[Wien] problem on electronic structure in slab

[Peter Blaha]

You have a small misconception.

Of course, if there is a state which is located somewhere (either at  the surface or in the
bulk), this state interacts with ALL other atoms. Eventually this interaction will be small,
but unless it is forbidden by symmetry, it will be there.

Thus if your surface layer has some DOS, the central layer MUST also have some (eventually small)
DOS. It is your responsibility to judge weather this is a "negligible" contribution (as you said, 10^-5)
or if it is sizable. If it is negligible, you still can interpret your results as being
halfmetallic in the bulk, but metallic at the surface .....

Am 29.07.2011 10:25, schrieb Aaron:
> Dear wien2k developers and users:
>
>      I've performed a calculation on a slab. My original purpose is to investigate the surface electronic structure, which should be different from the bulk one in my thought. The
> supercell is composed of a slab consisting 17 atomic layers and vacuum of 12 angstrom between periodic slab arrays. PBE parameterization of density functional has been used in
> calculation.
>      To my surprise, an unusual phenomena has been found in DOS (density of states) result that the position of valence band maximum (VBM) at surface and central layer are exactly
> the same. Actually, VBM in all LDOS (layer density of states) are precisely at the same energy position. Of course, the shape of DOS varies in different layers. But the edge of
> DOS peak are aligned to the same energy position from surface to central layer. That is to say, if DOS at a energy is zero in one layer, it will be zero at any other layer. And if
> DOS at a energy is not zero in one layer, it will never be zero at any other layer, even as small as 10^-5, but never would it be zero.
>      This phenomena brings a trouble that transitions such as metal to insulator or half-metal to metal from bulk to surface cannot be simulated in my slab model. For example,
> result (de Wijs and de Groot, PRB, 64, 020402) on (111) slab of NiMnSb cannot be well reproduced in my calculation, because DOS at the Fermi level in minority spin channel at
> central layer is not absolute zero in my result although it is very small.
>      So I post this letter here, asking for help. Is it a character of result from DFT program that DOS at different layer in a slab have the same edge, or did I make my
> calculation wrong?
>
> Sincerely yours,
> Aaron
>
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien

-- 

                                       P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------