I have been able to generate the Eliashberg functions for bulk Si. However, while trying to correlate the Fan and DW data as given in the "Electron phonon effects in semiconductors..." example file (SECTION A/ REFERENCE/r-03_OMS_RIM-gw0_el_ph) for bulk Si, I am having a few confusion:
1. In my case, I did a relax, scf and an nscf calculation at shift -k point of 1, 1, 1. The MP grids were 4,4,4 under tight conditions. There are two differences between my qp report file at 0K and the one in the above example. I have observed that when I compare the qp report files of mine and the example, the difference in my file is
a). X-grid is uniform :no
b) Using the Q-grid :no
I cant understand what I am missing here. I have used the uniform grids in espresso. The use of the shift-k grid increases the CBM energy value from 0.69 eV to much higher. If I dont use shift k grid, I am not able to generate the eliashberg data at CBM and VBM! Is this normal...?
2. If I plot the Eliasberg data and compare with the pic in the tutorial, I am getting a good answer but not the same as presented there. Is this happening due to the above two facts? I took 4 q points in the phonon calculation, but at tight conditions.
3. Even the spectral functions of the CBM and VBM done at 0K and 300K from my file seems to be logical. However, due to the use of shift-k points, the energies are higher. Is this fine to proceed for BSE?
With regards,
Sitangshu