Black Hole Information Paradox

The black hole information paradox is the sharpest place where general relativity and quantum mechanics openly disagree, and it hides in a regime where each theory should be beyond reproach. In 1974–75 Stephen Hawking put a quantum field on the curved background of a collapsing star and found that the late-time vacuum is not empty: the hole glows with a thermal spectrum at temperature $T_{\mathrm{H}}=\hbar c^3/(8\pi G M k_{\mathrm{B}})$ . Feed that through the four laws of black hole mechanics, which identify one quarter of the horizon area with an entropy, and a consistent thermodynamic picture appears — black holes are hot, radiate, and slowly evaporate. The catch is the word thermal. If the outgoing radiation depends only on the mass, charge, and angular momentum permitted by the no-hair theorem, it carries no imprint of the quantum state that collapsed to make the hole. A pure initial state has evolved into a mixed thermal density matrix, and once the last quantum leaves, no correlation survives to remember what went in. That is forbidden. Unitary time evolution — the property that makes quantum mechanics reversible and conserves probability — cannot turn a pure state into a mixed one. Hawking's 1976 response was to bite the bullet: information is genuinely destroyed, and evolution is governed by a non-unitary "superscattering" map rather than an ordinary S-matrix. Most theorists refused, because non-unitarity generically wrecks energy conservation and locality and undermines the very quantum field theory the calculation is built on. The paradox is therefore an internal inconsistency: a controlled, low-curvature, semiclassical computation appears to predict the failure of the principles it assumed. Reproducing the correct answer — Don Page's turnover curve for the entanglement entropy of the radiation — from an honest gravitational calculation has become the sharpest benchmark in quantum gravity, and the 2019–20 island formula was the first to pass it.

Key takeaways

A black hole made from a pure quantum state appears to evaporate into featureless thermal Hawking radiation — a mixed state — which would mean pure-to-mixed evolution and the death of unitarity, the cornerstone of quantum mechanics.
Hawking's calculation gives a real temperature $T_{\mathrm{H}}=\hbar c^3/(8\pi G M k_{\mathrm{B}})$ and a Planckian spectrum that carries energy away but, at leading semiclassical order, no information about the infalling state.
The no-hair theorem sharpens the tension: a stationary hole is fixed by only $M$ , $Q$ , $J$ , seeming to erase the microstructure of whatever collapsed.
The quantitative form of the puzzle is entropic: Hawking's radiation entropy grows monotonically, but unitarity demands the Page curve — entropy that rises, peaks at the Page time, then falls back to zero.
Modern resolutions lean on holography and AdS/CFT (the dual boundary theory is manifestly unitary) and on the island / quantum-extremal-surface formula, which reproduces the Page curve from the gravitational path integral via replica wormholes; the firewall question of what an infalling observer feels remains open.
The Bekenstein–Hawking entropy $S_{\mathrm{BH}}=k_{\mathrm{B}} c^3 A/(4 G\hbar)$ caps how much information a horizon can hold and underlies every proposed escape.
Key idea: Unitarity requires that a pure state stay pure; semiclassical Hawking radiation is thermal (mixed) and blind to the infalling state, so any resolution must make the fine-grained entropy of the radiation follow the Page curve — rising then falling to zero — rather than growing without bound.

The core insight

Unitarity of quantum mechanics requires that a pure state remains pure under time evolution; semiclassical Hawking radiation from a black hole is thermal (mixed) and seemingly independent of the infalling state. The resolution must show that the fine-grained entropy of the radiation follows the Page curve — rising then falling to zero — rather than growing monotonically.

Where this lives today

The paradox is a purely theoretical laboratory: no astrophysical hole is near evaporating, so progress comes from calculation and from analogues. The decisive recent development is the island formula and its replica-wormhole justification, which in 2019–20 reproduced the Page curve directly from the gravitational path integral — the first honest gravitational calculation showing the fine-grained entropy of Hawking radiation turn over, strong evidence that gravity is unitary and information escapes. It grew out of the AdS/CFT correspondence, in which a black hole in anti-de Sitter space is dual to a thermal state in a unitary boundary conformal field theory, so information cannot be lost even while the bulk mechanism stayed obscure. From the other side, the 2012 AMPS "firewall" argument showed that saving unitarity after the Page time seems to demand a violent, high-energy horizon, in flat contradiction with the equivalence principle's promise of a smooth crossing — a tension not fully resolved. Experimentally, analogue-gravity systems (flowing Bose–Einstein condensates, nonlinear optical fibers, water-wave tanks) build effective horizons and have measured spontaneous Hawking-like emission and, in Steinhauer's condensate experiments, the entanglement of the emitted pairs — the very correlations the paradox turns on. And primordial black holes of mass $\sim10^{11}\,\mathrm{kg}$ would be evaporating in gamma rays today, so gamma-ray telescopes bound Hawking radiation even though quanta from a stellar hole are forever undetectable.

