[Ifeffit] Difference spectra- Ifeffit Digest, Vol 159, Issue 21

Tue May 31 13:20:20 CDT 2016

Hi Stu,

In full agreement with Bruce, what you want to do can be done, and it was done a few times before.
I think the first group that applied conventional EXAFS analysis to differential signals was Pettifer et al in Nature (2005), to analyze EXAFS in magnetrostrictors measured at different values of magnetic field.  Then, C. Konig et al did a similar trick for a catalyst under different atmospheres. They also used conventional EXAFS equation to fit the difference because the model allowed them to do so (the changes in the amplitude factors was also an amplitude factor and could be obtained from the fit because the fitting program does not know what the meaning of amplitude factors is, you do.

I want to warn you that normalization of the amplitude of your EXAFS data will be the trickiest part because it depends on how many absorbing species changed structure during the excitation.

In my work we also used this method in a couple of papers, Spanjers et al (for Pd nanoparticles under Ar and inert environment), and Korobko et al, (2015) using electrostrictors with and without applied electric field. The latter example is here:

https://pubweb.bnl.gov/~frenkel/CGO/CGO-modulations.pdf

The references to Pettifer and Konig works are Refs 19 and 20.

Regards,
Anatoly

P.S. If anyone is interested, a postdoctoral position has just opened up in my group, for research of electromechanical materials. Please contact me for further information: frenkel at bnl.gov

________________________________________
From: Ifeffit [ifeffit-bounces at millenia.cars.aps.anl.gov] on behalf of Bruce Ravel [bravel at bnl.gov]
Sent: Tuesday, May 31, 2016 1:44 PM
To: XAFS Analysis using Ifeffit
Subject: Re: [Ifeffit] Difference spectra- Ifeffit Digest, Vol 159, Issue 21

On 05/31/2016 01:23 PM, sb2c08 at gmail.com wrote:
> To be clearer, can I make the excited state and ground state models
> in Artemis, take the data and difference (which can be done in
> Athena) and use that as a direct comparison, perhaps as you would
> with XANES data? I understand my data but I do not understand the
> best method to try get at least a first shell fit from the EXAFS
> region, collected by taking the difference of the ground state from
> the excited state, and can this be done in anyway via Artemis through
> some data analysis.

While it seems like Artemis specifically wants traditional EXAFS data as
its input, that's not really the case.  It just wants something wiggly
that can be Fourier transformed.

While it seems like Artemis wants a full Feff calculation and all the
path files that come from it with all the degeneracies that Feff thinks
it has, that also is not really the case.

I don't know what your data are or what they look like (you haven't told
us), so I will make up a scenario off the top of my head.  Hopefully
that will still be instructive.

Suppose that our ground state is a pure metal -- say iron metal.  Iron
metal has a near neighbor at about 2.5A.  Suppose that our secondary
state includes some iron oxide -- that has a neighbor at about 1.9A and
a correspond peak at a much shorter distance.

Suppose we can guess the amount of pure metal in the secondary state.
Then we could do some simple arithmetic to isolate (approximately!) the
oxide contribution.  This will resemble an iron oxide standard, but with
much smaller amplitude.

Seems to me that this difference spectrum could be fit in Artemis by
running Feff on the oxide, parameterizing the fit much like a
conventional spectrum, but giving the model enough freedom to have a
much smaller amplitude.  That is, you cannot force the fit to have an
S02 of around 0.9 -- like a normal fit -- and the full coordination
number of the normal oxide.  Something in the fitting model must be
introduced to allow the amplitude to be small enough to fit the
difference data properly.

You, then, need to have a way of interpreting this much smaller
amplitude in a way that makes sense for the problem at hand.

To say all that another way ... your difference data has Fourier
components ... Feff provides Fourier components ... you supply a model
(and an interpretation!) that fits Feff's Fourier components to the
Fourier components in your data.  If your data can be described by Feff,
Artemis can do the describing.

HTH,
B

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

  National Institute of Standards and Technology
  Synchrotron Science Group at NSLS-II
  Building 535A
  Upton NY, 11973

  Homepage:    http://bruceravel.github.io/home/
  Software:    https://github.com/bruceravel
  Demeter:     http://bruceravel.github.io/demeter/
_______________________________________________
Ifeffit mailing list
Ifeffit at millenia.cars.aps.anl.gov
http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Unsubscribe: http://millenia.cars.aps.anl.gov/mailman/options/ifeffit