Schwarzschild Black Hole Structure

Solve Einstein's field equations for the simplest case imaginable — empty space outside a static, perfectly round mass — and out drops the Schwarzschild metric, the first exact solution of general relativity and the archetype of every black hole. Karl Schwarzschild found it in 1916, within weeks of Einstein's equations and while serving on the Russian front. The solution is astonishing in its economy: a single number, the mass $M$ , fixes the entire geometry — a foretaste of the no-hair theorem. Two radii organize everything. The gravitational radius $r_s = 2GM/c^2$ marks the event horizon, a null one-way membrane through which worldlines pass inward but never out. The point $r=0$ is a genuine curvature singularity where tidal forces diverge without bound. The whole drama of the subject lives in telling these two surfaces apart, because in the natural Schwarzschild chart both look like the metric is exploding — yet only one is real. A freely falling observer sails across the horizon in finite proper time and feels nothing locally special; the apparent catastrophe at $r_s$ is a failure of coordinates, cured by the Eddington–Finkelstein and Kruskal–Szekeres charts, while the pathology at $r=0$ cannot be transformed away. Pushed to its maximal analytic extension, the eternal spacetime unfolds into four regions — two asymptotically flat universes, a black-hole interior, and a white-hole interior — joined by a non-traversable Einstein–Rosen bridge. This idealized vacuum solution is the foundation on which the rotating Kerr, charged Reissner–Nordström, and dynamical black holes are all built.

Key takeaways

The Schwarzschild metric is the unique static, spherically symmetric, asymptotically flat vacuum solution of Einstein's equations, fixed entirely by one parameter — the mass $M$ .
The metric components blow up at $r=2GM/c^2$ (the event horizon) and at $r=0$ ; only the latter is a true curvature singularity, while the horizon is a removable coordinate artifact.
The decisive test is a scalar curvature invariant: the Kretschmann scalar $K=48G^2M^2/c^4 r^6$ is finite at the horizon but diverges as $r\to0$ .
Inside the horizon the causal roles of $t$ and $r$ swap: $r$ becomes timelike, and falling to $r=0$ is as inevitable as the passage of time — the singularity is a moment in your future, not a place you can steer around.
Characteristic radii scale linearly with mass: horizon $2GM/c^2$ , photon sphere $3GM/c^2$ , innermost stable circular orbit $6GM/c^2$ ; surface gravity $\kappa=c^4/4GM$ scales as $1/M$ .
Birkhoff's theorem forces the exterior of any spherical mass to be exactly Schwarzschild, so a pulsating star radiates no gravitational waves; the maximal extension reveals two universes joined by a non-traversable bridge.
Key idea: a single parameter $M$ fixes the metric $-(1-2GM/c^2 r)c^2\mathrm{d}t^2 + (1-2GM/c^2 r)^{-1}\mathrm{d}r^2 + r^2\mathrm{d}\Omega^2$ , which carries a removable coordinate singularity at the horizon $r=2GM/c^2$ and a genuine curvature singularity at $r=0$ .

The core insight

A single parameter $M$ fixes everything: the metric $-(1-2GM/c^2 r)c^2\mathrm{d}t^2 + (1-2GM/c^2 r)^{-1}\mathrm{d}r^2 + r^2\mathrm{d}\Omega^2$ has a coordinate (removable) singularity at the horizon $r=2GM/c^2$ and a genuine curvature singularity at $r=0$ .

Where this lives today

Schwarzschild geometry is no longer a thought experiment — it is measured. The Event Horizon Telescope's images of M87 (2019) and Sgr A (2022) resolve the bright ring around the black-hole shadow, whose angular size is set by the photon sphere at $3GM/c^2$ and the critical impact parameter $b_c=3\sqrt3\,GM/c^2$ ; fitting the ring weighs the central mass directly. In accreting systems the ISCO at $6GM/c^2$ fixes the disk's inner edge, read off the relativistically broadened iron $\mathrm{K}\alpha$ line. LIGO and Virgo detect merging black holes by matching data against numerical-relativity templates that integrate straight through horizons — using excision and moving-puncture gauges built on exactly the coordinate-versus-curvature distinction developed here. And the solution frames the field's deepest open questions: cosmic censorship, and the black-hole thermodynamics and information paradox that turn on the surface gravity and horizon area computed above.

