ICP-AES测定高浓度基体下杂质元素的偏最小二乘法研究

建立了ICP-AES测定高浓度基体中微量杂质元素的偏最小二乘方法(PLS)。研究表明,PLS能有效校正高浓度基体干扰引起的测量误差,比多元光谱拟合法(MSF)能承受的基体浓度更高。当基体与杂质的含量比为1 000∶1～20 000∶1时,该方法的加标回收率在95%～105%之间。对于干扰效应与基体浓度呈非线性相关的体系,普通PLS的预测准确度不高,但使用基于样品浓度矩阵变换的偏最小二乘法(LIN-PPLS),则明显改善了预测的准确度。分别用MSF、普通PLS和LIN-PPLS对水系沉积物国家标准物质GBW07312中的Co,Pb和Ga进行测定,结果表明,LIN-PPLS的预测准确度优于普通PLS,而普通PLS的预测准确度优于MSF。

Research on partial least squares for determination of impurities in the presence of high concentration of matrix by ICP-AES

A method for detecting trace impurities in high concentration matrix by ICP-AES based on partial least squares (PLS) was established. The research showed that PLS could effectively correct the interference caused by high level of matrix concentration error and could withstand higher concentrations of matrix than multicomponent spectral fitting (MSF). When the mass ratios of matrix to impurities were from 1 000 : 1 to 20 000 : 1, the recoveries of standard addition were between 95% and 105% by PLS. For the system in which interference effect has nonlinear correlation with the matrix concentrations, the prediction accuracy of normal PLS method was poor, but it can be improved greatly by using LIN-PPLS, which was based on matrix transformation of sample concentration. The contents of Co, Pb and Ga in stream sediment (GBW07312) were detected by MSF, PLS and LIN-PPLS respectively. The results showed that the prediction accuracy of LIN-PPLS was better than PLS, and the prediction accuracy of PLS was better than MSF.