Research

The expanding complexity of material composition, morphology, and processing methods is making mesoscale material science increasingly complex. Current state-of-the-art atomistic theories need to continuously grow to keep up with the pace of modern-day materials science. I plan to focus my group’s effort on establishing the missing connection between phase transformations, transport, and reaction dynamics in mesoscale. For mesoscale evolution of new phases in a compositions space, it is important to know the accessible stable/metastable phases possible for a certain pressure/temperature condition. This data is absent in standard materials databases. Therefore, the first thrust area of theory development is about the high-throughput screening of the potential energy landscape of phases and phase transformation mechanisms powered by machine learning methods and physics-based simulations. The second thrust in the group is building exascale-ready computational tools in coupled multi-resolution mesoscale models for thermal-and-mass transport, phase nucleation, growth, and phase transitions under different thermo-chemo-mechanical conditions. Finally, multiscale theory cannot be a standalone part of building understanding without the integration of experimental data, sensors, and AI-based tools. The new part of our activities which integrated academic groups, national laboratories (ORNL and ANL), and NIST involves the integration of AI to reduce the time needed to predict outcomes using a purely physics-based approach. The third thrust in my research group is the use of AI for multiscale control of experiments for discovery and scale-up of advanced materials. Our latest work developing a manufacturing data and machine learning (MDML) platform can integrate experimental and simulated data in an AI-based platform. Ongoing work in autonomous steering of experiments, optimization of processes, and use of streaming data analytics on a low-latency framework are being tested by collaborators in Argonne National Laboratory. The NASEM committee on the future of manufacturing and chemical engineering invite me to present our approach for AI for Multiscale Control of Complex and Uncertain Process Environment and their report will highlight this as a frontier for engineering in the next decade.