[Ying-announcements] Fwd: Can you spare some exafs advice?

[Macon Abernathy]

Hey all,

This is for your EXAFS notes. Matt provided some really helpful tips on
manipulating S0^2 and N in artemis to get fractional coordination numbers,
and also just on the fitting process in general.

Macon

---------- Forwarded message ---------
From: Matt Siebecker <matthew.siebecker at ttu.edu>
Date: Wed, Jul 29, 2020 at 3:09 PM
Subject: Re: Can you spare some exafs advice?
To: Macon Abernathy <maber001 at ucr.edu>


Hi Macron,

It's nice to hear from you. And I'm glad the tutorials are helpful. Even
with the wealth of information out there on EXAFS fitting, when it comes to
the details, it is hard to find a specific answer.

For coordination numbers of the higher shells (like 2nd and 3rd shells), I
would still employ the same approach (i.e., setting CN to "guess" parameter
in Artemis as you described and fixing SO2 to some determined value or
0.9). I this method is appropriate for all coordination shells because it
lets the software do the heavy lifting, in terms of determining if a CN for
a particular scattering path is significant or not.

I also like this approach because it can help confirm a hypothesis if the
sorption complex is mononuclear or binuclear (which in this case the CN of
the second shell Fe or Mn would be close to 1 or 2). Or, if the CN is
calculated to be considerably higher, then perhaps a surface precipitate is
forming. If the CN for those shells is fixed, then that detail may be
missed. Additionally, manually iterating between 0.1 CN units to test for
fit improvement is tedious and you might not test all the variations.

That said, you could make the argument for not doing guessing CN (i.e.,
fixing CN). You will see many examples where people fix CN based on a
structural model (which can also be rationalized). Say for example, if you
think that it is a nonnuclear complex where the second shell would ideally
be 1. Or, if it is a binuclear complex, where the second shell would be 2.
You could make the case to "set" those to be 1 or 2 to see how the fit
behaves. I think there have also been attempts to assign a % of nonnuclear
or % binuclear complexes based on more complex fitting models with As on Fe
or Mn oxides. Often, as you know, sorption is not all entirely one species,
but a mix of those species.

If you have a mixed system with Mn and Fe oxides, it will be very
challenging to distinguish between those two adsorbents.

In general, I would advise to fix all other paths in a fit to focus on one
path at a time at a time. Say, for example, fix the first shell entirely
(after obtaining a reasonable first shell fit) and if your second shell has
2 or 3 paths, then work on them 1 at a time (i.e. fix them so as to not
have too many variables). Eventually, the goal is to produce a stable fit
while floating (i.e., guessing) as many variables as possible. But,
guessing them all at once (without being close to the best fit) will often
lead to strange fits and errors that are unreasonable.

If you can rationalize the reason for a fitting method (objectively), than
it is worth a try.

Best,
Matt

Matthew G. Siebecker, Ph.D.
Assistant Professor
Applied Environmental Soil Chemistry
Dept. Plant and Soil Science
Texas Tech University
Office +1-806-834-0266http://www.depts.ttu.edu/pss/ESC

On 7/27/2020 7:13 PM, Macon Abernathy wrote:

Hey Dr. Siebecker,

I'm a graduate student in Sam Ying's lab at UC Riverside, We had a
chance to meet briefly during ACS in San Diego last August, and I've
benefited greatly from your EXAFS tutorials on youtube.

I was wondering if you'd be able to answer a question for me that I've been
having a hard time answering based on what I can find in the ifeffit
mailing list, or in resources like Calvin's EXAFS for Everyone or Bunker's
Intro to EXAFS.

Basically, how do you get fractional coordination numbers out of
Artemis, particularly for 2nd and 3rd shell atoms? (i.e maybe a measurement
support 5.6±0.4 oxygens in a first shell rather than 6, and 1.6 ± 0.2 Fe in
the second shell rather than 2)

I've seen some answers for getting a fractional coordination number for the
first shell that involves setting "N" to 1, and then setting "amp" to
"0.9*CN", but have had a hard time finding out if that's appropriate for
2nd and 3rd shell atoms, which in my systems are the atoms of interest,
usually because I'm trying to figure out what types of adsorption complexes
form between my adsorbate (which I have exafs on-often As, V, Cr, etc) and
my sorbent surface (often Mn or Fe oxides).

I'd really appreciate any advice you have.

Thanks for taking the time,

Macon


-- 
Macon J. Abernathy
Ph.D Candidate
Department of Environmental Toxicology
University of California, Riverside
https://ucrsoils.weebly.com/
<https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fucrsoils.weebly.com%2F&data=02%7C01%7Cmatthew.siebecker%40ttu.edu%7Cddca640059fa487af02708d8328b2a2b%7C178a51bf8b2049ffb65556245d5c173c%7C0%7C1%7C637314920574274284&sdata=qa3yNDqePACN4qhZBPkUoL4uaIbpvBKinzKvutSbSO8%3D&reserved=0>





-- 
Macon J. Abernathy
Ph.D Candidate
Department of Environmental Toxicology
University of California, Riverside
https://ucrsoils.weebly.com/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ucr.edu/pipermail/ying-announcements/attachments/20200731/66f8a5f2/attachment.html>