Four Laws of Black Hole Mechanics

A stationary black hole is one of the simplest objects in physics. The no-hair theorem says it is fixed by just three numbers — mass $M$ , angular momentum $J$ , charge $Q$ — no matter how baroque the star that collapsed to make it. Write down how those numbers change when you feed the hole a little matter, and something uncanny happens: the differential relations, and the geometric quantities living on the event horizon, line up term for term with the laws of thermodynamics. The surface gravity $\kappa$ — the redshifted acceleration needed to hover just outside the horizon — plays the part of temperature and is uniform over the horizon, like a body in thermal equilibrium; the horizon area $A$ plays the part of entropy and never decreases; the change in mass obeys a Gibbs relation; and the extremal state $\kappa\to0$ is an unreachable floor, like absolute zero. Bardeen, Carter and Hawking proved these four statements in 1973 as exact theorems of differential geometry. Read classically, the parallel looks like a joke the equations are playing: a classical black hole swallows everything and emits nothing, so its temperature ought to be exactly zero and $T\propto\kappa$ pure numerology. The punchline arrived a year later. Hawking showed that quantum fields in the curved background make a black hole glow as a black body at $T_H=\hbar\kappa/2\pi k_B c$ , and once the temperature is real the first law forces the entropy to be one quarter of the horizon area in Planck units, $S_{BH}=k_B c^3 A/4G\hbar$ . The dictionary is no longer an analogy — it is physics, and it is the doorway to holography, the information paradox, and quantum gravity.

Key takeaways

The four laws of black hole mechanics are theorems about stationary horizons that mirror the four laws of thermodynamics, with surface gravity $\kappa$ in the role of temperature and horizon area $A$ in the role of entropy.
Zeroth law: $\kappa$ is constant over the event horizon of a stationary black hole, just as temperature is uniform in equilibrium.
First law: $\mathrm{d}M=\tfrac{\kappa}{8\pi}\,\mathrm{d}A+\Omega_H\,\mathrm{d}J+\Phi_H\,\mathrm{d}Q$ — a Gibbs relation balancing mass against area, spin, and charge.
Second law (Hawking's area theorem): the total horizon area never decreases, $\delta A\ge0$ , paralleling $\delta S\ge0$ .
Third law: the extremal state $\kappa\to0$ cannot be reached by any finite physical process, exactly as absolute zero cannot.
Hawking radiation turns the analogy literal: $T_H=\kappa/2\pi$ pins the entropy to $S_{BH}=A/4$ , so a solar-mass hole has $T\approx6.2\times10^{-8}$ K and $S\approx1.05\times10^{77}\,k_B$ .
Key idea: Replace (temperature, entropy, energy) with (surface gravity $\kappa/2\pi$ , area $A/4$ , mass $M$ ): the four laws of thermodynamics become exact theorems of stationary black hole geometry, and Hawking radiation makes the dictionary literal rather than analogical.

The core insight

Replace (temperature, entropy, energy) with (surface gravity $\kappa/2\pi$ , area $A/4$ , mass $M$ ): the four laws of thermodynamics then become exact theorems of stationary black hole geometry, and Hawking radiation makes the dictionary literal rather than analogical.

Where this lives today

The laws stopped being a formal analogy the moment gravitational-wave astronomy could test them. Every LIGO-Virgo-KAGRA binary-merger event is a laboratory for the second law: the remnant's area, inferred from its mass and spin, is checked against the sum of the progenitor areas, and Hawking's area theorem has now been confirmed directly from the GW150914 waveform by isolating the pre- and post-merger ringdown. On the theory side, $S=A/4$ is the target any candidate quantum gravity must reproduce by counting microstates: Strominger and Vafa's 1996 string calculation matched it exactly for a class of extremal holes, and loop quantum gravity recovers the same area law. Meanwhile the black hole information paradox — Hawking's own unfinished business — remains the sharpest open problem in fundamental physics, with recent island and replica-wormhole computations reproducing the expected Page curve.

