Probing Structural and Magnetic Phase Changes in the Shock Response of Iron with Molecular Dynamics

For ferromagnetic materials like iron which exhibit strong spin-lattice interactions, magnetic fluctuations have a significant impact on the underlying mechanical, thermal, and structural material properties. Utilizing a large training database of high temperature/pressure ab initio calculations we apply the machine-learned Spectral Neighbor Analysis Potential method to capture the magnon and phonon degrees of freedom within a simultaneous molecular- and spin- dynamics simulation. Our method has been validated by examining the thermal-mechanical properties and static transition pressures, which show good agreement with experiments. To further test our approach, we carry out large scale simulations that probe the shock response of iron for both single and polycrystalline domains. Doing so we characterize the shock driven bcc-to-hcp structural transition along with the corresponding ferro-to-paramagnetic transition which is directly observed in the simulations within the spin subsystem. This presentation will highlight the unprecedented predictive capability enabled for magnetic materials from atomistic molecular dynamics which is underpinned by advances in machine learned interatomic potentials.