Black Hole Spin Measurement

A black hole is the simplest macroscopic object in physics: strip away everything that fell in and only two numbers survive, the mass $M$ and the angular momentum $J$ . Charge leaks away astrophysically, so the entire spacetime outside a real black hole is the Kerr geometry fixed by $M$ and $J$ alone — the "no-hair" statement made quantitative. Packaging $J$ into the dimensionless spin $a_* = cJ/(GM^2)$ , bounded by $0 \le |a_*| \le 1$ , means that measuring $a_*$ is measuring one of only two intrinsic properties a black hole is allowed to have. It is also a clean strong-field test of general relativity, and a fossil record: a stellar-mass hole's spin was set at collapse, a supermassive hole's by billions of years of accretion and mergers. Spin is never read off directly. It is inferred from the one place where rotation leaves a sharp, calculable imprint — the innermost stable circular orbit (ISCO). General relativity predicts that circular geodesics become unstable inside a critical radius $r_{\mathrm{ISCO}}$ that depends monotonically on $a_*$ , shrinking from $6\,GM/c^2$ for a non-rotating hole to $1\,GM/c^2$ at extremal prograde spin. An accretion disk is assumed to terminate at the ISCO, so any observable that pins down the inner-disk radius becomes a spin gauge. Two techniques dominate: continuum fitting, which reads the emitting area of the thermal disk, and relativistic iron-line spectroscopy, which reads the gravitational-plus-Doppler distortion of the 6.4 keV Fe Kα fluorescence line. This article is the flagship synthesis — it derives the ISCO-to-spin map, works the efficiency and inversion numbers, and shows where the census stands today.

Key takeaways

Black-hole spin is the dimensionless $a_* = cJ/(GM^2)$ , with $0 \le |a_*| \le 1$ ; charge is negligible, so $M$ and $a_*$ fix the entire exterior Kerr spacetime.
General relativity fixes $r_{\mathrm{ISCO}}(a_*)$ : it falls monotonically from $6\,GM/c^2$ (Schwarzschild), through $2.32\,GM/c^2$ at $a_*=0.9$ , to $1\,GM/c^2$ (extremal prograde), and rises to $9\,GM/c^2$ for extremal retrograde.
Continuum fitting reads $r_{\mathrm{ISCO}}$ from the disk's thermal emitting area and needs independent mass $M$ , distance $D$ , and inclination $i$ — its systematics dominate the error, not photon counts.
Iron-line modeling reads $r_{\mathrm{ISCO}}/M$ from the redshifted, skewed profile of the 6.4 keV Fe Kα line — a dimensionless morphology, so it needs neither $D$ nor a precise $M$ .
Spin sets radiative efficiency $\eta = 1 - E_{\mathrm{ISCO}}$ : about 5.7% for Schwarzschild, rising toward 42% for extremal Kerr — the most efficient steady energy source known short of annihilation.
Both methods stand on the assumption that the disk truncates exactly at the ISCO and aligns with the spin; break that (hard states, warps, misalignment) and the inferred $a_*$ is biased high.
Key idea: general relativity uniquely fixes the ISCO radius as a monotonic function of $a_*$ ; because an accretion disk is assumed to truncate at the ISCO, any observable that measures the inner-disk radius — the thermal continuum's emitting area or the relativistically broadened iron line — measures the black-hole spin.

Where this lives today

Spin measurement has matured into a quantitative census spanning nine orders of magnitude in black-hole mass. For stellar-mass holes in X-ray binaries — Cygnus X-1, GRS 1915+105, LMC X-1 — continuum fitting in thermally dominant soft states, anchored by dynamically measured masses and Gaia or radio parallaxes, delivers spins ranging from modest to near-maximal; because these systems are short-lived, the result probes the natal angular momentum of the collapsing stellar core and, increasingly, is cross-checked against the effective spins that LIGO–Virgo–KAGRA infer from merging binary black holes. For supermassive holes in active galactic nuclei, the thermal peak lies in the unobservable far ultraviolet, so relativistic Fe Kα reflection is the primary tool: broad, skewed lines in Seyferts such as MCG−6-30-15 imply disks reaching small radii and hence high spin, feeding models of coherent versus chaotic accretion over cosmic time. The frontier is now multi-messenger and multi-technique — NASA's IXPE adds X-ray polarimetry that constrains inclination and geometry to break single-method degeneracies; XRISM brings microcalorimeter spectroscopy that resolves line profiles; and the Event Horizon Telescope, having imaged the shadows of M87 and Sgr A, is pushing toward direct horizon-scale constraints on spin. The measured spin distribution matters beyond the holes themselves: via the Blandford–Znajek mechanism the extractable jet power scales as $P_{\mathrm{jet}} \propto a_*^2 B^2$ , so spin sets the feedback that regulates galaxy growth, linking a strong-field general-relativistic quantity to the large-scale structure of the universe.

