Re: Amphiboles normalization

[Brian Joy]

Remember, for in depth evaluation of your amphibole mineralogy, one should instead utilize the Locock spreadsheet output format (seen in the previous post) and utilize Andrew's recalculation spreadsheets externally to PFE.

I don’t mean to be critical of Andrew, but keep in mind that not all of us agree that the Hawthorne et al. (2012, Am Min 97: 2031-2048) amphibole classification was a useful replacement for the Leake et al. (1997, Can Min 35: 219-246) classification.  The attached comment on it is a little convoluted, but Bernard and I had some important points to make.  I don’t think it’s a good idea to have Hawthorne et al. entrenched in PFE.

[John Donovan]

Remember, for in depth evaluation of your amphibole mineralogy, one should instead utilize the Locock spreadsheet output format (seen in the previous post) and utilize Andrew's recalculation spreadsheets externally to PFE.

I don’t mean to be critical of Andrew, but keep in mind that not all of us agree that the Hawthorne et al. (2012, Am Min 97: 2031-2048) amphibole classification was a useful replacement for the Leake et al. (1997, Can Min 35: 219-246) classification.  The attached comment on it is a little convoluted, but Bernard and I had some important points to make.  I don’t think it’s a good idea to have Hawthorne et al. entrenched in PFE.

Science is a moving target!   

To be clear, PFE only uses Andrew's calculation method to estimate ferrous/ferric ratios for amphiboles.  As far as classification of amphiboles goes, that aspect is left to those who utilize Andrew's spreadsheet externally to PFE.

I'll let you mineralogists sort it out!   

[Brian Joy]

Remember, for in depth evaluation of your amphibole mineralogy, one should instead utilize the Locock spreadsheet output format (seen in the previous post) and utilize Andrew's recalculation spreadsheets externally to PFE.

I don’t mean to be critical of Andrew, but keep in mind that not all of us agree that the Hawthorne et al. (2012, Am Min 97: 2031-2048) amphibole classification was a useful replacement for the Leake et al. (1997, Can Min 35: 219-246) classification.  The attached comment on it is a little convoluted, but Bernard and I had some important points to make.  I don’t think it’s a good idea to have Hawthorne et al. entrenched in PFE.

Science is a moving target!   

To be clear, PFE only uses Andrew's calculation method to estimate ferrous/ferric ratios for amphiboles.  As far as classification of amphiboles goes, that aspect is left to those who utilize Andrew's spreadsheet externally to PFE.

I'll let you mineralogists sort it out!   

Exactly what normalization scheme is used for each amphibole type listed?  For instance, for calcic amphiboles, is it assumed that the sum of cations per 23 anhydrous oxygens is equal to 13 + nCa²⁺ + nNa⁺ + nK⁺?  This requires some explanation, as not all normalization schemes are stoichiometrically allowable for a given analysis.  Further, in most cases, more than one normalization scheme is permissible, and this creates a range of “valid” Fe₂O₃ contents.  In truth, recalculated amphibole Fe₂O₃ content is a “moving target,” and the chosen value often is inconsistent with results from wet chemistry.  See, for instance, the attached paper by Cosca et al.  Further, assessment of the “correct” normalization must be guided petrologically.  For instance, if a given normalization places most Na⁺ on the M4 site rather than on the A site, then one might need to be able to provide evidence from the mineral assemblage that the amphibole crystallized at relatively high pressure.  Lastly, how is the normalization of “oxo-amphiboles” treated? 

The classification of Leake et al. is a petrologist’s classification, while that of Hawthorne et al. was created by mineralogists.  Hawthorne et al. undid precedents in amphibole classification that had been in place since 1968 and removed the notion that phase equilibria are an important consideration in amphibole nomenclature.  Aside from the fact that this new nomenclature requires all amphiboles to be reclassified (i.e., names that you see in the literature are no longer valid), it also results in silly situations such as former anthophyllite-gedrite pairs from a classic locality at Telemark now being classified as anthophyllite-anthophyllite pairs and ferrogedrite from the “Route 5 locality” in eastern Vermont now being nameless.

[John Donovan]

Attached is Andrew's amphibole normalization spreadsheet which is being distributed with the latest PFE version. 

