White Dwarf Cooling

A star that runs entirely on leftover heat

A white dwarf is what remains when a star like the Sun finishes burning. It has shed its outer layers as a planetary nebula and left behind a bare, exposed core: about 0.6 solar masses of carbon and oxygen packed into a sphere roughly the size of Earth, 10⁴ kilometres across. The density is staggering — a cubic centimetre of white dwarf matter weighs about a tonne. But here is the crucial point: there is no fusion happening anywhere inside it. The fuel is gone.

So what holds it up, and why does it shine? It is held up not by heat but by electron degeneracy pressure — a quantum-mechanical floor set by the Pauli exclusion principle, which forbids electrons from occupying the same state and resists compression even at zero temperature. This decouples the star's structure from its temperature: the white dwarf can cool indefinitely without collapsing. And it shines for the simplest reason imaginable. It is hot — born at around 10⁸ K in the core — and hot things radiate. With no internal furnace to replace what it loses, the white dwarf is an ember the size of a planet, slowly leaking the heat it was born with into the cold of space. That leak is white dwarf cooling, and because the reservoir is finite, the star can only ever get fainter.

The Mestel cooling law

The foundational theory was written down by Leon Mestel in 1952, and it is beautiful in its economy. The thermal energy of the white dwarf is stored almost entirely in its ions, because the degenerate electrons are quantum-frozen near the bottom of their Fermi sea and contribute almost no heat capacity. Treating the ions as a classical ideal gas with heat capacity (3/2) k_B per ion, the total stored thermal energy is

U ≈ (3/2) (M / A m_u) k_B T_c

where M is the stellar mass, A is the mean ion mass number (≈14 for a C/O mix), m_u is the atomic mass unit, and T_c is the central temperature. Heat escapes through a thin, non-degenerate surface layer whose opacity (Kramers-type) makes the photon luminosity scale steeply with central temperature:

L ≈ C · M · T_c^(7/2)

Setting the luminosity equal to the rate of loss of stored energy, L = −dU/dt, and integrating gives Mestel's celebrated result for the cooling time:

t_cool ∝ (M / A)^(5/7) · L^(−5/7)
       →   L ∝ t^(−7/5)

The luminosity falls off as time to the minus seven-fifths. The consequence is profound: a white dwarf races through its hot, bright phase and then crawls through an enormously long faint phase. The brighter half of its visible life lasts only a few hundred million years; the dim tail toward invisibility lasts longer than the present age of the Universe.

The four acts of cooling

Mestel's clean power law is the backbone, but a real white dwarf passes through distinct regimes as its temperature drops. The full cooling curve has four acts.

• Neutrino-dominated youth. Immediately after birth, with T_c above ~10⁷·⁵ K, the core loses energy faster through neutrino emission (mainly plasmon decay) than through surface photons. Neutrinos stream straight out and carry away energy from the deep interior, cooling the star rapidly for the first ~10⁷–10⁸ years before photon cooling takes over.
• The Mestel plateau. The classical regime, L ∝ t^(−7/5), where ions behave as an ideal gas and most observed white dwarfs live. This is the long, steady, predictable middle.
• The crystallization stall. When the core cools enough for the ions to freeze into a lattice (see below), latent heat is released, temporarily propping up the luminosity and slowing the descent.
• The Debye plunge. After crystallization is complete, the lattice heat capacity collapses as C ∝ T³ (the Debye law). With almost no heat left to give, the white dwarf cools dramatically faster, plunging toward the black-dwarf state it will not reach for ~10¹⁵ years.

Crystallization: latent heat from a freezing star

The most spectacular twist in the cooling story is that the core literally freezes solid. The ions sit in a sea of degenerate electrons, and the strength of their mutual electrostatic repulsion relative to their thermal jostling is measured by the Coulomb coupling parameter:

Γ = (Z e)² / (a k_B T)        a = (3 / 4π n_i)^(1/3)

Here Z is the ion charge, e the elementary charge, a the ion-sphere radius, and n_i the ion number density. When the plasma is hot, Γ is small and the ions form a disordered fluid. As T falls, Γ climbs, and at Γ ≈ 175 the one-component plasma undergoes a first-order phase transition to a body-centred-cubic crystal lattice. Freezing releases latent heat of roughly 0.77 k_B T per ion. On top of that, as a carbon-oxygen mixture crystallizes the heavier oxygen preferentially settles toward the centre, and this gravitational separation releases an additional, even larger, reservoir of energy. Together these temporarily offset the cooling, producing a pause that can add a billion years or more to the inferred age at fixed luminosity.

