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

We present an open Molecular Crystal (MC) database of Machine-Learned Interatomic Potentials (MLIPs) called MolCryst-MLIPs. The first release comprises fine-tuned MACE models for nine molecular crystal systems─Benzamide, Benzoic acid, Coumarin, Durene, Isonicotinamide, Nicotinic acid, Niacinamide, 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-1·atom-1 and a mean force MAE of 0.648 kJ·mol-1·Å-1 across all systems. Benchmarked against three state-of-the-art foundation models on the DFT-labeled polymorph set, only the fine-tuned models reliably identify the most stable polymorph, with a mean Kendall τ = 0.397 indicating a moderate rank correlation with the full DFT landscape. 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 data sets constitute a growing open database of validated MLIPs, ready for production MD simulations of molecular crystal polymorphism across the polymorphic landscape of each target compound under different thermodynamic conditions.