Stellar-Mass Black Holes

A stellar-mass black hole is what a massive star's core becomes when it runs out of ways to hold itself up. A white dwarf leans on electron degeneracy pressure, capped at the Chandrasekhar mass $\approx 1.44\,M_\odot$ . Push past that and the core collapses to nuclear density, where neutron degeneracy and the strong force take over — but they too have a ceiling, the Tolman–Oppenheimer–Volkoff limit near $2$ – $3\,M_\odot$ , set by the dense-matter equation of state. Above it no pressure can win, and general relativity is unambiguous: the core falls inside its own Schwarzschild radius and seals itself behind an event horizon. What emerges is startlingly simple. By the no-hair theorem an isolated astrophysical black hole is described completely by two numbers, its mass $M$ and spin $J$ — Schwarzschild for the non-rotating case, Kerr for the rotating one. Everything the progenitor was — composition, magnetic field, lumpy shape — is swallowed or radiated away. The horizon $R_S = 2GM/c^2$ is about $30\,\mathrm{km}$ for a $10\,M_\odot$ hole, and the innermost stable circular orbit at $6GM/c^2$ fixes how brightly infalling gas can shine before it plunges. Because the horizon emits nothing detectable — Hawking radiation from a stellar hole corresponds to $\sim 10^{-8}\,\mathrm{K}$ — these objects are found by their gravity. X-ray binaries betrayed the first ones, Cygnus X-1 above all, through a companion whipping around an unseen mass; since 2015, gravitational waves from merging pairs have opened a second, heavier population and turned black-hole demographics into a precision science.

Key takeaways

Stellar-mass black holes span roughly $3\,M_\odot$ to a few tens of $M_\odot$ ; they form when a stellar core exceeds the maximum mass ( $M_{\mathrm{TOV}}\approx 2$ – $3\,M_\odot$ ) that neutron degeneracy plus the strong force can support.
Their master scale is the Schwarzschild radius $R_S = 2GM/c^2 \approx 2.95\,\mathrm{km}\,(M/M_\odot)$ , the radius of the event horizon from which no signal escapes.
A non-rotating hole is the Schwarzschild metric; a spinning one is Kerr, with dimensionless spin $a_* = Jc/GM^2 \le 1$ shrinking the innermost stable orbit from $6GM/c^2$ toward $GM/c^2$ .
Emitting no light of their own, they are weighed dynamically: an X-ray binary's mass function sets a firm lower bound on the compact mass, and $f(M)\gtrsim 3\,M_\odot$ already means a black hole.
Accretion converts $5.7\%$ (Schwarzschild) to $42\%$ (extremal Kerr) of rest-mass energy to radiation — the most efficient steady power source in nature, roughly ten times nuclear fusion.
Gravitational-wave mergers (LIGO/Virgo since GW150914) reveal a population up to tens of $M_\odot$ , probing massive-star collapse and the pair-instability gap.
Key idea: when a stellar core exceeds the maximum mass any known pressure can support, gravitational collapse produces an event horizon — a black hole described by just its mass and spin.

The core insight

When a stellar core exceeds the maximum mass that any known pressure (electron degeneracy, then neutron degeneracy plus the strong force) can support, gravitational collapse produces an event horizon — a black hole described completely by just its mass and spin.

Where this lives today

Stellar-mass black holes are now studied across three complementary channels. X-ray binaries remain the workhorse for spins: fitting the disk's thermal continuum, or modeling the relativistically broadened iron K $\alpha$ line at $6.4\,\mathrm{keV}$ smeared by gravitational redshift near the ISCO, pins down $a_*$ — some holes, like GRS 1915+105, appear near-extremal. Gravitational-wave catalogs from LIGO–Virgo–KAGRA now hold roughly a hundred mergers, and their statistics are rewriting formation theory: a mass peak near $\sim 35\,M_\odot$ , a dearth of remnants in the $3$ – $5\,M_\odot$ "lower mass gap," and events like GW190521 with components inside the pair-instability gap ( $\sim 50$ – $130\,M_\odot$ ), hinting at hierarchical mergers in dense clusters. The third channel is astrometry: Gaia now reveals quiescent, non-accreting black holes (Gaia BH1, BH2, BH3) by the wobble they induce in ordinary companions, a population the X-ray binaries miss. Open problems remain — the true maximum neutron-star mass and hence the lower edge of the distribution, the origin of natal spins and kicks, and whether the horizon is truly featureless or carries quantum structure.

