Polyolefin formulation validation using infrared spectrometric classification techniques

Discriminant analysis using principal components of mid‐infrared spectral data is demonstrated as a powerful quality validation tool to augment sophisticated manufacturing processes where rigorous multicomponent analysis is not only costly but in many cases unwarranted. In order to balance competing requirements of validation, a multiple model approach is taken, which combines the best information available about the process, the chemistry and the spectroscopy of the formulation. A systematic approach to evaluate training data is discussed as a crucial step to insuring a diverse but homogenous calibration.