Algol Paradox

A star that aged out of turn

Point a spectroscope at Algol, the "Demon Star" in Perseus, and you find two stars locked in a 2.87-day orbit. One is a hot, blue-white, B8 main-sequence star of about 3.2 solar masses. The other is a cooler, dimmer K-type subgiant of only about 0.8 solar masses. There is nothing strange about a pair of unequal stars. The strangeness is which one is older — in the evolutionary sense.

The little 0.8 M☉ star has already exhausted the hydrogen in its core and swollen into a subgiant. It has, in the language of stellar evolution, left the main sequence. The hefty 3.2 M☉ star, by contrast, is still a perfectly ordinary main-sequence dwarf, quietly fusing hydrogen in its core just as the Sun does. So the lighter star is the more evolved one, and the heavier star is the more youthful one. That is exactly backwards from everything single-star theory predicts, and the contradiction is the Algol paradox.

Why heavier stars must die first

The whole puzzle hangs on one of the most robust results in stellar astrophysics: massive stars live fast and die young. A star's nuclear fuel is roughly proportional to its mass, but the rate at which it burns that fuel — its luminosity — climbs far more steeply. On the upper main sequence the mass-luminosity relation is approximately

L ∝ M^3.5

Main-sequence lifetime is fuel divided by burn rate, so

t_MS ∝ fuel / luminosity ∝ M / M^3.5 = M^(-2.5)

This single power law has dramatic consequences. Anchoring on the Sun (≈ 10 billion years):

Mass	Spectral type	Main-sequence lifetime
0.8 M☉	K0 V	≈ 18 billion yr
1 M☉	G2 V (Sun)	≈ 10 billion yr
2 M☉	A2 V	≈ 1.2 billion yr
3 M☉	B9 V	≈ 400 million yr
5 M☉	B5 V	≈ 90 million yr
10 M☉	B1 V	≈ 30 million yr

Two stars are born together from the same molecular cloud, at the same time, with the same composition. If one is even slightly heavier, it must reach the end of its main-sequence life first — without exception. A 0.8 M☉ star sitting next to a 3.2 M☉ star should, after only a few hundred million years, still be an unremarkable dwarf with more than 17 billion years of main-sequence life ahead of it, while the 3.2 M☉ star should be the one ballooning into a subgiant. Algol shows the reverse. Something must have changed the masses after the stars were born.

The resolution: the stars swapped roles

The escape from the paradox, worked out in the early 1950s once it became clear that close binaries can exchange matter, is that the masses we measure today are not the masses the stars were born with. The star that is now the 0.8 M☉ subgiant was originally the more massive of the two — perhaps it started near 3 M☉, with its companion near 1.5 M☉.

Being heavier, that original primary evolved first, exactly as the lifetime law demands. As it exhausted core hydrogen it expanded into a subgiant. But it was not alone in empty space: it sat in a tight orbit, surrounded by a teardrop-shaped region of gravitational dominance called its Roche lobe. When the swelling star grew to fill that lobe, gas at its surface was no longer bound to it more strongly than to its companion. Matter began streaming through the inner Lagrange point L1, the saddle in the gravitational potential between the two stars, and fell onto the partner.

Over time the originally massive star lost most of its hydrogen envelope this way. It dwindled to the stripped, low-mass subgiant we see now. The companion, fattening on the donated gas, climbed past it in mass to become today's 3.2 M☉ main-sequence star. The roles reversed. The star that is now light is the more evolved one precisely because it was once heavy and gave its mass away. The paradox dissolves the moment you stop assuming the stars evolved in isolation.

Roche lobes and the L1 funnel

The Roche lobe is the key piece of geometry. In the co-rotating frame of a binary, the effective potential (gravity from both stars plus the centrifugal term) has a figure-eight critical equipotential. The two lobes of that figure-eight meet at L1. Each star is gravitationally "in charge" of everything inside its own lobe; anything that drifts to L1 can fall freely onto the other star.

A convenient approximation for the radius of the Roche lobe, due to Peter Eggleton (1983), is

R_L / a = 0.49 q^(2/3) / [ 0.6 q^(2/3) + ln(1 + q^(1/3)) ]
       where q = m_donor / m_accretor,  a = orbital separation

A star evolves into Roche-lobe overflow in one of two ways: the star expands (as it leaves the main sequence) until its radius reaches R_L, or the orbit shrinks (by losing angular momentum) until R_L shrinks down to the star. In the classic Algol channel it is the first: the originally massive star simply got bigger than its lobe. Once contact is made, the system becomes semidetached — one star exactly fills its Roche lobe and is actively donating, while the other sits comfortably inside its own. Algol is the prototype semidetached binary, and "Algol-type" is now a standard class.

The runaway, and why it later calms down

The most counter-intuitive part is how the orbit responds. For conservative mass transfer — where the accretor catches everything the donor loses and the total mass and angular momentum stay fixed — the orbital separation obeys

a ∝ 1 / (m1 m2)^2     (total mass and angular momentum conserved)

so the separation is smallest when the product m1·m2 is largest, i.e. when the masses are most equal. Start with a heavy donor giving mass to a light accretor: at first the product m1·m2 is increasing, so the orbit shrinks. A shrinking orbit shrinks the donor's Roche lobe, which forces even more overflow — a self-accelerating runaway that transfers a large fraction of the envelope in a thermally short time (10⁴–10⁵ years). This rapid phase is why we so rarely catch a system mid-runaway.