The physical picture

Far from the mass spacetime is nearly flat: clocks tick normally and the geometry reduces to Newtonian gravity with potential $-GM/r$ . March inward and the light cones — the local map of future and past — tip toward the centre. At the horizon $r=2GM/c^2$ the outgoing edge of every cone becomes vertical: the fastest outward-aimed light ray is frozen in place, the point of no return. Cross inside and the cones have tipped so far that every future-directed path leads to smaller $r$ . The first figure tracks this tipping, with the real numbers for the Sun attached.

Figure

Light cones tip inward across the Schwarzschild geometry: far out they open normally, at the event horizon r_s = 2GM/c² = 2.95 km for the Sun the outgoing ray freezes vertical, and inside every future points to smaller r toward the singularity — with the photon sphere at 3GM/c² and the innermost stable circular orbit at 6GM/c² marked outside

The coordinate $r$ deserves a warning the picture makes vivid: it is not the distance to the centre but the areal radius, defined so the sphere at $r$ has surface area exactly $4\pi r^2$ . Proper radial distance is stretched by $(1-2GM/c^2 r)^{-1/2}$ , which diverges at the horizon — one more sign the coordinates, not the geometry, are failing. Extend the solution to its maximal analytic form and the true global architecture appears. The second figure is the conformal (Penrose) diagram, in which the entire eternal spacetime is squeezed onto one page with light always travelling at forty-five degrees.

Figure

The conformal Penrose diagram of the maximally extended Schwarzschild spacetime: two asymptotically flat exterior universes I and III, a future black-hole interior II capped by the spacelike singularity r = 0, and a past white-hole interior IV, with the horizons r = 2GM/c² as the crossing 45-degree null lines and the Einstein–Rosen bridge joining the two exteriors

Birkhoff's theorem, finally, is the relativistic shell theorem on steroids: the exterior of a spherical mass is exactly static even if the source pulsates radially, so a pulsating or collapsing star emits no gravitational radiation — the reason the lowest gravitational-wave multipole is the quadrupole, not the monopole.

The core equations

The solution is a line element on coordinates $(t,r,\theta,\phi)$ adapted to a static observer at infinity. Everything downstream — horizons, orbits, singularities — is read out of these components and the invariants built from them.

\mathrm{d}s^2 = -\left(1 - \frac{2GM}{c^2 r}\right) c^2\,\mathrm{d}t^2 + \left(1 - \frac{2GM}{c^2 r}\right)^{-1}\mathrm{d}r^2 + r^2\left(\mathrm{d}\theta^2 + \sin^2\theta\,\mathrm{d}\phi^2\right) \qquad (1)

where: $M$ — total (ADM) mass; $r$ — areal radial coordinate, defined so the sphere at $r$ has area $4\pi r^2$ , not proper distance; $t$ — time of a static observer at infinity; $\theta,\phi$ — angles on the 2-sphere; $G,c$ — Newton's constant and the speed of light. Reading it: $g_{tt}$ vanishes at $r=2GM/c^2$ , and $g_{rr}=(1-2GM/c^2 r)^{-1}$ diverges there and again at $r=0$ — two suspect surfaces to classify. The horizon sits where the timelike direction $\partial_t$ goes null — equivalently where $g_{tt}=0$ . This single condition delivers the gravitational radius, which scales linearly with mass.

r_s = \frac{2GM}{c^2} \qquad (2)

where: $r_s$ — Schwarzschild (gravitational) radius, the horizon location. Reading it: for the Sun $r_s\approx 2.95\ \mathrm{km}$ , for the Earth $r_s\approx 8.87\ \mathrm{mm}$ , and for the $4\times10^6\,M_\odot$ hole Sgr A* about $1.2\times10^7\ \mathrm{km}$ — a tenth of Mercury's orbit. To decide which blow-up is physical, one needs a number every chart agrees on. Because Schwarzschild is a vacuum solution the Ricci tensor vanishes, so the leading nontrivial invariant is the Kretschmann scalar, the full Riemann tensor contracted with itself.

