Neutron Star Cooling & the URCA Process

A star that cools by leaking, not by glowing

Pour boiling water into a thermos and it cools through the walls — heat conducts to the surface and radiates away as infrared light. A neutron star, freshly minted in a supernova at a hundred billion kelvin, cannot do this. Its interior is so dense — a few times nuclear saturation density, around 10¹⁴–10¹⁵ g/cm³ — that photons cannot escape; a photon born in the core would scatter for longer than the age of the universe before reaching the surface. If a neutron star could only cool the way a thermos does, it would stay blisteringly hot essentially forever.

Instead it cools the way no everyday object can: by emitting neutrinos. Deep in the core, weak-interaction beta reactions continuously create neutrinos, and because neutrinos barely interact with matter, they fly straight out of the star and into space — taking their energy with them. The star does not have to wait for heat to crawl to the surface; it bleeds energy directly out of its centre. For roughly the first hundred thousand years of a neutron star's life, this neutrino bleed completely dominates over the feeble light leaking from the surface. The headline reactions are collectively called the Urca process, and they are why a neutron star cools as fast as it does.

The Urca process: two beta reactions, two neutrinos

The Urca process is a cycle of two weak reactions that flip a nucleon back and forth between neutron and proton, emitting a neutrino each way:

n        →  p + e⁻ + ν̄ₑ     (beta decay)
p + e⁻   →  n + νₑ           (electron capture)
─────────────────────────────────────────────
net:  no change in composition, but TWO neutrinos escape

Run the cycle once and you are back where you started — the same number of neutrons and protons — but two neutrinos have left the star, each carrying away a chunk of thermal energy. The reactions are powered by the thermal motion of the degenerate particles, so the hotter the star, the faster the cycle and the faster the energy drain. The name is a famous in-joke: George Gamow and Mário Schenberg, working in Rio de Janeiro in 1941, named it after the Cassino da Urca, joking that the reaction drains energy from a star's core as relentlessly as the roulette wheel drained cash from the casino's patrons.

There is a catch that splits the process into two flavours. Inside a neutron star, neutrons, protons and electrons are all degenerate — only particles right at their Fermi surfaces can react. Conserving both energy and momentum among three Fermi-surface particles is geometrically demanding, and whether it is even possible depends on how many protons there are.

Direct Urca vs modified Urca

If the proton fraction is high enough that the three Fermi momenta can form a closed triangle (p_Fn ≤ p_Fp + p_Fe), the bare two-body reactions above proceed. This is direct Urca, and it is ferociously fast. The threshold is a proton fraction of roughly 11–15%, which only occurs above some critical density — and therefore only in the cores of the most massive neutron stars, where the equation of state allows it.

If the proton fraction is too low — as it is for ordinary ~1.4 M_☉ neutron stars — momentum cannot be conserved by three particles alone. Nature fixes this by adding a spectator nucleon that recoils to soak up the excess momentum:

n + n   →  n + p + e⁻ + ν̄ₑ
n + p + e⁻  →  n + n + νₑ

This is the modified Urca process. The extra particle in the reaction makes it slower by a large factor — roughly 10⁵–10⁶ — and changes its temperature dependence. The two emissivities scale very differently with temperature T (and so with the cooling clock):

Direct Urca:    Lν ∝ T⁶      (fast cooling — possible only above the proton-fraction threshold)
Modified Urca:  Lν ∝ T⁸      (standard cooling — works everywhere)

The steeper T⁸ law means modified-Urca cooling slows dramatically as the star cools — the neutrino faucet throttles itself. A star cooled by direct Urca, with its shallower T⁶ law and ~10⁶× higher prefactor, plunges in temperature so fast that within decades to centuries it is far colder than any modified-Urca star of the same age. Which channel a given neutron star uses is therefore one of the sharpest observational handles we have on the dense-matter equation of state inside it.

The cooling clock: from birth to old age

A neutron star's thermal life divides cleanly into two eras separated by a turnover near 10⁵ years.

