The ABINIT Users Mailing List ( CLOSED )

[matthieu verstraete <matthieu.jean.verstraete@gmail.com>]

Nicola is correct, these are free electron like (FEL) states, which I
have found as well, and also in a few published results see PHYSICAL
REVIEW B 68, 155424 2003 fig 7 (or PRL 101, 226405 (2008) fig 1). For
the GW correction I do not find the same sign as they do in the prb,
but it depends delicately on how fully converged you are wrt the FEL
states. The calculations always give perfectly reasonable BS near the
Fermi level. The states will keep coming down and create a dense
continuum as you increase the cell size. The lowest FEL state will
converge, but as you add bands and vacuum the rest will densify and
superimpose on the "normal" bandstructure of graphene. I insist they
are physical, and constitute a kind of unbound continuum of states,
corresponding to the interlayer graphite states.

Why VASP does not have them I don't know... very strange... perhaps
the way the states are represented based on the PAW, but that would be
wierd. The scf algorithm is a possibility, if for some reason it
prefers localized states and misses lower energy delocalized ones. How
far have you pushed the vacuum? I would expect problems with, e.g.,
siesta for delocalized states far from the atoms, but not with another
PW/PAW code.

Matthieu



On Wed, Aug 19, 2009 at 10:56 AM, Nicola Marzari<marzari@mit.edu> wrote:
>
>
> Dear Aaron,
>
>
> very interesting. Could it be that those bands are free electron like bands
> that live mostly in
> the vacuum region, and are lowered in energy when they are less confined
> (more spacing
> between the confining graphene layers) ? Interlayer bands in graphite have
> been discussed since
> ~30 years ago (I believe by Posternak and Baldereschi), and 2-3 years ago in
> relation with
> superconductivity in intercalated graphite (Csanyi et al - I believe Nature
> Materials).
>
> Why they do not appear in Vasp might be related to the DIIS algorithm ? I
> believe DIIS
> is not guaranteed to converge to the sequence of lowest eigenstates, just to
> eigenstates. Can
> you use a different algorithm ?
>
>            nicola
>
>
>> On Tue, Aug 18, 2009 at 11:53 PM, Aaron Zhou<fishzhou@gmail.com> wrote:
>>
>>>
>>> Dear all,
>>>
>>> I have a question when calculating graphene band structure. I leave a
>>> 10 angstrom vacuum in Z-direction. But I saw many extra bands which
>>> don't exist when doing the same thing in VASP. So I did a series of
>>> calculation only changing the vacuum size, from 3A to 15A. I saw some
>>> excited bands are pushed down from higher energy. Even 8 angstrom
>>> vacuum and 10 angstrom vacuum will make big difference. That is quite
>>> unexpected! I think the vacuum is already big enough to see the
>>> convergence in band structure.
>>>
>>> Could someone tell me why those conduction bands are pushed down when
>>> changing the vacuum size (acell(3) in input file)?
>>>
>
>
> --
> ---------------------------------------------------------------------
> Prof Nicola Marzari   Department of Materials Science and Engineering
> 13-5066   MIT   77 Massachusetts Avenue   Cambridge MA 02139-4307 USA
> tel 617.4522758 fax 2586534 marzari@mit.edu http://quasiamore.mit.edu
>



-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Dr. Matthieu Verstraete

European Theoretical Spectroscopy Facility (ETSF)
Dpto. Fisica de Materiales,
U. del Pais Vasco,
Centro Joxe Mari Korta, Av. de Tolosa, 72,   Phone: +34-943018393
E-20018 Donostia-San Sebastian, Spain        Fax  : +34-943018390

Mail : matthieu.jean.verstraete@gmail.com
http://www-users.york.ac.uk/~mjv500