Bader Analysis on CO2

[scafurin]

Good morning,

I am trying to perform a Bader analysis on CO2 molecule as preliminary test for future calculations.
I am using VASP and the script you developped. I am using a 15x15x15 Angstrom box and a soft PAW pseudopotential.
Below you can find my INCAR file:

SYSTEM =
IBRION=-1
POTIM=0.1
ISYM=0
#ISMEAR=-1;SIGMA=.1
ISIF=0
ISPIN=2
NSW=0
EDIFF=1E-06
EDIFFG=-0.001
#ENMAX=250
PREC=accurate
LWAVE=.FALSE.
LCHARG=.TRUE.
IALGO=48
NGXF=400
NGYF=400
NGZF=400
LAECHG = .TRUE.

I sum the charges in AECCAR0 and AECCAR2 files with the chgsum.pl script, then I launch "bader CHCCAR -ref CHGCAR_sum > bader_output
Here my output:

# X Y Z CHARGE MIN DIST ATOMIC VOL
--------------------------------------------------------------------------------
1 15.000000 15.000000 14.999505 7.039676 0.750495 36.095836
2 0.000000 0.000000 2.351745 7.035347 0.739245 36.049852
3 0.000000 15.000000 1.175625 1.898489 0.377099 9.082020
--------------------------------------------------------------------------------
VACUUM CHARGE: 0.0265
VACUUM VOLUME: 3293.7723
NUMBER OF ELECTRONS: 16.0000

I calculate the atoms charge as :q= valence electrons - charge (valence electrons: 6 for O and 4 for C).
The atoms charge are reasonable.
The problem is that the sum of the charges is not 0, but 0.026.
I tried to change the K-points (until 6x6x6), to put an hard pseudopotential, and to change the NGXX points, but nothing changes. The sum in all the cases is always 0.026. The sum does not change, but the charges yes, even if not so much. Is there any mistake in my calculations?
Thank you very much in advance.

[graeme]

Use the "-vac off"

[scafurin]

Ok, it seems to work good now! Thank you very much, you have been vert kind!