Superradiance and the Black Hole Bomb

Superradiance is what the Penrose process becomes when you replace the test particle with a wave. In the particle picture you split a trajectory inside the ergosphere into a positive- and a negative-energy fragment; in the wave picture you send a single monochromatic mode of a bosonic field at a spinning black hole and find the reflected wave carrying more energy than you sent in. Nothing has been created from nothing: the part of the wave that fell through the horizon carried negative energy and negative angular momentum, so the hole paid for the amplification out of its own rotation — it spun down a little, and its irreducible mass, hence its horizon area, still grew. The whole effect is governed by one inequality: a mode with time-dependence $e^{-i\omega t+im\phi}$ is amplified precisely when $0<\omega<m\,\Omega_H$ , where $\Omega_H$ is the angular velocity of the horizon. Below that threshold the wave co-rotates slower than the dragged spacetime, the horizon sees it as a negative-frequency disturbance, and it acts as a source rather than a sink. A single bounce yields at most a few percent for scalar and electromagnetic waves and up to $138\%$ for gravitational ones — modest, until you add a mirror. Confine the amplified wave so it scatters again and again and the gain compounds into an exponential instability: the "black hole bomb." A massive field supplies its own mirror through its rest mass, growing a macroscopic "gravitational atom" cloud around the hole — which makes spinning black holes the most sensitive detectors we have for ultralight bosons.

Key takeaways

Superradiance is the wave analog of the Penrose process: a bosonic mode of frequency $\omega$ and azimuthal number $m$ scattering off a Kerr hole reflects with $|\mathcal R|^2>1$ whenever $0<\omega<m\,\Omega_H$ , with $\Omega_H$ the horizon angular velocity.
The energy is mined from the hole's spin. The second law $\delta A\ge0$ survives because the transmitted part carries negative energy and negative angular momentum across the horizon.
The threshold $\omega=m\Omega_H$ is the statement that the flux measured by the horizon generator $\chi=\partial_t+\Omega_H\,\partial_\phi$ changes sign — below it, that flux is negative.
One scattering is weak (scalar $\sim0.4\%$ , electromagnetic $\sim4.4\%$ , gravitational up to $\sim138\%$ ), but a reflecting mirror turns per-bounce gain into exponential growth: the black hole bomb.
A massive boson provides its own mirror through the potential barrier of its rest mass, seeding a self-confined instability that grows a hydrogen-like "gravitational atom" and spins the hole down.
Charged superradiance is the electrostatic counterpart for Reissner-Nordström and Kerr-Newman holes, with $\omega<q\Phi_H$ replacing the rotational condition.
Key idea: energy at infinity is the flux of the timelike Killing vector $\partial_t$ , but the conserved quantity crossing a Kerr horizon is the flux of the horizon generator $\chi=\partial_t+\Omega_H\partial_\phi$ . When $\omega<m\Omega_H$ that flux is negative, so the hole gives up energy and angular momentum to the wave.

The core insight

Energy as measured at infinity is set by the timelike Killing vector $\partial_t$ , but the locally conserved quantity crossing the horizon is the flux of the horizon Killing vector $\chi=\partial_t+\Omega_H\partial_\phi$ ; when $\omega<m\Omega_H$ that flux is negative, so the hole gives up energy and angular momentum to the wave and the reflection is amplified.

Where this lives today

Superradiance has become a precision instrument for particle physics done with telescopes. The core argument is clean: any sufficiently light boson — a QCD axion, a string-theory axiverse field, a dark photon — with Compton wavelength near a black hole's size would trigger a self-confined instability, grow a macroscopic cloud, and spin the hole down on timescales from minutes to megayears. So the observed spins of real black holes exclude bosons in specific mass windows: roughly $10^{-13}$ – $10^{-11}\,$ eV from stellar-mass holes measured in X-ray binaries and gravitational-wave events, and $\sim10^{-18}$ – $10^{-16}\,$ eV from supermassive holes. This "black hole spin gap" is one of the few laboratory-free probes of physics at these masses. The clouds are also direct sources: a coherent cloud radiates near-monochromatic continuous gravitational waves at $f\approx2\mu/h$ , and LIGO-Virgo-KAGRA run dedicated searches for exactly these signals, while in a binary the cloud's level transitions imprint resonances on the inspiral — the emerging field of "gravitational collider physics." Superradiance is also the linear, test-field face of the same rotational-energy extraction that, in its magnetized nonlinear form, powers the Blandford-Znajek mechanism behind relativistic jets, both capped by the same $\delta A\ge0$ bound.

