Page 1 of 1

Polaron calculation in a selected site

Posted: Fri Oct 11, 2024 2:22 am
by Taifeng Liu
Dear developer,
When do the polaron localization in the epw2 step, how to make a displacement to let the polaron on a selected site. I find the defalt is the polaron localized on the first atom. I know in the Tutorial of exercise4, when doing the poalron transfering, it could include some displacement to move the polaron on an symmetry site. I want to know to make this displacement without the file dtau.plrn which is created after the epw2 calcualtion. Thank you.
Best wishes!
Taifeng Liu

Re: Polaron calculation in a selected site

Posted: Thu Oct 24, 2024 6:41 pm
by zhenbang
Hi Taifeng,

Since we don't know what the polarons look like a priori before we perform the epw2 calculation, it is not straightforward to manually set the position of the polarons. The default gaussian initialization of the A_nk vector in epw2 calculation is to set the polaron wavefunction at r=(0,0,0) in real space, but it does not guarantee that the converged polaron solution after self-consistency is right at the first atom, even thought it's very likely to be at the first atom because of the initialization.

To answer your question, I don't think there is a clear way of achieving what you described, but in the current release of EPW you can perhaps accomplish this task by creating a dtau_disp.plrn file manually, and setting all the entries to be zero except those corresponding to the atoms where you want your polarons to be centered at. You can play around with the numbers in those entries and to see if the polaron solutions will be centered at the desired site.

Best,
Zhenbang

Re: Polaron calculation in a selected site

Posted: Fri Oct 25, 2024 2:34 am
by Taifeng Liu
Hi Zhenbang,
Thank you. With the help of Prof. Jon Lafuente-Bartolome and Prof. Feliciano Giustino, we could localize the polaron at selected site by using the keyword dtau_disp.plrn. Of course the displacement must be made around the selected site.
But I am still keen to do this in the original epw2 calculation. Yes, you are right we don't know the polarons look like before we do calculations. But if there is a polaron forming, there must be displacement around. Therefore the initial displacement is very important. At least in the supercell approach, one can not get polaron localization even with the large fraction of HF exchange without displacement, i suppose it works in the epw calcualtions. Idealy, if there is a keyword in the epw2 calculations to control the polaron localization site, it would be very good and straightforward to do the following hoppling etc.
Anyway, thank you again, i am very happy that the epw method provides other way to investigate the polaron localization and hopping beyond the supercell approach.
Best wishes!
Taifeng