The physical picture

Gravity wins by default; it only loses when something pushes back hard enough. A star arranges that push — thermal pressure from fusion, then, once the fuel is gone, quantum degeneracy pressure. A white dwarf leans on its electrons' degeneracy pressure, but that saturates: above the Chandrasekhar mass the electrons would have to move faster than light, and the star collapses to neutron-star densities. There the neutrons take over, aided by the strong force's repulsive core, and again there is a ceiling — the TOV limit near $2$ – $3\,M_\odot$ . Cross it, and the ladder of supports is exhausted; nothing holds the core up, and it falls inward until the escape velocity at its surface reaches the speed of light. The Schwarzschild radius is a beautifully cheap way to see where that happens. Set the Newtonian escape speed $\sqrt{2GM/r}$ equal to $c$ and solve: $R_S = 2GM/c^2$ . The argument is only heuristic — it mixes Newtonian energy with a relativistic speed limit — yet it lands exactly on the horizon radius the full Schwarzschild solution gives. For a $10\,M_\odot$ hole that is a sphere just $30\,\mathrm{km}$ across, holding ten Suns of mass. Around it lies more structure fixed entirely by $M$ : a photon sphere at $1.5\,R_S$ where light itself can orbit, and an innermost stable circular orbit at $3\,R_S$ , the last ledge from which matter can circle before plunging.

Figure

Anatomy of a non-rotating 10 M⊙ black hole: the event horizon at R_S = 2GM/c² = 29.5 km, the photon sphere at 1.5 R_S = 44.3 km where light can orbit, and the ISCO at 3 R_S = 88.6 km — the inner edge of the accretion disk, which sets the 5.7% radiative efficiency

Everything a distant observer can measure reduces to two numbers. The progenitor's individuality — rotation profile, magnetic field, chemistry — is hidden behind the horizon or shaken off as gravitational radiation during formation. This no-hair simplicity is why black-hole physics is so predictive: give me $M$ and spin $a_*$ and I can compute the horizon, the orbits, the disk temperature, and a merger's waveform. Spin is the second lever, and it matters most where matter gives up its energy: as a hole spins faster the innermost stable orbit is dragged inward, letting gas circle deeper and radiate a larger fraction of its rest mass before it falls — the relation that makes black holes the brightest steady engines in the universe.

Figure

Radiative efficiency η = 1 − E_ISCO of a thin accretion disk versus black-hole spin a_: from η = 5.7% for a non-rotating Schwarzschild hole (a_ = 0, ISCO at 6GM/c²) up to η = 42% for extremal Kerr (a_ = 1, ISCO at GM/c²), with the realistic accretion-spun ceiling a_ ≈ 0.998 giving η ≈ 32%

We never see the hole itself; we read it from what it does to its neighborhood — the swing of a companion's orbit, the X-ray glow of gas heated to millions of kelvin near the ISCO, or the chirp of spacetime when two horizons merge and ring down into one.

The core equations

The horizon is the length against which everything else is measured. Equating the Newtonian escape speed to $c$ — a heuristic that reproduces the exact Schwarzschild result — gives the radius from which even light cannot climb out.

R_S = \frac{2GM}{c^2} \approx 2.95\,\mathrm{km}\left(\frac{M}{M_\odot}\right) \qquad (1)

where: $R_S$ — event-horizon radius of a non-rotating hole; $G$ — gravitational constant, $6.674\times10^{-11}\,\mathrm{m^3\,kg^{-1}\,s^{-2}}$ ; $M$ — black-hole mass; $c$ — speed of light, $2.998\times10^8\,\mathrm{m\,s^{-1}}$ . Reading: the horizon grows linearly with mass, so a $30\,M_\odot$ hole is only $\sim 90\,\mathrm{km}$ across — compactness, not bulk, is what defines it. The geometry outside a spherical, non-rotating mass is the Schwarzschild metric, the first exact solution of Einstein's equations. The factor $(1-R_S/r)$ appears twice: it slows clocks and stretches radial distances as $r\to R_S$ .

c^2\,\mathrm{d}\tau^2 = \left(1-\frac{R_S}{r}\right)c^2\,\mathrm{d}t^2 - \left(1-\frac{R_S}{r}\right)^{-1}\mathrm{d}r^2 - r^2\,\mathrm{d}\Omega^2 \qquad (2)

where: $\tau$ — proper time; $t$ — far-away coordinate time; $r$ — areal radial coordinate ( $4\pi r^2$ is the sphere's area); $\mathrm{d}\Omega^2 = \mathrm{d}\theta^2 + \sin^2\theta\,\mathrm{d}\varphi^2$ . Reading: at $r=R_S$ the $g_{tt}$ term vanishes and $g_{rr}$ diverges — a coordinate horizon, not a physical singularity; an infaller notices nothing local as they cross. Astrophysical holes rotate, and spin is a single dimensionless number normalized so the extremal Kerr hole sits at unity. Charge is negligible, so mass and spin are the whole story.

