In this light, if one has access to a microprobe with elemental mapping capabilities, a Ti map of a polished surface at appropriate scale would produce an image that could be used with image-processing software to get a very accurate estimate of areal proportion of rutile needles (assuming random and not preferred orientation of such needles). If I were doing this, I'd probably map several areas to test for consistency. But as Eric suggests, high-resolution BSE micrographs could also be used with image-processing to do the same thing, although the contrast between rutile and garnet would be more clearcut in an elemental map.
With regard to Eric's excellent point about "exsolving" an RO2 phase from an R2O3 one, there may be cases where "paired" exsolution may occur. We worked some years ago on Opx megacrysts from the Adirondacks which had apparent exsolution lamellae (they were definitely oriented lamellar intergrowths) of olivine, calcic plagioclase and ilmenite. The ilmenite was not a problem from the point of view Eric noted (R2O3 in R2O3), but olivine and plag were. However, an equi-molar mixture of anorthite and forsterite (modeling plag and olivine as pure end-members) would produce a hypothetical model composition of (CaAlMg2)(AlSi3)O12 which could be viewed as having a pyroxene-like stoichiometry. Not a real happy-looking stoichiometry, but if these megacrysts represented very rapid high-T crystallization (a Ron Emslie-like mechanism), unusual molecular constituents might have been included in initial magmatic crystallization.
Bob T.
Eric Essene wrote:
[log in to unmask]" type="cite">Alexander,
The term exsolution should not apply unless the structures of the host and newly formed phases are similar structures, e.g., feldspars, pyroxenes, amphiboles, carbonates. Some people use the term exsolution rather carelessly however. The term precipitate comes from metallurgy.
It is rather difficult to get rutile alone to form from garnet by a closed system process because rutile is an RO2 phase and garnet is R2O3. A small amount of reacted TiO2 could be accommodated by simultaneous oxidation of Fe in the garnet. Pate precipitation should also involve a halo around the needle, as has been shown by Dave Wark for rutile needles in some quartz. The amount one sees of needle rutile in transmitted light gives a mistakenly large overestimate of how much is really there because it includes needles from depth inside the host into what seems to be just two dimensions. This mistake has been made repeatedly by authors describing quartz in eclogitic Cpx, ilmenite in olivine, etc. The areal percent in reflected light or BSE images is a much better estimate. Spotting the needles in reflected light is OK once you learn how to identify the often 1-5 micron wide cross-sections of needles. They can be oriented on tetrahedral edges inside the garnet. So could they be epitaxial or topotaxial in origin, or do they represent a reaction upon cooling? Both mechanisms involve preferred orientation of the precipitate on the host.
cheers,
eric
On Sep 26, 2008, at 4:45 AM, proyer wrote
:
Hello everybody,
hope some of you can help me out on two questions related to each other:
1) Is there a source which clearly defines when to call a mineral phase unmixing from a solid solution "exsolution" or "precipitate"?
2) Do you know of any article expaining explicitly (reaction equations etc.) how rutile unmixes from high-tempertaure garnet?
I know of three such papers for ultrahigh-pressure garnets (Van Roermund, 2000; Zhang et all. 2003 and Yang & Liu, 2004),
but none for granulite facies rocks etc.
Please let me know if you have come across or even authored any such references.
Sincerely,
Alexander Proyer
University of Graz, Austria
--
Dr. Robert J. Tracy
Professor of Geosciences
Department of Geosciences
5064 Derring Hall, Virginia Tech
Blacksburg VA 24061-0420
Office Phone: 540-231-5980
Fax: 540-231-3386