Ergosphere and Frame Dragging

A rotating mass does something a static mass cannot: it drags the very fabric of spacetime into rotation with it. Far from an ordinary spinning body this "frame dragging" is a whisper — a gyroscope in orbit precesses by a few milliarcseconds a year. But wind the spin up around a black hole and the whisper becomes a command. In a shell just outside the horizon of a Kerr hole, the dragging is so violent that standing still is no longer an option: to hold fixed against the distant stars you would have to move backward faster than light, and since nothing outruns light, you are swept into corotation. That shell is the ergosphere, and its outer skin — the static limit — is the surface where the time-translation Killing vector $\partial_t$ , which far away points into the future along a possible worldline, tips over and becomes spacelike. The tipping-over is bookkept by a single sign flip: $g_{tt}$ changes from negative to positive. That one flip has an extraordinary consequence. The conserved energy of a particle, $E=-\xi\cdot p$ , is built from $\partial_t$ ; where $\partial_t$ is spacelike, $E$ is no longer a bona-fide energy but a component of momentum, and it can be negative for perfectly physical, future-pointing trajectories. Feed a negative-energy fragment into the horizon and the hole loses mass — you have mined its rotational energy. This is the seed of the Penrose process, of superradiance, and of the Blandford–Znajek engine that lights quasar jets. The ergosphere is where general relativity stops being a theory of how things fall and becomes a theory of how a black hole can be made to work.

Key takeaways

The ergosphere is the region between a rotating hole's event horizon and its static limit, where $\partial_t$ turns spacelike so $g_{tt}>0$ ; inside it no observer can remain at rest relative to infinity.
Its cause is frame dragging, encoded in the metric cross term $g_{t\phi}\neq0$ : the locally non-rotating frame (a ZAMO) itself circulates at $\omega=-g_{t\phi}/g_{\phi\phi}$ , rising from the Lense–Thirring tail to the rigid horizon rate $\Omega_H$ .
The static limit sits at $r_{\text{static}}=M+\sqrt{M^2-a^2\cos^2\theta}$ — it touches the horizon at the poles and bulges to $r=2M$ at the equator, for any spin.
The ergosphere lies entirely outside the horizon, so matter and light cross it both ways and escape; it is not a one-way membrane.
Because $\partial_t$ is spacelike inside, negative-energy orbits exist, capping the extractable energy at the irreducible-mass bound — up to $1-1/\sqrt2\approx29\%$ of $Mc^2$ for an extremal hole.
Key idea: where the time-translation Killing vector $\partial_t$ turns spacelike ( $g_{tt}>0$ ) you can no longer stand still relative to infinity — you are forced into corotation, and the conserved energy $-\xi\cdot p$ can go negative, which is the engine of black-hole energy extraction.

The core insight

Where the time-translation Killing vector $\partial_t$ turns spacelike ( $g_{tt}>0$ ) you can no longer stand still relative to infinity; you are forced into corotation, and the conserved energy $-\xi\cdot p$ can go negative — the engine of energy extraction.

Where this lives today

The ergosphere has moved from blackboard to observatory. The Blandford–Znajek mechanism — magnetic field lines threaded through the ergosphere and horizon, tapping the spin electromagnetically as the Penrose process suggests — is the leading model for the relativistic jets of active galactic nuclei, and in 2021 the Event Horizon Telescope's polarized image of M87's supermassive hole revealed an ordered magnetic field on the right scale to power its kiloparsec jet this way. Superradiance, the wave version of the Penrose process, amplifies any bosonic field with frequency $\omega<m\Omega_H$ ; an ultralight boson would grow a bound "cloud" around spinning holes and spin them down, so measured black-hole spins now set some of the tightest bounds on axion-like particles and ultralight dark matter. Frame dragging in its weak-field guise was confirmed in the Solar System: Gravity Probe B measured the Earth's Lense–Thirring gyroscope precession at $37.2\pm7.2$ milliarcseconds per year against a prediction of $39.2$ , and the LAGEOS/LARES satellites tracked the nodal drag of their orbits to a few percent. And every spin estimate from X-ray reflection spectroscopy or EHT imaging reads $a/M$ through frame dragging's imprint on the innermost stable orbit and the photon ring's asymmetry.