a_* \equiv \frac{Jc}{GM^2}, \qquad 0 \le a_* \le 1 \qquad (3)

where: $J$ — angular momentum; $a_*$ — spin normalized so $a_*=1$ is the maximal (extremal) Kerr value. Reading: $a_*>1$ would expose a naked singularity, forbidden by cosmic censorship; accreting holes are torqued up but stall near $a_*\approx 0.998$ (the Thorne limit). Rotation splits the single horizon into two. Setting the Kerr metric function $\Delta = r^2 - (2GM/c^2)r + (GMa_*/c^2)^2$ to zero gives the outer event horizon and inner Cauchy horizon.

r_{\pm} = \frac{GM}{c^2}\left(1 \pm \sqrt{1 - a_*^2}\right) \qquad (4)

where: $r_+$ — outer event horizon; $r_-$ — inner (Cauchy) horizon. Reading: at $a_*=0$ this recovers $r_+=R_S=2GM/c^2$ ; at $a_*=1$ the two horizons merge at $GM/c^2$ , half the Schwarzschild radius. Accretion cannot shine arbitrarily brightly: at the Eddington luminosity the outward radiation force on ionized gas (via Thomson scattering) balances gravity, and steady accretion self-regulates near this value.

L_{\mathrm{Edd}} = \frac{4\pi G M m_p c}{\sigma_T} \approx 1.26\times10^{38}\,\mathrm{erg\,s^{-1}}\left(\frac{M}{M_\odot}\right) \qquad (5)

where: $m_p$ — proton mass, $1.673\times10^{-27}\,\mathrm{kg}$ ; $\sigma_T$ — Thomson cross section, $6.652\times10^{-25}\,\mathrm{cm^2}$ ; $L_{\mathrm{Edd}}$ — luminosity where radiation pressure balances gravity. Reading: a $10\,M_\odot$ hole caps near $1.3\times10^{39}\,\mathrm{erg\,s^{-1}}$ , among the brightest point X-ray sources in a galaxy. The luminosity per unit accreted mass is set by a single efficiency $\eta$ , the fraction of rest-mass energy released, fixed by the ISCO binding energy. Inverting it gives the accretion rate a given luminosity demands.

\dot{M} = \frac{L}{\eta c^2}, \qquad \eta \approx 0.057\ (\text{Schwarzschild}),\ 0.42\ (\text{extremal Kerr}) \qquad (6)

where: $\dot{M}$ — mass accretion rate; $L$ — disk luminosity; $\eta$ — radiative efficiency set by the ISCO. Reading: because $\eta$ ranges over almost a factor of eight with spin, the same luminosity can imply very different fuel consumption — spin is a hidden variable in accretion budgets. To weigh an unseen compact object we watch its visible companion. Kepler's law plus the projected orbital velocity yields the mass function — built from observables alone and a strict lower bound on the compact mass.

f(M) = \frac{P_{\mathrm{orb}}\,K^3}{2\pi G} = \frac{M_{\mathrm{BH}}^3 \sin^3 i}{(M_{\mathrm{BH}} + M_c)^2} \qquad (7)

where: $P_{\mathrm{orb}}$ — orbital period; $K$ — radial-velocity semi-amplitude of the companion; $M_{\mathrm{BH}}$ — compact mass; $M_c$ — companion mass; $i$ — inclination. Reading: since $\sin i \le 1$ and $M_c>0$ , $f(M)\le M_{\mathrm{BH}}$ — so $f(M)\gtrsim 3\,M_\odot$ alone proves a black hole, with no need to know $i$ or $M_c$ . Finally, the horizon carries entropy proportional to its area, not its volume — the deepest hint that gravity, thermodynamics, and quantum theory are one subject.

S_{\mathrm{BH}} = \frac{k_B c^3 A}{4 G \hbar} = \frac{k_B A}{4\ell_P^2} \qquad (8)

where: $A = 4\pi R_S^2$ — horizon area; $\ell_P = \sqrt{G\hbar/c^3}\approx 1.6\times10^{-35}\,\mathrm{m}$ — Planck length; $k_B$ — Boltzmann constant. Reading: one bit lives on every $4\ell_P^2$ of horizon — a stellar hole holds more entropy than everything that ever fell into it, the seed of the holographic principle.

