Splitting of coarse el-ph matrix elements for polar material

General discussion around the EPW software

Moderator: stiwari

Post Reply
Vahid
Posts: 101
Joined: Fri Apr 08, 2016 11:02 pm
Affiliation:

Splitting of coarse el-ph matrix elements for polar material

Post by Vahid »

Dear EPW Users,

I am attempting to add electrostatic screening (similar to the procedure in Nature Communications,8,15769,2017) for polar materials to EPW. My understanding is that according to PRL (115,176401,2015), EPW does the following:

Code: Select all

(i) evaluate the complete matrix elements g on coarse Brillouin zone grids
(ii) subtract gL so as to obtain the short-ranged part of the matrix element, gS
(iii) apply Wannier-Fourier electron-phonon interpolation to the short-range matrix element, following Ref. [22]
(iv) add up the short-range part and the long-range part at arbitrary k and q points after interpolation.


So I would like to apply screening to both coarse and fine grid gL. However, the code does not seem to split the coarse grid g into gL and gS. It does add the gL for the fine grid in the rgd_blk_epw subroutine and ephwann_shuffle.f90. It also splits the dynamical matrix into L and S parts using the rgd_blk subroutine.

I have two questions:
1. Where does the code actually split the coarse matrix elements into the L and S components as suggested in (ii)?
2. Would it be sufficient to apply screening by dividing both the coarse and fine gL by epsilon(q), as suggested in the Nature Communications paper or are there other parts of the code that should be modified.

Thanks,

Vahid


Vahid Askarpour
Department of Physics and Atmospheric Science
Dalhousie University,
Halifax, NS, Canada

carla.verdi
Posts: 155
Joined: Thu Jan 14, 2016 10:52 am
Affiliation:

Re: Splitting of coarse el-ph matrix elements for polar mate

Post by carla.verdi »

Dear Vahid,

To answer your questions:
1) The 'full' matrix elements g are calculated on the coarse grid; then the step (ii) is performed inside the subroutine rotate_epmat by calling rgd_blk_epw .
2) It is sufficient to divide g by epsilon(q); to be precise, you just have to consider the full g on the fine grid, there is no need to divide g also on the coarse grid (if the coarse and fine grids were the same, those quantities would be the same).

Best,
Carla

Vahid
Posts: 101
Joined: Fri Apr 08, 2016 11:02 pm
Affiliation:

Re: Splitting of coarse el-ph matrix elements for polar mate

Post by Vahid »

Dear Carla,

Thank you so much for this clarification.

Best wishes,

Vahid

Post Reply