The physical picture

Imagine the inertial frames of spacetime as a thick fluid and the spinning hole as a paddle stirring it. Far out, the fluid barely moves and you can anchor against the fixed stars with an arbitrarily small rocket. Move inward and it circulates faster; to keep pointing at the same star you fire harder and harder against the swirl. At the static limit the swirl reaches the speed of light: a photon fired against the rotation is held exactly stationary at fixed angular coordinate. One step further in and even that photon is carried forward. A massive body, always slower than light, has no hope — it must corotate. Nothing about this requires the horizon; it happens in open spacetime you can still fly out of. The honest local standard of "not rotating" is the zero-angular-momentum observer, the ZAMO, who carries no angular momentum ( $L=p_\phi=0$ ) yet is seen from infinity to circulate at $\omega(r,\theta)$ . That $\omega$ is frame dragging made quantitative: it fades as $2Ma/r^3$ in the weak field — the Lense–Thirring frequency Gravity Probe B measured around the Earth — and climbs monotonically inward until, at the horizon, it locks to a single $\Omega_H$ shared by the whole surface. The horizon rotates rigidly, like a turntable, even though the dragging outside it varies with radius and latitude. The first figure shows the geometry; the second, the climb.

Figure

Poloidal cross-section of a Kerr hole with spin a = 0.9M: the event horizon is the sphere r₊ = 1.44M, and the ergosurface or static limit bulges to r = 2M at the equator while pinching onto the horizon at the poles. The amber crescent between them is the ergosphere, the region where g_tt > 0 and staticity is forbidden.

Figure

Equatorial frame-dragging rate ω(r) for spin a = 0.9M: it climbs from the Lense–Thirring tail ω ≈ 2Ma/r³ far out (dashed) to the rigid horizon value Ω_H = a/(2Mr₊) = 0.313/M at r₊ = 1.44M. Every ZAMO corotates with the hole, and the growth is monotonic all the way in.

Negative energy — the whole point of the ergosphere — is not a paradox. The energy an experimenter measures locally is $-u_{\text{obs}}\cdot p$ , contracted with the observer's future-timelike 4-velocity, and is always positive. The conserved energy is $-\xi\cdot p$ , contracted with the Killing vector $\partial_t$ . Outside the static limit the two agree in sign; inside, $\partial_t$ is spacelike, so $-\xi\cdot p$ is really a spatial momentum component, free to be negative. Nothing local ever sees a negative energy — the negativity is a statement about how the orbit connects to infinity.

The core equations

Everything begins with the Kerr metric, the unique stationary, axisymmetric, asymptotically flat vacuum solution with a horizon. In Boyer–Lindquist coordinates and units $G=c=1$ , its single off-diagonal term $g_{t\phi}$ is the entire origin of frame dragging; the rest is reading that term and the sign of $g_{tt}$ .

\mathrm{d}s^2=-\Big(1-\frac{2Mr}{\Sigma}\Big)\mathrm{d}t^2-\frac{4Mar\sin^2\theta}{\Sigma}\,\mathrm{d}t\,\mathrm{d}\phi+\frac{\Sigma}{\Delta}\mathrm{d}r^2+\Sigma\,\mathrm{d}\theta^2+\Big(r^2+a^2+\frac{2Ma^2r\sin^2\theta}{\Sigma}\Big)\sin^2\theta\,\mathrm{d}\phi^2 \qquad (1)

where: $M$ — mass; $a=J/M$ — spin per unit mass, $0\le a\le M$ ; $\Sigma=r^2+a^2\cos^2\theta$ ; $\Delta=r^2-2Mr+a^2$ ; $g_{t\phi}=-2Mar\sin^2\theta/\Sigma$ — the frame-dragging cross term, nonzero only for $a>0$ . Setting $a=0$ collapses this to Schwarzschild. The static limit is defined by the norm of $\xi=\partial_t$ . A static observer keeps $r,\theta,\phi$ fixed and so moves along $\xi$ ; this is timelike, hence physically possible, only where $g_{tt}=\xi\cdot\xi<0$ . The ergosphere is precisely the set $g_{tt}>0$ outside the horizon.