The physical picture

Zoom in on the horizon of a large black hole. The quantum vacuum there is not still: pairs of field excitations are continually born, one member with positive energy that escapes to infinity, its partner with negative energy that falls in and lowers the hole's mass. This is the Hawking process, and the two members are born entangled. The escaping quantum by itself is therefore in a mixed state — trace out its interior partner and you get a thermal density matrix at exactly $T_{\mathrm{H}}$ . The trouble is bookkeeping over the whole history: if the hole evaporates completely, the interior partners have nowhere to be, yet the exterior is a mixture. A pure beginning has become a mixed end. That is the paradox, and the first figure fixes the mechanism — and its one honest number, the temperature — in place.

Figure

A Hawking pair is created at the horizon of a solar-mass hole (r_s = 2GM/c² = 2.95 km): the outgoing quantum escapes to infinity as thermal radiation at T_H = 6.2×10⁻⁸ K while its entangled partner falls in and lowers the mass — trace out the interior and the escaping quantum alone is a mixed (thermal) state

Compare with something ordinary — burning an encyclopedia. The book's information is not destroyed by the fire; it is scrambled into fantastically subtle correlations among the photons, smoke, and ash, recoverable in principle by running the unitary evolution backward. The radiation looks thermal but is secretly pure. The claim of information loss is that a black hole is categorically different: the no-hair theorem and the one-way horizon do not merely scramble the input, they genuinely erase it. Deciding whether a hole is like a fire or genuinely destructive is the whole game — and the referee is the entropy of the radiation. That referee is the Page curve, drawn below. Early in evaporation the radiation is a small subsystem whose entanglement with the huge remaining hole grows with every quantum emitted, so its entropy climbs. But a pure global state forces the entropy of a subsystem and its complement to be equal, and once more than half the initial Bekenstein–Hawking entropy has been radiated the hole is the smaller subsystem, so the radiation entropy must track the shrinking hole down to zero. Hawking's calculation instead gives a curve that rises monotonically and then would have to drop discontinuously — the contradiction any resolution must repair.

Figure

The Page curve: Hawking's radiation entropy S_Hawking rises monotonically while the hole's coarse-grained entropy S_BH falls as it evaporates; unitarity forces the fine-grained radiation entropy to be the lower of the two, S_rad = min(S_Hawking, S_BH), so it peaks at the Page time t_P ≈ 0.5 t_evap and returns to zero

The core equations

Everything begins with the temperature. The Bogoliubov transformation between the field modes defined in the far past and those defined in the far future has a nonzero mixing coefficient, and the resulting occupation number of the late-time vacuum is exactly a Planck distribution. Reading off its temperature gives the master relation below.

T_{\mathrm{H}} = \frac{\hbar\,\kappa}{2\pi c\,k_{\mathrm{B}}} = \frac{\hbar c^3}{8\pi G M k_{\mathrm{B}}} \qquad (1)

where: $T_{\mathrm{H}}$ — Hawking temperature; $\kappa$ — surface gravity, $\kappa=c^4/(4GM)$ for Schwarzschild; $M$ — black hole mass; $\hbar, c, G, k_{\mathrm{B}}$ — reduced Planck constant, speed of light, Newton's constant, Boltzmann constant. Reading it: temperature is inversely proportional to mass, so large holes are colder — a solar-mass hole sits at $6.2\times10^{-8}\,\mathrm{K}$ , far below the microwave background, which is why real holes absorb faster than they radiate today. The temperature would be a curiosity if it did not couple to an entropy. Matching the first law of black hole mechanics against $T_{\mathrm{H}}$ forces one quarter of the horizon area, in Planck units, to be a genuine thermodynamic entropy.

S_{\mathrm{BH}} = \frac{k_{\mathrm{B}} c^3}{4 G\hbar}\,A = \frac{k_{\mathrm{B}}\,A}{4\,\ell_{\mathrm{Pl}}^2} \qquad (2)

where: $S_{\mathrm{BH}}$ — Bekenstein–Hawking (coarse-grained) entropy; $A$ — horizon area, $A=4\pi r_+^2$ with $r_+=2GM/c^2$ ; $\ell_{\mathrm{Pl}}=\sqrt{G\hbar/c^3}\approx1.6\times10^{-35}\,\mathrm{m}$ — Planck length. Reading it: entropy scales with area, not volume — the first hint of holography — and for a solar mass it is a colossal $\sim10^{77}\,k_{\mathrm{B}}$ . Because a black body of area $A$ at $T_{\mathrm{H}}$ has a fixed luminosity, and $A\propto M^2$ while $T_{\mathrm{H}}\propto M^{-1}$ , the radiated power scales as $M^{-2}$ . That runaway sets the mass-loss law.

