Formation of the Solar System: from protoplanetary disc to planets

At a glance

The Sun and its planets formed together from a rotating cloud of gas and dust about 4.6 billion years ago. The broad picture, the nebular hypothesis, goes back to Kant and Laplace in the eighteenth century, but the details are now tested in three independent ways: by dating the oldest solids in meteorites, by imaging planet-forming discs around young stars, and by analysing samples brought back from asteroids. The table collects key benchmarks used throughout this page.

Sources: Connelly et al. (Science, 2012) via AAAS/EurekAlert; ESO press release eso1436 (6 November 2014); Andrews et al. (DSHARP I, 2018); Kruijer et al. (PNAS, 2017) via LLNL; JAXA ISAS initial-analysis release (10 June 2022); NASA New Horizons mission page (accessed October 2026). CAI = calcium–aluminium-rich inclusion.

Sources: [1] [2] [3] [6] [9] [13]

From a collapsing cloud to a disc

Stars form inside cold molecular clouds (see our nebulae and clusters page). When a dense clump collapses, any slight rotation is amplified as it shrinks, just as a spinning skater turns faster with arms pulled in. Material falling along the rotation axis reaches the centre, where the protostar grows, but material with more angular momentum settles into a flattened, rotating protoplanetary disc. The young Sun was surrounded by such a disc, made mostly of gas with a small admixture of dust.

The clock for this era is set by meteorites. Primitive meteorites called chondrites contain millimetre-to-centimetre calcium–aluminium-rich inclusions (CAIs), the first solids to condense in the hot inner disc, and once-molten droplets called chondrules. Lead isotope dating by James Connelly and colleagues gave CAIs an age of 4,567.30 ± 0.16 million years, the conventional “time zero” of the Solar System, and showed that chondrules began forming at the same time and continued for about three million years, a span that matches the lifetimes of gas discs observed around other young stars.

Direct images of such discs arrived in 2014, when the Atacama Large Millimeter/submillimeter Array (ALMA) photographed the disc of HL Tauri, a star about 450 light-years away and no more than about a million years old. The image showed concentric bright rings separated by dark gaps, features widely interpreted as signs of forming planets. The DSHARP survey then imaged 20 nearby discs at a resolution of about 5 au and found that rings, gaps and spirals are common.

Sources: [1] [2] [3]

The snow line and the two kinds of planets

Temperature in the disc falls with distance from the young Sun. Close in, only rock- and metal-forming minerals can exist as solids; farther out, water vapour freezes onto dust grains. The boundary is called the snow line (or frost line). Beyond it, the amount of solid material available roughly multiplies, because ice is abundant, and this helps explain the basic architecture of the Solar System: small rocky planets inside, giant planets with massive cores outside.

The snow line is not fixed. Its position depends on the disc’s opacity, how fast gas is accreting onto the star and how much heat is generated inside the disc, and it moves inward as the disc ages and the young star dims or brightens. Snow lines exist for other ices too, such as carbon monoxide and carbon dioxide, each at its own condensation temperature. Chemical differences between bodies formed on either side of these lines are one of the main tools for reconstructing where things formed.

Meteorites preserve a striking version of this divide. Isotopic measurements split them into two families, “non-carbonaceous” from the inner disc and “carbonaceous” from the outer disc, which apparently stayed separated for millions of years. Thomas Kruijer and colleagues argued in 2017 that the separation was caused by Jupiter itself: its core reached about 20 Earth masses within the first million years and about 50 Earth masses by 3–4 million years, opening a gap that kept the two reservoirs apart. On that reading, Jupiter is the oldest planet.

Sources: [9]

Interactive lab: where is the snow line?

Change the luminosity of the central star or the condensation temperature of ice and see how the snow line and the equilibrium temperature at each planet’s distance respond. Compare the present-day Sun, the fainter young Sun and a red dwarf.

With the default settings the calculator places the snow line at about 2.7 au, inside the present asteroid belt; for a red dwarf a hundred times fainter it moves in to about 0.27 au. As the notes under the calculator stress, this is a simple radiative-equilibrium estimate: in a real disc the snow line also depends on internal heating and opacity, and it migrates over time.

From dust to planetesimals: streaming instability and pebbles

Growing from micrometre dust to kilometre-sized planetesimals is the hardest step to explain. Grains stick together easily at first, but once they reach centimetre sizes collisions tend to make them bounce or break, and gas drag makes them spiral rapidly towards the star. One leading solution is the streaming instability, described by Andrew Youdin and Jeremy Goodman in 2005: because solid particles and gas orbit at slightly different speeds, their mutual drag can concentrate particles into dense filaments, which may then collapse under their own gravity directly into planetesimals.

The Kuiper Belt object Arrokoth supports a gentle origin of this kind. NASA’s New Horizons flew past it on 1 January 2019 and found a two-lobed body about 35 km long. The mission team concluded that the lobes formed close together and merged at low speed, consistent with formation inside a gravitationally collapsing cloud of solid particles rather than through violent collisions. Arrokoth has been preserved in the cold, quiet outer Solar System ever since.

Once planetesimals exist, they can grow quickly by sweeping up the centimetre-to-metre “pebbles” that drift through the disc. Michiel Lambrechts and Anders Johansen showed in 2012 that this pebble accretion, aided by gas drag, can shorten the growth time of giant-planet cores by large factors compared with growth from planetesimals alone, making it possible for cores to form before the gas disc disappears.

Sources: [5] [6] [7] [8]

Giant and terrestrial planets, and the first 100 million years

Beyond the snow line, sufficiently massive cores can capture large envelopes of hydrogen and helium from the disc, producing gas giants like Jupiter and Saturn; Uranus and Neptune, which grew later or in thinner gas, kept smaller envelopes. All of this had to happen within the few million years before the gas disc dispersed.