We're more than happy to distribute alternative amphibole normalization spreadsheets if anyone is willing to provide one.

Who would have thought that amphibole mineralogy could get so exciting...   

[sem-geologist]

The classification of Leake et al. is a petrologist’s classification, while that of Hawthorne et al. was created by mineralogists.  Hawthorne et al. undid precedents in amphibole classification that had been in place since 1968 and removed the notion that phase equilibria are an important consideration in amphibole nomenclature.  Aside from the fact that this new nomenclature requires all amphiboles to be reclassified (i.e., names that you see in the literature are no longer valid), it also results in silly situations such as former anthophyllite-gedrite pairs from a classic locality at Telemark now being classified as anthophyllite-anthophyllite pairs and ferrogedrite from the “Route 5 locality” in eastern Vermont now being nameless.

Brian,
Thank You for these details, I had a gut feeling that Leake et al. classification is more sensible than newer Hawthorne et al. but had no explanation, hitherto (in Geology field I am more petrologist than mineralogist, this probably why Leake et al. classification looks for me making more sense).
Now talking about EPMA and Fe3+/Fe2+ estimation in amphiboles I think these approaches are worth looking:
* classical Fe Lα and Lβ flank method - https://doi.org/10.1016/j.chemgeo.2019.01.009
* classification based on principal component regression (PCR, machine learning) https://doi.org/10.1016/j.lithos.2020.105469

(Oh damn, I can't access those papers, I should probably look elsewhere (what a pun! - it sounds a bit like Elsevier  )

[Probeman]

The ferrous/ferric methods being discussed here are software methods based on site occupancy and charge balance.

For discussion about empirical methods for ferrous/ferric determinations see this topic here:

https://probesoftware.com/smf/index.php?topic=45.0

[AndrewLocock]

For many in our field of science, the nomenclature of amphiboles is viewed as the responsibility of the Subcommittee on Amphiboles of the CNMNC. However, as will be apparent below, the nomenclature is a set of recommended guidelines, not edicts.

Some background:
The Commission on New Minerals, Nomenclature and Classification (CNMNC) of the International Mineralogical Association is the successor to the Commission on New Minerals and Mineral Names, and the Commission on Classification of Minerals. The roles of the Commission and of the preceding CNMMN, are well described by Nickel & Grice (1998), and Hatert et al. (2017).

From the 1998 publication of Nickel and Grice (and reproduced on the CNMNC website at http://cnmnc.main.jp/):
The CNMNC “does not wish to impose an arbitrary set of rigid rules on the mineralogical community, but rather to provide a set of coherent guidelines that provide a reasonably consistent approach to the introduction of new minerals and the application of mineral nomenclature.”

From the article of Hatert et al. (2017):
“As in all scientific publications, it may appear that some … do not agree with our decisions”.
The CNMNC consists of “volunteers to ensure a consistent mineral nomenclature, in order to facilitate the progress of mineral science.”

Thus, one is not forced to agree with, or use, any particular decision of the CNMNC.
Some journals require rigorous adherence to IMA-approved nomenclature, whereas other do not (discussed in the editorial of Martin 2002).

The output formats recently added to the Probe for EPMA software (I hasten to add: not at my urging) are options for the convenience of the interested user. There are several other possible ways of exporting data from Probe for EPMA, and of looking at either garnet or amphibole compositions.


References:

Hatert, F., Pasero, M., Mills, S.J. and Hålenius, U., 2017. How to define, redefine or discredit a mineral species? Elements 13, 208.

Martin, R.F., 2002. Pondering nomenclature of minerals: a plea for consistency. Newsletter of the Mineralogical Association of Canada 66, 3.

Nickel, E.H. and Grice, J.D., 1998. The IMA Commission on New Minerals and Mineral Names: procedures and guidelines on mineral nomenclature, 1998. Canadian Mineralogist 36, 913-926.
Published also in: Mineralogical Record 30, 163-176. Mineralogy and Petrology 64, 237-263. Brazilian Journal of Geology 28, 229-242. Boletín de la Sociedad Española de Mineralogía 21, 177-202. Zapiski Vserossiiskogo Mineralogicheskogo Obshchestva 127, 51-65.

