Anorthosite assimilation and the origin of the Mg/Fe‐related bimodality of pristine moon rocks: Support for the magmasphere hypothesis

Abstract

The lunar magma ocean (or magmasphere) hypothesis, always controversial, has required considerable revision in recent years. The geochemical bimodality of pristine rocks, epitomized by a diagram of Na/(Na + Ca) versus mg′ ratio, has led to proposals that a major fraction of the crust (the Mg‐rich suite) formed as cumulates in numerous intrusions slightly younger than the magmasphere. The precise origin of the bimodality has been elusive, however. One previous suggestion was that Na was lost as a volatile before the ferroan anorthosites crystallized from the magmasphere, whereas the Mg‐rich intrusives retained all of their original Na. However, this model cannot fully explain the bimodality, because the same bimodal pattern is manifested by plotting Eu/Al (a ratio of involatile elements) versus mg′ ratio. Assimilation probably helped to engender the bimodal patterns. Mass/energy balance calculations indicate that large proportions of plagioclase were probably assimilated from the older (magmasphere‐generated) ferroan anorthosite crust by most of the Mg‐rich intrusive melts. The magmasphere, in the absence of assimilation, probably did not yield appreciable plagioclase (the ferroan anorthosite crust) until fractional crystallization of mafic silicates had diminished the melt mg′ ratio to about 0.42. However, assuming identical initial melt composition, an Mg‐rich intrusion assimilating ferroan anorthosite (and perhaps also a small proportion of urKREEP) would have reached plagioclase saturation at much higher mg′, about 0.66. The current version of the magmasphere hypothesis (ferroan anorthosites = magmasphere flotation cumulates; Mg‐rich rocks = products of slightly younger, localized intrusions) appears to be the only plausible mechanism for engendering the Mg/Fe‐related bimodality.