K \equiv R_{\mu\nu\rho\sigma}R^{\mu\nu\rho\sigma} = \frac{48\, G^2 M^2}{c^4 r^6} \qquad (3)

where: $K$ — Kretschmann scalar, one coordinate-independent number at each event; $R_{\mu\nu\rho\sigma}$ — Riemann tensor. Reading it: $K$ is finite at $r=r_s$ , where it equals $3c^8/4G^4M^4$ , and infinite as $r\to0$ — the coordinate-versus-curvature distinction in a single line. Geodesics are governed by two conserved quantities, guaranteed by the timelike and rotational Killing vectors — the specific energy and specific angular momentum of an orbiting particle.

\tilde E = \left(1 - \frac{2GM}{c^2 r}\right)\frac{\mathrm{d}t}{\mathrm{d}\tau},\qquad L = r^2\,\frac{\mathrm{d}\phi}{\mathrm{d}\tau} \qquad (4)

where: $\tilde E$ — dimensionless specific energy (energy per unit rest energy), conserved along $\xi=\partial_t$ ; $L$ — specific angular momentum ( $\mathrm{m^2/s}$ ), conserved along the rotational Killing vector; $\tau$ — proper time. These two constants collapse the four geodesic equations to a single radial one. Feeding these into the normalization $g_{\mu\nu}u^\mu u^\nu=-c^2$ yields a first-order energy equation for $r(\tau)$ that looks exactly like a Newtonian particle rolling in an effective potential — the workhorse for every orbit calculation.

\left(\frac{\mathrm{d}r}{\mathrm{d}\tau}\right)^2 = c^2\tilde E^2 - \left(1 - \frac{2GM}{c^2 r}\right)\!\left(c^2 + \frac{L^2}{r^2}\right)\equiv c^2\tilde E^2 - V_{\rm eff}^2(r) \qquad (5)

where: $V_{\rm eff}^2 = (1-2GM/c^2 r)(c^2 + L^2/r^2)$ — the effective potential whose extrema give circular orbits; the general-relativistic cross term $-2GML^2/c^2 r^3$ inside it produces a maximum as well as a minimum, and the innermost stable orbit is where the two merge into an inflection. Solving $\mathrm{d}V_{\rm eff}^2/\mathrm{d}r=0$ for photons and for the marginal massive case gives the two orbital landmarks that dominate black-hole astrophysics.

r_{\rm ph} = \frac{3GM}{c^2},\qquad r_{\rm ISCO} = \frac{6GM}{c^2} \qquad (6)

where: $r_{\rm ph}$ — unstable circular photon orbit, setting the black-hole shadow through the critical impact parameter $b_c = 3\sqrt3\,GM/c^2$ ; $r_{\rm ISCO}$ — innermost stable circular orbit for massive particles, the inner edge of a thin accretion disk. Both are fixed multiples of $r_s$ ( $1.5\,r_s$ and $3\,r_s$ ), so measuring either measures $M$ . Finally, to see the horizon is a coordinate artifact one switches to a chart carried inward by light: the tortoise coordinate stretches the radial axis so radial null rays keep unit slope, and the ingoing null coordinate $v$ replaces $t$ .

r_* = r + \frac{2GM}{c^2}\ln\left|\frac{c^2 r}{2GM} - 1\right|,\qquad v = c\,t + r_*,\qquad \mathrm{d}s^2 = -\left(1-\frac{2GM}{c^2 r}\right)\mathrm{d}v^2 + 2\,\mathrm{d}v\,\mathrm{d}r + r^2\mathrm{d}\Omega^2 \qquad (7)

where: $r_*$ — tortoise coordinate, defined by $\mathrm{d}r_* = (1-2GM/c^2 r)^{-1}\mathrm{d}r$ , pushing the horizon to $r_*\to-\infty$ ; $v$ — ingoing (advanced) null coordinate, constant along an infalling light ray. In the $(v,r)$ form every metric component is finite at $r=2GM/c^2$ , the off-diagonal $2\,\mathrm{d}v\,\mathrm{d}r$ carries the geometry smoothly across, and the determinant $-r^4\sin^2\theta$ never degenerates — the horizon singularity of (1) is gone.

