GGA direct band gap to G0W0 indirect band gap

[mailoliver]

Dear all:
I try to calculate the quasiparticle band structure of monolayer MoSe2. I use pwscf to product the database for yambo. The GGA band structure shows that monolayer MoSe2 is direct band gap at K point. However, G0W0 band structure shows this material is indirect band gap, VBM at K point , CBM at gamma-K. I think there is something wrong in my input file of G0W0 calculation. The input and output files are in attachment.

Thanks,
Tieyu Lu

[Daniele Varsano]

Dear Tieyu Lu,
I'm not expert on this system, so then may be some other can tells you more looking to your band structure.
I can't see anything wrong in your input, anyway at first glance,
here the problem it looks that the interpolation of ypp failed.
This utility has to be improved:
Anyway you can have look to this thread
this one or this one
Here you can find some information that can help you to do a better interpolation.
As first attempt I would try to remove the time-reversal symmetry from your system. This can be done by using:

Code: Select all

ypp -n

and uncomment the time-reversal removal. Then run ypp, and redo the setup in the new directory created by ypp.
Anyway I'm not totally sure it will work.
If you are just interested in the gap only and not in the whole band structure, you can include the K points in your sampling: you can use:

Code: Select all

ypp -k

to generate the K-grids containing the desired point.

Hope it helps,
Daniele

[mailoliver]

Thank you for your reply.

Tieyu Lu

[mmarsili]

Dear Tieyu Lu,
I could not find anything wrong in your GW input file as well.
Take into account that the direct-indirect character of the gap for this system is strongly dependent on the xc functional and on the details of the geometry of the slab (lattice constant, position of the S atoms).
Maybe you can look in the literature what functional is used and what lattice constant.
I hope this helps
Margherita