Hello, Everyone - This is my first time trying to work through EXAFS analysis of experimental data, and I've got what (I think) should be a fairly straightforward question. I'm trying to minimize my delta E0 by aligning the EXAFS data with my best fit (i.e. - align the k-grid of the data with the k-grid of the theory to obtain a small delta E0). However, I have been unsuccessful in "reading the theoretical chi(k) spectrum into Athenahttp://www.mail-archive.com/ifeffit%40millenia.cars.aps.anl.gov/msg02494.htm..." to accomplish this. I assume the crux of the problem is the format in which I am exporting my theoretical chi(k) spectrum from Artemis. The only way I have been able to export the data with Artemis so Athena can import it is to click the "Save next plot to a file" button, which then allows me to save the data as a text file. I know I need chi(k) data that is not k-weightedhttp://www.mail-archive.com/ifeffit%40millenia.cars.aps.anl.gov/msg01623.htm..., so I plot the best fit with k-weight of 0 in k space with deltaE0 equal to 0. Considering the problems I've had, I'm assuming somewhere in this process is where I'm going wrong. When I import this data, I have four columns: k, experimental data, fit, window. Since k is the wavenumber of the photoelectron, I choose that as "Energy." Then I select column 3, the best fit, as "Numerator," and change the data type to chi(k). I then import the data. I've attached two pngs to this e-mail which, assuming they come through, should display the Athena data import window and the corresponding plot of chi(k). When I plot this data in k-space or R-space using Artemis, the resulting plots look appropriate. When I try to align the experimental EXAFS data in Athena, I know it aligns the first derivative of E for the selected data to the first derivative of E for the standard. I want my best fit to be the standard to which the data is aligned. However, I cannot plot chi(k) data in E, which causes me to believe I shouldn't be able to align data to the first derivative of E. True enough, when I try to auto align, I get impossible values for E0 adjustment. Clearly I'm doing something wrong - possibly even several somethings wrong. Any guidance would be greatly appreciated. Thanks, - Carter
Dear Carter, the question is: Why? Why would you want to align your experimental spectra to a theoretical standard? For a number of reasons (the somewhat arbitrary definition of E0, the calibration of your monochromator, the way FEFF calculates the theoretical spectra...) delta E0 is almost never 0. Some(!) eV are normal and should just be repoerted in your publication. If delta E0 is larger than ~10 eV you should start to find a possible explanation for this behaviour. If you are confident about the calibration of your DCM it is usually indicating a problem with your fitting model. If you align your exp. spectra to the (so far...) best fit you are just obscuring these problems... Best regards Edmund On 04.06.2014 22:18, Carter Abney wrote:
Hello, Everyone --
This is my first time trying to work through EXAFS analysis of experimental data, and I've got what (I think) should be a fairly straightforward question. I'm trying to minimize my delta E0 by aligning the EXAFS data with my best fit (i.e. - align the k-grid of the data with the k-grid of the theory to obtain a small delta E0). However, I have been unsuccessful in "reading the theoretical chi(k) spectrum into Athena http://www.mail-archive.com/ifeffit%40millenia.cars.aps.anl.gov/msg02494.htm..." to accomplish this.
I assume the crux of the problem is the format in which I am exporting my theoretical chi(k) spectrum from Artemis. The only way I have been able to export the data with Artemis so Athena can import it is to click the "Save next plot to a file" button, which then allows me to save the data as a text file. I know I need chi(k) data that is not k-weighted http://www.mail-archive.com/ifeffit%40millenia.cars.aps.anl.gov/msg01623.htm..., so I plot the best fit with k-weight of 0 in k space with deltaE0 equal to 0. Considering the problems I've had, I'm assuming somewhere in this process is where I'm going wrong.
When I import this data, I have four columns: k, experimental data, fit, window. Since k is the wavenumber of the photoelectron, I choose that as "Energy." Then I select column 3, the best fit, as "Numerator," and change the data type to chi(k). I then import the data. I've attached two pngs to this e-mail which, assuming they come through, should display the Athena data import window and the corresponding plot of chi(k). When I plot this data in k-space or R-space using Artemis, the resulting plots look appropriate.
