Empirical temperature-dependent intermolecular potentials determined by data mining from crystal data

Highlights: • Crystal structure prediction. • Temperature dependent force field. • Data mining. Modern force fields are accurate enough to describe thermal effects in molecular crystals. Here, we have extended our earlier approach to discrete force fields for various temperatures to a force field with a continuous function. For the parametrisation of the force field, we used data mining on experimental structures with the temperature as an additional descriptor. The obtained force field can be used to minimise energy at a finite temperature and for molecular dynamics with zero-K potentials. The applicability of the method has been demonstrated for the prediction of crystal density, temperature density gradients and transition temperature.