The physical picture

Picture the ergosphere as a region of compulsory co-rotation: between the outer horizon and the static limit, frame dragging is so severe that no observer, and no wave, can stand still relative to infinity. A mode that tries to co-rotate more slowly than the dragging — one with $\omega/m<\Omega_H$ — looks, to the horizon's own clock, like a negative-frequency disturbance. Feeding it into the hole lowers the hole's energy, and conservation forces the balance back out through the reflected wave, amplified. The horizon is the perfect one-way absorber that completes the trick: it can only swallow, never classically emit, so it is the ideal place to dump the negative-energy part. The cleanest mechanical analog is Zel'dovich's rotating cylinder: shine a wave of frequency $\omega$ and angular momentum $m$ on an absorbing body spinning at $\Omega$ , and if $\omega<m\Omega$ the rotation Doppler-shifts it into "negative absorption" and the body amplifies it. Replace the cylinder with a Kerr hole and its horizon becomes the dissipator, its rotation the reservoir. The first figure fixes the geometry and the amplified reflection; the second is what a numerical relativist actually computes — the amplification factor versus frequency.

Figure

A wave of frequency ω and azimuthal number m scatters off a Kerr hole: inside the superradiant window with ω between 0 and mΩ_H the reflected wave carries more energy than the incident one, while a negative-energy flux crosses the horizon and spins the hole down. For a/M = 0.9 the m = 1 window is ω below Ω_H = 0.31/M

The gain of a single scattering is small for the light fields but reaches order unity for gravitational waves, and it turns off sharply at the threshold $\omega=m\Omega_H$ : below it the reflected energy exceeds the incident, above it ordinary absorption resumes and the curve crosses zero. That sign change — positive gain in a bounded window, zero at the edge — is the entire content of superradiance, and what a mirror or a field mass exploits to build an instability. The second figure plots the strongest case, the gravitational $\ell=m=2$ mode off a near-extremal hole.

Figure

Superradiant amplification of a gravitational ℓ = m = 2 wave off a near-extremal Kerr hole a/M → 1 versus Mω: the reflected energy gain Z, the fractional excess of reflected over incident energy, is positive throughout the window with ω between 0 and 2Ω_H, peaks near 138 percent, and crosses zero at the edge ω = 2Ω_H. Scalar waves reach only 0.4 percent and electromagnetic waves 4.4 percent

The core equations

The whole subject rests on the geometry of the Kerr horizon, so the first quantity to fix is its angular velocity. Because the horizon is a Killing horizon of $\chi=\partial_t+\Omega_H\partial_\phi$ , every point on it rotates rigidly at one rate $\Omega_H$ , the angular velocity a distant observer would ascribe to the hole's surface.

\Omega_H=\frac{a}{r_+^2+a^2}=\frac{a}{2Mr_+},\qquad r_+=M+\sqrt{M^2-a^2} \qquad (1)

where: $\Omega_H$ — angular velocity of the event horizon; $a=J/M$ — spin parameter (geometrized units $G=c=1$ ); $r_+$ — outer horizon Boyer-Lindquist radius; $M,\ J$ — hole mass and angular momentum. Reading: the two forms agree because $r_+^2-2Mr_++a^2=0$ on the horizon, so $r_+^2+a^2=2Mr_+$ ; $\Omega_H$ runs from $0$ (Schwarzschild) to $1/2M$ (extremal). The central result of the theory is the amplification window itself. A bosonic mode is superradiant exactly when its frequency, red-shifted into the horizon's co-rotating frame, comes out negative.

0<\omega<m\,\Omega_H \qquad (2)

where: $\omega$ — wave frequency measured at infinity (taken positive); $m$ — azimuthal quantum number, a positive integer for a co-rotating mode. Reading: the horizon generator acting on $e^{-i\omega t+im\phi}$ gives $\chi^a\partial_a=-i(\omega-m\Omega_H)$ , so $\omega-m\Omega_H$ is the frequency measured at the horizon; when it is negative the wave delivers negative energy there. Near the horizon the separated radial mode obeys a Schrödinger-like equation whose "energy" is the squared horizon frequency, which is what makes $\omega-m\Omega_H$ the natural wavenumber of the ingoing wave.

