[gpaw-users] Problem with charged slabs

[Heine Anton Hansen]

Hi Jaime,

Charged periodic systems can only be handled by adding a compensating
homogeneous background charge. This results in an error that scales as
V^(-1/3), where V is the volume of the unit cell. The leading error may be
removed by adjusting the average potential as described in
J. Paier, R. Hirschl, M.Marsman, G. Kresse,  J. Chem. Phys. 122, 234102
(2005). With this correction the error scales as V^(-1).

In my experience the error is still large. However, sometimes it is
possible to extrapolate the total energy to infinite volume. Maybe this is
possible for your system also.

To see if the correction above has been applied, look for the line
"Calculating charged periodic..."
in the output files.

Maybe you will find it useful to look at the source code:
https://trac.fysik.dtu.dk/projects/gpaw/browser/trunk/gpaw/poisson.py
(line 116-183)

Best Regards,
Heine



On Mon, March 15, 2010 5:50 pm, Jaime Gómez wrote:
>
>     Dear gpaw-users,
>
>     I've been employing GPAW for a few months now and  I have observed the
> following problem when attempting calculations with a charge. For
> instance:
>
>          For Au(111) neutral slab I obtain an energy of E(Au) = -99.09 eV
>          For the same Au(111) negatively charged slab (charge=-1) the
> energy is E(Au-) = -96.49 eV.
>
>     so adding an electron to the slab is endothermic, +2.5 eV, however,
> the energy released should be close to the workfunction value, -5.1
> eV.
>
>     The cell contains 36 atoms, periodic boundary conditions in three
> dimensions and probably a compensating background.
>
>
>     Is there any way to obtain the correct total energy for charged slabs?
>     Where can I read about how the compensating charge is added?
>
>     Thank you very much,
>
>     Jaime Gómez-Díaz.
>
> _______________________________________________
> gpaw-users mailing list
> gpaw-users at listserv.fysik.dtu.dk
> https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users
>