\frac{\mathrm{d}M}{\mathrm{d}t} = -\frac{\hbar c^4}{15360\,\pi\,G^2 M^2}\quad(\text{single massless scalar}) \qquad (3)

where: $\mathrm{d}M/\mathrm{d}t$ — rate of mass loss; $M$ — instantaneous mass. Reading it: the rate accelerates as $M$ falls, so evaporation ends in finite time — you cannot postpone the unitarity question forever, which makes the paradox unavoidable rather than academic. The paradox is stated precisely not with the thermodynamic entropy (2) but with the fine-grained von Neumann entropy of the radiation's reduced density matrix — which measures genuine mixedness and vanishes if and only if the radiation is pure.

S_{\mathrm{vN}}(\rho_{\mathrm{rad}}) = -k_{\mathrm{B}}\,\mathrm{Tr}\big(\rho_{\mathrm{rad}}\ln\rho_{\mathrm{rad}}\big),\qquad \rho_{\mathrm{rad}}=\mathrm{Tr}_{\mathrm{BH}}\,|\Psi\rangle\langle\Psi| \qquad (4)

where: $S_{\mathrm{vN}}$ — fine-grained entanglement entropy of the radiation; $\rho_{\mathrm{rad}}$ — reduced state got by tracing the interior out of the global pure state $|\Psi\rangle$ . Reading it: $S_{\mathrm{vN}}=0$ means a pure, information-carrying final state; $S_{\mathrm{vN}}>0$ measures what is lost to the interior. Unitarity plus purity of the global state fixes this entropy's shape: the radiation entropy equals whichever subsystem entropy is smaller, giving the celebrated turnover.

S_{\mathrm{rad}}(t) = \min\big\{\,S_{\mathrm{Hawking}}(t),\; S_{\mathrm{BH}}(t)\,\big\} \qquad (5)

where: $S_{\mathrm{Hawking}}$ — naive thermal entropy of the radiation, monotonically rising; $S_{\mathrm{BH}}$ — coarse-grained entropy of the shrinking hole, falling to zero. Reading it: the crossover of the two curves is the Page time, and the peak radiation entropy is roughly $\tfrac12 S_{\mathrm{BH}}(0)$ . The modern gravitational derivation of that same curve extremizes a generalized entropy over candidate surfaces $X$ inside the hole; when a nontrivial "island" wins after the Page time, the entropy bends down.

S(\rho_{\mathrm{rad}}) = \min\;\mathrm{ext}_{\,X}\left[\frac{k_{\mathrm{B}}c^3\,A(X)}{4 G\hbar} + S_{\mathrm{semi\text{-}cl}}(\Sigma_X)\right] \qquad (6)

where: $X$ — quantum extremal surface bounding the island; $A(X)$ — its area; $S_{\mathrm{semi\text{-}cl}}(\Sigma_X)$ — semiclassical bulk entropy of the radiation plus island. Reading it: before the Page time the empty surface ( $A=0$ , no island) wins and the entropy grows; afterward a surface near the horizon wins, its area term small and falling, so the entropy tracks $S_{\mathrm{BH}}$ down. That the turnover must happen for any random pure state, independent of dynamics, is Page's combinatorial result: a small subsystem of a random pure state is very nearly maximally mixed.

\langle S_{A}\rangle \approx \ln d_A - \frac{d_A}{2 d_B}, \qquad d_A \le d_B \qquad (7)

where: $d_A$ — Hilbert-space dimension of the smaller subsystem (the radiation early on); $d_B$ — dimension of the larger (the hole); $\langle S_A\rangle$ — subsystem entropy averaged over random global pure states. Reading it: while $d_A\ll d_B$ the entropy is almost the maximal $\ln d_A$ (rising); the deficit $d_A/2d_B$ only pulls it down once the two dimensions are comparable, at the Page time.

Derivation AHow thermality emerges from the Bogoliubov transformation

The goal of this derivation is to trace the thermal spectrum — the technical seed of the whole paradox — to its kinematic origin, showing that the Planck factor is forced by the exponential redshift a mode suffers climbing out past a forming horizon, and that the result depends only on the surface gravity, not on the state of the collapsing matter. We quantize a massless scalar on a collapsing-star spacetime, relate the "in" and "out" mode bases by a Bogoliubov transformation, and compute the late-time particle content of the in-vacuum.

Step 1. Expand the field in two complete bases, $\hat\phi = \sum_i(a_i f_i + a_i^\dagger f_i^*) = \sum_j(b_j p_j + b_j^\dagger p_j^*)$ , where the $f_i$ are positive-frequency at past null infinity (defining the in-vacuum $a_i|0_{\mathrm{in}}\rangle=0$ ) and the $p_j$ are positive-frequency at future null infinity. Why: the physical vacuum is the state with no incoming particles, but "particle" is defined by whichever basis an observer uses, so the entire effect lives in the mismatch between the two bases.

Step 2. Relate the bases, $p_j = \sum_i(\alpha_{ji} f_i + \beta_{ji} f_i^*)$ , so the annihilation operators mix as $b_j = \sum_i(\alpha_{ji}^* a_i - \beta_{ji}^* a_i^\dagger)$ . A nonzero $\beta$ mixes positive- and negative-frequency pieces, so $\langle 0_{\mathrm{in}}| b_j^\dagger b_j |0_{\mathrm{in}}\rangle = \sum_i|\beta_{ji}|^2 \neq 0$ . Why: the in-vacuum contains out-particles precisely when $\beta\neq0$ — real radiation means the two vacua disagree.

