AI-Assisted Autonomous Nanomaterials Discovery and Knowledge Generation

Paul Kenis and Baron Peters (Chemical & Biomolecular Engineering)
Moonsub Shim, Pinshane Huang, Daniel Shoemaker, and Jim Zuo (Materials Science & Engineering)

Research Problem

Presently, discovery and optimization of nanomaterials is carried out manually through a trial-and-error or design-of-experiments (DoE) approach. Unfortunately, the parameter space for nanomaterial synthesis is large (covering numerous different process parameters) and poorly understood, with a sparse number of optimal parameters that can generate nanomaterials of desired properties. As such, current approaches of intuition-guided and DoE-based experimentation are inadequate for screening for new recipes or generation of new hypotheses for materials discovery or knowledge generation. 

Vision

Develop the ML knowledge base and the AI-assisted autonomous platforms needed for nanomaterial discovery for various applications. This will be achieved by developing and applying machine learning (ML) and atomic-scale simulation approaches to predict the structure, properties, and optimal precursors for synthesizing novel nanomaterials. This information will then be leveraged with autonomous experimental platforms to systematically map vast synthesis parameter space, orders of magnitude larger in size than what current methods (ML aided or not) can generate. These vast parameter space maps will help refine subsequent synthesis planning algorithms while aiding domain experts in deducing new knowledge, such as new synthesis strategies and mechanistic insights into nanomaterial nucleation and growth. 

Larger Impact

Successful development of an efficient and generalizable autonomous workflow for nanoparticle synthesis incorporating comprehensive structure and composition characterization methods allows for developing a data-rich framework that can effectively solve immensely challenging problems. This will provide a template for autonomous workflows across various materials and insights that will be highly valuable for the broader scientific community. If offered, the results of this SRI grant will provide a solid footing for competitive proposals establishing Illinois as a leader in autonomous nanomaterials discovery.