Innermost Stable Circular Orbit

What the ISCO actually is

Drop a marble into a smooth bowl and it settles at the bottom; nudge it and it rolls back. That is a stable equilibrium. Now imagine a bowl whose floor curves down so steeply near the center that the rim of the bowl vanishes — past some radius there is no bottom to roll back to, only a slope leading to the drain. The radius where the stable bottom disappears is the innermost stable circular orbit. Around a black hole, the "drain" is the event horizon, and the steepening floor is the geometry of warped spacetime itself.

More precisely: a test particle in orbit feels an effective potential that combines gravity, its angular momentum, and — uniquely in general relativity — an extra relativistic attraction. For large orbits this potential has a stable minimum (a circular orbit) and an outer turning point, just like Newtonian gravity. As you shrink the orbit, the minimum and an accompanying maximum slide toward each other. At the ISCO they merge into a single inflection point: the orbit is now marginally stable. Below the ISCO the minimum is gone entirely, leaving only a potential peak, so no stable circular orbit exists. The ISCO is sometimes called the marginally stable orbit, r_ms, for exactly this reason.

The Schwarzschild number: 6GM/c²

For a non-rotating (Schwarzschild) black hole the answer is clean and famous:

r_ISCO = 6GM/c² = 3 r_s

where r_s = 2GM/c² is the Schwarzschild radius (the event horizon). So the closest stable circular orbit sits at exactly three times the horizon radius. A particle there is moving at half the speed of light (v = c/2 in the relevant local sense), and the binding energy it has released to get there is about 5.7% of its rest-mass energy — energy that, in a real disk, has been radiated away as light. For comparison, the more familiar photon sphere — where light itself can orbit, unstably — sits closer in at 1.5 r_s, half the ISCO radius. Massive particles can never reach the photon sphere on a stable orbit; the ISCO is their innermost refuge.

Scale this to a real object. Sagittarius A*, the Milky Way's 4.3-million-solar-mass black hole, has a Schwarzschild radius of about 12 million km (roughly 0.08 AU). Its non-spinning ISCO would lie near 38 million km (about 0.25 AU) — comfortably inside Mercury's perihelion, packed around an object the size of a small planetary orbit. For a 10-solar-mass stellar black hole, r_s is about 30 km and the ISCO sits near 90 km: the entire inner accretion engine of an X-ray binary fits inside a single city.

Spin moves the goalposts

Real astrophysical black holes rotate, often near the theoretical maximum. A spinning Kerr black hole drags spacetime around with it (frame dragging), and this changes the ISCO dramatically depending on which way the particle orbits relative to the spin. The dimensionless spin parameter a runs from 0 (non-rotating) to 1 (maximal).

Spin a	Prograde ISCO (GM/c²)	Retrograde ISCO (GM/c²)	Max radiative efficiency
0 (Schwarzschild)	6.00	6.00	5.7%
0.5	4.23	7.55	8.2%
0.9	2.32	8.72	15.6%
0.998 (Thorne limit)	1.24	8.99	~32%
1 (maximal)	1.00	9.00	~42%

Two things stand out. First, a prograde disk around a fast-spinning hole reaches almost down to the horizon, releasing far more energy than a non-spinning one — which is why spin is the single biggest lever on how brightly a black hole can shine for a given accretion rate. Second, the "Thorne limit" of a ≈ 0.998 is set by the disk itself: photons preferentially swallowed by the hole carry negative angular momentum and cap the achievable spin just short of maximal, so a true a = 1 is an idealization.

Why disks stop at the ISCO

A thin accretion disk is essentially a dense stack of nested circular orbits, each slightly inside the next, with gas slowly losing angular momentum to viscosity and drifting inward. That picture only works where stable circular orbits exist. At the ISCO the stack runs out of stable rungs: gas crossing it can no longer circle, so it transitions to a near-radial plunge, falling into the horizon in roughly one orbital period. The ISCO is therefore the natural inner edge of a standard (Novikov–Thorne) thin disk, and in disk models the radiative output is conventionally truncated there — the "zero-torque inner boundary condition."

This has observable consequences. The inner-disk temperature peaks just outside the ISCO; a smaller ISCO (high spin) means hotter inner gas and a bluer, more luminous disk. The total energy radiated per gram is set by the binding energy at the ISCO, which is exactly the radiative-efficiency column above. So when a quasar converts mass to light, the headline efficiency figure — far beyond anything fusion can manage — is fundamentally a statement about where the ISCO sits.

Measuring an orbit you can never see

Astronomers cannot resolve the ISCO directly for most black holes, but several independent methods read it off the light:

• Relativistic iron line. Fluorescent iron Kα photons (rest energy 6.4 keV) emitted from disk gas near the ISCO are smeared by Doppler shifts and gravitational redshift into a broad, skewed line. Its red wing reaches deepest for gas closest to the hole, so fitting the line profile yields the inner-disk radius — and hence the spin.
• Thermal continuum fitting. For stellar-mass black holes in the thermal state, the shape of the disk's blackbody spectrum, combined with a known mass and distance, pins the inner-disk radius to the ISCO.
• Quasi-periodic oscillations. High-frequency QPOs in X-ray binaries cluster near the orbital frequency of the innermost stable orbit, offering a clock tied to the ISCO.
• Direct imaging. The Event Horizon Telescope's images of M87* and Sgr A* probe the inner accretion flow on scales comparable to the ISCO and photon ring, increasingly constraining the geometry.

Common misconceptions

• The ISCO is the event horizon. No — for a non-spinning hole it sits at 3 r_s, three times farther out than the horizon. Only at maximal prograde spin do they nearly coincide.
• Nothing can orbit closer than the ISCO. Things can pass closer on unstable or plunging trajectories; what cannot exist below the ISCO is a stable circular orbit. Light orbits (unstably) at the photon sphere, well inside.
• The ISCO exists in Newtonian gravity too. It does not. Newtonian orbits are stable at every radius; the ISCO is a purely relativistic effect from the extra 1/r³ term in the effective potential.
• Crossing the ISCO means instantly hitting the horizon. The infall is fast but not instantaneous — matter spirals through a "plunging region" between the ISCO and horizon over about an orbital time.
• The ISCO is fixed. It depends entirely on the black hole's spin, scaling between 1 and 9 GM/c² as a ranges from maximal prograde to maximal retrograde.