As a planetary body forms, precious metals, such as gold and iridium, are stripped from its rocky mantle and passed into its metallic core. Successive impacts with other objects then re-enrich the mantle in these elements — a process known as late accretion1,2. Measurements of lunar rocks show that the Moon is greatly depleted in precious metals compared with Earth3,4. This deficit implies that the ratio of mass added to Earth during late accretion compared with that added to the Moon is more than 1,000:1, which is substantially different from the predicted ratio5 of about 20:1. In a paper in Nature, Zhu et al.6 show that inefficient delivery of material from glancing impacts, combined with an early hot, molten stage on the Moon, can explain this anomalous input-mass ratio.
An analysis of the mass and composition7 of material added to a planetary body can be used to examine the body’s formation. Moreover, late accretion is linked to the delivery of water and other volatile elements to Earth8, and such additions are probably a key factor in our planet’s habitability. The low abundances of precious metals in lunar rocks3,4,9 has prompted the proposal of competing models to explain the anomalous input-mass ratio between Earth and the Moon.
At one extreme, these models include delivery of material by a few massive impactors (larger than 2,500 kilometres in diameter) that preferentially struck Earth5. At the other, focusing of small objects (less than 10 m in diameter) on to Earth might have produced similar effects10. It has also been suggested that the difference in precious-metal abundances between Earth and the Moon was caused by a drop in the flux of impacts during the period between 4.5 billion and 4.1 billion years ago11, just after the Solar System formed. These models generally assume that the Moon retained about half of the mass that was transferred to it by impactors.
Using millions of computational impact simulations, Zhu and colleagues examined the fraction of impactor mass that could be retained by planetary bodies. The authors simulated impacts at different velocities (10–20 km per second), and at low angles (20°) to high angles (80°) with respect to the body’s surface (Fig. 1). They found that material from larger impactors is less effectively retained than that from smaller counterparts, and that high-angle impacts deliver a larger mass fraction to the body than do low-angle impacts.
In the case of Earth, these results imply that the retention of impactor mass is generally high for all but the most glancing impacts with the most massive objects. For the Moon, which has a mass only about 1% of that of Earth, the shallower the angle of impact, and the more massive the impactor, the greater the likelihood that material would be lost, never reaching the Moon’s surface or passing into its interior. Using crater diameters12 to establish the frequency and size of impactors striking the Moon, Zhu et al. discovered that impactor-mass retention probably changed modestly over time, and that the average retention was about 20%, which is around three times lower than previous estimates.
Inefficient retention of material from objects striking the Moon partially offsets the difference between the theoretically and geochemically determined Earth–Moon input-mass ratios. Zhu and colleagues then argue that about 50% of late-accretion input mass was lost to the Moon’s deep interior or core before 4.35 billion years ago, and that this loss explains any remaining discrepancy. Later, once the Moon had cooled, late-accretion input mass was distributed into the lunar mantle and crust. The authors further suggest that as many as 300 impact craters of more than 300 km in diameter might have existed on the Moon, but that fewer than 30% of these craters are preserved today owing to impact-derived erosion or to gradual subsidence (viscous relaxation) of the earliest craters in the hot lunar crust.
The suggestion of inefficient mass retention from glancing impacts negates the requirement for the proposed temporally varying impact fluxes11. However, the idea that precious metals were lost to the Moon’s deep interior or core before 4.35 billion years ago is more problematic. Without evidence from craters for the impact flux to the Moon at that time, geochemistry is the only valid test of this idea. Loss of precious metals to a metallic core can lead to these elements being separated (fractionated) from one another13. However, this chemical effect has not so far been detected in rocks from the lunar interior3,4. Furthermore, low precious-metal abundances estimated for the lunar mantle make it difficult to envisage how such fractionation signatures could have been erased by further late accretion.
Zhu et al. also assume that the penetration of impactors through the lunar crust, which is about 40 km thick14, would lead to all retained impactor material entering the mantle. But, in reality, this material would pollute both the crust and the mantle. Finally, because only a relatively small number of lunar rocks have been analysed, models such as the authors’ that can reproduce precious-metal abundances in the Moon through simulations have limited resolution.
Nevertheless, the new models will be of value in understanding the evolution of planetary bodies, especially Mars. Current estimates for late-accretion input mass expressed as a function of total body mass are 0.02%, about 0.5% and up to 0.7% for the Moon4, Earth2 and Mars15, respectively. The estimate for Mars has been explained by early formation of the planet relative to Earth and the Moon, in addition to a constant input-mass flux in the Solar System’s first 50 million to 100 million years15, and impacts involving massive objects5,16.
Extrapolation of Zhu and colleagues’ models would suggest that Mars, which has a mass only about 11% of that of Earth, retains less material from large impactors than does Earth. Assuming that the two bodies were subjected to a similar number of glancing impacts, these models would imply that Mars had a proportionally greater late-accretion flux than did Earth. The combination of geochemistry and formation models will undoubtedly continue to improve our understanding of how Earth and its nearest neighbours came to be.
Nature 571, 177-178 (2019)