GW POTCAR

[ahsan_javed]

MoS2 monolayer outputs a direct band gap for simple POTCAR but an indirect band gap for GW based POTCAR. What could be the possible reason??

KPOINTS = 11 11 1

POSCAR:
Mo S
1.0000000000000000
3.1841300000000001 0.0000000000000000 0.0000000000000000
-1.5920600000000000 2.7575300000000000 0.0000000000000000
0.0000000000000000 0.0000000000000000 27.3345999999999982
Mo S
1 2
Cartesian
-0.0000125872999998 1.8383625251000000 13.6973680599999987
1.5920825873000002 0.9191674749000001 15.2486066099999977
1.5920825873000002 0.9191674749000001 12.1461295099999997

[manuel_engel1]

Hi,

Thanks for reaching out. What you describe could hint at a convergence problem that incorrectly reproduces the band gap. However, in order to know for sure, I will need some more information. Could you please upload the necessary input and output files according to our forum posting guidelines.

[ahsan_javed]

PFA

[manuel_engel1]

I ran your calculation once with the supplied GW POTCAR and once with the most recent standard VASP POTCAR of the same kind (Mo_sv and S) with the same INCAR settings, and I could not reproduce your issue. The bands are slightly different but both calculations give the same band gap. What simple "non-GW-based" POTCAR are you using? Could you please also attach the OUTCAR of the standard non-GW run to see what's different?

[manuel_engel1]

I should note that there are two peaks at the valence band minimum in MoS2 and they are extremely close in energy. Even small numerical difference could push one peak over the other, changing the band gap from a direct to an indirect one, in the strictest sense. But since the energies are so close, I don't think there is reason to worry about this difference if it is small.