Fermi level shift to Valence Band in Hexagonal Cell

[Eugene]

Hello,

Recently I was performing VASP calculations of GaN (wurtzite, hexagonal) bulk unitcell. Each time DOS curve had Fermi level strongly shifted to the valence band maximum (VBM) at various parameters (PBE or HSE06, Gamma-centered or Monkhorst-Pack K-Point grids). I saw in literature that many people had the same effect having Fermi level shifted to VBM in DOS and band-structure calculations. I understand that placing Fermi level not at VBM, but in the bandgap is somewhat arbitrary. Very often VASP places Fermi level at the midgap point or close to it.
Why for GaN bulk unitcell (wurtzite) Fermi level is so strongly shifted to the valence band in DOS curves? Is it consequence of hexagonal geometry or some other effect? Any info or references would be appreciated. I use VASP 5.2.

Thanks.

[tlchan]

[quote="Eugene"]
I saw in literature that many people had the same effect having Fermi level shifted to VBM in DOS and band-structure calculations. I understand that placing Fermi level not at VBM, but in the bandgap is somewhat arbitrary. Very often VASP places Fermi level at the midgap point or close to it.
[/quote]
Technically, the Fermi level should be at the middle of the gap for a semiconductor. If the Fermi level moves towards the VBM, it may be due to defects, impurities or some expermental conditions that pin the Fermi level there.

From the perspective of a computer program, the Fermi level is determined by the total number of electrons in the system. Depending on the exact algorithm of computing the Fermi level, it can land anywhere within the gap.