[Pw_forum] Work function for Al100 in example files

Nicholas Edward Miller nedward at MIT.EDU
Fri Apr 1 00:42:09 CEST 2011


Hi Izaak,

Practically speaking the input values for a lot of the examples are either not
converged or do not lead to a well converged results.  This is definitely the
case for this example, where you would need a finer k-mesh for both
calculations, and more slab layers for the slab.

You are correct that having proportional k-point meshes for the corresponding
calculation should also help, and I would suggest that if you are going to use
a "bulk reference" for the Fermi energy, that you use the slab unit as your
bulk building block (you can match the k-point in plane exactly, and you can
have exactly matching FFT grids).  Although in my experience the use of the
bulk fermi energy in not strictly necessary, with enough slab layers you can
take the fermi energy and vacuum potential all from the slab calculation.

Hope that helps,

Nick

Quoting Izaak Williamson <izaakw89 at yahoo.com>:

> Dear all,
>
>
>
> I was looking at the example of the work function calculation for Al100 and
> noticed that the following celldm(1) and kpoints are used:
>
>
> For slab ---> celldm(1)= 5.4235090117,   kpoints=3x3x1
> For bulk ---> celldm(1)= 7.67000000,       kpoints=3x3x3
> With these values, one gets work functions of Wf1=4.15219 eV and 
> Wf2=4.54641 eV.
>
>
> Since, celldm(1) for slab calculation is smaller than celldm(1) for bulk,
> shouldn't we use a higher k-mesh for the slab calculation? Is this correct?
>
> When I use a kmesh of 4x4x1, I get very different values of work functions:
> Wf1=4.0412 eV
> Wf2=3.82069 eV
>
> Does this make sense?
>
> Thank you.
>
> --
> Izaak Williamson
> Research Assistant
> Physics Department
> Boise State University




More information about the Pw_forum mailing list