A Dual Projection Imaging System to Characterize Crystallization Processes: Design and Applications

Abstract

Crystallization is a separation process that has been practised and applied widely in fine chemical and pharmaceutical industries. In the majority of applications, it is used to obtain a high purity solid form, which often exhibits a distribution of sizes and shapes. Crystallization involves a number of fundamental phenomena -- often poorly understood -- like nucleation, growth, dissolution, agglomeration, to name a few. The lack of reliable monitoring tools and techniques inhibits attaining deeper insights into the mechanisms involved, which naturally affects the robust and optimal operation of these crystallization processes. Poor understanding coupled with stringent targets on the product quality for drugs in terms of purity, stability, and bioavailability, has attracted significant attention from both the academia and the industry. The work presented in this thesis led to improvements in the state-of-the-art imaging techniques for size and shape characterization of the solid phase in batch solution crystallization processes. Based on these improvements, studies making use of the shape information obtained from the imaging device were undertaken for several interesting and previously unexplored applications. The former point led to providing better characterization and understanding of the process from a macroscopic scale. While the latter point with the aid of the former led to several automated and controlled approaches to manipulate the size and shape of undesirable needle-like crystals to equant crystals. The key accomplishments of this thesis were • enhancements to the hardware of a stereoscopic imaging device and to the imaging analysis routines to classify crystals observed by the imaging device into five different shape classes and to obtain a three-dimensional reconstruction of these crystals. • assessing the reliability of commercial spectroscopic techniques to estimate solute concentration in batch crystallization processes and proposing a new approach based on volumetric reconstruction of crystals observed by the stereoscopic imaging device, to estimate the solute concentration. • transformation of needle-like crystals to more equant crystals in a multistage cyclic process consisting of wet milling, dissolution, and growth stages, exploiting the online monitoring capabilities of the imaging device and simple feedback control laws for the individual stages. To summarize, the results obtained certainly reinforce the potential and the competence of imaging tools to tackle a wide array of challenges faced by the crystallization community. Irrespective of the promising outcome, their potential pitfalls are definitely not overlooked and plausible proposals to overcome these are discussed diligently to assist future research on monitoring, on modeling and on control of crystallization processes.