[Ifeffit] Temp effect

Matt Newville newville at cars.uchicago.edu
Fri Jun 17 15:52:08 CDT 2011

```Hi Hana,

On Fri, Jun 17, 2011 at 2:30 PM, Hana <hana63.new at gmail.com> wrote:
> Dear EXAFS experts,
> I am quite a beginner, and that's might be an ignorant question but really
> important to my work, so I would be happy to get a direction:
> what would be the right way to weigh in the effect of temperature in soils
> samples, since solutions like Debye models will not fit to my understanding?
> specifically, I am trying to compare my measurements(collected at 10K) to
> literature data of these metals adsorbed on iron oxides (collected at room
> temperature).
>
> I would appreciate any advise,
> Hana

The Debye or Einstein model describes how sigma^2 changes with
temperature for a particular bond.  That can be helpful in many cases,
of course.  But for a soil sample or ion sorbed to metal oxide
surfaces, there are a few issues with this approach.

First, metal-oxygen bonds tend to be stronger than metal-metal bonds,
and so don't show a lot of temperature dependence to begin with.

Second, the Debye (Einstein) models consider only the thermal
component of sigma^2, whereas sorbates and metals substituted in metal
oxides are likely to have significant "static" disorder, as they have
a range of bond lengths.  In addition, it would probably be reasonable
to assume that for a metal sorbed on an insulating surface that each
bond would have it's own strength and so temperature dependence of
sigma^2.

Third, soil and sorption samples tend to be heterogeneous, and so
average over an unspecified number of local environments, further
contributing to static disorder.

It's a fine idea to measure any sample at low temperature as the
sigma^2 are smaller, and radiation damage reduced.  But I'm not sure
I've seen a work that compares sigma^2 for different temperatures in
soil samples or really even tries to interpret sigma^2 for such
samples.  I wouldn't expect it to tell you much, but I'd be happy to
be shown otherwise.

Cheers,

--Matt

```