Fourier transform Raman spectral reconstruction of inorganic lead mixtures using a novel band-target entropy minimization (BTEM) method

A novel method of band-target entropy minimization (BTEM) was successfully applied to spectral reconstruction of two sets of solid mixtures containing different species of lead compounds which were measured using Fourier transform (FT) Raman spectroscopy. As a first step, bands or regions containing significant spectral features in the vector-space decomposition of the observations were identified and selected to untangle the superposition of the pure component spectra. The targeted bands were retained and the pure component spectra of lead oxide and lead arsenate, both common in the two sets, lead carbonate (in Set 1) and lead hydroxide and pyromorphite (both in Set 2) were all readily recovered. The BTEM algorithm results in (i) a fast computational speed since one spectrum is generated at a time and (ii) high-quality spectral reconstructions. The algorithm allows maximum chemical information recovery from experimental data by utilizing more vectors in the vector-space decomposition than the actual number of species present in the observations. The results further indicate that the BTEM algorithm is viable for detecting species at low concentrations or which have low intensity. Most important, there is no recourse to any libraries and no a priori information is required for spectral reconstruction. This approach holds promise for the rapid analysis and identification of solid mixtures by FT Raman spectroscopy. Copyright (C) 2003 John Wiley Sons, Ltd.