[Wien] supercell calc / impurities in V

[Stefaan Cottenier]

> I am trying to calculate a 16 atom 2x2x2 bcc supercell of V
> with a central transition metal impurity. When introducing a
> Pd or Pt impurity atom in the supercell, I expect an expansion of
>  the lattice constant. However, my calculations give a
> contraction of about -3% by volume as a result.

In order for this to be meaningful, you should have relaxed also the nearest
neighbours (at least roughly). Then, if I look at the atomic radii of V and
Pd (http://www.webelements.com/webelements/elements/text/Pd/radii.html) -- 
this is a very crude estimate of the kind of volume effects you may
expect -- it appears that Pd is indeed smaller than V.

> Another surprising result is a very atomic-like 4d (resp 5d)
> band for the Pd impurity (resp Pt) in the DOS, which shows up
> as a very sharp peak in the middle of the DOS.
> The calculations seem well converged otherwise
> (DIS in the range of 1e-3 to 1e-4) and I have no ghost
> band warnings. I use 1e3 k-points and the
> following case.in1 file (constructed by following the example
> of S. Cottenier).

Can you send the scf of your last iteration, and an example of such a DOS
(graphical)?

Stefaan