The runaway only stops once the mass ratio passes through unity and reverses. After that, mass flows from the now-lighter donor to the now-heavier accretor, the product m1·m2 starts decreasing, the orbit widens, and the donor's Roche lobe expands away from it. Transfer slows to a gentle trickle governed by the donor's slow nuclear expansion. That long, placid phase — lasting tens of millions of years — is the state Algol is caught in today, which is exactly why we observe so many Algols: the slow phase is long-lived and therefore common.

Case A, B and C — when the overflow happens

Bohdan Kippenhahn and Alfred Weigert classified mass-transfer binaries in 1967 by the donor's evolutionary state at the moment it fills its Roche lobe. The category controls how much the star wants to expand and therefore how violent and how complete the transfer is.

Case	Donor's stage at overflow	Behaviour	Typical product
Case A	Core hydrogen burning (on main sequence)	Slow, prolonged; star barely larger than at birth	Algol systems, close binaries
Case B	Hydrogen shell burning, crossing Hertzsprung gap	Fastest, most common; star swells dramatically	Most Algols, He-star binaries, some X-ray binaries
Case C	After core helium burning, on the AGB	Often unstable; can lead to common envelope	Cataclysmic variables, double white dwarfs

Algol itself is generally modelled as a Case A or early Case B system, which is why its transfer is comparatively gentle and the donor is "only" a subgiant rather than a deeply evolved giant. If transfer onto a much-less-massive companion is too fast for the accretor to absorb — common in Case C — the donated gas can swamp the companion and engulf both stars in a shared common envelope, a very different and far more dramatic outcome.

The Algol system in numbers

Algol (Beta Persei) sits about 90 light-years away and is the best-known naked-eye example of an eclipsing binary, dipping from magnitude 2.1 to 3.4 every 2.867 days when the dim subgiant passes in front of the bright primary. John Goodricke correctly proposed the eclipse model in 1783, long before anyone understood the stars themselves. The modern parameters are well constrained by spectroscopy and interferometry:

Property	Algol A (accretor)	Algol B (donor)
Mass (now)	≈ 3.2 M☉	≈ 0.8 M☉
Spectral type	B8 V (main sequence)	K0 IV (subgiant)
Radius	≈ 2.7 R☉	≈ 3.5 R☉
Roche lobe	Underfilled (detached)	Exactly filled (donating)
Evolutionary state	Less evolved	More evolved

Note the radii: the lighter B star is actually the larger of the two, because it is a puffed-up subgiant filling its Roche lobe, while the heavier A star is a compact main-sequence ball. The orbital period of Algol is slowly lengthening, consistent with the post-reversal, orbit-widening phase, and radio and X-ray observations reveal an accretion stream and gas structure between the stars. (A third, wider star, Algol C, orbits the inner pair every 1.86 years but plays no role in the mass transfer.)

Where the paradox shows up beyond Algol

• Algol-type (EA) binaries. Hundreds are catalogued — semidetached systems where a low-mass evolved subgiant donates to a more massive main-sequence star. They are the direct descendants of the channel that produced Algol.
• Beta Lyrae (Sheliak). A more extreme, still-rapid mass-transfer system. Here the donor is being stripped so fast that a thick accretion disk and gas streams shroud the accretor, partly hiding it. Beta Lyrae is essentially a snapshot of the violent phase Algol has already passed through.
• Blue stragglers. In old clusters, some stars sit above the main-sequence turnoff — apparently too massive and too young to belong. Many are the accretors: rejuvenated by gaining hydrogen-rich fuel from an Algol-style donor, they look younger than the cluster. The donor often survives as a faint white dwarf.
• Cataclysmic variables and X-ray binaries. When the accretor is a compact remnant rather than a normal star, the same Roche-lobe-overflow machinery feeds a white dwarf (cataclysmic variables), neutron star, or black hole (X-ray binaries) — and the same mass-ratio-reversal bookkeeping applies.
• Type Ia supernova progenitors. One leading channel funnels mass onto a white dwarf via Roche-lobe overflow until it nears the Chandrasekhar limit and detonates — a cosmologically important application of exactly this physics.

Common misconceptions and edge cases

• "The stars were born with these masses." No — that assumption is what creates the paradox. The whole point is that the present masses are the result of transfer, not the initial conditions. The originally heavier star is the one that is now lighter.
• "More massive just means more evolved." Only for single stars of the same age. In an interacting binary, evolutionary state tracks the star's original mass and its transfer history, not its current mass.
• "Mass moving to the heavy star should push the orbit apart." Backwards. While mass flows from the heavier to the lighter star (before reversal) the orbit shrinks, driving a runaway. Only after the ratio reverses does the orbit widen.
• "It's the same as a common envelope." Not necessarily. Stable Roche-lobe overflow (the Algol channel) keeps the two stars distinct. A common envelope is what happens when transfer is dynamically unstable and the donated gas engulfs both stars — a separate, more catastrophic outcome.
• "All the donated mass is kept by the accretor." Only in the idealized conservative case. Real systems are partly non-conservative: some gas escapes, carrying angular momentum and altering the final period and mass ratio. The conservative formula is a useful first approximation, not a law.