[Ifeffit] Question about deglitching EXAFS data

Bruce Ravel bravel at bnl.gov
Tue Jul 7 16:11:14 CDT 2015


Hi Francesco,

This showed up while I was on vacation and it seems not to have gotten
an answer.  I'll give it a shot.

I think the most correct answer to your question would rely upon
Fourier analysis.  If you remove from the data a number of points
whose width is small compared to the wavelengths of the portion of the
data that you are attempting to analyze, then you are probably safe.

When the software moves the data from mu(E) to chi(k) after background
removal, it interpolates the data onto the "standard" k grid, which is
even spaced with points separated by 0.05 invAng.  This interpolation
grid is very fine compared to the frquencies of the data you will ever
analyze.

When you deglitch, the points are removed, then the background is
removed, then the interpolation happens.  If the width of the region
removed by deglitching is wide enough, then the interpolation will
necessarily change the Fourier spectrum.  If it is narrow enough, then
the interpolation will be a reliable approximation of the true data.

In your specific case, 1 and 2 point glitches are almost certainly
safe to remove.  A 10 point glitch is probably not safe to remove.

I am not going to commit to rule of thumb, however... :P

B

On 06/29/2015 03:41 PM, Francesco Marafatto wrote:
> Hello All,
>
> I have enjoyed this digest in the past and found it very helpful! I hope it may provide an answer to my question.
>
> I am collecting Mn EXAFS data out to k of 15 Å-1, and the data look fine. However, the monochromator (Si 111) has 5 strong 2-point glitches and 3 minor 1-point glitches in one of the regions that I am most interested in (k between 5 and 9 Å-1).
>
> My question relates to how many data points I can remove before influencing the data. In my previous experience I have removed 2 or 3 points but never more. I have read through the deglitching section of  “XAFS for Everyone”, and think that I could oversample the EXAFS (with k steps of 0.03 Å-1 vs 0.05 Å-1), but most of the data have already been collected.
> Is there a rule of thumb about deglitching? In other words, is there a limit to the number of bad points I can eliminate for an EXAFS scan to k of 15 Å-1?
>
> Thanks and looking forward to hearing your thoughts,
>
> Francesco
>
> Francesco Femi Marafatto
> francesco.marafatto at unil.ch
> FNS PhD student,
> Environmental Geochemistry Laboratory
> IDyST - FGSE
> Bureau 3233 Gèopolis
> Université de Lausanne
> CH-1015 Lausanne
> T. +41 (0)21 692 4451
>
>
> _______________________________________________
> Ifeffit mailing list
> Ifeffit at millenia.cars.aps.anl.gov
> http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
>


-- 
  Bruce Ravel  ------------------------------------ bravel at bnl.gov

  National Institute of Standards and Technology
  Synchrotron Science Group at NSLS-II
  Building 535A
  Upton NY, 11973

  Homepage:    http://bruceravel.github.io/home/
  Software:    https://github.com/bruceravel
  Demeter:     http://bruceravel.github.io/demeter/



More information about the Ifeffit mailing list