Silicon isotope constraints on terrestrial planet accretion

Understanding the nature and origin of the precursor material to terrestrial planets is key to deciphering the mechanisms and timescales of planet formation(1). Nucleosynthetic variability among rocky Solar System bodies can trace the composition of planetary building blocks(2–5). Here we report the nucleosynthetic composition of silicon (μ(30)Si), the most abundant refractory planet-building element, in primitive and differentiated meteorites to identify terrestrial planet precursors. Inner Solar System differentiated bodies, including Mars, record μ(30)Si deficits of −11.0 ± 3.2 parts per million to −5.8 ± 3.0 parts per million whereas non-carbonaceous and carbonaceous chondrites show μ(30)Si excesses from 7.4 ± 4.3 parts per million to 32.8 ± 2.0 parts per million relative to Earth. This establishes that chondritic bodies are not planetary building blocks. Rather, material akin to early-formed differentiated asteroids must represent a major planetary constituent. The μ(30)Si values of asteroidal bodies correlate with their accretion ages, reflecting progressive admixing of a μ(30)Si-rich outer Solar System material to an initially μ(30)Si-poor inner disk. Mars’ formation before chondrite parent bodies is necessary to avoid incorporation of μ(30)Si-rich material. In contrast, Earth’s μ(30)Si composition necessitates admixing of 26 ± 9 per cent of μ(30)Si-rich outer Solar System material to its precursors. The μ(30)Si compositions of Mars and proto-Earth are consistent with their rapid formation by collisional growth and pebble accretion less than three million years after Solar System formation. Finally, Earth’s nucleosynthetic composition for s-process sensitive (molybdenum and zirconium) and siderophile (nickel) tracers are consistent with pebble accretion when volatility-driven processes during accretion and the Moon-forming impact are carefully evaluated.