Structure and dynamics of supercooled liquid Ge2Sb2Te5 from machine‐learning‐driven simulations

Studies of supercooled liquid phase‐change materials are important for the development of phase‐change memory and neuromorphic computing devices. Here, we use a machine‐learning‐based interatomic potential for Ge2Sb2Te5 (GST) to carry out large‐scale molecular‐dynamics simulations of liquid and supercooled liquid Ge2Sb2Te5. We demonstrate a pronounced effect of the thermostat parameters on the simulation results, and we show how using a Langevin thermostat with optimized damping values can lead to excellent agreement with reference ab initio molecular dynamics (AIMD) simulations. Structural and dynamical analyses are presented, including studies of radial and angular distributions, homopolar bonds, and the temperature‐dependent diffusivity. Our work demonstrates the usefulness of ML‐driven molecular dynamics for further studies of supercooled liquid GST, with length and time scales far exceeding those that would be accessible to AIMD