Clarification needed on the structured 2nn peak
Greetings to all I am a beginner to EXAFS. Recently I had carried out Cd K-edge EXAFS measurements on Cd1-xCaxO solid solution. The end members CdO and CaO belong to same crystal structure Fm-3m. I had analysed the spectrum with Atthena and fitted with Arthemis. I am attaching the Athena project file (***.prj) and that of Arthemis project file (***.apr) for your reference. Athena: As per the crystal structure of CdO, the first nearest neighbour is at 2.345 A corresponding to Cd-O1 inter atomic distance and 2nn distance is 3.3.16 A corresponding to Cd-Cd1 inter atomic distance. Though I am able get the first peak in chi(R) plot at expected position, I am having some problems for the second peak. The second peak has features; main one and a shoulder. The main peak itself appears at lower R compared to expected one. Intensity of the shoulder increases with Ca substitution at the expense of main peak itself. Can any one comment on this? My predicament: If I associate, based on the structure of second peak, bimodal distribution for the Cd-Cd1 interatomic distance, the difference of 0.4 A is 11%. This value gives me an uneasy feeling. Is it possible to accommodate this difference within the setting of F m-3m? OR It is due the multi scattering is the problem? Arthemis: For CdO: The k^3 weighted chi fitting, I agree is far from satisfaction. Std of delR is a bothering point. I had included 20 paths. Even for the CdO with F m-3m, the calculated second peak has a shoulder at lower R. I could see a small contribution from path 4. But I am not sure if that is all. How to float the individual GDS parameter for each and every path? What can be done to improve the fitting? Thanking you all in advance Varadarajan Sridharan ---------------------------------- Condensed matter Physics Division Materials Science division Indira Gandhi Centre for Atomic Research Kalpakkam 603 102 India
Dear Varadarian,
I didn't look at your attachments, but have you considered that there
might be some Cd-Ca in the second shell? I didn't see what value of
x you had in Cd(1-x)CaxO, but if the oxides form a solid solution one
might naively expect the second shell to have (1-x) Cd and x Ca
coordination. Ca and Cd scatter photoelectrons differently enough
to be distinguishable in the EXAFS. It might be that there is extra
disorder too, say if the typical Cd-Ca distance was different from the
typical Cd-Cd distance.
Cheers,
--Matt
On Tue, Jul 6, 2010 at 7:29 AM, Varadarajan Sridharan
Greetings to all
I am a beginner to EXAFS.
Recently I had carried out Cd K-edge EXAFS measurements on Cd1-xCaxO solid solution. The end members CdO and CaO belong to same crystal structure Fm-3m.
I had analysed the spectrum with Atthena and fitted with Arthemis. I am attaching the Athena project file (***.prj) and that of Arthemis project file (***.apr) for your reference.
Athena:
As per the crystal structure of CdO, the first nearest neighbour is at 2.345 A corresponding to Cd-O1 inter atomic distance and 2nn distance is 3.3.16 A corresponding to Cd-Cd1 inter atomic distance. Though I am able get the first peak in chi(R) plot at expected position, I am having some problems for the second peak. The second peak has features; main one and a shoulder. The main peak itself appears at lower R compared to expected one. Intensity of the shoulder increases with Ca substitution at the expense of main peak itself.
Can any one comment on this?
My predicament: If I associate, based on the structure of second peak, bimodal distribution for the Cd-Cd1 interatomic distance, the difference of 0.4 A is 11%. This value gives me an uneasy feeling. Is it possible to accommodate this difference within the setting of F m-3m?
OR
It is due the multi scattering is the problem?
Arthemis:
For CdO: The k^3 weighted chi fitting, I agree is far from satisfaction. Std of delR is a bothering point. I had included 20 paths. Even for the CdO with F m-3m, the calculated second peak has a shoulder at lower R. I could see a small contribution from path 4. But I am not sure if that is all.
How to float the individual GDS parameter for each and every path? What can be done to improve the fitting?
Thanking you all in advance
Varadarajan Sridharan
----------------------------------
Condensed matter Physics Division
Materials Science division
Indira Gandhi Centre for Atomic Research
Kalpakkam 603 102
India
_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Dear Varadarajan, Infact if you look at the magnitude of FT's of all compounds together in Athena, you can see the evolution of "double peak" with increasing Ca in CdO. Therefore it should be most likely due to reasons pointed out by Matt in his mail. cheers Kaustubh ----- Original Message ----
From: Matt Newville
To: XAFS Analysis using Ifeffit Sent: Tue, 6 July, 2010 6:22:16 PM Subject: Re: [Ifeffit] Clarification needed on the structured 2nn peak Dear Varadarian,
I didn't look at your attachments, but have you considered that there might be some Cd-Ca in the second shell? I didn't see what value of x you had in Cd(1-x)CaxO, but if the oxides form a solid solution one might naively expect the second shell to have (1-x) Cd and x Ca coordination. Ca and Cd scatter photoelectrons differently enough to be distinguishable in the EXAFS. It might be that there is extra disorder too, say if the typical Cd-Ca distance was different from the typical Cd-Cd distance.
