Silicon isotopes in lunar rocks: Implications for the Moon's formation and the early history of the Earth

Silicon isotopic data from a range of lunar samples are presented to assess the degree of heterogeneity of the lunar mantle and its similarity to bulk silicate Earth (BSE). Multi-collector inductively-coupled-plasma mass spectrometry (MC-ICPMS) was used to analyse 24 samples, including both high and low-Ti basalts, as well as Highland anorthosites and picritic glasses, covering all the Apollo sample return missions. No systematic delta Si-30 differences are found between any of the bulk sample lithologies (+/- 2 sigma(SD)) (delta Si-30(Low-Ti) (basalt) = 0.29 +/- 0.06, delta Si-30(High-Ti) (basalt) = 0.32 +/- 0.09, delta Si-30(lunar) (glass) = 0.29 +/- 0.05 and delta Si-30(Highland) (rocks) = -0.27 +/- 0.10). The average of the lunar samples is delta Si-30 = -0.29 +/- 0.08 (2 sigma(SD)), which is identical to the composition of BSE, delta Si-30 = -0.29 +/- 0.08 (2 sigma(SD)), from Savage et al. (2010). The BSE Si isotope composition is thought to be the result of Si partitioning between metal and silicate, and consequent isotopic fractionation during core formation. The Moon-forming impactor would not be expected to share that composition, because it is thought to have been relatively small (similar to 0.1 Earth masses) like Mars and formed under relatively low temperatures and pressures that are insufficient for Si to partition into the core. Therefore, the identical lunar and BSE Si isotope data show that Si isotopes, like those of oxygen, must have homogenised in the aftermath of the Moon-forming impact, if smooth particle hydrodynamic simulations of the Giant Impact are correct in showing that most lunar material should have originated from the impactor rather than the Earth. The data presented here are in agreement with other isotope systems and experimental studies that indicate that the majority of core formation happened early and before the Giant Impact. It has been predicted (Pahlevan et al., 2011) that the Moon and the BSE should show a similar to 0.14 parts per thousand offset in delta Si-30 for the Moon to have an Fe/(Fe + Mg) ratio twice that of BSE. The current resolution and sample population size of the Si data for the Moon and Earth would allow such an offset to be detected. The fact that is it not observed can put constraints on the element ratios and lunar budgets as modelled by Pahlevan et al. (2011); in particular, it constrains the Fe/(Fe + Mg) ratio of the Moon to be only 1-1.3 BSE. (C) 2011 Elsevier Ltd. All rights reserved.