Black Hole Ringdown

The bell that is spacetime

Strike a bell and it rings: it deforms, then radiates that deformation away as sound at a few sharp tones that fade over a second or two. A black hole does the same thing — except the "bell" is the geometry of spacetime, the "sound" is gravitational waves, and the whole performance is over in milliseconds. When two black holes spiral together and their event horizons fuse, what is left for an instant is a single but badly distorted black hole. Its horizon is squashed and lopsided, lurching as it forms. Such a shape is not allowed to persist. The remnant promptly relaxes into the only configuration nature permits — a smooth, axisymmetric Kerr black hole — and the energy of the deformation escapes as a brief, dying chirp of gravitational radiation. That escape is the ringdown.

The reason the lumps cannot stay is the deepest fact about black holes: the no-hair theorem. A stationary, isolated black hole is fully described by just three numbers — mass, angular momentum, and electric charge (which is negligible for astrophysical holes). Everything else about whatever fell in — the shapes, the higher multipole moments, the memory of two separate objects — must be erased. The ringdown is that erasure caught in the act. Each "lump" is a higher multipole moment of the horizon, and each gets radiated away on its own timescale until the slate is wiped to bare mass and spin.

Inspiral, merger, ringdown

A black-hole coalescence has three movements, and the ringdown is the finale. During the inspiral, the two holes orbit each other, losing energy to gravitational waves and spiralling inward over millions of years, the frequency and amplitude both sweeping upward in the famous "chirp." The merger is the violent instant the horizons touch and combine — the most nonlinear, highest-amplitude part of the signal, where no clean analytic formula applies and only numerical relativity reproduces the waveform. The ringdown is the aftermath: the merged horizon settling, with a waveform that is, beautifully, analytic again — a simple sum of damped sinusoids.

This clean separation is what makes the ringdown so prized. The inspiral encodes the two initial masses and spins; the merger is a numerical-relativity puzzle; but the ringdown is governed by linear black-hole perturbation theory, where the equations are exactly solvable in the form of quasinormal modes. The waveform of the ringing is therefore predicted to exquisite precision from a single pair of numbers — the final mass and spin — making it the cleanest handle on the remnant.

The physics: damped quasinormal modes

Perturb the spacetime around a Kerr black hole — by, say, dumping in the asymmetry left over from a merger — and the curvature does not oscillate freely the way a struck tuning fork does in air. The horizon is a perfect one-way sink: any energy that crosses it is gone, and the oscillation is also leaking energy outward to infinity as gravitational waves. The result is an oscillation that is intrinsically damped. The natural modes of such a leaky system are called quasinormal modes (QNMs), and their frequencies are complex numbers:

ω_n = 2π f_n + i / τ_n        (complex QNM frequency)

h(t) = Σ_n  A_n e^(−t/τ_n) cos(2π f_n t + φ_n)     for t > t_peak

The real part 2πf_n sets the oscillation frequency (the "pitch"); the imaginary part 1/τ_n sets the exponential decay (the damping). Each mode is labelled by three indices: the angular harmonics ℓ and m (like the harmonics of a vibrating sphere) and an overtone number n = 0, 1, 2, ... that counts increasingly fast-decaying tones. The loudest by far is the fundamental ℓ = m = 2, n = 0 mode. For a non-spinning (Schwarzschild) hole the fundamental sits at

f_220 ≈ 0.3737 × c³ / (2π G M)   ≈ 12.0 kHz × (M☉ / M)
τ_220 = 1 / |Im ω| ≈ 11 × (GM/c³)   (a handful of light-crossing times)

Spin shifts both: as the dimensionless spin a = cJ/(GM²) rises from 0 toward its extremal value of 1, the fundamental frequency increases by roughly 50% and the damping time lengthens (an extremal hole rings nearly forever). Crucially, those shifts are fixed functions of a alone — which is exactly why two measured tones over-determine the system and test the theory.

