Making Reference Samples Redundant

Abstract
Chemometric methods to analyze spectroscopic data without using reference spectra are discussed. The data are first decomposed into principal components, and the number of contributing species is determined by statistical tests. The principal components are then rotated to produce spectroscopic responses and concentration profiles of the chemical species present. Samples that vary in a physical property like pH, total concentration, temperature, or ionic strength, are analyzed by regular 1-dimensional spectroscopy assuming that the components are in chemical equilibrium. Samples containing noninteracting compounds are analyzed by multidimensional spectroscopy, and the principal components are calculated by Procrustes rotation. Several applications of the two approaches on absorption and fluorescence data are presented.

This publication has 0 references indexed in Scilit: