Supplementary data for 'Accurate and efficient machine learning interatomic potentials for finite temperature modeling of molecular crystals' , arXiv:2502.15530

This repository contains supplementary data supporting the findings of the paper:

Della Pia F, Shi B X, Kapil V, Zen A, Alfè D, Michaelides A, Accurate and efficient machine learning interatomic potentials for finite temperature modeling of molecular crystals

Contents

*DFT-benchmark: Input/Output files for the benchmark of several DFT functionals on the lattice energies of X23 against reference quantum Diffusion Monte Carlo values;

*EXAMPLES: general examples of input files (PHON, ASE, i-PI) for the QHA, CMD, and QMD calculations of the sublimation enthalpy;

*QHA_OUTPUT_FILES: Output files of the QHA calculations of the sublimation enthalpy with DFT (vdW-DF2) and the fine tuned MACE models;

*CMD_QMD_INPUT_OUTPUT_FILES: Input/Output files of the CMD and QMD simulations run with the fine tuned MACE models to compute the sublimation enthalpies of X23;

*FINE_TUNED_MACE_MODELS: Training set (in the ASE .traj format) and fine tuned MACE models for each system in X23.