Early differentiation of the Moon: Evidence from trace elements in plagioclase

Abstract

Bulk samples and plagioclase mineral grains from cataclastic ferroan anorthosites (mainly from North Ray crater, Apollo 16) were analysed for trace elements by instrumental neutron activation analysis (INAA). Some pure plagioclase crystals have lower concentrations of Cr, Co, Sc, and heavy rare earth elements (REE) than previous analyses indicated. Concentrations of Sr, Eu, Na, Ga, and the REE in plagioclases from different anorthosites vary in a systematic way. Plagioclases with the lowest contents of these elements could have crystallized from a magma with chondritic relative abundances of Ca, Sr, Eu, and the REE and a tenfold enrichment relative to type 1 carbonaceous chondrites. Plagioclases with higher trace element concentrations crystallized from a more evolved magma. The trace element patterns of plagioclases from pristine norites and troctolites are very different from the plagioclase pattern of anorthosites. Besides having higher concentrations of incompatible elements, noritic, and probably also troctolitic plagioclases have much higher contents of Mg, Cr, and Co. This simultaneous enrichment in incompatible elements and Cr and Co can be explained by the mixing of a magma rich in incompatible elements (KREEP‐rich) with a Mg‐, Cr‐, and Co‐rich component supporting previous arguments made on bulk rock chemical grounds. The mafic component could either be late infalling material of bulk moon composition or may be derived from the large mafic layer that must exist in the interior of the moon. Mixing of KREEP and the mafic material probably occurred in the interior of the moon, because norites and troctolites could have formed from a magma of this composition. A similar mixture of KREEP and Mg‐rich material was distributed over the front side of the moon 3.9 to 4 b.y. ago, by large impacts and/or by volcanic eruptions. Anorthosites have Mg/Cr, Fe/Sc, and Co/Fe+Mg ratios different from all other highland rocks, indicating either that average highland rocks are not representative of the bulk moon system or that these ratios change during fractional crystallization of a magma of bulk moon composition.