[Ifeffit] delr and ss parameterization

[Bruce Ravel]

On Friday, March 30, 2012 12:57:21 am Ebrahim Rezaei wrote:
> I am working with bulk and nano structures of MnO, MnO2
> and Mn2O3. I have two questions regarding how to
> parameterize delr and ss in Artemis.
> 
> 
> It is recommended to
> use either delr=alfa*Reff or relate it to unit cell structure. Can anybody
> tell me how to do the second approach?
> 
> 
> Also is that a good idea to let all ss to float? Or it should
> be constrained based on type and distance of atoms? I have seen an example
> about ZnO just letting the first two shell’s ss to float and let the rest
> to be the same. I would appreciate it very much if anybody could help me
> with this too.

The alpha*reff trick is only strictly valid for cubic, monoatomic
materials.  For everything else in the world, it is an increasingly
poor approximation.  That's not to say it is not a helpful
parameterization, just that you need to be mindful of its limitations.

As for the sigma^2 parameter, you have to balance when is physically
reasonable with what can be reliably extracted from a real data
analysis.  

  - Two atoms of a different species at the same distance should have
    different sigma^2 because they have different masses.

  - Two atoms of the same species at different distances should have
    different sigma^2 because sigma^2 is a mean square deviation in
    distance.

So, in some sense, all scatterers might have their own sigma^2
parameters.  But that is not to say that you data are sufficient to
independently determine that many parameters.  When you run up against
the limited information content of your data, that is when you start
making approximations of the sort that lead to setting sigma^2
parameters for different scatterers to be the same.

However, there are sensible and physically justifiable reasons why the
sigma^2 for certain multiple scattering paths should be related to
those of single scattering paths.  See
http://dx.doi.org/10.1103/PhysRevB.54.156

Two of my favorite papers that show excellent data analysis solving
real-world problems and which are written in a way that is of
pedagogical value for XAS beginners are Scott Calvin's paper on doped
ferrites (http://dx.doi.org/10.1103/PhysRevB.66.224405) and Shelly
Kelly's paper on uranyl binding to biomass
(http://dx.doi.org/10.1016/S0016-7037(02)00947-X).  For more
interesting papers, look through the archives of the mailing list and
see who gives good answers to questions, then look up their papers.

Another extremely useful resource which includes extensive discussion
of the sorts of things you are asking about is Shelly's review
chapter.  Here's the amazon link to that book:
http://www.amazon.com/Methods-Soil-Analysis-Part-Mineralogical/dp/0891188460/
The book is a bit pricy, but if you are a clever googler, you may find
a PDF copy of her chapter.

Regards,
B


-- 

 Bruce Ravel  ------------------------------------ bravel at bnl.gov

 National Institute of Standards and Technology
 Synchrotron Methods Group at NSLS --- Beamlines U7A, X24A, X23A2
 Building 535A
 Upton NY, 11973

 My homepage:    http://xafs.org/BruceRavel
 EXAFS software: http://cars9.uchicago.edu/ifeffit/Demeter