Virtual Standards

[Joe Boesenberg]

John

Very cool you can generate a virtual standard for a noble gas. I read the help section of PfE. I am just starting with PfE, so forgive my idiot questions. I understand that you use the two adjacent elements peaks (K and Cl) to best determine the Ar peak position. I assume the virtual "standard" is considered 100% pure Ar, correct? The virtual intensity for the Ar is simply an estimated peak intensity based on the Cl and K standard peak heights, correct? Thus, your 10% error estimate.


Thanks
Joe

[John Donovan]

As seen in this screen shot, one actually needs three standards for this virtual standard calculation:

https://smf.probesoftware.com/index.php?topic=179.msg6743#msg6743

The first standard is the virtual standard itself which does not actually exist except as a composition in the standard database.  It could be a pure element, or any other composition, so long as the virtual element in question is present in the virtual standard composition.

The other two standards are standards that you actually have in your instrument. For example, for measuring Ar Ka you would want a Cl (or S) standard and a K standard, both measured on a PET spectrometer.

You tune your spectrometer to the Cl (or S) Ka peak (Cl ka on PET = ~151.3) and measure the Cl intensity on your actual Cl standard. Note the offset from the ideal position in PFE for the Cl Ka peak.

Then tune the same spectrometer to the K Ka peak (K ka on PET = ~119.7) and measure the K intensity on your actual K standard.  Again, note the offset from the ideal position in PFE for the K Ka peak position.

The reason for noting these offsets, is to improve accuracy in the interpolation of the Ar intensity, by averaging the two peak offsets and applying that to your "predicted" Ar Ka spectrometer position which  you will use in the Ar measurements on your unknowns as discussed here:

https://smf.probesoftware.com/index.php?topic=179.msg10578#msg10578

With the offsets from your Cl and K peaking applied to your Ar peak position, you might do better than 10% accuracy.

[aburnham]

Hopefully your user has done some background reading and alerted you to this study https://www.nature.com/articles/nature04583, which lacks much detail but basically uses a synthetic glass that was characterised by different methods (noble gas mass spectrometry).
Note of course that He does not have any X-ray emission lines (and in any case diffuses rapidly out of high pressure glasses).

[John Donovan]

Hopefully your user has done some background reading and alerted you to this study https://www.nature.com/articles/nature04583, which lacks much detail but basically uses a synthetic glass that was characterised by different methods (noble gas mass spectrometry).
Note of course that He does not have any X-ray emission lines (and in any case diffuses rapidly out of high pressure glasses).

The cool thing about the virtual standard method in Probe for EPMA is that it doesn't require an actual noble gas standard! 

Also it is worth mentioning that using a minor/trace element standard as a primary standard is problematic from a statistical point of view as explained in this presentation:

https://www.youtube.com/watch?v=mOca7-G4FvQ&ab_channel=ProbeSoftwareInc

So the virtual standard method solves both problems...