Peaks in MCA spectra

George Sterbinsky

31 Mar 2014 31 Mar '14

10:01 p.m.

Hello, I am writing with a general XAS question. It does not necessarily pertain to Ifeffit, however, I think the topic is something some, maybe most, list members will be knowledgeable about. So it seems like this list is a good place to post this question. On to the question. I have attached a plot of a MCA spectrum collected with a vortex silicon drift detector. The spectrum is actually the average of several spectra, all collected in the post edge region of the Co L-edge. The spectra were averaged to reduce noise. The three peaks result from fluorescence from carbon, oxygen, and cobalt. Low-energy shoulders on the Co and O peaks are also observed. These can be seen as the regions of the spectrum that are not well reproduced by the fit. The main reason I included the fit in the plot is to illustrate the presence of these shoulders, particularly in the oxygen florescence, where the additional intensity is not so obvious. I am writing to see if anyone has any suggestion as to what the origins of these peaks might be. They are not due to additional elements, as they appear at the same incident energies as the main florescence peaks, i.e. the Co shoulder appears at the same incident energies as the main Co peak, and the O shoulder appears at the same incident energies as the main O peak. It is possible that the peaks result form other transitions. Considering Co, the main peak is due to L3/L2-M4 transitions, and the shoulder is in a position that could be consistent with L3/L2-M1 transitions. However, by fitting the peaks with Gaussians, one finds an area for the shoulder that is about 25% of the area of the main peak. This is significantly larger than what one might expect from tabulated transition strengths like those given in Hephaestus. To summarize, does anyone know what these shoulders might result from if not lower energy transitions? If they are low energy transitions, why would the relative transition strengths differ from tabulated values? Thank you, George

Attachments:

attachment.html (text/html — 2.2 KB)
MCAspectrum.png (image/png — 26.5 KB)

Show replies by thread

Matthew Marcus

31 Mar 31 Mar

10:30 p.m.

Hmmm. An interesting brain-teaser. Scaling very roughly from the graph, the mystery peak shows up at about 715eV. This is consistent with Fe. Could there be a background from Fe somewhere in the detection systen? The intensity between the O and Co peaks doesn't really look like a peak, but just like a background tail, perhaps from the strong Co peak. It seems to be a little to the right of where N (from, for instance, a kapton window) would be. I could be wrong, in which case my Fe theory goes away. Now, I thought about an escape peak from Si, but I suspect that this is really small down here and anyway, the Si L-fluorescence energy is 92eV or so, and the extra peak is about 55eV below the main peak. Did you duck below the Fe edge to see if the peak to the L of Co goes away? The peaks in the fit seem to be consistently to the left of the ones in the data. Any idea why? Was the energy scale not a fit parameter? In order to analyze the 'extra' peaks, it may be useful to optimize the fit to the main peaks, then subtract to get the residual, including the 'extras'. What would happen if you added Fe and maybe N to the fit? I admit that I've never done soft X-ray fluorescence, so my ideas come from hard X-ray experience, which may not totally apply. mam On 3/31/2014 3:01 PM, George Sterbinsky wrote:

...

Hello,

I am writing with a general XAS question. It does not necessarily pertain to Ifeffit, however, I think the topic is something some, maybe most, list members will be knowledgeable about. So it seems like this list is a good place to post this question.

On to the question. I have attached a plot of a MCA spectrum collected with a vortex silicon drift detector. The spectrum is actually the average of several spectra, all collected in the post edge region of the Co L-edge. The spectra were averaged to reduce noise. The three peaks result from fluorescence from carbon, oxygen, and cobalt. Low-energy shoulders on the Co and O peaks are also observed. These can be seen as the regions of the spectrum that are not well reproduced by the fit. The main reason I included the fit in the plot is to illustrate the presence of these shoulders, particularly in the oxygen florescence, where the additional intensity is not so obvious.

I am writing to see if anyone has any suggestion as to what the origins of these peaks might be. They are not due to additional elements, as they appear at the same incident energies as the main florescence peaks, i.e. the Co shoulder appears at the same incident energies as the main Co peak, and the O shoulder appears at the same incident energies as the main O peak. It is possible that the peaks result form other transitions. Considering Co, the main peak is due to L3/L2-M4 transitions, and the shoulder is in a position that could be consistent with L3/L2-M1 transitions. However, by fitting the peaks with Gaussians, one finds an area for the shoulder that is about 25% of the area of the main peak. This is significantly larger than what one might expect from tabulated transition strengths like those given in Hephaestus.

