One Calibration for Multiple Metamorphic Mineral Phases?

[MelissaJDrignon]

Dear all,

I received a request for a single calibration that could be used to analyze oxides and both hydrous and anhydrous silicates in metamorphic thin sections, including garnet, chlorite, mica, feldspar, ilmenite, amphibole, and pyroxene.

The proposed elements are SiO₂, TiO₂, Al₂O₃, Cr₂O₃, FeO, MnO, MgO, CaO, Na₂O, and K₂O, with Cl added for the hydrous phases and La and Lu for garnet.

What are your thoughts on developing and using a single calibration for this combination of elements and phases?

Many thanks,
Melissa.

[Scott B.]

With that element list, it probably wouldn't be terribly difficult to create a single method, but in my opinion, it's not the most efficient way to go about it.

You'd be wasting a lot of time by counting elements you have no hope of measuring (e.g. Cl in feldspar) and counting for long periods on elements that are abundant (e.g. if you want Fe in feldspar, you need to count a lot longer than for Fe in ilmenite).

Instead, I suggest using "Sample Setups" in the Automate! window. You can make a setup for each phase (with different count times, elements, conditions, etc), and then assign those in Automate! when you run the analyses. Have the user include a label in their sample name (e.g. Ox, FSP, PYX, etc), then when you get ready to run, you can search by that character string and assign each different phase's analyses a specific setup.

Some tips and caveats to this...

1) If the different phases require spectrometer configurations with different crystals on particular channels, you will be forcing a crystal flip in mid run without repeaking. JEOL claims this can be done safely, but conventional wisdom is to not do this. I'd run a series of tests to prove to myself it can be done without significant peak shifting before I tried it. Even if it does work, there's likely a significant time penalty, which may wipe out any other benefits?

2) It may be most efficient to create a "Mother of All Recipes" sample setup that has all of the various elements you'll be targeting, so you can run all the standards only once, instead of multiple times for each different phase's setup. You'd then do all of the peaking and tuning on that MOAR, run all the standards a single time, then break up the MOAR into sub-routines for each phase and run the unknowns with the appropriate one. The efficiency is highly method dependent. Watch out for huge spectrometer runs or crystal flips or just general "incompatibility" which might slow the routine down so much that breaking it into separate runs would be more efficient (e.g. if a single method takes 6 minutes, but two individual methods are only 2 minutes each). You could still start with a MOAR for tuning and peaking, but then bust it up and apply it to different standards just like you can to unknowns (or use "Multiple Setups" on the standards to have it automatically run them multiple times with different setups).
 
3) Here's a YouTube video on setups if you need a refresher on using sample setups in PFE...

https://www.youtube.com/watch?v=NZc1Ey3V0ms

Hope that helps. Holler if you have more questions.

[John Donovan]

There's a nice discussion on using different samples setups for different minerals here:

https://smf.probesoftware.com/index.php?topic=1783.0