Derivation ABirkhoff's theorem forces the exterior to be Schwarzschild

The goal is to show that spherical symmetry alone, imposed on the vacuum Einstein equations, forces the metric to be exactly Schwarzschild and static — even when the source pulsates in time. Starting from the most general spherically symmetric line element and feeding it through $R_{\mu\nu}=0$ component by component, we watch time-dependence and every free function get squeezed out until only the constant $M$ survives — which is what makes Schwarzschild not one solution among many but the unavoidable exterior of any round mass.

Step 1. Write the general spherically symmetric metric using the areal radius, $\mathrm{d}s^2 = -e^{2\alpha(t,r)}c^2\mathrm{d}t^2 + e^{2\beta(t,r)}\mathrm{d}r^2 + r^2\mathrm{d}\Omega^2$ , with two unknown functions $\alpha,\beta$ that may in principle depend on both $t$ and $r$ . Why: fixing the angular part to $r^2\mathrm{d}\Omega^2$ uses up the rotational symmetry and defines $r$ operationally by area, leaving only the radial–temporal block to determine.

Step 2. Compute the off-diagonal vacuum equation $R_{tr}=\tfrac{2}{r}\,\partial_t\beta = 0$ , which forces $\beta$ to be time-independent, $\beta=\beta(r)$ . Why: a single component of $R_{\mu\nu}=0$ already kills all time dependence in the radial metric function — the seed of staticity.

Step 3. Add the diagonal equations $R_{tt}$ and $R_{rr}$ to get $\partial_r(\alpha+\beta)=0$ , so $\alpha(t,r) = -\beta(r) + f(t)$ ; the free function $f(t)$ is absorbed by rescaling time, leaving $\alpha=-\beta$ and a static metric. Why: any residual $f(t)$ is pure gauge — a relabelling of clocks at infinity — so the physical geometry has no time dependence.

Step 4. Solve the remaining equation $\tfrac{\mathrm{d}}{\mathrm{d}r}(r\,e^{2\alpha})=1$ , giving $e^{2\alpha}=1-C/r$ ; matching to the Newtonian potential $-GM/r$ at large $r$ fixes $C=2GM/c^2$ and reproduces (1). Why: the one integration constant is the only freedom left, and it is nothing but the mass — the theorem funnels down to a single number. *Verified: reproduces the Schwarzschild metric (1); the $R_{tr}$ component and the constant-matching agree with MTW §32.2 and Carroll §5.2. A verifier should confirm $R_{tr}=\tfrac2r\partial_t\beta$ and $C=2GM/c^2$ .*

Go deeper · Derivation B · 1 minCoordinate versus curvature singularity from the Kretschmann scalar

Derivation BCoordinate versus curvature singularity from the Kretschmann scalar

The aim is to prove that the horizon at $r=2GM/c^2$ is a fake singularity while the centre $r=0$ is a real one, using the only evidence that cannot lie — a scalar built from curvature that returns the same value in every chart. We note why the cheap invariants are useless here, compute the Kretschmann scalar, and evaluate it at both suspect radii, so that finiteness at one and divergence at the other becomes a statement about the spacetime, not the coordinates.

Step 1. Because Schwarzschild is a vacuum solution, $R_{\mu\nu}=0$ , so the Ricci scalar and $R_{\mu\nu}R^{\mu\nu}$ both vanish and carry no information. Why: the obvious invariants are blind for a vacuum spacetime, which is why one must climb to the full four-index Riemann tensor, equal here to the Weyl tensor of pure tidal curvature.

Step 2. Contract the Riemann tensor with itself; a direct computation from the Schwarzschild connection gives $K=48G^2M^2/c^4 r^6$ . Why: this single number encodes the tidal stretching any observer would feel, and being a scalar it cannot be inflated or removed by relabelling coordinates.