To summarize, does anyone know what these shoulders might result from if not lower energy transitions? If they are low energy transitions, why would the relative transition strengths differ from tabulated values?

Thank you, George

_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit

Matt Newville

1 Apr 1 Apr

12:10 a.m.

Hi George, Calibrated spectra would help, but if we guess the calibration is 0.56 bins/ eV, then we'd have line E (eV) bin # --------------------------------------------- Co La1,2 775 434 O Ka1,2 525 294 C Ka1,2 277 155 which looks pretty good. This puts the the unknown peak near 380/0.56 which is 678 eV. Tthat's very close to Co Ll (M1->L3), which is at 677 eV, and should be a bit less than 10% of Co La1 and La2, which is roughly right. So, I think it's Co Ll. That says the sample is just Co, C, and O. Is that reasonable? What surprises me is that there is no signal from the elastic peak. Was that somehow filtered out? The fact that the counts don't go to zero between C and O could be many factors, including incomplete charge collection. This (and Compton scattering) generally make peaks have a slightly non-Gaussian shape, with a low-energy tail. Cheers, --Matt On Mon, Mar 31, 2014 at 5:01 PM, George Sterbinsky wrote:

...

Hello,

I am writing with a general XAS question. It does not necessarily pertain to Ifeffit, however, I think the topic is something some, maybe most, list members will be knowledgeable about. So it seems like this list is a good place to post this question.

On to the question. I have attached a plot of a MCA spectrum collected with a vortex silicon drift detector. The spectrum is actually the average of several spectra, all collected in the post edge region of the Co L-edge. The spectra were averaged to reduce noise. The three peaks result from fluorescence from carbon, oxygen, and cobalt. Low-energy shoulders on the Co and O peaks are also observed. These can be seen as the regions of the spectrum that are not well reproduced by the fit. The main reason I included the fit in the plot is to illustrate the presence of these shoulders, particularly in the oxygen florescence, where the additional intensity is not so obvious.

I am writing to see if anyone has any suggestion as to what the origins of these peaks might be. They are not due to additional elements, as they appear at the same incident energies as the main florescence peaks, i.e. the Co shoulder appears at the same incident energies as the main Co peak, and the O shoulder appears at the same incident energies as the main O peak. It is possible that the peaks result form other transitions. Considering Co, the main peak is due to L3/L2-M4 transitions, and the shoulder is in a position that could be consistent with L3/L2-M1 transitions. However, by fitting the peaks with Gaussians, one finds an area for the shoulder that is about 25% of the area of the main peak. This is significantly larger than what one might expect from tabulated transition strengths like those given in Hephaestus.

To summarize, does anyone know what these shoulders might result from if not lower energy transitions? If they are low energy transitions, why would the relative transition strengths differ from tabulated values?

Thank you, George

_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit

-- --Matt Newville <newville at cars.uchicago.edu> 630-252-0431

George Sterbinsky

3:49 a.m.

Hi Matthew, Thanks for your reply. see below. On Mon, Mar 31, 2014 at 6:30 PM, Matthew Marcus wrote:

...

Hmmm. An interesting brain-teaser. Scaling very roughly from the graph, the mystery peak shows up at about 715eV. This is consistent with Fe. Could there be a background from Fe somewhere in the detection systen?

I don't think so. The shoulder appears when the incident energy corresponds to the Co L-edge, at the same time as the main Co peak. Also, MCA spectra collected above and below the Fe L-edge are pretty much identical.

...

The intensity between the O and Co peaks doesn't really look like a peak, but just like a background tail, perhaps from the strong Co peak. It seems to be a little to the right of where N (from, for instance, a kapton window) would be. I could be wrong, in which case my Fe theory goes away.

I'm assuming you mean the intensity between C and O. I've attached a plot of a MCA spectrum from the Co L pre-edge. This is again the sum of several spectra collected in the pre-edge region to reduce noise.The intensity between C and O is still present, so I don't think it is due to a tail from the Co peak. I also agree that is to the right of where one expects N.

...

Now, I thought about an escape peak from Si, but I suspect that this is really small down here and anyway, the Si L-fluorescence energy is 92eV or so, and the extra peak is about 55eV below the main peak.

Did you duck below the Fe edge to see if the peak to the L of Co goes away?

