Professor Rebecca Lindsey

Professor Rebecca Lindsey
Department of Chemical Engineering
University of Michigan

Combining Physics-Informed Machine Learning and Simulation to Characterize How Systems Evolve under Far-from-Equilibrium Conditions

Many rapid nanomaterial synthesis methods such as ultrasonication and laser-driven shocks function by inducing localized, short-lived extreme temperature and pressure conditions. The ensuing highly non-equilibrium evolution unfolds on sub-micron and sub-microsecond spatiotemporal scales, which makes direct experimental characterization difficult and leaves much of the underlying phenomena unclear. As a result, deploying these methods for design and synthesis of new materials requires considerable trial and error. Atomistic simulation holds promise for providing these missing insights but require interatomic models with both the computational efficiency of classical “force fields” (i.e., to reach the characteristic time and length scales) and accuracy of quantum-based methods (i.e., to describe the nominally reaction-mediated phenomena that underly these synthesis approaches). 

In this seminar, I will show that physics-informed machine- learning (ML) models can be used to bridge this gap. Practical considerations for construction and use of these models will be discussed, such as how model architecture and dataset selection modulate model efficiency, robustness, and predictive power. Sample applications will be provided, which include how we are using ML-enhanced simulations to guide development of an ultrafast laser shock-synthesis strategy for synthesis of bespoke covalent nanomaterials.

Rebecca Lindsey

Dr. Lindsey is an Assistant Professor of Chemical Engineering and by courtesy, of Applied Physics, Nuclear Engineering and Radiological Sciences, and Materials Science at the University of Michigan (UM). Prior to joining UM, Dr. Lindsey earned her B.S. in Chemical Engineering from Wayne State University and her M.S. and Ph.D. in Chemical Physics from the University of Minnesota, Twin Cities. Following, she worked as a postdoctoral scholar at Lawrence Livermore National Laboratory (LLNL), where she later converted to staff, leading a variety of research teams within the LLNL Energetic Materials Center. Her work in computational chemistry, for which applications have spanned sorption in soft materials, possible mechanisms for the origins of life, detonation synthesis of unusual carbon nanoparticles, and more, has been underpinned by a strong interest in developing tools enabling work in previously inaccessible problem spaces. In addition to her work in computational chemistry, she leverages data science and machine learning to aid in interpretation of large experimental datasets and to develop material performance models from them. Her research has been recognized by a number of awards, the most recent of which include the 2025 American Physical Society Neil Ashcroft Early Career Award for Studies of Matter at Extreme High Pressure Conditions and the 2024 American Institute of Chemical Engineering Computational Molecular Science and Engineering Forum Young Investigator Award.

Host: Professor Ilja Siepmann