SELF-MODELING CURVE RESOLUTION IN STUDIES OF SPECTROMETRIC TITRATIONS OF MULTI-EQUILIBRIA SYSTEMS BY FACTOR-ANALYSIS

Recently developed techniques of factor analysis are incorporated in a revised algorithm for resolving the spectra and concentration profiles of the species formed during a spectrometric titration. The techniques included here are: (1) sounder statistical tests and cross-validation techniques for estimating the number of species present in multi-equilibria systems; (2) target factor analysis of the spectra of suspected species; and (3) rank-annihilation estimation of the concentration profile of a detected species with known spectrum and subtraction of its spectral contribution from the complete data matrix. Although these methods can be applied to complement traditional least-squares methods, they can also be used quite independently in the spectrometric study of multi-equilibria systems in solution, without initial chemical modelling.