Abstract

Given the available data, we find that the wide range of mare basaltic material characteristics can be explained by a model in which: (1) The mare basalt magma source region lies between the crustmantle boundary and a maximum depth of 200 km and consists of a relatively uniform peridotite containing 73–80% olivine, 11–14% pyroxene, 4–8% plagioclase, 0.2–9% ilmenite and 1–1.5% chromite. (2) The source region consists of two or more density‐graded rhythmic bands, whose compositions grade from that of the very low TiO2 magma source regions (0.2% ilmenite) to that of the very high TiO2 magma source regions (9% ilmenite). These density‐graded bands are proposed to have formed as co‐crystallizing olivine, pyroxene, plagioclase, ilmenite, and chromite settled out of a convecting magma (which was also parental to the crust) in which these crystals were suspended. Since the settling rates of the different minerals were governed by Stoke's law, the heavier minerals settled out more rapidly and therefore earlier than the lighter minerals. Thus the crystal assemblages deposited nearest the descending side of each convection cell were enriched in heavy ilmenite and chromite with respect to lighter olivine and pyroxene and very much lighter plagioclase. The reverse being the case for those units deposited near the ascending sides of the convection cells. Simultaneously with this density controlled settling of the crystals, the heating of the magma at the bottom of the convection cells resulted in the partial remelting of significant amounts of suspended, but settling plagioclase, ilmenite, and pyroxene. This partial remelting process was also partially responsible for the decrease in the ilmenite content (from 9 to 0.2%) between the very high TiO2 and the very low TiO2 source regions and caused the plagioclases and pyroxenes in the former source regions to be more sodic and clinopyroxene richer, respectively, than in the latter. (3) During the mare basalt epoch, radiogenic induced remelting of various parts of the density‐graded bands led to the formation of the initial mare basalt magmas. The viscosity of melting systems, which decreases rapidly at about 30±5% partial melting, appears to have limited the degree of partial melting of the source region to about 30±2%. (4) These ∼30% partial melts rose to the crust‐mantle boundary where they pooled in magma storage chambers. The magmas remained in these chambers for different lengths of time, cooled to different degrees, and lost 0 to ∼30% olivine or ∼30% olivine plus 0 to ∼20% pyroxene by fractional crystallization. Those magmas which remained in the chambers for very short times and lost no, or essentially no olivine before eruption were the parental magmas of the pyroclastic glass units. Those magmas which lost increasing amounts of olivine ± pyroxene were the parental magmas of the Apollo 12 and 15 magmas, the Apollo 11 and 17 magmas, and the Luna 16 and 24 magmas, respectively. This fractional crystallization phase in the genesis of the magmas explains both the pattern of the mare basalt magmas in the pseudo‐quaternary phase diagrams and the decrease in the siderophile contents of the magmas as a function of their degree of fractionation. (5) The magma storage chambers, being located at the crust‐mantle boundary, were located in the zone where urKREEP formed during the initial differentiation of the moon. Due to the low melting point of KREEP compared with the initially high temperatures of the magmas, residuals of the urKREEP layer were assimilated by the cooling magmas. Since the urKREEP materials underwent varying degrees of fractional melting early in lunar history, these urKREEP residuals had varying degrees of light REE depletion. Thus the assimilation of up to a few percent of these residuals caused the magmas to have differing Eu and light REE depletion patterns, as well as a wide range in their REE abundances. (6) The fractional crystallization and urKREEP residual assimilation phase of mare basalt magma genesis ended with the eruption of the magmas onto the lunar surface.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call