[Brian Joy]

The ferrous/ferric methods being discussed here are software methods based on site occupancy and charge balance.

For discussion about empirical methods for ferrous/ferric determinations see this topic here:

https://probesoftware.com/smf/index.php?topic=45.0

Like I noted above, treatment of Fe₂O₃ recalculation in the amphiboles is not as straightforward as it is, say, for a spinel or garnet.  Various recalculation schemes may be applied, and, in most cases, only minimum and maximum Fe₂O₃ contents can be determined.  The amphibole recalculation options now present in the Analyze! Window require additional documentation.  If a specific radio button is selected, what exactly is the normalization procedure used?  Anyone who engages in amphibole recalculation needs to be aware of the underlying algebra, how it affects the distribution of cations on the various sites (A, M4, M13, M2, T), and what this distribution might imply about conditions under which the amphibole grew.

Further, any recalculation based on charge balance at least requires propagation of counting error in each measured element.  Uncertainty in estimated Fe₂O₃ content in many silicates tends to be dominated by error in SiO₂ and Al₂O₃ not just due to their concentrations, but also due to the relatively high charges of Si⁴⁺ and Al³⁺.

[Brian Joy]

For many in our field of science, the nomenclature of amphiboles is viewed as the responsibility of the Subcommittee on Amphiboles of the CNMNC. However, as will be apparent below, the nomenclature is a set of recommended guidelines, not edicts.

Hi Andrew,

The CNMNC certainly appeared to issue an edict to Bernard and me in their attached discussion.  In fact, they were quite patronizing.

Brian

[Brian Joy]

Now talking about EPMA and Fe3+/Fe2+ estimation in amphiboles I think these approaches are worth looking:
* classical Fe Lα and Lβ flank method - https://doi.org/10.1016/j.chemgeo.2019.01.009

The flank method will likely never be applied successfully to the amphiboles.  The compositional complexity is too great, and dehydrogenation under the beam at the large current required can result in oxidation of Fe²⁺.

[John Donovan]

The amphibole recalculation options now present in the Analyze! Window require additional documentation.  If a specific radio button is selected, what exactly is the normalization procedure used?  Anyone who engages in amphibole recalculation needs to be aware of the underlying algebra, how it affects the distribution of cations on the various sites (A, M4, M2, M13, T), and what this distribution might imply about conditions under which the amphibole grew.

I completely agree.  That is why I previously posted Andrew's Excel spreadsheet which is attached to this post here (which is what our PFE code is based on):

https://probesoftware.com/smf/index.php?topic=1460.msg10857#msg10857

And I also posted the code which documents which flags in the amphibole calculation are set for each of the amphibole groups in the Calculation Options dialog here in this post, which is in the PFE latest updates topic:

https://probesoftware.com/smf/index.php?topic=40.msg10853#msg10853

I will add a "interactive" Help button in PFE (and CalcZAF) linked to this topic so people can refer to these items easily.

In addition, the complete source code for these amphibole normalization calculations in PFE (and CalcZAF) are found on Github's Open Microanalysis in the CalcZAF project is linked here:

https://github.com/openmicroanalysis/calczaf

See ConvertFerrousFerricRatioFromComposition3 procedure in the Convert.bas module file.

[Brian Joy]

Hi John,

Below is a screenshot from my own amphibole recalculation and classification spreadsheet, which is not really fit for public distribution.  Each line represents a single analysis.  I’ve organized it in the manner shown because I consider it essential to have a summary of normalizations available at a quick glance along with an indication of which ones are “bad” (i.e., not stoichiometrically allowable) and which ones provide the respective upper and lower bounds on recalculated Fe2O3.  Any of these recalculations can be applied to (but might not necessarily be appropriate for) calcic amphiboles, sodic amphiboles, ferromagnesian amphiboles, etc.  The normalization factor shown multiplies each cation value in the Fe2O3-free formula unit (allowable or not), while nFe3+ per formula unit is equal to the total anion charge multiplied by (1 - factor).

Andrew’s spreadsheet is very comprehensive (more so than mine), but 1) I can’t find summaries such as this anywhere in it, and 2) it leads one into using the Hawthorne et al. classification, which requires an estimate of Fe2O3/FeO in the assignment of root names.

