GW corrections of tiny DFT gap

[sdwang]

Dear developers,
I am running a calculations with a tiny DFT gap (10^-2 eV) 2D system, readed from the report file (BoseTemp=0.0 eV also). But the GW correction indicates that the final gap is negative, e.g. the system is metalic. I also did the ev-GW calculations but the results still stick to the metalic behevior. How to please resolve this? (The DFT-PBE0 gives a finite gap, as large as ~0.9 eV)
Thanks!

Shudong Wang

[Daniele Varsano]

Dear Shudong,

can you post the report file of your calculation?
Anyway, it is possible that you have a band inversion and the starting DFT calculation (LDA/GGA) is inadequate, this would explain the qualitative difference
between a hybrid calculation and the GW one. You could inspect the VBM and CBM bands differences between the two calculations to understand if this is the case.
If this is the case, then the diagonal GW correction is not enough and you could try to perform a self-consistent COHSEX calculation.

Best,
Daniele

[sdwang]

Dear Daniele,
I have inspected the results and find the band inversion in the GW output. Is it possbile to deliminate the invesion?
Thanks!
Shudong

[Daniele Varsano]

Dear Shudong,

you can try to do a self-consistent COHSEX calculation using yambo_sc. Of course, the calculation will be much cumbersome.
Other solution could be to start from a hybrid calculation e.g. PBE0.
Maybe others can provide cheaper solutions.

Best,
Daniele