Dataset of Au atomic structures for training Machine Learning Interatomic Potentials

This dataset contains atomic structures of gold (Au) generated using density functional theory (DFT) calculations performed with the VASP package [1, 2]. The calculations were carried out using the projector-augmented wave (PAW) [3, 4] method and Perdew–Burke–Ernzerhof (PBE) pseudopotentials for gold, within the generalised gradient approximation (GGA) [5] for the exchange-correlation functional.

Molecular dynamics simulations were performed for at least 500 steps per structure. For bulk systems, the temperature range spans from 100 K to 1500 K, while for nanoparticles and slab structures, it extends from 100 K to 1000 K. For training our machine learning model (GAP [6]), the first 200 steps of each molecular dynamics trajectory were discarded to allow the thermostat to equilibrate the system to the target temperature. Using this dataset, we trained a GAP model for Au nanoparticles, whose applicability extends beyond the nanoparticle sizes included in the training set.

The dataset includes these starting configurations:

Small, low-energy Au nanoparticles (3 to 55 atoms)

Bulk Au in fcc, bcc, hcp, and simple cubic (sc) crystal structures

Slab models of fcc surfaces

The initial low-energy nanoparticle structures were adopted from a dataset reported in the literature [7].For each structure, we provide atomic coordinates along with corresponding total energies and per-atom forces. This dataset is suitable for training and validating machine learning interatomic potentials for gold.

References

[1] Kresse and Furthmüller. "Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set." Physical Review B 54.16 (1996): 11169.[2] https://www.vasp.at/[3] Blöchl. "Projector augmented-wave method." Physical Review B 50.24 (1994): 17953.[4] Kresse and Joubert. "From ultrasoft pseudopotentials to the projector augmented-wave method." Physical Review B 59.3 (1999): 1758.[5] Perdew, John P., Kieron Burke, and Matthias Ernzerhof. "Generalized gradient approximation made simple." Physical Review Letters 77.18 (1996): 3865[6] Bartók, Payne, et al. "Gaussian approximation potentials: The accuracy of quantum mechanics, without the electrons." Physical Review Letters 104.13 (2010): 136403.[7] Manna, Sukriti, et al. "A database of low-energy atomically precise nanoclusters." Scientific Data 10.1 (2023): 308.