Discrimination of crystal polymorphism in active pharmaceutical ingredients using time-domain 1H NMR relaxation combined with multivariate statistical process control.

This study investigated the feasibility of evaluating crystal polymorphs of active pharmaceutical ingredients (APIs) using 1H NMR relaxation measurements with time-domain NMR (TD-NMR). Desloratadine, linagliptin, and carbamazepine, each exhibiting multiple polymorphic forms, were selected as model APIs. T1 relaxation times were measured, and the results showed that T1 relaxation behavior differed for each crystal form of the APIs. For desloratadine, an amorphous sample was also prepared and evaluated. It was demonstrated that the T1 relaxation time of the amorphous form was shorter than those of the crystal polymorphs, indicating higher molecular mobility. T1 relaxation curves were further analyzed using multivariate statistical process control (MSPC) based on principal component analysis, and Hotelling's T² and Q statistics successfully distinguished different crystalline forms. In the latter part of this study, physical mixtures of crystal polymorphs were tested to evaluate the detection sensitivity of the MSPC approach. The results demonstrated that this method could identify polymorphic admixtures with a detection limit of approximately 5-10%, depending on the API. TD-NMR offers practical advantages, including rapid measurement, minimal sample preparation, and non-destructive analysis using low-field benchtop instruments. These results indicate that TD-NMR combined with MSPC provides a practical at-line method for monitoring crystal polymorphism and detecting polymorphic contamination during pharmaceutical manufacturing.