g_{tt}=-\Big(1-\frac{2Mr}{\Sigma}\Big)=\frac{2Mr-\Sigma}{\Sigma},\qquad r_{\text{static}}=M+\sqrt{M^2-a^2\cos^2\theta} \qquad (2)

where: $g_{tt}<0$ permits static observers; $g_{tt}=0$ marks the ergosurface; $g_{tt}>0$ is the ergoregion. Reading it: the static limit is a surface of revolution that sits at $r=2M$ on the equator ( $\cos\theta=0$ ) and shrinks to $r=M+\sqrt{M^2-a^2}$ at the poles. The inner boundary of the ergosphere is the outer event horizon, found where the coordinate radius degenerates, $\Delta=0$ . Comparing (2) and (3) shows the two surfaces meet exactly at the poles and separate everywhere else.

\Delta=r^2-2Mr+a^2=0\ \Rightarrow\ r_\pm=M\pm\sqrt{M^2-a^2} \qquad (3)

where: $r_+$ — outer (event) horizon, the inner wall of the ergosphere; $r_-$ — inner (Cauchy) horizon. Reading it: $r_{\text{static}}(\theta)\ge r_+$ always, with equality only at $\theta=0,\pi$ , so the ergosphere is a crescent thickest at the equator. Frame dragging is quantified by the angular velocity of the ZAMO, the observer whose angular momentum vanishes but who is nonetheless dragged around. It is fixed entirely by the metric, with no reference to any matter.

\omega(r,\theta)=-\frac{g_{t\phi}}{g_{\phi\phi}}=\frac{2Mar}{(r^2+a^2)\Sigma+2Ma^2r\sin^2\theta} \qquad (4)

where: $\omega=\mathrm{d}\phi/\mathrm{d}t$ of a zero-angular-momentum observer. Reading it: $\omega>0$ wherever $a>0$ (since $g_{t\phi}<0$ ), so the drag always follows the spin; far away $\omega\to2Ma/r^3$ , the Lense–Thirring law. Evaluated at the horizon, $\omega$ loses its $r,\theta$ dependence and becomes a single constant — the horizon rotates as a rigid body. This $\Omega_H$ is the frequency that appears in the first law of black-hole mechanics and sets the superradiant threshold.

\Omega_H=\omega(r_+)=\frac{a}{r_+^2+a^2}=\frac{a}{2Mr_+} \qquad (5)

where: $\Omega_H$ — the rigid angular velocity of the event horizon. Reading it: at extremal spin $a=M$ we have $r_+=M$ and $\Omega_H=1/2M$ , the fastest a horizon can turn. Finally, the conserved energy that can go negative. Because $\partial_t$ is a Killing vector, $E=-\xi\cdot p=-p_t$ is constant along any geodesic; where $\xi$ is spacelike this constant may be negative even for a future-directed particle.

E=-\xi_\mu p^\mu=-p_t,\qquad E<0\ \text{allowed where}\ g_{tt}>0;\qquad M^2=M_{\text{irr}}^2+\frac{J^2}{4M_{\text{irr}}^2} \qquad (6)

where: $E$ — conserved energy at infinity; $M_{\text{irr}}$ — irreducible mass, the part that cannot be extracted. Reading it: absorbing a negative- $E$ fragment lowers $M$ and $J$ together but never lowers $M_{\text{irr}}$ , so extraction halts at $a=0$ .

Derivation AThe static limit and the shape of the ergosphere

The goal is to locate the outer boundary of the ergosphere from first principles and explain its distinctive crescent shape — pinched onto the horizon at the poles, ballooning to twice the mass at the equator. We impose the physical condition defining a static observer, translate it into a condition on the metric, solve the resulting quadratic in $r$ , and compare the answer to the horizon radius latitude by latitude so the geometry becomes concrete rather than a formula.

Step 1. A static observer holds $r$ , $\theta$ , and $\phi$ fixed, so their 4-velocity is proportional to $\xi=\partial_t$ . For this worldline to be physical it must be timelike: $\xi\cdot\xi=g_{tt}<0$ . The boundary of possibility is therefore $g_{tt}=0$ . Why: "standing still relative to infinity" is not a coordinate choice but a geometric statement — you are following an orbit of the time-translation symmetry — so its failure is dictated purely by the sign of that vector's norm.

