[Ifeffit] How to perform normalisation and background removal of data containing L3 and L2 edges

Robert Gordon ragordon at alumni.sfu.ca
Mon Dec 7 16:35:25 CST 2020


In checking the literature, has anyone indicated what data collection 
ranges they used, and what pre- and post-edge
regions were selected for normalisation?


Do you have more data below 2800eV that you are not showing?
Starting data collection only 50 eV or so below the edge can be 
problematic for estimating background,
particularly so if a large feature (white line) is present at the edge.

Around 2810 eV, the step size has changed from coarser to finer steps in 
energy:



and that seems to be associated with a change in slope, but with more 
noise in the coarser step range.

If I had to pick something with just this data, I could tolerate using 
the first 10 eV of the scan for
the pre-edge...



If it is just this one sample, and you aren't trying to conclude 
anything from absolute intensities,
this could be reasonable. If you are comparing, be consistent in how you 
treat the data, and understand
(and admit in presenting the data) that there can be errors associated 
with normalisation.
If I vary the post-edge values a few eV to either direction, the edge 
step changes by +/- 0.03 (0.90+/-0.03).
The variation is even larger if I push the pre-edge upper limit to 
closer to the edge. (and I can convince myself
that the 1st ~15 eV of the scan are reasonable, for an edge step closer 
to 0.94)

Generally, for this and future reference, make sure you have sufficient 
data range for processing. This
will vary depending on what edge you are studying, but here, I would 
have started at least 100 eV below
the edge.

Understand that processing has an element of subjective interpretation 
to it - asking yourself what is reasonable.
Treat similar data consistently.

Probably paraphrasing Bruce a bit, but he has a mantra of "Data handling 
and interpretation should be done
in a manner that is reasonable, plausible, and defensible." Over time, 
through experience, you can gain
an understanding of what constitutes reasonable, plausible and 
defensible. Practice on some standards
at different edges to get a feel for it.

Some more unsolicited advice:
Avoid bias - let the data tell their story. Explain how the data were 
collected. Describe in detail how the spectra
were processed. Discuss the results - what you observe, how it compares 
to literature. If you see differences,
report differences, and consider what might be plausible explanations.




On 2020-12-07 8:55 a.m., Deka, N. wrote:
>
> Dear all
>
> I am a student trying to learn Athena for XAS data processing. 
> Attached is a data file which contains L3 and L2 edge of Ru. What 
> should be the right parameters for background removal and 
> normalization of this data set? In the videos from Bruce Ravel, I 
> could not find any example for such a data set containing two edges.
>
> Thank you
>
> Best Regards,
>
> Nipon Deka
>
> Leiden University
>
>
> _______________________________________________
> Ifeffit mailing list
> Ifeffit at millenia.cars.aps.anl.gov
> http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
> Unsubscribe: http://millenia.cars.aps.anl.gov/mailman/options/ifeffit

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://millenia.cars.aps.anl.gov/pipermail/ifeffit/attachments/20201207/06877aea/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ghbkhepbogidpiad.png
Type: image/png
Size: 22364 bytes
Desc: not available
URL: <http://millenia.cars.aps.anl.gov/pipermail/ifeffit/attachments/20201207/06877aea/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cdcnmnlceoeipgdg.png
Type: image/png
Size: 28667 bytes
Desc: not available
URL: <http://millenia.cars.aps.anl.gov/pipermail/ifeffit/attachments/20201207/06877aea/attachment-0003.png>


More information about the Ifeffit mailing list