[Ifeffit] Ifeffit Digest, Vol 49, Issue 20

Bruce Ravel bravel at anl.gov
Thu Mar 29 17:18:51 CDT 2007

On Thursday 29 March 2007 16:58, Ricardo Faccio wrote:
> I am performing a refinement for a  La2CuO4 in powder.
> I followed Bruce's course, and I tried determine S02 hopping its
> correlation with sigma^2.
> I am attaching a graph of this determination. Please see enclosed the file.
> My question is: Is common to obtain a S02 of 0.625? and is common to obtain
> a nearly sigma^2=0?

That doesn't seem right.  So2 and sigma^2 are correlated such that a
low value of one can be compensated by a low value of the other.  I'd
guess that is happening here.

> I didn't fit E0, alpha and degeneracy for the first single shell. What is
> the best strategy to fix this bad result?

It is worth mentioning that the document you are referencing is rather
old.  The issue of the technique you are asking about has come up
several times on this mailing list with various people casting doubt
on how reliable or believable it is.

I certainly prefer to use other tactics these days.  Much more
reliable, to my mind, is the multiple k-weights approach to the fit
where you set up the fit with all the parameters you think you need
then tell Artemis/Ifeffit to use many values of the k-weight.  See
page 8 of http://cars9.uchicago.edu/~ravel/talks/artemis.methyltin.pdf.

I think that is a much more reliable approach to figuring out the
correlations in So2 and sigma^2 in that the rest of the fit happens at
the same time.  As you suggest, not fitting the other parameters can
have an impact on your results.


 Bruce Ravel  ---------------------------------------------- bravel at anl.gov

 Molecular Environmental Science Group, Building 203, Room E-165
 MRCAT, Sector 10, Advanced Photon Source, Building 433, Room B007

 Argonne National Laboratory         phone and voice mail: (1) 630 252 5033
 Argonne IL 60439, USA                                fax: (1) 630 252 9793

 My homepage:    http://cars9.uchicago.edu/~ravel 
 EXAFS software: http://cars9.uchicago.edu/~ravel/software/

More information about the Ifeffit mailing list