The construction of near-infrared spectral libraries as an alternative to qualitative analysis methods for identifying pharmaceutical raw materials is proposed. Various conceptual and practical aspects of library construction are assessed and discussed. The procedure is demonstrated by constructing a library including NIR spectra for 125 different raw materials using the correlation coefficient as the discriminating criterion. Compounds with very similar spectra can be identified by constructing sub-cascading libraries branching off the main one that are developed by using chemometric procedures with higher discriminating ability. The construction of sub-libraries and their performance and discriminating power in three different situations are illustrated. The proposed methodology affords the expeditious unequivocal identification of all the compounds included in a library.
|Publication status||Published - 1 Dec 2001|