飞秒激光诱导击穿光谱对铀多金属矿石中铀的定量分析

Background] Uranium polymetallic ores are strategic emerging industry minerals, and the quantitative analysis of uranium (U) in ores is a crucial step in the development process of such ores. [Purpose] In this paper, a quantitative analysis of U in uranium polymetallic ores using femtosecond laser-induced breakdown spectroscopy (LIBS) combined with partial least squares regression (PLSR) modeling was conducted. [Methods] Initially, the concentrations of U in six samples were measured by high-purity germanium gamma spectrometer and set as reference values. Then femtosecond lasers were employed to ablate the samples to obtain LIBS spectra. Subsequently, two normalization methods were applied to preprocess the raw spectra, and the impact of the preprocessed spectra on PLS model prediction analysis was compared with that of the raw spectra. Subsequently, the spectral data from five sets of samples were utilized as a training dataset to construct a quantitative model, enabling the analysis and prediction of U concentration in sample 3# [Results] The results indicate that the relative standard deviation (RSD) and mean relative error (MRE) of 10 predicted values for sample 3# are merely 5.94% and 4.73%, respectively. [Conclusions] This demonstrates that the combination of PLSR algorithm and femtosecond LIBS exhibits high analytical accuracy and stability for samples. This finding can provide significant reference for practical applications of femtosecond LIBS technology in the quantitative analysis of U content in uranium polymetallic ores.