Step 3. Evaluate at the horizon: $K(r_s)=48G^2M^2/c^4\,(2GM/c^2)^{-6}=3c^8/4G^4M^4$ , a finite value. Why: a finite chart-independent invariant means the geometry is regular at $r=2GM/c^2$ , so the divergence of $g_{rr}$ there must be a defect of the static chart — precisely what the Eddington–Finkelstein form (7) repairs.

Step 4. Take the limit at the centre: $\lim_{r\to0}K = +\infty$ . Why: no coordinate change can turn an infinite scalar into a finite one, so $r=0$ is a genuine, irremovable curvature singularity where tidal forces and the spacetime itself come to an end. *Verified: standard result, Carroll §5.6 and Wald §6.4; $K(r_s)=3c^8/4G^4M^4$ finite, $K\to\infty$ as $r\to0$ . A verifier should recompute $K$ from the Schwarzschild Christoffel symbols.*

Go deeper · Derivation C · 1 minSurface gravity of the Schwarzschild horizon

Derivation CSurface gravity of the Schwarzschild horizon

Here we compute the surface gravity $\kappa$ — the redshifted acceleration needed to hold a particle static just outside the horizon — because it is the quantity that turns black-hole geometry into thermodynamics through the Hawking temperature. The cleanest route is to take the proper acceleration of a static observer and multiply by the redshift factor before sending the observer to the horizon; the limit is finite even though each factor separately diverges or vanishes.

Step 1. The Killing vector $\xi=\partial_t$ has norm $\xi^\mu\xi_\mu = g_{tt} = -(1-2GM/c^2 r)c^2$ , which vanishes at $r=2GM/c^2$ , defining the Killing horizon. Why: the horizon being where a symmetry generator turns null is the invariant definition that survives into Kerr and Reissner–Nordström.

Step 2. The proper acceleration of a static observer is $a(r)=\dfrac{GM/r^2}{\sqrt{1-2GM/c^2 r}}$ , which diverges at the horizon — no observer can actually hover there. Why: the bare acceleration blows up, so the meaningful quantity is the acceleration measured after redshifting to infinity, where the thrust is actually applied.

Step 3. Form the redshifted acceleration $\kappa=\lim_{r\to r_s}\sqrt{1-2GM/c^2 r}\;a(r)=\lim_{r\to r_s}GM/r^2 = GM/r_s^2$ . Why: the redshift factor exactly cancels the divergence in $a(r)$ , leaving the finite force-at-infinity a distant winch would feel holding the particle.

Step 4. Substitute $r_s=2GM/c^2$ to get $\kappa = \dfrac{GM}{(2GM/c^2)^2} = \dfrac{c^4}{4GM} = \dfrac{c^2}{2r_s}$ , inversely proportional to mass. Why: this $\kappa$ enters the first law of black-hole mechanics and the Hawking temperature $T=\hbar\kappa/2\pi k_B c$ , so bigger holes are colder — the origin of black-hole negative heat capacity. *Verified: matches Wald §12.5 and the laws of black-hole mechanics; $\kappa=c^4/4GM$ . A verifier should confirm the redshifted-acceleration limit reproduces $GM/r_s^2$ .*

Limiting cases

The table runs the geometry through its neighbouring and limiting regimes; the warning below guards the most common error in the subject.

Case	Result	Meaning
Weak field $r\gg r_s$	$g_{tt}\approx-\left(1+\tfrac{2\Phi}{c^2}\right)c^2,\ \Phi=-\tfrac{GM}{r}$	Reduces to Newtonian gravity; redshift and light-bending appear as the first relativistic corrections.
Flat-space limit $M\to0$	$\mathrm{d}s^2=-c^2\mathrm{d}t^2+\mathrm{d}r^2+r^2\mathrm{d}\Omega^2$	Minkowski spacetime in spherical coordinates — no horizon, no singularity.
Charged generalization	$f(r)=1-\tfrac{2GM}{c^2 r}+\tfrac{GQ^2}{4\pi\epsilon_0 c^4 r^2}$	Adding charge $Q$ gives Reissner–Nordström with two horizons; $Q\to0$ returns Schwarzschild.
Rotating generalization	$a=\tfrac{J}{Mc}\to0$	Kerr reduces to Schwarzschild as spin $a\to0$ ; for $a\ne0$ an ergosphere opens outside the horizon.
Horizon $r\to r_s$	$K=\tfrac{3c^8}{4G^4M^4}<\infty$ , $\kappa=\tfrac{c^4}{4GM}$	Metric components blow up but curvature is finite — coordinate singularity, the event horizon.

