change in quasiparticle band gap with kpoints

[manaswitakar]

Hi there,

I am running simple G0W0 calculation on CsSni3 cubic systems. it has 75 Kohn-Sham states using SOC. In the DFT level, I have 1300 bands in order to add more number of unoccupied bands in the Yambo run. I use 222 and 444 k-grid. With all parameters same, just by changing the k-grid from 222 to 444, the QP band gap increases by 100 meV. I am using NGBlks= 10 Ry and EXXVcs = 70 Ry. Is it normal to get this variation in gap with change in k- mesh?

Please let me know.

Thanks and Regards,
Manaswita

[Daniele Varsano]

Dear Manaswita,
yes, it is normal and k points are a critical parameter that needs to be converged.

Best,
Daniele