The physical picture

Picture the ISCO as the innermost ledge on which orbiting gas can safely stand. Outside it, a nudged orbit oscillates and returns; inside it, the same nudge grows without bound and the gas plunges through the horizon in a fraction of an orbit. A rotating hole drags spacetime around with it — frame dragging — and for gas circling in the same sense as the spin (prograde) this dragging props up stable orbits much closer in. So a faster-spinning hole has a smaller inner ledge, its disk reaches deeper into the potential well, and the light emitted there is hotter and more violently distorted by gravity. The inner glow is a thermometer and a speedometer at once.

Figure

Frame dragging shrinks the innermost stable orbit as spin rises. A non-rotating hole a_ = 0 keeps its ISCO out at 6 GM/c², far outside the 2 GM/c² horizon, so a wide gap of empty spacetime separates disk from hole. A near-extremal hole a_ → 1 drags orbits inward until the disk reaches 1.24 GM/c², almost onto the 1 GM/c² horizon.

The iron line is a natural relativistic ruler for that inner edge. In a laboratory the Fe Kα fluorescence line is a razor-sharp spike at 6.4 keV. Near a black hole it is torn open: gas sweeping toward us is Doppler-boosted to higher energy, gas receding is redshifted, and all of it climbs out of a deep gravitational well, losing energy to gravitational redshift on the way. The closer the emitting gas sits to the horizon — smaller ISCO, higher spin — the further the red wing stretches, from the sharp 6.4 keV rest energy down toward 3–4 keV. A key subtlety underlies both methods: they measure a radius in units of $GM/c^2$ , i.e. $r_{\mathrm{ISCO}}/M$ . Turning that dimensionless radius into a spin needs only general relativity; turning an observed flux or angular size into a physical radius needs $M$ , $D$ , and $i$ — which is why systematic errors in those priors, not photon statistics, usually set the spin uncertainty. The second figure is the curve every spin measurement inverts.

Figure

The general-relativistic map from spin to orbit. The ISCO radius in units of GM/c² falls monotonically from 9 at extremal retrograde spin, through 6 at Schwarzschild a_ = 0, to 1 at extremal prograde spin. Every spin measurement inverts this one curve: e.g. a measured inner edge of 2.32 GM/c² reads off as a_ = 0.9.

The core equations

Everything begins with the definition of spin. Angular momentum $J$ carries dimensions, so it is repackaged into a pure number by dividing out the natural angular-momentum scale $GM^2/c$ that a black hole of mass $M$ can hold. The bound $|a_*| \le 1$ is the statement that a Kerr horizon exists; beyond it the metric describes a naked singularity, which cosmic censorship forbids.

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

where: $a_*$ — dimensionless (Kerr) spin parameter; $J$ — black-hole angular momentum; $M$ — black-hole mass; $G,\ c$ — gravitational constant and speed of light. Reading: $a_*=0$ is non-rotating (Schwarzschild), $a_*=1$ is a maximally spinning (extremal) Kerr hole, and the sign distinguishes prograde from retrograde orbits. The central result is the closed form for the ISCO radius as a function of spin, derived by Bardeen, Press and Teukolsky from the equatorial Kerr geodesics. It is the curve plotted above and the one every measurement inverts.

r_{\mathrm{ISCO}} = \frac{GM}{c^2}\Big[\,3 + Z_2 \mp \sqrt{(3-Z_1)(3+Z_1+2Z_2)}\,\Big] \qquad (2)

where: $Z_1 = 1+(1-a_*^2)^{1/3}\big[(1+a_*)^{1/3}+(1-a_*)^{1/3}\big]$ ; $Z_2 = \sqrt{3a_*^2 + Z_1^2}$ ; the upper (minus) sign is prograde, the lower (plus) sign retrograde. Reading: set $a_*=0$ and both $Z_1,Z_2 \to 3$ , giving exactly $6\,GM/c^2$ ; push $a_*\to 1$ prograde and the bracket collapses to $1$ . Circular orbits carry a specific energy (energy per unit rest mass) that decreases as the orbit tightens. Its value at the ISCO is the deepest the gas gets before plunging, and it is what sets the luminosity.

