| Library search routines allow rapid identification, or conversely discrimination of target species. The ramifications to throughput and efficiency make these methods attractive over a diverse field of interest, such as forensics, industrial process control, and pharmaceutical development. Raman spectroscopy adds advantage when visible light excitation and no inherent sample preparation are considered. This study employs a spectral library for the analysis and deconvolution of white powder mixtures. Broadband backgrounds inherent to Raman spectra of white powders are removed via an automated process that does not corrupt the spectral information. This study incorporates important aspects of instrument response correction to the collection of a library. The major components of unknown mixtures are identified from their Raman spectra by use of a fuzzy logic and limited zone least squares (LZLS) algorithm. This protocol is shown to effectively extract components from mixture spectra in a stepwise fashion. Finally, we address the question of whether library references can be used as quantitative models of the qualitatively identified components. |
