[Pw_forum] bader population and pp.x nfile tool arge

[Alberto Otero de la Roza]

Hi Aldo,

The critic2 program can help you with this by calculating a core
contribution using its internal density tables:

https://github.com/aoterodelaroza/critic2

If in your system you have, say, C, H, O, and N, and the pseudos
represent the usual number of electrons, you can do:

crystal bleh.scf.out
zpsp c 4 h 1 n 5 o 6
load as core 100 100 100 id rhocore
cube grid field rhocore file core.cube

This writes a cube file ("core.cube") with 100 points in each
direction, zero core electrons on the H, 2 on the carbon, etc. You can
also load the pseudo-valence electron density and add the core
contribution, or do the Bader analysis with a few more commands. This:

crystal rho.cube
load rho.cube core id rho
zpsp h 1 c 4 n 5 o 6 
cube grid file rho_with_core.cube field rho

writes a cube file containing the density in rho.cube plus the core
contribution mentioned above. And this:

crystal rho.cube
load rho_with_core.cube 
load rho.cube id rhoval
integrable rhoval
yt

does the Bader integration of the valence pseudo-density in the basins
of the poor man's all-electron density you just calculated. As you
say, using PAW pseudos and reconstructing the all-electron density
would be the way to go, but this approach should give you a reasonable
approximation to the Bader charges if you can't follow that route.

Best,

Alberto

--
Dr. Alberto Otero de la Roza
Department of Chemistry, 
University of British Columbia, Okanagan
Kelowna, British Columbia, Canada V1V 1V7

* Aldo Ugolotti <a.ugolotti at campus.unimib.it> [2017-05-23 10:02:46 +0200]:
> Dear QE users,
> 
> I am trying to work out the Bader analysis on my system (molecules adsorbed
> on a metal surface) with the complication that I used US pseudos and I
> would not like to re-run it with PAW ones. So I tried the easiest option,
> i.e. with the charge density as-is but the result was not satisfying,
> suggesting that the core charge profile is needed to find the correct
> partitioning. I thought that, since that the valence charge should be
> similar in PAW and US calculations, I only needed to add the core
> contribution from every single atom, which was easy to calculate with PAW.
> Going straight to the end the problem now is that I need to sum all the
> core (here reads total - valence) charges saved in as many *.pp files as
> the number of atoms. Therefore I am setting nfile=241, but pp.x says that
> it is too large. In the source code I saw that nfile is defined to be
> maximum 7 files (in version 5.4.0).
> Changing that value and compiling again should given an impossibly long (or
> large) calculation, do you advice any other solution or I only have the
> option of using PAW pseudos?
> 
> Thanks in advance,
> 
> -- 
> Aldo Ugolotti
> 
> PhD student,
> Department of Materials Science,
> University of Milano-Bicocca,
> U5, Via R. Cozzi 55,
> 20125 Milano, Italy.

> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum