Thanks Bruce. Till the time i look into the suggested works, here is the missing informations. Question: how to check the appropriateness of the self absorption correction? I don't have corresponding transmission data. The question came to my mind when i compared the uncorrected (of courese corrected for dead time) data and the corrected (with SiO2Eu0.002 formula; the formula corresponds to the mole (roughly atom)percentage. The edge jump enhanced after correction to ~1 from ~0.04. Material: Eu(III) sorbed on silica. Sumit -- Sumit Kumar Scientific Officer(E) Radioanalytical Chemistry Division Bhabha Atomic Research Centre Mumbai, India-400 085. Phone: +91-22-2559 4093/0644 (O).
On Wednesday, January 09, 2013 10:28:17 AM sumitk@barc.gov.in wrote:
Thanks Bruce. Till the time i look into the suggested works, here is the missing informations.
That's kind of backwards. You should do your research *before* asking a question. One tends to ask better questions that way!
Question: how to check the appropriateness of the self absorption correction? I don't have corresponding transmission data. The question came to my mind when i compared the uncorrected (of courese corrected for dead time) data and the corrected (with SiO2Eu0.002 formula; the formula corresponds to the mole (roughly atom)percentage. The edge jump enhanced after correction to ~1 from ~0.04. Material: Eu(III) sorbed on silica.
Again, you are not providing enough information so that your question can be answered well. Where did that stoichiometry come from? It would be useful to know how thick the Eu film is on the silica substrate and then compare that to the information depth (aorund 50 microns, if I understand correctly what the sample is) of Eu around its L3 edge. Judging from amount the data changes when applying the correction, it would seem as if your film is quite thin. The XANES changes very little and the EXAFS not at all when the correction is applied. Are you fretting about the change in edge jump? If so, why? Edge jump is, in most cases, a pretty arbitrary number for a fluorescence measurement [1]. If you change the gain on I0 by one order of magnitude, the edge jump changes by a factor of 10. If you change the gasses in IO or change the gains on your energy discriminating detectors, the edge jump again changes. This is in contrast to a transmission experiment where, by virtue of the natural log, the edge jump tells you something about the physical thickness of the absorbing part of the sample. By applying the self-absorption correction, you did some kind of linear transform to the data, changing the absolute scale of the numbers in some way. Comparing the normalized data (remember, the normalization algorithm is used explicitly to remove all of these arbitrary scling factors from the data) shows that the correction did very little to your data. Perhaps I don't quite understand what your concern is with these data, but I doubt that self-absorption should be among them. Cheers, B [1] Here's a counter example http://dx.doi.org/10.1021/ac060476m, but note that great care was taken in the work to make the ensemble of measurements in a consistent fashion. -- Bruce Ravel ------------------------------------ bravel@bnl.gov National Institute of Standards and Technology Synchrotron Methods Group at NSLS --- Beamlines U7A, X24A, X23A2 Building 535A Upton NY, 11973 Homepage: http://xafs.org/BruceRavel Software: https://github.com/bruceravel
Hello: 1) Is there any reference which states what the error bars for Eo, Delr , SS and CN should be so that the fit with its parameters is regarded to be of publishable quality? 2) A lot of times we find So2 from the respective metal foil and fix it when we fit the sample data for the same element. Is there any reference for this one? I know about the book by Shelly, Hesterberg and Ravel. Would appreciate if I get some more references. Thanks. Soma
Hi Soma,
On Tue, Jan 22, 2013 at 11:11 AM, Soma Chattopadhyay
Hello:
1) Is there any reference which states what the error bars for Eo, Delr , SS and CN should be so that the fit with its parameters is regarded to be of publishable quality?
I don't think one can set a general threshold for what is and is not publishable according to the size of error bars. Data and results of analysis need to be of sufficient quality and robustness for the conclusions drawn. One can certainly have small error bars and be wrong!
2) A lot of times we find So2 from the respective metal foil and fix it when we fit the sample data for the same element. Is there any reference for this one?
I know about the book by Shelly, Hesterberg and Ravel. Would appreciate if I get some more references.
I'm not aware of a truly thorough study of this convention. Several of the early UWXAFS papers mention using this approach. Probably the clearest statement is from E A Stern, et al "The UWXAFS Analysis Package - Philosophy and Details", Physica B208 pp 117-120 (1995) [DOI: 10.1016/0921-4526(94)00826-H] which states (full paragraph): The set of parameters that most frequently have a strong enough correlation to give difficulty in obtaining their values separately are S02, the passive electron overlap constant[13], sigma^2, the disorder about the optimal distance, and N, the number of atoms in a particular coordination shell. The most reliable way to separate out sigma^2 from the other two parameters is to measure the temperature dependence of sigma^2 and vary the product S02*N so as to obtain a fit to an Einstein oscillator, particularly the zero point value. Since S02 is dependent essentially on only the center atom, it can be found in known structures and N determined from the S02*N value. Here Reference 13 is to Stern and Heald's chapter in the Handbook of Synchrotron Radiation Vol 1 (1983). Cheers, --Matt
Hi Soma!
