Gampp H, Maeder M, Meyer C J, Zuberbühler A D
Institute of Inorganic Chemistry, University of Basel, CH-4056 Basel, Switzerland.
Talanta. 1985 Feb;32(2):95-101. doi: 10.1016/0039-9140(85)80035-7.
Multiwavelength spectrophotometric and spectroscopic data in general contain considerably more information about complexation equilibria than potentiometric data do. With the construction of a fully automatic titration set-up built into a high-precision spectrophotometer, the problems related to the wider use of this method have shifted from the quality of the primary data to the complexity of their numerical treatment. Matrix algebra is used to show how these problems can be overcome. An algorithm is described for calculation of stability constants and absorption spectra, together with the associated standard errors, at a reasonable expense of computer time. Problems in finding the minimum in a multidimensional parameter space are reduced by elimination of the molar absorptivities from the algorithm for the iterative refinement. Numerical safety and speed of calculation are improved by use of analytical instead of numerical derivatives. The number of data to be fitted is decreased by principal-component analysis.
一般来说,多波长分光光度法和光谱学数据比电位法数据包含更多关于络合平衡的信息。随着高精度分光光度计内置全自动滴定装置的构建,与该方法更广泛应用相关的问题已从原始数据质量转移到其数值处理的复杂性。使用矩阵代数来展示如何克服这些问题。描述了一种算法,用于在合理的计算机时间成本下计算稳定常数和吸收光谱以及相关的标准误差。通过在迭代精化算法中消除摩尔吸光系数,减少了在多维参数空间中寻找最小值的问题。使用解析导数而非数值导数提高了计算的数值安全性和速度。通过主成分分析减少了要拟合的数据数量。