Multiparametric curve fitting—IX : Simultaneous regression estimation of stoichiometry and stability constants of complexes

Abstract

A chemical model ( i.e., the number of complexes, their stoichiometry and stability constants with molar absorptivities) in solution equilibria may be established by ( i) the trial-and-error method in which stability constants are estimated for an assumed set of complexes in the mixture and a fitness test is used to resolve a choice of plausible models to find the true one; ( ii) the simultaneous estimation of the stoichiometry and stability constants for species divided into “certain” species for which the parameters β pqr , ( p, q, r) are known and held constant, and “uncertain” species with unknown parameters which are determined by regression analysis. The interdependence of stability constants and particular sets of stoichiometric indices requires that the computational strategy should be chosen carefully for each particular case. The benefits and limitations of both approaches are compared by means of examples of potentiometric titration data analysis by the POLET(84) program and of spectrophotometric data analysis by the SQUAD(84) program. A strategy for efficient computation is suggested.