Hi Sophia,
As per my experience, you can reduce the k-mesh in the direction in which
you are increasing the size of your supercell. The k-mesh density, in
general, is inversely proportional to the size of the unit cell.
This means that you should be able to get the converged result for a single
gamma point calculation too, if you can manage to make your supercell large
enough.
Before following this suggestion, I would like you to wait to hear from
other experienced SIESTA users as I am also relatively new to this.
Best Regards,
Ravi
On Tue, May 19, 2009 at 10:01 AM, Sophia Nishad wrote:
> Hi Siesta users,
>
> I have a question regarding convergence of phonon calculation with respect
> to k-mesh.
>
> Suppose for an 8 atom (1x1x1x8) Si supercell Monkhrast k-mesh required is
> 16x16x16 for Gamma point phonons. If I now double the size of the supercell
> in one of the axial directions and have 16 atoms (2x1x1x8) how would the
> k-mesh be resolved for this larger supercell?
>
> To get the k-mesh denser like for the case of 8-atom supercell, we would
> need to use 8x16x16; but would it be necessary to use such equivalent dense
> mesh or due to the increase of the supercell size we can use less denser
> k-mesh, say, 8x8x4?
>
> Thanks in advance,
> --
> Sophia
>