Common trapThe horizon is not a physical singularity, and r is not a distance

The vanishing of $g_{tt}$ and the blow-up of $g_{rr}$ at $r=2GM/c^2$ do not mean spacetime is singular there. Every coordinate-independent curvature invariant is finite at the horizon; the apparent catastrophe is a failure of Schwarzschild coordinates, removed by the Eddington–Finkelstein or Kruskal–Szekeres charts, and a freely falling observer crosses feeling only gentle tides. Confusing this coordinate singularity with the genuine curvature singularity at $r=0$ is the single most common error in the subject. A second trap rides alongside it: the coordinate $r$ is not the proper distance to the centre but the areal radius, fixed so the sphere at $r$ has area $4\pi r^2$ , with proper radial distance stretched by $(1-2GM/c^2 r)^{-1/2}$ .

Worked examples

1 · Anatomy of a solar-mass black hole (fully worked)

Take $M=M_\odot=1.989\times10^{30}$ kg. Find the horizon, photon-sphere, and ISCO radii, the maximum survival time inside the horizon, and the accretion efficiency at the ISCO.

Step 1. The gravitational length is $GM/c^2=(6.674\times10^{-11})(1.989\times10^{30})/(2.998\times10^8)^2=\mathbf{1.477\times10^{3}\ m}$ , so $r_s=2GM/c^2=\mathbf{2.95\ km}$ , the photon sphere $r_{\rm ph}=3GM/c^2=\mathbf{4.43\ km}$ , and $r_{\rm ISCO}=6GM/c^2=\mathbf{8.85\ km}$ — a whole star's mass sculpted into orbital structure a few kilometres across.

Step 2. The longest proper time any observer can log between crossing the horizon and hitting the singularity comes from falling from rest at the horizon, integrating $\mathrm{d}r/\mathrm{d}\tau=-c\sqrt{r_s/r}$ from $r_s$ to $0$ : $\tau_{\max}=\dfrac{\pi GM}{c^3}$ . With $GM/c^3=4.93\times10^{-6}$ s, this is $\tau_{\max}\approx\mathbf{1.5\times10^{-5}\ s}$ . Struggling makes it worse: firing rockets only shortens this already fleeting interval.

Step 3. A particle spiralling down a thin disk radiates until it reaches the ISCO, where $\tilde E_{\rm ISCO}=\sqrt{8/9}=0.9428$ , so the fraction of rest energy released is $1-\sqrt{8/9}=\mathbf{5.72\%}$ . This dwarfs hydrogen fusion's 0.7%, making accretion the most efficient steady power source known — the engine behind quasars.

2 · Surface gravity and the Hawking temperature (fully worked)

For the same solar-mass hole, evaluate $\kappa$ and the Hawking temperature $T=\hbar\kappa/2\pi k_B c$ .

Step 1. From Derivation C, $\kappa=c^4/4GM$ . With $c^4=8.08\times10^{33}$ and $4GM=5.31\times10^{20}$ , $\kappa=\mathbf{1.52\times10^{13}\ m/s^2}$ — about $10^{12}$ times Earth's surface gravity, redshifted to infinity.

Step 2. Insert into the Hawking formula ( $\hbar=1.055\times10^{-34}$ , $k_B=1.381\times10^{-23}$ ): $\hbar\kappa=1.60\times10^{-21}$ over $2\pi k_B c=2.60\times10^{-14}$ gives $T=\mathbf{6.2\times10^{-8}\ K}$ . A solar-mass hole is colder than the cosmic microwave background by eight orders of magnitude, so it absorbs far more than it radiates today.

Step 3. Note the scaling $\kappa\propto1/M$ and hence $T\propto1/M$ . Bigger black holes are colder — a $10^9\,M_\odot$ hole has $T\sim6\times10^{-17}$ K, while a lunar-mass primordial hole would glow at billions of kelvin. This inverse relation is the source of the negative heat capacity that makes black-hole thermodynamics unlike an ordinary gas.

