(322e) Multistage, Multi-Zones Antisolvent-Cooling Crystallisation of a Proprietary API: Comparison of Two Continuous Crystalliser | AIChE

(322e) Multistage, Multi-Zones Antisolvent-Cooling Crystallisation of a Proprietary API: Comparison of Two Continuous Crystalliser


Ojo, E. - Presenter, University of Strathclyde
Siddique, H., University of Strathclyde
Houson, I., University of Strathclyde
O'Meadhra, R., Novartis
Anwar, A., University of Strathclyde
Schenkel, B., Novartis
Florence, A., University of Strathclyde
The study aimed to establish parallel antisolvent-cooling crystallisation of a proprietary active pharmaceutical ingredient (API) in (5-stages antisolvent additions and three temperature zones) continuous oscillatory baffled crystalliser (COBC) and mixed-suspension mixed-product removal (MSMPR) crystalliser with 1000 and 180 mL(s) for each stage respectively. Each stage comprises of known solvents-API compositions based on the provided solubility profile. Process platforms used were off shelf NITECH-COBC and CMAC-MSMPR multistage crystalliser. The crystallisation solvent mixture is a blend of two solvents namely – ethanol, tetrahydrofuran (THF) and water as antisolvent. The throughput in both systems was based on the matched residence time of approximately 6 hr. The three temperature zones of 40 ℃ (stages 1 – 3), 30 ℃ (stage 4) and 20 ℃ (stage 5) were spread across the stages as indicated. While this worked effectively with MSMPR, a linearised gradient was adopted for the COBC to minimise fouling at interstage nodes – a non-optimised condition. Temperatures, particle counts, chord length distributions, feed, seed and antisolvent flow rates, and blockages were consistently monitored throughout the continuous operations.

At matched residence times for the two crystallisers, the particle size distribution obtained from the COBC was bigger having enhanced flow properties, less agglomeration propensity, higher yield and lower surface area compared to MSMPR. In addition, slightly reduced impurity rejection was observed in the COBC; meanwhile, the MSMPR resulted in lesser yield but higher impurity rejection. In terms of platform operations, more interventions were required to keep the COBC running with significant impact on process steady-state and potential shutdown. Fouling and encrustation were observed in both systems but could be reduced if processing conditions are optimised for COBC. In conclusion, the COBC which was primarily built for flexibility and relatively new in terms of applications resulted in better crystal attributes and higher yield compared to the MSMPR platform which is more established in terms of existing knowledge and operations, and more plant-ready. Optimisation of the process conditions in the COBC is strongly encouraged coupled with the redesign of the seed transfer line to prevent blockages.