[Wien] slabs and cluster

[Stefaan Cottenier]

> If I want to calculate cluster or slab of BaTiO3, how
>I should make a slab(or cluster) struct file from the
>struc file of crystal BaTiO3? From supercell program, I
>should how to set parameter(repeat atoms in x,y,z
>direction? what  should the value of vavuum? )
> The mannual of supercell program is too simple.
>
> Below is structure parameter of crystal BaTiO3

First *you* have to decide which kind of surface you want to calculate. Let
us assume that you want a (001) surface (with as XYZ-axes the global
coordinate system as used in case.struct). For such a surface, you want to
introduce vacuum in the z-direction. You need a slab, so you also should
repeat your atoms in this z-direction. How much vacuum do you need? You can
e.g. test the value of :VZERO (one of the values there is the potential in
the center of the vacuum, see mailing list), which should converge to a
stable value with vacuum thickness. Or you can check the bands in a
spaghetti-plot along a direction perpendicular to the surface normal: they
shoul be flat. Or magnetic moments, field gradients,... for surface atoms
should not change any more upon increasing vacuum thickness. etc, etc. The
latter test must also be done for the slab thickness: observable quantities
for the atoms in the center of the slab should not change any more upon
increasing slab thickness.

The supercell program works only for a few types of lattices and for the
unit cell axes. For more complicated surfaces, you have to create the
relevant case.struct yourself from scratch.

See also the lecture of Peter Blaha during the latest wien workshop (can be
downloaded from www.wien2k.at) for a few slides on supercells and surfaces.

Stefaan