What Is a Protoplanetary Disk and How Do Planets Form?

Four and a half billion years ago, a cloud of interstellar gas and dust collapsed under its own gravity. Most of the material fell to the center and ignited as the Sun. But not all of it. Conservation of angular momentum — the same physics that makes a spinning skater speed up as they pull in their arms — forced the remaining material into a thin, rotating disk. From this disk emerged everything that orbits the Sun today: eight planets, hundreds of moons, millions of asteroids and comets, and the countless smaller bodies of the Kuiper Belt. This process, protoplanetary disk formation and planet formation, is one of the most consequential sequences of events in cosmic history.

Today, observatories like ALMA (the Atacama Large Millimeter/submillimeter Array) and JWST are imaging protoplanetary disks around other stars in extraordinary detail, catching planetary systems in the act of assembly. The results are reshaping theories developed over decades — and raising new questions about how common planets like Earth truly are.

The Starting Point: Molecular Clouds and Collapse

Stars — and their accompanying planet systems — begin in giant molecular clouds (GMCs): vast, cold (10–50 K), dense accumulations of molecular hydrogen (H₂), helium, and trace amounts of heavier elements in atomic, molecular, and solid (dust) form. GMCs range from 10⁴ to 10⁷ solar masses and span tens to hundreds of light-years.

Within GMCs, turbulence, magnetic fields, and self-gravity constantly compete. Dense cores — cloud subregions of 1–10 solar masses, temperatures of 10 K, sizes of ~0.1 parsec — are the sites where gravitational collapse begins. Collapse is triggered by:

• Loss of magnetic support (ambipolar diffusion): neutral gas slowly drifts inward despite magnetic pressure from ionized gas
• External compressions: shockwaves from nearby supernova explosions or spiral arm density waves
• Turbulent dissipation: as turbulent support decays, gravity wins

Once a core begins to collapse, it proceeds on the free-fall timescale:

$$t_{ff} = \sqrt{\frac{3\pi}{32G\rho}} \approx 10^5 \text{ yr} \left(\frac{n}{10^4 \text{ cm}^{-3}}\right)^{-1/2}$$

For a typical dense core, free-fall takes ~100,000 years.

From Cloud Core to Protostar: Angular Momentum Wins

A collapsing cloud core cannot fall directly to a point. Even a tiny initial rotation — inevitable in a turbulent, non-symmetric interstellar cloud — is amplified enormously as the gas contracts. If angular momentum (L = mvr) is conserved, and the cloud shrinks by a factor of 10,000 in radius, its rotation rate must increase by 10,000². A cloud rotating once per million years becomes a protostar spinning many times per second — unless something removes angular momentum.

The resolution is the protoplanetary disk. Gas cannot fall directly to the center; instead, it settles into a rotating disk around the central protostar, where the centrifugal force (resisting inward fall) balances gravity in the disk plane. The disk itself can then slowly transfer angular momentum outward via:

• Magneto-rotational instability (MRI): coupling between ionized gas and magnetic fields generates turbulence that transports angular momentum outward
• Gravitational torques: spiral density waves in massive disks transfer angular momentum
• Disk winds: magnetic field threading through the disk drives jets and outflows that carry away angular momentum

The protostar + disk system evolves through a sequence classified by the shape of its spectral energy distribution (SED):

• Class 0: The protostar is deeply embedded in its infalling envelope; essentially invisible in the optical. Strong submillimeter emission. Age: ~10,000–100,000 yr
• Class I: Still partly embedded, but bipolar outflows have cleared some material. Age: ~100,000–500,000 yr
• Class II (Classical T Tauri stars): The disk is no longer being fed by infalling material. The central star is visible but pre-main-sequence, actively accreting from the disk. This is the "protoplanetary disk" phase. Age: ~1–10 Myr
• Class III (Weak-line T Tauri stars): Most disk gas has been accreted, blown away, or incorporated into planetesimals. Only a tenuous debris disk remains. Age: ~10 Myr+

Protoplanetary Disk Structure

A typical Class II protoplanetary disk has:

• Radial extent: 10–200 AU (some extend to 500+ AU)
• Total mass: 0.001–0.1 M☉ in gas; dust mass ~1% of gas (the interstellar dust-to-gas ratio)
• Midplane temperature: 150–200 K at 1 AU, dropping to ~30 K at 50 AU
• Surface density profile: Σ ∝ r^(−1 to −1.5), with the inner disk denser
• Scale height: H/r ≈ 0.03–0.2, with the disk puffed up by thermal pressure (flared geometry)
• Snow lines: Radial locations where specific volatiles condense to ice — water ice at ~170 K (~2–3 AU in the solar nebula), CO₂ at ~70 K, CO at ~25 K. The water snow line is particularly important: beyond it, solid material is more abundant, explaining why the outer solar system has large, ice-rich planets

The disk midplane where planet formation occurs is relatively cold and dust-rich. The upper atmosphere is irradiated by the central star's UV radiation, creating a warm, ionized surface layer (the disk atmosphere). The midplane is protected from stellar radiation but may be warmed by viscous heating from accretion.

