[Ifeffit] Theoretical value of S02 for different absorber

[Matt Newville]

Hi Jatin,

On Thu, Apr 5, 2012 at 5:44 AM, Rana, Jatinkumar Kantilal
<jatinkumar.rana at helmholtz-berlin.de> wrote:
> Dear Dr. Ravel,
>
> Thank you very much for your prompt reply. I followed the instruction on (http://leonardo.phys.washington.edu/feff/wiki/index.php?title=EXAFS_calculation) and activated S02 card by making S02 0.0 in feff.inp file. As a result, FEFF calculated S02.
>
> However, I get confused when I compare FEFF calculated S02 with reported value in literature and the one obtained from my own fit.
>
> For example,
>
> S02 for Mn absorber
>
> FEFF calculated                           0.94
> Fitted value for my sample               ~0.75
> Fitted value for various Mn-oxides       ~0.75
> Other reported value from Mn fit         ~0.70-0.80 (http://www.sciencedirect.com/science/article/pii/S0020169397057496)
>
> S02 for Fe absorber
>
> FEFF calculated                          ~0.94
> Fitted value for Fe-foil                 ~0.70
> Fitted value for various Fe-oxides       ~0.70
> Reported value by calculation             0.69 (http://prola.aps.org/abstract/PR/v169/i1/p27_1)
>
>
> It is well known that S02 is a property of central absorbing atom
> and is considered to be constant between 0.7 - 1.0. However, in real
> world, S02 is affected by various systematic errors like thicker
> samples (in transmission), absorption by other elements in sample at
> given energy range and many more. These factors can suppress the
> amplitudes of EXAFS oscillations resulting in lower value of fitted
> S02 than it should be for a given absorber.

Yes, that is all correct. You don't mention it explicitly, so I'll
also remind you that energy resolution also couples closely with S02
(especially if you're not carefully investigating energy resolution)
and poorer-than-ideal resolution would also tend to make S02 lower
than the "theoretical" value, which takes into account only the
passive electron effect.  Taken together, these could easily explain
why you see values closer to 0.7 than 0.9.


> The reason I am so much curious about S02 is, if I know what should
> be (or should be considered) the value of S02 for a given absorber
> than I can possibly understand the reasons of fitted S02 being too
> low or too high. For instance, S02 and Sigma2 are highly correlated
> for first shell. So from fitted value of S02 and Sigma2 of first
> shell, I could possibly understand changes in local structure.

Yes, this is why it is often recommended to measure a "standard" (a
sample with well-constrained first shell coordination) so that an S02
that includes many of the above effects (notably, energy resolution)
can be determined.

> So I would like to put forward very simple but still confusing
> question that which value of S02 should be considered, FEFF
> calculated or the one obtained by fitting standards (either reference
> foils or reference compounds measured on same beamline) and why.

The latter.   It still might miss some effects that are
sample-specific (pinholes, over-absorption), but it does include
beamline-related effects that the calculation does not account for.

Cheers,

--Matt