|The Earth’s Moon is thought to have formed by accretion from a disk generated by a giant impact onto the Earth. In the canonical case, the impactor is a Mars-size object and the disk is composed primarily of impactor material (e.g. Canup 2004, 2008). Since the impactor likely had a composition different from that of the Earth, this seems at odds with the identical isotopic compositions of the Earth’s mantle and the Moon. Pahlevan & Stevenson (2007) suggested that material exchange between the disk’s and Earth’s atmospheres could modify the composition of the disk to match that of the silicate Earth, resulting in compositional equilibration in O(100) years, a timescale much longer than that predicted for lunar accumulation from the disk (Ida et al. 1997, Kokubo et al. 2000).
Using a more accurate modeling of the Moon’s accretion from the protolunar disk, accounting in particular for the presence of vapor in the disk, we have shown that the Moon’s accretion from the disk occurs on a timescale compatible with that required for equilibration to occur (Salmon & Canup 2012). However, it may be difficult for equilibration to occur without simultaneously depleting the disk of its mass (Melosh 2009). In addition, in our model a substantial portion of the Moon accumulates rapidly after the impact from material placed into distant orbits, and at least this portion appears unlikely to equilibrate with the Earth.
Recently, new types of impacts have been proposed, involving either larger impactors (Canup 2012) or high-velocity impacts on a fast-spinning Earth (Cuk & Stewart 2012), and resulting in a protolunar disk whose composition is much closer to that of the post-impact Earth. These impacts, however, leave the Earth-Moon system with an excess of angular momentum. Subsequent capture of the Moon into the evection resonance has been argued to be capable of reducing the angular momentum of the Earth-Moon system by a factor of 2, making it compatible with its current value (Cuk and Stewart 2012).
We have identified two main concerns with these non-canonical impacts: 1) they form more compact disks, with most of the mass located inside the Roche limit at 2.9 R_Earth. Incorporation of material from this inner region into the Moon is rather inefficient (Salmon and Canup 2012), so that formation of Moon-size objects may be compromised; 2) capture into the evection resonance seems possible only for a narrow range of orbital parameters. Previous work assumed that the Moon formed around 3.8 R_Earth (Cuk & Stewart 2012), while we found that the Moon forms in fact around 6 R_Earth.
We have modeled the accretion of the Moon from non-canonical disks, and find that forming a Moon-size object requires very massive disks that may only be achievable by the impact-scenario of Canup (2012). We also find that the Moon is driven even farther away than in canonical cases, which may compromise subsequent capture into the evection resonance.