Integrated Self-Optimisation of API Synthesis and Crystallisation Using Machine-Learning

Pharmaceutical manufacturing demands precise control over chemical reactions and crystallisation processes to ensure the quality and efficacy of drug products. A holistic, self-optimising system that integrates these processes can enhance efficiency, reduce waste, and ensure consistent product quality. This PhD project will be focused on developing such a system for a coupled reaction and crystallisation process, incorporating advanced process control, machine learning, and real-time analytics. The project will optimise impurity profiles to target product properties critical for drug formulation and delivery, including particle size distribution, crystal habits and polymorphic forms.

The primary objectives are:

To Develop a Coupled Reaction and Crystallisation Kinetic Model

To Implement Advanced Process Control for real-time optimisation.

To Utilise Machine Learning for Self-Optimisation of process parameters.

To Optimise Impurity Profiles to achieve targeted product properties.

To Validate the Holistic Approach to ensure the developed system meets pharmaceutical quality standards and regulatory requirements.

The outcomes will include a validated kinetic and thermodynamic model of the coupled reaction and crystallisation process, with an advanced process control system capable of real-time optimisation. Machine learning models will enable autonomous, self-optimising process control, with optimised impurity profiles to control crystallisation kinetics and polymorph formation. Improvements in process efficiency, product quality, and operational cost savings will be demonstrated, providing a framework for holistic self-optimisation applicable to other pharmaceutical processes.

Thereby, this research will significantly contribute to pharmaceutical process development by demonstrating the feasibility and benefits of a holistic, self-optimising approach.