Predicting the phase behaviors of high-pressure materials

Experimental synthesis and characterization of high-pressure materials are extremely difficult, so first-principles methods based on quantum mechanics can play a crucial role. However, the high computational costs of these methods typically prevent rigorous predictions of macroscopic quantities at finite temperatures including the phase behaviors. In this talk, I will discuss how to enable such predictions by combining advanced statistical mechanics with data-driven machine learning interatomic potentials. As an example, we computed the low-pressure phase diagram of water from density functional theory at the hybrid level, accounting for thermal fluctuations, proton disordering and nuclear quantum effects. We then mapped the phase diagram of superionic water, which helps resolve the composition of ice giants. As a final example, we simulated the high-pressure hydrogen system with converged system size and simulation length, and found, contrary to established beliefs, supercritical behaviour of liquid hydrogen above the melting line.