Contact Binary

Two stars touching

Picture two suns so close that they are not really two suns any more. Each has bloated outward until it spills onto the other, and the boundary between them has dissolved into a single, hourglass-shaped cloud of gas. They whirl around their common centre in eight hours, their facing surfaces nearly grazing, dragging a shared atmosphere with them. That is a contact binary — a pair of stars that have merged their outer layers without yet merging their cores.

It is one of the strangest stable configurations in stellar astrophysics. A normal binary is two distinct stars in clean orbits. A contact binary blurs that distinction: the gas at the surface does not know which star it belongs to, because at the connecting point it belongs equally to both. The English-language shorthand for the most famous class is "W UMa stars," after the prototype W Ursae Majoris, a sixth-magnitude variable in the bowl of the Big Dipper that dims and brightens twice every eight hours.

The Roche lobe and the L1 neck

To understand contact, you first need the geometry of a close binary. In a frame co-rotating with the orbit, the effective gravitational potential — true gravity plus the centrifugal term — has a characteristic figure-eight cross-section. Each star sits inside one loop of the eight, called its Roche lobe: the largest closed equipotential surface that still wraps a single star. The two lobes meet at exactly one saddle point, the inner Lagrange point L1, where the pull of the two stars and the centrifugal force cancel.

The Roche-lobe radius of a star depends almost entirely on the mass ratio q = M₂/M₁ and the orbital separation a. A useful approximation due to Eggleton (1983) is

R_L / a = 0.49 q^(2/3) / [ 0.6 q^(2/3) + ln(1 + q^(1/3)) ]

A detached binary has both stars comfortably inside their lobes. A semi-detached binary has one star filling its lobe and pouring gas through L1 onto the other — this is ordinary Roche-lobe overflow, the engine of Algol systems and cataclysmic variables. A contact binary is the next step: both stars fill and overflow their lobes, so the surface is now described not by two separate lobes but by a single equipotential that surrounds the pair — bounded on the outside by the next critical surface that passes through the outer Lagrange point L2. The shared layer of gas above the inner critical surface is the common envelope, and its thickness (the "fill-out factor") measures how deep into contact the system has fallen.

Why they orbit in hours

Contact binaries orbit fast because they are small. Kepler's third law fixes the relationship between period P, separation a, and total mass:

P² = 4π² a³ / [ G (M₁ + M₂) ]

For two solar-mass stars almost touching, the separation is only a couple of solar radii. Put a ≈ 2 R☉ ≈ 1.4 × 10⁹ m and M₁ + M₂ ≈ 2 M☉ into Kepler's law and the period comes out near a quarter of a day. Concretely, the W UMa class spans orbital periods from roughly 0.22 days at the short end to about 1 day at the long end, with the bulk piled up around 0.3 to 0.4 days. The prototype W UMa itself has P = 0.3336 day — exactly 8.01 hours — so it runs through a full eclipse cycle three times a night.

Because the stars are tidally locked (synchronised by the same tides that distort them), each rotates once per orbit. A solar-radius surface turning around in eight hours has an equatorial speed of order 100–150 km/s, hundreds of times faster than the Sun's 2 km/s. That rapid rotation drives a vigorous magnetic dynamo, which is why W UMa stars are spotted, flaring, X-ray-bright, and chromospherically active — and why they steadily lose angular momentum, as we will see.

The equal-temperature paradox

Here is the feature that makes contact binaries genuinely peculiar. The two stars usually have unequal masses — mass ratios q from about 0.1 to 0.9 are common — and an isolated main-sequence star's surface temperature climbs steeply with mass. Two stars differing by a factor of two in mass ought to differ by thousands of kelvin in surface temperature. Yet observed W UMa systems show the two eclipse minima at nearly the same depth, implying the two surfaces differ by only a few hundred kelvin. A 0.9 M☉ and a 0.4 M☉ star can share an effective temperature of, say, 5800 K.

