[Wien] bandstructure on slabs

[Michael Gurnett]

Yes, first thing I did was optimised the bulk unit cell, than created slab
and then ran mini to minimise forces. The position of the surface atoms
agree well with previously reported data. The structure in the z direction
does not have inversion symmetry, which means the only way I could have the
same reconstructions on both sides is to created an artificial intephase in
the middle in which I simply mirror the slab (this did not seem like a good
idea at the time).

The slab is relatively thick so there is no "leackage" between the H side
and the top "interesting side". I'll give it a go at see what I can figure
out.

Thanks for the helping words

Mick


----- Original Message ----- 
From: "Lukasz Plucinski" <pluto at buphy.bu.edu>
To: <wien at zeus.theochem.tuwien.ac.at>
Sent: Saturday, November 08, 2003 8:46 PM
Subject: Re: [Wien] bandstructure on slabs


> > I think you can remove surface state in the data file after you fingure
out
> > which one is surface state. I usually use atomic LDOS to locate the
surface states'
> > location in the band structure and removed it from original data file.
> >
> > Good luck.
> >
> > Xiliang Nie
>
> Hello,
>
> I make slab band structure calculations. WIEN is very good program to do
> it, since it can calculate automatically k- and eigenvalue-resolved charge
> inside the spheres. This gives you automatically surface states - its
> implemented in w2web, and one can even make plots immediately and
> automatically (called character plots).
>
> I don't think it would be easy to separate H related states from the rest
> of states. This is because none of the states are localized just on one
> atom - surface states are the ones that have significantly more charge on
> surface atoms (or/and first, second layer). But they still have some
> amount of charge 'inside' the slab.
>
> What you could do (and what is typically done) is to find out the surface
> states one the interesting side of the slab (i.e. the side without H
> atoms). Just make some arbitral criteria and seprate the surface states
> from the rest of states. I don't know how much you know - technically at
> first step you should play with the parameters in case.insp file
> (mainly in line  # jatom, jtype, size of heavier ...) - then plot the band
> structure. Later you probably need to write some script that takes the
> eigenvalues and partial densities from one of the output files (it was
> discussed in this list before which files to use).
>
> In principle you could also just plot the whole thing including H atoms,
> however, this is not the most elegant. But its ok I believe if you only
> mention it.
>
> I personally do not use H atoms to terminate slab - just use larger slab
> with reconstructed surface on both sides. Its quite longer calculation but
> as you see some problems are avoided. Its also difficult to say what
> actually H atoms do to the slab. I also have a feeling that the larger the
> slab is the easier for WIEN (or any other DFT code) to separate surface
> and bulk states. This is because larger slab means more eigenvalues, means
> more choice.
>
> Again I don't know how much you know but there are
> two more important things: do you know that its better to find theoretical
> lattice constant for the bulk material and use it in slab claculation
> (optimize in WIEN) ? Do you know you should relax the surface
> theoretically before making slab band structure calculation (mini.
> positions in WIEN) ?
>
> You don't have to do these 2 things, however, if you don't do it you
> should have a good answer why.
>
> Best regards,
>
> Lukasz
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>