The physical picture

Lower a mass slowly on a string toward the horizon and read the tension at infinity. What you feel is the local acceleration of a hovering observer, redshifted by the deep gravitational well; in the limit where the mass touches the horizon that redshifted force per unit mass is the surface gravity $\kappa$ . Walk to a different spot and repeat: you measure the same $\kappa$ . That equality everywhere on the horizon is the zeroth law — the horizon sits in a mechanical equilibrium, the precise counterpart of a body in thermal equilibrium having one temperature throughout. The first figure collects the geometry each law speaks about, and prices a change of the hole with the first-law relation.

Figure

A stationary horizon carries a uniform surface gravity κ over its whole surface (zeroth law) while its area A can only grow, δA ≥ 0 (second law); the first law dM = (κ/8π) dA + Ω_H dJ + Φ_H dQ prices each change, and for Schwarzschild κ = 1/(4M) with A = 16πM²

Feed the hole. Throw in matter that satisfies the ordinary energy conditions and the area grows; collide two black holes and the remnant's area exceeds the sum of the areas that went in. Area behaves like entropy — an irreversible ledger of everything that ever crossed the horizon, a number that can climb but never fall. The mass locked behind that ledger is the irreducible mass $M_{irr}=\sqrt{A/16\pi}$ : energy you can never extract, because getting it back would require shrinking the area. The remainder — rotational and electrostatic energy — is extractable, and the first law is the exact accounting of which is which. The second figure turns the temperature side of the dictionary into hard numbers across sixty orders of magnitude of mass.

Figure

Hawking temperature versus mass on a log-log scale, T_H = ℏc³/(8πG k_B M) so T_H ∝ 1/M: a 10¹² kg primordial hole blazes near 10¹¹ K, a solar-mass hole sits at 6.2×10⁻⁸ K, Sagittarius A near 1.5×10⁻¹⁴ K, and M87 near 10⁻¹⁷ K — larger black holes are colder

The core equations

Surface gravity is the single quantity that makes the whole dictionary work, so it is defined first. On a Killing horizon the null generators are orbits of a Killing field $\xi^a$ , but that field is not affinely parametrized along them; the failure of affinity is $\kappa$ .

\xi^a\nabla_a\xi^b = \kappa\,\xi^b \quad\text{on the horizon} \qquad (1)

where: $\xi^a$ — the Killing vector that is null on, and generates, the event horizon; $\kappa$ — surface gravity, the non-affinity of the horizon generators; $\nabla_a$ — the spacetime (Levi-Civita) covariant derivative. Reading: the horizon-generating flow accelerates itself by a factor $\kappa$ , one number for the whole horizon. The zeroth law elevates that observation to a theorem: for a stationary black hole (with matter obeying the dominant energy condition, or in vacuum) the surface gravity does not vary from place to place on the horizon.

\kappa = \text{constant over } \mathcal{H} \qquad (2)

where: $\mathcal{H}$ — the event horizon of a stationary black hole. Reading: as temperature is uniform in thermal equilibrium, $\kappa$ is uniform in the mechanical equilibrium of a stationary horizon. The first law is the heart of the subject. It says that a quasi-static change of the hole balances the mass gained against three exchanges: an irreversible increase of area at the price $\kappa/8\pi$ , plus reversible work done against frame dragging and against the horizon potential.

\mathrm{d}M = \frac{\kappa}{8\pi}\,\mathrm{d}A + \Omega_H\,\mathrm{d}J + \Phi_H\,\mathrm{d}Q \qquad (3)

where: $M$ — ADM mass; $A$ — horizon cross-sectional area; $\Omega_H$ — angular velocity of the horizon; $J$ — angular momentum; $\Phi_H$ — horizon electrostatic potential; $Q$ — electric charge. Reading: $\tfrac{\kappa}{8\pi}\,\mathrm{d}A$ is the gravitational $T\,\mathrm{d}S$ ; the other terms are ordinary work. The second law is Hawking's area theorem — the classical statement that the total area of all horizons cannot decrease in any physical process.

\delta A \geq 0 \qquad (4)

where: $\delta A$ — change in total horizon area in any classical process obeying the null energy condition. Reading: area is the black hole's entropy ledger, the geometric shadow of the second law of thermodynamics. The third law forbids reaching the extremal, zero-temperature state in a finite sequence of steps — the counterpart of Nernst's unattainability of absolute zero.

