[Pw_forum] Fermi Energy

[Paolo Giannozzi]

On Thu, 2013-01-31 at 17:58 +0330, mohaddeseh abbasnejad wrote:

> Regarding some calculations for semiconductors, the calculations are
> not converged and we need to consider smearing.
> In that case, in the output file, it prints the Fermi energy instead
> of highest occupied and lowest unoccupied energy.
> How should we interpret the result?

if the semiconductor has a gap, the Fermi energy will be 
somewhere in the gap. Its exact position will depend upon 
the smearing and the gap. It can be anywhere in the energy
region of the gap where the (smeared) DOS is negligibly small.

> On the other hand, when calculating the electronic density of states,
> the integrated DoS doesn't reach to the total number of valance
> electrons exactly near the Fermi energy.

the integrated DOS is not calculated exactly in the same way 
as during the self-consistent calculation, so a small difference
is to be expected.

> In that case, how should we consider the Fermi energy and obtain the
> band gap? 

by counting the occupied states and looking at the highest occupied
and lowest unoccupied electron state

P.

-- 
Paolo Giannozzi, IOM-Democritos and University of Udine, Italy