Where did the Moon come from?

The Moon as a whole holds less than a tenth of its mass in iron, although its mantle is relatively rich in it: what is missing is the metal. Its rocks are also depleted in volatile elements relative to Earth, most probably by evaporation, during the impact or in a later ocean of magma. The iron core reaches only about 330 kilometres out, a solid inner core of 240 inside a liquid shell 90 thick, in a body whose radius is about 1,740. Proportionally it is smaller than any other rocky world’s, and far smaller than Earth’s, which holds nearly a third of the planet’s mass.

Its oxygen and titanium isotope ratios, though, match Earth’s to high precision. The Moon’s 50Ti/47Ti sits within about four parts per million of Earth’s, a hundred and fiftieth of the range meteorites cover, and a 2024 oxygen study, working at a precision earlier ones could not reach, found no difference down to a part or two per million. Mars and most meteorites carry isotopic fingerprints of their own, and the Moon carries ours.

In the standard version a body about the size of Mars struck the young Earth obliquely. The collision threw mantle material into orbit, most of it the impactor’s, where it gathered into the Moon, and left Earth spinning fast.

That accounts for the Moon’s mass, the angular momentum of the pair and the missing iron, and offers one route to the missing volatiles. It is not a fringe idea and it is not in trouble as a class of explanation.

Simulations of the classic version put most of the Moon’s material from the impactor rather than from Earth. Unless that body happened to be isotopically almost identical to Earth, or the two were mixed thoroughly afterwards, the samples should show it. Some researchers think such a twin unlikely; others argue that neighbours built from the same material would match.

Two fixes were published together in 2012. In one the impactor is much larger than the canonical Mars-sized body, comparable in mass to the proto-Earth, and the collision leaves a disc with the composition of Earth’s mantle. In the other the Earth is spinning very fast before a smaller body strikes it, and the Moon forms largely from Earth’s own mantle. Both leave the pair turning too fast, and both rely on a later resonance between the Moon and the Sun to shed the excess. Later work finds the resonance can do that only in a limited range of conditions, and a 2026 study that follows the young Earth as it cools finds it unlikely to manage alone.

Other routes exist: vapour mixing the disc with Earth after the impact, a Moon built from many smaller impacts, and a hot, vaporised, fast-spinning body called a synestia. In 2022 simulations run with far more particles than the earlier ones found that some impacts place a satellite of about the right mass and iron content into orbit almost at once, rather than accreting it from a disc. Its outer layers come out around sixty per cent proto-Earth material, though the satellite as a whole is still mostly impactor.

Each works under some set of conditions. None has been singled out by the samples, and a review written by several of the models’ own authors says all of them involve “conditions whose likelihood appears small or remains highly uncertain”.

The models differ in how much impactor they leave in the Moon, so measuring the isotopes more precisely in more elements is one lever. Another is Venus: if a planet that size proved isotopically Earth-like, an Earth-like impactor would look less like luck. So is sampling the parts of the Moon the Apollo and Luna collections do not cover.

The far side has now been sampled. Chang’e-6 returned material from the South Pole-Aitken basin, and a 2024 study dated basalt from it at 2,830 million years, give or take five. The deep interior is still out of reach.

It is waiting on ideas as much as on rocks. Whether vapour could really have mixed the Moon and Earth, and whether the resonance could shed as much spin as the models ask of it, are both still open.