Workfunction of compound slab

[qql_2011]

When I try to simulate the workfunction of the compound slab like TiN, the vacuum level split to two sides (like dipole)if I choose the same number of atoms for Ti and N. And the delta of the energy is big: ~3.2eV. Isn't it artificial? Since the dipoles should be cancelled due to the symmetry. So how can I get an unique workfunction number as in real case? Thanks!

The POSCAR I used is as below:
fcc:
4.18
0.70710678 0.0 0.0
-0.35355339 0.6123724 0.0
0.0 0.0 5.1961523999999998
4 4
Direct
0.0 0.0 0.0
0.33333 0.66666 0.11111
0.66666 0.33333 0.22222
0.00000 0.00000 0.33333
-0.33333 -0.66666 0.055555
0.0 0.00 0.16666
0.33333 -0.33333 0.27777
-0.33333 -0.66666 0.38888

The INCAR file is:
ENMAX = 400
general:
SYSTEM = clean TiN (111) surface
ISTART = 0
ICHARG = 2
ISMEAR = 2 ; SIGMA = 0.2
ALGO=V
NBANDS = 35
special:
LVTOT = .TRUE.

And KPOINTS file is:
K-Points
0
Monkhorst-Pack
11 11 1
0 0 0



<span class='smallblacktext'>[ Edited ]</span>

[tlchan]

From your POSCAR, it seems like one side of the slab is terminated by Ti, while the other is terminated by N. This gives rise to the dipole you observed. You can try the LDIPOL flag mentioned in the manual to obtain the work function.