Physics of the portrait

The Schwarzschild solution is where the abstract machinery of general relativity first touched something you could compute a number for — and it carries the name of a man who found it under conditions no theorist should have to work in.

Origins · 1 minKarl Schwarzschild and how this was discovered

Karl Schwarzschild

1873–1916 · the first exact solution of Einstein's field equations

Karl Schwarzschild was already a distinguished astronomer — director of the Potsdam Observatory, a pioneer of photographic photometry and stellar-atmosphere theory — when the First World War sent him to the Russian front as an artillery officer. There, in late 1915, he read Einstein's just-published field equations and, within weeks and between computing shell trajectories, produced the first exact solution: the vacuum geometry outside a spherical mass. He mailed it to Einstein in January 1916, who replied that he had not expected the exact solution to be so simple and presented it to the Prussian Academy on his behalf. He did not live to see what he had begun — he contracted pemphigus, a rare autoimmune disease, at the front and died in May 1916, aged 42. For decades the surface he exposed at $r=2GM/c^2$ was misread as an edge of the world and even called "the Schwarzschild singularity," before Eddington, Lemaître, Finkelstein, Kruskal and Szekeres showed it was a one-way horizon and that the only true singularity lay at the centre. The black hole his mathematics predicted would not be named for another half-century nor imaged until 2019, but its geometry is still, a century on, the first thing every relativist computes.

History

1915

— Einstein completes the field equations of general relativity

1916

— Schwarzschild derives the exact vacuum solution and dies months later

1923

— Birkhoff proves any spherical vacuum field is static and Schwarzschild

1958

— Finkelstein recasts r = 2GM as a one-way causal horizon in regular coordinates

1960

— Kruskal and Szekeres construct the maximal analytic extension

Going further

Texts. Misner, Thorne & Wheeler, Gravitation (1973), Chs. 23, 25, 31 · Wald, General Relativity (1984), Ch. 6 and §12.5 · Carroll, Spacetime and Geometry (2004), Ch. 5 · Hartle, Gravity (2003), Chs. 9, 12 · Schutz, A First Course in General Relativity (2009), Ch. 11 · Schwarzschild, K. (1916), Sitzungsber. Preuss. Akad. Wiss. 189.

Provenance. The metric, $r_s$ , photon sphere, ISCO, Kretschmann scalar, surface gravity and Birkhoff's theorem are standard textbook results tracing to MTW (1973), Wald (1984) Ch. 6, Carroll (2004) Ch. 5, Hartle (2003) and Schutz (2009) Ch. 11. Numbers checked here: solar $r_s\approx2.95$ km, ISCO efficiency $1-\sqrt{8/9}\approx5.7\%$ , maximum interior proper time $\pi GM/c^3\approx1.5\times10^{-5}$ s, surface gravity $\kappa=c^4/4GM\approx1.5\times10^{13}\ \mathrm{m/s^2}$ and Hawking temperature $\approx6.2\times10^{-8}$ K for a solar mass. A verifier should recompute $K$ from the Schwarzschild Christoffel symbols, confirm the surface gravity via the redshifted-acceleration limit, and check the ISCO specific energy $\sqrt{8/9}$ from $\mathrm{d}V_{\rm eff}^2/\mathrm{d}r=0$ at the inflection.

Prerequisites. Einstein Field Equations · Spacetime Metric and Line Element · Geodesics and the Geodesic Equation · Killing Vectors and Symmetry. Leads to. Event Horizon and Apparent Horizon · Killing Horizons and Surface Gravity · Reissner–Nordström Charged Black Hole · Kerr Metric and Rotating Black Holes · No-Hair Theorem and Black Hole Uniqueness · Four Laws of Black Hole Mechanics · Trapped Surfaces and Closed Trapped Regions.

Ask the tutor: ask me to walk you from the single line element to why the very same blow-up is a harmless horizon at $r=2GM/c^2$ but the end of spacetime at $r=0$ , and I will show you the invariant that decides it.

