Application of Artificial Neural Networks for Analysis of Highly Overlapped and Disturbed Differential Pulse Polarographic Peaks in the Region of Hydrogen Evolution

Multivariate calibration based on a suitable experimental design (ED) and soft modelling with artificial neural networks (ANNs) is proposed for quantitative analysis of highly overlapped and disturbed differential pulse polarographic (DPP) peaks that occur in the region of a hydrogen evolution. It is demonstrated that analysis of mixtures, even if some of the constituents undergo an irreversible reduction and the background current varies significantly with a composition of a sample, can be quantified with reasonable accuracy using a combination of ED and ANNs. Examples of DPP examination of ZnII and CrIII mixtures and/or simultaneous determination of metal ions and a strong acid concentration are presented. The possibility of an on-line monitoring is suggested. It is demonstrated that standard hard model based refinement procedures perform much worse than ANNs combined with ED and, in principle, proved to be unsuitable for the purpose. South African Journal of Chemistry Vol.53(3) 2000: 213-23