Moonstone from the Moon: Coordinated Analysis of Exsolved Super-Ternary Feldspars in Lunar Felsites

Although the Moon is generally thought of as a basaltic body, the first Apollo-era studies found that the Moon was also capable of forming granitic rocks as evidenced by small but significant numbers of felsite clasts in lunar breccias. Many of the alkali feldspar compositions in these rocks are notable, and unusual, in having An + Ab components that pushed grain compositions well into the feldspar ternary miscibility gap. In terrestrial rocks, such compositions typically undergo exsolution during sub-solidus cooling to form cryptoperthites visible by transmission electron microscopy (TEM), or coarser perthites visible under the petrographic microscope. Although Apollo-era work only hinted at the presence of exsolution in the lunar feldspars, we have now used coordinated optical, SEM-EBSD and TEM observations to confirm a diversity of exsolution microstructures in the lunar ternary feldspars on the SEM and/or TEM scale. The most unusual of these are compositions in sample 72255 that lie near in the center of the An – Or binary join, but which are nevertheless stoichiometric, homogeneous single crystals based on EPMA, SEM and EBSD analyses. TEM observations further confirm these near-binary K-Ca feldspars are single crystals, but with nanoscale compositional modulations consistent with having undergone incipient exsolution by spinodal decomposition during subsolidus cooling. Ternary feldspars with zoned compositions in samples 15405 and 12013 were also found to have developed exsolution microstructures on the TEM, SEM and even optical scale, typically in whichever zoned portion (core or rim) plotted most towards the center of the ternary miscibility gap. The sub-micrometer lamellar microstructures in our lunar samples are typical of what TEM studies have found in terrestrial non-ternary alkali feldspar cryptoperthites from volcanic rhyolites and granitic shallow intrusives. However, both microstructural coarsening and compositional equilibration in the ternary lunar feldspars can be expected to be much slower because of the need to segregate the slowly diffusing Al-Si cations along with Na, K and Ca. As a result, the felsites hosting the exsolved lunar ternary feldspars are more likely to have been intrusive, with a subsolidus cooling history on the 102-103 year timescale. If exsolution occurred subsequent to clast formation and cooling in the host impact breccia, the same type of 102-103 cooling timescale would have required emplacement in a km-scale impact melt sheet, or similar sized ejecta blanket.