I keep intending to modify my own matrix correction program in order to incorporate these amphibole recalculations within the iterations required to determine the matrix correction factors so that I can automate determination of bounds on Fe2O3 and also automate determination of wt% H2O (with appropriate caveats), but I simply haven’t had the time to do so; my workload is oppressive. 

Brian

[John Donovan]

I keep intending to modify my own matrix correction program in order to incorporate these amphibole recalculations within the iterations required to determine the matrix correction factors so that I can automate determination of bounds on Fe2O3 and also automate determination of wt% H2O (with appropriate caveats), but I simply haven’t had the time to do so; my workload is oppressive. 

I'm sure many mineralogists are looking forward to your amphibole recalculations.

As I mentioned previously, we've implemented Andrew's amphibole codes merely to perform the ferrous/ferric calculations as part of our ferrous/ferric options within the matrix corrections in Probe for EPMA (and CalcZAF). As you indicated it's important to iterate the matrix corrections when adding in oxygen from ferric iron.

Of course for most amphiboles, it's a relatively small effect, but for Fe-Ti oxides and magnetites the effect is surprisingly large. For hematites/magnetites it's a >1% change in the Fe concentrations:

https://probesoftware.com/smf/index.php?topic=92.msg8593#msg8593

For those specifically interested in the amphibole (or garnet) classification schemes we provide an export format for use with Andrew's Excel spreadsheets:

https://probesoftware.com/smf/index.php?topic=40.msg10846#msg10846

[AndrewLocock]

The calculation of the proportion of ferric iron in an amphibole composition is most thoroughly examined by Schumacher (1997) in the Appendix 2 of the Leake et al. (1997) report on amphibole nomenclature. Table A-2 of Schumacher (1997) uses a particular analysis from Appendix 1 of Deer et al. (1966, p. 515); the subtitle of this latter appendix is “Calculation of a chemical formula from a mineral analysis, hornblende analysis”.

This topic is reprised in Appendix III: Calculation of Fe³⁺ and (OH) in Amphiboles of the Hawthorne et al. (2012) report on amphibole nomenclature. Appendix Table 1 of Hawthorne et al. (2012) uses the same analysis from Deer et al. (1966, p. 515)

For the Excel spreadsheet to classify chemical analyses of amphiboles following the IMA 2012 recommendations that I authored, the topics of formula normalization and Fe³⁺ estimation are discussed in the paper published in Computers & Geosciences 62, 1–11.

To clarify: in that spreadsheet it is possible to force the ratio of Fe³⁺/ΣFe to 0 (all ferrous iron), or to 1 (all ferric iron), or to any particular value-of-interest, by entering the appropriate wt% FeO and/or Fe₂O₃ values and setting the option “use initial M³⁺/ΣM? (TRUE or 1/FALSE or 0)” to True or to 1. (Recall the conversion wt% FeO = 0.8998085 Fe₂O₃).

However, in general, the ratio of Fe³⁺/ΣFe is not known a priori, and thus some sort of formula normalization must be used, following Schumacher (1997). In the spreadsheet, 4 schemes of cation normalization are possible (modified from Schumacher 1997 to include known Li-content):
1. Sum of all cations from Si to K = 16 apfu.
2. Sum of cations from Si to Na = 15 apfu.
3. Sum of cations (includes Li) from Si to Ca = 15 apfu.
4. Sum of cations (includes Li) from Si to Mg = 13 apfu.
Methods of estimation of Fe³⁺ contents of amphibole are generally inaccurate in comparison to measured values, but are usually better than no estimate at all.

As stated in Computers & Geosciences 62, 1–11:
“How should an algorithm determine which schemes are most appropriate for a given analysis? Hawthorne et al. (2012) showed that the constraints on the amphibole formula arising from the various cation normalization schemes could be treated as criteria. As the criteria are not each satisfied by every amphibole endmember, and as real analyses are imperfect, there will usually be deviations from the criteria. In the spreadsheet, for each of the four normalization schemes, the maximum magnitude of the deviations of the formula proportions from the following criteria is determined: Si < 8 apfu; non-H cations < 16 apfu; sum Si to Ca (+Li) < 15 apfu; sum Si to Mg (+Li) > 13 apfu; sum Si to Na > 15 apfu. The normalization schemes with the smallest maximum deviations are used. To allow for the imperfection of real data, a threshold of 0.005 apfu is used for the deviations, and for the separation of the normalization schemes.”