Yes, as noted above, the attached MCA spectrum from the Co pre-edge is pretty much identical to spectra collected below the Fe L-edge. A peak is seen at energies consistent with Fe and/or Co florescence. I think it is actually a combination of 2nd order cobalt, a small amount of fluorine, and maybe some elastic scatter. However regardless of its origin, it is much weaker than the Co L3/L2-M4 and the shoulder on that peak, so I don't think they are related.

...

The peaks in the fit seem to be consistently to the left of the ones in the data. Any idea why? Was the energy scale not a fit parameter?

My guess is that they are shifted because they are trying to fit the intensity of the shoulders. In order to analyze the 'extra' peaks, it may be useful to optimize the fit

...

to the main peaks, then subtract to get the residual, including the 'extras'. What would happen if you added Fe and maybe N to the fit?

Yes, I do get a good fit when I do that. In fact, that is what I originally thought before I started to look at the data a little more closely and ruled out Fe in the ways explained above. I'll take a look at the residuals, but I need to play around with the fit to get rid of the shifting to left of the peaks. Thanks, George

...

I admit that I've never done soft X-ray fluorescence, so my ideas come from hard X-ray experience, which may not totally apply. mam

On 3/31/2014 3:01 PM, George Sterbinsky wrote:

...
Hello,

I am writing with a general XAS question. It does not necessarily pertain to Ifeffit, however, I think the topic is something some, maybe most, list members will be knowledgeable about. So it seems like this list is a good place to post this question.

On to the question. I have attached a plot of a MCA spectrum collected with a vortex silicon drift detector. The spectrum is actually the average of several spectra, all collected in the post edge region of the Co L-edge. The spectra were averaged to reduce noise. The three peaks result from fluorescence from carbon, oxygen, and cobalt. Low-energy shoulders on the Co and O peaks are also observed. These can be seen as the regions of the spectrum that are not well reproduced by the fit. The main reason I included the fit in the plot is to illustrate the presence of these shoulders, particularly in the oxygen florescence, where the additional intensity is not so obvious.

I am writing to see if anyone has any suggestion as to what the origins of these peaks might be. They are not due to additional elements, as they appear at the same incident energies as the main florescence peaks, i.e. the Co shoulder appears at the same incident energies as the main Co peak, and the O shoulder appears at the same incident energies as the main O peak. It is possible that the peaks result form other transitions. Considering Co, the main peak is due to L3/L2-M4 transitions, and the shoulder is in a position that could be consistent with L3/L2-M1 transitions. However, by fitting the peaks with Gaussians, one finds an area for the shoulder that is about 25% of the area of the main peak. This is significantly larger than what one might expect from tabulated transition strengths like those given in Hephaestus.

To summarize, does anyone know what these shoulders might result from if not lower energy transitions? If they are low energy transitions, why would the relative transition strengths differ from tabulated values?

Thank you, George

_______________________________________________ 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

George Sterbinsky

3:57 a.m.

Hi Matt, Thanks for your reply. Please see below. On Mon, Mar 31, 2014 at 8:10 PM, Matt Newville wrote:

...

Hi George,

Calibrated spectra would help, but if we guess the calibration is 0.56 bins/ eV, then we'd have

I've attached a plot of the data showing the x-axis in keV. Also, as requested by Zack, I've attached a two column data file.

...

line E (eV) bin # --------------------------------------------- Co La1,2 775 434 O Ka1,2 525 294 C Ka1,2 277 155

which looks pretty good. This puts the the unknown peak near 380/0.56 which is 678 eV. Tthat's very close to Co Ll (M1->L3), which is at 677 eV, and should be a bit less than 10% of Co La1 and La2, which is roughly right.

Based on fitting the spectrum with Gaussians, the Co LI is 24% of the La2. As you mention, a value of roughly 10% is expected. What could cause such a discrepancy?

...

So, I think it's Co Ll. That says the sample is just Co, C, and O. Is that reasonable?

Yes, it is. I think maybe a little fluorine too, but it is very weak as I mention in my response to Matthew.

...

What surprises me is that there is no signal from the elastic peak. Was that somehow filtered out?

No, if there is an elastic peak it is probably lost under the Co La2. Thanks, George

...

The fact that the counts don't go to zero between C and O could be many factors, including incomplete charge collection. This (and Compton scattering) generally make peaks have a slightly non-Gaussian shape, with a low-energy tail.

...

Cheers,

--Matt

...
Hello,

I am writing with a general XAS question. It does not necessarily

...
Ifeffit, however, I think the topic is something some, maybe most, list members will be knowledgeable about. So it seems like this list is a good place to post this question.