The terrestrial planets grew more slowly, through collisions between planetesimals and Moon- to Mars-sized embryos, over tens of millions of years. A late giant impact is thought to have formed the Moon: according to the widely accepted hypothesis, a Mars-sized body struck the young Earth and the debris coalesced in orbit. The precise date of that event is still debated.

Giant planets did not necessarily stay where they formed. In the Grand Tack model of Kevin Walsh and colleagues (2011), Jupiter migrated inward to about 1.5 au while the gas disc was present and then moved back out, pulled by Saturn; this would have truncated the inner disc of planetesimals at about 1 au, explaining Mars’ small mass and the mixed composition of the asteroid belt. In their simulations the terrestrial planets finished forming over the following 30–50 million years. The Grand Tack is a model, not an observation, and competing scenarios also reproduce a small Mars.

Sources: [10]

The Nice model and the bombardment debate

A second group of models deals with a later rearrangement of the outer planets. The Nice model, named after the city of the Côte d’Azur Observatory where it was developed and published in three Nature papers in 2005, proposes that after a long stable period the giant planets’ orbits became unstable and they migrated rapidly, scattering planetesimals throughout the Solar System. Gomes and colleagues linked this event to a Late Heavy Bombardment, an impact spike on the Moon about 700 million years after the planets formed.

The bombardment itself is now contested. Its main evidence was a clustering of impact-melt ages around 3.9 billion years in Apollo lunar samples. Patrick Boehnke and Mark Harrison showed in 2016 that such peaks in argon–argon age histograms can arise from partial argon loss in rocks reheated by later impacts, so that a smoothly declining impact rate could produce an apparent spike. Many researchers now favour either an earlier giant-planet instability or a gradually declining bombardment, and the question remains open.

Whatever the timing, the leftovers of planet formation are still with us. Planetesimals that never became planets survive as the asteroids between Mars and Jupiter, as the icy bodies of the Kuiper Belt beyond Neptune, among them the dwarf planets Pluto and Eris, and as comets scattered into the distant Oort cloud by the giant planets. Their orbits record the migrations described above, and their compositions record conditions in the disc where they formed, which is why small-body missions and sample returns are central to this field.

Sources: [11] [12]

Evidence in hand: meteorites and asteroid samples

Meteorites are free samples, but they are altered by atmospheric entry and contamination on Earth. Sample-return missions remove those doubts. JAXA’s Hayabusa2 collected material from the carbonaceous asteroid Ryugu and brought it to Earth. The initial chemical analysis team reported in 2022 that Ryugu closely resembles the rare CI (Ivuna-type) chondrites; its minerals were altered by liquid water about 5 million years after the Solar System formed, at around 40 °C, and the material has not been heated above about 100 °C since. JAXA described the samples as among the most primitive natural material available for study.

NASA’s OSIRIS-REx delivered samples of the asteroid Bennu to Earth in 2023. Results published in January 2025 in Nature Astronomy and Nature reported 14 of the 20 amino acids used by life on Earth to build proteins, all five nucleobases used in DNA and RNA, and 11 minerals that form when salty water evaporates, including calcite, halite, sylvite and trona. The findings show that Bennu’s parent body hosted brines and that the raw ingredients of biochemistry were widespread in the early Solar System. They are not evidence of life. For the wider family of small bodies see our small bodies page.

Sources: [13] [14]

Current research (2024–2026)

• The first solids in another system (2025). In July 2025 a team led by Melissa McClure (Leiden University) reported in Nature that JWST and ALMA had caught the protostar HOPS-315, about 1,300 light-years away, at the moment its first solids condense: silicon monoxide gas and crystalline silicate minerals just beginning to solidify, in a region comparable to the Sun’s asteroid belt. It offers a direct analogue of the CAI era.
• Bennu’s brines and organics (2025). The January 2025 papers on Bennu, based on material never exposed to Earth’s environment, showed that salty water and a broad set of prebiotic molecules existed together on a primitive parent body, sharpening questions about how such ingredients were delivered to the young Earth.
• New Horizons (2025–2026). The spacecraft that visited Arrokoth remains active in the Kuiper Belt; NASA reports that it entered a long hibernation in August 2025 and has since woken up in good health.

Sources: [4] [6] [14]

Known vs open questions

Well established:

• The Sun and planets formed together from a disc about 4.57 billion years ago; the oldest solids are dated to ±0.16 million years.
• Planet-forming discs with rings and gaps are common around young stars.
• Rocky planets formed inside the snow line and giant planets beyond it; primitive asteroids preserve water-altered, organic-rich material.

Still open:

• Exactly how centimetre pebbles become planetesimals in real discs, and how important the streaming instability is.
• Whether, when and how far the giant planets migrated (Grand Tack, Nice model and alternatives are models, not observations).
• Whether a distinct Late Heavy Bombardment occurred, and the precise date of the Moon-forming impact.
• How typical the Solar System’s architecture is compared with the exoplanet systems now known.

How to go further: data archives

Hayabusa2 sample information and curation data are published through JAXA’s DARTS Ryugu sample dataset. Spacecraft data on asteroids, comets and Kuiper Belt objects, including New Horizons results, are archived by NASA’s Planetary Data System Small Bodies Node, and JWST observations of young discs are in MAST. Our research dossier on planet formation discusses how disc observations and models are compared.

Related on this site

• Research dossier: Planet formation
• Small bodies: asteroids and comets
• Dwarf planets
• Exoplanets
• Stars and stellar evolution
• Interstellar objects
• Orbit lab