\frac{\mathrm{d}^2 R}{\mathrm{d}r_*^2}+\Big[(\omega-m\Omega_H)^2-V_{\rm eff}\Big]R=0\quad (r\to r_+) \qquad (3)

where: $R$ — radial part of the separated mode; $r_*$ — tortoise coordinate, $\mathrm{d}r_*=\frac{r^2+a^2}{\Delta}\,\mathrm{d}r$ ; $\Delta=r^2-2Mr+a^2$ , vanishing at $r_+$ ; $V_{\rm eff}$ — effective scattering potential, which tends to $0$ at the horizon and to $\omega^2$ (massless) or $\mu^2$ (massive) at infinity. Reading: because $r_*\to-\infty$ at the horizon, the horizon is pushed to infinite tortoise distance and the mode there is a pure plane wave with wavenumber $k_H=\omega-m\Omega_H$ . Matching the ingoing horizon wave to the incident-plus-reflected wave at infinity, and using flux conservation, gives the amplification relation directly.

|\mathcal R|^2=1-\frac{\omega-m\Omega_H}{\omega}\,|\mathcal T|^2 \qquad (4)

where: $\mathcal R$ — reflection amplitude at infinity; $\mathcal T$ — transmission amplitude at the horizon. Reading: the coefficient $(\omega-m\Omega_H)/\omega$ is negative throughout the window, so $|\mathcal R|^2>1$ ; outside the window it is positive and $|\mathcal R|^2<1$ , ordinary absorption. Treating one scattering as a quasi-static change of the hole, the mode's fixed ratio of angular momentum to energy ties $\delta J$ to $\delta M$ and both to the energy handed to the field.

\delta M=\frac{\omega}{m}\,\delta J,\qquad \delta M=\Omega_H\,\delta J-\delta E_{\rm wave} \qquad (5)

where: $\delta M$ — change in hole mass; $\delta J$ — change in hole angular momentum; $\delta E_{\rm wave}>0$ — energy delivered to the radiation. Reading: a quantum $\hbar\omega$ carries angular momentum $\hbar m$ , so $\delta J/\delta M=m/\omega$ ; in the window both $\delta M<0$ and $\delta J<0$ — the hole loses mass and spin. The area theorem is not violated, and equation (6) shows why: the combination that controls the area is a perfect sign-definite object in the window.

\frac{\kappa}{8\pi}\,\delta A=\delta M-\Omega_H\,\delta J=\frac{\omega-m\Omega_H}{m}\,\delta J\ \ge 0 \qquad (6)

where: $A$ — horizon area; $\kappa=\dfrac{r_+-r_-}{2(r_+^2+a^2)}$ — surface gravity; $r_-=M-\sqrt{M^2-a^2}$ — inner horizon. Reading: with $\delta J<0$ and $(\omega-m\Omega_H)<0$ the product is positive, so $\delta A\ge0$ even as $M$ falls — energy is mined but irreversibility holds. For a massive field the rest mass acts as a confining wall, binding the mode into hydrogen-like levels. The controlling parameter is a gravitational fine-structure constant, and the spectrum is Coulombic.

\alpha\equiv\frac{GM\mu}{\hbar c}=\frac{r_g}{\lambda_C},\qquad \omega_R\simeq\frac{\mu c^2}{\hbar}\Big(1-\frac{\alpha^2}{2n^2}\Big) \qquad (7)

where: $\mu$ — boson rest mass; $\alpha$ — gravitational fine-structure constant; $r_g=GM/c^2$ ; $\lambda_C=\hbar/\mu c$ — reduced Compton wavelength; $n$ — principal quantum number; $\omega_R$ — real part of the bound-state frequency, whose rest-mass floor is the Compton frequency $\mu c^2/\hbar$ . Reading: the far-field $1/r$ potential produces exactly the hydrogen spectrum, with $\alpha$ as the fine-structure constant. The imaginary part of the bound-state frequency is the growth rate, and its steep power-law dependence on $\alpha$ is why the instability is efficient only in a narrow mass window.

\Gamma\propto(m\Omega_H-\omega_R)\,\alpha^{4\ell+5},\qquad \tau_{\rm inst}\sim\frac{1}{\Gamma} \qquad (8)

where: $\Gamma$ — imaginary part of the mode frequency (growth rate when positive); $\ell$ — orbital quantum number; $\tau_{\rm inst}$ — e-folding time. Reading: for the fastest mode $\ell=1$ the rate scales as $\alpha^9$ , so pushing $\alpha$ from $0.1$ to $0.4$ speeds growth by more than four orders of magnitude — the window is sharply peaked around $\alpha\sim O(0.1\text{–}1)$ . Finally, charge plays the same role as spin. A charged field scattering off a charged hole is superradiant below the chemical potential set by the horizon's electrostatic potential.

\omega<m\,\Omega_H+q\,\Phi_H \qquad (9)

where: $q$ — charge of the field quantum; $\Phi_H$ — electrostatic potential of the horizon. Reading: $\Omega_H$ and $\Phi_H$ are the two "intensive" chemical potentials of the first law of black hole mechanics, and each opens its own superradiant channel.

Derivation ASuperradiance from a conserved Wronskian

The goal is to prove $|\mathcal R|^2>1$ for a massless scalar scattering off Kerr using nothing but the radial Teukolsky equation, its two asymptotic regions, and the fact that a real potential forces the Wronskian of two solutions to be constant. We separate the field, read off the ingoing horizon wave and the incident-plus-reflected wave at infinity, form the Wronskian of the mode with its conjugate, and equate its value at the two ends. The sign of the reflected flux falls out with no appeal to energy conditions or the first law.