Step 2. Insert the metric component $g_{tt}=-(1-2Mr/\Sigma)=(2Mr-\Sigma)/\Sigma$ and set the numerator to zero, giving $2Mr=\Sigma=r^2+a^2\cos^2\theta$ . Rearranged this is the quadratic $r^2-2Mr+a^2\cos^2\theta=0$ . Why: $\Sigma>0$ everywhere outside the ring singularity, so the sign of $g_{tt}$ is carried entirely by its numerator $2Mr-\Sigma$ , and the zero set is a clean quadratic.

Step 3. Take the outer root, $r_{\text{static}}=M+\sqrt{M^2-a^2\cos^2\theta}$ . This is the ergosurface. Compare it to the horizon $r_+=M+\sqrt{M^2-a^2}$ from $\Delta=0$ : the only difference is $\cos^2\theta$ under one square root versus $1$ under the other. Why: writing both surfaces in the same closed form makes their relationship transparent — they differ only through the latitude factor multiplying $a^2$ .

Step 4. Evaluate the latitudes. At the poles $\cos^2\theta=1$ , so $r_{\text{static}}=r_+$ — the surfaces touch. At the equator $\cos\theta=0$ , so $r_{\text{static}}=M+\sqrt{M^2}=2M$ , independent of spin. In between, $r_{\text{static}}>r_+$ strictly, so the ergosphere is a crescent that vanishes at the spin axis and is thickest, $2M-r_+$ , on the equator. Why: this is the whole geometry of the figure — the "pumpkin" ergosurface enclosing the smaller horizon sphere, kissing it only at the poles — read directly off two square roots.

Go deeper · Derivation B · 2 minWhy every observer inside the ergosphere must corotate

Derivation BWhy every observer inside the ergosphere must corotate

The goal is to prove the defining kinematic fact of the ergoregion: not merely that you cannot stay still, but that you are forced forward in the direction of the spin, with a strictly positive minimum angular velocity. We restrict to circular motion, impose the timelike condition, read off the allowed band of angular velocities as the roots of a quadratic, and show that inside the ergosphere both roots lie on the same, positive, side of zero. Along the way the static limit reappears as the surface where the counter-rotating photon is marginally frozen.

Step 1. Consider an observer at fixed $r,\theta$ circling with angular velocity $\Omega=\mathrm{d}\phi/\mathrm{d}t$ . Dividing the line element by $\mathrm{d}t^2$ , the timelike condition $\mathrm{d}s^2<0$ becomes $g_{tt}+2g_{t\phi}\Omega+g_{\phi\phi}\Omega^2<0$ . Why: any physical circular orbit must have a future-timelike tangent, and for purely azimuthal motion that reduces to a single quadratic inequality in $\Omega$ .

Step 2. The quadratic $g_{\phi\phi}\Omega^2+2g_{t\phi}\Omega+g_{tt}=0$ has roots $\Omega_\pm=\big(-g_{t\phi}\pm\sqrt{g_{t\phi}^2-g_{tt}g_{\phi\phi}}\big)/g_{\phi\phi}=\omega\pm\sqrt{\omega^2-g_{tt}/g_{\phi\phi}}$ . Since $g_{\phi\phi}>0$ , the inequality holds between the roots: $\Omega_-<\Omega<\Omega_+$ . Why: $\Omega_\pm$ are exactly the angular velocities of the co- and counter-rotating photons emitted tangentially, so every material observer's angular velocity is trapped inside the light cone they bound.

Step 3. Outside the static limit $g_{tt}<0$ , so the product of roots $\Omega_-\Omega_+=g_{tt}/g_{\phi\phi}<0$ : the roots straddle zero and $\Omega=0$ (staying still) is allowed. On the static limit $g_{tt}=0$ , so $\Omega_-=0$ exactly — the counter-rotating photon is momentarily frozen at fixed $\phi$ . Why: the vanishing of the lower root is an equivalent definition of the ergosurface, and it is the precise sense in which "light itself can just barely hold still" there.

