ANALYSIS OF COMPLEX SPECTRA VIA VISIBLE S PECTROSCOPY - DECONVOLUTION, MAXIMUM-ENTROPY METHOD AND LINEAR PREDICTION

The analysis of spectrum change in visible spectroscopy is a common a approach in photobiology research. This analysis is made difficult by the overlapping of spectral bands which is due not to lack of instrume ntal resolution but rather to intrinsic properties of the examined bio logical material. To minimize the overlap bands, an FT-IR spectroscopy method called Line Shape Optimized Maximum Entropy Linear Prediction has been proposed. This method is based on the spectral deconvolution, maximum entropy and linear prediction methods (J.K. Kauppinen et al, Appl. Spectrosc., 45, 411, 1991). We present a model study which for t he first time enables the use of the deconvolution, method and autoreg ression in visible spectroscopy By analyzing the phototransformation k inetics of pigment-protein complexes in etiolated barley leaves, we de monstrate the advantages of this new analytical approach compared to t he methods used to date (D.G. Cameron et al., 41, 539, 1987; R.D.B. Fr aser et al., Anal. Chem., 38, 1770, 1966).