Laser-induced Breakdown Spectroscopy for Quantitative Analysis of Multi-Target Elements in Uranium Ore

Uranium is one of the most important strategic nuclear resources, which plays an important role in national strategy and nuclear energy. Therefore, it is significant to explore uranium ore precisely and smelt effective. Laser-induced breakdown spectroscopy (LIBS) is a process analytical technology, which is good at quantitative analysis of multiple elements quickly and in situ. It has unique advantages in detection of multiple elements in uranium ore. How to break through the problem of accurate quantification is still the key research content of LIBS technology. In this work, the quantitative models of U, Si, Al and Ti in uranium ore were established, based on univariate analysis (UVA), principal component regression (PCR) and support vector regression (SVR). The predictive coefficients (R2) of the two multivariate models for U, Si, Al and Ti were all better than 0.99. The quantitative accuracy of the SVR model for Si, Al and Ti was better, and the relative errors (Er) were 0.01%, 1.41% and 0.13%, respectively. The quantitative accuracy of the PCR model for U was the best, with the minimum Er of 0.64%. Compared with UVA, multivariate analysis (MVA) could overcome the quantitative bias caused by spectral line interference and matrix effect. Hence, it made the LIBS technique more accurate in quantification of multiple elements in uranium ore.