[Wien] bcc.case klist

[Elisabeth Magerl]

Dear professor Blaha,

thank you for your answer. I am sorry I asked something which seems to be
already discussed, unfortunately I was not able to find it in the previous
posts.
Maybe the term 'energy surface' was misleading. I am calculating group
velocities from the bandstructure, so I have to know which k-points lie
next to each other to get the curvature right. My idea was to take the
eigenvalues in case.energy, relate them to the kpoints in case.klist and
calculate the gradient.
I used xcrysden to get a distribution of k-points along symmetry lines,
but now I would like to sample the whole Brillouin zone.

Thank you,
Elisabeth


Peter Blaha pblaha at theochem.tuwien.ac.at wrote:

As mentioned before, for centered lattices the k-mesh is generated in the
"primitive cell", symmetry is applied to reduce this mesh, and than it is
transfered into carthesian coordinates. In this step, it happens that some
k-points are given outside the first BZ. This does not influence any
results.

  > interested in the shape of the energy surface,

You mean the fermi surface ??
Consider xcrysden for 3d-plots or the fsgen program for 2d cuts.

Elisabeth Magerl schrieb:
> Dear Wien2k users,
>
> I have a problem with understanding the klist-file I get out for a bcc
> structure. I set up a case for a normal 3d tungsten structure, the
> results for the bandstructure are in good agreement with published data
> - so everything works properly.
>
> If I now look at the distribution of the k-points in my initial
> klist-file, I find that they do not resemble the irreducible wedge but
> are buidling up two isolated blocks (I attached a list of the k-points
> and a plot showing the distribution mapped on the xy-plane). It seems
> that the second block is lying outside the first Brillouin zone. Do I
> have to apply a symmetry operation to get it in the right place? I am
> interested in the shape of the energy surface, so I have to get the
> k-points related to each other correctly.
>
> Thank you in advance,
> Elisabeth