Exoplanets
Boil-Off Escape: How a Newborn Planet Sheds Its Primordial Envelope
Within roughly 100,000 years of its birth disk evaporating, a young sub-Neptune can blow away more than 90% of the hydrogen-helium atmosphere it spent millions of years accreting — a violent, spontaneous outflow driven not by fresh stellar X-rays but by the simple removal of the gas that was holding it in. This is boil-off (also called spontaneous mass loss): a brief hydrodynamic escape phase that a low-mass planet undergoes the moment its protoplanetary disk disperses and the confining ambient pressure vanishes.
First worked out in detail by James Owen and Yanqin Wu in 2016, boil-off explains why newborn planets are hugely inflated — up to ~10 R⊕ at 20–30 Myr — yet mature into the compact ~2–3 R⊕ sub-Neptunes and stripped ~1.5 R⊕ super-Earths that dominate the Kepler census. It sets the initial conditions for everything that follows.
- RegimeLow-mass planets (3–10 M⊕ cores) at disk dispersal
- Driven byLoss of disk confining pressure → hydrostatic disequilibrium
- Key criterionRp ~ RB = GMp/cs²; mass loss stalls at Rp ~ 0.1 RB
- Timescale≲1 Myr, set by inner-disk dispersal ~10⁵ yr
- Mass lost≳90% of primordial H/He envelope
- First describedOwen & Wu 2016 (ApJ 817, 107)
- Observed withKepler/K2/TESS young transiting planets; radius valley at ~1.8 R⊕
- Matters forSub-Neptune vs super-Earth demographics; radius valley initial conditions
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
What boil-off is and why it matters
A giant planet's low-mass cousin — a rocky-icy core of a few Earth masses — grows inside the protoplanetary disk by pulling in hydrogen and helium until it holds an envelope of roughly 10–30% of its mass. While the disk is present, the surrounding gas presses inward and confines that envelope in hydrostatic balance. Boil-off is what happens the instant that support is withdrawn: as the inner disk clears, the ambient pressure at the planet's outer boundary plunges toward vacuum, the puffed-up envelope can no longer stay bound, and it pours off in a transonic hydrodynamic wind.
The phase lasts under a million years but is decisive. A planet that accreted a fat envelope emerges from boil-off with only ~1% of its mass in gas, or none at all. Because this sets the atmosphere every planet starts its long life with, boil-off is the true zero point of exoplanet thermal evolution — the initial condition for photoevaporation, core-powered mass loss, and the observed split between sub-Neptunes and super-Earths.
The mechanism, step by step
The controlling scale is the Bondi radius, R_B = GM_p / c_s², where c_s = √(k_B T_eq / μ m_H) is the isothermal sound speed of gas heated by the star. R_B marks where thermal energy equals gravitational binding energy. A newborn planet is inflated to a size of order its Bondi radius, R_p ~ R_B, because it accreted while embedded in warm disk gas.
While the disk confines it, the atmosphere sits in equilibrium. Remove the confining pressure and hydrostatic balance is impossible: the only steady solution is a Parker-type wind, transonic outflow whose sonic point sits at r_s = GM_p / 2c_s² = R_B/2. Gas streams outward, carrying away mass and — crucially — internal energy, so the planet also cools and contracts rapidly. The mass-loss rate, Ṁ = 4πR_p² ρ c_s ℳ, is exponentially sensitive to R_p/R_B, so as the envelope shrinks the escape self-limits. Stellar bolometric light keeps the wind roughly isothermal, but the energy for escape comes mainly from the envelope's own contraction (binding-energy release), not from photons.
Characteristic numbers, scales, and the shut-off criterion
The escape runs away until the planet has contracted to about R_p ~ 0.1 R_B. Because Ṁ falls off roughly as exp(3 R_p/R_B), the mass-loss timescale skyrockets once the planet reaches ~10% of its Bondi radius, and boil-off effectively switches off there. This shut-off radius is only logarithmically sensitive to model details, which is why the outcome is so robust.
Typical parameters: cores of 3–10 M⊕ at 0.1–1 AU, with equilibrium temperatures ~500–900 K. A 5 M⊕ planet can stay inflated near R_B after disk removal only if its envelope fraction exceeds ~10% (a 3 M⊕ core needs ~5%); cores above ~10 M⊕ instead trigger runaway gas accretion and become giants. During boil-off, outward advection of heat exceeds radiative cooling by ~100×, so the planet's Kelvin–Helmholtz time jumps to ~10⁷–10⁸ yr — it becomes 'a planet that is young in age but looks old.' Final envelope fractions cluster near ~1%, and after Gyr of cooling radii converge toward ~2.5 R⊕.
How it is detected
Boil-off itself is too fast (≲1 Myr) to catch a single planet mid-outflow, so it is inferred through two observational fingerprints. First, the inflated radii of young transiting planets. Systems like V1298 Tau (a ~20–30 Myr solar analog in Taurus with four transiting planets, several near ~10 R⊕, found in Kepler's K2 extended mission) and DS Tuc A show exactly the swollen sizes that thermal-evolution models predict before contraction completes. JWST and ground-based spectroscopy (e.g. metal-poor, haze-free atmospheres on V1298 Tau b) further constrain their still-hot, evolving interiors.
Second, and more powerfully, the radius valley — the deficit of planets near ~1.8 R⊕ (the Fulton gap) sharply mapped by Kepler, refined with K2 and TESS radial-velocity follow-up. Boil-off supplies the initial envelope masses that later escape carves into the observed bimodal split between ~1.4 R⊕ bare cores and ~2.4 R⊕ sub-Neptunes. Owen (2020) also showed young mass+radius+age measurements can bound a planet's formation entropy through how much envelope it could retain against escape.
