Deep machine learning potential for atomistic simulation of Fe-Si-O systems under Earth's outer core conditions

Using artificial neural-network machine learning (ANN-ML) to generate interatomic potentials has been demonstrated to be a promising approach to address the longstanding challenge of accuracy vs efficiency in molecular dynamics (MD) simulations. Here, taking the Fe-Si-O system as a prototype, we show that accurate and transferable ANN-ML potentials can be developed for reliable MD simulations of materials at high-pressure and high-temperature conditions of the Earth's outer core. The ANN-ML potential for the Fe-Si-O system is trained by fitting the energies and forces of related binaries and ternary liquid structures at high pressures and temperatures obtained by first-principles calculations based on density functional theory (DFT). We show that the generated ANN-ML potential describes well the structure and dynamics of liquid phases of this complex system. In addition to binary systems (${\mathrm{Fe}}_{189}{\mathrm{Si}}_{61}, {\mathrm{Fe}}_{189}{\mathrm{O}}_{61}$, and ${\mathrm{Si}}_{80}{\mathrm{O}}_{160}$) and ternary systems (${\mathrm{Fe}}_{189}{\mathrm{Si}}_{38}{\mathrm{O}}_{23}$), whose snapshots are included in the training dataset, the reliability of the ANN-ML potential is validated in two other ternary systems (${\mathrm{Fe}}_{189}{\mathrm{Si}}_{23}{\mathrm{O}}_{38}$ and ${\mathrm{Fe}}_{158}{\mathrm{Si}}_{14}{\mathrm{O}}_{28}$), whose snapshots are not included in the training dataset. The efficient ANN-ML potential with DFT accuracy provides a promising scheme for accurate atomistic simulations of structures and dynamics of the complex Fe-Si-O system in the Earth's outer core.