Data for the publication High-accuracy thermodynamic properties to the melting point from ab initio calculations aided by machine-learning potentials, npj Comput. Mater., DOI: 10.1038/s41524-022-00956-8 (2023)
This data set contains
- the training sets (VASP files),
- the low moment-tensor-potentials (MTPs) and high-MTPs,
- the effective quasiharmonic (QH) potentials,
- the free energies,
- and the final thermodynamic database (properties) for the four unaries:
Nb bcc PBE, Ni fcc PBE, Al fcc PBE, Al fcc LDA, Mg hcp PBE, and Mg hcp LDA. (niobium, nickel, aluminum, and magnesium)
A flowchart and step-by-step methodology are provided in the Supplementary Information of the publication, and a short method in its main text, for the reproduction of the results.
VASP, 5.3.5
MLIP (Machine-Learning Interatomic Potentials), 2
Phonopy, 2.15
Hiphive, 0.7.1
LAMMPS, 23 Jun 2022
; Forslund, Axel