Ergosphere

A second surface around a spinning hole

The Schwarzschild black hole has exactly one surface of interest — the event horizon at r = 2M, a perfect sphere from which nothing returns. The moment you allow the hole to rotate, that simple picture splits. The horizon survives, but it shrinks (a maximally spinning Kerr hole has r_+ = M, half the Schwarzschild value). A second surface appears outside it, no longer spherical: an oblate, pumpkin-shaped envelope that bulges at the equator and touches the horizon at the poles. This outer surface is the static limit, and the volume between it and the horizon is the ergosphere.

The name comes from the Greek ἔργον, "work". Inside, you can do work on the black hole — or, more interestingly, the black hole can do work on you. The hole is no longer just a sink. It has become a battery.

Where the surface comes from

In Boyer–Lindquist coordinates, the exterior Kerr metric has a metric component

g_tt = −(1 − 2Mr / Σ),       Σ = r² + a²cos²θ

where M is the mass and a = J/Mc is the angular momentum per unit mass (with G = c = 1, a ranges from 0 to M). Setting g_tt = 0 and solving for r gives the static limit

r_static(θ) = M + √(M² − a²cos²θ)

and the event horizon is the inner root of g_rr⁻¹ = 0, at

r_+ = M + √(M² − a²)

At the poles (θ = 0 or π) the two surfaces coincide. At the equator (θ = π/2) the static limit sits at r_static = 2M while the horizon sits at r_+ = M + √(M² − a²) ≤ 2M, with equality only when a = 0 (no ergosphere, no rotation). The two roots open up a finite gap whenever the hole spins, and that gap is the ergosphere.

The physical content: frame dragging

What is happening physically inside that surface? In the Kerr geometry, a massive observer is described by a 4-velocity u^μ. To "stand still relative to infinity" means u^μ ∝ ∂/∂t, i.e. zero spatial velocity in Boyer–Lindquist coordinates. For such an observer to be physically real, the worldline must be timelike: g_tt < 0. But inside the static limit g_tt > 0. There simply is no timelike worldline of constant (r, θ, φ); any massive observer must have nonzero dφ/dt, and the sign must match the hole's spin. Lense and Thirring had derived a weak-field version of this dragging effect in 1918; in Kerr's full solution the dragging becomes so violent that there is a region where it cannot be resisted.

The frame-dragging angular velocity is

ω(r, θ) = −g_tφ / g_φφ = 2Mar / (Σ(r² + a²) + 2Ma²r sin²θ)

which goes from zero at infinity to a finite, positive value at the horizon (ω → Ω_H = a / (r_+² + a²) on the horizon itself). Locally non-rotating observers — observers who carry orthonormal tetrads at constant r, θ — are nevertheless carried around the spin axis at angular speed ω as seen from infinity. They are the "ZAMOs" (zero-angular-momentum observers) that appear throughout the Kerr literature.

The Penrose process

In 1969, in a remarkable paper aimed largely at general readers, Roger Penrose pointed out a startling consequence. Far from the hole, the Killing energy of a particle is

E = −p · ∂_t  ≡  −p_t

which is positive for any physical worldline reaching infinity. Inside the ergosphere, where ∂_t is spacelike, the quantity −p_t can be either sign while still corresponding to a timelike worldline. There exist physical particle trajectories with negative Killing energy as measured from infinity.

The setup is then almost embarrassingly simple. Send a single particle of energy E_0 into the ergosphere on a carefully chosen trajectory. Inside, arrange for it to fission into two pieces, daughter A and daughter B, with energies E_A and E_B. By 4-momentum conservation E_A + E_B = E_0. Arrange the geometry so that daughter A's worldline has E_A < 0 and is on a trajectory that crosses the horizon. The other daughter has E_B = E_0 − E_A > E_0 and is on an outgoing trajectory escaping to infinity. The escaping piece carries more energy than was delivered. The black hole, in absorbing a particle of negative energy, loses mass: δM = E_A < 0. It also loses angular momentum, because the only trajectories with E < 0 in the ergosphere are counter-rotating, so the absorbed angular momentum δJ is also negative.

Energy and angular momentum are conserved overall. The escaping daughter has mined the rotational energy of the hole.

How much can you extract?

