Symbiotic Star

Two stars that have no business sharing a spectrum

Point a spectrograph at a symbiotic star and you see something that should not be possible from a single object. There are the deep, cool molecular bands of titanium oxide — the unmistakable fingerprint of a red giant's ~3,000 K photosphere, the same bands you see in Betelgeuse or any garden-variety M giant. Superimposed on top of them are sharp, bright emission lines of helium and oxygen that have lost five electrons, lines that demand a radiation field hotter than 100,000 K. Cool molecules and ferociously hot ions, in the same beam of light, at the same time.

The resolution is that you are not looking at one star. You are looking at a wide binary: an evolved red giant losing mass through a slow, dense wind, and a tiny white dwarf companion buried in that wind, hot enough to ionise it. The giant supplies the cool molecular bands and the raw gas; the white dwarf supplies the hard ultraviolet that lights the gas up. The two stars are not touching — they are often separated by several astronomical units — but they are physically coupled through the giant's wind, and that coupling is what we call symbiosis. The American astronomer Paul Merrill, who catalogued the first members in the 1930s and 1940s, borrowed the biological word precisely because two organisms of completely different character seemed to be living together in one object.

The components: a giant, a white dwarf, and a wind

Every symbiotic star has three essential ingredients.

The cool giant. This is an evolved star — a first-ascent red giant or, more often, an asymptotic-giant-branch (AGB) star — with a radius of tens to hundreds of solar radii and an effective temperature near 2,800–3,500 K. Crucially, evolved giants lose mass copiously: a slow wind at 10–30 km/s carrying anywhere from 10⁻⁸ to 10⁻⁶ solar masses per year. In the "S-type" (stellar) symbiotics the giant is a normal red giant; in the rarer "D-type" (dusty) symbiotics it is a pulsating Mira variable shrouded in a thick dust shell that radiates strongly in the infrared.

The hot companion. In about 95% of confirmed systems this is a white dwarf with a surface temperature of 10⁵ to 2×10⁵ K and a luminosity of roughly 100 to 10,000 L☉ — far brighter than an isolated cooling white dwarf, because it is powered by accretion and, frequently, by quasi-steady nuclear burning of the hydrogen it accretes. A handful of systems instead host a neutron star (a "symbiotic X-ray binary", prototype GX 1+4), and these reach far higher X-ray luminosities.

The shared nebula. The white dwarf's ionising photons carve an H II region out of the giant's wind. The fraction of the wind that lies inside the Strömgren-like ionised zone produces the emission-line spectrum; the neutral remainder produces, among other things, the Raman-scattered lines discussed below. The geometry is genuinely three-dimensional and orbital-phase dependent: as the white dwarf moves around the giant, our sightline sweeps through regions of different ionisation, and the emission lines vary in strength over the orbit.

How the giant feeds the dwarf

Mass transfer in a symbiotic system is fundamentally different from the tidy Roche-lobe stream of a close binary. The giant is usually not large enough relative to its orbit to overflow its Roche lobe directly. Instead the white dwarf captures part of the giant's wind. The classic estimate is Bondi–Hoyle–Lyttleton accretion: a body of mass M_WD moving with relative velocity v_rel through gas of density ρ gravitationally focuses material within an accretion radius

R_acc = 2 G M_WD / v_rel²

Ṁ_acc ≈ π R_acc² ρ v_rel = 4π G² M_WD² ρ / v_rel³

Because the giant's wind is slow (v_wind ~ 10–30 km/s, comparable to the orbital speed), the relative velocity is small, R_acc is large, and the captured fraction can be substantial — of order 1–10% of the total mass-loss rate. That places typical accretion rates onto the white dwarf in the range 10⁻⁹ to 10⁻⁷ M☉/yr. Modern hydrodynamic simulations show that when the wind is this slow it is gravitationally beamed toward the white dwarf and funnelled through the inner Lagrange point in a process now called wind Roche-lobe overflow — intermediate between a pure spherical wind and classic Roche-lobe overflow, and far more efficient than simple Bondi capture. Whether an accretion disk forms depends on the specific angular momentum of the captured gas; many but not all symbiotics show disk signatures.

