Abstract:
The decay of the short-lived isotope 182Hf to the isotope 182W early in solar system history provides clues to the origin and evolution of the Earth and Moon. Current time-scales for the formation of proto-planets in the solar system require early accretion of the proto-Earth, accompanied with core formation, which should lead to non-chondritic radiogenic W isotope ratios (182W/184W) in the Earth's mantle. Surprisingly, the observed ratio of 182W/184W in the Earth's mantle is chondritic, which requires special circumstances to explain (Halliday et al., 1996; Lee and Halliday, 1997; Lee et al., 1997). Previous explanations include delayed core formation until 50 M yrs. after the origin of the solar system, or a very long time-scale for accretion of the Earth (Halliday et al., 1996; Lee and Halliday, 1997; Lee et al., 1997). In this paper, we explore models consistent with the early formation of the proto-Earth, a giant impact origin of the Moon, and substantial additional accretion to the Earth following the impact. The post-impact processes responsible for the chondritic W isotopes may include late core-mantle equilibration in the hot Earth following the giant impact, and later metal segregation and accretion.In contrast to the Earth, various igneous rocks on the Moon exhibit non-chondritic, positive anomalies in the ratio of 182W/184W. The explanation for the lunar samples with chondritic W isotopes may also involve late metal segregation and accretion, as in the Earth. The positive anomalies in lunar samples may be derived in part or whole through cosmogenic processes on the lunar surface, Hf-W fractionation during core formation, or Hf-W fractionation during lunar magma ocean crystallization. If the latter, the positive anomaly places constraints on the early differentiation of the lunar mantle and the duration of the crystallization of the lunar magma ocean. The observed W-Hf systematics are best explained if the W anomalies are carried by late-crystallizing ilmenite and high-Ca pyroxene, which is later mixed into the source regions of the mare basalts. In this model, the crystallization of the magma ocean must occur in less than 40 million years. This rapid crystallization implies that the very early lunar crust was not as stable and insulating as previously suggested, and that the fundamental change from magmasphere to serial magmatism processes occurred very early in the evolution of the Moon.