\kappa \to 0 \text{ unattainable in finite steps} \qquad (5)

where: $\kappa\to0$ — the extremal limit $M^2\to a^2+Q^2$ , at which the horizon has zero surface gravity. Reading: you can approach extremality but never land on it, every step toward $\kappa=0$ demanding finer tuning than the last. Integrating the first law using the scaling dimensions of the variables gives an algebraic identity, the Smarr formula, that holds exactly on any Kerr-Newman black hole.

M = \frac{\kappa A}{4\pi} + 2\Omega_H J + \Phi_H Q \qquad (6)

where: $M$ — total mass as an algebraic combination of horizon quantities. Reading: the differential first law (3) has an integrated partner, just as $\mathrm{d}U=T\,\mathrm{d}S-p\,\mathrm{d}V$ has the Euler relation $U=TS-pV+\mu N$ . The area repackages as an irreducible mass — the part of $M$ no classical process can extract — and $M$ decomposes into irreducible, charge, and rotational pieces (Christodoulou-Ruffini).

M_{irr}=\sqrt{\frac{A}{16\pi}},\qquad M^2 = \left(M_{irr}+\frac{Q^2}{4M_{irr}}\right)^2 + \frac{J^2}{4M_{irr}^2} \qquad (7)

where: $M_{irr}$ — irreducible mass, monotonic in $A$ and hence non-decreasing. Reading: because $A$ can only grow, $M_{irr}$ can only grow, and the extractable energy is capped at $M-M_{irr}$ . Finally, Hawking's quantum calculation supplies the proportionality constants the classical laws leave open, fixing the temperature and, through the first law, the entropy.

T_H = \frac{\hbar\,\kappa}{2\pi k_B c},\qquad S_{BH}=\frac{k_B c^3 A}{4 G \hbar} \qquad (8)

where: $T_H$ — Hawking temperature; $S_{BH}$ — Bekenstein-Hawking entropy, one quarter of the horizon area in Planck units; $\hbar,k_B,c,G$ — the Planck, Boltzmann, light, and gravitational constants. Reading: $T_H=\kappa/2\pi$ and $\mathrm{d}M=T\,\mathrm{d}S$ together force $\mathrm{d}S=\tfrac14\mathrm{d}A$ , so $S=A/4$ .

Derivation ASurface gravity of the Schwarzschild horizon

The goal is to turn the abstract definition $\xi^a\nabla_a\xi^b=\kappa\,\xi^b$ into a single number for the simplest black hole, and to show that this number is $\kappa=1/4M$ . We take the static time-translation Killing vector, use the fact that for a static metric the surface gravity reduces to a derivative of the metric function at the horizon, and check the answer against the known Hawking temperature — seeing on the way why the zeroth law is trivial here, since there is nowhere on this horizon for $\kappa$ to vary.

Step 1. Take Schwarzschild in the usual coordinates with $f(r)=1-2M/r$ , so the metric is $\mathrm{d}s^2=-f\,\mathrm{d}t^2+f^{-1}\mathrm{d}r^2+r^2\mathrm{d}\Omega^2$ , and let $\xi^a=(\partial_t)^a$ be the timelike Killing vector. Its norm is $\xi^a\xi_a=g_{tt}=-f(r)$ , which vanishes at $r_+=2M$ : that null surface is the Killing horizon. Why: the horizon is exactly where the static observer's Killing time becomes null, so $\kappa$ must be read off there.

Step 2. For a static, spherically symmetric metric the coordinate-free formula $\kappa^2=-\tfrac12(\nabla^a\xi^b)(\nabla_a\xi_b)\big|_{\mathcal H}$ collapses to $\kappa=\tfrac12 f'(r_+)$ , because the only non-trivial gradient of $\xi$ is the radial one built from $g_{tt}=-f$ . Why: all the angular and time derivatives of the static metric vanish, leaving a single radial derivative of $f$ to carry the entire surface gravity.

