Although the giant impact hypothesis is the most accepted model for the formation of the Moon the origin of its volatile depletion is still matter of debate. The heavy isotopic composition of moderately volatile elements like K, Rb or Zn is often interpreted as supporting evidence for a large-scale volatile depletion event. On the other hand, elevated abundances of water and other volatile elements in some lunar rocks are interpreted in favour for a less volatile depleted interior and more local volatile-loss processes like magmatic degassing. In the latter model, the heavy stable isotope composition of lunar rocks would be a result of late stage magmatic degassing into vacuum. Here we explore the processes affecting volatile elements in mare basalts through the scope of copper and zinc isotopes.

We report new data for mass-dependent stable isotopes of copper and zinc determined from the same rock aliquot of low- and high-Ti mare basalts. Thanks to the combine data set and the high quality of double spike Zn isotopic data, we resolve the effects of fractional crystallization and late magmatic degassing. Based on these results, fractional crystallization and late-stage magmatic degassing cannot explain volatile depletion and the heavy isotopic composition of most mare basalts and their mantle sources. The homogeneous Zn isotopic composition of low and high-Ti basalt mantle sources suggest that volatile loss and mass-dependent isotope fractionation occurred before the formation of these lunar mantle reservoirs, likely during or briefly after the giant impact.

Elevated contents of water and moderately volatile elements in some lunar materials have invigorated the discussion on the volatile content of the lunar interior and on the extent to which the volatile element inventory of lunar magmatic rocks is controlled by element volatility and degassing. In order to constrain magmatic processes and mantle source compositions, we obtained a comprehensive data set for mass fractions of the moderately to highly volatile elements Cu, Se, Ag, S, Te, Cd, In, and Tl in various low- and high-Ti mare basalts. Mass fractions of Cu, S, Se, and Ag in each suite are mainly controlled by fractional crystallization. In contrast, Te, Cd, In, and Tl display disturbed fractional crystallization trends, most likely due to late magmatic degassing and recondensation of volatile species of these elements. Low-Ti mare basalt suites display constant ratios of specific siderophile volatile elements (e.g. Cu/Ag, Cu/S, S/Se), which we interpret as characteristics of their mantle sources. High-Ti mare basalt suites differ from low-Ti mare basalts by their significantly lower Cu/S, but higher S/Se ratios. Fractional crystallization modeling reveals that these differences are inherited mainly from their source regions in the lunar mantle. Despite the systematically different element ratios, low- and high-Ti mare basalt source compositions are characterized by consistently low mass fractions of siderophile volatile elements. Our new data support the hypothesis of volatile loss prior to formation of the lunar mantle sources and reveal element ratios in the lunar mantle that are significantly different from the terrestrial mantle.

Understanding the early lunar bombardment history hinges on reliable formation ages of the large lunar basins. Recent studies have shown that Zr minerals and Ca phosphates in petrographically, chemically, and microstructurally well-characterized lunar impactites can yield easily reproducible U-Pb ages that can be related to impacts. Geological, textural, and chemical arguments, plus impact melt distribution models, suggest that the widespread 3.92 Ga and 4.21 ages in breccias at the Apollo 14-17 landing sites likely reflect the Imbrium and Serenitatis impacts, respectively. However, the ages of other basins are more uncertain. Lunar granulites – impactites that were thermally metamorphosed by impact melt sheets – also should record basin-forming events. We have combined precise in situ Pb-Pb dates with Pb-Pb isochron dating to constrain the early history of the Apollo 17 granulite 77017. The rock contains annealed anorthositic gabbro clasts with relict igneous textures and a finer, thermally annealed matrix. High Ir abundances and the presence of metal indicate that the gabbro crystallized from impact melt. Baddeleyites define a homogeneous Pb-Pb age distribution (4175±3 Ma, n = 5), which is interpreted to date the crystallization of this impact melt. The phosphate Pb-Pb ages range between 4.18 and 4.13 Ga and are consistent with variable resetting by the metamorphic heating event. A Pb-Pb isochron yields a precise date for the granulite facies metamorphism (4143±1 Ma). The heating event may either relate to the impact that formed the Crisium basin or to another nearby basin-forming impact on the feldspathic highland terrane (Nectaris or Smythii).