E_{\mathrm{ISCO}} = \frac{r^{3/2} - 2Mr^{1/2} \pm a_* M^{1/2}}{r^{3/4}\left(r^{3/2} - 3Mr^{1/2} \pm 2a_* M^{1/2}\right)^{1/2}} \quad (G=c=1) \qquad (3)

where: $E_{\mathrm{ISCO}}$ — specific orbital energy at the ISCO, in units of $c^2$ ; $r$ — Boyer–Lindquist radius evaluated at $r_{\mathrm{ISCO}}$ ; upper sign prograde. Reading: at $a_*=0$ , $r=6M$ this gives $\sqrt{8/9}\approx0.943$ ; at extremal prograde it drops to $1/\sqrt{3}\approx0.577$ . The radiative efficiency is the binding energy left behind at the ISCO. Gas arrives from infinity with unit specific energy, spirals down radiating the difference, and carries $E_{\mathrm{ISCO}}$ across the horizon.

\eta = 1 - E_{\mathrm{ISCO}} \qquad (4)

where: $\eta$ — fraction of accreted rest-mass energy radiated before plunge. Reading: $\eta$ runs from $\approx 0.057$ (Schwarzschild) up to $\approx 0.42$ (extremal Kerr) — spin makes accretion up to seven times more luminous per gram. Continuum fitting rests on the Novikov–Thorne thin-disk temperature profile, the relativistic descendant of the Shakura–Sunyaev disk. Each annulus radiates as a blackbody, and the run of temperature with radius fixes the emitting area.

F(R) = \frac{3GM\dot M}{8\pi R^3}\left(1 - \sqrt{\frac{R_{\mathrm{in}}}{R}}\right), \qquad T_{\mathrm{eff}}(R) = \left[\frac{F(R)}{\sigma}\right]^{1/4} \qquad (5)

where: $F(R)$ — local flux from the annulus at radius $R$ ; $\dot M$ — mass accretion rate; $R_{\mathrm{in}} = r_{\mathrm{ISCO}}$ — inner disk edge; $\sigma$ — Stefan–Boltzmann constant. Reading: the flux vanishes at $R_{\mathrm{in}}$ (zero-torque inner boundary), peaks just outside it, and falls as $R^{-3}$ far out — so most thermal light comes from a few $r_{\mathrm{ISCO}}$ . Integrating that profile, the disk luminosity is the emitting area times a characteristic $T_{\mathrm{in}}^4$ — a Stefan–Boltzmann statement that isolates $R_{\mathrm{in}}$ , hence the spin.

L_{\mathrm{disk}} = 4\pi \int_{R_{\mathrm{in}}}^{\infty} F(R)\,R\,\mathrm{d}R \;\propto\; R_{\mathrm{in}}^2\,T_{\mathrm{in}}^4, \qquad L_{\mathrm{disk}} = \frac{4\pi D^2 F_{\mathrm{obs}}}{\cos i} \qquad (6)

where: $L_{\mathrm{disk}}$ — thermal disk luminosity; $T_{\mathrm{in}}$ — characteristic inner temperature from the spectral rollover; $D$ — distance; $i$ — inclination. Reading: measure $L_{\mathrm{disk}}$ and $T_{\mathrm{in}}$ from the spectrum, solve for the area $\propto R_{\mathrm{in}}^2$ , and with $M$ known convert $R_{\mathrm{in}} = r_{\mathrm{ISCO}}(a_*)\,GM/c^2$ into $a_*$ . Finally the iron line encodes spin through a schematic combination of gravitational and Doppler shifts, both of which sharpen as the emitting radius shrinks.

\frac{E_{\mathrm{obs}}}{E_0} \approx \frac{\sqrt{1 - \tfrac{2GM}{c^2 R}}}{1 \pm \tfrac{v}{c}\sin i}, \qquad E_0 = 6.4\ \mathrm{keV} \qquad (7)

where: $E_0$ — rest-frame Fe Kα energy; $E_{\mathrm{obs}}$ — observed energy after redshifts; $v$ — orbital speed of the emitting gas; $i$ — inclination. Reading: the deepest gravitational redshift comes from the smallest $R = r_{\mathrm{ISCO}}$ , so the extent of the red wing directly gauges the spin — and it does so through the dimensionless $2GM/c^2R = 2/(r/M)$ , needing no distance.

Derivation AThe ISCO from marginal stability of circular orbits

The goal is to show where equation (2) comes from — not to reproduce the full algebra, but to expose the single physical condition behind it. A circular orbit needs the effective radial potential to have an extremum; a stable one needs that extremum to be a minimum. The ISCO is the knife-edge where the minimum and the neighbouring maximum merge into an inflection, so that the last stable orbit and the first unstable one coincide. Everything about the spin dependence of $r_{\mathrm{ISCO}}$ flows from imposing three conditions on one function of radius.

