Peaking for silicate minerals

[BenjaminWade]

Hi all
I was wondering as to what peoples protocols are for peaking a "usual" suite of major elements for silicate minerals (Si,Al,Fe,Na etc). Are people using the ROM peaking method of their instrument then manually reviewing/manipulating their peak position after, or are they using one of the peaking methods in PfE (Maxima, Gaussian, etc) and finding one is better than another?

Due to some reasonable peak shifts in the different silicate minerals for elements like Si/Al, are people peaking on matrix matched standards as best they can and having individual sample setup files for different minerals?

I found a great talk online by John Fournelle given in 2007 on this, but was wondering what other people are doing at the moment.

Cheers

[Ben Buse]

Hi,

Yes John Fournelle's talk is very good and interesting. Fortunately most of the time the peak shifts are so small it is generally not necessary to repeak for different silicates. Exceptions being S. That said its good where possible to use similar standards and unknowns, I guess for feldspars and glasses I would use albite/sanidine for Si & Al, whereas olivine for Si; key being the standard having similar or greater intensity than standard.

I tend to do the auto peaking and use maximum intensity - I've set the peak windows as to have lots of points over the peak. I will review peak positions but rarely change them.

Ben

[Probeman]

Hi Ben and Ben,
My method is normally to just peak on my primary standards (using the Peak On Assigned Standards checkbox in Automate!).  I almost always also check the box to acquire a PHA scan after the peaking.

I always use the Maxima option for peak fitting.

Then I review the peaking scans and PHA scans from the Run | Display, Fit and Export Spectrometer Peaking and PHA Scans menu dialog.

If necessary (sometimes on a very narrow peak there might not be quite enough points for an ideal fit depending on how wide I've set my peaking scan range and how many peak scan points I'm using), I might right click to adjust the peak position slightly.  I almost never need to do that however.
john

[Probeman]

...Exceptions being S.

My "rule of thumb" for peaking sulfur (for trace sulfur in glasses) is to peak on pyrite. From prior efforts I know that anhydrite will always be 30 points lower (in Cameca units).

Our experience is that for most of the basaltic glasses we analyze for the Wallace group is that the sulfur peak in these glass will almost always be slightly oxidized and therefore we "de-tune" the sulfur peak down by about 10 points (in Cameca units).  I'm not a geologist but they tell me this works for them.

Of course until one knows for sure where the sulfur peak position exactly is for ones own samples, you will need to perform a high precision scan on your glass to determine the exact sulfur peak position for your particular glasses.

Note that when you do this it's very important  to utilize the "Use increment during scanning" option, because the sulfur will oxidize over long exposure to the beam (from heating I assume).
john

[BenjaminWade]

Hi all
Awesome, thankyou for the feedback. Its pretty much what I was doing (I use albite for Si/Al), however I am using the ROM peaking. I might try it with the Maxima option and see how it goes.

The reason I was asking is that I know I don't get good results using albite as the standard for a number of silicate elements (aluminosilicates, biotite etc), which I am sure is due to the peakshift (see attached images for my quick and dirty maxima values I measured on some of my silicates). Doing some back of the envelope calculations these shifts can affect the Si/Al values markedly, which is quite obvious in real data when you change standards around. As a result I have been switching standards to matrix match, which in effect I guess is like a "fudge" factor as that matrix matched standard would have its peak measured slightly "off peak" similar to your unknown.

Was just wondering how people might be getting around this, but sounds like people are also running one set of peaking and switching standards. Thanks again!

Cheers

[Ben Buse]

Hi Ben,

If you repeak on each of the materials do you see a significant change intensity. How large a shift in intensity do these peak shifts equate to. Although it will depend on spectrometer resolution

Ben

[BenjaminWade]

Hi Ben
I confess that it was very quick and dirty and only had the time to repeak and check on a couple of other minerals at the time. There was some movement in the peak maxima after repeaking again but only of a couple of units.
Just to take the extreme for Si, looking at my wavescans if measure at the albite Si maxima position in kyanite, my intensity is ~4.5% lower than it should be (275 CPS/nA vs 288 CPS/nA). This is almost exactly what I see in spot data if I use albite as the Si standard for kyanite (I get 16.55wt% Si when it should be 17.33wt%).

I guess this is the extreme though. Was curious as to how people were dealing with it.

Cheers