Derivation AThe innermost stable circular orbit in Schwarzschild geometry

The goal is to show that in the Schwarzschild spacetime a freely orbiting particle has a hard inner limit at $r=6GM/c^2$ , below which no stable circular orbit exists at all — a purely relativistic effect with no Newtonian analogue, since Newtonian gravity permits stable circular orbits at every radius. This single radius sets the inner edge of an accretion disk and the binding energy available to be radiated, and so controls the efficiency of the brightest engines in the cosmos. We build it from the effective potential for a timelike geodesic.

Step 1. Conservation of energy and angular momentum along a geodesic collapses the radial equation of motion into $\big(\tfrac{\mathrm{d}r}{\mathrm{d}\tau}\big)^2 = \tilde E^2 c^2 - \tilde V_{\mathrm{eff}}^2$ , with $\tilde V_{\mathrm{eff}}^2(r) = \left(1-\tfrac{R_S}{r}\right)\!\left(1 + \tfrac{\tilde L^2}{r^2 c^2}\right)$ and $\tilde L = L/(mc)$ the specific angular momentum. Why: a one-dimensional potential problem lets us find orbits by the same extremum reasoning as Newtonian mechanics — the relativity is entirely inside the shape of $\tilde V_{\mathrm{eff}}$ .

Step 2. A circular orbit sits at a turning point, so impose $\mathrm{d}\tilde V_{\mathrm{eff}}^2/\mathrm{d}r = 0$ , giving the required $\tilde L^2 = \dfrac{GMr}{c^2}\,\dfrac{1}{1 - 3GM/(rc^2)}$ , which diverges at $r\to 3GM/c^2$ — the photon sphere. Why: the blow-up shows circular orbits demand ever more angular momentum as $r$ shrinks, unlike Kepler's gentle $1/r$ well.

Step 3. Stability requires a genuine minimum, so the marginally stable orbit is where minimum and maximum merge: $\mathrm{d}^2\tilde V_{\mathrm{eff}}^2/\mathrm{d}r^2 = 0$ . Imposing this with Step 2 eliminates $\tilde L$ and leaves one equation for $r$ . Why: an inflection point is the boundary between orbits that recover after a nudge and orbits that spiral away — the definition of "innermost stable".

Step 4. Solving the pair yields $r_{\mathrm{ISCO}} = 6GM/c^2 = 3R_S$ . Inside it the potential has no minimum, so any perturbed orbit plunges. Why: this is the disk's inner wall — gas drifts inward through near-circular orbits until $3R_S$ , then falls in freely, carrying its binding energy into the hole. *The ISCO at $6GM/c^2=3R_S$ is a standard result of Schwarzschild geodesic theory; for $M=10\,M_\odot$ this is $88.6\,\mathrm{km}=3\times29.5\,\mathrm{km}$ .*

Go deeper · Derivation B · 1 minThe radiative efficiency of black-hole accretion

Derivation BThe radiative efficiency of black-hole accretion

The goal is to compute the fraction of rest-mass energy that accreting matter radiates before crossing the horizon, and to show it equals the binding energy per unit mass at the ISCO — because a thin disk radiates the energy lost as gas spirals from rest at infinity down to the last stable orbit, then swallows the rest whole. The result, $\eta\approx 5.7\%$ for a non-spinning hole, makes black holes the most efficient steady energy source known, roughly an order of magnitude above hydrogen fusion. We evaluate the conserved specific energy of a circular orbit at the ISCO.

Step 1. The specific energy of a circular geodesic is $\tilde E(r) = \dfrac{1-R_S/r}{\sqrt{1 - 3GM/(rc^2)}}$ , with $\tilde E = E/(mc^2)$ so that $\tilde E\to 1$ for a particle at rest infinitely far away. Why: normalizing to rest-mass energy makes the deficit $1-\tilde E$ read directly as the fraction of $mc^2$ given up to radiation.

Step 2. Evaluate at $r_{\mathrm{ISCO}}=6GM/c^2$ , so $R_S/r = 1/3$ and $3GM/(rc^2)=1/2$ : $\tilde E_{\mathrm{ISCO}} = \dfrac{1-1/3}{\sqrt{1-1/2}} = \dfrac{2/3}{1/\sqrt2} = \sqrt{\tfrac{8}{9}}$ . Why: the ISCO ends the disk, so the energy still bound in orbit there is the most that could have been radiated on the way in — anything deeper is lost across the horizon.

