Variations In Major Element Compositions Research Articles

Special Papers-Apollo 11 Symposium: Petrogenesis of Apollo 11 Basalts and Implications for Lunar Origin

The origin of Apollo 11 basalts is discussed in terms of two hypotheses: (1) formation by a small degree of partial melting in the lunar interior, and (2) formation by prolonged near-surface crystallization differentiation in a lava lake. The second hypothesis is rejected on the following grounds: Most Apollo 11 magmas are far removed from the plagioclase-pyroxene-ilmenite cotectic; fractional crystallization cannot explain the large variations in concentrations of incompatible trace elements in conjunction with the small variations in major-element compositions, particularly, Mg/Fe ratios; experimentally determined partition coefficients show that the high abundances of Cr and V in Apollo 11 rocks cannot be reconciled with the previous separation of large quantities of ore minerals and pyroxenes. On the other hand, the major-element and trace-element contents of Apollo 11 rocks can be readily explained by partial melting of source material that buffers the major-element compositions and causes enrichments of incompatible elements according to the degree of partial melting (first hypothesis). Two alternative sources have been suggested for Apollo 11 basalts formed by partial melting: (1) unfractionated pyroxenite source region at depths of 200–600 km, and (2) fractionated source region with incompatible elements (e.g. Ba, U, and rare earths) strongly enriched over chondritic abundances and containing plagioclase (approximate eucritic composition). Mass-balance calculations and plagioclase-stability conditions show that the second hypothesis requires Apollo 11 basalts to be generated by partial melting in the outer 150 km of the moon. This is very difficult to achieve one billion years after the moon's formation, since the outer 200 km will have cooled well below the solidus by conduction. Furthermore, magmas generated by partial melting of a plagioclase-bearing source region should have plagioclase on the liquidus, which is contrary to observation. The second hypothesis accordingly appears improbable. The first hypothesis is capable of explaining the major-element chemistry and the trace-element abundances (Eu; see below) in terms of a simple, single-stage model that is consistent with the moon's density, moment of inertia, and inferred thermal history. A possible explanation of the europium anomaly is suggested on the basis of the first hypothesis. It will be necessary to determine the appropriate partition coefficients in order to test this explanation. If the lunar highlands are anorthositic, extensive differentiation of the outer 150 km of the moon is required. This may have been caused by heating arising from partial conservation of gravitational potential energy during the final stage of accretion. Formation of Apollo 11 basalts by partial melting 3.7 billion years ago was probably the result of radioactive heating (U, Th) in the deep interior of the moon. A two-stage magmatic history for the moon is thus required. Similarities between compositions of Apollo 11 and terrestrial basalts and between their respective source regions are suggestive of a genetic relationship between moon and earth. Nevertheless, important differences in trace-element abundances, major-element compositions, and oxidation states exist. These abundance patterns are unfavorable to the traditional fission, binary planet, and capture hypotheses of lunar origin. However, they may be explicable in terms of the precipitation hypothesis proposed by the author. This maintains that during the later stages of accretion of the earth, a massive primitive atmosphere developed that was hot enough to evaporate selectively a substantial proportion of the silicates that were accreting on the earth. Subsequently, the atmosphere was dissipated and the relatively nonvolatile silicate components were precipitated to form a swarm of planetesimals or moonlets, from, which the moon accreted.