What can be done in this case, is not to use either flourescence or
transmission data, - because both are bad by the reasons Bruce explained - but
use corrections, if you cannot change your experimental geometry (different
concentration of the sample or different thickness or different reactor cell).
What we do in these cases, if we must keep original thick sample for
fluorescence measurements, we make the same sample prepared for transmission on
a tape or as a pellet, that gives proper thickness to be free of any leakage or
thickness effects that reduce oscillation intensity or introduce noise due to
poor statistics. Then we compare EXAFS oscillations plotted together for the
two samples: one measured in fluorescence in the cell you want to use in your
in situ experiment, and the other meausred in tranmission in ideal conditions.
The former will have reduced intensity. You can manually find scaling factor
that is needed to match two intensities. Based on many early papers (e.g., Kim,
Stern and Heald, Physical Review B, Thickness effect in EXAFS data or something
like that), this scaling factor is approximately constant. Once you found it,
then you can take all your in situ data in flourescence and later on, during
data processing and analysis, correct all the EXAFS data by scaling them up
with the same factor.
Don't change the sample geometry midway, and do not apply similar correction to
recover XANES. Other strategies should be used for XANES corrections.
Anatoly
From: ifeffit-boun...@millenia.cars.aps.anl.gov on behalf of Andrew Campos
Sent: Wed 5/5/2010 11:24 AM
To: ifeffit@millenia.cars.aps.anl.gov
Subject: [Ifeffit] Differences between fluorescence and transmission of thesame
sample
Dr. Ravel,
Thanks so much for the link and the advice! I appreciate it greatly. I
will advise my lab mates as such and may have to only use the
fluorescence data if that is indeed the case.
I also included the file where the lower temperature is included and
you might come to the same conclusion. The samples that I ran were
pre-sieved, and the ones included in the .prj file aren't so that
should be pursued prior to running the experiment. If they crush the
particle and sieve the sample, I think that we can be more certain
that this is not the case. This was very helpful!
Andrew
<>___
Ifeffit mailing list
Ifeffit@millenia.cars.aps.anl.gov
http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit