Plastic Reference Material (PRM) Combined With Partial Least Square (PLS) in Laser-Induced Breakdown Spectroscopy (LIBS) in the Field of Quantitative Elemental Analysis

作者：Wang Bin; Zheng Shao-feng; Gan Jiu-lin; Liu Shu; Li Wei-cai; Yang Zhong-min; Song Wu-yuan^*

来源：SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(7): 2124-2131.

DOI：10.3964/j.issn.1000-0593(2023)07-2124-08

• Plastic
• Reference materia
• Partial least squares
• Laser-induced breakdown spectroscopy

Laser-induced breakdown spectroscopy was widely used for quantitative elemental analysis of the field for its rapid, efficient, harmless, full spectrum advantage of direct reading and almost do not need sample preparation. In order to establish a simple method, developed by using polypropylene (PP) plastic reference material (RM), combined with partial least squares (PLS), using laser-induced breakdown spectroscopy (LIBS), the PRM-PLS-LIBS model of Pb and Cr was established. PP reference materials were developed by the requirements of international standards. The content gradient of Pb and Cr were set in the range of 0 similar to 1 000 mg center dot kg(-1) according to the limitation requirements of various countries and regions. The certification results combined with uncertainty gave the specific values, and the certified reference material had good homogeneity and stability. The correlation coefficients of the standard curves of Pb and Cr were 0. 999 2 and 0. 998 9, respectively, and the detection limits were 35 and 28 mg center dot kg(-1), respectively, which had reached the quantitative analysis ability. In order to improve the accuracy of the quantitative analysis model, it was necessary to optimize the data of standard curve further. Partial least squares (PLS) and classical least squares (CLS) were compared by multiple optimizations of baseline types for small data volume. Experimental results showed that the root mean square error of calibration (RMSEC) and correction coefficient (Corr. Coeff) of PLS were better than that of CLS. By optimizing the analytical wavelength range and baseline type of Pb and Cr, the correction coefficient between the given value and the predicted value of the correction curve reached 1. 000 0, which further improved the model's accuracy. A set of PP-certified reference materials were then selected to verify the calibration curve. High content samples (PP-306) and low-content samples (PP-302) were taken for determination, and Pb and Cr determination data were substituted into the PRM-PLS model. The Pb and Cr determination values in PP-306 were 998 and 96 mg center dot kg(-1), respectively. The determination values of Pb and Cr in PP-302 are 980 and 95 mg center dot kg(-1), respectively, within the given range. The method was effective and reliable.