Step 4. Inside the ergosphere $g_{tt}>0$ , so $\Omega_-\Omega_+=g_{tt}/g_{\phi\phi}>0$ : both roots share a sign. The upper root $\Omega_+$ is manifestly positive, hence $\Omega_->0$ as well, and the allowed band $\Omega_-<\Omega<\Omega_+$ lies entirely at positive $\Omega$ . Every observer must corotate; standing still ( $\Omega=0$ ) has dropped out of the light cone. Why: this is corotation as a theorem — no rocket, however powerful, can supply an angular velocity outside $(\Omega_-,\Omega_+)$ , because doing so would make the worldline spacelike.

Go deeper · Derivation C · 2 minNegative energy, the Penrose process, and the irreducible-mass bound

Derivation CNegative energy, the Penrose process, and the irreducible-mass bound

The goal is to show why the ergosphere permits rotational energy to be mined and to derive the hard limit on how much. We start from the distinction between locally measured energy and the conserved energy at infinity, show the latter can go negative inside the ergoregion, follow a fragmentation that carries surplus energy out, and track the hole's response to obtain Christodoulou's irreducible mass, which never decreases and caps the yield near twenty-nine percent for a maximally spinning hole.

Step 1. The energy a local observer measures is $-u_{\text{obs}}\cdot p$ , contracted with a future-timelike 4-velocity, and is always positive for a future-directed $p$ . The conserved energy is $E=-\xi\cdot p$ , contracted with the Killing vector. Inside the ergosphere $\xi$ is spacelike, so $\xi\cdot p$ can be positive — that is, $E<0$ — with $p$ still future-directed. Why: negativity of $E$ is not a local pathology but a statement about the Killing structure; it is impossible to detect with any local energy measurement.

Step 2. Send in a particle of energy $E_0>0$ that splits inside the ergoregion into two fragments, $E_0=E_1+E_2$ . Arrange the split so the second fragment has $E_2<0$ ; conservation then forces $E_1=E_0-E_2>E_0$ . Let fragment 2 fall through the horizon and fragment 1 escape to infinity. Why: the escaping fragment leaves with more energy than the whole system brought in — the surplus $-E_2$ has been drawn from the hole, which is the Penrose process in one line.

Step 3. The captured fragment changes the hole by $\delta M=E_2<0$ and $\delta J=L_2$ . Future-directedness across the horizon requires the fragment's momentum to satisfy $p\cdot(\xi+\Omega_H\psi)<0$ , where $\psi=\partial_\phi$ , which reduces to $E_2\ge\Omega_H L_2$ , i.e. $\delta M\ge\Omega_H\,\delta J$ . Because $E_2<0$ this drives $J$ down too. Why: the horizon's rigid rotation $\Omega_H$ is exactly the exchange rate between the mass you remove and the angular momentum you must remove with it.

Step 4. Rewrite the inequality using the irreducible mass defined by $M^2=M_{\text{irr}}^2+J^2/(4M_{\text{irr}}^2)$ , equivalently $M_{\text{irr}}^2=\tfrac12\big(M^2+\sqrt{M^4-J^2}\big)$ and $A=16\pi M_{\text{irr}}^2$ . The bound $\delta M\ge\Omega_H\,\delta J$ becomes $\delta M_{\text{irr}}\ge0$ : the irreducible mass, and with it the horizon area, can never decrease. Why: this is Hawking's area theorem seen through the Penrose process — the extractable energy is $M-M_{\text{irr}}$ , and once $J\to0$ you are stuck with $M_{\text{irr}}$ .

Step 5. At extremal spin $J=M^2$ , so $M_{\text{irr}}=M/\sqrt2$ and the maximum yield is $(M-M_{\text{irr}})/M=1-1/\sqrt2\approx0.29$ . Up to twenty-nine percent of a maximal Kerr hole's mass-energy is rotational and extractable. Why: this fixes the astrophysical ceiling — black-hole rotation is one of the most concentrated energy reservoirs known, and this number is why it can power the universe's brightest engines.