Cheers,
--Matt
On Tue, Jul 6, 2010 at 7:29 AM, Varadarajan Sridharan
wrote: Greetings to all
I am a beginner to EXAFS.
Recently I had carried out Cd K-edge EXAFS measurements on Cd1-xCaxO solid solution. The end members CdO and CaO belong to same crystal structure Fm-3m.
I had analysed the spectrum with Atthena and fitted with Arthemis. I am attaching the Athena project file (***.prj) and that of Arthemis project file (***.apr) for your reference.
Athena:
As per the crystal structure of CdO, the first nearest neighbour is at 2.345 A corresponding to Cd-O1 inter atomic distance and 2nn distance is 3.3.16 A corresponding to Cd-Cd1 inter atomic distance. Though I am able get the first peak in chi(R) plot at expected position, I am having some problems for the second peak. The second peak has features; main one and a shoulder. The main peak itself appears at lower R compared to expected one. Intensity of the shoulder increases with Ca substitution at the expense of main peak itself.
Can any one comment on this?
My predicament: If I associate, based on the structure of second peak, bimodal distribution for the Cd-Cd1 interatomic distance, the difference of 0.4 A is 11%. This value gives me an uneasy feeling. Is it possible to accommodate this difference within the setting of F m-3m?
OR
It is due the multi scattering is the problem?
Arthemis:
For CdO: The k^3 weighted chi fitting, I agree is far from satisfaction. Std of delR is a bothering point. I had included 20 paths. Even for the CdO with F m-3m, the calculated second peak has a shoulder at lower R. I could see a small contribution from path 4. But I am not sure if that is all.
How to float the individual GDS parameter for each and every path? What can be done to improve the fitting?
Thanking you all in advance
Varadarajan Sridharan
----------------------------------
Condensed matter Physics Division
Materials Science division
Indira Gandhi Centre for Atomic Research
Kalpakkam 603 102
India
_______________________________________________ Ifeffit mailing list 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
Hi Varadarajan, I am by no means an expert but here are a few things I noticed & changes I made. In regards to your fitting, from what I can see, you haven't defined separate delrs & ss for each path that you want to use. Assuming the artemis file was for Cd80Ca20O, I've made the following changes: tweaked your feff.inp file to add 2 Ca in the second shell (as suggested by Matt) Introduced separate delr and ss for Cd-Cd and Cd-Ca paths k-range of 3.2-16; dk = 0.6 R-range of 1.4-3.6; dr = 0.7 Changed fitting space to R Visually the fit seems a lot better with these changes though you would be in a better position to judge if the actual numbers seem realistic. You could probably make the fit a bit better by introducing some constrains (ss_Cd = ss_Ca maybe) and floating the coordination number for the NNN Cd and Ca rather than leaving it fixed at 10 & 2. Hope that helps. Zohair From: ifeffit-bounces@millenia.cars.aps.anl.gov [mailto:ifeffit-bounces@millenia.cars.aps.anl.gov] On Behalf Of Varadarajan Sridharan Sent: Tuesday, 6 July 2010 10:29 PM To: ifeffit@millenia.cars.aps.anl.gov Subject: [Ifeffit] Clarification needed on the structured 2nn peak Greetings to all I am a beginner to EXAFS. Recently I had carried out Cd K-edge EXAFS measurements on Cd1-xCaxO solid solution. The end members CdO and CaO belong to same crystal structure Fm-3m. I had analysed the spectrum with Atthena and fitted with Arthemis. I am attaching the Athena project file (***.prj) and that of Arthemis project file (***.apr) for your reference. Athena: As per the crystal structure of CdO, the first nearest neighbour is at 2.345 A corresponding to Cd-O1 inter atomic distance and 2nn distance is 3.3.16 A corresponding to Cd-Cd1 inter atomic distance. Though I am able get the first peak in chi(R) plot at expected position, I am having some problems for the second peak. The second peak has features; main one and a shoulder. The main peak itself appears at lower R compared to expected one. Intensity of the shoulder increases with Ca substitution at the expense of main peak itself. Can any one comment on this? My predicament: If I associate, based on the structure of second peak, bimodal distribution for the Cd-Cd1 interatomic distance, the difference of 0.4 A is 11%. This value gives me an uneasy feeling. Is it possible to accommodate this difference within the setting of F m-3m? OR It is due the multi scattering is the problem? Arthemis: For CdO: The k^3 weighted chi fitting, I agree is far from satisfaction. Std of delR is a bothering point. I had included 20 paths. Even for the CdO with F m-3m, the calculated second peak has a shoulder at lower R. I could see a small contribution from path 4. But I am not sure if that is all. How to float the individual GDS parameter for each and every path? What can be done to improve the fitting? Thanking you all in advance Varadarajan Sridharan ---------------------------------- Condensed matter Physics Division Materials Science division Indira Gandhi Centre for Atomic Research Kalpakkam 603 102 India
participants (4)
-
Kaustubh Priolkar
-
Matt Newville
-
Varadarajan Sridharan
-
Zohair.Hussain@csiro.au