If I fell through the horizon, what would I feel at the crossing?

Locally, nothing dramatic. The event horizon of a large black hole is an ordinary region of spacetime — the curvature invariant $K$ is finite there, and by the equivalence principle a small freely falling laboratory registers only tidal forces, gentle enough for a supermassive hole that you would cross without noticing. There is no wall, no surface, no jolt; you cannot even locally tell the moment you crossed. What has changed is global and causal: every future-directed path now leads to smaller $r$ , and no signal you send can ever again reach the outside. For a stellar-mass hole the tides near the horizon are already lethal, but that is a statement about its small size, not about the horizon being a physical barrier.

Why is the singularity described as a time rather than a place?

Inside the horizon the factor $(1-2GM/c^2 r)$ is negative, so $g_{tt}>0$ and $g_{rr}<0$ : the radial coordinate $r$ has become timelike and $t$ spacelike. Because $r$ is now a time coordinate that can only decrease along any future-directed worldline, reaching $r=0$ is as unavoidable as reaching next Tuesday. You cannot steer away from the singularity any more than from your own future — every direction you can accelerate is a direction in space, and $r=0$ lies in time. This is why "the singularity sucks you in" is the wrong picture: it is not a location in front of you but a moment ahead of you, and firing your engines only shortens the proper time you have first.

Does a black hole pull things in more strongly than an ordinary star?

No — at a given distance, not at all. Far from the hole the metric is indistinguishable from that of any mass $M$ : replace the Sun with a solar-mass black hole and Earth's orbit would not shift by a millimetre, because Birkhoff's theorem makes the exterior field depend only on the enclosed mass. What differs is that a black hole has no surface to stop you, so you can approach far closer than $r_s$ — down to the photon sphere and the ISCO — where the field becomes overwhelming. The "irresistible pull" is a feature of getting close, which only a black hole lets you do, not of it reaching out further than a star of equal mass. Distant orbits around Sgr A* are, to excellent approximation, ordinary Keplerian ellipses.

What are the white hole and the second universe in the Penrose diagram?

They emerge when you extend the solution to its maximal analytic form — the largest spacetime consistent with the same metric and no artificial edges. Region II is the familiar black-hole interior; its time-reverse, region IV, is a white hole, a past singularity that only expels matter and can never be entered; region III is a second asymptotically flat universe, joined to ours at a moment by the Einstein–Rosen bridge, a wormhole throat. Crucially, this is a feature of the eternal, exactly static vacuum solution, not of a real black hole. A hole formed by stellar collapse replaces regions III and IV with the infalling star, so no white hole and no other universe actually exist — and even in the idealized case the bridge pinches off faster than any timelike curve can cross, so it is non-traversable.

Why does distant light from an infalling object appear to freeze at the horizon?

Because two different clocks are involved. In the infalling object's own proper time the crossing of $r=2GM/c^2$ is quick and unremarkable. But the Schwarzschild coordinate $t$ is the distant observer's time, and $t\to\infty$ as the object approaches the horizon, so each successive photon it emits takes longer to climb out and arrives more redshifted. The distant observer therefore sees the image slow, redden, and dim toward invisibility, asymptotically frozen at the horizon but never quite reaching it. Neither description is wrong — they are the same worldline in two charts. The freezing is a coordinate effect of $t$ , the same pathology that makes $g_{rr}$ diverge, which is why the infaller, using proper time, sails straight through.

Why is there no monopole gravitational radiation from a pulsating star?

This is a direct corollary of Birkhoff's theorem, which says any spherically symmetric vacuum region is static Schwarzschild, determined solely by the enclosed mass. So if a star pulsates while staying perfectly spherical, its exterior field cannot change at all — it stays exactly Schwarzschild with the same $M$ throughout. A time-varying field is what would carry gravitational waves away, and here there is none, so a radially oscillating star radiates nothing gravitationally. The lesson generalizes: gravitational radiation requires a changing quadrupole moment or higher, because the monopole is fixed by mass conservation and the dipole by momentum conservation. This is the relativistic reason the leading emission from any source is quadrupolar, as LIGO's binary-inspiral signals confirm.