The resolution is energy transfer through the common envelope. The more massive primary generates far more luminosity in its core than its undersized partner. Part of that energy is carried sideways through the L1 neck into the secondary's envelope — by large-scale circulation, by the contact layer acting as a thermal bridge, or both — so that the secondary radiates light it did not itself generate. The shared envelope behaves like a single radiating surface trying to settle to one temperature. The two subtypes encode which star is hotter at the surface: in A-type W UMa systems the more massive star is slightly hotter; in W-type systems the less massive star is hotter, a sign the energy transfer has over-corrected. Either way, the near-equality of the minima is the fingerprint of a shared envelope and remains an active modelling problem.

Reading the light curve

You can classify an eclipsing binary by eye from its light curve, and contact binaries have an unmistakable shape. Compare the three Roche states:

State	GCVS type	Light-curve shape	Minima	Prototype
Detached	EA (Algol)	Flat between sharp, narrow eclipses	Often unequal; can be near-equal	Algol (β Persei)
Semi-detached	EB (β Lyrae)	Continuously varying, rounded	Clearly unequal depths	β Lyrae
Contact (overcontact)	EW (W UMa)	Continuous sinusoid, no flat part	Nearly equal depths	W Ursae Majoris

The continuous variation of an EW system comes from the fact that the stars are not spheres but distorted ellipsoids. As they orbit, the projected area you see changes constantly — there is "ellipsoidal variation" even between eclipses — so the brightness never flattens out. The two minima are nearly equal because the two surfaces are nearly the same temperature, so it barely matters which star is in front. By contrast a detached EA system spends most of its period showing two unblended round stars at constant total brightness, dipping only briefly when one passes in front of the other.

Numbers for a typical W UMa system

To anchor the picture, here are representative figures for the class and for the prototype itself.

Quantity	Typical W UMa	W Ursae Majoris (the prototype)
Orbital period	0.22 – 1.0 day	0.3336 day (8.01 h)
Combined mass	1 – 3 M☉	≈ 1.7 M☉ total (≈ 1.14 + 0.55)
Mass ratio q = M₂/M₁	0.1 – 0.9	≈ 0.49
Separation a	1 – 3 R☉	≈ 2.4 R☉
Surface temperature	4500 – 7000 K (both)	≈ 6300 K (both)
Spectral type	F, G, K dwarfs	≈ F8/G V
Absolute magnitude MV	+4 to +6	≈ +4.8
Distance (prototype)	—	≈ 52 pc (170 ly)

Notice how compact these systems are: a separation of just a few solar radii means the centre-to-centre distance is comparable to the stars' own diameters, which is exactly why their surfaces are in contact. The whole binary would fit comfortably inside the orbit of Mercury (about 83 solar radii), with enormous room to spare — the two stars and the gap between them span only a few times the Sun's width.

How they form and how they die

Contact binaries are not born in contact. They begin as ordinary, somewhat close detached binaries — two cool dwarfs orbiting in one to a few days. Two processes then squeeze them together. First, both stars are sun-like and magnetically active, so they drive magnetised stellar winds; the wind carries away angular momentum even though it carries away almost no mass, and this magnetic braking bleeds orbital angular momentum out of the system. Second, as the orbit shrinks, the stars grow to fill their Roche lobes and contact is established. From then on the system loses angular momentum on a timescale of order a billion years, the orbit slowly tightens, and the contact deepens.

The end is dramatic. As angular momentum drains, the mass ratio and orbital separation evolve toward a configuration where the orbit can no longer hold against the tidal coupling between spin and orbit. This is the Darwin instability: when the spin angular momentum the stars demand exceeds roughly a third of the orbital angular momentum available, tides cannot keep the stars synchronised, the orbit decays catastrophically, and the two stars spiral together and merge in a matter of days. The product is a single, fast-spinning, rejuvenated star — exactly the kind of object thought to explain some blue stragglers in old clusters, stars that look bluer and younger than their neighbours because they were assembled from two older stars.

