EWALD CUTOFF: Difference between revisions
No edit summary |
No edit summary |
||
| Line 10: | Line 10: | ||
For surface calculations with large cells and thick vacuum regions, however, some necessary G-vectors may be cut off for the default value of {{TAG|EWALD_CUTOFF}}, and variations in total energy might increase to the order of 10 meV. | For surface calculations with large cells and thick vacuum regions, however, some necessary G-vectors may be cut off for the default value of {{TAG|EWALD_CUTOFF}}, and variations in total energy might increase to the order of 10 meV. | ||
This can spoil convergence with respect to the number of layers or the size of the vacuum | This can spoil convergence with respect to the number of layers or the size of the vacuum for weakly bound surfaces. | ||
{{NB|tip|For highly accurate slab calculations, set {{TAG|EWALD_CUTOFF|6.0}}.}} | {{NB|tip|For highly accurate slab calculations, set {{TAG|EWALD_CUTOFF|6.0}}.}} | ||
{{available|6.6.0}} | {{available|6.6.0}} | ||
Revision as of 09:21, 17 March 2026
EWALD_CUTOFF = [real]
| Default: EWALD_CUTOFF | = 4.0 |
Description: EWALD_CUTOFF sets the unified cutoff radius for the Ewald summation of the electrostatic interaction. It controls both the number of cells that contribute (in real space) as well as the number of G-vectors that are considered (in reciprocal space).
The default value of EWALD_CUTOFF is a safe choice in nearly all cases. For bulk systems, increasing it will not change the total energy by more than a few $\mu$eV.
For surface calculations with large cells and thick vacuum regions, however, some necessary G-vectors may be cut off for the default value of EWALD_CUTOFF, and variations in total energy might increase to the order of 10 meV. This can spoil convergence with respect to the number of layers or the size of the vacuum for weakly bound surfaces.
Tip: For highly accurate slab calculations, set EWALD_CUTOFF = 6.0.
|
| Mind: Available as of VASP 6.6.0 |