Hf-W Isotope Geochemistry, Core Formation, the Lead Paradox and the Origin of the Moon

Alex Halliday Department of Geological Sciences, University of Michigan, Ann Arbor, MI 48109-1063, USA


Mark Rehkämper Department of Geological Sciences, University of Michigan, Ann Arbor, MI 48109-1063, USA

Der-Chuen Lee Department of Geological Sciences, University of Michigan, Ann Arbor, MI 48109-1063, USA

Wen Yi Department of Geological Sciences, University of Michigan, Ann Arbor, MI 48109-1063, USA

Hafnium and W are both highly refractory elements. However, Hf is lithophile whereas W is moderately siderophile such that the Hf/W ratio of the Earth's primitive mantle or BSE is about a factor of 20 greater than chondritic following core formation. 182Hf decays to 182W with a half life of 9 Myrs. Therefore, if the Earth's core formed early, during the lifetime of 182Hf, an excess of 182W would be recorded in the W isotopic composition of the BSE, relative to that found in chondrites, the magnitude of this excess being a function of the time of core formation. Harper and co-workers (1991) made the first W isotopic measurements of iron meteorites and demonstrated a significantly lower 182W abundance in the iron meteorite Toluca relative to that found in the silicate Earth. Lee and Halliday (1995) found that such 182W deficiencies were present in other iron meteorites and also showed that the W isotopic composition of the BSE is identical to that of the carbonaceous chondrites Allende and Murchison. From this it follows that core formation must be late and postdate iron meteorites by 62±10Myrs. This conclusion holds even if the silicate Earth's inventory of W is dominated by late additions (Newsom, 1990). The age of core formation deduced from Hf-W is in fact fully consistent with models based on Pb isotopes (Allègre et al., 1982) and helps to provide rigorous time constraints for these models. Many models of terrestrial accretion and core formation involve an early core that developed during the first 90% of accretion history (Lee and Halliday, 1995; Stevenson, 1981; Sasaki and Nakazawa, 1986), generally considered to correspond to a time span (Wetherill, 1980) significantly shorter than that permitted by the W isotopic data. These models can only be reconciled with the W isotopic data if a proto-Earth with an early core was somehow rehomogenized, such as by a major impactor.

There is evidence that the Moon was derived from the Earth after terrestrial core formation. Isotopic ages for lunar highlands rocks (Carlson and Lugmair, 1988; Shih et al., 1993; Alibert et al., 1994; Tera et al., 1974; Hanan and Tilton, 1987) combined with the W isotopic age constraints indicate that terrestrial core segregation and the formation of the Moon took place within < 80 Myrs at 4.47±0.04 Ga. Certain isotopic ages for lunar rocks (Alibert et al., 1994) would be consistent with a more restricted time window of 4.50±0.01 Ga. The entire accretion history of the earth may have been completed at this time (Ringwood, 1992). Potassium (Humayun and Clayton, 1995) and Cr (Lugmair and MacIsaac, 1995) isotopic data indicate early volatile depletion of the material from which the Earth and Moon formed and constrain models of pre-core Pb isotopic evolution. Using the 4.47±0.04 Ga age of the core, the second stage Pb isotopic evolution reproduces reasonable estimates for the present day BSE Pb isotopic composition if the second stage 238U/204Pb (m) is in the range of 8.9±0.5. The "lead paradox" (Allègre, 1982) is entirely predictable from the 4.47±0.04 Ga age of core formation and does not require explanations such as late accretion, storage of unradiogenic Pb in the lower continental crust, recycling of U-enriched altered MORB or upper continental crust, or contamination of a highly depleted MORB reservoir by OIB components. The similarity in age of the Earth's core, the Moon and degassing of the terrestrial mantle (Allègre et al., 1995) would be consistent with a genetic link to a single late event. However, in detail it would appear that the history may have been significantly more complex (Ringwood, 1989). For example, the Hf-W data may define the age of a core that formed as a result of a giant impact, prior to the addition of volatile-rich impactors that formed the late veneer and triggered the creation of the Moon. However, even if these more complex scenarios are correct, the time span is still within 4.47±0.04 Ga.


Alibert, C., Norman, M.D. & McCulloch, M.T., Geochim. Cosmochim. Acta 58, 2921-2926 (1994).

Allègre, C.J., Tectonophysics 81, 109-132 (1982).

Allègre, C.J., Dupré, B. & Brévart, O., Phil. Trans. R. Soc. Lond. A 306, 49-59 (1982).

Allègre, C.J., Manhès, G. & Göpel, C., Geochim. Cosmochim. Acta 59, 1445-1456 (1995).

Carlson, R.W. & Lugmair, G.W., Earth Planet. Sci. Lett. 90, 119-130 (1988).

Hanan, B.B. & Tilton, G.R., Earth Planet. Sci. Lett. 84, 15-21 (1987).

Harper, C.L., Völkening, J., Heumann, K.G., Shih, C-Y. & Wiesmann, H., Lunar Planet. Sci. XXII, 515-516 (1991).

Humayun, M. & Clayton, R.N., Geochim. Cosmochim. Acta 59, 2131-2151 (1995).

Lee, D-C. & Halliday, A.N., Nature, (1995, in press).

Lugmair, G.W. & MacIsaac, Ch., Lun. Plan. Sci. Conf. XXVI, 879-880 (1995).

Newsom, H.E., In: Origin of the Earth (Newsom, H. E. & Jones, J. H., eds.) 273-288 (Oxford University Press, Oxford, 1990).

Ringwood, A.E., Earth Planet. Sci. Lett. 95, 208-214 (1989).

Ringwood, A.E., Earth Planet. Sci. Lett. 111, 537-555 (1992).

Sasaki, S. & Nakazawa, K., J. Geophys. Res. 91, B9231-9238 (1986).

Shih, C.-Y., Nyquist, L.E., Dasch, E.J., Bogard, D.D., Bansal, B.M. & Wiesmann, H., Geochim. Cosmochim. Acta 57, 915-931 (1993).

Stevenson, D.J., Science 214, 611-619 (1981).

Tera, F., Papanastassiou, D.A. & Wasserburg, G.J., Earth Planet. Sci. Lett., 22, 1-21,(1974).

Wetherill, G.W., Ann. Rev. Astron. Astrophys. 18, 77-113 (1980).