Principal component analysis of FT-IR spectra for cationic photopolymerization of mixtures of two monomers

PCA models of the overall data set for different compositions are able to clearly separate compositional effects from time varying conversion effects. Multivariate curve resolution (MCR) on individual runs was found to provide no additional information due to rotational and intensity ambiguities that remain in spite of the imposition of different constraints. For PCA models built on runs with the same initial monomer composition ratios, loading plots and contribution plots are used to obtain a much clearer interpretation of the reactions involved in the polymerization. Furthermore, important structural differences occurring among “replicate” runs are uncovered providing further insight into the reaction mechanisms, and into reasons for batch-to-batch inconsistencies that may occur in a manufacturing process.