Resolution of multicomponent mixture spectra in mid-infrared spectroscopy using spherical projection factor analysis: application to real data including a six-component mixture set

Abstract

Spherical projection factor analysis is a self-modelling curve resolution method used for the deconvolution of a set of mixture spectra into the pure component spectral and concentration profiles. A modification of the previously proposed method is described, consisting of four steps, namely principal components analysis, spherical projection including wavenumber weighting, location of composition one points and regression of composition one points. The method was applied to real and simulated mixture sets, including one which contained six components.