Premium
Plasma spectroscopy + chemometrics: An ideal approach for the spectrochemical analysis of iron phosphate glass samples
Author(s) -
Devangad Praveen,
V K Unnikrishnan,
M. Yogesha,
Kulkarni Suresh D.,
Chidangil Santhosh
Publication year - 2020
Publication title -
journal of chemometrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.47
H-Index - 92
eISSN - 1099-128X
pISSN - 0886-9383
DOI - 10.1002/cem.3310
Subject(s) - chemometrics , partial least squares regression , principal component analysis , calibration , linear discriminant analysis , support vector machine , analytical chemistry (journal) , mean squared error , analyte , chemistry , laser induced breakdown spectroscopy , spectroscopy , biological system , artificial intelligence , chromatography , mathematics , computer science , statistics , physics , quantum mechanics , biology
The penultimate aim of all analytical techniques is to provide best quantitative information with high sensitivity and accuracy. Such techniques then can be ideal to estimate the concentration of different species in a sample or to measure the surface concentration and so on. The preliminary step involved in quantitative analysis is evaluation of the system response for a given amount of the analyte under investigation. Different methodologies are proposed and executed for the above mentioned purpose. In this work, we have demonstrated the performance of chemometric methods to study the atomic emissions from simulated nuclear waste glasses. Five samples of iron‐phosphate glasses were synthesized by doping Cr, Sr, and Ti at various concentrations. The spectra recorded using laser‐induced breakdown spectroscopy (LIBS) system were analyzed using principal component analysis (PCA), partial least square‐discriminant analysis (PLS‐DA), and support vector machines (SVM) for the classification. As compared with the PLS‐DA, SVM provided better results for classification with accuracy of 100% on both calibration and validation sets, respectively. The quantitative analysis of glass samples is carried out using partial least square regression (PLSR) and support vector regression (SVR). The root mean squared error of prediction (RMSEP) is found to be 0.16, 0.20, and 0.08 (wt%) for Cr, Sr, and Ti using PLSR. The detailed investigation elucidates the advantages of chemometrics in handling the complex LIBS spectrum of glass samples and their significance in process control of nuclear waste, nuclear forensics, and so on.