UT theoretical chemistry code forum

Hello,

I tried to calculate the Bader charges for a number of systems, but I keep getting some very bizarre charges. This is an example ACF. dat file from a 4 layer 2x2 Ag(111) slab.

# X Y Z CHARGE MIN DIST ATOMIC VOL
--------------------------------------------------------------------------------
1 0.0000 0.0000 23.8566 46.4902 1.3046 72.6097
2 2.9416 0.0000 23.8566 46.4902 1.3046 71.3996
3 1.4708 2.5475 23.8566 46.4902 1.3046 70.3827
4 4.4123 2.5475 23.8566 46.4902 1.3046 69.1742
5 1.4708 0.8492 21.4774 50.2055 1.2829 17.7726
6 2.9416 3.3966 21.4774 50.2055 1.2829 17.7726
7 4.4124 0.8492 21.4774 50.1980 1.2829 17.7179
8 5.8831 3.3966 21.4774 50.1980 1.2829 17.7179
9 2.9416 1.6983 19.1127 770.5519 1.2886 17.7519
10 4.4123 4.2458 19.1127 770.5519 1.2886 17.7519
11 5.8831 1.6983 19.1127 770.5519 1.2886 17.7519
12 7.3539 4.2458 19.1127 770.5519 1.2886 17.7519
13 0.0000 0.0000 16.7236 770.5857 1.3084 74.5285
14 2.9416 0.0000 16.7236 770.5857 1.3084 73.1634
15 1.4708 2.5475 16.7236 770.5857 1.3084 72.1133
16 4.4123 2.5475 16.7236 770.5857 1.3084 70.7498

What am I doing wrong?

There looks to be some problem with the normalization or the input file. We'll need to know how the input file was generated, or see it.

If it was from vasp, and you are using the core charge, make sure that you follow the instructions on this page carefully:
http://theory.cm.utexas.edu/bader/vasp.php

The integration of charge should be done on the valance charge density, and the charge should sum almost exactly to the total number of valance electrons.

If the charge density was converted from some other program, we're always interested to hear about problems that arise -- particularly with regard to the cube file format.

It was generated from VASP. I had relaxed the top two layers and used those positions. I followed the directions from the web site, using the script to add the valence and core charges to generate the CHGCAR_sum. I didn't play around with the FFT grid dimensions, however. I guess I should have tried, but I don't have a feel for how large to make them. Should this be my next step? Thanks!

No, I think that something is very strange. These charges are far too large to make sense.

Are you using vasp5? This new version has a slightly different format and is not read properly by our program yet.

If you can make the file available to use, we'll take a look at it.

Sorry for the long lapse.

I'm definitely not using version 5. It's the latest version of vasp 4.6 (I think 4.6.26, but I'm not sure). Here's the INCAR file for Ag(111).

SYSTEM = BADER CHARGE
ISTART = 0
ENCUT = 400 eV
PREC = Normal
GGA = 91
ISMEAR = 0
SIGMA = 0.2
NSW = 0
IBRION = 1
POTIM = .5
NELMIN = 4
EDIFFG = -0.01
LDIPOL = .TRUE.
IDIPOL = 3
VOSKOWN = 1
IALGO = 48
LREAL = Auto
LAECHG=.TRUE.