The Frost Line

A line where the disk turns to ice

Picture a newborn star surrounded by a flat, rotating disk of gas and dust — the protoplanetary disk it will spend a few million years dispersing. Close to the star the disk is hot, hundreds to thousands of kelvin; far out it is bitterly cold, only a few tens of kelvin. Somewhere in between is a special radius where the temperature crosses the value at which water can no longer stay a vapour. Inside that radius, water exists only as gas. Outside it, water vapour freezes onto dust grains and coats them in a mantle of ice. That radius is the frost line — also called the snow line or ice line.

The frost line matters far out of proportion to how thin it is, because of one fact about cosmic chemistry: oxygen is the third most abundant element in the universe, after hydrogen and helium, and a huge fraction of it ends up bound into water. Inside the frost line that water is locked in vapour and adds nothing you can build a planet out of. Cross the line and the water suddenly precipitates as solid ice, and the budget of condensed, planet-building material jumps. The frost line is, in effect, the place where a young planetary system stops being able to make only rock and starts being able to make rock-plus-ice — and that single difference shows up in the architecture of essentially every planetary system, including our own.

The physics: condensation in a low-pressure disk

The location of a frost line is set by where the local disk temperature equals the condensation temperature of the volatile in question. For a species in equilibrium between vapour and solid, the partial pressure of the vapour just balances the saturation vapour pressure of the ice, which depends exponentially on temperature through the Clausius-Clapeyron relation:

P_sat(T) = P_0 · exp( − L_sub / (R T) )

Here L_sub is the latent heat of sublimation (for water ice, about 51 kJ/mol), R is the gas constant, and P_0 is a reference pressure. Condensation happens where the actual partial pressure of water vapour in the disk exceeds P_sat(T) — that is, where it is cold enough that ice no longer sublimates faster than it forms.

The crucial subtlety is the pressure. We are used to water freezing at 273 K, but that is the value at one atmosphere. The midplane of a protoplanetary disk has gas pressures of only nanobars to microbars — roughly a billion times thinner than Earth's surface air. At those pressures the vapour-solid balance for water shifts down to a condensation temperature of about 150 to 170 K. (The precise figure depends on the assumed water abundance and the local gas density, which is why model values scatter over a ~20 K range.)

To turn a temperature threshold into a distance, you need the disk's temperature profile. A passively heated, flat disk reradiating absorbed starlight follows approximately

T(r) ≈ 280 K · (L_⋆ / L_⊙)^(1/4) · (r / 1 AU)^(−1/2)

so temperature falls off as the inverse square root of distance. Setting T(r) equal to the condensation temperature and solving for r gives the frost-line radius. In a real disk, viscous heating from accretion adds to this and dominates in the inner regions during the early, high-accretion phase — which pushes the frost line outward when the star is young and rapidly accreting, then lets it migrate inward as the accretion rate decays.

The key numbers

For water in the early solar nebula the standard placement of the frost line is near 2.7 AU — in the outer asteroid belt, comfortably between Mars at 1.5 AU and Jupiter at 5.2 AU. The most consequential number, though, is the size of the jump in solid surface density Σ at the line. The mass available as condensed solids depends on the elemental abundances; adding water ice to the rock and metal that condense inside the line increases the solid surface density by a factor of roughly

Σ_solid(outside) / Σ_solid(inside) ≈ 2 to 4

with the exact value depending on the assumed oxygen and carbon abundances and how much oxygen is sequestered into CO versus H₂O. Below is a rough ranking of how much condensable mass the major species contribute and where they freeze.

Volatile	Condensation T (disk)	Frost line (solar nebula)	Role
Silicates / iron	~1300–1500 K	≲ 0.5 AU	Rock; condenses almost everywhere
Water (H₂O)	~150–170 K	~2.7 AU	The dominant ice; doubles solids
Carbon dioxide (CO₂)	~70 K	~10 AU	Adds carbon-bearing ice
Carbon monoxide (CO)	~20–25 K	~30 AU	Sets gas-phase C/O ratio
Nitrogen (N₂)	~20 K	~30–40 AU	Most volatile; freezes only far out

The timescale matters too. Protoplanetary disks live only about 3 to 10 million years before photoevaporation and accretion clear the gas. In the core-accretion picture a giant-planet core must reach roughly 10 Earth masses within that window to trigger runaway gas capture. The extra ice beyond the frost line is what makes hitting that mass in time plausible — which is the quantitative heart of why the frost line and the giant planets are linked.

How the frost line is observed and located

For our own system, the asteroid belt is a fossil record of the frost line. The inner belt is dominated by dry, rocky S-type asteroids; the outer belt is rich in carbon-and-water-bearing C-type asteroids. The transition between them, around 2.5 to 3 AU, traces roughly where ice could survive in the young nebula. Carbonaceous chondrite meteorites, fragments of those outer-belt bodies, carry several percent water by mass locked in hydrated minerals — direct samples of material that condensed beyond the line.

In other planetary systems, frost lines are now seen directly. The headline case is V883 Orionis: in 2016, ALMA caught its water frost line pushed out to roughly 40 to 100 AU because an FU Orionis-type accretion outburst had heated the disk and sublimated ice across a broad region, making the boundary imageable. More commonly the outer CO frost line is mapped indirectly — molecules such as N₂H⁺ and DCO⁺ become abundant only where CO has frozen out of the gas, so a ring of their emission marks the line. This was used to place the CO snow line in disks like TW Hydrae (around 30 AU) and HD 163296. These detections confirm that nested ice lines are a genuine, resolvable feature of real disks.

