Nnls and mcr-als decomposition of raman and ftir spectra of multicomponent liquid solutions

Abstract

The possibilities of the decomposition by non-negative least squares (NNLS) and Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS) methods of determination of spectra and concentrations of species in multicomponent (up to 8) liquid solutions by analyzing FTIR ATR and Raman spectra are discussed. It is shown that the NNLS decomposition of Raman spectra provides better results than the decomposition of the same mixture measured with the FTIR ATR method. This fact is connected with narrower width and smaller overlap of vibrational peaks obtained with the use of Raman spectroscopy comparing to IR spectroscopy for molecular liquids. It is shown that using the known spectra can increase the total number of species, for which a quantitative information can be successfully obtained. The unknown concentration profiles of an N = 8-component liquid mixture can be found correctly if the N − 1 spectra of pure components are used as spectral constrains during the MCR-ALS analysis.