Step 1. Separate the Klein-Gordon field as $\Phi=\sum R_{\ell m}(r)\,S_{\ell m}(\theta)\,e^{-i\omega t+im\phi}$ , so the radial factor obeys $\mathrm{d}^2R/\mathrm{d}r_*^2+V(r)R=0$ with $V\to\omega^2$ as $r_*\to+\infty$ and $V\to(\omega-m\Omega_H)^2$ as $r_*\to-\infty$ . Why: pushing the horizon to $r_*=-\infty$ turns both ends into constant-potential regions where the solutions are plane waves, which is what makes the flux bookkeeping exact.

Step 2. Impose physical boundary conditions: purely ingoing at the horizon and incident-plus-reflected at infinity,

R\sim\begin{cases}\mathcal T\,e^{-ik_H r_*}, & r_*\to-\infty\\[2pt] e^{-i\omega r_*}+\mathcal R\,e^{+i\omega r_*}, & r_*\to+\infty\end{cases}\qquad k_H=\omega-m\Omega_H.

Why: causality forbids anything emerging from the classical horizon, so the horizon side is a single ingoing wave — this is the one physical input that breaks the symmetry between the two ends.

Step 3. Because $V$ is real, $R$ and its conjugate $\bar R$ both solve the same equation, so their Wronskian $W=R\,\mathrm{d}\bar R/\mathrm{d}r_*-\bar R\,\mathrm{d}R/\mathrm{d}r_*$ has $\mathrm{d}W/\mathrm{d}r_*=0$ . Why: a constant Wronskian is a conservation law in disguise — it is the conserved radial number-flux of the wave, and equating it at the two ends is flux conservation.

Step 4. Evaluate $W$ at each end using the asymptotic forms:

W(-\infty)=-2i\,k_H\,|\mathcal T|^2,\qquad W(+\infty)=-2i\,\omega\,(1-|\mathcal R|^2).

Setting them equal and dividing by $-2i\omega$ gives $1-|\mathcal R|^2=\dfrac{\omega-m\Omega_H}{\omega}|\mathcal T|^2$ . Why: this is equation (4); every symbol in it is now pinned to a boundary amplitude, so its sign is unambiguous.

Step 5. In the window $0<\omega<m\Omega_H$ the factor $(\omega-m\Omega_H)/\omega$ is negative, hence

|\mathcal R|^2=1+\frac{m\Omega_H-\omega}{\omega}\,|\mathcal T|^2>1.

Why: the reflected flux exceeds the incident flux by exactly the transmitted flux times a positive number — the wave is amplified, and the "extra" energy is the negative flux that went down the horizon. *The numerical anchors $r_+/M=1.436$ and $\Omega_H M=0.313$ at $a/M=0.9$ confirm the analytic result.*

Go deeper · Derivation B · 1 minEnergy balance and the area theorem

Derivation BEnergy balance and the area theorem

The goal is to show that a superradiant scattering lowers the hole's mass and spin yet still increases its horizon area, so that mining rotational energy is consistent with the second law of black hole mechanics. We treat one bounce as a quasi-static transition $(M,J)\to(M+\delta M,J+\delta J)$ , fix the ratio $\delta J/\delta M$ from the mode numbers, feed it through the first law, and read off the sign of $\delta A$ . The same inequality then caps the extractable energy, recovering Christodoulou's irreducible mass and the $29\%$ ceiling for extremal Kerr.

Step 1. Fix the flux ratio. A wave with energy $\hbar\omega$ per quantum carries angular momentum $\hbar m$ , and the classical fluxes share the same ratio, so $\delta J/\delta M=m/\omega$ , i.e. $\delta M=(\omega/m)\,\delta J$ . Why: this single ratio is what links the two conserved charges the wave removes, and it is set purely by the mode's $(\omega,m)$ , not by the hole.

Step 2. Use the first law $\delta M=\dfrac{\kappa}{8\pi}\,\delta A+\Omega_H\,\delta J$ and solve for the area term: $\dfrac{\kappa}{8\pi}\,\delta A=\delta M-\Omega_H\,\delta J$ . Why: the first law is the exact differential relation among $M$ , $A$ , and $J$ , so it converts a statement about mass and spin into a statement about area.

Step 3. Substitute the flux ratio:

\frac{\kappa}{8\pi}\,\delta A=\Big(\frac{\omega}{m}-\Omega_H\Big)\delta J=\frac{\omega-m\Omega_H}{m}\,\delta J.