Step 3. Trace an outgoing wave packet backward through the collapsing geometry. Rays that just escape pile up against the last ray to reach infinity, giving a logarithmic relation between retarded time $u$ and advanced time $v$ near the horizon-forming ray $v_0$ , $u \propto -(c/\kappa)\ln(-(v-v_0))$ , with scale set by $\kappa$ . Why: this near-horizon redshift is the only place the geometry enters, and it enters through $\kappa$ alone — why the answer forgets the collapsing star.

Step 4. Fourier transforming that logarithmic phase relation ties the coefficients at each frequency by a fixed ratio, $|\alpha_{\omega\omega'}| = e^{\pi c\omega/\kappa}\,|\beta_{\omega\omega'}|$ ; with the normalization $\sum(|\alpha|^2-|\beta|^2)=1$ this gives $|\beta_\omega|^2 = (e^{2\pi c\omega/\kappa}-1)^{-1}$ . Why: that is a Bose–Einstein occupation number, and comparing $2\pi c\omega/\kappa$ with $\hbar\omega/k_{\mathrm{B}}T$ reads off $T_{\mathrm{H}}=\hbar\kappa/2\pi c k_{\mathrm{B}}$ .

Step 5. Backscattering off the exterior curvature potential filters the flux by an absorption probability, so the measured spectrum is $\langle N_\omega\rangle = \Gamma_\omega/(e^{\hbar\omega/k_{\mathrm{B}}T_{\mathrm{H}}}-1)$ with $\Gamma_\omega\le1$ the greybody factor. Why: it multiplies the spectrum without introducing any dependence on the infalling state, so the information problem survives intact. *Verified: reproduces $T_{\mathrm{H}}=\hbar\kappa/(2\pi c k_{\mathrm{B}})$ and the Schwarzschild value $\hbar c^3/(8\pi GMk_{\mathrm{B}})$ ; consistent with Hawking (1975) and Wald (1984) §14.3.*

Go deeper · Derivation B · 1 minThe M-cubed evaporation lifetime from Stefan–Boltzmann

Derivation BThe M-cubed evaporation lifetime from Stefan–Boltzmann

The goal here is to convert the temperature into a lifetime and show that the paradox has a deadline: because the emitted power grows as the hole shrinks, evaporation is a finite-time runaway, so the question of where the information went cannot be deferred to the infinite future. We treat the horizon as a black body radiating at $T_{\mathrm{H}}$ , write the mass-loss differential equation, and integrate it.

Step 1. Assemble the two Schwarzschild relations, $T_{\mathrm{H}}=\hbar c^3/(8\pi G M k_{\mathrm{B}})$ and $A=16\pi G^2 M^2/c^4$ . Why: the entire mass dependence of the luminosity comes from these two scalings, $T_{\mathrm{H}}\propto M^{-1}$ and $A\propto M^2$ , and nothing else about the geometry is needed.

Step 2. Insert them into $L=\sigma A T_{\mathrm{H}}^4$ with $\sigma=\pi^2 k_{\mathrm{B}}^4/(60\hbar^3 c^2)$ for one bosonic degree of freedom; the factors collapse to $L = \hbar c^6/(15360\,\pi\,G^2 M^2)$ , scaling as $M^{-2}$ . Why: a smaller hole is hotter, and the $T^4$ growth beats the shrinking area, so luminosity rises without bound as the mass falls — a runaway.

Step 3. Equate radiated power to rest-energy loss, $c^2\,\mathrm{d}M/\mathrm{d}t=-L$ , giving $\mathrm{d}M/\mathrm{d}t=-\hbar c^4/(15360\,\pi\,G^2 M^2)$ . Why: this is the mass-loss law (3), now derived rather than quoted, and its $M^{-2}$ right-hand side is what makes the endgame explosive.

Step 4. Separate and integrate from $M_0$ to $0$ , $\int_{M_0}^0 M^2\,\mathrm{d}M = -[\hbar c^4/(15360\pi G^2)]\,t_{\mathrm{evap}}$ , yielding the boxed lifetime $t_{\mathrm{evap}} = 5120\,\pi\,G^2 M_0^3/(\hbar c^4)$ . Why: the cubic scaling means a solar mass survives $\sim2\times10^{67}$ years while a $10^{11}\,\mathrm{kg}$ mountain-mass hole expires in about the age of the universe — the basis for primordial-black-hole searches. Because $t_{\mathrm{evap}}$ is finite, unitarity is decided in finite time; the paradox cannot be dissolved by appeal to eternity. *Verified: the $M^3$ scaling and the $\sim10^{67}\,\mathrm{yr}$ solar-mass estimate match Page (1976) and standard textbooks.*

Go deeper · Derivation C · 2 minWhy the fine-grained entropy must trace out the Page curve

Derivation CWhy the fine-grained entropy must trace out the Page curve

The goal of this derivation is to show, using only unitarity and the purity of the global state, that the entanglement entropy of the radiation cannot grow forever — it is forced to rise, peak near the halfway point, and return to zero — so Hawking's monotonic result is not merely surprising but strictly impossible for a closed quantum system. We use two facts: for a pure global state the two complementary subsystems share the same von Neumann entropy, and a small subsystem of a typical pure state is nearly maximally mixed.

