Recurrent Nova: Repeated Thermonuclear Flashes on a Massive White Dwarf

What a recurrent nova is

A recurrent nova is a type of cataclysmic variable: a close binary in which a white dwarf accretes hydrogen-rich matter from a companion. When enough fuel piles onto the white dwarf's surface, it ignites in a thermonuclear runaway, blowing off a shell and flaring by 7–15 magnitudes (a brightening of hundreds to a million times). What sets recurrent novae apart from ordinary classical novae is timescale: at least two eruptions have been observed on human timescales, with recurrence intervals from about 1 to 100 years.

• Same physics, faster clock: both classical and recurrent novae are surface hydrogen flashes; recurrent ones simply reach ignition far sooner.
• The star survives: unlike a supernova, the white dwarf is not destroyed. It ejects only a thin envelope (typically 1e-7 to 1e-6 M_sun) and resumes accreting.
• Extreme progenitors: rapid recurrence demands both a high accretion rate and a white dwarf near the maximum mass a white dwarf can carry, the Chandrasekhar limit of ~1.4 M_sun.

Only about ten are confirmed in the Milky Way, though many "classical" novae are likely recurrent on longer, unobserved cycles.

The mechanism: degenerate ignition and the recurrence clock

Accreted hydrogen settles into a thin layer at the base of which density and temperature climb as more mass arrives. Because the white dwarf interior is electron-degenerate, pressure barely responds to temperature, so when hydrogen (via the CNO cycle) ignites, the layer cannot expand and cool. Temperature and burning rate spiral upward in a runaway until the layer reaches ~1e8 K and finally becomes non-degenerate, expanding explosively and driving mass loss at 1,000–4,000 km/s.

Ignition requires a roughly fixed critical envelope mass, M_ign. A more massive white dwarf has a smaller radius and far stronger surface gravity (g scales steeply with mass near Chandrasekhar), so the base of even a thin shell reaches ignition pressure with less accreted mass. The recurrence time follows the simple accounting relation:

• t_rec ≈ M_ign / Ṁ — recurrence time equals the fuel needed divided by the accretion rate.

Since M_ign falls sharply as white dwarf mass rises, and Ṁ is large in these systems, t_rec collapses from ~10,000 yr (classical) to years or decades. Short recurrence therefore simultaneously requires a high Ṁ and a near-Chandrasekhar mass.

Characteristic numbers and a worked example

Consider U Scorpii, the fastest-recurring bright example, with eruptions in 1863, 1906, 1936, 1979, 1987, 1999, 2010 and 2022 — roughly every 10 years.

• White dwarf mass: ~1.37 M_sun (essentially at the Chandrasekhar limit).
• Ignition mass: at this mass, M_ign is only ~1e-6 M_sun of hydrogen.
• Accretion rate: to burn through that in ~10 years, Ṁ ≈ 1e-6 / 10 ≈ 1e-7 M_sun/yr — indeed the measured range.

The eruption itself is fast and hot: U Sco declines 3 magnitudes (t3) in only ~2.6 days, among the quickest of any nova. Peak absolute magnitudes reach roughly M_V ≈ -8 to -9, so at outburst these systems briefly emit 10^38 erg/s or more, near the Eddington luminosity of a solar-mass object. Ejecta velocities of thousands of km/s and ejected kinetic energies near 1e44 erg are typical — energetic, but a hundred-thousand times below a supernova.

How they are observed and detected

Recurrent novae are found and characterized across the spectrum:

• Optical light curves: amateur and survey photometry (AAVSO, ASAS-SN) catch the sudden rise and the rapid, often plateaued decline; long historical records let astronomers count cycles back to the 1800s.
• Spectroscopy: broad, high-velocity emission lines reveal the fast ejecta; in symbiotic systems the red-giant wind is shock-heated by the blast.
• X-rays and the supersoft phase: after ejection, residual surface burning makes the white dwarf a bright supersoft X-ray source for weeks, seen by Swift and Chandra.
• Gamma rays: Fermi-LAT detected GeV gamma rays from RS Ophiuchi's 2021 eruption, direct evidence that the blast wave accelerates particles as it plows into the red giant's dense wind — a nova acting as a miniature supernova-remnant shock.
• Radio and interferometry: VLBI and instruments like the VLTI/AMBER resolved RS Oph's expanding, bipolar shock only days after outburst.

Pre-eruption warning signs also exist: T CrB brightened into a "high state" around 2015 and then dipped in 2023–2024, patterns tied to changing mass-transfer that flag an imminent flash.

How recurrent novae differ from their cousins

It helps to place recurrent novae against neighboring phenomena on the same white dwarf:

• vs classical novae: identical hydrogen-flash physics, but classical novae sit on lower-mass white dwarfs with modest Ṁ, so they recur only every ~1e4–1e5 yr and have been seen erupting just once.
• vs dwarf novae: those are accretion-disk instabilities (a brightening of the disk, not the star), fainter and far more frequent; no thermonuclear burning is involved.
• vs Type Ia supernovae: a nova ejects a thin skin and the white dwarf survives; a Type Ia disrupts the entire star in a carbon-oxygen detonation releasing ~1e51 erg — 10 million times more energy.

Recurrent novae also subdivide by donor: the RS Oph / symbiotic class feeds off a red giant (long orbits of hundreds of days); the U Sco class has an evolved subgiant on a ~1-day orbit; the T Pyx class resembles a classical-nova binary with a short, hours-long period. Whether each nova gains net mass depends on whether ejected mass exceeds accreted mass — the crux of the supernova question.

Significance, open questions, and famous cases

The stakes are cosmological. If a recurrent nova's white dwarf retains more mass than it ejects each cycle, it creeps toward the Chandrasekhar limit and may end as a Type Ia supernova — the standardizable explosions used to measure cosmic distances and discover dark energy. This is the "single-degenerate" progenitor channel, and RS Oph (WD ~1.2–1.35 M_sun) and U Sco (~1.37 M_sun) are textbook candidates.

• The mass-balance debate: hydrodynamic models disagree on whether nova eruptions are net-erosive (dredging up core material and shedding mass) or net-accumulating. If erosive, recurrent novae may never reach the limit.
• Famous systems: T CrB (~80 yr, eruptions 1866, 1946, and expected mid-2020s), RS Oph (~15 yr; eruptions include 1898, 1933, 1958, 1985, 2006, 2021), U Sco (~10 yr), and T Pyx, whose eruptions have slowed and whose ejected shell is directly imaged.
• Extragalactic recurrents: the nova M31N 2008-12a in Andromeda erupts almost annually, implying a white dwarf right at 1.38 M_sun and an exceptionally high Ṁ — the most extreme recurrence known.

Each new eruption, especially the anticipated T CrB flash, is a live test of these ideas.

The three classes of Galactic recurrent novae, distinguished by donor type and orbital period, with representative systems and typical recurrence intervals.

Class	Donor star	Orbital period	Recurrence time	Examples
RS Oph / symbiotic	Red giant	~100–520 days	~9–80 yr	RS Oph, T CrB, V745 Sco, V3890 Sgr
U Sco	Subgiant	~1–1.4 days	~10 yr	U Sco, V394 CrA, V2487 Oph
T Pyx	Main-sequence dwarf (short period)	~1.8–3.4 hr	~12–24 yr	T Pyx, IM Nor, CI Aql
Classical nova (for contrast)	Dwarf/subgiant	hours–days	~10,000–100,000 yr	GK Per, DQ Her, V1500 Cyg