Improvement of sample discrimination using laser-induced breakdown spectroscopy with multiple-setting spectra
Yuzhou Song,Weiran Song,Xiang Yu,Muhammad Sher Afgan,Jiacen Liu,Weilun Gu,Zongyu Hou,Zhe Wang,Zheng Li,Gangyao Yan,Qing Ye,Zijun Liu,Hongqi Zheng,Junsheng Fan,Yuchun Yu,Liang Li
DOI: https://doi.org/10.1016/j.aca.2021.339053
IF: 6.911
2021-11-01
Analytica Chimica Acta
Abstract:Laser-induced breakdown spectroscopy (LIBS) is a promising multi-elemental analysis technique and has the advantages of rapidness and minimal sample preparation. In traditional LIBS measurement, sample spectra are generally collected based on a single set of fixed experimental parameters, such as laser energy and delay time. When samples have the same main components and similar component concentrations, the difference in their spectral intensities becomes less obvious. This can lower the sensitivity of LIBS measurement and pose a threat to the accuracy and robustness of LIBS qualitative analysis. In this work, we propose a new method to increase the spectral difference between similar samples, namely multiple-setting spectra. For each sample, it adopts different sets of experimental parameters and obtains a group of spectra to increase the fingerprint spectral information. The effectiveness of the proposed method is theoretically verified and then tested on 11 similar coal samples. Specifically, the sample spectra were collected with different laser energy and delay time, and processed by principal component analysis (PCA) and Davies-Bouldin index (DBI). The results show that the use of multiple-settings spectra can significantly improve the sample discrimination accuracy from 81.8% to 96.4%. In addition, the proposed method can maintain the efficiency and cost of LIBS measurement.
chemistry, analytical