• Proto-neutron-star phase (first ~minute). Immediately after core collapse the star is a hot, lepton-rich ball at internal temperatures of 10¹¹–10¹² K (tens of MeV). It deleptonises in ~10–20 s, releasing the famous ~3 × 10⁵³ erg of gravitational binding energy as neutrinos — the burst detected from SN 1987A by Kamiokande-II and IMB.
• Neutrino-cooling era (~10 s to ~10⁵ yr). Neutrino luminosity L_ν vastly exceeds the surface photon luminosity L_γ. The interior is nearly isothermal (held together by the high thermal conductivity of degenerate electrons), and Urca emission drives it from ~10¹⁰ K down to ~10⁸ K. The surface temperature, set by the thin insulating envelope, sits around a few × 10⁵–10⁶ K — neutron stars of this age glow in soft X-rays.
• Photon-cooling era (after ~10⁵ yr). Once the interior drops low enough, the throttled neutrino emission finally falls below surface photon emission, and ordinary blackbody radiation from the surface takes over. The star fades through the X-ray and then the optical/UV, reaching ~10⁵ K by ~10⁶ yr and a few × 10⁴ K for the oldest isolated neutron stars.

The transition is where the neutrino luminosity curve, falling steeply (T⁸ or T⁶), crosses the more gently declining photon luminosity. That crossover is the single most diagnostic feature of any theoretical cooling curve.

By the numbers: temperatures, timescales, energies

Phase / quantity	Value	Note
Birth (proto-NS) interior T	~10¹¹–10¹² K	tens of MeV; deleptonising
Binding energy released as ν	~3 × 10⁵³ erg	~10% of M c²; SN 1987A burst
SN 1987A ν burst duration	~10 s	~19 events at Kamiokande-II + IMB (~24 with Baksan)
Interior T at ~100 yr	~10⁹ K	modified-Urca cooling
Surface T during neutrino era	~(0.5–3) × 10⁶ K	soft-X-ray emitter, ~0.1–0.3 keV
Neutrino-to-photon turnover	~10⁵ yr	standard (modified-Urca) cooling
Surface T at ~10⁶ yr	~(2–6) × 10⁵ K	fading toward optical/UV
Direct-Urca proton-fraction threshold	~11–15%	only in massive-NS cores
Superfluid critical T	~10⁸–10⁹ K	opens PBF neutrino channel
Cassiopeia A NS age	~340 yr	fastest measured cooling

For orientation: a canonical neutron star packs ~1.4 M_☉ (2.8 × 10³³ g) into a radius of ~11–12 km, so its mean density exceeds 10¹⁴ g/cm³. That extreme density is exactly what makes the interior opaque to light and transparent to neutrinos — the two facts that together force the cooling to be neutrino-dominated.

Superfluidity: the plot twist that speeds it up

Below critical temperatures of order 10⁸–10⁹ K, the neutrons and protons in the core pair into Cooper pairs, just as electrons do in a terrestrial superconductor. The neutrons become a superfluid; the protons become a superconductor. Pairing opens an energy gap Δ at the Fermi surface, and this has two competing effects on cooling.

First, it suppresses the Urca rates and the nucleon heat capacity by a Boltzmann-like factor ~exp(−Δ/k_BT). A fully paired core is hard to cool through beta reactions, because the reacting particles are locked into pairs.

Second — and counter-intuitively — the very onset of pairing triggers a brief, intense burst of neutrino emission. As the star cools through the critical temperature, thermally broken pairs continually re-form, and each formation event can emit a neutrino–antineutrino pair. This pair-breaking-and-formation (PBF) process is a powerful transient cooling channel that switches on precisely when the temperature crosses T_c. The PBF burst is the leading explanation for the anomalously fast cooling seen in the youngest well-studied neutron star, in Cassiopeia A.

Cassiopeia A: watching a star cool in real time

The central compact object in the Cassiopeia A supernova remnant is the youngest known neutron star with a measured surface temperature, born in a supernova whose light reached Earth around 1680 — making it roughly 340 years old. Because it is so young and nearby (~3.4 kpc, ~11,000 light-years), the Chandra X-ray Observatory has been able to monitor its surface temperature over more than two decades.

Several analyses report a surprisingly rapid decline — of order a few percent in surface temperature over a decade. No ordinary modified-Urca star should cool that fast at 340 years old. The favoured interpretation is that the core is right now passing through the neutron superfluid transition: the PBF neutrino burst is switching on, dumping extra energy and producing a brief epoch of accelerated cooling. If correct, Cas A is direct, real-time evidence for neutron superfluidity at supranuclear density — a measurement no laboratory on Earth can reproduce. The exact rate remains debated (calibration of Chandra's detector matters at the percent level), but the case is the textbook example of cooling as a probe of dense matter.