Why: this is the crux — the area change is proportional to the product of two quantities whose signs we control in the superradiant window.

Step 4. Analyze signs. For the hole to lose energy we need $\delta M<0$ , hence $\delta J<0$ (it spins down); in the window $\omega-m\Omega_H<0$ ; the product of two negatives is positive, so $\delta A>0$ . Why: mass and spin fall while area rises — the hole becomes less energetic but more irreversible, exactly as the area theorem demands.

Step 5. Bound the yield. Since $\delta A\ge0$ always, the irreducible mass $M_{\rm irr}=\sqrt{A/16\pi}$ never decreases; only the rotational energy $M-M_{\rm irr}$ is extractable. For extremal Kerr $A=8\pi M^2$ gives $M_{\rm irr}=M/\sqrt2$ , so at most $1-1/\sqrt2\approx29.3\%$ of the mass-energy can be mined. Why: the second law is not a bystander here — it is the quantitative cap on every rotational-energy engine, superradiance and Blandford-Znajek alike.

Go deeper · Derivation C · 2 minThe black hole bomb as a feedback instability

Derivation CThe black hole bomb as a feedback instability

The goal is to turn the per-bounce gain $|\mathcal R|>1$ into an exponential-in-time instability by trapping the wave, and to see why a massive field does this for free. We surround the hole with a reflecting mirror at radius $r_m$ , quantize the cavity, demand that each round trip both re-amplify and reflect the mode, and show the resonance condition forces a complex frequency whose positive imaginary part is the growth rate. Replacing the mirror by the potential barrier of a rest mass gives the same structure without a wall — the mechanism behind gravitational-atom clouds.

Step 1. Build the cavity. A mirror at $r_m$ plus the horizon form a resonant box whose real eigenfrequencies are quantized roughly as $\omega_R\sim n\pi/r_m$ for large $r_m$ . Why: confinement is what lets the wave return; without a second boundary the amplified reflection simply escapes to infinity and nothing accumulates.

Step 2. Amplify each cycle. In the superradiant window every reflection multiplies the amplitude by $|\mathcal R|=\big(1+\tfrac{m\Omega_H-\omega_R}{\omega_R}|\mathcal T|^2\big)^{1/2}>1$ . Why: a gain greater than one per round trip is the defining condition of any oscillator or laser — here the "gain medium" is the hole's spin.

Step 3. Promote the frequency to complex, $\omega=\omega_R+i\Gamma$ , so the field evolves as $\Phi\sim e^{-i\omega_R t}\,e^{+\Gamma t}$ . Matching the horizon condition to the mirror condition forces $\Gamma>0$ precisely when $\omega_R<m\Omega_H$ . Why: an exponentially growing mode is exactly an eigenfrequency in the upper-half complex plane, so instability and superradiance become the same statement.

Step 4. Scale the rate. The growth rate behaves as $\Gamma\sim\dfrac{m\Omega_H-\omega_R}{r_m}\,|\mathcal T|^2$ , which vanishes both when the mirror is far (weak feedback) and when the box is so small that $\omega_R$ leaves the window; an optimal $r_m$ maximizes it. Why: there is a sweet spot — too loose and nothing builds, too tight and the resonance frequency exceeds $m\Omega_H$ and the gain shuts off.

Step 5. Replace the mirror with mass. A field of rest energy $\mu$ has an effective potential rising to $\mu$ at large $r$ , so modes with $\omega_R<\mu$ are bound and reflected back automatically; the instability saturates when back-reaction has reduced $\Omega_H$ to $\omega_R/m$ . Why: nature supplies the mirror, so no artificial wall is needed — the massive-field instability is the astrophysically realized black hole bomb.

Limiting cases

The window $0<\omega<m\Omega_H$ and its consequences take clean forms in the characteristic regimes:

Case	Result	Meaning
Schwarzschild	$a\to0\Rightarrow\Omega_H\to0$	No rotation, no ergosphere; the window closes entirely and a non-rotating neutral hole cannot superradiate.
Extremal Kerr	$a\to M\Rightarrow\Omega_H\to\tfrac{1}{2M}$	Widest window and largest gain; gravitational $\ell=m=2$ reaches $\sim138\%$ per scattering, electromagnetic $\sim4.4\%$ , scalar $\sim0.4\%$ .
Low frequency	$M\omega\ll1,\ \ell=m$	Gain is a high power of frequency, $\|\mathcal R\|^2-1\propto(M\omega)^{2\ell+2}$ — tiny per bounce, hence the need for confinement.
Massive field	$\omega_R<\mu$ bound	Rest mass builds its own mirror; self-confined instability with rate $\Gamma\propto\alpha^{4\ell+5}$ grows a gravitational-atom cloud.
Fermionic field	$\|\mathcal R\|^2\le1$ always	Dirac fields never superradiate: the conserved current is positive-definite and the Pauli principle forbids stimulated amplification.
Charged, non-rotating	$\omega<q\Phi_H$	A Reissner-Nordström hole superradiates charged bosons even at zero spin, the electrostatic analog of the rotational window.

