Chen Xing-Long, Fu Hong-Bo, Wang Jing-Ge, Ni Zhi-Bo, He Wen-Gan, Xu Jun, Dong Rui-Zhong
Guang Pu Xue Yu Guang Pu Fen Xi. 2014 Nov;34(11):3100-3.
Most quantitative models used in laser-induced breakdown spectroscopy (LIBS) are based on the hypothesis that laser-induced plasma approaches the state of local thermal equilibrium (LTE). However, the local equilibrium is possible only at a specific time segment during the evolution. As the populations of each energy level does not follow Boltzmann distribution in non-LTE condition, those quantitative models using single spectral line would be inaccurate. A multivariate nonlinear model, in which the LTE is not required, was proposed in this article to reduce the signal fluctuation and improve the accuracy of quantitative analysis. This multivariate nonlinear model was compared with the internal calibration model which is based on the LTE condition. The content of Mn in steel samples was determined by using the two models, respectively. A minor error and a minor relative standard deviation (RSD) were observed in multivariate nonlinear model. This result demonstrates that multivariate nonlinear model can improve measurement accuracy and repeatability.
激光诱导击穿光谱（LIBS）中使用的大多数定量模型基于激光诱导等离子体接近局部热平衡（LTE）状态的假设。然而，局部平衡仅在演化过程中的特定时间段内才有可能。由于在非LTE条件下每个能级的粒子数不遵循玻尔兹曼分布，那些使用单条谱线的定量模型将不准确。本文提出了一种不需要LTE的多元非线性模型，以减少信号波动并提高定量分析的准确性。将该多元非线性模型与基于LTE条件的内部校准模型进行了比较。分别使用这两种模型测定了钢样中锰的含量。在多元非线性模型中观察到较小的误差和较小的相对标准偏差（RSD）。这一结果表明，多元非线性模型可以提高测量精度和重复性。