The key numbers

The single most useful fact is that everything in the ringdown scales with the remnant mass. The light-crossing time of a black hole is GM/c³, and both the period and the damping time of the fundamental mode are small multiples of it:

GM/c³ = 4.93 µs × (M / M☉)

For M = 62 M☉ (GW150914 remnant):
  GM/c³ ≈ 0.31 ms
  f_220 ≈ 250 Hz       (spin a ≈ 0.67 raises it above the 194 Hz Schwarzschild value)
  τ_220 ≈ 4 ms         (≈ 1 oscillation period × Q, quality factor Q ≈ 3–4)
  rings for ≈ 10 ms before fading to the noise

So a stellar-mass remnant rings in the hundreds of hertz — within the audible range, which is why a frequency-shifted LIGO recording genuinely sounds like a "thud-and-chirp." A supermassive black hole, with mass a million to a billion times larger, rings a million to a billion times lower and longer:

Remnant mass	Fundamental f₂₂₀	Damping τ	Detector band	Example
3 M☉	~4 kHz	~0.2 ms	High-freq edge of LIGO	GW170817 (if it formed a BH)
62 M☉	~250 Hz	~4 ms	LIGO / Virgo sweet spot	GW150914
140 M☉	~110 Hz	~9 ms	LIGO low band	GW190521 (IMBH)
10⁴ M☉	~1.5 Hz	~0.6 s	Deci-Hertz (gap)	IMBH merger
10⁶ M☉	~0.015 Hz	~1 min	LISA (space)	Sgr A*-class
10⁹ M☉	~1.5 × 10⁻⁵ Hz	~12 h	Below LISA / PTA edge	Quasar SMBH merger

Energetically, the ringdown is the quiet movement. An equal-mass binary radiates roughly 5% of its total mass-energy across the whole coalescence; only a fraction of that — typically of order 1% of the total, or a few tenths of a solar mass — comes out during the ringdown itself. The bulk is emitted at merger. For GW150914 the full event radiated about 3 M☉ as gravitational waves, peaking at a luminosity near 3.6 × 10⁴⁹ W — for that instant brighter than the combined light of every star in the observable universe.

Black-hole spectroscopy and the no-hair test

Here is the payoff. For a Kerr black hole, the full table of QNM frequencies and damping times is a fixed function of two parameters — the final mass M and spin a. The fundamental mode alone gives you (M, a). But every other mode is then completely predicted. So the test is a consistency check:

1. Measure the dominant ℓ=m=2, n=0 tone → infer the final (M, a).
2. Predict where the first overtone (n=1) or the next angular harmonic (ℓ=m=3) must lie if the object is a Kerr black hole.
3. Independently measure that second tone in the data.
4. If the two measurements of (M, a) agree, general relativity and the no-hair theorem pass. If they disagree, the remnant is not a Kerr black hole.

This is the gravitational-wave analogue of atomic spectroscopy — the spacing of a black hole's tones is its "spectral fingerprint," and reading two lines instead of one is what turns a detection into a test of strong-field gravity. It is one of the only ways to probe whether the object that formed is really the smooth horizon GR predicts, or something more exotic — a horizonless "ECO" (exotic compact object), a boson star, or a fuzzball — which would imprint subtle "echoes" in the late ringdown.

Worked example: the GW150914 ringdown

Take the first-ever detection, GW150914 (14 September 2015). The two inspiralling holes were about 36 and 29 M☉; they merged into a remnant of roughly 62 M☉ spinning at a ≈ 0.67, with about 3 M☉ radiated away. Let's reconstruct the ringdown tone.

First the characteristic time:

GM/c³ = 4.93 µs × (M/M☉) = 4.93 µs × 62 ≈ 0.306 ms

For a Schwarzschild hole of this mass the fundamental frequency would be

f = 12.0 kHz × (M☉/M) = 12.0 kHz / 62 ≈ 194 Hz

But the remnant spins at a ≈ 0.67, which raises the ℓ=m=2 fundamental by about 25–30%, landing it near

f_220 ≈ 250 Hz

The damping time follows from the mode's quality factor Q ≈ 3–4 at this spin, via τ = Q / (π f):

τ_220 ≈ Q / (π f) ≈ 3.3 / (π × 250 Hz) ≈ 4.2 ms

So the remnant rang at about a quarter of a kilohertz and fell silent in roughly 10 ms — only two or three full oscillations. That is precisely what LIGO recorded in the final few cycles of the chirp, and fitting a single damped sinusoid to that tail returned a mass and spin consistent with the values inferred from the inspiral. The two independent estimates agreeing was the first direct, observational confirmation that the object that formed was a Kerr black hole.

