Quench-rate and size-dependent behaviour in glassy Ge<sub>2</sub>Sb<sub>2</sub>Te<sub>5</sub> models simulated with a machine-learned Gaussian approximation potential
Felix C. Mocanu, Konstantinos Konstantinou, Stephen R. Elliott
Abstract
Abstract Phase-change memory materials are promising candidates for beyond-silicon, next-generation non-volatile-memory and neuromorphic-computing devices; the canonical such material is the chalcogenide semiconductor alloy Ge 2 Sb 2 Te 5 . Here, we describe the results of an analysis of glassy molecular-dynamics models of this material, as generated using a newly developed, linear-scaling (O( N )), machine-learned, Gaussian approximation potential. We investigate the behaviour of the glassy models as a function of different quench rates (varied by two orders of magnitude, down to 1 K ps −1 ) and model sizes (varied by two orders of magnitude, up to 24 300 atoms). It is found that the lowest quench rate studied (1 K ps −1 ) is comparable to the minimum cooling rate needed in order completely to vitrify the models on quenching from the melt.