人工神经网络分光光度法同时测定锆和铬的研究

在锆 (铬 ) -二溴羟基苯基荧光酮 - Triton X- 1 0 0显色体系中 ,应用三层人工神经网络解析锆和铬的吸收光谱 ,分光光度法同时测定锆和铬。锆和铬分别在 0～ 0 .48mg/ L和 0～ 1 .5 6 mg/ L浓度范围内服从 Beer定律 ,其表观摩尔吸光系数分别为ε53 4=1 .48× 1 0 5L/ (mol· cm) ,ε50 0 =2 .3×1 0 4 L/ (mol·cm)。使用改进的 BP算法 ,避免了人工神经网络学习过程中可能产生的麻痹现象。对合成样品分析 ,结果满意。

Study on Simultaneous Determination of Zirconium and Chromium by Artifical Neural Network Spectrophotometry

A three-layer artificial neural network (ANN) was applied to simultaneous determination of zirconium and chromium by spectrophotometry. The conditions of simultaneous determination by color reaction of Zr(Ⅵ), Cr(Ⅵ) with dibromo-hydroxyphenyl fluorone in the presence of Triton X-100 were investigated. The apparent molar absorbance of Zr(Ⅵ) and Cr(Ⅵ) complexes are found to be ε 534 =1.48×10 5 L/(mol·cm), ε 500 =2.3×10 4 L/(mol·cm), respectively. Beer's law is obeyed in the concen-tration range of 0～0.48 mg/L and 0～1.56 mg/L, respectively. The paralysis in the procedure of training was avoided with an improved back propagation algorithms. This method was applied to the determination of the synthetic samples with satisfactory results.