On to the question. I have attached a plot of a MCA spectrum collected with a vortex silicon drift detector. The spectrum is actually the average of several spectra, all collected in the post edge region of the Co L-edge. The spectra were averaged to reduce noise. The three peaks result from fluorescence from carbon, oxygen, and cobalt. Low-energy shoulders on

...
and O peaks are also observed. These can be seen as the regions of the spectrum that are not well reproduced by the fit. The main reason I included the fit in the plot is to illustrate the presence of these shoulders, particularly in the oxygen florescence, where the additional intensity is not so obvious.

I am writing to see if anyone has any suggestion as to what the origins of these peaks might be. They are not due to additional elements, as they appear at the same incident energies as the main florescence peaks, i.e.

...
Co shoulder appears at the same incident energies as the main Co peak, and the O shoulder appears at the same incident energies as the main O peak. It is possible that the peaks result form other transitions. Considering Co, the main peak is due to L3/L2-M4 transitions, and the shoulder is in a position that could be consistent with L3/L2-M1 transitions. However, by fitting the peaks with Gaussians, one finds an area for the shoulder

...
about 25% of the area of the main peak. This is significantly larger than what one might expect from tabulated transition strengths like those given in Hephaestus.

To summarize, does anyone know what these shoulders might result from if not lower energy transitions? If they are low energy transitions, why would

On Mon, Mar 31, 2014 at 5:01 PM, George Sterbinsky wrote: pertain to the Co the that is the

...
relative transition strengths differ from tabulated values?

Thank you, George

_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit

-- --Matt Newville <newville at cars.uchicago.edu> 630-252-0431 _______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit

Matthew Marcus

3:45 p.m.

OK, I think the "other Matt" has solved it - it's the L1 peak. I was confused by an inaccurate attempt at reading the energy scale. Also, I somehow didn't read correctly the branching ratio from Hephaestus. mam On 3/31/2014 8:57 PM, George Sterbinsky wrote:

...

Hi Matt,

Thanks for your reply. Please see below.

On Mon, Mar 31, 2014 at 8:10 PM, Matt Newville mailto:newville@cars.uchicago.edu> wrote:

Hi George,

Calibrated spectra would help, but if we guess the calibration is 0.56 bins/ eV, then we'd have

I've attached a plot of the data showing the x-axis in keV. Also, as requested by Zack, I've attached a two column data file.

line E (eV) bin # --------------------------------------------- Co La1,2 775 434 O Ka1,2 525 294 C Ka1,2 277 155

which looks pretty good. This puts the the unknown peak near 380/0.56 which is 678 eV. Tthat's very close to Co Ll (M1->L3), which is at 677 eV, and should be a bit less than 10% of Co La1 and La2, which is roughly right.

Based on fitting the spectrum with Gaussians, the Co LI is 24% of the La2. As you mention, a value of roughly 10% is expected. What could cause such a discrepancy?

So, I think it's Co Ll. That says the sample is just Co, C, and O. Is that reasonable?

Yes, it is. I think maybe a little fluorine too, but it is very weak as I mention in my response to Matthew.

What surprises me is that there is no signal from the elastic peak. Was that somehow filtered out?

No, if there is an elastic peak it is probably lost under the Co La2.

Thanks, George

The fact that the counts don't go to zero between C and O could be many factors, including incomplete charge collection. This (and Compton scattering) generally make peaks have a slightly non-Gaussian shape, with a low-energy tail.

Cheers,

--Matt

On Mon, Mar 31, 2014 at 5:01 PM, George Sterbinsky mailto:GeorgeSterbinsky@u.northwestern.edu> wrote: > Hello, > > I am writing with a general XAS question. It does not necessarily pertain to > Ifeffit, however, I think the topic is something some, maybe most, list > members will be knowledgeable about. So it seems like this list is a good > place to post this question. > > On to the question. I have attached a plot of a MCA spectrum collected with > a vortex silicon drift detector. The spectrum is actually the average of > several spectra, all collected in the post edge region of the Co L-edge. The > spectra were averaged to reduce noise. The three peaks result from > fluorescence from carbon, oxygen, and cobalt. Low-energy shoulders on the Co > and O peaks are also observed. These can be seen as the regions of the > spectrum that are not well reproduced by the fit. The main reason I included > the fit in the plot is to illustrate the presence of these shoulders, > particularly in the oxygen florescence, where the additional intensity is > not so obvious. > > I am writing to see if anyone has any suggestion as to what the origins of > these peaks might be. They are not due to additional elements, as they > appear at the same incident energies as the main florescence peaks, i.e. the > Co shoulder appears at the same incident energies as the main Co peak, and > the O shoulder appears at the same incident energies as the main O peak. It > is possible that the peaks result form other transitions. Considering Co, > the main peak is due to L3/L2-M4 transitions, and the shoulder is in a > position that could be consistent with L3/L2-M1 transitions. However, by > fitting the peaks with Gaussians, one finds an area for the shoulder that is > about 25% of the area of the main peak. This is significantly larger than > what one might expect from tabulated transition strengths like those given > in Hephaestus. > > To summarize, does anyone know what these shoulders might result from if not > lower energy transitions? If they are low energy transitions, why would the > relative transition strengths differ from tabulated values? > > Thank you, > George > > > > > _______________________________________________ > 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 >