The composite spectrum and the Raman lines

The defining observable is the spectrum itself. Three components blend together:

• The cool continuum and molecular bands of the giant, dominating the red and infrared, with TiO and VO absorption.
• The hot continuum of the white dwarf, rising into the ultraviolet, plus a recombination nebular continuum.
• The emission-line spectrum of the ionised wind: hydrogen Balmer lines, He I and He II (the He II 4686 Å line requires photons above 54 eV), and forbidden lines like [O III] 5007 Å and [Ne V].

The crown jewel is a pair of broad emission features at 6825 Å and 7082 Å that defied identification for decades. In 1989 Hans Martin Schmid showed they are Raman-scattered O VI lines (later modelled in detail by Hee-Won Lee and others). Far-ultraviolet O VI resonance photons at 1032 and 1038 Å, produced near the white dwarf, strike neutral hydrogen atoms in the dense giant wind. Instead of being absorbed and re-emitted at the same wavelength (resonant scattering), the atom is left in an excited 2s state and re-emits a much redder photon — Raman scattering. The wavelength conversion is exact:

1/λ_out = 1/λ_in − 1/λ_Lyα
e.g.  1031.9 Å (O VI) → 6825 Å
      1037.6 Å (O VI) → 7082 Å

Because the process needs both a hard O VI–producing source and a thick neutral-hydrogen screen along the same path, the Raman lines appear in roughly half of all symbiotic stars and in essentially nothing else. Their polarization and profile encode the orbital geometry, making them a uniquely powerful diagnostic of the binary's three-dimensional structure.

A symbiotic system by the numbers

Concrete values for a representative S-type symbiotic, alongside the contrasting D-type:

Property	S-type (e.g. Z And)	D-type (e.g. R Aqr)
Cool component	Red giant, ~3,200 K	Mira variable + dust
Giant radius	~50–100 R☉	~300–500 R☉ (pulsating)
White dwarf T_eff	~1.5×10⁵ K	~10⁵ K
White dwarf L	~1,000 L☉	~100–1,000 L☉
Orbital period	~760 days	~44 years
Wind mass-loss	~10⁻⁷ M☉/yr	~10⁻⁶ M☉/yr
Accretion rate	~10⁻⁸ M☉/yr	~10⁻⁷ M☉/yr
Infrared excess	Modest	Strong (dust at ~10³ K)
Separation	~2–3 AU	tens of AU

The orbital period is the cleanest divider: S-types cluster near 1–3 years, D-types stretch to decades because the dusty Mira must sit in a much wider orbit. R Aquarii, the nearest symbiotic at only ~200 pc, additionally drives a spectacular bipolar jet and surrounding nebula visible in Hubble and Chandra imaging.

Outbursts: shell burning versus thermonuclear runaway

Symbiotic stars are variable on every timescale, but their dramatic brightenings fall into two physically distinct classes.

Classical (Z And–type) outbursts. The white dwarf brightens by 1–3 magnitudes over months to years, recurring irregularly. These are not nuclear explosions. They are driven by changes in the accretion rate and quasi-steady shell burning: when accretion temporarily exceeds the rate at which hydrogen can burn steadily, the burning envelope expands, the photosphere cools and swells, and the optical brightness rises while the effective temperature drops. The prototype Z Andromedae has cycled through such events for over a century.

Symbiotic novae. These are true thermonuclear runaways, the same physics as a classical nova eruption but with a giant donor. Hydrogen accumulates on the white dwarf until the base of the accreted layer reaches ~10⁷ K and ignites in a degenerate flash, brightening the system by 4–7 magnitudes and ejecting a shell. The decisive parameter is the white dwarf mass: the ignition column mass scales steeply with mass, so a near-Chandrasekhar white dwarf needs to accrete very little before it erupts.

ΔM_ig ∝ R_WD⁴ / M_WD   (ignition mass falls steeply with M_WD)
t_rec  = ΔM_ig / Ṁ_acc  (recurrence time)

For an ordinary ~0.6 M☉ white dwarf, t_rec is 10⁴–10⁵ years (a "slow nova" seen once). But for the ~1.3 M☉ white dwarf in RS Ophiuchi, ΔM_ig is tiny and t_rec collapses to about 15 years; RS Oph has erupted in 1898, 1933, 1958, 1967, 1985, 2006, and 2021. T Coronae Borealis, the "Blaze Star", recurs roughly every 80 years (1866, 1946) and has been widely anticipated to erupt again imminently.

