[Ifeffit] XANES pre-edge peak intensities

Matt Newville newville at cars.uchicago.edu
Sat Jul 13 11:47:19 CDT 2013


Hi Paul,

On Sat, Jul 13, 2013 at 7:47 AM, Bingham, Paul A <P.A.Bingham at shu.ac.uk> wrote:
>
> Dear Ifeffit forum users
>
> I have been working on a batch of Fe K-edge XANES data for dilute glasses
> some time and I have extracted pre-edge peak positions and intensities, as
> has been widely used in the literature, to get qualitative analysis of the
> Fe coordination in my samples. I have also done this for a number of mineral
> "standards", again as has widely been done in the literature (Grandidierite,
> Yoderite, Staurolite, etc). However, I have a question: as my samples are
> considerably more dilute (ca. 0.15 mol% Fe) than all of the mineral
> standards (these vary in Fe content between a few percent and 10's of
> percent), how realistic is it to use mineral standards such as these to
> provide reference points to "calibrate" the integrated pre-peak intensities
> from the dilute samples?
>
> I have found nothing conclusive on this in the literature and would ask this
> forum for help, or any thoughts you may have on this issue. My own feeling
> is that concentration will have an effect on integrated pre-edge intensities
> and I suspect that more dilute samples are likely to give slightly higher
> integrated intensities than concentrated samples, but it is difficult to
> unpick and any thoughts or useful references dealing with such issues would
> be most welcome.
>
> Many thanks in advance for taking the time to read this message and thanks
> for your responses!
>
> Kind Regards
>
> Paul Bingham
>

I completely agree with Scott's and Carlo's points, but will add a
couple.  First, to add to what Scott said, the structures of the
absorption coefficient (including pre-edge peak intensities) are
independent of metal concentration.  The *measurements* of mu(E) in
fluorescence may not be independent of concentration, but in the
dilute limit (below 1%, typically), the correction is tiny.  But, no,
lower concentrations of Fe should definitely not give higher pre-edge
peaks relative to the edge step.

There is pretty extensive literature of analyzing Fe pre-edge peaks in
minerals and mineral glasses.  See Elizabeth Cottrell, et al in
Science from last month for example, but also the work of Andrew
Berry, et al and Darby Dyar, et al, George Calas, et al, etc -- also,
don't miss the 1997 JACS of Westre et al.  The standard practice *is*
to relate pre-edge intensities and energy centroids to values based on
higher concentration mineral phases, though many works also use
glasses with known fugacities (in the end, those point back to higher
concentration mineral phases and the XANES (and Mossbauer)
measurements of oxidation state).  That's all to say that it is
absolutely fine to use samples with higher concentrations of Fe for
standards.

However, the minerals you've picked may not be great choices as XANES
standards.  They all appear to be mixed metal-site silicates, and so
with rather complex structures and coordination environments.  While
these may be "standards" in some sense (perhaps crystals with similar
chemical compositions), generally what one wants for a XANES standard
is "rock solid" (pun intended, I am sorry to say) coordination
chemistry and local environment, including how perfect the octahedral
and tetrahedral coordination would be.  I''m not much of a
mineralogist, but it's not clear to me that the minerals you've picked
meet those criteria. Again, I'd guess they were chosen for their
composition, not local environment.   I'd also caution you to be
careful of minerals that are not centro-symmetric, as the polarization
dependencies can give changes in the pre-edges of the same size as
what you're likely looking for.  Generally, this is more of a problem
for crystalline inclusions in the glasses, and less of a problem for
standards, but do make sure to grind such minerals well and disperse
them with random orientation.

Using glasses at well-constrained fO2 or simpler silicates (say,
fayalite), or even metal oxides (hematite, wustite) may be worth
considering as standards.

--Matt



More information about the Ifeffit mailing list