Wrong VBM and CBM in mobility calculation

[sponce]

Does you calculation include SOC ?

If I understand correctly you have 32 electrons and 32 bands in your nscf. This means you only have occupied band. You have no unoccupied band.
Is this correct ?

If this is the case, you cannot expect to have a CB Fermi level.

Now if this is without SOC and you therefore have only 16 bands that are populated in NSCF.

In that case, the question is: did you skip band for doing the wannierization. If so, you have to indicate that with nbnskip.

The easiest would be to show us a plot with the nscf bandstructure and the Wannierzied bandstructure on top of it.

[sdwang]

I have not included SOC, and there is 16 bands and I set nbnd=32 in nscf. In epw.in the nbndsub=20 and nbndskip=0.
Now the problem is wrong VBM and CBM is OK since I moved scissor.
The band structures of DFT and Wannier interpolated is here., or https://drive.google.com/file/d/1bBWtmgaI8Y0f_EQHIayxxclrl6FeIxNH/view?usp=sharing
Thanks.

[sdwang]

Dear Samuel,
I use the dev-version of EPW and 'Warning: too many iterations in bisection' disappear now
but the VBM is still wrong.
Is there any suggestion?

Thanks!

SD