(660b) Improving Product Quality and Process Robustness with PAT-Guided Crystallization Process Development

During the isolation of an intermediate compound and to purge a key impurity, a fit-for-purpose crystallization process was developed. However, the process exhibited high sensitivity to input material variability such as seeds property and lacked the desired robustness and most importantly, poor filtration performance. Through a combination of offline microscopy, online monitoring, kinetic measurements, and process modeling, a new crystallization process was developed to improve the filtration performance and polymorph control during scale up.

Through lab scale crystallization experiments, it was discovered that the original crystallization process suffered from a very slow growth kinetics where the seed growth depended significantly on the seed properties. To improve the growth kinetics, desaturation was monitored using Crystalline or/and EasyViewer. It was identified that by shifting the solvent composition at the seed point, desaturation can occur much faster for a wide range of seed particle sizes.

Although the form was shown to be stable during slurry stability experiments, it was discovered later on through imaging that a new polymorph co-exists under certain solvent composition and temperature as a result of primary nucleation. This polymorph reduced product quality as well as filtration performance. Through analysis of EasyViewer results, the process was changed from cool-then-add to add-then-cool to avoid the risk of the undesired polymorph.

Finally, temperature cycles were added to the new crystallization process to increase the particle size and thus improve the filtration rate. The filtration performance is modeled and characterized with Dynochem and the cake resistance and compressibility were used to guide a robust scale up strategy and equipment selection.