This thesis explores the design and monitoring of crystallization processes with multiple solid-state forms, which play a crucial role in determining the quality of crystalline products. The research is divided into two main parts, each focusing on distinct aspects of crystallization processes while sharing the common goal of enhancing process understanding and monitoring.

The first part investigates the temperature cycling induced deracemization (TCID) process with concurrent degradation to elucidate the impact of degradation on the TCID process and its economic implications. This understanding is crucial for informed process design and optimization decisions. Enantiomeric separations are vital for ensuring the safety and efficacy of active ingredients in natural products and pharmace...[ Read more ]

This thesis explores the design and monitoring of crystallization processes with multiple solid-state forms, which play a crucial role in determining the quality of crystalline products. The research is divided into two main parts, each focusing on distinct aspects of crystallization processes while sharing the common goal of enhancing process understanding and monitoring.

The first part investigates the temperature cycling induced deracemization (TCID) process with concurrent degradation to elucidate the impact of degradation on the TCID process and its economic implications. This understanding is crucial for informed process design and optimization decisions. Enantiomeric separations are vital for ensuring the safety and efficacy of active ingredients in natural products and pharmaceuticals. However, some active ingredients may degrade at elevated temperatures. Despite extensive literature on ideal TCID processes under various conditions, systematic studies on the potential impacts of degradation during TCID processes are limited. Chapter 2 presents an experimental characterization of the deracemization of p-synephrine through a TCID process in the presence of dimerization-driven degradation, leading to the discovery of a new solid-state form of a synephrine dimer. The results demonstrate that an 86% enantiomeric excess of R-(-)-p-synephrine can be achieved under optimal conditions. The systematic identification of the design space reveals that degradation significantly narrows the operating window when aiming for high enantiomeric excess, emphasizing the importance of batch time control. Building upon these experimental findings, Chapter 3 assesses the economic implications of degradation on the TCID process for p-synephrine through a conceptual process design and economic analysis over various operational scales. The economic analysis develops correlations between TCID operational cost can serve as a methodological reference for investigating the economic implications of degradation in other TCID processes.

The second part of the thesis focuses on protein crystallization, an increasingly important alternative purification technique to chromatography due to its potential for cost-effectiveness and scalability. In this part, the development of a soft sensor for monitoring the mass fractions of different solid-state forms in protein crystallization is presented, contributing to improved process monitoring and understanding of factors influencing the formation of different solid-state forms, such as shear rate. Chapter 4 introduces a novel soft sensor based on convolutional neural networks for quantifying the mass fraction of distinct solid-state forms of lysozyme crystals from microscopy images of a crystal slurry. This soft sensor addresses the significant challenge of reliably quantifying the mass fractions of different solid-state forms of protein crystals during operation, which is crucial for process control and optimization. The optimized models exhibit good accuracy, with a mean absolute error of 3%, and successfully describe the dynamic development of solid-state forms in both batch and continuous processes. Leveraging this new methodology, Chapter 5 quantitatively correlates shear rate with the mass fractions of resulting solid-state forms of lysozyme crystals in continuous segmented flow crystallization. The results demonstrate that shear rate and lysozyme concentration are critical factors in determining the attainable solid-state forms, with higher shear rates favouring the formation of needle-like crystals under low lysozyme concentrations, whereas higher lysozyme concentrations promote the formation of tetragonal crystals, even at high shear rates. The developed soft sensor enables continuous monitoring and quality assurance, enhancing the control and predictability of crystalline protein product production with desired solid-state forms, representing a step towards the implementation of advanced monitoring techniques in protein-based biopharmaceutical manufacturing.

The methodologies and insights presented in this thesis lay the groundwork for future research on the design and optimization of crystallization processes and the development of advanced monitoring techniques. The findings have the potential to contribute to ensuring the quality and consistency of crystalline products in various industries, including natural products, proteins, and pharmaceuticals.