MolCryst-MLIPs: A Machine-Learned Interatomic Potentials Database for Molecular Crystals

TL;DR

This paper introduces MolCryst-MLIPs, an open database of machine-learned interatomic potentials for molecular crystals, developed using an automated pipeline and validated through molecular dynamics simulations.

Contribution

It presents a reproducible workflow for developing and validating MLIPs for molecular crystals, with the first release covering nine systems and providing a growing open database.

Findings

Mean energy MAE of 0.141 kJ/mol/atom across systems

Mean force MAE of 0.648 kJ/mol/Angstrom

Validated models maintain structural stability in MD simulations

Abstract

We present an open Molecular Crystal (MC) database of Machine-Learned Interatomic Potentials (MLIP) called MolCryst-MLIPs. The first release comprises fine-tuned MACE models for nine molecular crystal systems -- Benzamide, Benzoic acid, Coumarin, Durene, Isonicotinamide, Niacinamide, Nicotinamide, Pyrazinamide, and Resorcinol -- developed using the Automated Machine Learning Pipeline (AMLP), which streamlines the entire MLIP development workflow, from reference data generation to model training and validation, into a reproducible and user-friendly pipeline. Models are fine-tuned from the MACE-MH-1 foundation model (omol head), yielding a mean energy MAE of 0.141 kJ/mol/atom and a mean force MAE of 0.648 kJ/mol/Angstrom across all systems. Dynamical stability and structural integrity, as assessed through energy conservation, P2 orientational order parameters, and radial distribution functions, are evaluated using molecular dynamics simulations. The released models and datasets constitute a growing open database of validated MLIPs, ready for production MD simulations of molecular crystal polymorphism under different thermodynamic conditions.