Molecular Dynamics and Deep Learning for Materials Including TMDC & Oxide Moire Structures: II

Materials simulations under extreme conditions are essential for advancing our understanding of planetary interiors. For over two decades, combining ab initio static calculations with harmonic or anharmonic phonon frequencies has enabled remarkable predictions of minerals' thermodynamic and thermoelastic properties, facilitating direct planetary interior modeling. This approach has traditionally been preferred over ab initio molecular dynamics (MD) due to the computational challenge of achieving highly converged free energy calculations across the required range of conditions. However, recent advancements in deep learning (DL)-based force fields for MD simulations now allow for routine exploration of complex systems and processes critical to geophysics and geochemistry. Geochemical research, in particular, involves thermochemical equilibrium in multi-phase, multi-component systems, necessitating robust free energy databases under extreme conditions.

In this talk, I will discuss the application of materials simulations to longstanding questions in geophysics and demonstrate their growing integration into planetary models. I will also highlight how this approach breaks new ground in geochemistry, enabling studies previously constrained by computational limits.

*Research supported by DOE and NSF