Reading the cooling curve

The central tool of the field is the cooling curve: surface temperature (or luminosity) plotted against age. Theory predicts a family of curves — one for each assumed composition, mass, superfluid model, and envelope chemistry — and observers overlay measured neutron stars on top.

The picture splits broadly into standard (slow) cooling and enhanced (fast) cooling:

Scenario	Dominant ν channel	Emissivity	Surface T at 10³ yr	Where it applies
Standard / slow	Modified Urca	∝ T⁸	~10⁶ K	Typical ~1.4 M☉ stars
Enhanced / fast	Direct Urca	∝ T⁶	~3 × 10⁵ K	Massive cores above threshold
Exotic fast	Pion/kaon condensate, quark	∝ T⁶	~few × 10⁵ K	Hypothetical dense phases
Superfluid-modulated	PBF + suppressed Urca	transient burst	variable dip	Cas A-type young stars
Bremsstrahlung floor	n-n / n-p nucleon brems.	∝ T⁸	backup channel	When Urca is paired-out
Photon era	Surface blackbody	∝ Ts⁴	(post-10⁵ yr)	All old stars

A star that lands on the fast track at a given age is telling you its core is dense enough — and massive enough — to have crossed the direct-Urca threshold or to host an exotic phase. A star on the slow track has a more modest central density. Cooling, in other words, weighs the equation of state of matter we cannot otherwise reach.

Where neutron star cooling shows up

• Cassiopeia A neutron star. The ~340-year-old central compact object whose decade-scale temperature drop is the cleanest evidence for fast cooling and superfluid onset.
• Isolated thermal neutron stars (the "Magnificent Seven"). Nearby (~100–500 pc) middle-aged isolated neutron stars such as RX J1856.5−3754, detected purely by their thermal X-ray glow at surface temperatures of ~(0.5–1) × 10⁶ K. They are clean cooling-curve calibrators because they have no accretion or magnetospheric contamination.
• Soft X-ray transients in quiescence. Accreting neutron stars in binaries that, between outbursts, glow from heat deposited deep in the crust by nuclear reactions. Their quiescent luminosity tracks the long-term average accretion rate and tests crust-to-core thermal coupling — and some, like SAX J1808.4−3658, are so cold in quiescence that they require enhanced (direct-Urca) core cooling.
• SN 1987A. The ~10-second neutrino burst detected on 23 February 1987 was the deleptonisation of a newborn proto-neutron star — the only direct detection of the very first instant of neutron-star cooling, and the founding event of neutrino astronomy.
• Magnetars. Young neutron stars with 10¹⁴–10¹⁵ G fields whose surfaces are hotter than passive cooling predicts, because ohmic decay of the colossal magnetic field continually reheats the crust — a reminder that cooling competes with internal heating.

Common misconceptions and edge cases

• "The star cools by radiating light from its surface." Only after ~10⁵ years. For most of the dramatic early cooling, the surface photons are a tiny side-channel; the real energy loss is neutrinos pouring out of the core. The surface temperature you measure in X-rays is a proxy for the interior, not the main cooling route.
• "All neutron stars cool the same way." No — the dominant channel depends on central density and therefore mass. A massive star can switch on direct Urca and plunge in temperature; a low-mass star never crosses the threshold and cools slowly. Two stars of the same age can differ in surface temperature by a factor of several.
• "Superfluidity only slows cooling." It does suppress the Urca rates, but its onset also triggers the PBF neutrino burst that accelerates cooling — the very effect invoked to explain Cas A. The net effect is non-monotonic and time-dependent.
• "The Urca process changes the star's composition." Each full Urca cycle returns the nucleon to its starting state; the composition is in beta equilibrium and stays there. The only net export is energy, in the form of escaping neutrinos.
• "Neutrino cooling and the SN 1987A burst are the same thing." They are related but distinct. The 10-second SN 1987A burst was deleptonisation — the proto-neutron star shedding its lepton number and gravitational binding energy. Urca cooling is the much slower, centuries-long thermal relaxation that follows once the star has settled.
• "A reheated magnetar is still on the standard cooling curve." No. When magnetic-field decay or accretion deposits heat faster than neutrinos remove it, the surface is hotter than passive cooling predicts and the star sits above the curve. Cooling models must include such heating sources to interpret the hottest young neutron stars.