Premium
Quantitative analysis of polymorphic mixtures of carbamazepine by Raman spectroscopy and principal components analysis
Author(s) -
Strachan Clare J.,
Pratiwi Destari,
Gordon Keith C.,
Rades Thomas
Publication year - 2004
Publication title -
journal of raman spectroscopy
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.748
H-Index - 110
eISSN - 1097-4555
pISSN - 0377-0486
DOI - 10.1002/jrs.1140
Subject(s) - raman spectroscopy , wavenumber , principal component analysis , analytical chemistry (journal) , chemistry , quantitative analysis (chemistry) , spectral line , spectroscopy , mathematics , chromatography , statistics , physics , optics , quantum mechanics , astronomy
Abstract The polymorphic behaviour of drugs is a major concern of the pharmaceutical industry as it may have considerable formulation, therapeutic, legal and commercial implications. It is therefore crucial to be able adequately to identify and quantify different polymorphic forms of drugs as early as possible in the drug discovery and development process. The aim of the present study was to investigate the application of Raman spectroscopy and principal components analysis (PCA) to the quantitative analysis of polymorphic mixtures of carbamazepine forms III and I, based on both individual peak areas and heights in the Raman spectra and whole wavenumber ranges. Various linear regressions indicated that in general using every fourth wavenumber value of a larger region of the Raman spectrum gives better quantitative results than choosing defined peak areas and peak heights, as the spectral differences between the polymorphs were only subtle. Baseline correction of the spectral data did not appear to improve the quantitative analysis, with detection and quantification limits being the lowest (0.33 and 1.09%, respectively) and R 2 being the highest (0.96) when using every fourth wavenumber value from 2950 to 3100 cm −1 as input data for the PCA analysis without baseline correction. This study demonstrated that PCA of Raman spectroscopic data provides a sensitive method for the quantitative analysis of polymorphic forms, with a quantitation limit of <2%, if the spectral input data are selected carefully. In the absence of clear and distinct spectral differences at certain wavenumbers between the polymorphic forms, selection of larger spectral regions that contain most of the detectable differences between the two forms appears to be the best data selection technique. Copyright © 2004 John Wiley & Sons, Ltd.