Step 1. Reduce equatorial Kerr geodesics to a one-dimensional problem: the conserved energy $E$ and axial angular momentum $L$ let the radial equation be written as $\dot r^2 = E^2 - V_{\mathrm{eff}}(r; E, L, a_*)$ , a particle rolling in an effective potential. Why: turning a curved-spacetime orbit into a rolling-marble picture is what makes "stable versus unstable" a statement about the shape of a single curve.

Step 2. Impose circularity: a circular orbit sits at a turning point that does not move, so $V_{\mathrm{eff}} = E^2$ and $\partial_r V_{\mathrm{eff}} = 0$ simultaneously. These two equations solve for the circular-orbit functions $E(r)$ and $L(r)$ at every radius. Why: this is what makes energy and angular momentum functions of radius alone — the input to both the ISCO and the efficiency.

Step 3. Impose marginal stability: demand additionally $\partial_r^2 V_{\mathrm{eff}} = 0$ , the vanishing curvature that marks an inflection. Solving all three conditions together yields the closed form (2); it is a cubic-in-disguise whose real root is the bracket $3 + Z_2 \mp \sqrt{(3-Z_1)(3+Z_1+2Z_2)}$ . Why: the third condition is the whole content of "innermost stable" — drop it and you get any circular orbit; keep it and you get the unique boundary between stable and plunging.

Step 4. Check the anchors: $a_*=0$ gives $Z_1=Z_2=3$ and $r_{\mathrm{ISCO}}=6\,GM/c^2$ ; prograde $a_*\to1$ gives $Z_1=1$ , $Z_2=2$ and $r_{\mathrm{ISCO}}\to GM/c^2$ (coincident with the horizon in Boyer–Lindquist radius); retrograde $a_*\to1$ gives $r_{\mathrm{ISCO}}\to9\,GM/c^2$ . Why: these three exact values are the calibration points of the entire spin-measurement enterprise, and any code must reproduce them before it can be trusted on data.

Go deeper · Derivation B · 1 minRadiative efficiency of a spinning disk

Derivation BRadiative efficiency of a spinning disk

The goal is to compute how much rest-mass energy a black hole liberates as light, and to show that spin is the knob that controls it. The key idea is that a steady thin disk radiates exactly the binding energy the gas has lost by the time it reaches the ISCO: it starts at rest at infinity with specific energy one, spirals down through a sequence of near-circular orbits radiating the energy difference, and finally falls from the ISCO through the horizon carrying the leftover $E_{\mathrm{ISCO}}$ with it. So the efficiency is simply $\eta = 1 - E_{\mathrm{ISCO}}$ , and the spin enters entirely through how deep the ISCO sits.

Step 1. Fix the non-rotating baseline. At $a_*=0$ the ISCO is at $r=6\,GM/c^2$ ; substituting into (3) gives $E_{\mathrm{ISCO}} = \sqrt{8/9} \approx 0.9428$ . Why: the Schwarzschild value is the reference every spin efficiency is quoted against, and it already dwarfs other cosmic engines.

Step 2. Read off the Schwarzschild efficiency: $\eta = 1 - \sqrt{8/9} \approx 0.057$ , so 5.7% of the rest mass is radiated. Compare hydrogen fusion, which liberates only $\approx 0.7\%$ . Why: even a non-spinning black hole is nearly an order of magnitude more efficient than the nuclear burning that powers stars — accretion is the brightest steady process in the universe.

Step 3. Take the extremal prograde limit. As $a_*\to1$ the ISCO shrinks to $r\to GM/c^2$ and the specific energy drops to $E_{\mathrm{ISCO}}\to 1/\sqrt{3} \approx 0.577$ , so $\eta \to 1 - 1/\sqrt{3} \approx 0.42$ . Why: a maximally spinning hole converts up to 42% of infalling mass to radiation, second only to matter–antimatter annihilation — this is why the brightest quasars are believed to host fast-spinning holes.

Step 4. Note the retrograde branch: a counter-rotating disk has its ISCO pushed out to $9\,GM/c^2$ , where $E_{\mathrm{ISCO}}\approx0.962$ and $\eta\approx0.038$ . Why: the same hole is more than ten times more efficient prograde than retrograde, so orbit sense — not just spin magnitude — matters for the energetics, and a chaotically fed hole averages down toward low efficiency.

