A question about SiC charge transfer by Bader

[speedxyl]

Dear all
We have used Bader to calculate the charge transfer on VASP CHGCAR of SiC system, the result is listed as below

# X Y Z CHARGE MIN DIST ATOMIC VOL
--------------------------------------------------------------------------------
1 13.0831 10.0000 1.7771 6.5321 1.0028 108.0190
2 12.6736 11.5434 4.4379 6.5294 0.9955 149.1135
3 11.5434 12.6736 1.7773 6.5294 0.9955 149.1526
4 10.0000 13.0831 4.4378 6.5321 1.0028 108.0369
5 8.4566 12.6736 1.7773 6.5294 0.9955 148.7406
6 7.3264 11.5434 4.4379 6.5294 0.9955 152.9039
7 6.9169 10.0000 1.7771 6.5321 1.0028 112.1081
8 7.3264 8.4566 4.4379 6.5294 0.9955 152.3904
9 8.4566 7.3264 1.7773 6.5294 0.9955 152.8687
10 10.0000 6.9169 4.4378 6.5321 1.0028 112.0956
11 11.5434 7.3264 1.7773 6.5294 0.9955 152.4465
12 12.6736 8.4566 4.4379 6.5293 0.9955 148.7979
13 12.9471 10.0000 5.3176 1.4715 0.6644 33.9201
14 12.5523 11.4742 2.6574 1.4688 0.6596 44.4063
15 11.4742 12.5523 5.3181 1.4688 0.6596 44.4115
16 10.0000 12.9471 2.6569 1.4715 0.6644 33.9210
17 8.5258 12.5523 5.3181 1.4688 0.6596 44.3937
18 7.4477 11.4742 2.6574 1.4688 0.6596 43.3919
19 7.0529 10.0000 5.3176 1.4715 0.6644 31.4187
20 7.4477 8.5258 2.6574 1.4688 0.6596 43.5411
21 8.5258 7.4477 5.3181 1.4688 0.6596 43.4039
22 10.0000 7.0529 2.6569 1.4715 0.6644 31.4181
23 11.4742 7.4477 5.3181 1.4688 0.6596 43.3861
24 12.5523 8.5258 2.6574 1.4688 0.6596 44.2572
--------------------------------------------------------------------------------
VACUUM CHARGE: 0.0000
VACUUM VOLUME: 0.0000
NUMBER OF ELECTRONS: 96.0000

and we have followed the instruction of http://theory.cm.utexas.edu/vtsttools/bader/vasp.php.
Does above mean the the charge transfer amount is about 2.5? If yes, it is too much for SiC system.
Please tell me how we can fix this. Thank you!

[graeme]

I don't see any problems with these numbers. The interpretation ... well, that's up to you.

[speedxyl]

We have calculated ZnO and SiC naotube by VASP, and analyzed the charge transfer by Bader. The result for ZnO is about 1.2 e, this makes sense and verified by other work. But the charge transfer for SiC is 2.5, much higher than ZnO. It is a little weird. Did we miss anything in dealing with the covalency material like SiC?

[graeme]

I really don't know. I generally use the Bader analysis to quantify changes in the charge distribution. As far as the absolute numbers, I don't have a good intuition for what to expect. You might want to look at the Bader volumes around Si and C, and just make sure they are a sensible shape -- they should, if you included the core charge. Then, I would not be surprised if Si is getting more charge simply because it is larger than C.

[speedxyl]

Maybe I don't know how to analyze the data.
Does above ACF.dat show that the net charge of C atom is about 6.53, and for Si atom is 1.47?
Thank you!

[graeme]

The net charge is the Bader charge minus the number of valence electrons modeled in the DFT calculate for the atom. If both C and Si had 4 electrons, than your analysis was correct:

C_net_charge = 1.47 - 4 = -2.53 e
Si_net_charge = 6.53 - 4 = 2.53 e

Note: this charge is the net number of electrons; the net charge is really the negative of this.