Development of Machine-Learned Force Field for Shock-Driven Dynamics of Low-Density Foams in Inertial Fusion Energy Conditions

The targets for inertial confinement energy (IFE) are primarily made of low-density foams. The shock response of the foam microstructures thus influence critically influence target performance. However, accurate simulation of these materials across the wide range of temperatures ranging to several eVs relevant to IFE remains computationally challenging with Kohn-Sham density functional theory (DFT). In this work, we develop a machine-learned force field (MLFF) for poly(methyl methacrylate) (PMMA) using on-the-fly machine learning strategy, spanning a dense temperature grid (0.08 to 25 eV) across the warm dense matter regime. The MLFF is being validated by benchmarking against the FPEOS DFT equation of state database [1], capturing pressure–volume–temperature behavior with high fidelity. We will then deploy the MLFF for large simulation cells to perform large-scale shock simulations of PMMA foams using different shock compression simulation strategies. These simulations will reveal the role of morphology in the equation-of-state of these foams. Our results will highlight the predictive capabilities of MLFFs for modeling realistic foams under dynamic loading, offering insights into material design and target optimization for high-gain IFE experiments.



Reference:

1. 1. Militzer et al., Physical Review E, 103 (2021) 013203.