Optimal Operation of a Seeded Pharmaceutical Crystallization with Growth-Dependent Dispersion

A cooling profile for a seeded pharmaceutical crystallization designed to achieve a desired mean crystal size is found by minimizing the batch operating time subject to a crystallization model. In many systems, optimizing a characteristic of the product crystal size density is the usual objective. This objective is not feasible for this system because growth-dependent dispersion occurs in the absence of secondary nucleation. Instead, the batch operating time is minimized, and the resulting characteristics of the crystal size density are then fixed by the operating policy. To prevent undesired secondary nucleation, both a cooling-rate constraint and a supersaturation constraint are investigated. The optimal policies are implemented and verified experimentally. The experimental measurements include solution concentration, slurry transmittance, and video microscopy. The video microscopy is used to determine the mean crystal size and standard deviation. These values are found to agree closely with the model predictions for the optimal operating policies. In addition to determining the optimal operating policies, this study is the first to use video images for monitoring crystal size density mean and standard deviation as part of crystallizer model validation.