[Ifeffit] XANES temperature estimation (Robert Gordon)

[Matt Newville]

Hi Gabriele,


On Mon, Nov 21, 2022 at 12:59 PM Gabriele Garofalo <
gabriele.garofalo at esrf.fr> wrote:

> Hi Gordon,
>
> I have just one spectrum per sample, all taken at different T (unknown,
> due to technical problems) and P (known). I'd like to find T and using
> an EoS is not a viable option.
> I was thinking that maybe, assuming that the Debye-Waller term is
> explained only by thermal disorder, the Debye temperature is P
> independent, and there are no structural or electronic transformations,
> I could retrieve T by fitting the Debye-Waller term and using the
> reference spectra. Is this possible?
>


It seems unlikely to me.  XANES is typically described as probing the
electronic states of a sample.  There is, of course, a strong connection
between atomic/nuclear structure and electronic structure, but the changes
in thermal structural disorder with temperature are actually pretty small.
Interference-sensitive measurements (EXAFS, XRD) will be sensitive to such
changes, but XANES is much less so.  When describing an EXAFS
Deybe-Waller model, the amplitude decays as exp(-2k^2*sigma^2).  When k is
very small (ie, XANES), you need to have sigma be very large for that to
make much of an impact in the XANES region.

Some XANES features (such as pre-edge peak intensities) can have some
temperature dependence - these are sensitive to small changes in the
overlap of specific electronic levels.  But, in general, the XANES features
don't have a strong temperature dependence.

But, I guess we should ask: do you see a significant variation in your
XANES spectra with temperature?

--Matt
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://millenia.cars.aps.anl.gov/pipermail/ifeffit/attachments/20221121/256474d3/attachment.htm>