Common trapSuperradiance breaks the area theorem or makes energy from nothing

It does neither. The hole's mass and angular momentum both decrease, but the quantity that the second law constrains is the area, and $\dfrac{\kappa}{8\pi}\delta A=\dfrac{\omega-m\Omega_H}{m}\delta J$ is non-negative in the window (both factors after $\kappa$ are negative). So $\delta A\ge0$ : the horizon grows even as $M$ falls, because the extracted energy is rotational, not irreducible. Nor is the amplification free — every joule that comes back in the reflected wave was paid for by the negative-energy flux that crossed the horizon, exactly as in the Penrose process. A second trap: superradiance is a bosonic effect. For a Dirac field the conserved current is positive-definite, the flux through the horizon cannot reverse sign, and $|\mathcal R|^2\le1$ always — fermions do not superradiate, because stimulated amplification needs Bose statistics.

Worked examples

1 · The superradiant window of a spinning stellar black hole (fully worked)

A $10\,M_\odot$ black hole spins at $a/M=0.9$ . Find its horizon angular velocity in physical units and the frequency band of $m=1$ modes it can amplify.

Step 1. Locate the horizon. In units of $M$ , $r_+=M\big(1+\sqrt{1-(a/M)^2}\big)=M\big(1+\sqrt{1-0.81}\big)=M(1+0.436)=\mathbf{1.436\,M}$ .

Step 2. Horizon angular velocity from equation (1): $\Omega_H=\dfrac{a}{2Mr_+}=\dfrac{0.9M}{2M\cdot1.436M}=\dfrac{0.9}{2.872\,M}=\mathbf{0.313/M}$ (geometrized). Note this is well below the extremal ceiling $1/2M=0.5/M$ .

Step 3. Convert to seconds. For $M=10\,M_\odot$ the time unit is $GM/c^3=(4.925\times10^{-6}\,\mathrm{s})\times10=4.925\times10^{-5}\,\mathrm{s}$ , so $\Omega_H=\dfrac{0.313}{4.925\times10^{-5}\,\mathrm{s}}=\mathbf{6.36\times10^{3}\ rad/s}$ .

Step 4. Read off the window. For $m=1$ the amplified band is $0<\omega<\Omega_H=6.36\times10^3\,\mathrm{rad/s}$ , i.e. $f<\Omega_H/2\pi=\mathbf{1.0\ kHz}$ . Only sub-kilohertz co-rotating waves are amplified; higher- $m$ modes widen the band to $m\times1.0$ kHz. A megahertz wave sent at this hole is simply absorbed — it sits far above the window and $|\mathcal R|^2<1$ .

2 · Matching an ultralight boson to a black hole (fully worked)

The superradiant instability of a massive field grows fastest when the gravitational fine-structure constant is of order $\alpha\sim0.1$ . What boson mass does that pick out for the same $10\,M_\odot$ hole, and what is its Compton wavelength?

Step 1. Write $\alpha=r_g/\lambda_C$ with $r_g=GM/c^2$ . For $10\,M_\odot$ , $r_g=(1.477\ \mathrm{km})\times10=1.477\times10^4\ \mathrm{m}$ .

Step 2. Solve for the boson rest energy from $\alpha=GM\mu/\hbar c=r_g\,\mu c/\hbar$ , i.e. $\mu c^2=\dfrac{\alpha\,\hbar c}{r_g}$ . With $\hbar c=1.973\times10^{-7}\ \mathrm{eV\cdot m}$ and $\alpha=0.1$ : $\mu c^2=\dfrac{0.1\times1.973\times10^{-7}}{1.477\times10^4}=\mathbf{1.3\times10^{-12}\ eV}$ .

Step 3. Its reduced Compton wavelength is $\lambda_C=r_g/\alpha=1.477\times10^4/0.1=1.48\times10^5\ \mathrm{m}\approx\mathbf{150\ km}$ — comparable to the hole's gravitational radius, which is exactly the resonance condition.

Step 4. Interpret. A boson of mass $\sim10^{-12}$ eV, sitting squarely in the axion-like window $10^{-13}$ – $10^{-11}$ eV probed by stellar-mass holes, would grow a cloud and spin this hole down. The steep scaling $\Gamma\propto\alpha^9$ (for $\ell=1$ ) means that raising $\alpha$ to $\sim0.4$ shortens the e-folding time by more than $10^{4}$ , which is why measured black-hole spins carve out sharp exclusion bands rather than fuzzy ones.

