README for dataset: Pure Magnesium DFT calculations for interatomic potential
fitting

This dataset provides DFT (density functional theory as implemented in VASP,
Vienna Ab Initio Simulation Package) calculations for pure Magnesium. It was
designed by Binglun Yin, Markus Stricker and William A. Curtin for fitting a
neural network potential with Behler-Parinello symmetry functions. Binlung Yin
carried out the calculation.

It corresponds to a dataset that is commonly used to fit interatomic potentials
for mechanics applications and includes structure-energy relationships for
structures used to calculate:

1. Bulk properties
   - Elastic constants
   - Cohesive energy
   - Energy-volume relationship

2. Generalized stacking fault energies
   - Curve for basal stacking fault
   - Curve for pyramidal I stacking fault
   - Curce for pyramidal II stacking fault
   - Stable stacking faults with \sigma_{3j} relaxation

3. Decohesion and relaxed surfaces
   - Basal plane
   - Prism plane
   - Pyramidal I plane
   - Pyramidal II plane
   - Relaxed surfaces

4. Dimer
   - 2 atom from small to large distance

5. Corner and rod geometries
NOTE: these structure are not relaxed.
   - Basal: bulk, dot, rod
   - Pyramidal I: bulk, dot, rod
   - Pyramidal II: bulk, dot, rod

6. Vacancy formation energy
   - Several structures with varying number of atoms

The dataset provides the raw unaltered VASP output.