Go deeper · Derivation C · 2 minInverting the continuum to a spin

Derivation CInverting the continuum to a spin

The goal is to trace, quantity by quantity, how a measured thermal spectrum becomes a number for $a_*$ , and to see why the answer is only as good as the mass, distance, and inclination that feed it. The physical content is Stefan–Boltzmann: a blackbody-like inner disk of projected area proportional to $R_{\mathrm{in}}^2\cos i$ radiating at temperature $T_{\mathrm{in}}$ has a luminosity $\propto R_{\mathrm{in}}^2 T_{\mathrm{in}}^4$ , so if two of luminosity, temperature, and area are known the third follows — and the area hands us $R_{\mathrm{in}} = r_{\mathrm{ISCO}}$ .

Step 1. Deproject the flux. The observed thermal flux $F_{\mathrm{obs}}$ becomes a luminosity through $L_{\mathrm{disk}} = 4\pi D^2 F_{\mathrm{obs}}/\cos i$ , which demands the distance $D$ and the inclination $i$ independently, from parallax, binary dynamics, or radio jets. Why: this is the step that injects the dominant systematic — an error in $D$ enters $L$ as $D^2$ and propagates straight into $R_{\mathrm{in}}$ .

Step 2. Extract the temperature. Fit the thermal continuum's rollover to get a colour temperature $T_{\mathrm{col}}$ , then apply a spectral-hardening factor $f = T_{\mathrm{col}}/T_{\mathrm{eff}} \approx 1.5$ – $1.7$ from disk-atmosphere models, since electron scattering makes the disk radiate hotter than a true blackbody. Why: skipping the hardening correction would overestimate the temperature, shrink the inferred area, and bias the spin upward — the correction is not optional.

Step 3. Solve for the inner radius. Stefan–Boltzmann gives $R_{\mathrm{in}} \propto L_{\mathrm{disk}}^{1/2}\,T_{\mathrm{in}}^{-2}$ , corrected by a relativistic factor $g$ that folds in light bending and Doppler boosting of the inner-disk photons. Why: the same general relativity that defines the ISCO also warps the photons leaving it, so the naive flat-space area must be relativistically corrected before it means anything.

Step 4. Convert to a spin. With the dynamical mass $M$ from the binary orbit, form $r_{\mathrm{ISCO}}/M = R_{\mathrm{in}}\,c^2/(GM)$ and invert the monotonic curve (2) numerically for $a_*$ . The method is trustworthy only in disk-dominated soft states, where $R_{\mathrm{in}}$ stays constant as the luminosity varies — the observational signature that the disk really does reach the ISCO. Why: constancy of the inferred radius across a range of luminosities is the internal check that the ISCO assumption holds, distinguishing a genuine spin from a truncated-disk artifact.

Limiting cases

The ISCO radius, its energy, and the efficiency depend sharply on both spin magnitude and orbit sense:

Case	Result	Meaning
Schwarzschild, $a_*=0$	$r_{\mathrm{ISCO}} = 6\,GM/c^2$ , $\eta \approx 0.057$	non-rotating baseline; disk sits far from the $2\,GM/c^2$ horizon; 5.7% efficiency
Moderate prograde, $a_*=0.9$	$r_{\mathrm{ISCO}} = 2.32\,GM/c^2$ , $\eta \approx 0.156$	disk reaches well inside $6\,GM/c^2$ ; efficiency nearly tripled
Extremal prograde, $a_*\to1$	$r_{\mathrm{ISCO}} \to GM/c^2$ , $\eta \to 1-1/\sqrt3 \approx 0.42$	disk touches the horizon; efficiency approaches its physical ceiling
Extremal retrograde, $a_*\to-1$	$r_{\mathrm{ISCO}} \to 9\,GM/c^2$ , $\eta \approx 0.038$	counter-rotating disk pushed far out; efficiency falls below Schwarzschild
Truncated disk (hard state)	$r_{\mathrm{in}} > r_{\mathrm{ISCO}}$	ISCO assumption fails; inferred spin becomes an unreliable lower limit

Common trapThe ISCO is not automatically the inner edge of the disk

Both techniques assume the disk truncates exactly at the ISCO and lies in the spin equatorial plane. Neither is guaranteed. In hard or low-luminosity states the thin disk recedes and is replaced by a hot, geometrically thick, advection-dominated flow whose inner edge sits outside the ISCO — read naively, this masquerades as a large ISCO and a low spin. Conversely, if the model, hardening factor, or the $M$ – $D$ – $i$ priors are wrong, or if the disk is warped or misaligned (the Bardeen–Petterson effect tips the inner disk into the equatorial plane only within some radius), the inferred $a_*$ is biased — and in practice the bias usually runs high, toward artificially near-maximal spins. The defence is to work in thermally dominant soft states and to demand that the inferred inner radius stay constant as the luminosity changes.