3 · Maximum extractable energy from extremal Kerr

How much of an extremal Kerr hole's mass can superradiance (or any classical process) ever remove?

Step 1. For extremal Kerr $r_+=M$ , so the area is $A=4\pi(r_+^2+a^2)=4\pi(M^2+M^2)=8\pi M^2$ .

Step 2. The irreducible mass is $M_{\rm irr}=\sqrt{A/16\pi}=\sqrt{8\pi M^2/16\pi}=M/\sqrt2$ .

Step 3. Since $M_{\rm irr}$ cannot decrease, the extractable fraction is $\dfrac{M-M_{\rm irr}}{M}=1-\dfrac{1}{\sqrt2}=\mathbf{0.293}$ . At most $\sim29\%$ of an extremal Kerr hole's mass-energy is available, and a real process reaches this bound only in the reversible ( $\delta A\to0$ ) limit.

Physics of the portrait

Superradiance has an unusually clean origin story: a single insight, that a rotating absorber can amplify rather than absorb, carried from a laboratory cylinder to the event horizon of a black hole in the space of two years.

Origins · 1 minYakov Zel'dovich and how this was discovered

Yakov Zel'dovich

1914–1987 · rotational superradiance and black-hole energy extraction

Yakov Borisovich Zel'dovich was the towering polymath of Soviet physics — nuclear weapons designer, combustion theorist, cosmologist, astrophysicist — and it was he who, in 1971, realized that a rotating absorbing body would amplify an incident wave whenever $\omega<m\Omega$ , drawing on his intuition about how rotation Doppler-shifts a wave into "negative absorption." He immediately saw the astrophysical target: a rotating black hole should do the same, its horizon the ideal dissipator. Charles Misner reached a compatible conclusion at nearly the same moment, and the underlying reservoir — the rotational energy inside the ergosphere — had already been exposed by Roger Penrose's 1969 particle-splitting argument. The rigorous computation for Kerr came from Alexei Starobinsky in 1973, who solved the scalar wave equation and confirmed the amplification; Teukolsky and Press then tabulated the numbers in 1974, finding the striking $138\%$ maximum for gravitational waves. Meanwhile Press and Teukolsky had already, in 1972, imagined the runaway: enclose the hole in a mirror and superradiance becomes the "black hole bomb." Steven Detweiler showed in 1980 that a massive field supplies its own mirror, and the modern program — using observed black-hole spins to constrain ultralight bosons — grew directly from these roots. Zel'dovich's spinning can turned into one of the sharpest probes of physics beyond the Standard Model.

History

1969

— Penrose shows rotational energy can be extracted from a Kerr black hole

1971

— Zel'dovich predicts wave amplification by a rotating absorbing body, and by a black hole

1972

— Press and Teukolsky propose the black hole bomb

1973

— Starobinsky computes the superradiant amplification for a Kerr hole

1974

— Teukolsky and Press tabulate amplification factors, up to 138 percent for gravitational waves

1980

— Detweiler shows a massive field seeds a self-confined superradiant instability

Going further

Texts. Misner, Thorne & Wheeler, Gravitation (1973), §33 · Wald, General Relativity (1984), Ch. 12 · Carroll, Spacetime and Geometry (2004), Ch. 6 · Hartle, Gravity (2003), Ch. 15 · Frolov & Novikov, Black Hole Physics (1998), Ch. 4.

Papers. Press & Teukolsky, Nature 238, 211 (1972) · Starobinsky, Sov. Phys. JETP 37, 28 (1973) · Teukolsky & Press, ApJ 193, 443 (1974) · Detweiler, Phys. Rev. D 22, 2323 (1980) · Cardoso, Dias, Lemos & Yoshida, Phys. Rev. D 70, 044039 (2004) · Brito, Cardoso & Pani, Superradiance, Lect. Notes Phys. 906 (2015) · Arvanitaki & Dubovsky, Phys. Rev. D 83, 044026 (2011). The maximum single-scattering amplification factors are $\sim0.4\%$ (scalar), $\sim4.4\%$ (electromagnetic), and $\sim138\%$ (gravitational, near-extremal $\ell=m=2$ ); note that $4.4\%$ is the electromagnetic value, not the gravitational one. The extractable fraction $1-1/\sqrt2=0.293\approx29\%$ follows from Christodoulou's irreducible mass. The instability scaling $\Gamma\propto\alpha^{4\ell+5}$ and the hydrogenic bound-state spectrum are standard results. Numerical anchors used throughout: $r_+/M=1.436$ , $\Omega_H M=0.313$ (at $a/M=0.9$ ), $\mu c^2\sim10^{-12}$ eV (at $\alpha=0.1$ , $10\,M_\odot$ ), $1-1/\sqrt2=0.293$ .