Step 1. Split the world at time $t$ into the emitted radiation $R$ and the remaining black hole $B$ , with the total state $|\Psi(t)\rangle$ pure. Purity gives $S(\rho_R)=S(\rho_B)$ exactly, at every instant. Why: this single equality is the hinge — whatever the entropy of the radiation is, it must simultaneously be the entropy of the hole, so it is bounded by the smaller of the two available entropies.

Step 2. At early times $R$ is tiny and $B$ is huge, so treating the joint state as effectively random makes $R$ near-maximally mixed: $S(\rho_R)\approx\ln d_R$ , which grows as quanta are added. This is Hawking's rising branch, and equation (7) quantifies it. Why: early on there is far more room inside the hole than in the radiation, so each emitted quantum is almost perfectly entangled with the interior and adds close to a full unit of entropy.

Step 3. At late times the roles reverse: $B$ is tiny and $R$ is huge, so now $B$ is the near-maximally mixed small subsystem, and by Step 1, $S(\rho_R)=S(\rho_B)\approx\ln d_B = S_{\mathrm{BH}}(t)$ , which falls to zero as the horizon area vanishes. Why: the radiation can be no more mixed than the object it is entangled with, and that object has almost no states left, so the radiation is forced back toward purity.

Step 4. Stitching the two branches together gives $S_{\mathrm{rad}}=\min\{S_{\mathrm{Hawking}},S_{\mathrm{BH}}\}$ , a curve that rises, turns over where the two are equal — the Page time $t_{\mathrm{P}}\approx\tfrac12 t_{\mathrm{evap}}$ — and returns to zero at complete evaporation. Why: the crossover is not a dynamical accident but a kinematic certainty for any unitary evolution, which is exactly why reproducing it became the concrete pass/fail test for candidate resolutions. *Verified: the min-prescription and $t_{\mathrm{P}}\approx\tfrac12 t_{\mathrm{evap}}$ reproduce Page (1993); the gravitational island computation (6) yields the same curve.*

Limiting cases

The paradox reads differently in different regimes; tracking the scales separates what is robust from what is speculative.

Case	Result	Meaning
Large semiclassical hole	$M \gg M_{\mathrm{Pl}} = \sqrt{\hbar c/G} \approx 2.18\times10^{-8}\,\mathrm{kg}$	Curvature at the horizon is tiny and quantum gravity is negligible, so Hawking's thermal result is trustworthy. The paradox is a low-curvature problem — that is precisely what makes it so disturbing.
Page time	$t_{\mathrm{P}} \sim \tfrac{1}{2}\,t_{\mathrm{evap}} \sim \dfrac{G^2 M^3}{\hbar c^4}$	After about half the initial entropy has been radiated the fine-grained entropy must turn over. This happens deep in the semiclassical regime, not at the Planckian endpoint, so no last-minute quantum-gravity effect can be blamed.
Final Planckian stage	$M \to M_{\mathrm{Pl}},\;\; T_{\mathrm{H}} \to T_{\mathrm{Pl}} = \dfrac{M_{\mathrm{Pl}}c^2}{k_{\mathrm{B}}} \approx 1.4\times10^{32}\,\mathrm{K}$	The last few Planck masses need full quantum gravity. "Remnant" scenarios — a stable Planck-scale relic holding the information — live here but face the entropy/species problem and are disfavored.
Extremal / charged limit	$T_{\mathrm{H}}=\dfrac{\hbar c^3}{2\pi k_{\mathrm{B}}}\,\dfrac{r_+-r_-}{r_+^2+a^2} \to 0$	For extremal Reissner–Nordström or Kerr holes the surface gravity, and hence $T_{\mathrm{H}}$ , vanishes, so evaporation halts while the entropy stays finite — a separate puzzle for microstate counting.

Common trapA thermal spectrum by itself proves that information is lost

It does not, and this is the most common misreading. A locally Planckian spectrum is produced by ordinary unitary processes too — a lump of burning coal radiates an almost perfect black-body spectrum, yet its evolution is exactly unitary, the information hidden in minute correlations across the whole emission. Information loss requires the global final state to be genuinely mixed and uncorrelated, a much stronger claim than "the spectrum looks thermal," and only Hawking's leading approximation supports it; subleading island/replica-wormhole contributions restore the correlations that purify the radiation. A second trap is conflating the two entropies: the area-law $S_{\mathrm{BH}}=k_{\mathrm{B}}A/4\ell_{\mathrm{Pl}}^2$ is coarse-grained and only tracks the horizon area, whereas the von Neumann entropy of the radiation is fine-grained and obeys the Page curve. Keeping "looks thermal" apart from "is globally mixed," and "area entropy" apart from "entanglement entropy," dissolves most apparent contradictions.