Step 3. Differentiate: $f'(r)=2M/r^2$ , so $\kappa=\tfrac12\cdot\dfrac{2M}{r_+^2}=\dfrac{M}{(2M)^2}=\dfrac{1}{4M}$ . Because $f'$ is evaluated at the single radius $r_+=2M$ and does not depend on the angular position, $\kappa$ is automatically constant over the horizon — the zeroth law holds by inspection. Why: for the round Schwarzschild sphere there is literally no angular variable in $\kappa$ , so uniformity is manifest rather than earned.

Step 4. Cross-check through the temperature: $T_H=\kappa/2\pi=1/8\pi M$ in geometrized units, and restoring constants gives $T_H=\hbar c^3/8\pi G k_B M\approx6.2\times10^{-8}\,(M_\odot/M)$ K. This is the standard Hawking value, confirming both the formula and the factor of $1/2$ . Why: an independent quantity computed from the same $\kappa$ reproduces the textbook temperature, closing the loop.

Go deeper · Derivation B · 1 minThe first law as an identity for a Schwarzschild black hole

Derivation BThe first law as an identity for a Schwarzschild black hole

The goal is to show that the first law $\mathrm{d}M=\tfrac{\kappa}{8\pi}\,\mathrm{d}A$ is not an extra postulate but an algebraic identity forced by the geometry, once we know how the horizon area depends on the mass. We compute the area of the Schwarzschild horizon, differentiate it with respect to $M$ , insert the surface gravity found above, and watch the coefficient $\kappa/8\pi$ emerge with no freedom left to adjust. This is the cleanest place to see that "energy in equals area times its price".

Step 1. The horizon is the sphere $r=r_+=2M$ , whose proper area is $A=4\pi r_+^2=4\pi(2M)^2=16\pi M^2$ . Why: the induced metric on the $t=\text{const}$ , $r=2M$ surface is that of an ordinary round two-sphere of radius $2M$ , so its area is elementary.

Step 2. Differentiate the area-mass relation: $\mathrm{d}A=32\pi M\,\mathrm{d}M$ , hence $\mathrm{d}M=\dfrac{\mathrm{d}A}{32\pi M}$ . Why: with $J=Q=0$ the only way to change the hole is to change its mass, so $A$ and $M$ are locked together and one differential determines the other.

Step 3. Insert the surface gravity $\kappa=1/4M$ into the proposed coefficient: $\dfrac{\kappa}{8\pi}=\dfrac{1}{8\pi}\cdot\dfrac{1}{4M}=\dfrac{1}{32\pi M}$ , which is exactly the prefactor found in Step 2. Therefore $\mathrm{d}M=\dfrac{\kappa}{8\pi}\,\mathrm{d}A$ holds identically. Why: the same $M$ -dependence appears on both sides, so the first law is a rearrangement of $A=16\pi M^2$ rather than new physical input.

Step 4. Read the physics: every unit of mass dropped in enlarges the area at the rate set by $\kappa$ , and since $\kappa$ falls as $1/M$ , a bigger hole stores each added unit of energy in more new area — the gravitational analogue of a large heat capacity. Why: interpreting $\tfrac{\kappa}{8\pi}\mathrm{d}A$ as $T\,\mathrm{d}S$ shows the accretion is irreversible entropy production, not reversible work.

Go deeper · Derivation C · 1 minThe Smarr formula from Euler scaling

Derivation CThe Smarr formula from Euler scaling

The goal is to derive the integrated first law — the Smarr relation $M=\tfrac{\kappa A}{4\pi}+2\Omega_H J+\Phi_H Q$ — from nothing more than the differential first law and the way the variables scale under a change of units. In geometrized units $M$ , $A$ , $J$ , $Q$ carry definite dimensions of length, so $M(A,J,Q)$ is a homogeneous function and Euler's theorem converts the intensive first law into an extensive algebraic identity. We then note it can be checked head-on against Kerr-Newman.

Step 1. In units $G=c=1$ the mass has dimension of length, the area of length squared, the angular momentum of length squared, and the charge of length: $[M]=L$ , $[A]=L^2$ , $[J]=L^2$ , $[Q]=L$ . Under a rescaling $M\to\lambda M$ every geometric quantity rescales by its dimension, so $M(\lambda^2 A,\lambda^2 J,\lambda Q)=\lambda\,M(A,J,Q)$ — $M$ is homogeneous of degree one when $A,J$ count with weight $2$ and $Q$ with weight $1$ . Why: the horizon quantities are pure geometry, so their behaviour under a similarity transformation is fixed entirely by their length dimensions.

