[Pw_forum] degauss in DOS calculation.

[O. Baris Malcioglu]

Dear Jiayu,

I think there might be a misunderstanding due to disambiguation here.

"degauss" in DOS is a parameter that controls the broadening of the
otherwise singular peaks.

while "degauss" in PW.x is something different that controls the
occupation of bands itself.

in short they serve different purposes. Or, let me say, different
approaches to complications that arises from possible over-reduction
in describing the system.

The broadening in energy levels that you see in DOS might arise from
many physical phenomena, like electron-phonon coupling (which
dominates in higher temperatures where the experiments are usually
performed, but still significant at lower temperatures), a plethora of
quasi-particle effects, crystal field effects, etc.

Often such effects are prohibitively expensive to add in a calculation
that involves even moderately sized systems.

If you can not afford such an intricate calculation, you can add
some artificial broadening to your DFT results, until it looks "good"
(i.e. something resembles more what an experimental scientist would
see).

In some cases, this works, that is, all the complicated phenomena in a
system can be reduced to a Gaussian around a certain central value,
but in some cases it does not, for example a strongly interacting
exciton may shift your energy level significantly, and if it is a
significant enough peak, you can not account it using broadening
alone.

Thus, it might not be so correct to infer if the smearing you used in
PW.x is applicable or not by interpreting how the artificial
broadening you add changes DOS.x results alone.

If you are interested in the matter, some literature can be found here
A. Marini et al. Phys. Rev. Lett. 101, 106405
Manuel Cardona, Solid State Communications, 133, 3-18 (2005).
M. Cardona et al. Phys. Rev. B 36, 4821–4830




2010/1/19 Paolo Giannozzi <giannozz at democritos.it>:
>
> On Jan 19, 2010, at 7:27 , Jiayu Dai wrote:
>
>> Thanks a lot. The figures of DOS using huge smearing is very
>> different.
>
> if you use a huge smearing, you get a hugely smeared DOS
>
>> So, i do not think we can use the "tetrahedra" method to calculate
>> the DOS
>> now.
>
> why not?
>
> P.
> ---
> Paolo Giannozzi, Dept of Physics, University of Udine
> via delle Scienze 208, 33100 Udine, Italy
> Phone +39-0432-558216, fax +39-0432-558222
>
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>