Google "S02 transferability site:http://millenia.cars.aps.anl.gov/" and
you'll find lots of messages on the subject if you haven't already.
There's a post from Bruce on May 4 2010 where he writes:
Recent work from John Rehr's group has elaborated on this,
demonstrating that S02 is, in the strictest sense, neither
transferable nor constant in energy. However, these effects are small
such that chemical transferability remains a useful and defensible
practice in many EXAFS analyses.
http://millenia.cars.aps.anl.gov/pipermail/ifeffit/2010-May/009378.html
I don't know what specific work he's referring to but that might be a good
direction.
-Jason
On Tue, Jan 22, 2013 at 1:15 PM, Matt Newville
Hi Soma,
On Tue, Jan 22, 2013 at 11:11 AM, Soma Chattopadhyay
wrote: 2) A lot of times we find So2 from the respective metal foil and fix it when we fit the sample data for the same element. Is there any reference for this one?
I know about the book by Shelly, Hesterberg and Ravel. Would appreciate if I get some more references.
I'm not aware of a truly thorough study of this convention. Several of the early UWXAFS papers mention using this approach. Probably the clearest statement is from
E A Stern, et al "The UWXAFS Analysis Package - Philosophy and Details", Physica B208 pp 117-120 (1995) [DOI: 10.1016/0921-4526(94)00826-H]
which states (full paragraph):
The set of parameters that most frequently have a strong enough correlation to give difficulty in obtaining their values separately are S02, the passive electron overlap constant[13], sigma^2, the disorder about the optimal distance, and N, the number of atoms in a particular coordination shell. The most reliable way to separate out sigma^2 from the other two parameters is to measure the temperature dependence of sigma^2 and vary the product S02*N so as to obtain a fit to an Einstein oscillator, particularly the zero point value. Since S02 is dependent essentially on only the center atom, it can be found in known structures and N determined from the S02*N value.
Here Reference 13 is to Stern and Heald's chapter in the Handbook of Synchrotron Radiation Vol 1 (1983).
Cheers,
--Matt _______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Thanks, Matt and Jason. I very much appreciate the prompt replies from both of you. best Soma Jason Gaudet wrote:
Hi Soma!
Google "S02 transferability site:http://millenia.cars.aps.anl.gov/" and you'll find lots of messages on the subject if you haven't already.
There's a post from Bruce on May 4 2010 where he writes:
Recent work from John Rehr's group has elaborated on this, demonstrating that S02 is, in the strictest sense, neither transferable nor constant in energy. However, these effects are small such that chemical transferability remains a useful and defensible practice in many EXAFS analyses.
http://millenia.cars.aps.anl.gov/pipermail/ifeffit/2010-May/009378.html
I don't know what specific work he's referring to but that might be a good direction.
-Jason
On Tue, Jan 22, 2013 at 1:15 PM, Matt Newville
mailto:newville@cars.uchicago.edu> wrote: Hi Soma,
On Tue, Jan 22, 2013 at 11:11 AM, Soma Chattopadhyay
mailto:soma@agni.phys.iit.edu> wrote: > > 2) A lot of times we find So2 from the respective metal foil and fix it > when we fit the sample data for the same element. Is there any reference for > this one? > > I know about the book by Shelly, Hesterberg and Ravel. > Would appreciate if I get some more references.
I'm not aware of a truly thorough study of this convention. Several of the early UWXAFS papers mention using this approach. Probably the clearest statement is from
E A Stern, et al "The UWXAFS Analysis Package - Philosophy and Details", Physica B208 pp 117-120 (1995) [DOI: 10.1016/0921-4526(94)00826-H]
which states (full paragraph):
The set of parameters that most frequently have a strong enough correlation to give difficulty in obtaining their values separately are S02, the passive electron overlap constant[13], sigma^2, the disorder about the optimal distance, and N, the number of atoms in a particular coordination shell. The most reliable way to separate out sigma^2 from the other two parameters is to measure the temperature dependence of sigma^2 and vary the product S02*N so as to obtain a fit to an Einstein oscillator, particularly the zero point value. Since S02 is dependent essentially on only the center atom, it can be found in known structures and N determined from the S02*N value.
Here Reference 13 is to Stern and Heald's chapter in the Handbook of Synchrotron Radiation Vol 1 (1983).
Cheers,
--Matt _______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov mailto:Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
------------------------------------------------------------------------
_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
participants (5)
-
Bruce Ravel
-
Jason Gaudet
-
Matt Newville
-
Soma Chattopadhyay
-
sumitk@barc.gov.in