Demetrios Christodoulou (1970) and, later, Christodoulou and Ruffini (1971), proved that the total mass of a Kerr black hole decomposes as

M² = M_irr² + J² / (4 M_irr²)

where M_irr is the irreducible mass, related to the area A_H of the horizon by 16π M_irr² = A_H. The second law of black-hole mechanics forbids A_H from decreasing, so M_irr can only grow. The rotational energy that can be liberated is therefore at most

E_rot = M − M_irr  ≤  M (1 − 1/√2)  ≈  0.2929 M

for a maximally spinning hole (a = M). About 29 percent of the rest mass of a Kerr hole is available as extractable spin energy. Once it has all been mined the hole becomes Schwarzschild — the same mass A_H, no spin, no ergosphere, no more energy on offer. Beyond that, the area theorem locks the remaining energy in.

Spin parameter a*	r_+ / M	r_static (eq) / M	Max extractable	Comment
0.0	2.000	2.000	0.0 %	Schwarzschild — no ergosphere
0.3	1.954	2.000	1.2 %	Slow rotator
0.5	1.866	2.000	3.4 %	Typical AGN spin estimates begin here
0.7	1.714	2.000	7.7 %	Most observed AGN
0.9	1.436	2.000	15.0 %	M87* central spin estimate ~0.9
0.998	1.063	2.000	27.7 %	"Thorne limit" — accretion-spin equilibrium
1.0	1.000	2.000	29.3 %	Maximal Kerr — extremal, non-physical end-state

Notice that the equatorial static limit is fixed at 2M regardless of spin — the ergosphere's outer radius does not move; its inner boundary (the horizon) recedes inward, and the volume of the ergosphere grows as spin grows. A non-spinning hole has zero ergosphere; a maximal Kerr hole has a fat one.

Blandford-Znajek: the astrophysical Penrose

The single-particle Penrose process is theoretically elegant but practically difficult: the relative velocity of the two fragments at the splitting point must exceed about 0.5 c, which is hard to engineer naturally. Real astrophysical systems exploit a continuous, fluid analogue.

In 1977 Roger Blandford and Roman Znajek considered a Kerr hole threaded by a magnetic field anchored in the surrounding plasma — typically an accretion disk in the magnetically-arrested-disk (MAD) configuration. The rotation of spacetime within the ergosphere acts on the magnetic field lines as if the hole were a giant unipolar inductor. The result is an outgoing Poynting flux along the polar field lines, carrying energy and angular momentum away electromagnetically. The hole spins down; the jet flies out.

The luminosity is, to leading order,

L_BZ ≈ (κ / 4πc) Φ_B² Ω_H²    where    Ω_H = a / (r_+² + a²)

and Φ_B is the magnetic flux threading the horizon. κ ≈ 1/6 in the slow-rotation limit. For a 10⁹ M☉ supermassive black hole threaded by 10⁴ G near the horizon — typical MAD conditions — L_BZ ≈ 10⁴⁶ erg/s, comfortably enough to power a quasar jet.

Crucially, no matter has to cross the horizon for L_BZ to flow. The energy is paid in spin, not in rest mass. This is what makes it the astrophysical Penrose process: it is the ergosphere doing work on the external electromagnetic field, exactly as Penrose's particles did work on the external trajectory.

Where the ergosphere shows up in the sky

• M87*. The 6.5 × 10⁹ M☉ supermassive black hole imaged by the Event Horizon Telescope. EHT polarisation maps released in 2021 show a magnetic-field geometry consistent with a MAD configuration; modelling of the kilo-parsec-scale jet implies a spin parameter a* ≈ 0.9 and a BZ-powered launching site within a few r_g of the hole. The jet is the most direct piece of evidence we have for ergospheric energy extraction in nature.
• SS 433. A galactic microquasar 5.5 kpc away — a stellar-mass black hole (≈ 8–16 M☉) accreting super-Eddington from a massive companion, launching baryon-loaded jets at 0.26 c that precess on a 162-day clock. The jets visibly trace ballistic blobs over months of monitoring; spin estimates and X-ray polarimetry support an ergospheric launching mechanism.
• Cygnus X-1. Stellar-mass BH in a high-mass X-ray binary; iron-line and continuum-fitting spectral modelling consistently give spin parameter a* > 0.95, near maximal. The presence of compact radio jets when in the hard state is consistent with BZ.
• GRS 1915+105. Microquasar with one of the best-measured stellar-mass spins (a* ≈ 0.98) and dramatic relativistic ejection events. The 1994 superluminal jet — moving at 92 % c in projection — is the textbook example of a BZ-driven relativistic outflow from a near-extremal hole.
• Gravity Probe B (Earth). Not a black hole, but the same physics: the satellite's gyroscopes measured the frame-dragging precession of spacetime around the rotating Earth in 2011 at 37.2 ± 7.2 mas/yr, matching general relativity at the 19 % level. Earth has no ergosphere — its r_static would be inside its centre — but the underlying Lense–Thirring effect is the same.