Step 3. The radiated fraction is the binding energy $\eta = 1 - \tilde E_{\mathrm{ISCO}} = 1 - \sqrt{8/9} \approx 0.057$ . For a maximal prograde disk the ISCO reaches $GM/c^2$ , $\tilde E_{\mathrm{ISCO}}=1/\sqrt3$ , and $\eta = 1 - 1/\sqrt3 \approx 0.42$ . Why: spin drags the last stable orbit inward, so gas falls deeper and sheds more binding energy — the climb toward extremal Kerr. *The efficiencies work out to $\eta_{\rm Schw}=1-\sqrt{8/9}=0.0572$ and $\eta_{\rm extremal}=1-1/\sqrt3=0.4226$ , compared with hydrogen fusion at $\sim0.7\%$ .*

Go deeper · Derivation C · 1 minThe binary mass function as a lower bound on the compact mass

Derivation CThe binary mass function as a lower bound on the compact mass

The goal is to derive the mass function — the quantity that turned "there is an unseen companion" into "the companion must be a black hole" — and to prove it is a rigorous lower limit on the compact mass built from two observables alone: the orbital period and the companion's line-of-sight velocity amplitude. This matters because inclination and companion mass are usually unknown, yet the mass function needs neither. We combine Kepler's third law with the geometry of the companion's projected orbit.

Step 1. For a circular orbit Kepler's third law reads $P_{\mathrm{orb}}^2 = \dfrac{4\pi^2 a^3}{G(M_{\mathrm{BH}}+M_c)}$ , with $a$ the total separation. Why: this ties the timed period directly to total mass and separation, the anchor every dynamical mass estimate hangs from.

Step 2. Spectroscopy measures only the visible star's line-of-sight velocity; its orbit about the center of mass has radius $a_c = a\,M_{\mathrm{BH}}/(M_{\mathrm{BH}}+M_c)$ , giving $K = 2\pi a_c \sin i/P_{\mathrm{orb}}$ . Why: the $\sin i$ enters because we see only the projection — a face-on orbit ( $i=0$ ) shows no Doppler swing at all.

Step 3. Eliminating $a$ and $a_c$ packages the observables into $f(M) = \dfrac{P_{\mathrm{orb}}K^3}{2\pi G} = \dfrac{M_{\mathrm{BH}}^3\sin^3 i}{(M_{\mathrm{BH}}+M_c)^2}$ . Since $\sin i\le 1$ and $M_c>0$ , the right side never exceeds $M_{\mathrm{BH}}$ , so $f(M)\le M_{\mathrm{BH}}$ . Why: the inequality is the payoff — $f(M)$ is a floor on the compact mass no geometry can lift; if it sits above the neutron-star ceiling, the case is closed. *The inequality $f(M)\le M_{\mathrm{BH}}$ holds for all geometries, with equality only when $i=90^\circ$ and $M_c\to 0$ . Applied to V404 Cygni ( $P=6.47\,\mathrm{d}$ , $K=208.5\,\mathrm{km\,s^{-1}}$ ) it gives $f(M)=6.08\,M_\odot$ — above the TOV limit, hence a black hole.*

Limiting cases

Black-hole phenomenology is organized by two dials, mass and spin, through a handful of characteristic scales:

Case	Result	Meaning
Non-rotating (Schwarzschild)	$a_* = 0,\quad r_{\mathrm{ISCO}} = 6GM/c^2 = 3R_S$	Reference case: horizon at $R_S$ , photon sphere at $1.5R_S$ , ISCO at $3R_S$ , efficiency $\eta\approx 5.7\%$ .
Maximally rotating (extremal Kerr)	$a_* \to 1,\quad r_{\mathrm{ISCO}} \to GM/c^2$	Prograde ISCO shrinks toward the horizon; efficiency rises to $\eta\approx 42\%$ . Cosmic censorship forbids $a_*>1$ .
Weak field / far away	$g_{tt} \to -\left(1 - \tfrac{2GM}{c^2 r}\right),\quad r\gg R_S$	Spacetime reduces to Newtonian gravity; orbits are Keplerian ellipses — the regime in which a companion's swing weighs the hole.
Quantum / Hawking limit	$T_H = \dfrac{\hbar c^3}{8\pi G M k_B}\approx 6\times10^{-8}\,\mathrm{K}\left(\dfrac{M_\odot}{M}\right)$	For stellar masses this is far below the $2.7\,\mathrm{K}$ CMB, so the hole absorbs more than it emits and grows; evaporation is astronomically irrelevant.

