Data-driven simulation and characterisation of gold nanoparticles melting

We develop efficient, accurate, transferable, and interpretable machine learning force fields for Au nanoparticles, based on data gathered from Density Functional Theory calculations. We then use them to investigate the thermodynamic stability of Au nanoparticles of different sizes (1 to 6 nm), containing up to 6266 atoms, with respect to a solid-liquid phase change through molecular dynamics simulations. We predict nanoparticle melting temperatures in good agreement with respect to available experimental data. Furthermore, we characterize in detail the solid to liquid phase change mechanism employing an unsupervised learning scheme to categorize local atomic environments. We thus provide a rigorous and data-driven definition of liquid atomic arrangements in the inner and surface regions of a nanoparticle, and employ it to show that melting initiates at the outer layers.

The record contains all the MD simulation trajectories carried out using mapped machine learning force fields in LAMMPS, the machine learning force fields as LAMMPS pair potentials, and the ab initio training data used to construct such force fields.