[Wien] Energy level

[Lukasz Plucinski]

> The only explanation I would have (except that you have made some mistake)
> is: By choosing a particular slab with a certain number of layers you
> "define" the k-mesh in the kz direction. Suppose you use 3 layers, this
> means that you have Gamma and kz=1/3 and 2/3 (of your bulk cell) in your
> Gamma-point of the slab.
> If it happens that the VBMax/min is e.g. at kz=0.5, then this point is NOT in
> your k-mesh and the corresponding width is smaller.
> Maybe an even number of layers helps in this case, but this needs further
> analysis.

Hello,

In this e-mail I discuss some of my results - sorry its so long...

I looked in my resutls again and I guess now everything is consistent.
Before I was comparing valence band (VB) width in bulk calculation with
exp. lattice const (5.67 A) to the slab caculation with theoretical
lattice constant (5.747 A). Theoretical lattice constant was found by
volume optimization of bulk. When I compare surface and bulk
calculations done at theoretical lattice constant is looks better. Here
are some numbers:

Latt. const.	5.67 A bulk	5.747 A bulk	5.747 slab

width at X	4.776 eV	4.599 eV	4.51 eV
width at L	5.244 eV	4.986 eV	4.87 eV