Melt Migration in the Early Lunar Crust: Formation of the Primitive, Pure Lunar Ferroan

Introduction: Over the years, there have been many proposed mechanisms for moon formation, including fission, capture, and coaccretion, but the most popular and ever-evolving model at present is the giant impact origin [1]. The impact origin points to the existence of a hot, young moon, likely with a global magma ocean [2]. The next logical step, based on the mantle cumulate crystallization sequence [3], would be a resultant flotation crust made of plagioclase. Early calculations of the expected anorthosite content of the Moon [4] did not match initial measurements of Apollo samples [5], and more recently have not matched Clementine measurements [6] or SELENE measurements [7] because they assumed interstitial melt would freeze. By considering a physical model of a magma ocean with an accumulating flotation lid, we find that significant escape of melt can take place for reasonable physical parameters and timescales of melt migration, thus allowing for more nearly pure lunar anorthosites, consistent with observations. Compositions of Ferroan Anorthosites: Measured plagioclase contents of lunar anorthosites. Generally, anorthosites are 80-100% plagioclase feldspar with the remaining interstitial volume consisting of mafic minerals like pyroxene, ilmenite, magnetite, and olivine. Plagioclase is a type of feldspar ranging from albite (100% NaAlSi2O8) to anorthite (100% CaAlSi2O8). Unique properties of lunar ferroan anorthosites include high Al2O3, high CaO, low Mg number, low levels of incompatible elements, a large Eu anomaly, and rare earth element abundances less than chondritic [8]. The bulk of the Apollo anorthosite samples contain up to 95% plagioclase by volume [5]. Recent data taken by SELENE (Selenological and Engineering Explorer) on the Moon Mineralogy Mapper suggest that plagioclase percentages of crustal anorthosite may be as high as 98%, if not 100% [7]. The lunar magma ocean model would predict that the anorthosites were formed early as a flotation crust with later intrusions by the rocks known as the Mg-suite. Mg-suite rocks are characterized by their high Mg/Fe ratios for their mafic components and are usually found in the lunar highlands. This hypothesis is thrown into question with the dating of lunar anorthosites and Mg-suite rocks; Sm-Nd dating predicts that lunar anorthosites are between 4.29 and 4.56 Ga, the youngest of which overlap the oldest of the Mg-suite rocks in time [9]. Initial calculations for lunar anorthosites. [4] carried out a mass-balance calculation of plagioclase and the interstitial melt remaining the the lunar crust. Assuming the density of the moon’s flotation crust was equivalent to that of the underlying magma ocean, then using accepted values for the densities of plagioclase, mafic mineral, and melt, the volume of plagioclase in the lunar crust should be at least 81%. See figure 1 for a summary of these measurements and calculations. Proposed formation models for lunar ferroan anorthosites. There exist many hypotheses for the formation of the lunar ferroan anorthosites (LFA). The high plagioclase content may suggest that some secondary process occurred which enriched primitive, perhaps less pure, anorthosite. [12] suggested that the lunar ferroan anorthosites are consistent with the postcumulus removal of pyroxene through multiple melting episodes. [13] suggested that diapirism may enrich the LFA in anorthosite, eliminating the need altogether for the primary LFA to be exceptionally pure. During this process, pyroxene-rich extract would sink out of the anorthosite, and maybe could even by lost by the crust to return to the active magma ocean. This pluton model predicts anorthosite with 85% plagioclase by volume. [14] proposes a formation process via adcumulus growth at the base of the lunar crust. We take this idea as the starting point of our model. Methods and Calculations: We propose to revisit the idea of a primary mechanism for the formation of the lunar anorthosites to explain the disparity between measurements and predictions of their purities. We apply the concepts of Darcy flow and thermal diffusion to a region of early lunar flotation crust, squeezing mafic melt out and leaving behind the buoyant plagioclase crystals to freeze in place. The process of the formation of the lunar anortho-