[Pw_forum] smearing at zero K

[Paolo Giannozzi]

Jiri Houska wrote:

> when running a wavefunction optimization for a metal
> at zero K, should the result be the same with 
> (1) occupations smearing & practically-zero degauss
 > and with (2) occupations fixed?

no, because fixed occupations means "at each k, occupy
the N lowest states", zero smearing means "occupy all
states starting from the lowest-energy ones and going
to higher energy until you get the correct number of
electrons". For a metal the latter procedure will in
general yield a different number of filled bands for
different k.

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