Using the "Flank" method in Probe for EPMA

[John Donovan]

Some of you may be familiar with the so called "Flank" method for the determination of Fe2+/Fe3+ in some minerals by comparing subtle peak shifts in the region of the Fe La/Lb emission lines for different Fe oxidation states.

Höfer, Heidi E., and Gerhard P. Brey. "The iron oxidation state of garnet by electron microprobe: Its determination with the flank method combined with major-element analysis." American Mineralogist 92.5-6 (2007): 873-885.

I myself have never attempted these measurements but I hear that they can be useful.  Anyway, Stuart Kearns, Ben Buse, Ery Hughes and I were chatting about some possible tweaks to PFE that might make these "Flank" determinations easier and the general consensus was that these measurements are so esoteric that there probably isn't much value in making changes to the software.

But I decided that one small change might be helpful and not just for "Flank" method determinations, and that is adding a "difference" plot option to the Plot! window as shown here:



Using the this new feature I plotted two samples from Ery's run and we obtain this plot where blue is the difference plot:



Maybe this will aid in the selection of these peak "Flanks"... though in fact Ery apparently averaged many of these plots together to get better counting statistics (too long of a beam exposure and one risks oxidizing the sample further).

But this feature may also be useful in other studies such as this plot comparing SiO2 and SiO:



Then by also checking the "Normalize Samples" option we obtain this:



These differences become even more clear when using a PET rather than a TAP crystal:



Maybe this "difference wavescan" option will be useful to some of you...

[John Donovan]

We recently added a new output method for so called "Flank" intensities as requested by Ben Buse, Stuart Kearns and Ery Hughes:



With this output method one can export the unknown intensities and also the time dependent intensities (TDI) associated with those elements which were acquired with that option.  Also, one does not require standard intensities to export data, unlike most of the other export methods in Probe for EPMA.

This output method can be used from the Output menu for all unknown samples, for selected samples from the Analyze! window using a right mouse click on the sample list.

[Ben Buse]

Thank you John,

We've recently been doing some work as part of a masters student project, this addition exporting the data required without standards greatly helps

The protocol we've been doing, inspired by Ery Hughes paper is as follows:

1 - Acquire wavescans on two extremes - ideally of same mineral type - otherwise garnets - one extreme high Fe content and high Fe2+ (which gives max absorption of Fe La), other extreme lower Fe content and highest Fe3+ (which gives minimum absorption of Fe Lb - although in both cases high Fe content is good for producing clean spectra. To achieve clean spectra either need high probe current + long dwell time - which on a single spectra necessitates stop start motion of spectrometer, otherwise multiple spectra which aggregate outside PFE. Must ensure Fe does not change oxidisation state whilst acquiring spectra - a problem for some sample types - requiring appropriate beam conditions. - eg glass defocused beam, move stage whilst acquire spectra (using PFE option), lower beam currents, multiple spectra aggregated

2 - Normalise two spectra and calculate difference spectra - if single spectra for each can be done within PFE Plot window

3 - Select flank positions and enter as element peak positions. (Either doing both flank positions on each spectrometer, or one flank position on one spectrometer, and another flank position on another spectrometer, to avoid spectrometer movement). Multiple spectrometers used to increase precision as increase count time. Select MAN backgrounds to avoid measuring background positions

4 - Acquire measurements on standard samples (minerals of same type for which Fe2+/Fe3+ ratio and iron content known) and on unknown samples (for which total Fe content must be measured separately). Depending on sample type instability must be considered reflected in choice of beam conditions (e.g. Large spot size if possible), use of TDI to correct back to initial value etc.

5 - Export data using custom output analysis #12 flank

6 - Use your preferred program to plot flank ratio against Fe2+ content. Use standard samples as calibration curve. Determine Fe2+ content for unknowns which when compared against total Fe content gives Fe3+/Fe2+ ratio