-- --Matt Newville http://cars.uchicago.edu> 630-252-0431 tel:630-252-0431 _______________________________________________ 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

George Sterbinsky

4:56 p.m.

Hi Matthew, Hephaestus shows the strength of the Ll emission to be about 10% of the La2 emission. When fitting the spectrum as a sum of Gaussians, I find that the area of of the Gaussian used to fit the Ll is 24% of that used to fit the La2. What would cause the ratio determined from tabulated data to differ from that found in the data I collected? Am I misunderstanding the meaning of the "strength" value in Hephaestus? Thanks, George On Tue, Apr 1, 2014 at 11:45 AM, Matthew Marcus wrote:

...

OK, I think the "other Matt" has solved it - it's the L1 peak. I was confused by an inaccurate attempt at reading the energy scale. Also, I somehow didn't read correctly the branching ratio from Hephaestus. mam

On 3/31/2014 8:57 PM, George Sterbinsky wrote:

...
Hi Matt,

Thanks for your reply. Please see below.

On Mon, Mar 31, 2014 at 8:10 PM, Matt Newville < newville@cars.uchicago.edu mailto:newville@cars.uchicago.edu> wrote:

Hi George,

Calibrated spectra would help, but if we guess the calibration is 0.56 bins/ eV, then we'd have

I've attached a plot of the data showing the x-axis in keV. Also, as requested by Zack, I've attached a two column data file.

line E (eV) bin # --------------------------------------------- Co La1,2 775 434 O Ka1,2 525 294 C Ka1,2 277 155

which looks pretty good. This puts the the unknown peak near 380/0.56 which is 678 eV. Tthat's very close to Co Ll (M1->L3), which is at 677 eV, and should be a bit less than 10% of Co La1 and La2, which is roughly right.

Based on fitting the spectrum with Gaussians, the Co LI is 24% of the La2. As you mention, a value of roughly 10% is expected. What could cause such a discrepancy?

So, I think it's Co Ll. That says the sample is just Co, C, and O. Is that reasonable?

Yes, it is. I think maybe a little fluorine too, but it is very weak as I mention in my response to Matthew.

What surprises me is that there is no signal from the elastic peak. Was that somehow filtered out?

No, if there is an elastic peak it is probably lost under the Co La2.

Thanks, George

The fact that the counts don't go to zero between C and O could be many factors, including incomplete charge collection. This (and Compton scattering) generally make peaks have a slightly non-Gaussian shape, with a low-energy tail.

Cheers,

--Matt

On Mon, Mar 31, 2014 at 5:01 PM, George Sterbinsky > wrote: > Hello, > > I am writing with a general XAS question. It does not necessarily pertain to > Ifeffit, however, I think the topic is something some, maybe most, list > members will be knowledgeable about. So it seems like this list is a good > place to post this question. > > On to the question. I have attached a plot of a MCA spectrum collected with > a vortex silicon drift detector. The spectrum is actually the average of > several spectra, all collected in the post edge region of the Co L-edge. The > spectra were averaged to reduce noise. The three peaks result from > fluorescence from carbon, oxygen, and cobalt. Low-energy shoulders on the Co > and O peaks are also observed. These can be seen as the regions of the > spectrum that are not well reproduced by the fit. The main reason I included > the fit in the plot is to illustrate the presence of these shoulders, > particularly in the oxygen florescence, where the additional intensity is > not so obvious. > > I am writing to see if anyone has any suggestion as to what the origins of > these peaks might be. They are not due to additional elements, as they > appear at the same incident energies as the main florescence peaks, i.e. the > Co shoulder appears at the same incident energies as the main Co peak, and > the O shoulder appears at the same incident energies as the main O peak. It > is possible that the peaks result form other transitions. Considering Co, > the main peak is due to L3/L2-M4 transitions, and the shoulder is in a > position that could be consistent with L3/L2-M1 transitions. However, by > fitting the peaks with Gaussians, one finds an area for the shoulder that is > about 25% of the area of the main peak. This is significantly larger than > what one might expect from tabulated transition strengths like those given > in Hephaestus. > > To summarize, does anyone know what these shoulders might result from if not > lower energy transitions? If they are low energy transitions, why would the > relative transition strengths differ from tabulated values? > > Thank you, > George > > > > > _______________________________________________ > Ifeffit mailing list > Ifeffit@millenia.cars.aps.anl.gov > http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit >

-- --Matt Newville http://cars.uchicago.edu>> 630-252-0431 tel:630-252-0431 _______________________________________________ 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

_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit

Matt Newville

5:26 p.m.

Hi George, On Tue, Apr 1, 2014 at 11:56 AM, George Sterbinsky wrote:

...

Hi Matthew,

Hephaestus shows the strength of the Ll emission to be about 10% of the La2 emission. When fitting the spectrum as a sum of Gaussians, I find that the area of of the Gaussian used to fit the Ll is 24% of that used to fit the La2. What would cause the ratio determined from tabulated data to differ from that found in the data I collected? Am I misunderstanding the meaning of the "strength" value in Hephaestus?

Thanks, George

Strength here means "fraction of the X-rays emitted from the decay of the initial level". So, well above L3 edge, Ll should be ~10% of La1,2. Well above L2 edge, Ln should be ~10% of Lb1,2. Those values of branching could be somewhat approximate. It's also possible that they don't properly account for the Coster-Kronig process (inter-shell transitions, known to be important for L shells, especially low energy lines where Auger transitions dominate). I thought they did take this into account, but perhaps that's not correct. of the correction is It wasn't completely clear what your incident energy is, but there are also elastic and Compton peaks to consider. I think the elastic line being folded into your main peak would tend to make the ratio of La1,2 to Ll (Lb1,2 and Ln) bigger, which seems to be the wrong way. It seems the incident energy was below the L1 edge (925 eV), because the Lb3,4 peaks (~855 eV) are missing... Then again, if the detector was at 90deg in the horizontal plane from a synchrotron source, the elastic peak may be tiny. Compton scattering might appear to increase the low energy side of the peak. I think you'd have to do a more careful analysis with the energy well above all Co L edges with a well-resolved Compton peak separated from the L lines to do a more quantitative analysis. Finally, it's known (perhaps not well-known) that the peaks are slightly asymmetric, with a low energy tail. This is usually a small effect, but perhaps it's part of what you're seeing. All that said, I don't have a lot of experience working at such low energies. --Matt

Matthew Marcus

10:07 p.m.

I don't think so. This may be a better question for M. Newville, or whoever wrote the program. mam On 4/1/2014 9:56 AM, George Sterbinsky wrote:

...

Hi Matthew,

Hephaestus shows the strength of the Ll emission to be about 10% of the La2 emission. When fitting the spectrum as a sum of Gaussians, I find that the area of of the Gaussian used to fit the Ll is 24% of that used to fit the La2. What would cause the ratio determined from tabulated data to differ from that found in the data I collected? Am I misunderstanding the meaning of the "strength" value in Hephaestus?

Thanks, George

On Tue, Apr 1, 2014 at 11:45 AM, Matthew Marcus mailto:mamarcus@lbl.gov> wrote:

OK, I think the "other Matt" has solved it - it's the L1 peak. I was confused by an inaccurate attempt at reading the energy scale. Also, I somehow didn't read correctly the branching ratio from Hephaestus. mam

On 3/31/2014 8:57 PM, George Sterbinsky wrote:

Hi Matt,

Thanks for your reply. Please see below.

On Mon, Mar 31, 2014 at 8:10 PM, Matt Newville mailto:newville@cars.uchicago.edu mailto:newville@cars.uchicago.edu>> wrote:

Hi George,

Calibrated spectra would help, but if we guess the calibration is 0.56 bins/ eV, then we'd have

I've attached a plot of the data showing the x-axis in keV. Also, as requested by Zack, I've attached a two column data file.

line E (eV) bin # ------------------------------__--------------- Co La1,2 775 434 O Ka1,2 525 294 C Ka1,2 277 155

which looks pretty good. This puts the the unknown peak near 380/0.56 which is 678 eV. Tthat's very close to Co Ll (M1->L3), which is at 677 eV, and should be a bit less than 10% of Co La1 and La2, which is roughly right.