When I try to align the experimental EXAFS data in Athena, I know it aligns the first derivative of E for the selected data to the first derivative of E for the standard. I want my best fit to be the standard to which the data is aligned. However, I cannot plot chi(k) data in E, which causes me to believe I shouldn't be able to align data to the first derivative of E. True enough, when I try to auto align, I get impossible values for E0 adjustment.
Clearly I'm doing something wrong -- possibly even several somethings wrong. Any guidance would be greatly appreciated.
Thanks,
- Carter
_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Hi Carter, My interpretation of (for example) Shelly Kelly's tutorial on varying experimental E0 to fit the theory was to do this with the background subtraction E0 in Athena. E.g. http://www.xafs.org/Tutorials?action=AttachFile&do=get&target=Basics_of_XAFS_analysis_2009.pdf I believe that the idea is not to move the whole spectrum in energy, but to change E0 in the background removal routine. You vary the E0 in the background removal box (as well as Rbkg, probably) to make the experimental EXAFS spectrum more closely match the theoretical first shell. It sounds like you are trying move the whole spectrum with respect to energy calibration. However, someone who knows better than I may say that these two things are equivalent... But I would think if you were trying to fit some Cu compound [I gather people that do EXAFS have some sort of thing for Cu ;-) ] you'd still want to leave the whole spectrum calibrated to Cu foil, and move the background subtraction E0 around to best fit the theoretical first shell. I suspect there are also cases where you might not be inclined to think that deltaE0 should be close to 0. For example, if you use a quasi-reference for energy calibration. Arsenic and uranium are common cases in my research, where oxides or higher valence compounds serve as references. Best, Drew Latta From: ifeffit-bounces@millenia.cars.aps.anl.gov [mailto:ifeffit-bounces@millenia.cars.aps.anl.gov] On Behalf Of Carter Abney Sent: Wednesday, June 04, 2014 3:19 PM To: ifeffit@millenia.cars.aps.anl.gov Subject: [Ifeffit] Aligning EXAFS E0 to Theory Hello, Everyone - This is my first time trying to work through EXAFS analysis of experimental data, and I've got what (I think) should be a fairly straightforward question. I'm trying to minimize my delta E0 by aligning the EXAFS data with my best fit (i.e. - align the k-grid of the data with the k-grid of the theory to obtain a small delta E0). However, I have been unsuccessful in "reading the theoretical chi(k) spectrum into Athenahttp://www.mail-archive.com/ifeffit%40millenia.cars.aps.anl.gov/msg02494.htm..." to accomplish this. I assume the crux of the problem is the format in which I am exporting my theoretical chi(k) spectrum from Artemis. The only way I have been able to export the data with Artemis so Athena can import it is to click the "Save next plot to a file" button, which then allows me to save the data as a text file. I know I need chi(k) data that is not k-weightedhttp://www.mail-archive.com/ifeffit%40millenia.cars.aps.anl.gov/msg01623.htm..., so I plot the best fit with k-weight of 0 in k space with deltaE0 equal to 0. Considering the problems I've had, I'm assuming somewhere in this process is where I'm going wrong. When I import this data, I have four columns: k, experimental data, fit, window. Since k is the wavenumber of the photoelectron, I choose that as "Energy." Then I select column 3, the best fit, as "Numerator," and change the data type to chi(k). I then import the data. I've attached two pngs to this e-mail which, assuming they come through, should display the Athena data import window and the corresponding plot of chi(k). When I plot this data in k-space or R-space using Artemis, the resulting plots look appropriate. When I try to align the experimental EXAFS data in Athena, I know it aligns the first derivative of E for the selected data to the first derivative of E for the standard. I want my best fit to be the standard to which the data is aligned. However, I cannot plot chi(k) data in E, which causes me to believe I shouldn't be able to align data to the first derivative of E. True enough, when I try to auto align, I get impossible values for E0 adjustment. Clearly I'm doing something wrong - possibly even several somethings wrong. Any guidance would be greatly appreciated. Thanks, - Carter
Yes, Shelly's tutorial was to obtain a better starting point for E0 so that it is minimized in the modeling. Since E0 and delr are highly correlated, this will serve to improve the value obtained for the path length. Carlo On Thu, 5 Jun 2014, Latta, Drew E wrote:
Hi Carter,
My interpretation of (for example) Shelly Kelly's tutorial on varying experimental E0 to fit the theory was to do this with the background subtraction E0 in Athena. E.g. http://www.xafs.org/Tutorials?action=AttachFile&do=get&target=Basics_of_XAFS_analysis_2009.pdf
I believe that the idea is not to move the whole spectrum in energy, but to change E0 in the background removal routine. You vary the E0 in the background removal box (as well as Rbkg, probably) to make the experimental EXAFS spectrum more closely match the theoretical first shell. It sounds like you are trying move the whole spectrum with respect to energy calibration.