Worked example: locating the water frost line

Suppose we want the water frost line of a passively heated disk around a solar-luminosity star, taking the condensation temperature to be 160 K. Start from the temperature profile:

T(r) = 280 K · (L_⋆ / L_⊙)^(1/4) · (r / 1 AU)^(−1/2)

With L_⋆ = L_⊙ the prefactor is just 280 K. Set T = 160 K and solve for r:

160 = 280 · (r / 1 AU)^(−1/2)
(r / 1 AU)^(1/2) = 280 / 160 = 1.75
r = 1.75² AU ≈ 3.1 AU

So a purely passive disk puts the line near 3 AU — close to, but a bit beyond, the canonical 2.7 AU. Now ask what happens for a faint young star at one tenth the Sun's luminosity, L_⋆ = 0.1 L_⊙. The prefactor scales as L^(1/4):

prefactor = 280 · (0.1)^(1/4) = 280 · 0.562 = 157 K
160 = 157 · (r / 1 AU)^(−1/2)
(r / 1 AU)^(1/2) = 157 / 160 = 0.98
r ≈ 0.97 AU

The frost line collapses inward to under 1 AU around a dim star — a direct consequence of the weak L^(1/4) dependence: a tenfold drop in luminosity moves the line by only a factor of about 1.8 in temperature scaling, but because temperature goes as r^(−1/2) that translates into the line shrinking from ~3 AU to ~1 AU. This is why low-mass M dwarfs, which are intrinsically faint, have frost lines snuggled up close to the star — with real consequences for where ice-rich planets can form around them.

Discovery and the people behind it

The idea grew out of the condensation-sequence work of the 1970s. Lewis (1972) and the equilibrium-condensation calculations of the era worked out which compounds condense at which temperatures in a cooling solar nebula, naturally predicting a radial sequence from refractory rock near the Sun to volatile ices far out. The frost line emerged as the boundary in that sequence where water — the most abundant condensable after the rock-formers — could freeze.

The concept became central to planet formation through the core-accretion model, articulated by Pollack and colleagues (1996) and rooted in earlier work by Safronov, Hayashi, Mizuno, and Stevenson in the 1970s and 80s. Hayashi's (1981) "minimum-mass solar nebula" explicitly built in a surface-density jump at the snow line near 2.7 AU to account for the extra solids needed to grow Jupiter's core. The recognition that the line moves — outward when accretion heating is strong, inward as the disk fades, modelled in detail by authors such as Garaud & Lin (2007) and in viscous-evolution studies through the 2000s and 2010s — turned the frost line from a fixed radius into a dynamic, time-evolving boundary. The 2016 ALMA imaging of the V883 Ori line and the molecular-tracer mapping of CO snow lines in TW Hya and HD 163296 moved it from theory into directly observed astrophysics.

Nested frost lines and related boundaries

• The CO₂ and CO snow lines. Beyond the water line, carbon dioxide freezes near 10 AU and carbon monoxide near 30 AU. Because CO carries much of the disk's carbon, where it freezes out controls the gas-phase carbon-to-oxygen (C/O) ratio at each radius — a quantity that planets imprint on their atmospheres. A giant planet that accretes gas between the water and CO lines inherits gas that is oxygen-poor and carbon-rich, a link Öberg, Murray-Clay & Bergin (2011) proposed as a way to read a planet's formation location from its atmospheric C/O.
• The soot or tar line. An inner boundary, around 300–500 K, where refractory carbon-rich organics are destroyed; it helps explain why the inner solar system, including Earth, is carbon-poor relative to cosmic abundances.
• Frost-line traffic jams. Ice-coated grains drifting inward sublimate as they cross the water line, releasing vapour that can recondense and pile up; this "cold finger" or pile-up effect is invoked to concentrate solids and help trigger the streaming instability that forms planetesimals.
• The habitable zone. Often confused with the frost line, but it is a different boundary set by a planet's surface conditions (liquid water under an atmosphere), located near 1 AU for the Sun rather than at 2.7 AU.

Common misconceptions and subtleties

• "The frost line is at 0 °C." No — at the disk's nanobar pressures water condenses near 150–170 K, far below 273 K. Freezing temperature is pressure-dependent, and disk pressures are roughly a billion times lower than at Earth's surface.
• "There's only one frost line." Each volatile has its own. The unqualified term means the water line because water dominates the condensable mass, but the CO₂ and CO lines are equally real and shape planetary chemistry.
• "The frost line is fixed in place." It migrates. Early, strong accretion heating pushes it outward; as the disk cools and the accretion rate decays over millions of years, it sweeps inward — possibly to ~1–2 AU by the time the gas clears.
• "The frost line is where life can exist." That conflates it with the habitable zone. The frost line is about ice condensing on dust during planet formation, not about a planet's surface being warm enough for oceans.
• "Earth formed beyond the frost line because it has oceans." Earth almost certainly formed inside the line, dry, and was supplied water afterward by icy and hydrated bodies scattered inward from beyond it — the frost line is upstream of Earth's water, not its birthplace.