Common trapThe event horizon is a "surface" where infalling matter is crushed against a wall

It is not. The horizon is a global causal boundary, not a material surface: for a large hole a freely falling observer crosses it without any local signal — no wall, no flash, no crush. Its specialness is only global: no future-directed path from inside can ever reach infinity. Two further traps ride along. First, the "size" of a black hole is its horizon radius $R_S\propto M$ , not a mass distribution — in classical GR the mass sits at the central singularity, not spread over a sphere of radius $R_S$ . Second, that singularity is where classical general relativity predicts its own breakdown; it is a signpost that new physics is needed, not a well-understood point mass. Crushing does happen — but at the singularity deep inside, and for a stellar hole the tidal "spaghettification" sets in outside the horizon precisely because the hole is so small.

Worked examples

1 · Proving a black hole from a mass function (fully worked)

The X-ray transient V404 Cygni has orbital period $P_{\mathrm{orb}}=6.47\,\mathrm{days}$ and companion radial-velocity semi-amplitude $K=208.5\,\mathrm{km\,s^{-1}}$ . Show the compact object must be a black hole, using observables alone.

Step 1. In SI: $P_{\mathrm{orb}} = 6.4714\times 86\,400 = 5.591\times10^5\,\mathrm{s}$ , $K = 2.085\times10^5\,\mathrm{m\,s^{-1}}$ , so $K^3 = 9.06\times10^{15}\,\mathrm{m^3\,s^{-3}}$ .

Step 2. Mass function $f(M) = P_{\mathrm{orb}}K^3/(2\pi G)$ . Numerator $5.591\times10^5 \times 9.06\times10^{15} = 5.07\times10^{21}$ ; denominator $2\pi G = 4.19\times10^{-10}$ . So $f(M) = 1.21\times10^{31}\,\mathrm{kg}$ .

Step 3. In solar masses: $f(M) = 1.21\times10^{31}/1.989\times10^{30} = \mathbf{6.08\,M_\odot}$ .

Step 4. Interpret. Because $f(M)\le M_{\mathrm{BH}}$ for any inclination and any $M_c>0$ , we have $M_{\mathrm{BH}}\ge 6.08\,M_\odot$ — more than double the TOV ceiling, so the object is a black hole regardless of the unknown geometry. (Full modeling gives $M_{\mathrm{BH}}\approx 9\,M_\odot$ ; the mass function alone already settles it.)

2 · Horizon, ISCO, and Eddington limit of a 10 M⊙ hole (fully worked)

For a non-rotating $10\,M_\odot$ hole, find the horizon radius, ISCO radius, Eddington luminosity, and the accretion rate needed to shine at Eddington with Schwarzschild efficiency.

Step 1. Horizon: $R_S = 2.95\,\mathrm{km}\times(M/M_\odot) = \mathbf{29.5\,km}$ ; photon sphere $1.5R_S = 44.3\,\mathrm{km}$ .

Step 2. ISCO: $r_{\mathrm{ISCO}} = 3R_S = \mathbf{88.6\,km}$ — the inner edge of any thin accretion disk.

Step 3. Eddington luminosity: $L_{\mathrm{Edd}} = 1.26\times10^{38}\,\mathrm{erg\,s^{-1}}\times 10 = \mathbf{1.26\times10^{39}\,erg\,s^{-1}} = 1.26\times10^{32}\,\mathrm{W}$ — brighter than $3\times10^5$ Suns from a city-sized region.

Step 4. Accretion rate at Eddington, using $\dot M = L/(\eta c^2)$ with $\eta = 0.057$ : $\dot M = 1.26\times10^{32}/(0.057\times 8.99\times10^{16}) = 2.46\times10^{16}\,\mathrm{kg\,s^{-1}} \approx \mathbf{3.9\times10^{-7}\,M_\odot\,yr^{-1}}$ — under a millionth of a solar mass per year sustains a supersolar X-ray source. A maximal-spin hole ( $\eta=0.42$ ) needs about eight times less fuel for the same glow.

3 · The energy released in a black-hole merger

GW150914 merged a $36\,M_\odot$ and a $29\,M_\odot$ hole into a $\approx 62\,M_\odot$ remnant. How much energy left as gravitational waves, and how does its peak power compare to starlight?