However, someone who knows better than I may say that these two things are equivalent... But I would think if you were trying to fit some Cu compound [I gather people that do EXAFS have some sort of thing for Cu ;-) ] you'd still want to leave the whole spectrum calibrated to Cu foil, and move the background subtraction E0 around to best fit the theoretical first shell.
I suspect there are also cases where you might not be inclined to think that deltaE0 should be close to 0. For example, if you use a quasi-reference for energy calibration. Arsenic and uranium are common cases in my research, where oxides or higher valence compounds serve as references.
Best,
Drew Latta
From: ifeffit-bounces@millenia.cars.aps.anl.gov [mailto:ifeffit-bounces@millenia.cars.aps.anl.gov] On Behalf Of Carter Abney Sent: Wednesday, June 04, 2014 3:19 PM To: ifeffit@millenia.cars.aps.anl.gov Subject: [Ifeffit] Aligning EXAFS E0 to Theory
Hello, Everyone -
This is my first time trying to work through EXAFS analysis of experimental data, and I've got what (I think) should be a fairly straightforward question. I'm trying to minimize my delta E0 by aligning the EXAFS data with my best fit (i.e. - align the k-grid of the data with the k-grid of the theory to obtain a small delta E0). However, I have been unsuccessful in "reading the theoretical chi(k) spectrum into Athenahttp://www.mail-archive.com/ifeffit%40millenia.cars.aps.anl.gov/msg02494.htm..." to accomplish this.
I assume the crux of the problem is the format in which I am exporting my theoretical chi(k) spectrum from Artemis. The only way I have been able to export the data with Artemis so Athena can import it is to click the "Save next plot to a file" button, which then allows me to save the data as a text file. I know I need chi(k) data that is not k-weightedhttp://www.mail-archive.com/ifeffit%40millenia.cars.aps.anl.gov/msg01623.htm..., so I plot the best fit with k-weight of 0 in k space with deltaE0 equal to 0. Considering the problems I've had, I'm assuming somewhere in this process is where I'm going wrong.
When I import this data, I have four columns: k, experimental data, fit, window. Since k is the wavenumber of the photoelectron, I choose that as "Energy." Then I select column 3, the best fit, as "Numerator," and change the data type to chi(k). I then import the data. I've attached two pngs to this e-mail which, assuming they come through, should display the Athena data import window and the corresponding plot of chi(k). When I plot this data in k-space or R-space using Artemis, the resulting plots look appropriate.
When I try to align the experimental EXAFS data in Athena, I know it aligns the first derivative of E for the selected data to the first derivative of E for the standard. I want my best fit to be the standard to which the data is aligned. However, I cannot plot chi(k) data in E, which causes me to believe I shouldn't be able to align data to the first derivative of E. True enough, when I try to auto align, I get impossible values for E0 adjustment.
Clearly I'm doing something wrong - possibly even several somethings wrong. Any guidance would be greatly appreciated.
Thanks,
- Carter
-- Carlo U. Segre -- Duchossois Leadership Professor of Physics Director, Center for Synchrotron Radiation Research and Instrumentation Illinois Institute of Technology Voice: 312.567.3498 Fax: 312.567.3494 segre@iit.edu http://phys.iit.edu/~segre segre@debian.org
participants (4)
-
Carlo Segre -
Carter Abney -
Edmund Welter -
Latta, Drew E