negative DOS below the valence band

[Dhirendra]

Strangely in Ge (or in Si) I see that the DOS projected onto p-orbitals is negative much below the valence band of Ge(or Si). What could be the reason for it?

[Dhirendra]

Below -10eV, there is only one band, and if you look at the Fig 3 of this paper http://dx.doi.org/10.1103/PhysRevB.80.115205, it seems that the band is of 4s and 3d states. Is it a bug in the routine that calculates DOS?

The calculation above are simple bandstructure and DOS calculations on 2atom diamond lattice of Ge. With Silicon the same is observed.
Exc is GGA.PBE
Density mesh cutoff is 110 Hartree
Kmesh - 18x8x18 and for DOS 25x25x25
basis set is BasisLDAPZHGH.Germanium_4_Tier_4

[Anders Blom]

Small artifacts can be seen in such projections, as well as in Mulliken populations. It's not really a bug, more a numerical issue. It has no physical consequence.