Separating Fe into FeO & Fe2O3

[jeb]

What is the best/most accepted way to separate Fe into it's different oxidation states?  Is there a formula that doesn't involve a lot of hand waving?  I have a customer who wants to do this for his garnet analyses.

[JohnF]

Here is a paper which has a complete recalculation of garnet formula from EPMA data:

A. Locock, Computers & Geosciences, 2008, v 34, p 1734-1780: "An Excel spreadsheet to recast analyses of garnet into end-member components, and a synopsis of the crystal chemistry of natural silicate garnets."

If you can't locate a copy, let me know and I'll send you the pdf (with link to the excel spreadsheet).

John F

[jeb]

Great, this looks like it should help.  I found it no problem and forwarded it on, thanks!

[jeb]

For anyone else looking for this:
http://www.sciencedirect.com/science/article/pii/S0098300408001064

PDF is also posted here, but the Excel file is too large to post.

[Jeremy Wykes]

If you would like something rigorous I suggest:

Dollase W. A. and Newman W. I. (1984) Statistically most probable stoichiometric formulae. American Mineralogist 69, 553

[Probeman]

Another method for FeO/Fe2O3 determination is the so called "flank method" (see attached below), but it is rather complicated and depends on a suitable set of calibration standards. 

I have not tried it myself but I'd be interested to hear if others have tried it and on what materials.

[Frank Spear]

There is also a section in my Phase Equilibria book (pages 101-103) on simultaneous cation and charge normalizations (the basis for this method). And I have a paper (Spear & Kimball 1984 Computers in Geoscience) on amphibole recalculation schemes (I think John Schumacher also has a publication on this).

But be wary - errors propagate very substantially using the stoichiometric method because the result (Fe2+/Fe3+) incorporates the errors in the measurement of all other cations. If one element is systematically off a bit this can lead to rather erroneous Fe2+/Fe3+ values.

I have also tried to measure Fe2+/Fe3+ in sodic amphiboles using a modification of the flank method described above with limited success. I don't think it is worth the effort at this point. We simply need a better way to make the measurement directly.

[Les Moore]

What ever happened about the postulated Fe2+/Fe3+ state being determined by the ratio of the L lines?

I know there was an issue with self fluorescence but has it ever been positively dismissed as an analytical technique?

XANES can do it nicely but who has a synchrotron at hand.

You can get a good standard by oxidising a pure iron standard.  Mostly FeO then Fe3O4 then a thin layer of Fe2O3.  Trouble is, the FeO varies from one side to the other of the wustite phase.  If there is Mn in the steel, you get a fascinating pile up rejection in the wustite at the magnetite boundary as it flees the higher oxidation states.

[Jeremy Wykes]

The group at Frankfurt¹ does quite a bit of Flank method Fe²⁺/Fe³⁺ work on mantle minerals with the electron probe.

It turns out that XANES is not as straightforward as you would expect for Fe²⁺/Fe³⁺ in minerals like garnet. A Moessbauer/XANES cross calibration for peridotitic garnet does not work for eclogitic garnet, as the differences in major element chemistry of the garnets are sufficient to screw everything up, because the Fe XANES signal is the result of both chemical/structural environment and valence state.

Synchrotron moessbauer^2,3 is the way to go for Fe²⁺/Fe³⁺ in materials with natural ⁵⁷Fe abundances, but who has a spare ⁵⁷Fe borate crystal handy in addition to a synchrotron?

¹ http://www.mineralogie.uni-frankfurt.de/petrologie-geochemie/mitarbeiter/hoefer/publikationen/

² http://scripts.iucr.org/cgi-bin/paper_yard?vv5038

³ https://epub.uni-bayreuth.de/172/1/DissVP.pdf