Vacuum charge slab model cell

[unk]

I have just tryed bader to analyse surface adsorption. My slab model cell contains thick (i hope sufficient) vacuum region to avoid interactions between slabs in periodic code. I analyzed all-electron valence charge density and got AVC file with vacuum charge
VACUUM CHARGE: 0.9912
VACUUM VOLUME: 6920.9169
NUMBER OF ELECTRONS: 137.9998 (real number 138)
And adatom seems to lose electrones, but it is not experimental-like result. Than I made -vac off bader, and adatom's charge came better.
What the nature of this value? is it real electronic density spreaded on vacuum? Does it mean that vacuum region in cell is too small?

[graeme]

Use the "-vac off" flag for higher accuracy. This has no bearing on the size of the vacuum region, the setting to ignore charge in the vacuum is only to improve the speed of the Bader calculation.