Polymer Membranes and Nanostructured Materials: AI/ML in Advanced Manufacturing, Characterization, and Self-Driving Labs

Abstract

Membrane and separations technology is a driving force for both fundamental and applied research that impacts the economy. Polymer materials are a vast design space that can be further exploited towards high-performance membrane and filtration applications. Machine learning (ML) can be used to improve membrane materials and design using datasets and data analytics with agentic AI tasks from large language models (LLM). The applications of AI/ML workflows include membrane polymer structure-composition-processing-property (SCPP) relationships, membrane testing, and performance validation. Characterization methods to quantify transport, permeance, selectivity, and environmental stability of membranes need to go beyond statistical methods. Generative AI workflows are important for science discovery and rapid feedback. Exploring bio-inspired designs – de novo membrane design, and future directions will depend on how fast we can implement these workflows. By designing and updating research end-to-end, it is possible to leverage the new tools and datasets becoming available to the membrane community.