UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Developing a self-driving robotic platform for rapid screening of solid purification methods Depner, Noah Brendan

Abstract

Crystallization is a highly effective method of both isolating and purifying a compound of interest following synthesis. In industry, the development of a crystallization method is a time and labor-intensive process, often requiring a significant amount of impure solid and a strong understanding of the many factors that complicate crystallization processes. Due to these limitations, crystallization design may not be feasible in time or material-limited situations. In this thesis, I outline the development of a self-driving platform that can rapidly screen purification attempts on an impure solid. The platform autonomously collects solubility information and conducts crystallization and washing experiments on the solid, evaluating the yield, purity, and impurity rejection from each attempt. The solvent system is the primary variable screened by the platform. Following a decision tree, a user will arrive at a potential purification method for the process solid, requiring minimal characterization of impurities and using significantly less material than a traditional high-throughput screen. The platform itself is highly modular and reconfigurable; important considerations for integrating filtration, agitation and analytical modules are discussed in detail. Three industry-relevant solids were used to demonstrate the ability of the platform to model solubility data and acquire reproducible data on crystallizations and washing experiments. In one of these case studies, the full decision tree was used to successfully find high-purity crystallization conditions. Information about the impurity profile of the solid is another important factor in designing a purification. The platform is able to autonomously conduct powder dissolution studies on a solid, identifying the retention mechanism for each impurity. This was successfully demonstrated on one test system and three industry-relevant solids, identifying impurity retention mechanisms across a range of categories. Development of the platform is still in progress; future directions are outlined for integrating machine learning abilities, fully automating all decision points, and increasing robustness and accessibility.

Item Media

Item Citations and Data

Rights

Attribution-NonCommercial-NoDerivatives 4.0 International