KINETIC MODELING OF DISSOLUTION AND CRYSTALLIZATION OF BATCH REACTIONS WITH IN SITU SPECTROSCOPIC MEASUREMENTS

Abstract

The use of Process Analytical Technology (PAT) recommended by the Food and Drug Administration (FDA) has significantly increased during the past years in the design, control and monitoring of pharmaceutical or chemical manufacturing processes. Nowadays, PAT is also commonly used in Good Manufacturing Practices (GMPs). Some PAT techniques employ on-line fiber-optic sensors to acquire non-destructive measurements of physical properties. In this thesis we demonstrated that these methods can also be used to obtain kinetic information of dissolved and solid fractions of molecular substances in slurries in real time. The main objective of this project was to develop a comprehensive model for DuPont's sulfonylurea coupling reaction for monitoring purposes (e.g. detect process upsets, detect endpoints, and forecast changes). The comprehensive model will allow us to estimate the kinetics of the reaction, the kinetics of dissolution, and the kinetics of crystallization. Before such a complex model can be developed, it was necessary to conduct experiments in a simpler system (e.g. salicylic acid in water-ethanol mixtures). These experiments were designed so the process of dissolution and the process of crystallization could be observed independently of one another and independently of chemical reactions. Consequently, we were able to elucidate an appropriate model for each process in our small-scale semi-batch reactor. For the simpler system, this study used attenuated total reflectance ultra-violet visible (ATR UV-vis) spectroscopy for kinetic modeling of the dissolution and crystallization of salicylic acid in ethanol-water. The dissolution model, which relied on a power-law equation, was obtained by adding aliquots of an ethanol-water mixture into a salicylic acid slurry. Near-infrared (NIR) diffuse reflectance spectroscopy was used to detect and quantify the solid fraction present in a slurry. Using a Partial Least-Squares (PLS) calibration, we were able to verify and validate the kinetic model. A temperature probe was also used to monitor heat changes involved in these experiments.