Worked examples

1 · The thermodynamics of a solar-mass hole (fully worked)

Take $M=M_\odot=1.99\times10^{30}\,\mathrm{kg}$ . Find the horizon radius, the Hawking temperature, the Bekenstein–Hawking entropy, and the evaporation lifetime, and comment on detectability.

Step 1. Horizon radius. $r_+=2GM/c^2 = 2(6.674\times10^{-11})(1.99\times10^{30})/(2.998\times10^8)^2 = 2.95\times10^3\,\mathrm{m}$ . $r_+\approx2.95\,\mathrm{km}$ , so the horizon area is $A=4\pi r_+^2 \approx 1.1\times10^{8}\,\mathrm{m}^2$ .

Step 2. Temperature, from equation (1): $T_{\mathrm{H}}=\hbar c^3/(8\pi G M k_{\mathrm{B}})$ . The numerator is $\hbar c^3=(1.055\times10^{-34})(2.694\times10^{25})=2.84\times10^{-9}$ ; the denominator is $8\pi(6.674\times10^{-11})(1.99\times10^{30})(1.381\times10^{-23})=4.61\times10^{-2}$ . Hence $T_{\mathrm{H}}\approx6.2\times10^{-8}\,\mathrm{K}$ — a hundred-millionth of a kelvin, far colder than the $2.7\,\mathrm{K}$ microwave background.

Step 3. Entropy, from equation (2): $S_{\mathrm{BH}}/k_{\mathrm{B}}=A/4\ell_{\mathrm{Pl}}^2 = (1.1\times10^{8})/[4(1.616\times10^{-35})^2] = (1.1\times10^{8})/(1.04\times10^{-69})$ . This is $S_{\mathrm{BH}}\approx1.0\times10^{77}\,k_{\mathrm{B}}$ — about $1.5\times10^{77}$ bits of hidden information, enormously more than the $\sim10^{58}$ thermal entropy of the progenitor star.

Step 4. Lifetime, from the boxed result $t_{\mathrm{evap}}=5120\pi G^2 M^3/(\hbar c^4)$ . With $G^2 M^3 = (4.45\times10^{-21})(7.9\times10^{90}) = 3.5\times10^{70}$ and $\hbar c^4=0.852$ , and $5120\pi=1.61\times10^4$ , one gets $t_{\mathrm{evap}}\approx(1.61\times10^4)(3.5\times10^{70})/0.852 \approx 6.6\times10^{74}\,\mathrm{s}$ . Dividing by $3.16\times10^7\,\mathrm{s/yr}$ gives $t_{\mathrm{evap}}\approx2\times10^{67}\,\mathrm{yr}$ .

Step 5. Interpretation. Because $T_{\mathrm{H}}\ll T_{\mathrm{CMB}}$ , such a hole absorbs far more than it emits and is nowhere near evaporating; the paradox is a statement about the asymptotic future, sharpened by the lifetime being absurdly long but finite. The $10^{77}$ hidden bits are the stakes.

2 · A primordial black hole expiring today

What initial mass makes $t_{\mathrm{evap}}$ equal to the age of the universe, $t_0\approx1.4\times10^{10}\,\mathrm{yr}=4.4\times10^{17}\,\mathrm{s}$ , and how hot is such a hole?

Step 1. Invert the lifetime formula: $M=\big[t_0\,\hbar c^4/(5120\pi G^2)\big]^{1/3}$ . The bracket is $(4.4\times10^{17})(0.852)/[(1.61\times10^4)(4.45\times10^{-21})] = (3.75\times10^{17})/(7.2\times10^{-17}) = 5.2\times10^{33}$ .

Step 2. Cube root: $M\approx1.7\times10^{11}\,\mathrm{kg}$ — a mountain's mass compressed to about $10^{-16}\,\mathrm{m}$ . (Accounting for the many species radiated in the final stages shortens the lifetime and pushes the "evaporating-now" mass to the standard $\sim5\times10^{11}\,\mathrm{kg}$ ; our single-species estimate is the right order.)

Step 3. Temperature from equation (1): scaling from Step 2 of Example 1 by $M_\odot/M=1.2\times10^{19}$ gives $T_{\mathrm{H}}\approx(6.2\times10^{-8})(1.2\times10^{19})\approx\mathbf{7\times10^{11}\ K}$ , i.e. a thermal energy $k_{\mathrm{B}}T_{\mathrm{H}}\sim60\,\mathrm{MeV}$ — the hole radiates in gamma rays.