Worked examples

1 · ISCO radius and efficiency at a_* = 0.9 (fully worked)

A stellar-mass black hole is measured to spin at $a_* = 0.9$ (prograde). Find its ISCO radius in units of $GM/c^2$ , the specific energy there, and the radiative efficiency. Then compare with the non-rotating case.

Step 1. Build the auxiliary quantities. With $a_* = 0.9$ : $(1-a_*^2)^{1/3} = 0.19^{1/3} = 0.575$ , $(1+a_*)^{1/3} = 1.9^{1/3} = 1.239$ , $(1-a_*)^{1/3} = 0.1^{1/3} = 0.464$ . Then $Z_1 = 1 + 0.575\,(1.239 + 0.464) = 1 + 0.575\times1.703 = \mathbf{1.979}$ , and $Z_2 = \sqrt{3(0.81) + 1.979^2} = \sqrt{2.43 + 3.916} = \mathbf{2.519}$ .

Step 2. Insert into the prograde ISCO formula (2): $r_{\mathrm{ISCO}} = 3 + 2.519 - \sqrt{(3-1.979)(3+1.979+2\times2.519)} = 5.519 - \sqrt{1.021\times10.017} = 5.519 - \sqrt{10.23} = 5.519 - 3.198 = \mathbf{2.32\ GM/c^2}$ . The disk reaches less than half the Schwarzschild inner radius.

Step 3. Evaluate the specific energy (3) at $r = 2.321$ , $a_*=0.9$ (units $G=c=M=1$ ): numerator $= r^{3/2} - 2r^{1/2} + a_* = 3.536 - 3.047 + 0.9 = 1.389$ ; the inner bracket $= r^{3/2} - 3r^{1/2} + 2a_* = 3.536 - 4.571 + 1.8 = 0.766$ ; denominator $= r^{3/4}\sqrt{0.766} = 1.880\times0.875 = 1.645$ . So $E_{\mathrm{ISCO}} = 1.389/1.645 = \mathbf{0.844}$ .

Step 4. Efficiency: $\eta = 1 - 0.844 = \mathbf{0.156}$ , i.e. 15.6% of the accreted rest mass is radiated — versus $1 - \sqrt{8/9} = 5.7\%$ for the same hole non-rotating. Spinning up from $a_*=0$ to $0.9$ nearly triples the light output per gram and moves the inner edge from $6$ to $2.32\,GM/c^2$ . $r_{\mathrm{ISCO}} = 2.32\,GM/c^2$ , $E_{\mathrm{ISCO}} = 0.844$ , $\eta = 15.6\%$ .

2 · From a measured inner radius to a spin (continuum fitting)

Continuum fitting of a soft-state X-ray binary returns a stable inner-disk radius $R_{\mathrm{in}} = 4.23\,GM/c^2$ across a factor of several in luminosity. What prograde spin does this imply, and what is the efficiency?

Step 1. The constancy of $R_{\mathrm{in}}$ with luminosity is the signal that the disk truly reaches the ISCO, so we set $r_{\mathrm{ISCO}} = 4.23\,GM/c^2$ and invert curve (2). Scanning the prograde branch (from Example 1 we know $a_*=0.9$ gives $2.32$ , and $a_*=0$ gives $6$ ), the value $4.23$ falls at $\mathbf{a_* \approx 0.5}$ — indeed the exact formula gives $r_{\mathrm{ISCO}}(0.5) = 4.233\,GM/c^2$ .

Step 2. Efficiency at $a_*=0.5$ : evaluating (3)–(4) gives $E_{\mathrm{ISCO}} = 0.918$ and $\eta = 1 - 0.918 = \mathbf{0.082}$ , about 8.2% — modestly above the Schwarzschild 5.7%.

Step 3. Sanity-check the sensitivity. Because the curve is steep at moderate spin, a 10% error in $R_{\mathrm{in}}$ (say $4.23 \to 4.65$ ) shifts the spin to $a_*\approx0.42$ — a several-percent change in $a_*$ . Since $R_{\mathrm{in}} \propto L^{1/2}T^{-2} \propto D$ , that 10% could come from a 10% distance error alone, illustrating why $D$ , $M$ , and $i$ dominate the error budget. $a_* \approx 0.5$ , $\eta \approx 8.2\%$ ; the spin error is set by the priors, not the photons.