Where it operates and how it differs from related escape
Boil-off is universal for close-in low-mass planets: essentially every planet born with an H/He envelope inside ~1 AU passes through it at disk dispersal. Beyond ~1 AU around Sun-like stars the gas is cool enough (~200–250 K for 5–10 M⊕) that the isothermal Parker-wind picture breaks down and loss becomes far less efficient, so distant planets keep their primordial atmospheres.
It is important not to conflate boil-off with the two escape channels it precedes. XUV photoevaporation is driven by the star's high-energy photons over the first ~100 Myr; core-powered mass loss is powered by the planet's own leftover interior heat plus bolometric light over ~Gyr. Boil-off, by contrast, is triggered purely by the disappearance of the confining disk, acts in under 1 Myr, and is over before either of the others begins in earnest. In the language of the field, boil-off sets the initial conditions; photoevaporation and core-powered loss finish the job of excavating the radius valley.
Open questions and significance
Boil-off reframed exoplanet evolution: the atmospheres we see today are not what planets accreted but what survived a violent infancy. Its biggest payoff is demographic — it naturally funnels a wide range of birth envelope masses toward the ~1% remnant that, after cooling, reproduces the observed radius distribution without fine tuning.
Key open problems remain. How exactly does boil-off hand off to core-powered mass loss and photoevaporation — do they add coherently, and which dominates the final radius valley? (Recent work argues core-powered loss leaves 'minimal long-lived' loss once boil-off is accounted for, still debated.) The transition from subsonic, disk-coupled 'breeze' outflows to fully transonic escape depends sensitively on the disk dispersal timescale, which is poorly constrained observationally. Radiation-hydrodynamic simulations are now mapping the bolometric-to-photoevaporative transition in real time. And the role of magma-ocean outgassing, which can replenish H/He dissolved in a molten core after boil-off strips the surface, is an active frontier for interpreting sub-Neptune interiors.
| Property | Boil-off (spontaneous) | XUV photoevaporation | Core-powered mass loss |
|---|---|---|---|
| Trigger / energy source | Loss of disk confining pressure; bolometric heating + envelope contraction | Stellar X-ray/EUV (XUV) photon flux ionizing/heating the upper atmosphere | Planet's residual internal (core) heat + bolometric irradiation |
| When it acts | At disk dispersal (planet age ~1–10 Myr) | First ~100 Myr, while stellar XUV is high | ~0.1–1 Gyr timescales |
| Duration | ≲1 Myr (very brief) | ~10⁸ yr | ~10⁹ yr |
| Fraction of envelope removed | Up to ≳90% of primordial mass | Can strip remaining few % down to bare core | Comparable; strips low-mass remnant envelopes |
| Key discoverers | Owen & Wu (2016) | Lopez, Owen, Jackson, Wu (2012–2017) | Ginzburg, Schlichting, Sari (2016–2018) |
| Role | Sets initial conditions for the others | Carves radius valley (XUV route) | Carves radius valley (thermal route) |
Frequently asked questions
What actually triggers boil-off?
The disappearance of the protoplanetary disk. While the planet is embedded, disk gas presses in on its atmosphere and holds it in hydrostatic balance. When the inner disk clears — on a timescale of only ~10⁵ years — that confining pressure drops toward zero, hydrostatic equilibrium becomes impossible, and the inflated envelope escapes as a transonic Parker-type wind. No new energy input is needed; removing the support is enough.
How is boil-off different from photoevaporation?
They differ in energy source and timing. Photoevaporation is driven by the star's X-ray and extreme-UV photons heating and ionizing the upper atmosphere, and it operates over the first ~100 Myr after the disk clears. Boil-off is driven by the loss of disk confining pressure plus the planet's own contraction, and it happens in under 1 Myr, right at disk dispersal — before photoevaporation gets going. Boil-off sets the atmosphere that photoevaporation then works on.
What is the Bondi radius and why does it matter here?
The Bondi radius R_B = GM_p/c_s² is the distance from a planet where the thermal energy of star-heated gas equals its gravitational binding energy. A newborn planet is inflated to roughly R_B. Because the mass-loss rate depends exponentially on the ratio R_p/R_B, boil-off runs away while R_p is near R_B and shuts off sharply once the planet contracts to about 0.1 R_B.
How much atmosphere does a planet lose to boil-off?
A great deal — typically more than 90% of the primordial hydrogen-helium envelope. Planets born with envelopes of ~10–30% of their mass emerge from boil-off with only about 1% (or, for the lightest cores, essentially none). This is why the phase is so important: it largely determines whether a planet ends up a sub-Neptune or a stripped super-Earth.
Can we see boil-off happening?
Not in a single planet mid-outflow — the phase is too brief (≲1 Myr). Instead it is inferred. Young transiting planets like those around V1298 Tau (~20–30 Myr) and DS Tuc A are observed to be hugely inflated (up to ~10 R⊕), exactly as pre-contraction models predict. And on population scales, the radius valley near ~1.8 R⊕ seen by Kepler and TESS reflects the initial envelope masses that boil-off, then later escape, sculpted.
Does boil-off explain the radius valley by itself?
Not entirely. Boil-off sets the initial envelope masses that planets start with after disk dispersal, but the sharp gap near ~1.8 R⊕ between bare ~1.4 R⊕ super-Earths and ~2.4 R⊕ sub-Neptunes is finished by later escape — XUV photoevaporation over ~100 Myr and/or core-powered mass loss over ~Gyr. Which of those two dominates is still actively debated; boil-off provides the common starting point for both.