Limiting cases

Case	Result	Meaning
Schwarzschild limit $a\to0$	$r_{\text{static}}=M+\sqrt{M^2-a^2\cos^2\theta}\to2M=r_+$	Static limit falls onto the horizon; $g_{t\phi}=0$ , no frame dragging, no ergosphere.
Extremal Kerr $a\to M$	$r_+=M,\quad r_{\text{static}}=M+M\sin\theta$	Horizon at $r=M$ ; ergosphere persists, reaching $2M$ at the equator and pinching to $M$ at the poles; $\Omega_H=1/2M$ .
Weak field / large $r$	$\omega\approx\dfrac{2Ma}{r^3}=\dfrac{2GJ}{c^2r^3}$	Lense–Thirring dragging falls as $r^{-3}$ ; the regime probed by Gravity Probe B and LAGEOS.
Equatorial plane $\theta=\pi/2$	$r_{\text{static}}=2M$ (any $a$ )	Outer ergosurface always at $r=2M$ on the equator — the ergosphere is thickest there.

Worked examples

1 · The ergosphere of a rapidly spinning hole (confidence builder)

A Kerr black hole has spin $a=0.9M$ . Find the horizon radius, the equatorial and $45^\circ$ static-limit radii, the equatorial ergosphere thickness, and the horizon angular velocity, in units of $M$ .

Step 1. Horizon: $r_+=M+\sqrt{M^2-a^2}=M(1+\sqrt{0.19})=\mathbf{1.436\,M}$ . The inner wall of the ergosphere sits here.

Step 2. Static limit at the equator ( $\cos\theta=0$ ): $r_{\text{static}}=M+\sqrt{M^2}=\mathbf{2M}$ , exactly, for every spin. At $45^\circ$ ( $\cos^2\theta=\tfrac12$ ): $r_{\text{static}}=M(1+\sqrt{1-0.81\cdot0.5})=M(1+\sqrt{0.595})=\mathbf{1.771\,M}$ , between the polar and equatorial extremes.

Step 3. Equatorial thickness: $\Delta r=r_{\text{static}}-r_+=2M-1.436M=\mathbf{0.564\,M}$ — a substantial shell, over half a gravitational radius deep.

Step 4. Horizon angular velocity: $\Omega_H=a/(2Mr_+)=0.9M/(2M\cdot1.436M)=\mathbf{0.313/M}$ . As a check, $a/(r_+^2+a^2)=0.9/(2.062+0.81)=0.313/M$ — the two forms agree, since $\Delta=0$ gives $r_+^2+a^2=2Mr_+$ . Key results: r₊ = 1.436M, static limit = 2M (equator), thickness = 0.564M, Ω_H = 0.313/M.

2 · How much energy can you mine from a spinning black hole?

Take the same $a=0.9M$ hole with $M=10\,M_\odot$ . Find the irreducible mass, the maximum extractable energy in joules, compare it to the Sun's lifetime output, and find how fast the horizon spins.

Step 1. With $J=aM=0.9M^2$ , the irreducible mass is $M_{\text{irr}}^2=\tfrac12\big(M^2+\sqrt{M^4-J^2}\big)=\tfrac12 M^2\big(1+\sqrt{1-0.81}\big)=0.718\,M^2$ , so $M_{\text{irr}}=\mathbf{0.847\,M}$ . The extractable fraction is $1-M_{\text{irr}}/M=\mathbf{0.153}$ , about 15% of the total mass-energy.

Step 2. In joules, $Mc^2=10(1.989\times10^{30})(2.998\times10^8)^2=1.79\times10^{48}\ \mathrm{J}$ , so the extractable rotational energy is $0.153\times1.79\times10^{48}=\mathbf{2.7\times10^{47}\ J}$ .

Step 3. For scale, the Sun radiates about $1.2\times10^{44}\ \mathrm{J}$ over a $10^{10}$ -year main-sequence lifetime. This single stellar-mass hole holds $2.7\times10^{47}/1.2\times10^{44}\approx\mathbf{2000}$ times the Sun's entire lifetime output as extractable spin energy.

Step 4. Physical horizon spin: $\Omega_H=0.313/M$ becomes $\Omega_H=0.313\,c^3/(GM)$ in SI. With $GM/c^3=10\times4.93\times10^{-6}\,\mathrm{s}=4.93\times10^{-5}\,\mathrm{s}$ , $\Omega_H=0.313/(4.93\times10^{-5})=\mathbf{6.4\times10^{3}\ rad/s}$ , a horizon rotation period $2\pi/\Omega_H\approx\mathbf{0.99\ ms}$ . Key results: M_irr = 0.847M, extractable ≈ 15% ≈ 2.7×10⁴⁷ J ≈ 2000 solar lifetimes, horizon period ≈ 1 ms.