Superradiance: the wave version

The Penrose process has a direct analogue for waves. When a wave of frequency ω and azimuthal quantum number m scatters off a Kerr hole, the scattered wave has greater amplitude than the incident wave whenever the superradiance condition

0 < ω < m Ω_H

is satisfied. Energy is extracted from the spin; the scattered wave carries it away. For massless scalar, electromagnetic, and gravitational fields the effect is small but real; for massive bosons of mass μ such that μ < m Ω_H, the extracted energy can build up in bound states between successive scatterings, producing a "black-hole bomb" — a self-amplifying cloud that spins the hole down on a timescale much shorter than the Hubble time. This sets some of the tightest astrophysical constraints on ultralight bosons, including QCD axions and dark-photon candidates.

In 2017, Torres et al. demonstrated the hydrodynamic analogue: surface waves on a draining bathtub vortex were amplified by scattering off the vortex's rotating core, with the amplification matching the predicted superradiance gain to within the experimental error.

A trajectory worth tracing

Consider a particle dropped from rest at infinity onto the equatorial plane of a Kerr hole with a* = 0.998 (Thorne limit). The particle's angular momentum, fixed at infinity, drops into the deep gravity well; by the time it crosses the static limit at r = 2M it is moving at roughly 0.7 c azimuthally as measured by a ZAMO, even though it had zero initial angular momentum. Frame dragging has done all the work. If the particle is at this point pulsed apart by an internal spring with relative speed 0.5 c — say a chemical decay or a programmed engine — one fragment can be put on a marginally bound retrograde trajectory with E < 0 that crashes through the horizon, and the other on a strongly hyperbolic prograde escape with E up to about 1.21 E_0. A 21 % energy gain on a single shot, drawn from the hole.

Stack 10⁶ such shots and you have, in principle, a civilisation-scale power station. Christodoulou's irreducible-mass bound says the gravy stops at 29 %, but until then each pulse is essentially free.

Common pitfalls

• Confusing the static limit with the event horizon. The static limit is not a one-way membrane. Massive observers can and do cross r_static in both directions — provided that while inside they co-rotate with the hole. The event horizon at r_+ is the true point of no return.
• Thinking the ergosphere requires extreme spin. Any spin at all (a > 0) produces an ergosphere. The size of the gap grows with a, but a slowly spinning Kerr hole has a thin, real ergosphere with the same property — no static observers allowed.
• Quoting "29 percent of the mass" as the energy in the spin alone. The 29 % figure is the fraction of the total mass that is extractable for a maximally spinning hole. It is not the spin energy as a fraction of M c² in some Newtonian sense; it is the irreducible mass theorem's strict upper bound, set by horizon-area conservation.
• Assuming Penrose requires touching the horizon. The Penrose process operates entirely in the ergosphere, outside the horizon. Only the negative-energy fragment needs to be absorbed; the splitting event and the escaping fragment can stay well outside r_+.
• Treating Blandford-Znajek and Blandford-Payne interchangeably. Blandford-Payne (1982) extracts energy from the accretion disk via magnetic centrifugal acceleration, not from the hole. Blandford-Znajek (1977) extracts spin energy from the hole itself through the ergosphere. Both contribute to real jets; only BZ depends on the existence of an ergosphere.
• Forgetting that extracting energy spins the hole down. The Penrose process and BZ both extract angular momentum proportional to the energy. There is no way to mine ergospheric energy without reducing a*, and once a* reaches zero the ergosphere is gone.