Data Analytics-Enabled High-Throughput Material Discovery and Computational Analyses for Copolymer Membrane Innovations

The industrial adoption of novel membrane materials remains a significant challenge despite their potential to contribute to sustainability and energy efficiency. This dissertation, Data Analytics-Enabled High-Throughput Material Discovery and Computational Analyses for Copolymer Membrane Innovations, addresses these challenges by integrating advanced data analytics, computational modeling, and experimental design to accelerate material discovery and optimize membrane fabrication processes. Through the use of model-based design of experiments (MBDoE), this work provides novel insights into reactive ink formulations for membrane functionalization and the rapid characterization of membrane transport properties, laying the groundwork for the development of self-driving laboratories (SDLs) to explore the structure-property-performance relationships in copolymer membranes. The dissertation first establishes a comprehensive model-based data analytics framework, employing techniques such as maximum likelihood estimation, uncertainty quantification, Fisher Information Matrix (FIM) analysis to improve parameter estimability, and Akaike Information Criterion (AIC) for reaction pathway identification. These methodologies are applied to the development of mathematical models for reactive ink formulations and the optimization of membrane functionalization processes, with a focus on balancing environmental sustainability and industrial scalability. In addition, the work introduces a high-throughput diafiltration apparatus that significantly reduces the time required to characterize membrane properties five-fold, leveraging real-time data analytics for enhanced process optimization. Further applications of MBDoE refine this apparatus to maximize parameter precision while minimizing experimental effort. Finally, an integrated framework for the design of surface-charged nanofiltration membranes is presented, utilizing dynamic experiments and computational analysis to accelerate the development of membranes tailored for specific industrial applications. The findings of this research demonstrate the critical role of data-driven approaches in transforming membrane science, providing a pathway for reducing experimental effort, improving membrane performance, and enabling the transition of innovative materials from the laboratory to industrial scale.