Physics of the portrait

The ergosphere arrived backward: the Kerr solution came first, in 1963, and its most consequential feature was read out only afterward, by a topologist asking how to get energy out of a rotating hole. The story braids three names — Kerr, who wrote the metric; Lense and Thirring, who computed the weak-field drag half a century earlier; and Penrose, who saw that the drag, taken to its extreme, hands you a machine.

Origins · 1 minRoger Penrose and how this was discovered

Roger Penrose

b. 1931 · black-hole energy extraction and the ergosphere

Roger Penrose is the mathematical physicist who, more than anyone, turned general relativity's black holes from curiosities into engines and inevitabilities. Trained as a pure mathematician, he brought topology and global methods to a subject dominated by explicit coordinate solutions. In 1965 his singularity theorem showed that gravitational collapse must produce a spacetime singularity, as a generic consequence of trapped surfaces — work recognized with the 2020 Nobel Prize in Physics. In 1969, in a short paper in Rivista del Nuovo Cimento, he pointed out that the region outside a Kerr hole's horizon where $\partial_t$ turns spacelike — soon named the ergosphere, from the Greek ergon, "work" — permits orbits of negative conserved energy, so a body fragmenting there could send a piece to infinity with more energy than it arrived with, at the hole's expense. The Penrose process was the first concrete mechanism for tapping a black hole's spin. His student Demetrios Christodoulou quantified the limit in 1970 with the irreducible mass — a result Hawking soon subsumed into the area theorem and the thermodynamic analogy. Penrose's ergosphere is the conceptual hinge on which superradiance, the Blandford–Znajek jet mechanism, and black-hole thermodynamics all turn.

History

1918

— Lense and Thirring derive the frame-dragging precession around a rotating mass

1963

— Kerr finds the exact metric of a rotating black hole

1969

— Penrose shows rotational energy can be extracted via negative-energy orbits in the ergosphere

1970

— Christodoulou defines the irreducible mass, bounding the extractable energy

2011

— Gravity Probe B reports the measured frame-dragging of the Earth

Going further

References. Misner, Thorne & Wheeler, Gravitation (1973), §33 · Wald, General Relativity (1984), §12.3–12.4 · Carroll, Spacetime and Geometry (2004), §6.6–6.7 · Hartle, Gravity (2003), Ch. 15 · Schutz, A First Course in General Relativity, 2nd ed., §11 · Bardeen, Press & Teukolsky, ApJ 178 (1972) 347 · Penrose, Riv. Nuovo Cimento 1 (1969) 252.

Provenance. The Kerr metric and Boyer–Lindquist form follow MTW §33 and Wald §12.3. The ergosurface $r=M+\sqrt{M^2-a^2\cos^2\theta}$ , horizons $r_\pm=M\pm\sqrt{M^2-a^2}$ , $\Omega_H=a/(2Mr_+)$ , and $\omega=-g_{t\phi}/g_{\phi\phi}$ are verifiable from Carroll §6.6–6.7. The irreducible-mass relation and the $29\%$ extremal yield are due to Christodoulou (1970) and Wald §12.4. The worked numbers ( $a=0.9M$ : $r_+=1.436M$ , $\Omega_H=0.313/M$ , $M_{\text{irr}}=0.847M$ , extractable $15.3\%$ ) and the physical horizon period ( $\approx1\,\mathrm{ms}$ at $10\,M_\odot$ ) were recomputed for this article; the Lense–Thirring $\omega\approx2GJ/c^2r^3$ and the Gravity Probe B value ( $37.2\pm7.2\,\mathrm{mas/yr}$ ) are standard. A verifier should recheck $\Omega_H=1/2M$ at extremality and $\Omega_-=0\Leftrightarrow g_{tt}=0$ .

Builds on. Kerr Metric and Rotating Black Holes · Event Horizon and Apparent Horizon · Killing Horizons and Surface Gravity · Schwarzschild Black Hole Structure. Leads to. Penrose Process and Energy Extraction · Superradiance and the Black Hole Bomb · Four Laws of Black Hole Mechanics · Kerr–Newman Solution.

Ask the tutor: Ask me to walk you from the sign flip in $g_{tt}$ all the way to a working Penrose engine — or to show why the horizon spins as one rigid piece while the drag outside it does not.