Stage 1: Dust Grain Growth to Pebbles

The first challenge of planet formation: go from sub-micron interstellar dust grains to kilometer-scale planetesimals. This is a journey across nine orders of magnitude in size, and it is the most poorly understood phase of the entire process.

Brownian motion causes sub-micron grains to collide and stick via van der Waals forces and electrostatic interactions. Laboratory experiments show that this works well up to millimeter sizes — collisions are gentle, and grains stick and form fluffy aggregates. But problems emerge at larger sizes:

The fragmentation barrier: When pebbles (millimeter to centimeter sizes) collide at velocities of meters per second (driven by turbulence-induced relative motions), they shatter rather than stick. For silicate grains, the fragmentation velocity is ~1 m/s; for ice grains beyond the snow line, it may be ~10 m/s. This creates a pile-up of particles at ~millimeter to centimeter sizes — the "pebbles" that modern planet formation theory has come to embrace.

The radial drift barrier: Pebbles in a disk are partially supported against gravity by gas pressure, but the gas itself rotates slightly slower than Keplerian speed (the pressure gradient provides partial centrifugal support). Pebbles, less pressure-supported, rotate at near-Keplerian speed and feel a headwind from the slower-rotating gas. This headwind robs them of angular momentum, causing them to drift inward — potentially spiraling from 50 AU to the star in as little as 100,000 years. The drift velocity peaks for meter-sized objects ("the meter-size barrier"), which drift at up to 50 m/s.

Stage 2: Streaming Instability and Planetesimal Formation

The mechanism that most successfully bridges the barrier between pebbles and planetesimals is the streaming instability (Youdin & Goodman 2005). Here is the key insight: the headwind experienced by pebbles also causes a back-reaction on the gas. In dense regions where the ratio of pebble density to gas density is high (near the midplane and near pressure maxima in the disk), pebbles and gas can enter a coupled state where pebbles concentrate into dense filaments. When the local pebble-to-gas density ratio exceeds ~1, gravitational collapse of these filaments proceeds rapidly, bypassing the fragmentation and drift barriers entirely.

The result: gravitational collapse of pebble filaments directly forms planetesimals of 100–1000 km size. This is consistent with observations of trans-Neptunian objects (like Arrokoth, visited by New Horizons), which show that Kuiper Belt objects preferentially have sizes of ~100 km and appear to have formed as contact binaries from gentle collapse rather than violent collision.

Pressure maxima in the disk — created by transitions in disk opacity, magnetohydrodynamic effects, or gaps opened by forming planets — are particularly effective pebble traps. The rings and gaps seen in ALMA disk images may mark the locations of pressure bumps where pebbles accumulate and planetesimal formation is enhanced.

Stage 3: Planetesimal Accretion — Building Planetary Embryos

Once planetesimals of ~100 km exist, conventional gravitational accretion takes over. Larger planetesimals grow faster because their gravitational cross-sections exceed their geometric cross-sections by a factor called gravitational focusing:

$$\sigma_{eff} = \sigma_{geom}\left(1 + \frac{v_{esc}^2}{v_{rel}^2}\right)$$

where v_esc is the planetesimal's escape velocity and v_rel is the relative velocity. Larger bodies have higher escape velocities and thus larger effective cross-sections — they accrete even faster. This runaway growth rapidly produces a few dominant planetary embryos (also called oligarchs or protoplanets) of Moon-to-Mars mass from a swarm of smaller planetesimals within ~10⁵ years.

Pebble accretion (Ormel & Klahr 2010; Lambrechts & Johansen 2012) dramatically accelerates this phase. An embryo several thousand kilometers in size can capture pebbles drifting inward with extraordinarily high efficiency: the pebble's trajectory is curved by the embryo's gravity while gas drag prevents it from escaping, leading to accretion rates many orders of magnitude higher than classical planetesimal accretion. Pebble accretion can grow a planetary embryo to Earth mass in ~1 Myr — far faster than previous estimates of 10–100 Myr.

