the sign of the potential alignment term in the Freysoldt correction scheme.

[Shuaishuai Yuan]

Hi!

I had a question about the sign of the potential alignment term in the Freysold correction scheme. 

The related code is around here:

https://github.com/materialsproject/pymatgen/blob/v2020.11.11/pymatgen/analysis/defects/corrections.py#L263-L270

In this paper Freysoldt, C., Neugebauer, J. & Van de Walle, C. G. Electrostatic interactions between charged defects in supercells. Phys. Status Solidi Basic Res. 248, 1067–1076 (2011):

I am still new to this. Please forgive me if this is a dumb question. I think C denotes the potential difference here and + q \Delta V is equal to - q C.  While I am not sure why in the codes, C is defined as negative of the potential difference “C = -np.mean(tmppot)”  and when calculating the potential alignment, there is a negative sign again “self.pot_corr = -q * C”. I think \Delta V in this paper is defined a little bit different as others, and C here is actually the potential difference. So I am thinking should we remove the negative sign, such that “C = np.mean(tmppot)“? I am also a VASP user, I am not sure whether the negative sign is related to the sign of LOCPOT? Your code runs perfectly on my side and I got all correction numbers, but I just want to understand everything. Thanks a lot for your time and help!

Sincerely,

Shuaishuai

[PyCDT forum]

Shuaishuai, Thanks for using PyCDT. 

This is not a dumb question and was one we had to look into further in the very early days of the code. If my memory serves me right, I believe the negative sign has to do with the fact that the LOCPOT file assigns a positive sign for electron charge. This result can also be checked by running one of the simple examples from Freysoldt's original paper. 

Hope this helps,

Danny  

[syua...@gmail.com]

Hi Danny,

Thanks very much for the quick reply! It helps a lot! I just run a calculation of the GaAs with V_Ga 3-.

Yes,  VASP  indeed gives the potential of a negative charge instead of the potential for a conventional positive test charge, and I also see that you have added a negative sign on the model charge potential to match with that. 

Yeah, everything makes sense to me now.

The negative sign is to reverse that potential to the conventional potential.

-(dft_dff - V_R) = C

In the figure generated by PyCDT code, we plot dft_diff, V_R - C, and dft_diff + C

Thanks a lot for your help! I will keep learning your codes! Thanks!

Sincerely,

Shuaishuai