The Moon's cratered surface preserves the record of impacts that occurred during the late stages of its accretion. New simulations show that a collision with a companion moon may have formed the lunar farside highlands. See Letter p.69
The Moon dramatically exemplifies the role that collisions played in shaping the solid planets, particularly during the earliest history of the Solar System. The Moon's very existence is believed to be a consequence of debris accreting in Earth's orbit after a Mars-sized body collided with Earth shortly after it formed1,2 — a theory known as the giant-impact hypothesis. Lava-filled basins on the lunar nearside, and the massive South Pole–Aitken basin on the farside, mark the locations of major asteroid-sized impacts after the Moon accreted, melted and cooled. And thousands of craters all over the surface represent 'scars' of abundant smaller impacts, mostly accumulated during the waning stage of planet formation. The common attribute of the geological record of impact events is that they produced 'holes' — craters and basins. Now Jutzi and Asphaug (page 69 of this issue3) present intriguing simulations suggesting that a collision by a lunar companion that was created in the Moon-forming impact produced the extensive mountainous terrain on the lunar farside, rather than a crater or basin (Fig. 1).
The idea that multiple moons could have formed after the giant impact had been demonstrated by previous two-dimensional simulations4 of the dynamics of the Earth-encircling debris disk that was generated by the impact. Calculations4 showed that perturbations caused by the largest moonlet (our Moon) as it receded from Earth effectively removed material from the disk inside the Moon's orbit; a companion in this region would probably not have survived long.
The situation could have been different, however, for a companion at Earth–Moon Trojan points — locations ahead of and behind the orbit of the Moon that mark positions where the gravitational attraction of the Earth and Moon balance. Objects at these points are more stable than those at other parts of the Earth–Moon system, and could survive for up to tens of millions of years. At such time after accretion, the lunar crust would have been in place, probably consisting of a global 'magma ocean' with a thin, overlying crystallized crust5.
So a lunar companion could conceivably survive long enough for its eventual collision to occur after the Moon was fully accreted and partially cooled. But how could the impact make mountains rather than a hole? To do so would require special conditions, which Jutzi and Asphaug elucidated using calculations performed by a computer code6 of the kind used to study the giant impact. The authors' simulation considered the effect of planet-scale collisions by taking into account impact forces and gravity, as well as the deformational properties of relevant geological materials. The authors found that, to create elevated terrain of the volume and position of the farside highlands, a companion moon with a diameter about a third that of the current Moon would have been needed.
The companion, because it would have formed from the same debris disk as the Moon, would have had a similar composition, but would have solidified before the Moon did, so the age at which its rocks crystallized would be older than for lunar rocks. The authors deduced that the companion would need to have impacted the Moon at a velocity lower than the speed of sound, and that the volume of the lunar crust excavated must have been much less than the volume of the impactor. In addition, most of the material from the impacting companion must have remained near the impact point. This is in contrast to impacts that produce craters and basins, where the impactor exceeds the sound speed. In these hypervelocity events, material is excavated and ejected away from the impact point, and target rocks are melted and vaporized in amounts that scale with the kinetic energy of the event.
In the highlands-forming collision, the addition of material from the companion would have thickened the lunar crust, and such thickening is observed from analyses of lunar topography and gravity7. The collision would also have preferentially redistributed the underlying magma ocean to the nearside of the Moon, consistent with the distribution of heat-producing elements observed8 by the Lunar Prospector spacecraft.
The origin of the farside highlands has been a topic of speculation since the first global measurement9 of the Moon's shape. Ideas have included piling up of material excavated from the adjacent South Pole–Aitken impact9 basin, variations in tidal heating10, and a variety of other processes that produce spatial heterogeneities, for example shallow melting, crustal crystallization, interior dynamics and impact bombardment. The challenge is that the distributions of the Moon's elevation and crustal thickness calculated from orbital observations broadly fit sinusoids, and because numerous processes in nature can produce such a pattern, the ability to fit does not in itself distinguish between different mechanisms. Consequently, the current study3 demonstrates plausibility rather than proof. Because the accreted companion would have formed earlier than the Moon, demonstrating that the farside highlands have older crystallization ages would bolster support for this idea. However, that would require returning samples of the farside highlands to Earth to calculate the absolute ages.
More feasible in the near term will be to match simulations such as those carried out by the authors to the detailed internal structure of the Moon that will be obtained by topography (Fig. 1) from the Lunar Reconnaissance Orbiter mission and by high-resolution gravity mapping, which will be obtained next year by the Gravity Recovery and Interior Laboratory (GRAIL) mission. In any case, Jutzi and Asphaug raise3 the legitimate possibility that, after the giant impact, our Earth perhaps fleetingly possessed more than one moon. Furthermore, significant remnants of this long-departed member of the Earth–Moon collisional family may be preserved today on the lunar farside.
Cameron, A. G. W. in Origin of the Moon (eds Hartmann, W. K., Phillips, R. J. & Taylor, G. J.) 609–620 (Lunar Planet. Inst., 1986).
Canup, R. M. & Esposito, L. W. Icarus 119, 427–446 (1996).
Jutzi, M. & Asphaug, E. Nature 476, 69–72 (2011).
Canup, R. M., Levison, H. F. & Stewart, G. R. Astron. J. 117, 603–620 (1999).
Warren, P. H. Annu. Rev. Earth Planet. Sci. 13, 201–240 (1985).
Benz, W. & Asphaug, E. Comput. Phys. Commun. 87, 253–265 (1995).
Neumann, G. A., Zuber, M. T., Smith, D. E. & Lemoine, F. G. J. Geophys. Res. doi:10.1029/96JE01246 (1996).
Lawrence, D. J. et al. Science 281, 1484–1489 (1998).
Zuber, M. T., Smith, D. E., Lemoine, F. G. & Neumann, G. A. Science 266, 1839–1843 (1994).
Garrick-Bethell, I., Nimmo, F. & Wieczorek, M. A. Science 330, 949–951 (2010).
Smith, D. E. et al. Geophys. Res. Lett. doi:10.1029/2010GL043751 (2010).