Builds on: Kerr Metric and Rotating Black Holes · Ergosphere and Frame Dragging · Penrose Process and Energy Extraction · Killing Horizons and Surface Gravity · Four Laws of Black Hole Mechanics. Leads to: Hawking Area Theorem · No-Hair Theorem and Black Hole Uniqueness · Reissner-Nordström Charged Black Hole · Kerr-Newman Solution.

Ask the tutor: Ask me to walk through why $\omega<m\Omega_H$ is exactly a negative-energy-flux condition at the horizon, or to estimate the spin-down time of a real black hole for a boson of your chosen mass.

Is superradiance the same thing as the Penrose process

They are the same physics in two descriptions. The Penrose process is the particle picture: a body splits inside the ergosphere into a negative-energy fragment that falls in and a positive-energy fragment that escapes with more energy than the original. Superradiance is the wave (field) picture: a mode scatters and the reflected wave is amplified while a negative-energy flux crosses the horizon. Both mine the hole's rotational energy, both are bounded by $\delta A\ge0$ , and both close off when the hole has spun down. The wave version adds one crucial feature the particle version lacks — coherence — which is what makes confinement and the exponential black hole bomb possible.

Where does the "extra" energy in the reflected wave actually come from

From the hole's rotation. In the superradiant window the transmitted part of the wave carries negative energy and negative angular momentum as measured at infinity, so when it crosses the horizon the hole's mass $M$ and spin $J$ both go down, and conservation puts that energy into the reflected wave. The bookkeeping is exact: $\delta M=\Omega_H\delta J-\delta E_{\rm wave}$ , with $\delta E_{\rm wave}>0$ the energy handed to the field. The irreducible mass — hence the horizon area — never decreases, so the extraction is a genuine conversion of rotational energy, capped at $29\%$ for extremal Kerr.

Why can't fermions superradiate

Because superradiance is stimulated emission, and stimulated emission needs Bose statistics. Technically, the conserved current of a Dirac field is positive-definite, so the flux crossing the horizon can never reverse sign the way a bosonic energy flux does; the amplification relation becomes $|\mathcal R|^2=1-\frac{\omega-m\Omega_H}{\omega}|\mathcal T|^2$ with the transmitted term bounded so that $|\mathcal R|^2\le1$ . Physically, the Pauli principle forbids piling an unlimited number of quanta into a single mode, which is exactly the runaway that lets a bosonic wave grow. So a Dirac field scattering off Kerr is always at least partly absorbed — a rotating hole can still transfer angular momentum to fermions, but it cannot amplify them.

What is a "gravitational atom" and how big does the cloud get

When the field has mass, its rest energy $\mu c^2$ builds a potential well outside the ergoregion, and the bound states are hydrogen-like with the gravitational fine-structure constant $\alpha=GM\mu/\hbar c$ playing the role of the electromagnetic one. Superradiance pumps a level with $\omega_R<m\Omega_H$ until it grows into a macroscopic cloud — a "gravitational atom" of Bohr radius $\sim r_g/\alpha^2$ . The cloud can hold up to a few percent of $M$ before back-reaction halts growth, making it literally a hydrogen atom the size of a star with a black hole for a nucleus, which then radiates continuous gravitational waves and can be detected directly.

How does an observed black-hole spin rule out a boson

If a boson exists whose mass places $\alpha=GM\mu/\hbar c$ in the efficient window ( $\sim0.1$ – $1$ ) for a hole of mass $M$ , the superradiant instability would have spun that hole down long ago — faster than the hole could be re-spun by accretion. So finding an old, rapidly spinning hole of mass $M$ is evidence against any boson in the corresponding mass band. Stellar-mass holes (a few to tens of $M_\odot$ ) probe $\mu\sim10^{-13}$ – $10^{-11}$ eV; supermassive holes probe $\mu\sim10^{-18}$ – $10^{-16}$ eV. The measured spins of real black holes therefore carve out exclusion regions in boson mass, complementary to laboratory and cosmological searches.

Does charge play the same role as spin

Yes. The first law of black hole mechanics has two "intensive" potentials, the horizon angular velocity $\Omega_H$ and the horizon electrostatic potential $\Phi_H$ , and each opens a superradiant channel. For a charged scalar of charge $q$ the general condition is $\omega<m\Omega_H+q\Phi_H$ , so a Reissner-Nordström hole (no spin at all) can still amplify a charged bosonic wave via $\omega<q\Phi_H$ . Confining that instability — for instance with anti-de Sitter boundary conditions, which act as a perfect reflecting box — is the mechanism behind holographic superconductors and behind questions about the stability of near-extremal charged holes.