Step 2. Euler's theorem for a homogeneous function then reads $M=2\dfrac{\partial M}{\partial A}A+2\dfrac{\partial M}{\partial J}J+\dfrac{\partial M}{\partial Q}Q$ , with the weights $2,2,1$ inherited from Step 1. Why: Euler's identity is the exact statement that a degree-one homogeneous function equals the weighted sum of its variables times its partial derivatives.

Step 3. The partial derivatives are precisely the intensive variables the first law identifies: $\dfrac{\partial M}{\partial A}=\dfrac{\kappa}{8\pi}$ , $\dfrac{\partial M}{\partial J}=\Omega_H$ , $\dfrac{\partial M}{\partial Q}=\Phi_H$ . Substituting gives $M=2\cdot\dfrac{\kappa}{8\pi}A+2\Omega_H J+\Phi_H Q=\dfrac{\kappa A}{4\pi}+2\Omega_H J+\Phi_H Q$ . Why: reading the slopes of $M$ off the differential first law is what fuses thermodynamic content with the pure scaling argument.

Step 4. Sanity-check on Schwarzschild, where $J=Q=0$ : the formula demands $M=\kappa A/4\pi=\tfrac{1}{4M}\cdot 16\pi M^2/4\pi=M$ , an identity, confirming the weight- $2$ bookkeeping for the area term. Why: the simplest member of the family must satisfy the integrated law trivially, and it does.

Limiting cases

The horizon quantities take simple closed forms across the Kerr-Newman family (geometrized units $G=c=1$ ).

Case	Result	Meaning
Schwarzschild	$\kappa=\tfrac{1}{4M},\ A=16\pi M^2,\ T_H=\tfrac{1}{8\pi M}$	Surface gravity falls with mass, so larger holes are colder. With $J=Q=0$ the first law is just $\mathrm{d}M=\tfrac{\kappa}{8\pi}\mathrm{d}A=T_H\,\mathrm{d}S$ .
Kerr	$\kappa=\tfrac{r_+-M}{r_+^2+a^2},\ \Omega_H=\tfrac{a}{r_+^2+a^2},\ r_+=M+\sqrt{M^2-a^2}$	With $a=J/M$ and $A=4\pi(r_+^2+a^2)$ . The rotational term $\Omega_H\,\mathrm{d}J$ now appears; $\Omega_H$ is the angular velocity of the horizon.
Reissner-Nordström	$\kappa=\tfrac{r_+-r_-}{2r_+^2},\ \Phi_H=\tfrac{Q}{r_+},\ r_\pm=M\pm\sqrt{M^2-Q^2}$	The charge term $\Phi_H\,\mathrm{d}Q$ enters, with area $A=4\pi r_+^2$ and $\Phi_H$ the horizon electrostatic potential.
Extremal limit	$\kappa\to0$ as $M^2\to a^2+Q^2$	Zero surface gravity and zero Hawking temperature, but finite area. The third law forbids reaching this state in finitely many steps.

Common trapReading the laws as thermodynamics before Hawking radiation fixes the constants

Classically $\kappa$ and $A$ are pure geometry, and the constants between $(T,S)$ and $(\kappa,A)$ are completely undetermined — you could equally well write $T=\alpha\kappa$ and $S=A/8\pi\alpha$ for any $\alpha$ , and the first law would still hold. Worse, a classical black hole has temperature zero, because it absorbs and never emits, so the identification looks like a formal coincidence with no temperature behind it. It is Hawking's quantum calculation — a universal, model-independent spectrum at $T_H=\kappa/2\pi$ — that pins $\alpha=1/2\pi$ and forces $S=A/4$ . Do not present $S=A/4$ as a classical result; it is the one place where $\hbar$ is indispensable. A milder caution: the first law is written in geometrized units, and restoring $G,c$ (or adding matter and cosmological-constant terms) must be done with care.