Stage 4: Gas Giant Formation — The Race Against Disk Dispersal

Giant planets like Jupiter and Saturn must form before the disk gas dissipates — a process that typically takes 3–10 million years (inferred from the measured fraction of protoplanetary disks as a function of star cluster age). This creates a fundamental time constraint.

Core accretion model: A rocky/icy core grows by pebble and planetesimal accretion to a critical mass of 10–15 Earth masses. At this point, the core's gravity is sufficient to capture a massive gaseous envelope from the surrounding disk. Once the envelope mass equals the core mass, runaway gas accretion begins — the planet grows from ~20 M_Earth to a Jupiter mass in perhaps 10,000–100,000 years. The disk gap opened by the growing planet eventually starves it of gas, halting accretion.

Disk instability model: In massive, cold disks, gravitational fragmentation can directly produce giant planet-mass clumps on timescales of ~100 years. This may form the most massive directly-imaged giant planets (>10 Jupiter masses) found at large separations (>50 AU), where core accretion is too slow.

Jupiter's core likely formed beyond the water snow line (~3–5 AU in the early solar system), where the solid surface density was enhanced by ice condensation. Its formation may have happened remarkably fast — within the first 1–2 Myr — acting as a Grand Tack: Jupiter's growth opened a gap in the disk and then migrated inward to ~1.5 AU before being caught in a 3:2 resonance with the forming Saturn, after which both migrated outward to their current positions. This "Grand Tack" scenario explains the small mass of Mars and the depletion of the asteroid belt.

Stage 5: Terrestrial Planet Formation — The Violent Final Stage

After the giant planets form and the disk gas dissipates, the inner solar system is left with dozens of Mars-to-Earth-mass planetary embryos and a swarm of smaller planetesimals. What follows is a period of giant impacts lasting 10–100 million years — the most violent phase of terrestrial planet formation.

Embryos on increasingly excited orbits (stirred by the giant planets) collide and merge. These are not gentle encounters: impacts occur at velocities of several km/s, partially vaporizing both bodies and thoroughly mixing their compositions. The Moon is believed to be the product of a giant impact between the proto-Earth and a Mars-sized impactor (Theia) ~45 million years after the solar system formed, 4.51 billion years ago. Isotopic analyses of lunar samples confirm the Moon is largely made of Earth material, consistent with a grazing collision that ejected mantle material into orbit.

Mercury's high density — it is ~60% iron by mass compared to ~30% for Earth — may reflect a giant impact that stripped away most of its silicate mantle, leaving the iron core behind. Recent simulations suggest Mercury formed closer to its current composition through "accretion of efficiently differentiated planetesimals" without needing a stripping impact.

The Late Heavy Bombardment (or Late Accretion) at ~3.9–4.0 Ga — a period of elevated impact rates on the Moon recorded in crater statistics and lunar sample ages — may reflect the dynamical instability of the outer solar system (the "Nice Model") that scattered Neptune outward and destabilized the Kuiper Belt, sending a barrage of comets and asteroids into the inner solar system.

What ALMA Has Revealed

The Atacama Large Millimeter/submillimeter Array has transformed understanding of protoplanetary disks since its early science operations in 2012. Operating at millimeter and submillimeter wavelengths, ALMA detects thermal emission from dust grains in the cool disk midplane — the very region where planet formation occurs.

The landmark DSHARP survey (Disk Substructures at High Angular Resolution Project, 2018) imaged 20 protoplanetary disks around nearby young stars at ~35 milliarcsecond resolution (~5 AU at 140 pc). The results were stunning: virtually every disk showed rings, gaps, spirals, and asymmetries. The smooth, featureless disks of theoretical models were nowhere to be found.

These substructures are attributed to:

• Gaps opened by forming planets (even Earth-mass planets can produce detectable features)
• Snow line chemistry: Ice condensation and sublimation alter particle sizes and opacity
• Zonal flows: Magnetic pressure gradients create pressure maxima that trap pebbles into rings
• Gravitational instability spirals in the most massive disks

The implication is remarkable: planets begin forming very early — while the disk is still young and gas-rich — not after it has dispersed. Some disk systems show features at only 0.5–1 Myr old.

HL Tau (a ~1 Myr old star) was the first dramatic example: ALMA revealed a series of bright rings and dark gaps at 10–100 AU, suggesting planets were clearing gaps in the disk at an age when classical theory said only small planetesimals should exist.

