[Ifeffit] [IFEFFIT] question about chemically unordered systems

[Matt Newville]

Hi Falk,

On Tue, Aug 7, 2012 at 8:42 AM, Meutzner, Falk <f.meutzner at hzdr.de> wrote:
> Dear IFEFFIT Community,
>
> is there a way to calculate the spectrum for chemically disordered systems?
> We have a simple binary bcc structure. It crystallises either in
> superstructure (CsCl) or chemically disordered (W type) with both atoms
> randomly taking part of either (0,0,0) or (0.5,0.5,0.5). It is easy to get
> the superstructured model for artemis, but we have no clue how to model the
> disordered system...
>
> Thanks a lot for your help, we really appreciate it.
>
> Best regards
> Falk

Feff generates the EXAFS contributions for an individual atom in a
cluster of atoms.  To model a range of local structures that an atomic
species might have in a disordered (or even moderately complex ordered
structure, such as two inequivalent sites in a unit cell), one has to
do a calculation for each candidate environment for the absorbing
species and average the results accordingly.

To model the disordered system, I think you would need to run one
calculation with the central atom at the (0, 0, 0) site and one at the
(1/2, 1/2, 1/2) site.  You can then add the weighted contributions
from the corresponding paths from each calculation.     One of the
things EXAFS might be able to distinguish (depending on the Z contrast
of the atoms in your structure) is whether one site is preferred over
another.

I hope that helps get you started.  If not, perhaps a more concrete
question would help.

--Matt