Worked examples

1 · Temperature and entropy of a solar-mass black hole (fully worked)

Take a non-rotating, uncharged black hole of one solar mass, $M=M_\odot=1.99\times10^{30}$ kg. Find its Hawking temperature and its Bekenstein-Hawking entropy, and compare the temperature with the cosmic microwave background at 2.7 K.

Step 1. Horizon radius and area. The Schwarzschild radius is $r_+=2GM/c^2=2(6.67\times10^{-11})(1.99\times10^{30})/(3.00\times10^8)^2=\mathbf{2.95\times10^{3}\ m}$ , so the horizon area is $A=4\pi r_+^2=\mathbf{1.10\times10^{8}\ m^2}$ — a sphere a few kilometres across.

Step 2. Temperature. Using $T_H=\hbar c^3/8\pi G k_B M$ and plugging the constants gives $T_H=\mathbf{6.2\times10^{-8}\ K}$ . This is a hundred-millionth of a kelvin — nine orders of magnitude below the 2.7 K microwave background, so a real solar-mass hole absorbs vastly more radiation than it emits and grows rather than evaporates.

Step 3. Entropy. The Planck area is $\ell_P^2=G\hbar/c^3=2.61\times10^{-70}\ \mathrm{m}^2$ , so the dimensionless entropy is $S/k_B=A/4\ell_P^2=(1.10\times10^{8})/(4\times2.61\times10^{-70})=\mathbf{1.05\times10^{77}}$ .

Step 4. Perspective. The thermodynamic entropy of the Sun as a ball of gas is about $10^{58}\,k_B$ ; collapsing that mass to a black hole raises its entropy by nearly twenty orders of magnitude. A solar-mass hole has $T_H=6.2\times10^{-8}$ K and $S=1.05\times10^{77}\,k_B$ — cold, and staggeringly high in entropy. The scalings $T\propto1/M$ , $S\propto M^2$ then place Sagittarius A* ( $4\times10^6\,M_\odot$ ) at $T\approx1.5\times10^{-14}$ K.

2 · Maximum energy extractable from an extremal Kerr black hole

A maximally spinning ("extremal") Kerr black hole has $a=M$ , i.e. $J=M^2$ . Using the Christodoulou-Ruffini relation, find what fraction of its mass-energy can in principle be extracted while respecting the area theorem.

Step 1. At $Q=0$ the mass formula (7) is $M^2=M_{irr}^2+J^2/4M_{irr}^2$ . The irreducible mass is locked by the area and cannot be lowered, so the best case is to spin the hole all the way down to $J=0$ reversibly, leaving only $M_{irr}$ .

Step 2. Evaluate $M_{irr}$ for the extremal hole. With $a=M$ the horizon is at $r_+=M$ and the area is $A=4\pi(r_+^2+a^2)=8\pi M^2$ , so $M_{irr}=\sqrt{A/16\pi}=\sqrt{M^2/2}=\mathbf{M/\sqrt2}$ .

Step 3. The extractable energy is the difference $M-M_{irr}=M\!\left(1-\tfrac{1}{\sqrt2}\right)=\mathbf{0.293\,M}$ . At most about 29% of an extremal Kerr hole's mass is available as rotational energy, via the Penrose process or superradiance; the area theorem guarantees no scheme can do better. For the analogous extremal Reissner-Nordström hole one finds $M_{irr}=M/\sqrt2$ as well, with up to 29% stored as electrostatic energy.

3 · Area theorem check on the GW150914 merger

GW150914 joined black holes of roughly $36\,M_\odot$ and $29\,M_\odot$ into a remnant of about $62\,M_\odot$ spinning at $a_*\equiv a/M\approx0.67$ . Verify that the horizon area increased, as the second law requires.

Step 1. For a Kerr hole the area is $A=8\pi M^2\!\left(1+\sqrt{1-a_*^2}\right)$ ; the two initial holes are taken non-spinning ( $a_*=0$ ), so each has $A=16\pi M^2$ . Dropping the common $8\pi$ , the initial total is $A_i\propto 2(36^2+29^2)=\mathbf{4274}$ .

