For decades, researchers have debated how planetary bodies in the Solar System with diameters of more than a few hundred kilometres were formed. Current models1 suggest that such bodies accumulated (accreted) small particles known as pebbles that were drifting towards the Sun in a cloud of gas and dust called the protoplanetary disk. However, these models have difficulty in explaining the diverse composition of objects in the Solar System: if all such bodies grew by accretion from the same flow of pebbles, then why do they have different compositions? In a paper in Nature, Schiller et al.2 provide a possible answer to this question, on the basis of their observation that the composition of calcium isotopes in planetary bodies correlates with the masses of these bodies.
The authors measured the calcium-isotope content of samples from the parent bodies of types of meteorite known as angrites and ureilites, as well as from Earth, Mars and the asteroid Vesta. They present their data in terms of the value μ48Ca, which expresses the ratio of two calcium isotopes (48Ca and 44Ca) in a sample relative to that in a terrestrial reference standard, and is given as parts per million (p.p.m.). The authors identified a positive correlation between μ48Ca and planetary-body mass for Earth, Mars and Vesta, which have known masses, and for the parent bodies of angrites and ureilites, the masses of which were inferred from thermal models3,4.
Schiller et al. account for this correlation by assuming that the pebble-accretion scenario is correct, but by rejecting the conventional idea that bodies of different masses grew at different rates throughout the lifetime of the protoplanetary disk. Instead, the authors suggest that all bodies grew at the same rate, but stopped growing at different times: smaller bodies ceased accretion earlier than did larger ones.
This unorthodox view of growth is, in fact, supported by numerical simulations5, which show that growing bodies perturb each other’s orbits around the Sun. Bodies that acquire eccentric or inclined orbits stop accumulating pebbles, whereas those that remain on circular trajectories in the mid-plane region of the disk continue to grow. The correlation of μ48Ca values with planetary-body masses therefore also becomes a correlation with the timescale of the growth of such bodies. The authors confirm this using growth timescales for angrites, ureilites, Vesta and Mars that were inferred by nuclear chronometry (a dating technique that uses the decay of radioactive isotopes) and thermal models.
Schiller and colleagues propose that material in the inner part of the disk initially had low μ48Ca values (about −150 p.p.m.). Planetary bodies grew from this matter until they reached the size of the ureilite parent body (about 200 km in diameter). The inner disk then started to be fed with pebbles that drifted in from the outer part of the disk. Such pebbles had high μ48Ca values (about 200 p.p.m.), which are typical of the most primitive meteorites (carbonaceous chondrites) that are thought to have formed beyond Jupiter6. The mean μ48Ca value in the inner disk therefore increased progressively. As a consequence, bodies that continued to grow to the size of Vesta (530 km in diameter) reached μ48Ca values of −100 p.p.m., and those that grew to the size of Mars (6,800 km in diameter) reached μ48Ca values of −20 p.p.m (Fig. 1).
Earth and the Moon have μ48Ca values of about 0 p.p.m. It is generally assumed that Earth formed after the disappearance of the protoplanetary disk, as a result of collisions of planetary embryos with masses similar to that of Mars; the body that collided with early Earth to give rise to the Moon was one of these embryos. However, this formation scenario is not possible if Schiller and colleagues’ proposal is correct. Planetary embryos with the mass of Mars would have μ48Ca values of −20 p.p.m., and therefore an ‘Earth’ resulting from the merger of these bodies would have a similar composition.
Instead, for Earth to have a μ48Ca value of about 0 p.p.m., the embryos would need to have grown by pebble accretion to masses roughly half that of Earth. In turn, the Moon-forming impactor would need to have had a mass comparable to that of the embryos to have the same μ48Ca value, supporting the theory that the Moon resulted from the collision of two bodies with half-Earth masses7.
A potential problem is that hafnium–tungsten radioactive chronometry indicates that Earth reached 63% of its present mass only after a duration of between 11 million and 24 million years, depending on the type of core–mantle equilibration that occurred during the collision of the embryos8. It is therefore difficult to imagine that such embryos grew to bodies with half-Earth masses in the disk’s putative lifetime of 5 million years9.
But perhaps the embryos did reach the size of one-third of Earth’s mass (three times the mass of Mars). Such a proposal cannot be discounted as a compromise between the authors’ correlation of calcium-isotope compositions with planetary masses and the chronological constraints on Earth’s accretion. The fraction of Earth’s mass that would have been accreted from the outer disk (estimated at about 40% in the authors’ study) would be higher than calculated previously10,11, but such computations did not consider some of the building blocks of Earth to have ureilite-like compositions.
Schiller and colleagues’ view of accretion in the Solar System is in sharp contrast with that presented by two previous studies6,12. These reports concluded that the flux of pebbles from the outer disk was shut down during the first million years of the disk’s lifetime by the formation of Jupiter. This prevented bodies in the inner Solar System from accumulating large amounts of water ice, explaining why such bodies are mostly dry12, and maintained an isotopic dichotomy between two types of meteorite: ordinary and carbonaceous chondrites6.
The composition of ordinary chondrites is difficult to account for using Schiller and colleagues’ model. Unlike large bodies, which should grow continuously by pebble accretion, the smaller parent bodies of chondrites should form suddenly, from clusters of pebbles that are generated by a mechanism known as a streaming instability1. In the case of ordinary chondrites, this would have happened in the inner Solar System at a late stage in its formation, after the time by which Mars had accreted most of its mass. Ordinary chondrites should therefore represent a snapshot of the composition of the late disk, and have a positive value of μ48Ca in the authors’ model. But, in reality, such chondrites have μ48Ca values of −35 p.p.m.
Schiller et al. explain this discrepancy by speculating that pebble accretion and the streaming instability have differing preferences for pebble size. The pebbles that came from the outer disk were small and, although they were efficiently accreted by large bodies, they barely participated in the streaming instability. Consequently, ordinary chondrites formed mostly from larger pre-existing pebbles called chondrules, which typically have negative values of μ48Ca. The validity of this proposal will need to be checked using high-resolution numerical simulations of the pebble-accretion and streaming-instability processes.
The authors’ work adds a missing piece to the jigsaw puzzle of planet formation that will need to be connected with the other pieces provided by isotopic, chemical, chronological and dynamical constraints. Although the puzzle seems more complete than before, perhaps some other key pieces are still missing.
Nature 555, 451-452 (2018)
Sign up for the daily Nature Briefing email newsletter
Stay up to date with what matters in science and why, handpicked from Nature and other publications worldwide.