Page 1 of 1

Regrading mode-resolved Eliashberg function

Posted: Sun Dec 31, 2017 1:48 am
by hellolori
Dear all,

How can I get the phonon mode resolved Eliashberg function from the EPW package ? I can only get the total Eliashberg function in the file "$prefix.a2f.XX" and I can't find any tag regarding the output of the mode-resolved function. Do I need to modify the code or use the output of the e-ph constant to compute it myself ? Is there any simpler way?

Best,
Feipeng

Re: Regrading mode-resolved Eliashberg function

Posted: Wed Jan 03, 2018 8:48 pm
by carla.verdi
Hi Feipeng,
I think you would need to modify the code, for example by restricting the summation over phonon indices to a subset of indices.
Best,
Carla

Re: Regrading mode-resolved Eliashberg function

Posted: Fri Jan 12, 2018 3:52 pm
by hellolori
carla.verdi wrote:Hi Feipeng,
I think you would need to modify the code, for example by restricting the summation over phonon indices to a subset of indices.
Best,
Carla

Dear Carla,

Thank you very much for your kindness. I have another question. As is known to all that band crossing is commonly seen. I would like to know that how does EPW determine brach indexes for all phonon modes for a given q point? Simply by the rank of frequencies? I'm afraid if this is the case, the mode-resolved Eliashberg spectrum would be insignificantly in the presence of band crossing.


bests,
Feipeng

Re: Regrading mode-resolved Eliashberg function

Posted: Fri Jan 12, 2018 3:59 pm
by sponce
Dear Feipeng,

This is correct. It is simply ordered by rank index. Therefore if there are band crossing you need to change the mode index you are following.

For this reason we do not provide the feature.

The correct automatic way would be to compute wf overlap and determine band character to follow the bands. This is complicated to implement and therefore not done.
You can do it by hand but must be careful to follow the correct band.

Best,
Samuel