Famous examples and where they sit

• Z Andromedae. The class prototype. An S-type system, ~760-day orbit, white dwarf + red giant, the namesake of classical accretion/shell-burning outbursts. The whole class of "Z And–type" outbursts is defined by it.
• R Aquarii. The nearest symbiotic (~200 pc), a D-type Mira + white dwarf in a ~44-year orbit, surrounded by the Cederblad 211 nebula and driving a precessing bipolar jet imaged in X-rays by Chandra. The textbook picture of a symbiotic with extended structure.
• RS Ophiuchi. A recurrent symbiotic nova; its ~1.3 M☉ white dwarf erupts every ~15 years, the blast wave ploughs into the red giant's dense wind, and the 2021 eruption was detected in TeV gamma rays by H.E.S.S. and MAGIC — direct evidence of particle acceleration in a nova shock.
• T Coronae Borealis. The "Blaze Star", a recurrent symbiotic nova with ~80-year recurrence (last in 1946), one of only a handful of naked-eye recurrent novae and a prime Type Ia progenitor candidate.
• CH Cygni. A bright, nearby, complex symbiotic notorious for radio jets and rapid flickering attributed to an accretion disk around its white dwarf.
• GX 1+4. The prototype symbiotic X-ray binary: an accreting neutron star, not a white dwarf, fed by an M-giant wind — a reminder that the "hot companion" is usually but not always a white dwarf.

Only ~300 symbiotic stars are confirmed in the Milky Way, yet population synthesis predicts anywhere from a few thousand to 10⁵, because their long periods and the overwhelming brightness of the giant make the hot companion easy to miss. Surveys exploiting the Raman O VI lines and H-alpha excess (and now Gaia astrometry) are steadily closing the gap.

Why symbiotic stars matter

Beyond being a spectroscopic curiosity, symbiotic stars are laboratories for several of astrophysics' hardest problems. They are the slowest, widest accreting white-dwarf binaries, so they probe the wind-accretion regime that fast cataclysmic variables cannot reach. The recurrent symbiotic novae are front-line single-degenerate candidates for Type Ia supernovae: their white dwarfs already sit near the Chandrasekhar mass, and whether they are net mass-gainers (marching toward detonation) or simply eject everything they accrete is a central, unresolved question that feeds directly into the cosmology of standardisable supernova distances. Their shock-driven outbursts, lit up in radio, X-rays, and gamma rays, are nearby analogues of the much more distant nova and supernova shocks we cannot resolve.

Common misconceptions and edge cases

• "Symbiotic" is not a biological description. The term refers strictly to the two-temperature composite spectrum. The stars do not exchange material symmetrically; mass flows one way, from giant to white dwarf.
• The hot star is not always a white dwarf. The defining feature is a hot, compact ioniser; in symbiotic X-ray binaries it is a neutron star, which dramatically changes the X-ray output and accretion physics.
• Not all outbursts are novae. Z And–type outbursts are accretion/shell-burning events of 1–3 magnitudes, fundamentally different from the 4–7-magnitude thermonuclear runaways of symbiotic novae. Conflating the two leads to wrong recurrence-time and energetics estimates.
• The giant rarely fills its Roche lobe. Mass transfer is dominated by wind capture and wind Roche-lobe overflow, not a focused Roche stream — which is exactly why orbital periods are so long compared to cataclysmic variables.
• The white dwarf's high luminosity is not residual cooling. An isolated old white dwarf is faint (~10⁻³ L☉). The 10²–10⁴ L☉ seen in symbiotics is powered by accretion and surface hydrogen burning; switch off the wind and the hot component would fade dramatically.
• D-type ≠ "more evolved" S-type. The S/D split is about the cool star (normal giant vs dusty Mira) and the orbital scale, not a single evolutionary sequence; the two types occupy different period and infrared regimes.