Scissor Corrections for LiF Tutorial

[c3031929]

Hi,

I've been doing calculations on LiF from scratch using exactly the same Abinit set-up as the Yambo tutorials, and I've been using the Plasmon Pole GW run level to determine the QP gap corrections, and I get a gap correction of ~4.4eV. I have attached the input file for the GW run (with NGsBlkXp= 100 (which goes up to 113)), along with the other files obtained from that run. I push the GW runs as far as NGsBlkXp= 283.

In the tutorial the values used for the QP gap correction are 5.19 for RPA and 5.80 for BSE .

1) Would it be possible for someone to explain to me how these values were obtained and give some guidance as to why my values are different.

2) Also why were different values used for the RPA and BSE calculations?

Help is most appreciated.

Regards,

Daniel Walton

[myrta gruning]

Dear Daniel

I am not sure on how the values in the tutorial were obtained, but tutorials usually have an illustrative purpose, and the values used are not at convergence, so they should not be taken as reference for calculations.
From my GW calculations on LiF I remember that at a level of G0W0 one get a smaller gap by about 0.5-1 eV with respect to the experimental value.

I recommend you to study the behavior of the gap with respect to the other parameters as well (e.g. also wrt the polarization functions and the G bands range and k point mesh).
Also if you are just interested in the gap you don't need to calculate the corrections for all the bands!!

Code: Select all

%QPkrange                    # [GW] QP generalized Kpoint/Band indices
  1| 19|  1| 50|
%

You can limit the range to bands close to the gap. In convergence studies you can also choose just a particular k-point.

Regards,
Myrta