Factor Analysis and Least-Squares Curve-Fitting of Infrared Spectra: An Application to the Study of Phase Transitions in Organic Molecules

Abstract

The significance of eigenvectors and eigenspectra used in factor analysis, the relation of least-squares coefficients to the concentration of individual components, and the role of difference spectrum in the least-squares method have been examined theoretically. The equivalence of factor analysis and least-squares curve-fitting has been established. Factor analysis and least-squares curve-fitting have been attempted on the Fourier transform Infrared spectra of nonadecane, nonadecane-d2, nonadecane-d2 melt, and C24H49,COOH, obtained as a function of temperature. Experimental results have been used to substantiate theoretical conclusions.