I want to use the optPBE functional and PBE_PAW POTCAR files. I see in the wiki that I should set GGA=OR.
What is confusing me is this note in the wiki:
"Important: VASP recalculates the exchange-correlation energy inside the PAW sphere and corrects the atomic energies given by the POTCAR file. For this to work, the original LEXCH tag must not be modified in the POTCAR file."
I don't see any mention of LEXCH in the section on the POTCAR and the GGA section does not say how LEXCH should be set for the different GGA's. The section on LEXCH seems to indicate that PAW_PBE potentials are the default for GGA. Further, inside the PAW-PBE files LEXCH is set to PE.
So, can I just ignore the note so long as I am using the PAW_PBE potentials?
Thanks
LEXCH tag in POTCAR
Moderators: Global Moderator, Moderator
-
fabien_tran1
- Global Moderator
- Posts: 367
- Joined: Mon Sep 13, 2021 11:02 am
Re: LEXCH tag in POTCAR
Hi,
The note on the wiki at the page GGA is correct and is there to avoid that users modify LEXCH. The LEXCH tag in the POTCAR indicates which functional was used to generate the data (pseudopotential) in the POTCAR. Therefore, LEXCH (and the POTCAR in general) must never be modified. If a functional other than PBE (e.g., GGA=OR) is specified in the INCAR, then a correction (OR minus PBE) for the energy of the atoms is calculated during the calculation. Using a POTCAR generated by PBE for a calculation with another functional (another GGA, a MGGA or a hybrid) is somehow inconsistent, but this is usually not a problem for accuracy. More discussion can also be found at https://www.vasp.at/forum/viewtopic.php?t=18315.
The note on the wiki at the page GGA is correct and is there to avoid that users modify LEXCH. The LEXCH tag in the POTCAR indicates which functional was used to generate the data (pseudopotential) in the POTCAR. Therefore, LEXCH (and the POTCAR in general) must never be modified. If a functional other than PBE (e.g., GGA=OR) is specified in the INCAR, then a correction (OR minus PBE) for the energy of the atoms is calculated during the calculation. Using a POTCAR generated by PBE for a calculation with another functional (another GGA, a MGGA or a hybrid) is somehow inconsistent, but this is usually not a problem for accuracy. More discussion can also be found at https://www.vasp.at/forum/viewtopic.php?t=18315.