GW for 2D or 1D materials

[Dhanjit]

Dear Daniele,
Thank you for your kind reply. It has cleared my queries.

[Dhanjit]

Dear Daniele,
In the HF calculation I get the following informations:

[X] Direct Gap : 1.681847 [eV]
[X] Direct Gap localized at k-point : 31
[X] Indirect Gap : 1.668276 [eV]
[X] Indirect Gap between k-points : 31 5
[Hartree-Fock] Direct Gap : 9.895720 [eV]
[Hartree-Fock] Direct Gap localized at k-point : 61
[Hartree-Fock] Indirect Gap : 9.895720 [eV]
[Hartree-Fock] Indirect Gap between k-points : 61 61

For the further calculations should I take K-point 31 or 61 for direct band gap?

[Daniele Varsano]

Dear Dhanjit,

the gap indicated in the report file can be misleading unless you have calculated the HF correction for all the K points of your grid.
I suggest you to calculate it for few bands across the gap to all the k point and see if the gap character does really change.
The problem here is that if you do calculate it for only few k points/bands, when recalculating the gaps KS and HF energies are mixed and the report is erroneous (this is a part of the report we should correct).

Best,
Daniele

[Dhanjit]

Dear Daniele,
This is the input file for the HF calculation. I have considered all the k-points and the bands under consideration are the filled band and start of the empty band.

HF_and_locXC # [R] Hartree-Fock
EXXRLvcs= 10 Ry # [XX] Exchange RL components
VXCRLvcs= 5000 RL # [XC] XCpotential RL components
%QPkrange # [GW] QP generalized Kpoint/Band indices
1|164|252|253|
%

Should I do the HF calculation considering a higher number of bands? like 250|255.

[Daniele Varsano]

Dear Dhanjit,
indeed, it is possible that higher conduction band and lower valence band (not HF corrected) are seen now forming the gap.
You can either inspect if this is the case, looking at the energies in the report file, either add more bands in the calculation and see if the gap do change.
Actually, I think that this is what's happening as HF provide a very large correction.
You can also inspect the HF corrected bands vs KS bands, if the profile of the bands is similar, than you can expect that correction effect is a scissor plus eventually a stretching of the bands. In this case, the KS bands are qualitatively correct, i.e. the correction does not change the character of the gap.
Best,
Daniele

[Dhanjit]

Dear Daniele,
I have tried using bands 250|255, yet the results are same

[X] Direct Gap : 1.681847 [eV]
[X] Direct Gap localized at k-point : 31
[X] Indirect Gap : 1.668276 [eV]
[X] Indirect Gap between k-points : 31 5
[Hartree-Fock] Direct Gap : 9.895720 [eV]
[Hartree-Fock] Direct Gap localized at k-point : 61
[Hartree-Fock] Indirect Gap : 9.895720 [eV]
[Hartree-Fock] Indirect Gap between k-points : 61 61

I think HF band nature differs from KS bands. What do you suggest?

[Daniele Varsano]

Dear Dhanjit,

please note that this is just a matter of reporting for the gap (I will fix it soon). You have the entire result of the calculation in the report, so you can
plot the KS and HF band structure and see if it is just a matter of bad reporting, i.e. higher KS band that are found below the HF ones, or it is a meaningful result.

Best,
Daniele