[Ifeffit] Peaks in MCA spectra

[George Sterbinsky]

Hi Matthew,

Thanks for your reply. see below.

On Mon, Mar 31, 2014 at 6:30 PM, Matthew Marcus <mamarcus at lbl.gov> 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 at millenia.cars.aps.anl.gov
>> http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
>>
>>  _______________________________________________
> Ifeffit mailing list
> Ifeffit at millenia.cars.aps.anl.gov
> http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://millenia.cars.aps.anl.gov/pipermail/ifeffit/attachments/20140331/5dfdadb4/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: MCA_Co_preEdge.png
Type: image/png
Size: 16396 bytes
Desc: not available
URL: <http://millenia.cars.aps.anl.gov/pipermail/ifeffit/attachments/20140331/5dfdadb4/attachment.png>