3 · The red wing of the iron line (order-of-magnitude)

Estimate how far gravitational redshift alone pushes the 6.4 keV Fe Kα line for gas at the Schwarzschild ISCO ( $r=6\,GM/c^2$ ) versus a high-spin ISCO ( $r=2.32\,GM/c^2$ , i.e. $a_*=0.9$ ).

Step 1. Use the gravitational redshift factor $\sqrt{1 - 2GM/(c^2 R)} = \sqrt{1 - 2/(r/M)}$ from (7), holding Doppler aside. At $r=6$ : $\sqrt{1 - 2/6} = \sqrt{2/3} = 0.816$ , so $E_{\mathrm{obs}} = 6.4\times0.816 = \mathbf{5.2\ keV}$ .

Step 2. At the high-spin edge $r=2.32$ : $\sqrt{1 - 2/2.32} = \sqrt{0.138} = 0.372$ , so $E_{\mathrm{obs}} = 6.4\times0.372 = \mathbf{2.4\ keV}$ — the line's red wing stretches down past 3 keV.

Step 3. Interpret: a non-rotating hole can redshift the line only to $\approx5.2$ keV, but a fast-spinning one drags emitting gas to $r\approx2.3\,GM/c^2$ , extending the red wing to $\approx2.4$ keV before Doppler and transverse effects are even added. The extent of the red wing is thus a near-model-independent spin diagnostic — and, crucially, it depends only on $r/M$ , so no distance is required. Schwarzschild: down to $\approx5.2$ keV; $a_*=0.9$ : down to $\approx2.4$ keV.

Physics of the portrait

The whole edifice of spin measurement rests on one exact solution of Einstein's equations, and on the man who, against the prevailing pessimism of 1963, found it.

Origins · 1 minRoy Kerr and how this was discovered

Roy Kerr

1934– · the exact geometry of a rotating black hole

By the late 1950s general relativists had the Schwarzschild solution for a non-rotating mass and almost nothing else — the Einstein equations for a spinning body were thought too nonlinear to solve in closed form, and most physicists doubted a clean rotating analogue existed. Roy Kerr, a New Zealand mathematician then at the University of Texas, approached the problem through the algebraic classification of the Weyl tensor, seeking metrics of a special "algebraically degenerate" type. In 1963 he published a two-page paper giving the exact, stationary, axisymmetric vacuum solution now bearing his name — the geometry of every uncharged rotating black hole in the universe. Its importance was not immediately grasped; Subrahmanyan Chandrasekhar later called it the most important exact solution ever discovered, and remarked on the "shuddering" realisation that this abstract mathematics describes real objects scattered through the cosmos. The Kerr metric introduced frame dragging as an exact effect, the ergosphere from which rotational energy can be extracted (Penrose, 1969), and — through the geodesics worked out by Carter, Bardeen, Press and Teukolsky — the spin-dependent ISCO that every continuum-fitting and iron-line measurement now inverts. Kerr received the 2016 Crafoord Prize for the work; the solution stands as the foundation on which the entire enterprise of measuring a black hole's spin is built.

History

1918

— Lense and Thirring predict frame dragging by a rotating mass in general relativity

1963

— Kerr finds the exact metric for a rotating black hole

1969

— Penrose shows rotational energy can be extracted from the ergosphere

1972

— Bardeen, Press and Teukolsky derive the ISCO and orbital energetics of Kerr

1989

— Fabian and collaborators predict the relativistically broadened iron line

2019

— the Event Horizon Telescope resolves the shadow of M87 at horizon scale

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., Ch. 17 (general relativity and black holes) · Binney & Tremaine, Galactic Dynamics, 2nd ed. (orbits and dynamics context) · Bardeen, Press & Teukolsky, ApJ 178, 347 (1972) — Kerr geodesics and the ISCO formula · McClintock et al., Class. Quantum Grav. 28, 114009 (2011) — continuum-fitting review · Reynolds, ARA&A 59, 117 (2021) — iron-line / X-ray reflection spin measurements. The key anchor values from equations (2)–(4) are: $r_{\mathrm{ISCO}} = 6\,GM/c^2$ and $\eta = 1-\sqrt{8/9}\approx0.057$ at $a_*=0$ ; $r_{\mathrm{ISCO}} = 2.32\,GM/c^2$ and $\eta\approx0.156$ at $a_*=0.9$ ; $r_{\mathrm{ISCO}}\to GM/c^2$ and $\eta\to1-1/\sqrt3\approx0.42$ at extremal prograde; and $r_{\mathrm{ISCO}}\to9\,GM/c^2$ , $\eta\approx0.038$ at extremal retrograde. The horizon relation $r_H = (GM/c^2)(1+\sqrt{1-a_*^2})$ and the spin definition $a_* = cJ/GM^2$ are exact results of the Kerr metric. The thin-disk flux law (5) is the Shakura–Sunyaev / Novikov–Thorne result.

