The calculated charge is quite different from the given ACF.dat in example

[cugmsn]

Hi, I am using bader to calculate the charge based on the charge density from CASTEP. We tried the AlN, Al2O3, BN, and SiO2 crystal. The bader runs very well. The calculated charge shows all atoms have much higher ionicity (As shown below) by comparing with those in NaCl example.
So, we calculated the bader charge using CHGCAR in NaCl.tar.gz(example). The result is not consistent with value in the ACF.dat. Is there something wrong?
------------------------------------------------------------
atoms charge
AlN N 7.939
Al 0.061
Al2O3 O 7.962
Al 0.056
Si2O O 7.955
Si 0.091
BN B 0.000
N 8.000
---------------------------------------------------------------

NaCl
--------------------------------------------------------------
NaCl_Example ACF.dat
Na 0.1776
Cl 7.8181/7.8297

Calculated ACF.dat using CHGCAR in NaCl_example
Na 0.000144
Cl 7.993543/8.012879
--------------------------------------------------------------
Thanks
Bingtian

[graeme]

Make sure that you are doing the partitioning on the total charge density (CHGCAR_sum) and the integration on the valance charge (CHGCAR). The appropriate command is:

bader CHGCAR -ref CHGCAR_sum

This should reproduce the results in the ACF.dat file.

[cugmsn]

Thank you! Yes, we reproduced the results.

It seems like the CASTEP_user can just get the total charge density (CHGCAR_sum) from *.den_fmt file.
I don't know how to obtain the valance charge (CHGCAR)...

[graeme]

I'm not an expert at castep so you are on your own there. You can, however, do the analysis on the total charge density, it is just typically hard to converge because of the charge density cusps at the atomic centers.