Step 1. Mass deficit: $36 + 29 - 62 = 3\,M_\odot$ , radiated as gravitational waves.

Step 2. Energy: $E = 3\,M_\odot c^2 = 3\times 1.989\times10^{30}\times(2.998\times10^8)^2 = \mathbf{5.4\times10^{47}\,J}$ — comparable to the Sun's total electromagnetic output over its entire $\sim10^{10}$ -year life, released here in under a tenth of a second.

Step 3. Peak luminosity reaches $\sim 3.6\times10^{49}\,\mathrm{W}$ , briefly exceeding the combined light of every star in the observable universe — yet purely in spacetime strain, detectable only as a $\sim 10^{-21}$ fractional stretch by LIGO.

Physics of the portrait

The stellar-mass black hole was theory long before observation: a chain of collapse calculations, from white dwarfs to neutron stars to horizons, worked out on paper decades before any instrument could see one. The figure who first followed that chain to its end is J. Robert Oppenheimer.

Origins · 1 minJ. Robert Oppenheimer and how this was discovered

J. Robert Oppenheimer

1904–1967 · gravitational collapse and the maximum mass of degenerate stars

Oppenheimer is known to the public for the Manhattan Project, but among astrophysicists his most enduring work is a pair of 1939 papers that founded the theory of gravitational collapse. With George Volkoff, using the relativistic hydrostatic equation Richard Tolman had derived, he integrated the structure of a neutron star and found it too has a maximum mass — the Tolman–Oppenheimer–Volkoff limit. With his student Hartland Snyder he then answered what happens above it, in "On Continued Gravitational Contraction": a sufficiently massive pressureless sphere collapses without limit, its surface appearing to a distant observer to freeze at the Schwarzschild radius as its light reddens to nothing, while an observer riding the surface sails smoothly through in finite proper time. This is the modern black hole — frozen-star illusion and benign horizon crossing alike — written down in 1939. The work was then set aside for two decades, doubted even by Einstein and Eddington. The revival came in the 1960s: Kerr found the rotating solution in 1963, John Wheeler coined "black hole" around 1967, and Cygnus X-1 gave the first strong candidate. The ladder Oppenheimer climbed — Chandrasekhar mass, TOV limit, continued collapse — is exactly the argument this article rests on.

History

1916

— Schwarzschild solves Einstein's equations for a point mass, giving the horizon radius

1931

— Chandrasekhar derives the maximum mass of a white dwarf

1939

— Oppenheimer, Volkoff and Snyder establish the neutron-star mass limit and continued collapse

1963

— Kerr finds the exact solution for a rotating black hole

1971

— Cygnus X-1 is identified as a stellar-mass black-hole candidate

2015

— LIGO detects GW150914, the first binary black-hole merger

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., Ch. 16–17 & 24 · Shapiro & Teukolsky, Black Holes, White Dwarfs, and Neutron Stars (1983), Ch. 9 & 12 · Misner, Thorne & Wheeler, Gravitation, Ch. 25, 31–33 · Frank, King & Raine, Accretion Power in Astrophysics, 3rd ed. · Binney & Tremaine, Galactic Dynamics, 2nd ed. The Schwarzschild radius $R_S=2GM/c^2$ , ISCO $=6GM/c^2=3R_S$ , and efficiency $\eta=1-\sqrt{8/9}\approx0.057$ follow from the Schwarzschild effective potential; Kerr horizons from $\Delta=0$ ; extremal efficiency $1-1/\sqrt3\approx0.42$ . The Eddington luminosity $1.26\times10^{38}(M/M_\odot)\,\mathrm{erg\,s^{-1}}$ comes from $4\pi GMm_pc/\sigma_T$ with $\sigma_T=6.652\times10^{-25}\,\mathrm{cm^2}$ . The Chandrasekhar mass $1.44\,M_\odot$ and TOV limit $\sim2$ – $3\,M_\odot$ are standard results. V404 Cyg: $f(M)=6.08\,M_\odot$ from $P=6.47\,\mathrm{d}$ , $K=208.5\,\mathrm{km\,s^{-1}}$ . GW150914: $36+29\to62\,M_\odot$ , $3\,M_\odot c^2$ radiated. Cyg X-1: $\sim21\,M_\odot$ . Hawking temperature: $T_H=\hbar c^3/(8\pi GMk_B)\approx6\times10^{-8}(M_\odot/M)\,\mathrm{K}$ .