Is the ergosphere the same thing as the event horizon?

No, and keeping them distinct is the whole point. The event horizon at $r_+$ is where $\Delta=0$ — a one-way membrane you cannot escape once crossed. The ergosphere is the shell outside that horizon, bounded above by the static limit where $g_{tt}=0$ . You can enter and leave the ergosphere freely; light and rockets cross its outer skin in both directions. The only thing you forfeit inside it is the ability to remain at a fixed angular position: frame dragging forces you to orbit in the hole's direction. The two surfaces touch at the poles and separate elsewhere, and the gap between them — the ergosphere proper — is exactly the region where energy extraction becomes possible.

Why can't the most powerful rocket imaginable hold you still inside the ergosphere?

Because staticity is a geometric condition, not a matter of thrust. To "stand still relative to infinity" means to follow an orbit of $\partial_t$ , and that worldline is timelike — a possible path for a massive body — only where $g_{tt}<0$ . Inside the ergosphere $g_{tt}>0$ , so $\partial_t$ is spacelike and a static worldline would be spacelike too, i.e. faster than light. No finite proper acceleration turns a spacelike curve into a timelike one; the light cones have tilted so far in the $\phi$ direction that "not moving in $\phi$ " falls outside every cone. The rocket is irrelevant — the constraint is on the causal structure.

How can conserved energy be negative without violating local physics?

There are two different "energies" and only one can go negative. The energy a local observer measures is $-u_{\text{obs}}\cdot p$ , formed with a future-timelike 4-velocity; it is strictly positive for any real particle, everywhere, including deep inside the ergosphere. The conserved quantity is $E=-\xi\cdot p$ , formed with the Killing vector $\partial_t$ . Outside the static limit $\partial_t$ is timelike and the two nearly coincide; inside, $\partial_t$ is spacelike, so $E$ is really a spatial momentum component, and momentum components are freely signed. A particle whirling forward can have negative $E$ while every local measurement stays positive — two inequivalent projections of the same 4-momentum, no contradiction.

What is a ZAMO and why does it matter?

A ZAMO is a zero-angular-momentum observer: one whose conserved $p_\phi$ is zero, yet who is dragged around the hole at $\omega=-g_{t\phi}/g_{\phi\phi}$ . It is the closest thing Kerr spacetime has to a "locally non-rotating" frame — a gyroscope it carries points at fixed local landmarks even though, seen from infinity, the whole frame revolves. ZAMOs matter because relative to them physics looks clean and local, while relative to infinity it carries the full imprint of frame dragging. The ZAMO's $\omega$ is the practical face of frame dragging, interpolating from the Lense–Thirring rate far out to the rigid $\Omega_H$ at the horizon.

Does frame dragging require a black hole?

No. Any rotating mass drags spacetime; the effect is universal and was derived by Lense and Thirring in 1918, long before Kerr's solution. The Earth drags a gyroscope's axis by about $39$ milliarcseconds per year, which Gravity Probe B measured directly, and the same drag shifts the orbital planes of laser-ranged satellites. What is special about a black hole is only the magnitude: near a fast-spinning compact object the dragging grows so strong it forbids staticity outright, creating an ergosphere. Around the Earth it is a tiny perturbation; around a maximal Kerr hole it dominates the causal structure. Same physics, wildly different regime.

Why does the extractable energy stop at the irreducible mass?

Because each act of extraction lowers the hole's spin, and once the spin is gone there is no ergosphere left to work with. The negative-energy fragment fed into the horizon must satisfy $\delta M\ge\Omega_H\,\delta J$ ; rewriting this via the irreducible mass — defined so $M^2=M_{\text{irr}}^2+J^2/(4M_{\text{irr}}^2)$ — shows $M_{\text{irr}}$ can never decrease. The removable energy is $M-M_{\text{irr}}$ , pure rotational energy; when $J\to0$ it vanishes and $M=M_{\text{irr}}$ . For a maximal hole the ceiling is $1-1/\sqrt2\approx29\%$ . Since $M_{\text{irr}}$ is proportional to the square root of the horizon area, "cannot extract below $M_{\text{irr}}$ " is the same statement as "the horizon area cannot shrink."