Evolving Factor Analysis of Spectrophotometric Titrations: Forget About the Law of Mass Action?

Abstract

Evolving Factor Analysis (EFA) is a new and powerful mathematical algorithm for the model-free analysis of any ordered set of multiwavelength data. Applied to spectrophotometric titrations, even strongly overlapping concentration profiles and the spectra of all absorbing species are correctly calculated without making use of the law of mass action and without even defining the stoichiometric composition of the species in question. The basic idea behind EFA is repetitive application of factor analysis to all subsets of 1, 2, ...N measurements and iterative refinement of the eigenvalues thus obtained into concentration profiles with concomitant least-squares estimation of the species spectra. EFA has been tested successfully with a dozen chemical systems as well as with model data for strongly overlapping concentration profiles.