Stability and accuracy improvement of element analysis in steel alloys using polarization-resolved laser-induced breakdown spectroscopy

The existence of the background signal is one of the factors leading to the uncertainty of spectral intensity, which dramatically affects the stability, repeatability, and quantitative accuracy of laser-induced breakdown spec-troscopy (LIBS) measurements. In this work, the polarization-resolved LIBS (PR-LIBS) and partial least squares (PLS) regression models were proposed to improve the stability and accuracy of qualitative and quantitative analysis of steel alloy elements. The obtained spectra were pre-processed to select an appropriate number of variables, and the relationship between the spectral intensity and the certified concentration information of steel alloy samples was modeled using PLS. Compared to LIBS measurements, the relative standard deviation (RSD) and root-mean-square error (RMSE) of PR-LIBS decreased, while the signal-to-background ratio (SBR) and co-efficient of determination (R2) increased. Finally, the quantitative results of LIBS and PR-LIBS of Si, Mn, Cr, and Ni in steel alloy samples using PLS were displayed and discussed. For the PR-LIBS dataset, the R2 has increased by at least 3.4%, and the root mean square error of calibration (RMSEC) and root mean square error of prediction (RMSEP) have improved by at least 6.3% and 19.0%, respectively. These findings coupled with the simplicity of the experimental process indicate that PR-LIBS is a simple and low-cost solution that can increase the stability and accuracy of LIBS measurements.

Automated summary

The polarization-resolved LIBS (PR-LIBS) and partial least squares (PLS) regression models were proposed to improve the stability and accuracy of qualitative and quantitative analysis of steel alloy elements. The relationship between the spectral intensity and the certified concentration information of steel alloy samples was modeled using PLS. PR-LIBS showed decreased relative standard deviation (RSD) and root-mean-square error (RMSE), and increased signal-to-background ratio (SBR) and co-efficient of determination (R2) compared to LIBS measurements. Quantitative results of Si, Mn, Cr, and Ni in steel alloy samples using PLS were displayed and discussed. For the PR-LIBS dataset, R2 has increased by at least 3.4%, and the root mean square error of calibration (RMSEC) and root mean square error of prediction (RMSEP) have improved by at least 6.3% and 19.0%, respectively. PR-LIBS is a simple and low-cost solution that can increase the stability and accuracy of LIBS measurements.