How to Combine LOCPOT Files for Two DFT Runs?

[rmz4ed]

Hi,

I performed a relaxation and a static calculation in VASP and need to combine their LOCPOT files for calculating the defect formation energy FNV correction term using Sxdefectalign, as it requires a single LOCPOT file starting from the unrelaxed structure. I am wondering if it is possible to do this?

Thanks!

[fabien_tran1]

Hi,

According to the example for VASP in Sec. 4.3 of the manual (https://sxrepo.mpie.de/attachments/down ... manual.pdf), two LOCPOT files are necessary (located in two different directories).

[rmz4ed]

Yes - Two LOCPOT files of the bulk and vacancy cell are needed for the FNV correction calculation. However, as I used a used a two-step calculation for my vacancy cell (I first performed a relaxation using LDA, then a more expensive static calculation with HSE), I am unsure how to combine the LOCPOT files of the vacancy cell.

[fabien_tran1]

Why do you want to combine the LOCPOT files of the two calculations of the vacancy cell? What does it mean? Logically, you should choose the LOCPOT file of the calculation that is supposed to be the most interesting.

Besides this, there is an important point:
From the manual of Sxdefectalign, I understand that it is not really possible to use this program with the hybrid HSE. The hybrid potential is of nonlocal type and cannot be represented in a LOCPOT file. Only LDA and GGA potentials can be fully represented by a LOCPOT file.

Thus, it seems that the proper way to use Sxdefectalign is to get two LOCPOT files: one from the bulk calculation and one from the vacancy cell calculation. These two calculations should be done with the same potential (GGA is usually more accurate than LDA).