Search

The phase-change material, Ge2Sb2Te5, is the canonical material ingredient for next-generation storage-class memory devices used in novel computing architectures, but fundamental questions remain regarding its atomic structure and physico-chemical properties. Here, we introduce a machine-learning (ML)-based interatomic potential that enables large-scale atomistic simulations of liquid, amorphous, and crystalline Ge2Sb2Te5 with an unprecedented combination of speed and density-functional theory (DFT) level of accuracy. Two applications exemplify the usefulness of such an ML-driven approach: we generate a 7,200-atom structural model, hitherto inaccessible with DFT simulations, that affords new insight into the medium-range structural order; and we create an ensemble of uncorrelated, smaller structures, for studies of their chemical bonding with statistical significance. Our work opens the way for new atomistic insights into the fascinating and chemically complex class of phase-change materials that are used in real non-volatile memory devices.

We provide an introduction to Gaussian process regression (GPR) machine-learning methods in computational materials science and chemistry. The focus of the present review is on the regression of atomistic properties: in particular, on the construction of interatomic potentials, or force fields, in the Gaussian Approximation Potential (GAP) framework; beyond this, we also discuss the fitting of arbitrary scalar, vectorial, and tensorial quantities. Methodological aspects of reference data generation, representation, and regression, as well as the question of how a data-driven model may be validated, are reviewed and critically discussed. A survey of applications to a variety of research questions in chemistry and materials science illustrates the rapid growth in the field. A vision is outlined for the development of the methodology in the years to come.

Structurally disordered materials pose fundamental questions1–4, including how different disordered phases (‘polyamorphs’) can coexist and transform from one phase to another5–9. Amorphous silicon has been extensively studied; it forms a fourfold-coordinated, covalent network at ambient conditions and much-higher-coordinated, metallic phases under pressure10–12. However, a detailed mechanistic understanding of the structural transitions in disordered silicon has been lacking, owing to the intrinsic limitations of even the most advanced experimental and computational techniques, for example, in terms of the system sizes accessible via simulation. Here we show how atomistic machine learning models trained on accurate quantum mechanical computations can help to describe liquid–amorphous and amorphous–amorphous transitions for a system of 100,000 atoms (ten-nanometre length scale), predicting structure, stability and electronic properties. Our simulations reveal a three-step transformation sequence for amorphous silicon under increasing external pressure. First, polyamorphic low- and high-density amorphous regions are found to coexist, rather than appearing sequentially. Then, we observe a structural collapse into a distinct very-high-density amorphous (VHDA) phase. Finally, our simulations indicate the transient nature of this VHDA phase: it rapidly nucleates crystallites, ultimately leading to the formation of a polycrystalline structure, consistent with experiments13–15 but not seen in earlier simulations11,16–18. A machine learning model for the electronic density of states confirms the onset of metallicity during VHDA formation and the subsequent crystallization. These results shed light on the liquid and amorphous states of silicon, and, in a wider context, they exemplify a machine learning-driven approach to predictive materials modelling. Machine learning models enable atomistic simulations of phase transitions in amorphous silicon, predict electronic fingerprints, and show that the pressure-induced crystallization occurs over three distinct stages.