LFOCKAEDFT: Difference between revisions
No edit summary |
No edit summary |
||
| Line 7: | Line 7: | ||
This flag should be set only in exceptional cases. The Hartree as well | This flag should be set only in exceptional cases. The Hartree as well | ||
as the DFT part are usually calculated very accurately using the one-centre | as the DFT part are usually calculated very accurately using the one-centre | ||
PAW spheres. Restoring the all-electron charge | PAW spheres. Restoring the all-electron charge accurately on the plane | ||
wave grid adds potentially noise, but should not change the results. | wave grid adds potentially noise, but should not change the results (relative energies, | ||
forces etc.). | |||
The flag, however, needs to be set in optimized potential methods, which | The flag, however, needs to be set in optimized potential methods, which | ||
are supported by VASP but not documented yet. | are supported by VASP but not documented yet. | ||
Revision as of 15:50, 14 March 2017
LFOCKAEDFT = [logical]
Default: LFOCKAEDFT = .FALSE.
Description: LFOCKAEDFT forces VASP to use the same charge augmentation for the Hartree and exchange part as is used for the Fock exchange and the many body beyond DFT methods.
This flag should be set only in exceptional cases. The Hartree as well as the DFT part are usually calculated very accurately using the one-centre PAW spheres. Restoring the all-electron charge accurately on the plane wave grid adds potentially noise, but should not change the results (relative energies, forces etc.). The flag, however, needs to be set in optimized potential methods, which are supported by VASP but not documented yet.
Related Tags and Sections
LMAXFOCKAE, NMAXFOCKAE, QMAXFOCKAE