[Pw_forum] Total DOS and sum of PDOS don't overlap

Masoud Nahali masoudnahali at live.com
Sat Nov 6 13:56:02 CET 2010


Dear Luis

If I am correct , I feel that the non-orthonormality of atomic states on
different atoms may affect the projected properties like PDOS.

**
 Best Wishes
*

-----------------

Masoud Nahali

PhD Student of Physical Chemistry
Sharif University of Technology
*




>
> Stefano de Gironcoli :
>
>
> the integral of the dos on the energy axis (from -infty to +infty) is
> the total number of states in your system ... (the number of plane waves
> in your basis) .... can you figure out why ?
> the integral of the total pdos over the same interval is the total
> number of atomic wavefunctions over which one is projecting ...   again,
> can you figure out why ?
> As NPW >> NATOMWFC no surprise they differ in some portion of the energy
> axis.
> Since approximations like LCAO are usually reasonable the occupied
> portion of the energy axis is usually  well described but sooner or
> later the pdos must drop to zero while the dos likely goes on for a
> while (forever for infinite cutoff).
>
> hope this helps
> stefano
>
> Luis  wrote:
>


> > Dear QE users,
> >
> > I have performed a DOS calculation of a metal oxide system with some
> > adsorbates.
> > Then I plot the total DOS using the file myfile.pdos_tot, and I also plot
> > the sum of the PDOS generated by the program sumpdos.x included in the
> > Quantum Espresso package.
> > In principle both plots should give the same, but what I see is that the
> DOS
> > plots for the zone of the unoccupied bands do not overlap. In particular,
> > the DOS area generated with the sumpdos.x program is considerably
> smaller.
> > I can't understand why do I encounter this difference. Is there a reason
> for
> > that?
> >
> > Thank you
> >
> >
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20101106/032b9766/attachment.html>


More information about the users mailing list