Step 2. The remnant with $a_*=0.67$ gives $\sqrt{1-0.67^2}=0.742$ , so $A_f\propto 62^2(1+0.742)=\mathbf{6697}$ in the same units.

Step 3. The ratio is $A_f/A_i=6697/4274=\mathbf{1.57}$ : the final area exceeds the sum of the initial areas by 57%, satisfying $\delta A\ge0$ even though the merger radiated about $3\,M_\odot$ as gravitational waves. The area law holds — and it is the reason the radiated energy is capped: you cannot emit so much that the remnant's area would fall below $A_i$ .

Physics of the portrait

Every strand of this subject — the area theorem, the temperature, the entropy, the information paradox it opened — runs through one person, who turned a geometric curiosity into the deepest known clue about quantum gravity.

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 thermodynamics but with singularities: working with Roger Penrose in the late 1960s, he proved that gravitational collapse and the Big Bang generically produce singularities in general relativity. Turning to horizons, in 1971 he proved the area theorem — that the total horizon area can never decrease — as a rigorous consequence of the null energy condition and the focusing of light rays. At first he resisted its thermodynamic reading; when Jacob Bekenstein argued in 1972 that the area was a physical entropy, Hawking objected, because a body with entropy must have a temperature, and a temperature means radiation, which a black hole was supposed never to emit. Pursuing that objection, in 1974 he did the calculation that made his name: quantizing fields on a collapsing background, he found that a black hole emits a genuinely thermal spectrum at $T_H=\hbar\kappa/2\pi k_B c$ . This inverted his own view — the area really is entropy, $S=A/4$ in Planck units, and the four laws he had helped prove as a formal analogy became literal thermodynamics. The same calculation raised the black hole information paradox, which drove decades of work on holography and remains unresolved. Hawking pursued all of this while living for more than half a century with motor neurone disease, and he regarded the formula $S=A/4$ as the result he most wanted on his memorial.

History

1971

— Hawking proves the area theorem: the total horizon area never decreases

1972

— Bekenstein conjectures that black hole entropy is proportional to horizon area

1973

— Bardeen, Carter and Hawking establish the four laws of black hole mechanics

1974

— Hawking discovers thermal radiation, fixing the temperature at κ over 2π

1975

— the entropy is confirmed as A over 4 in Planck units, making the analogy literal

Going further

Texts. Bardeen, Carter & Hawking, "The Four Laws of Black Hole Mechanics", Commun. Math. Phys. 31, 161 (1973) · Wald, General Relativity (1984), Ch. 12 · Misner, Thorne & Wheeler, Gravitation (1973), Ch. 33 · Carroll, Spacetime and Geometry (2004), Ch. 6 · Hawking & Ellis, The Large Scale Structure of Space-Time (1973), Ch. 9 · Wald, Quantum Field Theory in Curved Spacetime and Black Hole Thermodynamics (1994). Key reference values: Schwarzschild $\kappa=1/4M$ , $A=16\pi M^2$ ; Kerr $\kappa=(r_+-M)/(r_+^2+a^2)$ , $\Omega_H=a/(r_+^2+a^2)$ , $r_+=M+\sqrt{M^2-a^2}$ ; Reissner-Nordström $\kappa=(r_+-r_-)/2r_+^2$ , $\Phi_H=Q/r_+$ ; Smarr relation via Euler scaling; irreducible mass $M_{irr}=\sqrt{A/16\pi}$ and the 29% maximal-Kerr extraction; solar-mass numbers $T_H=6.2\times10^{-8}$ K and $S=1.05\times10^{77}\,k_B$ ; GW150914 area ratio $A_f/A_i\approx1.57$ . Units geometrized ( $G=c=1$ ) unless constants are shown.

Prerequisites. Killing Horizons and Surface Gravity · Event Horizon and Apparent Horizon · Kerr Metric and Rotating Black Holes · Trapped Surfaces and Closed Trapped Regions. Leads to. Hawking Area Theorem · No-Hair Theorem and Black Hole Uniqueness · Penrose Process and Energy Extraction · Reissner-Nordström Charged Black Hole.

Ask the tutor: Ask me to walk through the Kerr surface gravity, the generalized second law, or how string theory counts the $A/4$ microstates.