Builds on: Neutron Stars · Tolman–Oppenheimer–Volkoff Equation · Dense-Matter Equation of State · Chandrasekhar Mass Limit. Leads to: Supermassive Black Holes · Gravitational Waves · Accretion Disks.

Ask the tutor: Want me to walk through why the ISCO sits at exactly $6GM/c^2$ , or how a merger waveform's chirp encodes both masses and spins?

If nothing escapes a black hole, how do we know they exist?

We never detect the hole directly; we detect its gravity. In an X-ray binary an ordinary star orbits an invisible companion, and its periodic Doppler shift yields a mass function that floors the companion's mass — when that floor exceeds the neutron-star maximum ( $\sim3\,M_\odot$ ), only a black hole remains. Gas falling in heats to millions of kelvin and radiates X-rays varying on millisecond timescales, implying an emitter tens of kilometers across. And since 2015, gravitational waves let us "hear" two horizons spiral together, reading their masses and spins from the waveform. Three independent methods, one conclusion.

What is the difference between the Schwarzschild and Kerr solutions?

Both are exact vacuum solutions of Einstein's equations for an isolated hole; the difference is spin. Schwarzschild depends on mass alone: one spherical horizon at $R_S=2GM/c^2$ and an ISCO at $3R_S$ . Kerr adds angular momentum through the dimensionless spin $a_*=Jc/GM^2$ . Rotation splits the horizon into outer (event) and inner (Cauchy) horizons, drags spacetime around with the hole (frame dragging, creating an ergosphere), and pulls the prograde ISCO inward from $6GM/c^2$ toward $GM/c^2$ as $a_*\to1$ . Real astrophysical holes are Kerr — they inherit and accrete angular momentum — so spin is a genuine, measurable second parameter.

Why is accretion onto a black hole so much more efficient than fusion?

Efficiency means the fraction of rest-mass energy $mc^2$ turned into radiation. Hydrogen fusion converts about $0.7\%$ — the mass difference between four protons and a helium nucleus. Accretion instead taps gravitational binding energy: gas must shed energy to spiral deeper, and by the innermost stable orbit it has radiated $\approx5.7\%$ of $mc^2$ for a non-spinning hole, up to $\approx42\%$ for a maximal Kerr hole — roughly $8$ to $60$ times fusion. That is why a hole sipping under a millionth of a solar mass per year can outshine hundreds of thousands of Suns, and why accretion powers the most luminous persistent objects in the universe.

Do stellar-mass black holes evaporate by Hawking radiation?

In principle yes, in practice never on any relevant timescale. Hawking radiation gives a hole a temperature $T_H\propto1/M$ , about $6\times10^{-8}\,\mathrm{K}$ for a solar-mass hole — far colder than the $2.7\,\mathrm{K}$ cosmic microwave background bathing it. A stellar hole therefore absorbs vastly more from the CMB than it radiates, so it grows rather than shrinks, and cannot begin to evaporate until the universe cools below its temperature, aeons hence. The evaporation lifetime of a solar-mass hole is around $10^{67}$ years, dwarfing the age of the universe. Profound theory, but for stellar holes entirely academic.

What sets the boundary between a neutron star and a black hole?

The maximum mass a neutron star can support — the Tolman–Oppenheimer–Volkoff limit — found by integrating the relativistic hydrostatic-equilibrium equation with a realistic dense-matter equation of state. Its value, around $2$ – $3\,M_\odot$ , is uncertain because the equation of state above nuclear density is not fully known: stiffer matter supports more mass. The heaviest well-measured neutron stars sit near $2\,M_\odot$ , and tidal constraints from GW170817 favor the lower end. A core below the TOV limit settles as a neutron star; above it, no pressure can win and it becomes a black hole. Pinning down this boundary is a central open problem linking nuclear physics and gravitational-wave astronomy.

Does an infalling astronaut get destroyed at the horizon?

For a stellar-mass hole the danger comes before the horizon, not at it. The horizon is not a physical surface; crossing it feels locally like nothing. What destroys an infalling body is the tidal force — the difference in gravity between head and feet — which for a small hole is already lethal outside the horizon, stretching the astronaut into a thin stream ("spaghettification"). Counterintuitively, the larger the hole, the gentler the horizon: tidal forces at $R_S$ scale as $1/M^2$ , so a supermassive hole can be crossed unharmed while a $10\,M_\odot$ hole shreds you far outside it. A distant observer, meanwhile, never sees the crossing — infalling light is redshifted without limit, so the astronaut appears to freeze and fade, the "frozen star" of Oppenheimer and Snyder's 1939 paper.