Bader failure for Si(100)-2x1

Bader charge density analysis

Moderator: moderators

Post Reply
imcnab
Posts: 14
Joined: Sat Jul 24, 2010 4:15 pm

Bader failure for Si(100)-2x1

Post by imcnab »

Hi everyone,

I have found that VASP/Bader analysis appears to give a pathologically incorrect result for the buckled dimers of Si(100)-2x1. I am using VASP 5.2.11 with the latest version of your add-ins for CI-NEB etc, and the latest version of "bader", the latest versions of your perl scripts.
(as you can see, I am pretty much totally dependent upon the Henkelman group's wonderful generosity).

It is well known in the literature that the UP atoms of the silicon dimers are partially negatively charged (between 0.1 and 0.3 excess electrons, depending
on what experiment / theory is used).

When I do a BADER analysis, I find that the UP Si atoms have about defecit of 0.25 electrons, the DOWN Si atoms have excess 0.25 electrons - the wrong way around.

I tried increasing NGxyz / NGFxyz, number of Kpoints (now 2x2x1), used both ISPIN=0 and ISPIN=1, all without greatly affecting this result.

I also did a VORONOI analysis, which gives a better result. the charge is the right way around: 0.07e excess on UP atoms, 0.07e deficit on DOWN atoms.

Probably, I am not using bader correctly - I am worried about the vacuum charge, but the output says this is small
VACUUM CHARGE: 0.1817
VACUUM VOLUME: 3987.0025
NUMBER OF ELECTRONS: 480.0000

Any advice on how to do this BADER analysis better will be most gratefully received.

Also, please can you tell me if there is an option in bader (I can't find one), to integrate charge within atom centered spheres? It might be a useful check sometimes.

Thanks in advance for any and all help,
best regards... Iain

DETAILS::::::::
I attach the INCAR/POSCAR/KPOINTS files below.

CHGCAR_sum was generated using your script
chgsum.pl AECCAR0 AECCAR2

I used bader with the following flags:
bader CHGCAR -ref CHGCAR_sum -vac auto
OR bader CHGCAR -ref CHGCAR_sum -vac auto -n bader -c voronoi
Attachments
MCNAB_bader_problem.7z
zip of the INPUT files for VASP
(73.99 KiB) Downloaded 1019 times
kai
Posts: 56
Joined: Sun Oct 31, 2010 11:01 pm

Re: Bader failure for Si(100)-2x1

Post by kai »

I think i finally understand it now, perhaps....

Bader is talking about excess valence electron-charge. so this is a relative quantity; a positive bader charge could mean an absolute negative charge.
Post Reply