Dear Ho,
In electronic structure calculations with atoms fixed at their positions the mass does not play a role, so anything can be written there.
Yes, it is normal to use a smaller grid on the nscf step before wannierization
Different pseudos may include or not more deep core states. Probably some states are presented by a large negative number (like -100eV), so they did not fit into the format. Anyway, such states should be excluded from wannierization
Stepan Tsirkin,
Centro de Física de Materiales
San Sebastián, Spain.
12 сент. 2024 г. 18:40 пользователь Ho Viet Thang - Khoa Hoa <[email protected]> написал:
Dear Wannier developers,
I have tried to run the Wannier interpolated electronic band structure of Cu in tutorial6 with the input files given in this link:https://github.com/wannier-developers/wannier90/tree/develop/tutorials/tutorial06. I realized that:1) the mass of Cu in scf (63.546) and nscf (55.85) file is different and the number of kpoints in scf (16x16x6) and nscf (64) are also different.2) I used ultrasoft pseudo cu_pbe_v1.2.uspp.F.UPF, this gave rise to completely different band structures as Cu.pz-n-van_ak.UPF pseudo (even the copper_band.gnu file did not present the lowest value in yrange, just *******:)
Would you please so kind give me some explanation on these issues?Thank you very much in advance
--------------------------------------
HoThe University of Danang, Vietnam
_______________________________________________ Wannier mailing list [email protected] https://lists.quantum-espresso.org/mailman/listinfo/wannier
