Extended spectrometer

[staciag]

Does anyone have experience with the extended spectrometer offered by Cameca? Cameca has suggested that the extended spectrometer with a LTAP crystal has the best energy resolution and they have argued that the peak shift in oxygen will be measurable with the LTAP but not with PC1. For example, in their brochure, they show CuO versus Cu2O data. Does anyone have the extended spectrometer with the LTAP setup?

Thanks,
Stacia

[Probeman]

Hi Stacia,
I've never had access to an extended range spectrometer on a Cameca, so I'd be interested in what others' experiences are with this device. But the sensitivity for oxygen is going to be very poor at such a high sin theta.

It's true that the TAP (or LTAP) analyzing crystal will have better spectra resolution than a PC1 (or LDE1) multi-layer "crystal", but can you provide more details as to why you want to see the peak shift in oxygen?  Most of the time we just want to correct for the oxygen peak shift (or perform integrated intensity scans) to measure elemental concentrations and in these cases, the lower the spectral resolution, the smaller the peak shift correction.

Of course, sometimes on the microprobe we do want to measure the peak shifts, e.g., sulfur in basaltic glasses to try and determine the sulfur oxidation state, but usually such low energy peaks shifts are a hindrance rather than a help when attempting to measure elemental concentrations. In Oregon we've resorted to measuring sulfur at several different peak positions on multiple spectrometers (5 spectrometers tuned to different offset peak positions every 10 spectro units on the Cameca, from pyrrhotite to anhydrite), so we can determine both the peak shift (from the channel with the highest k-ratio) and also the concentration (by aggregating the intensities from all spectrometers for best sensitivity) at the same time.

Back to oxygen, what sort of applications are you thinking about?  Even using a PC1/LDE1 crystal one can see oxygen peak shifting and the intensities are awesome.  Might be better, hard to say, will depend on the concentrations and degree of expected peak shift.

Maybe also relevant, here is a topic on area peak factors (peak shift corrections) for boron:

https://probesoftware.com/smf/index.php?topic=536.0

and here is a comparison of PC1 and PC2 for oxygen:

https://probesoftware.com/smf/index.php?topic=261.0
john

[staciag]

Hi John,

We were discussing with Cameca different setups that would be ideal for economic geology research. They had suggested the extended spectrometer would be ideal for detecting oxygen for mapping between CuO vs. Cu2O (this is their main example and shown in their brochure) and for mapping between hematite and magnetite.

It sounds like there are likely other ways to get after this information (e.g., your example of sulfur) and that the extended spectrometer is not worth the extra cost and potential other problems (e.g., potential of wires getting hung up).

Thanks for the info,
Stacia

[John Donovan]

Hi John,

We were discussing with Cameca different setups that would be ideal for economic geology research. They had suggested the extended spectrometer would be ideal for detecting oxygen for mapping between CuO vs. Cu2O (this is their main example and shown in their brochure) and for mapping between hematite and magnetite.

It sounds like there are likely other ways to get after this information (e.g., your example of sulfur) and that the extended spectrometer is not worth the extra cost and potential other problems (e.g., potential of wires getting hung up).

Thanks for the info,
Stacia

Hi Stacia,
OK, I understand now what you want to do.

So if you want to distinguish between different oxides, e.g., CuO and Cu2O or hematite and magnetite, the most effective method I think would be to do quantitative x-ray mapping of these phases, using our fully quantitative methods on each pixel in our CalcImage app. It's super easy and the accuracy is as good as point analyses.  The difference in oxygen concentration between these phases is easy distinguishable once the x-ray maps are quantified.

But particularly for oxygen, you really want to rely on full quantification methods at the pixel level because the matrix corrections are quite large and calibration curves simply won't do the trick. Here is an example from Anette von der Handt comparing raw oxygen intensities and full quant oxygen concentrations in some oxides:



The interesting thing about measuring oxygen in oxides (and silicates) is that it automatically takes care of the "excess oxygen" issues that geologists often need to deal with, when both ferric and ferrous iron are present!

For point analyses we can deal with the "excess oxygen" if we know the mineral as described here:

https://probesoftware.com/smf/index.php?topic=92.msg8612#msg8612

But for an x-ray map it would be more difficult to specify a cation/oxygen ratio for the various phases (it could be done, but only with significant additional effort once a modal analysis has been performed), so the easiest way would be to simply include oxygen in ones x-ray mapping and perform full quantification at the pixel level.

[sem-geologist]

We have such spectrometer from 2014 on our SXFiveFE.
I can tell you that I am having love-hate relation with it.

cons:
1) I think the mapping approach (on PC0) is better than Oxygen peak shift, it is not clear how much it is shifted, in case of mixed oxidation states the peak will be broadened. Generally I see O Ka shift to be really complicated to have any use.
2) on extended spectrometer the oxygen Ka is situated at the very edge of "spectrum", it is enough to get some mechanical de-calibration and oxygen peak is no more accessible (until Spectrometer is serviced).
3) Best spectral resolution (?) – only for Oxygen! Please mind, that normally TAP's suffer from too broad peaks - they (mounted diff. crystals) are normally smaller than PET or LiF and is further narrowed by blanking at the edges, so that there would be even narrower peaks and less overlaps. And here you have LTAP - at anything else than Oxygen (and Florine) with complicated composition that crystal is a complete nightmare! Well that can be mitigated a bit if You use ProbeSoftware and can use MAN background (unfortunately, we don't have ProbeSoftware), because it is impossible to find any place for classical background measurements for Si, Al, Mg, M lines of REE, Hf- W, Y, Sr, Rb. That spectrometer is idle for most of our analysis, unless very simple minerals are analysed.
4) If You are going to have LTAP, you can have only one more crystal mounted on the turret, large crystals are mounted only on the double and not quadruple turrets. Those actually are much better designed than new type (2014) quadruple, and so I would not fear for wires. We have this extended spectrometer with LTAP and LPC0. We are often switching and had no wire failure (differently than with quadruple turrets...)

pros:
1) I prefer this crystal over LPC0 for F measurements. Maybe I would have different opinion if I would use Probesoftware and its MAN background modeling, however with classical background measurement LTAP have enough of space (for F) out from interference for background measurement in complicated minerals (containing REE, Fe (try that on LPC0, PHA wont help you there)).
2) I think LTAP can have some potential for measuring Fe L lines for Fe oxidation state determination (it also suffers from similar uncertainties as Oxygen peak shift), That needs to be clarified. As for oxygen I find LPC0 quite impressive for that.

Concluding:
I would prefer instead of current extended spectrometer with x2 turret (LPC0, LTAP) - the spectrometer with x4 turret (TAP, PC0, PET). It would then could work, instead of being 90% of time idle.