Temperature in electron-phonon calculations

Theory Electron-phonon coupling
1

Hello Everyone,

I have some questions from S. Ponc et al, Comput. Phys. Commun. 209, 116 (2016),

  1. In Eq. (3) which parameter takes care of summation over m, is it fsthick?
  2. I calculate phonon frequency using QE code and from my understanding its always at 0K and in Eq. 3 there is a temperature dependence on phonon occupation, will it not effect the phonon frequencies? Or is there some assumption which I am not aware of?
  3. Is there any binary to generate random k/q-grids? or we can simply use kgrid.pl from wannier90?

Thanks in advance.

Best wishes,
DInesh Yadav
University of Konstanz

2

Hello,

  1. It will be summed over all the bands in your calculations (nbndsub)
  2. It is indeed a first-order approximation. In principle, the phonon frequencies also get renormalized by el-ph coupling and should depend on (\omega,T). You get QP phonon eigenstates.
  3. Random grid are generated with the input parameters rand_k and rand_q from EPW (look at input variable list on the website).
    However, for that paper I used quasi-random Sobol points. In that case, I wrote a python script to do this. The script is not public as it is quite dirty.
3

Dear Dr. Sponce,

Thank you very much for your reply. I have further questions from the same paper of yours mentioned in the last post.

  1. in Eq. 3 on R.H.S. there is no term with index (n,k) then how does sigma depends on (n,k)? is it typo with ef in the denominator?
  2. As the sigma depends upon the electronic occupation, if we think of relaxation of carriers via e-ph interaction like PhysRevLett. 112. 257402. Should it not be that sigma must be calculated again and again w.r.t. to different occupation depending upon the time?

Best wishes,
Dinesh
Uni. of Konstanz

4

Dear Dinesh,

  1. There is no typo. There is an explicit nk index on thr rhs of Eq. 3 through Eq.2.
    Moreover, except when you do spectral function, you use \omega = \varepsilon_nk.

  2. Yes, in the case of carrier dynamics you need an occupation that change with time. We did not investigate that in the EPW paper.

Best,

Samuel

5

Dear Dr. S. Ponce,

Thank you very much for your quick reply.
Now I understand that Eq. 3 in much better way.
Enjoy your weekend.

Best wishes,
Dinesh
Uni. of Konstanz