建立了一种新的稳健回归方法,用于解析显色剂二溴对甲偶氮甲磺(DBM-MSA)与铅、铈配合物的重叠吸收光谱,不经分离分光光度法同时测定了铅和铈.铅和铈的配合物最大吸收波长分别为640和642 nm,表观摩尔吸光系数分别为7.31×104和7.80×104L·mol-1·cm-1.铅和铈配合物分别在0～9 μg/25 mL和0-11 μg/25 mL范围内服从比耳定律.将人工神经网络BP算法(ANN-BP)应用于该体系的解析结果与新的稳健回归方法进行比较,表明两种方法用于铅和铈的同时测定,结果满意.

A chemometrics method of robust regression analysis (RRA) applied to spectrophotometric determination of components in a binary system is reported in this paper.Both Pb and Ce are reacted with dibromo-p-methyl-methylsulfonazo (DBM-MSA) to form color complexes,having similar absorption spectra which were seriously overlapped each other.The absorption peaks are located at 640 nm with molar absorptivity 7.31×104L·mol-1·cm-1 for Pb and at 642 nm with molar absorptivity 7.80×104L·mol-1·cm-1 for Ce,respectively.By using the RRA method,the absorption spectra were resolued and Pb and Ce determined simultaneously without separation.Beer′s law was obeyed in the range of 0-9 μg/25 mL for Pb and 0-11 μg/25 mL for Ce.The method has been applied to the analysis of simulated samples and SRM of steels,the results were compared to those obtained by the Artificial Neural Network-BP method,showing that both the methods gave consistent results.