The spreadsheet therefore automatically determines which normalization scheme or schemes are appropriate, based on the smallest maximum deviations from the criteria listed above.
However, the user can force the use of any, some, or all of the normalization schemes, by setting the following options to True or 1:
Require Si–Ca&Li<=15? (TRUE or 1/FALSE or 0)
Require Si–Mg&Li>=13? (TRUE or 1/FALSE or 0)
Require Si–Na>=15? (TRUE or 1/FALSE or 0)
Require Si–K<=16? (TRUE or 1/FALSE or 0)
If one is deeply interested in the details of this calculation, I recommend to look in the spreadsheet ( version 9.8 ) at the 3.Calculation worksheet, specifically rows 5495 to 5554.

Schumacher (1997) did list additional stoichiometric constraints for certain metamorphic amphiboles.
This is partially captured in Warnings section of the Output of the spreadsheet, specifically if the sum of high-valence C cations (M³⁺ and M⁴⁺) is greater than 2 apfu.

The calculation of the formula proportions of an amphibole does not depend on the subsequently determined name of that amphibole.

References:

Deer, W.A., Howie, R.A., Zussman, J., 1966. An Introduction to the Rock-Forming Minerals. Longman Group Limited, London, U.K.

Hawthorne, F.C., Oberti, R., Harlow, G.E., Maresch, W.V., Martin, R.F., Schumacher, J.C., Welch, M.D., 2012. IMA report, nomenclature of the amphibole supergroup. American Mineralogist 97, 2031–2048.

Leake, B.E., Woolley, A.R., Arps, C.E.S., Birch, W.D., Gilbert, M.C., Grice, J.D., Hawthorne, F.C., Kato, A., Kisch, H.J., Krivovichev, V.G., Linthout, K., Laird, J., Mandarino, J.A., Maresch, W.V., Nickel, E.H., Rock, N.M.S., Schumacher, J.C., Smith, D.C., Stephenson, N.C.N., Ungaretti, L., Whittaker, E.J.W., Guo, Y., 1997.
Nomenclature of amphiboles: report of the subcommittee on amphiboles of the International Mineralogical Association, Commission on New Minerals and Mineral Names. Canadian Mineralogist 35, 219–246.

Schumacher, J.C., 1997. Appendix 2. The estimation of the proportion of ferric iron in the electron-microprobe analysis of amphiboles. In: Leake, B.E., et al. (Eds.), Nomenclature of Amphiboles, vol. 35. Canadian Mineralogist, pp. 238–246.

[Brian Joy]

As stated in Computers & Geosciences 62, 1–11:
“How should an algorithm determine which schemes are most appropriate for a given analysis? Hawthorne et al. (2012) showed that the constraints on the amphibole formula arising from the various cation normalization schemes could be treated as criteria. As the criteria are not each satisfied by every amphibole endmember, and as real analyses are imperfect, there will usually be deviations from the criteria. In the spreadsheet, for each of the four normalization schemes, the maximum magnitude of the deviations of the formula proportions from the following criteria is determined: Si < 8 apfu; non-H cations < 16 apfu; sum Si to Ca (+Li) < 15 apfu; sum Si to Mg (+Li) > 13 apfu; sum Si to Na > 15 apfu. The normalization schemes with the smallest maximum deviations are used. To allow for the imperfection of real data, a threshold of 0.005 apfu is used for the deviations, and for the separation of the normalization schemes.”

The spreadsheet therefore automatically determines which normalization scheme or schemes are appropriate, based on the smallest maximum deviations from the criteria listed above.

I'd rather see the results of all normalizations and then choose for myself while applying some petrological guidance.  There is no simple means of extracting an accurate value of Fe2O3 from an electron microprobe analysis of an amphibole.  This has been my point from the beginning.

And what about propagated counting error?  In addition, systematic error in SiO2 (due to choice of standard or matrix corrections) will contribute additional significant uncertainty.