose of this work is to provide detailed study of statistical behavior of different types of analytical signals in typical of Laser-Induced Breakdown Spectroscopy (LIBS) measurements. The main goal of this work is to justify usage of arithmetic mean and standard deviation as statistical estimates of expected value of selected analytical signal. In contrary to the general assumption that LIBS data show Gaussian distribution, this paper deals with the hypothesis that the data rather demonstrate Generalized Extreme Value Distribution. The study is realized on 10 selected lines measured on NIST glass standard. In order to cover wide range of possible applications three different spectra internal standardization techniques and their influence on distribution were studied. Finally, assuming that the data comes from a single distribution and the central limit theorem is valid, the influence of accumulations on the line distribution is examined and discussed. Statistical tools used and described in this paper can be utilized by other researchers to confirm their hypotheses and verify utilization of Gaussian distribution or even novel data processing methods.