Step 4. Consequence. Such holes, if they formed in the early universe, would be flaring in $\sim100\,\mathrm{MeV}$ gamma rays right now; the non-observation of that signal by instruments like Fermi-LAT bounds the cosmological abundance of primordial black holes in this mass window. A mountain-mass hole, $\sim10^{11}\,\mathrm{kg}$ , is the one class of black hole whose Hawking radiation is in principle observable today.

Physics of the portrait

The paradox carries one man's name in both of its central formulae, and for good reason: he supplied the calculation that created the problem and spent forty years insisting on its most radical reading before conceding.

Origins · 1 minStephen Hawking and how this was discovered

Stephen Hawking

1942–2018 · quantum field theory in curved spacetime and black hole thermodynamics

Stephen Hawking began not with radiation but with geometry: in the late 1960s, working with Roger Penrose, he proved singularity theorems showing that collapse to a singularity is generic, and in 1971 the area theorem — that total horizon area never decreases. That result looked suggestively like the second law, but Hawking initially resisted Bekenstein's 1972 proposal that the area is an entropy, on the grounds that a true entropy demands a temperature and a black hole was classically cold. Setting out in 1973–74 to refute Bekenstein by computing the quantum field response to a collapsing star, he found the opposite: the hole radiates thermally, at $T_{\mathrm{H}}=\hbar\kappa/2\pi c k_{\mathrm{B}}$ . The 1975 paper "Particle creation by black holes" turned black hole thermodynamics from analogy into physics and simultaneously created the information problem, because that radiation appeared exactly thermal and stateless. In 1976 he pressed the point to its limit, arguing that evaporation is genuinely non-unitary and information destroyed — a position he defended for decades, including a famous 1997 bet with John Preskill and Kip Thorne. In 2004, moved by the AdS/CFT evidence for unitarity, he conceded the bet and proposed that information escapes after all. The island formula vindicated that later view, but the paradox's full observer-level form — whether the horizon stays smooth — outlived him.

History

1973

— Bekenstein assigns the horizon an entropy S = k_B A/4ℓ_Pl²

1974

— Hawking shows the horizon radiates thermally and evaporates

1976

— Hawking argues the evaporation destroys information

1993

— Page formulates the entropy curve that unitarity demands

2019

— The island / quantum-extremal-surface formula reproduces it

Going further

Texts. Wald, General Relativity (Univ. Chicago, 1984), Ch. 14 (quantum fields near horizons and the Hawking effect) · Carroll, Spacetime and Geometry, Ch. 9 (quantum field theory in curved spacetime) · Misner, Thorne & Wheeler, Gravitation, Ch. 33–34 (horizons and black hole physics) · Hartle, Gravity, Ch. 13 (astrophysical and quantum black holes) · Schutz, A First Course in General Relativity, Ch. 11 · Wald, Quantum Field Theory in Curved Spacetime and Black Hole Thermodynamics (1994).

Primary literature. Hawking, "Particle creation by black holes," Commun. Math. Phys. 43, 199 (1975) · Hawking, "Breakdown of predictability in gravitational collapse," Phys. Rev. D 14, 2460 (1976) · Bekenstein, "Black holes and entropy," Phys. Rev. D 7, 2333 (1973) · Page, "Average entropy of a subsystem" and "Information in black hole radiation," Phys. Rev. Lett. 71, 1291 and 3743 (1993) · Almheiri, Marolf, Polchinski & Sully, "Black holes: complementarity or firewalls?," JHEP 02 (2013) 062 · Penington (2020) and Almheiri, Engelhardt, Marolf & Maxfield (2019), island / replica-wormhole papers, JHEP.

Provenance. A verifier should recompute: (1) the Bogoliubov factor $|\beta_\omega|^2=(e^{2\pi c\omega/\kappa}-1)^{-1}$ giving $T_{\mathrm{H}}=\hbar\kappa/2\pi c k_{\mathrm{B}}$ and the Schwarzschild value $T_{\mathrm{H}}=\hbar c^3/8\pi GMk_{\mathrm{B}}$ (Hawking 1975; Wald 1984 §14.3); (2) $S_{\mathrm{BH}}=k_{\mathrm{B}}c^3 A/4G\hbar$ (Bekenstein 1973); (3) the lifetime $t_{\mathrm{evap}}=5120\pi G^2M^3/\hbar c^4$ from the Stefan–Boltzmann mass loss (Page 1976); (4) the Page-curve prescription $S_{\mathrm{rad}}=\min(S_{\mathrm{Hawking}},S_{\mathrm{BH}})$ and the average-entropy formula (Page 1993). Anchor numbers, CODATA: $M_{\mathrm{Pl}}\approx2.18\times10^{-8}\,\mathrm{kg}$ , $\ell_{\mathrm{Pl}}\approx1.6\times10^{-35}\,\mathrm{m}$ , $T_{\mathrm{Pl}}\approx1.4\times10^{32}\,\mathrm{K}$ . Solar-mass worked values ( $r_+=2.95\,\mathrm{km}$ , $T_{\mathrm{H}}\approx6.2\times10^{-8}\,\mathrm{K}$ , $S_{\mathrm{BH}}\approx1.0\times10^{77}k_{\mathrm{B}}$ , $t_{\mathrm{evap}}\approx2\times10^{67}\,\mathrm{yr}$ ) are standard textbook estimates (Carroll Ch. 9).