This was theoretical for decades until Gaia's precise parallaxes and photometry made it visible. In 2019, Tremblay and collaborators reported a pile-up of white dwarfs on the colour-magnitude diagram — an over-density along a near-vertical strip nicknamed the "Q branch" — exactly where theory predicts crystallization stalls the cooling and causes white dwarfs to linger. It was the first direct observational confirmation that white dwarf cores really do freeze into crystals, decades after the prediction.

The key numbers

White dwarf cooling is anchored by a small set of well-measured physical quantities. The table collects the values that recur throughout the literature.

Quantity	Typical value	Note
Typical WD mass	0.6 M☉	Sharp peak in the mass distribution
Chandrasekhar limit	1.44 M☉	Maximum mass; degeneracy can't support more
Radius	~6000–9000 km	Earth-sized; smaller for higher mass
Mean density	~10⁶ g/cm³	A tonne per cubic centimetre
Birth core temperature	~10⁸ K	Cools without contracting
Hottest observed (e.g. central stars)	~150,000 K (T_eff)	Young, just unveiled from nebula
Crystallization onset Γ	≈ 175	BCC lattice forms in the core
Latent heat at freezing	~0.77 k_B T / ion	Plus gravitational O settling energy
Coolest disk WDs	~3000–4000 K	Cooling ages of 9–10 Gyr
Faint-end of WD luminosity function	L ≈ 10⁻⁴·⁵ L☉	Sets the disk age, ~8–10 Gyr

Two facts stand out. First, the white dwarf shrinks as it gains mass — a counter-intuitive mass-radius relation that is a direct fingerprint of degeneracy. Second, the cooling has produced no truly cold remnant anywhere: the coolest known white dwarfs are still around 3000 K, because the Universe simply has not existed long enough to make a black dwarf.

Worked example: how old is a 0.005 L☉ white dwarf?

Suppose we observe a 0.6 M☉ white dwarf and measure its luminosity at L = 0.005 L☉ (a typical disk white dwarf). We can estimate its cooling age from Mestel's scaling. A convenient calibration of the Mestel law, normalised so a 0.6 M☉ remnant reaches L = 10⁻² L☉ at about 6 × 10⁷ years, is

t_cool ≈ 6 × 10⁷ yr × (M / 0.6 M☉)^(5/7) × (L / 10⁻² L☉)^(−5/7)

For M = 0.6 M☉ the mass factor is 1. The luminosity ratio is L / 10⁻² L☉ = 0.5, so we need 0.5^(−5/7):

0.5^(−5/7) = 2^(5/7) = e^(0.714 × ln2) = e^(0.495) ≈ 1.64

t_cool ≈ 6 × 10⁷ yr × 1.64 ≈ 1.0 × 10⁸ yr

So this remnant has been cooling for roughly 100 million years. Now push the same formula to a much fainter object at L = 10⁻⁴·⁵ L☉ ≈ 3 × 10⁻⁵ L☉, near the faint end of the luminosity function:

(3 × 10⁻⁵ / 10⁻²)^(−5/7) = (3 × 10⁻³)^(−5/7)
                          = (333)^(5/7) ≈ 63

t_cool ≈ 6 × 10⁷ yr × 63 ≈ 4 × 10⁹ yr  (Mestel only)

The pure Mestel estimate gives ~4 Gyr, but the real cooling ages reach 9–10 Gyr once crystallization latent heat, gravitational settling, and the detailed surface physics are included — which is precisely why the simple power law underestimates the oldest ages and why full cooling models (e.g. by Fontaine, Brassard & Bergeron, or the La Plata and Montreal grids) are needed for precision cosmochronology. The lesson: the scaling law tells you the shape of the clock, but dating the Galaxy requires the full physics.

How we read the clock

To turn a white dwarf into a clock you need its luminosity and an age-luminosity relation. The luminosity follows from the surface temperature (from its colour or spectrum) and the radius (from the mass-radius relation, with the mass pinned by gravity-sensitive spectral line widths, by an astrometric companion, or by gravitational redshift). Astronomers then read the cooling age in two complementary ways.

