[Wien] questions about "supercell"

[Lukasz Plucinski]

Hello,

> 1. how can I add the "normal" H atom for termination? I was only asked if
> I add the atom at x=0 (y=0, z=0) for the termination of the layers during
> "x supercell".

H atoms are used when simulating surfaces - to passivate one side of the
slab. From your e-mail it does not look like you are looking for surface -
seems you are looking for bulk.

> 2. I read the paper you recommanded (Fu-He Wang, Peter Kruger & Johannes
> Pollmann), I don't know exactly how the authors created the super unit
> cell and terminated the layers.

They were simulating the surface.

> 3. I don't understand well what you meant by "If you simulate the surface
> z is perpendicular to the surface. You are not 'choosing' x and y - they
> are defined by the symmetry of the system".

It seems you are not simulating the surface but bulk properties of
'kaolinite'. From basics it is obvious that the supercell must be the
primitive cell for the studied SYSTEM no matter what system you are
studying. Making supercell other than primitive cell is wasting computer
time (by the way WIEN should detect primitive cell even if you set
non-primitive one in w2web). It would also lead to wrong Brillouin zone.

I don't know how far you are, so let me ask stupid question: do you know
what is primitive cell for 'fcc - cubic all faces centered' lattice? Its
the relation between the primitive and cubic cell that is important.

> Now I am dealing with mineral kaolinite, which has special layer
> structure, i.e., the AlO6 octahedral layers are seperated from each other
> by certain distances in the mineral structure. Now I try to choose one of
> such layers from the structure and calculate the electronic properties of
> the layer.  Therefore, surface characters for my layers are not
> important.

I did some work on layered crystals: PRB68, 125304 (2003). Maybe it helps
a bit. I can't help you much more - it just takes months to understand the
geometry especially for complicated lattices.

Regards and good luck,

Lukasz