Based on fitting the spectrum with Gaussians, the Co LI is 24% of the La2. As you mention, a value of roughly 10% is expected. What could cause such a discrepancy?

So, I think it's Co Ll. That says the sample is just Co, C, and O. Is that reasonable?

Yes, it is. I think maybe a little fluorine too, but it is very weak as I mention in my response to Matthew.

What surprises me is that there is no signal from the elastic peak. Was that somehow filtered out?

No, if there is an elastic peak it is probably lost under the Co La2.

Thanks, George

The fact that the counts don't go to zero between C and O could be many factors, including incomplete charge collection. This (and Compton scattering) generally make peaks have a slightly non-Gaussian shape, with a low-energy tail.

Cheers,

--Matt

On Mon, Mar 31, 2014 at 5:01 PM, George Sterbinsky mailto:GeorgeSterbinsky@u.northwestern.edu mailto:GeorgeSterbinsky@u.northwestern.edu>> wrote: > Hello, > > I am writing with a general XAS question. It does not necessarily pertain to > Ifeffit, however, I think the topic is something some, maybe most, list > members will be knowledgeable about. So it seems like this list is a good > place to post this question. > > On to the question. I have attached a plot of a MCA spectrum collected with > a vortex silicon drift detector. The spectrum is actually the average of > several spectra, all collected in the post edge region of the Co L-edge. The > spectra were averaged to reduce noise. The three peaks result from > fluorescence from carbon, oxygen, and cobalt. Low-energy shoulders on the Co > and O peaks are also observed. These can be seen as the regions of the > spectrum that are not well reproduced by the fit. The main reason I included > the fit in the plot is to illustrate the presence of these shoulders, > particularly in the oxygen florescence, where the additional intensity is > not so obvious. > > I am writing to see if anyone has any suggestion as to what the origins of > these peaks might be. They are not due to additional elements, as they > appear at the same incident energies as the main florescence peaks, i.e. the > Co shoulder appears at the same incident energies as the main Co peak, and > the O shoulder appears at the same incident energies as the main O peak. It > is possible that the peaks result form other transitions. Considering Co, > the main peak is due to L3/L2-M4 transitions, and the shoulder is in a > position that could be consistent with L3/L2-M1 transitions. However, by > fitting the peaks with Gaussians, one finds an area for the shoulder that is > about 25% of the area of the main peak. This is significantly larger than > what one might expect from tabulated transition strengths like those given > in Hephaestus. > > To summarize, does anyone know what these shoulders might result from if not > lower energy transitions? If they are low energy transitions, why would the > relative transition strengths differ from tabulated values? > > Thank you, > George > > > > > _________________________________________________ > Ifeffit mailing list > Ifeffit@millenia.cars.aps.anl.__gov mailto:Ifeffit@millenia.cars.aps.anl.gov mailto:Ifeffit@millenia.cars.aps.anl.gov> > http://millenia.cars.aps.anl.__gov/mailman/listinfo/ifeffit http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit >

-- --Matt Newville http://cars.uchicago.edu http://cars.uchicago.edu> 630-252-0431 tel:630-252-0431 tel:630-252-0431> _________________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.__gov mailto:Ifeffit@millenia.cars.aps.anl.gov mailto:Ifeffit@millenia.cars.aps.anl.gov> http://millenia.cars.aps.anl.__gov/mailman/listinfo/ifeffit http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit

_________________________________________________ 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 http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit

_________________________________________________ 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 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

George Sterbinsky

2 Apr 2 Apr

7:56 p.m.

Matt and Matthew, Thanks for your suggestions. I will try to collect some spectra well above the Co L-edge and see if they provide additional insight. Best, George On Tue, Apr 1, 2014 at 6:07 PM, Matthew Marcus wrote:

...

I don't think so. This may be a better question for M. Newville, or whoever wrote the program. mam

On 4/1/2014 9:56 AM, George Sterbinsky wrote:

...
Hi Matthew,

Hephaestus shows the strength of the Ll emission to be about 10% of the La2 emission. When fitting the spectrum as a sum of Gaussians, I find that the area of of the Gaussian used to fit the Ll is 24% of that used to fit the La2. What would cause the ratio determined from tabulated data to differ from that found in the data I collected? Am I misunderstanding the meaning of the "strength" value in Hephaestus?