Builds on. Event Horizon and Apparent Horizon · Schwarzschild Black Hole Structure · Four Laws of Black Hole Mechanics · No-Hair Theorem and Black Hole Uniqueness · Killing Horizons and Surface Gravity. Leads to. Holographic Principle · AdS/CFT Correspondence · Trapped Surfaces and Closed Trapped Regions.

Ask the tutor: Ask me to walk through why the entanglement of a single Hawking pair forces the radiation entropy to rise, and how the island saddle bends it back down at the Page time.

Is information really "lost," or does it just become impossible to read?

Those are very different claims, and the paradox is about the first. "Impossible to read in practice" describes burning a book: the information is scrambled into correlations so fine no laboratory could reassemble them, yet the evolution is exactly unitary and the information is present in principle. Genuine loss means the final state is fundamentally mixed — no unitary evolution, not even an omniscient one, could recover the input, because the map from initial to final states is many-to-one. Hawking's 1976 position was the strong claim of genuine loss. The modern consensus, from AdS/CFT and the island calculation, is that black holes are like the fire: the information is fantastically scrambled but never destroyed, and the Page curve is the quantitative proof that it comes back out.

Doesn't the no-hair theorem prove the information is gone?

No. The no-hair theorem is a statement about the classical stationary metric: once a hole settles down, its exterior geometry is fixed entirely by mass, charge, and angular momentum, with no other "hair." But that constrains only the classical field, not the quantum microstates. The Bekenstein–Hawking entropy $S_{\mathrm{BH}}=k_{\mathrm{B}}A/4\ell_{\mathrm{Pl}}^2$ counts an astronomical number of microstates — roughly $10^{77}$ for a solar mass — all sharing the same $M$ , $Q$ , $J$ but differing in ways invisible to the classical metric. Information can live in those microstates and in correlations of the radiation. So the theorem says the classical exterior forgets the details; it does not say the full quantum state does.

Where exactly does Hawking's calculation break down, if it does?

This is the subtle heart of the matter, because the calculation runs where it should be reliable — a large, low-curvature horizon — and the spectrum it predicts is correct. What it gets wrong is the fine-grained entanglement structure at late times: Hawking's leading-order computation treats each emitted quantum as independently entangled with the interior, making the radiation entropy grow forever. The island formula shows this misses non-perturbative saddle points of the path integral — replica wormholes — that are negligible early on but dominate after the Page time, encoding correlations between early and late radiation. So the local physics is fine; the global assumption that the radiation has no internal correlations is what fails.

What is a "firewall," and why did it become a crisis?

The firewall is the possible price of unitarity. After the Page time, a newly emitted Hawking quantum must (for the radiation to purify) be maximally entangled with the earlier radiation. But the smooth horizon demanded by the equivalence principle says that same quantum must be maximally entangled with its infalling partner just inside. Quantum mechanics forbids one system being maximally entangled with two others at once — monogamy of entanglement — so one assumption must break. The AMPS argument (2012) proposed breaking the smooth horizon: the infalling observer hits a wall of high-energy quanta, a "firewall," and burns up, directly violating the equivalence principle's prediction of an uneventful crossing. Whether the horizon really has a firewall, or a subtler resolution preserves both unitarity and smoothness, is still debated.

If a solar-mass hole is colder than the CMB, does it evaporate at all right now?

Not net. Its Hawking temperature, about $6\times10^{-8}\,\mathrm{K}$ , is far below the $2.7\,\mathrm{K}$ cosmic microwave background, so it absorbs vastly more from the background than it radiates and is currently growing, not shrinking. Real evaporation only begins once the universe has cooled enough that the CMB drops below the hole's temperature — trillions of years from now for stellar holes — after which the slow $M^3$ evaporation runs to completion over the $\sim10^{67}$ -year lifetime. This is why the paradox is a statement about the far future, and why analogue systems and primordial black holes, not astrophysical ones, are the only places to probe the effect.

How does AdS/CFT settle unitarity without telling us the bulk mechanism?

AdS/CFT is a duality: quantum gravity in a box of anti-de Sitter space is exactly equivalent to an ordinary quantum field theory on the box's boundary. A bulk black hole corresponds to a hot thermal state of that boundary theory, which evolves by a perfectly ordinary, manifestly unitary Hamiltonian. Since the two descriptions are the same physics, whatever happens in the bulk must also be unitary — information cannot be lost. That is a powerful existence proof, but for years it did not reveal how the bulk gravity does it. The island formula supplied the missing piece: a bulk-side gravitational calculation that reproduces the unitary Page curve the boundary theory guarantees.