Application of a Semiautomated Crystallizer to Study Oiling-Out and Agglomeration Events—A Case Study in Industrial Crystallization Optimization

Abstract

A case study on a challenging industrial crystallization is presented. The crystallization under investigation was prone to oiling-out, which led to agglomeration and stalling of crystallization. A semiautomated crystallizer—designed and built iteratively as the crystallization needs changed—was used to study the crystallization and identify the failure modes. A fiber optic probe was included in this setup to track turbidity, which was found to drop in signal whenever an oiling-out/agglomeration event was observed with in situ video microscopy. With this knowledge, the turbidity was monitored to detect oiling-out and was subsequently used as a feedback mechanism to change the temperature profile to reverse oiling-out and resume crystal growth during the crystallization. The automated sampler, coupled with off-line HPLC analysis, provided valuable insight about the behavior of impurities during oiling-out and identified one specific impurity that influenced the formation of the second liquid phase and crystallization rate. The data-rich experimentation performed enabled the expedient optimization of crystallization parameters such as the temperature profile, seed loading, and cycle time. The final crystallization was robust, predictable, and capable of self-correction should a potential failure mode be encountered.