Dear ifeffit users, This posting is with quite basic questions on XANES data processing. I am a somewhat new XANES experimentalist, in a group that does not typically use XANES, so I am lacking some basic knowledge. For normalization of XANES, my impression is that people generally normalize the spectra from ca. 50 eV up to 100 eV or more after the edge onset. Is there a standard acceptable normalization range for K edges? I have noticed that when I change the start of the normalization window by even a few eV, this can make substantial differences in the XANES when comparing two similar spectra (for example, change in peak and pre-edge peak intensities between XANES of one sample at two temperatures). I keep second-guessing my choice of normalization range. Second, I am also looking at the effect of the various ways to set E0. When I start with ifeffit's default (how is this determined?) and then proceed to set E0 to the zero-crossing of 2nd derivative, I'll get value X. When I change it to a set fraction of the edge step (is this legitimate?) I typically get a lower value, and then when I proceed again to set it to the 2nd derivative zero-crossing, I get a value different from the first time it was set to the zero- crossing. For example, the E0 at an Fe K edge: ifeffit default = 7128.9 eV 2nd derivative zero = 7128.8 set fraction = 7125.1 2nd derivative zero = 7125.5 The value given by the 2nd derivative zero-crossing is dependent on the previous E0 setting. I am having a hard time to move on to the actual interpretation of the data while I am still uncertain about such details. Thanks for your patience with some simple questions. Best, Ashley
Hi Ashley: I think you are struggling with one of the most difficult aspects of XAS analysis. There is no single, automatic way of setting these parameters although Athena and Ifeffit do a good job of initial guesses. It is rather trite to say that it is a question of experience but that is exactly what it is. The important thing is consistency throughout a data set and this is much more difficult if there are large changes in the absorption as a function of sample or measurement conditions. So, it is difficult to answer you in the abstract. It would be good for you to send a project file to the list and then you would be able to get some direct feedback on _your_ data. I am sure that you will get several comments from the experts on this list. Cheers, Carlo On Sat, 14 Jul 2007, Ashley Harvey wrote:
Dear ifeffit users,
This posting is with quite basic questions on XANES data processing. I am a somewhat new XANES experimentalist, in a group that does not typically use XANES, so I am lacking some basic knowledge.
For normalization of XANES, my impression is that people generally normalize the spectra from ca. 50 eV up to 100 eV or more after the edge onset. Is there a standard acceptable normalization range for K edges? I have noticed that when I change the start of the normalization window by even a few eV, this can make substantial differences in the XANES when comparing two similar spectra (for example, change in peak and pre-edge peak intensities between XANES of one sample at two temperatures). I keep second-guessing my choice of normalization range.
Second, I am also looking at the effect of the various ways to set E0. When I start with ifeffit's default (how is this determined?) and then proceed to set E0 to the zero-crossing of 2nd derivative, I'll get value X. When I change it to a set fraction of the edge step (is this legitimate?) I typically get a lower value, and then when I proceed again to set it to the 2nd derivative zero-crossing, I get a value different from the first time it was set to the zero- crossing. For example, the E0 at an Fe K edge: ifeffit default = 7128.9 eV 2nd derivative zero = 7128.8 set fraction = 7125.1 2nd derivative zero = 7125.5 The value given by the 2nd derivative zero-crossing is dependent on the previous E0 setting.
I am having a hard time to move on to the actual interpretation of the data while I am still uncertain about such details. Thanks for your patience with some simple questions.
Best, Ashley
_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
-- Carlo U. Segre -- Professor of Physics Associate Dean for Special Projects, Graduate College Illinois Institute of Technology Voice: 312.567.3498 Fax: 312.567.3494 segre@iit.edu http://www.iit.edu/~segre segre@debian.org
participants (2)
-
Ashley Harvey -
Carlo Segre