posted on 2021-09-13, 18:33authored byAldo Glielmo, Federico Bianchini
The files consist of picoseconds-long canonical (and thermalised) trajectories of 4 metallic crystalline systems. Within each file, positions and forces of all the atom are saved.
The time-step was chosen to be 2 fs.
The temperature was controlled by a loosely coupled Langevin thermostat.
The periodic cell was taken of dimension 4x4x4.
Details of each file:
Ni_500K.xyz: Nickel, 500K.
Ni_1700K.xyz: Nickel, 1700K.
Fe_500K.xyz: Iron, 500K.
Fe_500K_vac.xyz: Iron, 500K, with a single vacancy.
Utility:
The data can be used to reproduce the results of the associated publication and for further developments of closely related research.
A. Glielmo, P. Sollich, A. De Vita, “Accurate Interatomic Force Fields via Machine Learning with Covariant Kernels”, Physical Review B. Submitted
Funding
Office of Naval Research Global, ONRG (Award No. N62909-15-1-N079)
EPSRC Centre for Doctoral Training in Cross-Disciplinary Approaches to Non-Equilibrium Systems
Engineering and Physical Sciences Research Council
The data was obtained via the Vienna Ab inito SImulation Package (VASP) implementation of Density Functional Theory (DFT).
Plane-waves were used as basis set and the PBE/ GGA approximation was adopted for the exchange and correlation energy.