[Ifeffit] Energy shift

[Bruce Ravel]

On Tuesday, October 04, 2011 09:25:18 pm JeongEunSuk wrote:
> Hell all
> I have the study for PtO nanoparticles with EXAFS.
> When I fitted the data to model, I had a problem for energy shift.
> I thought that the energy shift obtained from fitting must be below White
> line. However it was over white line. Although I removed background
> carefully and changed bond length in model, the energy shift was still
> big. I want to know other factors which affect energy shift.

As Darek said, it is hard to know what you are talking about without
some more information.

There are a lot of things that can lead to weird values of E0 in an
EXAFS fit, for instance:

 * You expect an oxide, but your sample has a large metallic component

 * You expect an oxide, but your sample has a large sulfide component

 * You run Feff on the Pt K edge rather than the Pt L3 edge, resulting
   in an incorrect phase shift in the theory

 * You mistyped a number while inputing crystal data into Atoms,
   resulting in icorrect distances in the feff.inp file

 * You took crystal data from an untrustworthy source,
   resulting in icorrect distances in the feff.inp file

 * Your material as a nanoparticle is substantially different from the
   bulk analogue

I, of course, have no idea if any of those suggestions apply.  The
bottom line is that a disturbingly large E0 value is usually a
suggestion that you have made an error of procedure or assumption
while creating your fitting model.  Sometimes it indicates a
systematic problem in the data collection -- certain monochromator
problems can introduce a periodic but non-EXAFS signal into the data.
On occassion it even suggests a bug in the software -- but usually
not.  Usually there is some problem in the fitting model.

I should also point out that going through the excercise of
documenting the problem clearly, as Darek suggested, is valuable not
only because it provides the information that makes it likely that
someone can actually answer your question, but because doing so forces
you to carefully examine your work, which may turn expose the problem.

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