Improvement in the analytical performance of underwater LIBS signals by exploiting the plasma image information

• Published in: Journal of analytical atomic spectrometry Vol. 35; no. 2; pp. 366 - 376
• Main Authors: Li, Qingyang, Tian, Ye, Xue, Boyang, Li, Nan, Ye, Wangquan, Lu, Yuan, Zheng, Ronger
• Format: Journal Article
• Language: English
• Published: London Royal Society of Chemistry 12.02.2020
• Subjects: Correlation analysis; Laser plasmas; Line spectra; Regression analysis; Underwater; Variation
• ISSN: 0267-9477; 1364-5544
• DOI: 10.1039/c9ja00367c

Laser-induced plasma in water always suffers from strong pulse-to-pulse fluctuations due to the multiple breakdown phenomenon, leading to a poor stability of underwater LIBS signals. The traditional normalization method by using the internal standard element is often limited in some practical cases due to the lack of a suitable element as a reference. In this work, we developed an effective normalization method by using the plasma image information for underwater LIBS analysis. Correlations between the plasma images and LIBS spectra were firstly studied, showing a good linear relationship between the spectral line intensity and plasma image intensity. Subsequently, the spectral line intensities were standardized by using the corresponding image intensities and then used for quantitative analysis. A good normalization model was established by using partial least squares regression (PLSR). With the proposed method, the average relative standard deviations (RSDs) of validation samples were significantly reduced from 10.71% to 5.76%, and the average relative errors (AREs) of the validation samples were also reduced from 7.80% to 7.55%. Moreover, by combining the proposed method with the internal standard method, the average RSD and ARE can be further reduced to 4.07% and 4.86%, respectively, both of which are better than those obtained using the internal standard method only. A normalization method for underwater LIBS signals by using a synchronous plasma image is proposed and used for quantitative analysis.