Prerequisites. Neutron Stars · Tolman–Oppenheimer–Volkoff Equation · Equation of State of Dense Matter. Leads to. Kerr Metric · Accretion Disk Theory · Event Horizon Telescope Imaging.

Ask the tutor: ask me to invert the ISCO curve for any spin, work a continuum or iron-line measurement end to end, or show why frame dragging makes a fast hole the most efficient engine in the universe.

Why can't we just watch a black hole spin and read off its angular momentum?

Because a black hole has no surface to track. Everything that fell in is hidden behind the horizon, and the no-hair theorems guarantee the exterior spacetime is the Kerr geometry set by $M$ and $J$ alone — there are no features to clock. So spin is inferred indirectly, from how the Kerr geometry shapes the orbits of the accreting gas just outside the horizon. The sharpest lever is the ISCO radius, which general relativity ties to $a_*$ by a fixed, monotonic function. Measure where the disk's inner edge sits, and you have measured the spin.

Why does a higher spin push the ISCO inward for prograde orbits but outward for retrograde ones?

Frame dragging. A rotating hole drags the local inertial frames around with it in the direction of spin. Gas orbiting the same way (prograde) is partly "carried" by this dragging, so it needs less support from its own orbital motion and can remain on a stable circular orbit much closer in — down to $GM/c^2$ at extremal spin. Gas orbiting against the drag (retrograde) is fighting the swirl of spacetime and loses stability farther out, so its ISCO grows to $9\,GM/c^2$ . Same hole, opposite orbit sense, a factor of nine difference in inner radius.

Continuum fitting and iron lines both use the ISCO — so why keep two methods?

Because they fail differently and apply in different regimes. Continuum fitting needs the thermal disk peak in your band (great for stellar-mass holes, useless for supermassive ones whose peak is in the far UV) and independent $M$ , $D$ , $i$ . Iron-line modeling reads a dimensionless line shape, so it works across the whole mass range and needs no distance, but it depends on the geometry of the hard-X-ray "corona" that lights the disk. Where both apply — a handful of stellar-mass sources — their agreement is a genuine cross-check on the Kerr paradigm itself. Disagreement flags a broken assumption in one of them.

Is efficiency really as high as 42 percent? That seems impossibly large.

It is real, and it is why accreting black holes are the brightest steady engines in the universe. The 42% is the fraction of infalling rest mass radiated before the gas crosses the ISCO of an extremal prograde hole, computed straight from $\eta = 1 - E_{\mathrm{ISCO}}$ with $E_{\mathrm{ISCO}} = 1/\sqrt3$ . For comparison, hydrogen fusion converts only about 0.7%. The energy comes from gravitational binding, not from any new physics — deep in the Kerr potential well the gas has shed almost half its rest energy as light. Only matter–antimatter annihilation does better. Real disks reach the theoretical ceiling only for near-extremal spin; a typical measured $\eta$ is 10–20%.

What actually sets the error on a spin measurement?

Almost never the photon statistics. For continuum fitting the budget is dominated by the priors — mass, distance, and inclination — because $R_{\mathrm{in}}\propto L^{1/2}\propto D$ , so a 10% distance error is a 10% radius error and a several-percent spin error, and mass and inclination enter comparably. On top of that sit modeling systematics: the spectral-hardening factor, the assumption that the disk truncates exactly at the ISCO, and the relativistic ray-tracing corrections. For iron lines the analogous systematics are the assumed disk emissivity profile and the geometry of the illuminating corona. This is why the field emphasises soft-state selection, constancy of $R_{\mathrm{in}}$ with luminosity, and cross-checks between independent methods.

How do these spins connect to the black holes LIGO detects merging?

They are two windows on the same population. Electromagnetic methods measure the spin of a single accreting hole from its disk; gravitational-wave detectors measure the effective spin of a merging binary from the waveform's phase. Comparing the two distributions tests formation channels: stellar-mass X-ray binary spins reflect natal collapse plus any accretion spin-up, while binary-merger spins encode whether the pair formed together (aligned spins) or captured each other dynamically (random orientations). The tension and agreement between the electromagnetic and gravitational-wave spin distributions is now one of the most active questions in compact-object astrophysics, and it all rests on the same underlying quantity, $a_*$ .