Posts Tagged ‘plagioclase’

Today’s mineral is a common sight to igneous and metamorphic petrologists, though how we ID plagioclase (or plag for short) in a gabbro vs. a garnet schist will vary quite a bit.

Plag varies from Na-rich (albite or Ab-rich) to Ca-rich (anorthite or An-rich) with complete solid solution between the two end-members.   The difference in composition is reflected in a variation between low refractive indices (Ab-rich) and “higher” refractive indices (An-rich).


Nesse (2009); Troger marks the diagram with quartz (~1.55), epoxy (~1.54), and orthoclase (~1.525) refractive indices for comparison

In PPL, plag is almost a dead-ringer property-wise for the K-feldspars, quartz & cordierite.   Plag is usually clear with low positive relief, though Ab-rich plag will have a low negative relief (see figure above).   Similar to the K-feldspars, plagioclase is more likely to alter to a clay (in this case a white mica called sericite), which you may be able to see in plain light as little dots or imperfections.   Though plag also has two good directions of cleavage like K-feldspar, because of the low relief we rarely (if ever) can distinguish the 90 degree planes.   In mafic igneous rocks, plagioclase is more likely to be lath-like than the other other clear, low-relief minerals, but tends towards more equant grains in felsic igneous & metamorphic rocks.


Clear mineral is plagioclase, brown is an orthopyroxene (opx); norite (yes, I renamed according to the IUGS classification!) from http://www.geolab.unc.edu/Petunia/IgMetAtlas/plutonic-micro%7F/hypgabbro.UX.html

In XPL, we can have more hope of distinguishing between plag and the other minerals–at least in igneous rocks.   Just like twinning is commonly found in igneous K-feldspars, plag also frequently is forms twins during growth from a magma.   In plag’s case, the twins are called “polysynthetic” and kind of look like jail bars.   Plag can also have simple twins (usually in very An-rich grains), but the polysynthetic ones are more common.

same norite as above, but in XPL; plagioclase is 1st order greys/whites with polysynthetic twins while the opx is 2nd-3rd order colors (TS is too thick): http://www.mpch-mainz.mpg.de/~jesnow/Ozeanboden/2001/Lecture3/iugs-gabbro-640.gif

same norite as above, but in XPL; plagioclase is 1st order greys/whites with polysynthetic twins while the opx is 2nd order colors (TS is too thick): http://www.mpch-mainz.mpg.de/~jesnow/Ozeanboden/2001/Lecture3/iugs-gabbro-640.gif


olivine (2nd - 3rd), clinopyroxene (cpx; 2nd-3rd - yes, at this scale we can't tell the diff), and plag laths (1st greys/whites with polysynthetic twins) in a diabase: http://www.geolab.unc.edu/Petunia/IgMetAtlas/plutonic-micro%7F/diabase.X.html

Plag has complete solid solution between the An & Ab end-members, but the diffusion between Na1+ & Ca2+ is fairly slow within a single crystal because the charge imbalance requires coupled substitution.   As plagioclase grows out a melt, the melt’s composition will change and become more enriched in whatever ISN’T being used by the plag (and anything else that is growing).   An crystallizes at higher temps, so the melt will be first depleted in Ca.   The next plag that grows won’t have the same ratio of Ca/Na to incorporate, so the next layer that grows will be more of a Ca-Na plag.   As this continues, each additional layer will be more & more Na-rich assuming that no Ca is magically injected back into the magma chamber.

We can actually see this in thin section, since each type of composition of plagioclase has a slightly different set of crystallographic axes (= varying refractive indices).   Each orientation of the specific plag compositions will have a slightly different extinction angle (almost all of which are non-parallel)–we look for the maximum EA possible in the thin section.

On the practical side, this means if the plag is zoned, we’ll be able to see the variation from core to rim.   We can also estimate what the composition of the plag is without trying to find an optic axis figure (which would also work if we note the change in 2V on the first figure way up at the top of this post), so that usually saves us some time.

zoned plagioclase crystal that was later fractured: http://www.geolab.unc.edu/Petunia/IgMetAtlas/minerals/plagioclase.X.html

On some occasions, the melt will manage to get back into the middle of the crystal.   That melt & the interior portion of the crystal are not thermodynamically at equilibrium with each other (which usually the rim & the melt are), so the interior starts to melt, while the rim remains untouched.   This can result in a “sieve” texture:

Ok, this works well for igneous plag, but there are some issues when we move over into metamorphic rocks.   The process of metamorphism usually either erases what zoning or twinning was present & plag that grows due to solid-state reactions doesn’t form twins, so plagioclase most common looks just like quartz, K-feldspar, or cordierite.   Luckily, cordierite is only present in low-pressure, high-temperature rocks & forms pleochroic haloes around U/Th bearing minerals, so we can rule that out quickly.   K-feldspar may occur in metamorphosed felsic igneous rocks, but is rare in metasedimentary rocks until high temperatures, so usually we don’t have to worry about it either.   However, quartz is almost everywhere.   And differentiating between quartz & plag in metamorphic rocks is probably one of the most annoying things to do with a PLM.

garnet (isotropic) schist with muscovite (2nd-3rd), quartz, plagioclase and biotite (2nd-3rd)--but which grains are qtz? which are plag?: http://www.geolab.unc.edu/Petunia/IgMetAtlas/meta-micro/porphyroblastic.X.html

There is some hope:

  • if the rock has been deformed, quartz may either have undulose extinction or subgrains — rare (if ever) in plag
  • if the thin section is slightly too thick, quartz has a tendency to become 1st order yellows instead of greys / whites (only very, very Ca-rich plag will become yellow)
  • quartz is uniaxial, plagioclase is biaxial–don’t hang your hat on this one, since we purposely cut metamorphic thin sections parallel to lineations & perpendicular to orientations, which makes finding an optic axis ridiculously difficult (if you have perp to both foliation & lineation, it works fine)
  • plagioclase may have deformation twins if 1) the rock has been deformed and 2) hasn’t been heated to >~400 C since the deformation; deformation twins aren’t as straight, tend to taper, and may only be in part of the grains in contrast to the polysynthetic growth twins in igneous rocks

Anorthosite (Ca-plagioclase) rock from Ceilidh Hill, central Australia; plagioclase contains deformation twins that bend, taper out, and terminate at cross-cutting twin faces: http://www.geo.umn.edu/orgs/struct/microstructure/images/047.html

  • if you’re lucky, the plagioclase has enough of a shift in composition between the core & rim that you’ll be able to see core-rim extinction differences–this is easy to mix up with undulose extinction in quartz unless you’ve been trained to look for it–unfortunately, I don’t have a digital picture at this point, though my master’s rocks from the Bronson Hill were filled with them; if I find time to scan, I’ll add one at a later time
  • at the end of the day, usually we end up taking our thin section and finding the closest SEM, because they are ridiculously easy to differentiate under an electron beam–this is part of that last 5% of nailing down what exactly is in the rock via more expensive methodology than the PLM

Next week, we’re returning to a mineral that most of you won’t recognize in thin section (which probably means Chris Rowan will be happy!): pyrite.

Read Full Post »