What JWST Has Added

JWST probes protoplanetary disks at near- and mid-infrared wavelengths, sensing warm inner disk material and disk atmospheres. Key results:

• Inner disk water vapor: Mid-infrared spectroscopy of disks around low-mass M-dwarf stars (which are the most common star type) reveals abundant water vapor in the innermost disk regions (~0.1–1 AU). The terrestrial planet-forming zone of M-dwarf systems appears water-rich, potentially favorable for habitable planet formation.

• Carbon-to-oxygen ratio: JWST has measured C/O ratios in disk atmospheres, providing constraints on the composition of planets forming within these disks. Elevated C/O ratios suggest carbon-rich chemistry in some systems.

• Outflows and jets: JWST's infrared sensitivity images the jets and outflows from young protostars that carry away angular momentum, with unprecedented detail in the shocked gas emission.

The Diversity of Planetary Systems

The discovery of thousands of exoplanets has revealed that planetary systems are spectacularly diverse:

• Hot Jupiters (gas giants orbiting <0.1 AU): Form beyond the snow line via core accretion or disk instability, then migrate inward via Type II migration (planet opens a gap, migrates with the viscous flow). Occur in ~1% of solar-type stars.
• Super-Earths and mini-Neptunes (1–4 R_Earth, 2–20 M_Earth): The most common planet type in the galaxy, with no solar system analog. May form near their current orbits via pebble accretion, or migrate from beyond the snow line.
• Wide-separation giant planets (>10 AU): Directly imaged systems like HR 8799 challenge core accretion (too far, too slow); disk instability or scattering from closer orbits are proposed formation mechanisms.
• Compact multi-planet systems in tight resonances (like TRAPPIST-1): Likely formed further out and migrated inward through the disk, parking in resonance chains. TRAPPIST-1's seven planets are locked in a complex resonance chain that constrains their formation to have occurred in a gas-rich disk.

The solar system's architecture — with rocky planets close in, giant planets far out, and a relatively clean inner system — appears to be somewhat unusual. Most observed planetary systems have multiple super-Earths within 0.5 AU, suggesting Jupiter's formation and subsequent behavior (possibly the Grand Tack migration) gave our solar system a peculiar character.

Disk Dispersal: The End of Planet Formation (for Gas Worlds)

The disk gas must eventually disappear. Disk dispersal occurs via:

• Viscous accretion: Gas spirals inward and accretes onto the star
• Photoevaporation: UV and X-ray radiation from the central star (and neighboring massive stars) ionizes disk gas, launching a thermal wind at up to 10 km/s
• Giant planet gaps: Massive planets starve the inner disk of gas supply from the outer disk, accelerating inner disk clearing
• Disk winds: Magnetohydrodynamic winds driven by magnetic fields threading the disk surface efficiently remove disk mass

The combination of these processes typically disperses disk gas in 3–10 Myr — a relatively short window by cosmic standards. After gas dispersal, only solid bodies remain: rocky planets, planetesimal belts (asteroid and Kuiper Belt analogs), and debris disks of dust generated by ongoing collisions among planetesimals.

The Solar System's Own Disk: The Minimum Mass Solar Nebula

By adding enough hydrogen and helium to each planet to restore it to solar composition, and smearing that material into a smooth disk, one can reconstruct the minimum mass solar nebula (MMSN) — the minimum amount of solid-forming material the early solar disk must have contained:

$$\Sigma_{solid} \approx 7 \left(\frac{r}{1,\rm AU}\right)^{-3/2} \rm g,cm^{-2}$$

This gives a total disk mass of ~0.01–0.07 M☉, consistent with observed Class II disk masses. The MMSN assumes no material was lost to inward drift, giant planet gaps, or early dispersal — making it a lower limit. The true nebula may have been several times more massive.

Meteorites preserve chemical and isotopic records of the solar nebula's earliest moments. Calcium-aluminum-rich inclusions (CAIs) in carbonaceous chondrites are the oldest solar system solids, formed 4.5672 billion years ago within the first ~100,000 years of the nebula's existence. Chondrules — millimeter-sized silicate spheres in chondrites — formed slightly later (within 2–4 Myr) by brief high-temperature melting events whose cause (shock waves, protoplanetary impacts, lightning) remains debated.

---

Related Reading

• What Are Exoplanets and Could Any of Them Host Life?
• What Is Stellar Evolution? The Complete Life Cycle of Stars
• What Is Planetary Migration? How Planets Move in Solar Systems
• What Is the Habitable Zone and Why Earth Is In It?
• How Do We Detect Exoplanet Atmospheres and What Are Biosignatures?