Thanks, George

On Tue, Apr 1, 2014 at 11:45 AM, Matthew Marcus > wrote:

OK, I think the "other Matt" has solved it - it's the L1 peak. I was confused by an inaccurate attempt at reading the energy scale. Also, I somehow didn't read correctly the branching ratio from Hephaestus. mam

On 3/31/2014 8:57 PM, George Sterbinsky wrote:

Hi Matt,

Thanks for your reply. Please see below.

On Mon, Mar 31, 2014 at 8:10 PM, Matt Newville < newville@cars.uchicago.edu mailto:newville@cars.uchicago.edu mailto:newville@cars.uchicago.edu>> wrote:

Hi George,

Calibrated spectra would help, but if we guess the calibration is 0.56 bins/ eV, then we'd have

I've attached a plot of the data showing the x-axis in keV. Also, as requested by Zack, I've attached a two column data file.

line E (eV) bin # ------------------------------__---------------

Co La1,2 775 434 O Ka1,2 525 294 C Ka1,2 277 155

which looks pretty good. This puts the the unknown peak near 380/0.56 which is 678 eV. Tthat's very close to Co Ll (M1->L3), which is at 677 eV, and should be a bit less than 10% of Co La1 and La2, which is roughly right.

Based on fitting the spectrum with Gaussians, the Co LI is 24% of the La2. As you mention, a value of roughly 10% is expected. What could cause such a discrepancy?

So, I think it's Co Ll. That says the sample is just Co, C, and O. Is that reasonable?

Yes, it is. I think maybe a little fluorine too, but it is very weak as I mention in my response to Matthew.

What surprises me is that there is no signal from the elastic peak. Was that somehow filtered out?

No, if there is an elastic peak it is probably lost under the Co La2.

Thanks, George

The fact that the counts don't go to zero between C and O could be many factors, including incomplete charge collection. This (and Compton scattering) generally make peaks have a slightly non-Gaussian shape, with a low-energy tail.

Cheers,

--Matt

On Mon, Mar 31, 2014 at 5:01 PM, George Sterbinsky mailto:GeorgeSterbinsky@u.northwestern.edu>> wrote: > Hello, > > I am writing with a general XAS question. It does not necessarily pertain to > Ifeffit, however, I think the topic is something some, maybe most, list > members will be knowledgeable about. So it seems like this list is a good > place to post this question. > > On to the question. I have attached a plot of a MCA spectrum collected with > a vortex silicon drift detector. The spectrum is actually the average of > several spectra, all collected in the post edge region of the Co L-edge. The > spectra were averaged to reduce noise. The three peaks result from > fluorescence from carbon, oxygen, and cobalt. Low-energy shoulders on the Co > and O peaks are also observed. These can be seen as the regions of the > spectrum that are not well reproduced by the fit. The main reason I included > the fit in the plot is to illustrate the presence of these shoulders, > particularly in the oxygen florescence, where the additional intensity is > not so obvious. > > I am writing to see if anyone has any suggestion as to what the origins of > these peaks might be. They are not due to additional elements, as they > appear at the same incident energies as the main florescence peaks, i.e. the > Co shoulder appears at the same incident energies as the main Co peak, and > the O shoulder appears at the same incident energies as the main O peak. It > is possible that the peaks result form other transitions. Considering Co, > the main peak is due to L3/L2-M4 transitions, and the shoulder is in a > position that could be consistent with L3/L2-M1 transitions. However, by > fitting the peaks with Gaussians, one finds an area for the shoulder that is > about 25% of the area of the main peak. This is significantly larger than > what one might expect from tabulated transition strengths like those given > in Hephaestus. > > To summarize, does anyone know what these shoulders might result from if not > lower energy transitions? If they are low energy transitions, why would the > relative transition strengths differ from tabulated values? > > Thank you, > George > > > > > _________________________________________________ > 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 http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit >

-- --Matt Newville http://cars.uchicago.edu> http://cars.uchicago.edu> 630-252-0431 tel:630-252-0431> _________________________________________________ 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 < 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 < 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 < 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

_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit

4238

Age (days ago)

4240

Last active (days ago)

List overview

Download

9 comments

3 participants

participants (3)

George Sterbinsky
Matt Newville
Matthew Marcus

Peaks in MCA spectra

tags

participants (3)