• The cooling sequence of a star cluster. All the stars in a globular or open cluster formed at essentially the same time, so the faintest, oldest white dwarfs in the cluster mark how long the very first ones have been cooling. The terminus of the cooling sequence gives the cluster's white-dwarf age, independent of main-sequence turnoff fitting. Hubble imaged the cooling sequence of the globular cluster M4 in 2002 and derived an age of about 12.7 ± 0.7 billion years.
• The disk white-dwarf luminosity function. Count white dwarfs as a function of luminosity in the solar neighbourhood. Because cooling slows so dramatically, white dwarfs pile up toward faint luminosities — until you hit the very oldest ones, which abruptly cut off the count. That sharp downturn near L ≈ 10⁻⁴·⁵ L☉ dates the onset of star formation in the Galactic disk to about 8–10 billion years ago. This was pioneered by Winget and collaborators in 1987 and refined enormously by SDSS and Gaia.

History: from Sirius B to Gaia's crystals

The white dwarf story begins with a wobble. In 1844 Friedrich Bessel inferred an unseen companion tugging on Sirius from its proper-motion wiggle; in 1862 Alvan Graham Clark first saw it — Sirius B, the nearest white dwarf. Walter Adams measured its spectrum in 1915 and found a star as hot as Sirius A but a ten-thousandth as luminous, implying it had to be tiny and absurdly dense. Arthur Eddington called the conclusion "nonsense," but the physics held.

The structure was explained by R. H. Fowler in 1926, who applied the brand-new Fermi-Dirac statistics to show that electron degeneracy pressure supports the star. In 1930–31 Subrahmanyan Chandrasekhar added special relativity and discovered the mass ceiling — the Chandrasekhar limit of 1.44 M☉ — work that won him the 1983 Nobel Prize. The cooling theory itself came from Leon Mestel in 1952, who realised that a degenerate star is essentially a heat reservoir leaking through an insulating envelope and derived L ∝ t^(−7/5). Crystallization of the core was predicted by Abrikosov, Kirzhnits, and Salpeter around 1960–1961. Don Winget's group turned cooling into cosmochronology in 1987. And in 2019 the Gaia mission's parallaxes finally made the crystallization pile-up — the Q branch — directly visible, closing a 60-year loop between prediction and observation.

Variants: not all white dwarfs cool alike

• DA vs DB atmospheres. The cooling rate depends on the envelope's insulating power, which depends on composition. Hydrogen-atmosphere (DA) and helium-atmosphere (DB) white dwarfs have different opacities and therefore slightly different cooling tracks at the same mass and temperature. Convective coupling between the surface convection zone and the degenerate core, occurring around 10,000–6000 K, also briefly accelerates cooling.
• Pulsating white dwarfs (ZZ Ceti, DBV, DOV). In narrow temperature strips, white dwarfs pulsate. Asteroseismology of these g-mode oscillations probes the interior layering and the crystallized fraction directly, providing an independent check on cooling models. See asteroseismology.
• Helium-core and ultra-massive white dwarfs. Low-mass helium-core white dwarfs (from binary stripping) and ultra-massive O/Ne white dwarfs (above ~1.05 M☉) follow different cooling tracks. Ultra-massive remnants crystallize earlier (higher density) and figure prominently among the Q-branch population.
• Accreting white dwarfs. A white dwarf in a close binary is not left to cool quietly — it accretes, brightens, and may erupt as a cataclysmic variable or nova, or accumulate toward the Chandrasekhar limit and detonate as a Type Ia supernova. Cooling theory applies only to isolated remnants.

Common misconceptions and subtleties

• "A white dwarf contracts as it cools." No. Degeneracy pressure is nearly independent of temperature, so the white dwarf's radius is essentially fixed by its mass. It cools at constant size — unlike a normal star, which would shrink and reheat. This decoupling of pressure from temperature is the whole reason the cooling is so clean.
• "Cooling means it dims uniformly toward zero." No. The cooling rate is wildly non-uniform. The hot phase is over in a geological eyeblink; crystallization briefly stalls the descent; and the final Debye phase accelerates it again. The luminosity history is a multi-act curve, not a simple exponential.
• "Black dwarfs exist." Not yet. A black dwarf — a white dwarf cooled to invisibility — would take of order 10¹⁵ years to form, roughly 100,000 times the current age of the Universe. Every white dwarf in existence is still glowing.
• "The light comes from slow residual fusion." No. The carbon-oxygen core is below the ignition temperature and density for further fusion; the luminosity is purely stored thermal (mostly ionic) energy plus latent heat of crystallization. There is no nuclear source.
• "The electrons store the heat." Almost the opposite. The degenerate electrons are quantum-frozen and contribute very little specific heat; the heat reservoir is the non-degenerate ions. That is exactly why crystallizing those ions — changing their heat capacity — has such a dramatic effect on the cooling rate.