Tungsten isotope evidence for post-giant impact equilibration of the Earth and Moon

Abstract:

The Moon is thought to have formed by re-accretion of material ejected by a giant impact on Earth [e.g., 1]. This model, at least in its classical form, predicts an isotopic difference between the Earth and Moon, because the Moon would largely consist of impactor material. Yet Earth and Moon show an unexpected isotopic similarity for many elements [e.g., 2]. Here we use variations in ¹⁸²W—the decay-product of short-lived ¹⁸²Hf (t_1/2~9 Myr)—between the Moon and the bulk silicate Earth (BSE) to shed new light on this issue. We precisely determined the lunar ¹⁸²W value by analysing KREEP-rich samples with MC-ICPMS and a new approach for quantifying cosmogenic ¹⁸²W variations using Hf isotopes [6].

We find that the Moon shows a 27±4 ppm ¹⁸²W excess over the modern BSE, in excellent agreement with [7]. This excess agrees with the predicted ¹⁸²W change resulting from disproportional late accretion to the Earth and Moon after Earth’s core had fully formed [6,7]. Thus, the pre-late-veneer BSE and the Moon were indistinguishable in ¹⁸²W. However, the giant impact itself should have caused a notable Earth–Moon ¹⁸²W difference by (1) changing the ε¹⁸²W of the proto-Earth mantle by adding impactor mantle and core material, both carrying distinct ¹⁸²W anomalies, and (2) by supplying W-rich but ¹⁸²W-depleted impactor core material into the lunar accretion disk [6]. Thus, the Earth–Moon ¹⁸²W homogeneity is an unexpected outcome of the giant impact. Unlike for Ti and O isotopes, the ¹⁸²W homogeneity cannot be explained by accretion of impactor and proto-Earth from a homogeneous inner disk reservoir [3] or by making the Moon fully from proto-Earth mantle [4,5]. Thus, the ¹⁸²W results require an efficient post-impact isotopic equilibration of the BSE and the Moon, but the mechanism for this has yet to be explored. One option is that Earth’s mantle and its vapour atmosphere remained connected with the lunar accretion disk just after the giant impact [8].

[1] Canup R.M. & Asphaug E. (2001) Nature 412, 708-712. [2] Zhang J. et al (2012) Nature Geosci. 5, 251-255. [3] Dauphas N. et al (2014) Phil Trans R. Soc. [4] Ćuk, M. & Stewart S.T. (2012) Science 338, 1047-1052. [5] Canup R.M. et al (2012) Science 338, 1052-1055. [6] Kruijer T.S. et al. (2015) Nature, 520, 534-537. [7] Touboul et al. (2015) Nature 520, 530-533. [8] Lock S.J. et al. (2015) LPSC #2193.