This is no longer hypothetical. In 2008 the nova-like transient V1309 Scorpii brightened by several magnitudes. When astronomers went back through six years of OGLE survey photometry, they found it had been an eclipsing contact binary whose orbital period was visibly decreasing — the orbit shrinking — right up until the outburst. V1309 Sco was a contact binary caught merging in real time, the first direct confirmation that W UMa systems end as stellar mergers. The outburst was a "red nova" (a luminous red transient), the merger signature.

Famous and extreme examples

• W Ursae Majoris. The prototype, in the bowl of the Big Dipper, P = 0.3336 day, about 52 pc away. Two F/G dwarfs of total mass ≈ 1.7 M☉ (a ≈ 1.14 M☉ primary and a ≈ 0.55 M☉ secondary) sharing one envelope. Visible in a small telescope changing brightness within a single evening.
• VFTS 352. In the Tarantula Nebula of the Large Magellanic Cloud, the most massive and hottest overcontact binary known — two O-type stars of about 28 and 29 M☉ orbiting in just 1.12 days, with about 30% of their material shared. A test case for whether massive contact binaries merge or evolve chemically homogeneously toward a double black hole.
• V1309 Scorpii (2008). The contact binary merger caught on camera: a period-shrinking W UMa system whose orbit decayed into a red-nova outburst, confirming the merger fate.
• 44 Boötis (i Boötis). A naked-eye-bright, nearby W-type W UMa system at about 12.8 pc, period 0.27 day, useful for high-precision modelling because of its brightness and proximity.
• OGLE / Kepler / ZTF samples. Wide-field surveys have catalogued tens of thousands of W UMa stars; they are so numerous that they dominate the short-period end of any eclipsing-binary catalogue.

Contact binaries as distance markers

Because W UMa stars obey a tight relation between their orbital period, intrinsic colour, and absolute magnitude, they double as standard candles for relatively nearby distances. The empirical relation has the form

M_V = a · log P + b · (B − V) + c

where P is the period in days and (B − V) the colour index. Calibrated against systems with trigonometric parallaxes from Hipparcos and Gaia, this period–luminosity–colour relation pins a W UMa star's absolute magnitude to within a few tenths of a magnitude. Since these systems are common and easily found in star clusters and the field, they provide an independent, abundant rung for measuring distances out to the nearer parts of the Galaxy and to Local Group clusters — complementary to the more famous Cepheid and RR Lyrae distance indicators.

Misconceptions and edge cases

• "Contact binary" in stellar physics is not the same as in planetary science. Comets and Kuiper-belt objects like Arrokoth are also called "contact binaries" — two solid lobes touching gently. That is a coincidence of geometry, not of physics. A stellar contact binary is two gravitationally bound stars sharing a gaseous envelope; the planetary version is two rubble piles resting against each other.
• The cores have not merged. In a contact binary the outer envelopes are shared, but each star still has its own dense core orbiting the centre of mass. Full merger of the cores comes later, at the Darwin instability — it is the end state, not the definition.
• Overcontact is not the giant-phase common-envelope. The phrase "common envelope" also describes the brief, unstable phase when a giant swallows its companion during common-envelope evolution. A W UMa contact binary is a long-lived, stable shared envelope of two main-sequence dwarfs, not the dramatic in-spiral of a giant. Same words, very different timescales: billions of years versus hundreds of years.
• Equal minima do not mean equal stars. The near-equal eclipse depths reflect equal surface temperatures enforced by energy transfer, not equal masses or radii. The stars are typically quite unequal in mass.
• Not every short-period eclipser is a contact binary. A near-equal-depth, sinusoidal light curve is the signature; a system with flat stretches and sharp eclipses at the same period is a detached EA binary, not an EW contact system. The Roche state, not just the period, defines the class.