Are the four laws an analogy or literal thermodynamics?

Both, in sequence. Bardeen, Carter and Hawking proved them in 1973 as a formal analogy: exact theorems of stationary black hole geometry that happen to have the same structure as the four laws of thermodynamics, with $\kappa$ where temperature sits and $A$ where entropy sits. At that stage the identification carried no factors — you could not say what $\kappa$ was the temperature of, and classically the temperature was zero anyway. Hawking's 1974 discovery that black holes radiate thermally at $T_H=\kappa/2\pi$ promoted the analogy to physics: the temperature is real, the entropy is genuinely $A/4$ , and the black hole obeys the same thermodynamics as a box of gas. So the laws are a mathematical analogy that quantum mechanics made literal.

What exactly is surface gravity, physically?

Imagine holding a test mass stationary just outside the horizon with a long ideal string and measuring the tension from far away, at infinity. The local force needed to hold the mass diverges as you approach the horizon, but the redshifted force measured at infinity stays finite, and its value per unit mass in the limit is the surface gravity $\kappa$ . Equivalently, $\kappa$ measures how fast the horizon-generating Killing field fails to be affinely parametrized — a purely geometric statement. For Schwarzschild $\kappa=1/4M$ , so bigger holes have weaker surface gravity, which is why they are colder. The zeroth law is the non-obvious fact that $\kappa$ is the same everywhere on a stationary horizon, even a distorted rotating one.

How can area be entropy when a black hole seems featureless?

That is precisely the puzzle Bekenstein and Hawking uncovered. The no-hair theorem fixes the exterior geometry by $M$ , $J$ , $Q$ , so the hole looks featureless from outside — yet it formed from an enormous number of distinct microscopic configurations. The entropy $S=A/4$ counts the logarithm of that hidden multiplicity: $\exp(S/k_B)\approx\exp(10^{77})$ internal states for a solar-mass hole. Reproducing that count from an actual enumeration of quantum states is the benchmark for quantum gravity, and string theory has succeeded for special (extremal) black holes. So "featureless outside" and "vast entropy" are consistent: the information is there, just not visible in the classical exterior.

Why can't you extract all of a black hole's energy?

Because the area can never decrease, and the area is packaged as the irreducible mass $M_{irr}=\sqrt{A/16\pi}$ . Any energy you pull out — by the Penrose process, superradiance, or a Blandford-Znajek magnetic torque — has to leave $A$ , and hence $M_{irr}$ , at least as large as before. So the floor on the mass is $M_{irr}$ , and the most you can extract is $M-M_{irr}$ : about 29% for a maximally spinning Kerr hole, and zero for a Schwarzschild hole, which has no spin or charge energy to give up. This is the black hole version of the statement that you cannot extract heat from a system without paying an entropy cost.

Does the third law forbid extremal black holes from existing?

No — it forbids manufacturing one in a finite number of steps. Extremal black holes ( $M^2=a^2+Q^2$ , $\kappa=0$ ) are perfectly good solutions of Einstein's equations; the third law says only that you cannot reach that state by any finite physical process, just as Nernst's theorem says you cannot cool a system to absolute zero in finitely many steps. Every attempt to spin up or charge up a hole toward extremality requires ever finer tuning and returns ever less, so the boundary is approached asymptotically. This dovetails neatly with cosmic censorship: overshooting to $M^2<a^2+Q^2$ would strip the horizon and expose a naked singularity, and the third law makes that threshold unreachable, protecting the conjecture that horizons always hide singularities.

Does Hawking radiation violate the second (area) law?

The classical area theorem does decrease under Hawking radiation — an evaporating hole shrinks, so $A$ falls, seemingly breaking $\delta A\ge0$ . The resolution is that the classical law was only ever the $\hbar\to0$ shadow of a deeper statement. Bekenstein proposed the generalized second law: the sum of black hole entropy and the ordinary entropy of matter and radiation outside, $S_{BH}+S_{matter}$ , never decreases. When a hole evaporates its $A/4$ drops, but the thermal radiation it emits carries even more entropy into the exterior, so the total still climbs. This generalized law has survived every proposed counterexample and is now regarded as the correct, quantum-corrected form.