JOURNAL OF RAMAN SPECTROSCOPY J. Raman Spectrosc. 2003; 34: 282–289 Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jrs.986 Fourier transform Raman spectral reconstruction of inorganic lead mixtures using a novel band-target entropy minimization (BTEM) method Lih Ren Ong, Effendi Widjaja, Robert Stanforth and Marc Garland * Department of Chemical and Environmental Engineering, 4 Engineering Drive 4, National University of Singapore, Singapore 119260 Received 6 June 2002; Accepted 5 December 2002 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  2003 John Wiley & Sons, Ltd. KEYWORDS: pure component spectral reconstruction; singular value decomposition; entropy minimization; lead compounds; Fourier transform Raman spectroscopy INTRODUCTION A novel pure component spectral reconstruction algorithm called band-target entropy minimization (BTEM) has been highly successful in recovering the Fourier transform (FT) IR spectra of the constituent organometallic species in solution. 1 This method has been further applied to the rhodium-catalyzed hydroformylation of alkenes, where all the anticipated major and minor component FTIR spectra were also reconstructed. 2 The BTEM algorithm has the advantage of achieving accurate spectral reconstructions without recourse to any libraries or a priori knowledge. More important, it is sensitive enough to recover pure component spectra of species possessing very low absorbance due to their low concentrations. The key features of the algorithm are (i) the targeting of a local feature in the vector-space L Correspondence to: Marc Garland, Department of Chemical and Environmental Engineering, 4 Engineering Drive 4, National University of Singapore, Singapore 119260. E-mail: chemvg@nus.edu.sg decomposition of the mixture spectral data and (ii) the reconstruction of the simplest function associated with the targeted local spectral feature, namely the underlying pure component spectrum. Inorganic lead compounds are toxic contaminants found commonly in soil and industrial wastes. Lead is carcinogenic and is known to inhibit the mental development of children. 3 It can also accumulate in human bones, releasing slowly over time to cause a host of health problems. Although the use of lead has drastically decreased in many areas of the world as the use of leaded gasoline and leaded paint has been phased out, the problem of lead pollution remains significant in many regions. In this work, the applicability of the BTEM algorithm to recover the pure component spectra of inorganic lead compounds in mixtures measured using FT Raman spec- troscopy was evaluated. Typical lead compounds such as lead arsenate, lead oxide (PbO), lead carbonate (PbCO 3 ), lead hydroxide [Pb(OH) 2 ] and pyromorphite [Pb 5 (PO 4 ) 3 OH] were combined in varying proportions to form two sets Copyright  2003 John Wiley & Sons, Ltd.