Discovery, missions, and people

The mathematics came decades before the data. C. V. Vishveshwara showed in 1970 that a Gaussian pulse scattered off a Schwarzschild black hole rings down in a characteristic damped oscillation. William Press named these quasinormal modes in 1971, and Saul Teukolsky's 1972–73 master equation made the Kerr case tractable. Subrahmanyan Chandrasekhar and Steven Detweiler tabulated the Schwarzschild mode frequencies in 1975. Throughout the 1990s and 2000s the frequencies were computed to high precision (Leaver's continued-fraction method, 1985) and Kip Thorne and others proposed black-hole spectroscopy as a test of the no-hair theorem.

The observational era opened on 14 September 2015, when the twin LIGO detectors in Hanford, Washington and Livingston, Louisiana recorded GW150914 — the merger of two stellar-mass black holes 1.3 billion light-years away. The discovery, announced in February 2016, won Rainer Weiss, Barry Barish, and Kip Thorne the 2017 Nobel Prize in Physics. The ringdown was visible in the waveform's tail. Since then the LIGO–Virgo–KAGRA network has logged the post-merger ringing of dozens of mergers; GW190521 (2020) produced an intermediate-mass remnant near 140 M☉ that rang especially low. Future facilities — the underground Einstein Telescope, the US Cosmic Explorer, and the space-based LISA mission (launch ~2035, targeting supermassive-black-hole ringdowns at millihertz frequencies) — are designed in large part to resolve multiple ringdown tones and make black-hole spectroscopy routine.

Related phenomena and variants

• Overtones (n ≥ 1). Beyond the fundamental, each (ℓ, m) has a tower of overtones that decay even faster. They are loudest right at the merger peak and have been controversially used to push spectroscopy to the earliest possible time, where the signal is strongest but linear perturbation theory is least trustworthy.
• Angular harmonics (ℓ=m=3, ℓ=m=4). Higher angular modes, excited more strongly in unequal-mass and highly inclined mergers. The cleanest independent second tone for a no-hair test, since it differs from the fundamental in both frequency and shape.
• Gravitational-wave echoes. If the remnant were a horizonless exotic compact object rather than a true black hole, a partly reflective "surface" near where the horizon should be would send delayed pulses back out — "echoes" in the late ringdown. Searches have so far found no statistically compelling evidence.
• Tidal ringing of neutron-star remnants. A neutron-star merger can produce a hypermassive neutron star that oscillates at kilohertz frequencies before collapsing — a different, matter-driven kind of post-merger ringing, distinct from the vacuum black-hole ringdown.
• Superradiance and the bomb. An almost-extremal Kerr hole can amplify rather than damp certain modes by extracting rotational energy, the seed of "black-hole bomb" instabilities and ultralight-boson clouds — the flip side of the usual damping.

Common misconceptions and subtleties

• "Something physical is vibrating." No matter is oscillating. It is the curvature of spacetime just outside the horizon — the gravitational field — that rings. The black hole has no surface to wobble.
• "Quasinormal modes are normal modes." They are not. A normal mode (like a guitar string) conserves energy and has a purely real frequency. QNMs leak energy down the horizon and out to infinity, so their frequencies are complex and the modes are not even orthogonal in the usual sense. This is why the math is genuinely harder than for a bell.
• "The ringdown carries most of the energy." The opposite — merger dominates the energy budget. The ringdown is the low-amplitude tail; it matters because it is clean, not because it is loud.
• "You can read off the mass from the frequency directly." Only if you know the spin, because frequency depends on both M and a. A single tone gives a degenerate M–a track; you need the damping time (or a second mode) to break the degeneracy.
• "Ringdown begins at a sharp, well-defined moment." The transition from the nonlinear merger to the linear ringing is fuzzy. Choosing the start time too early contaminates the fit with merger nonlinearity and overtones; too late, and the signal has decayed into the noise. This start-time ambiguity is the central methodological headache of black-hole spectroscopy.
• "It is the same thing as the inspiral chirp." The inspiral frequency rises (a chirp); the ringdown frequency is essentially constant while the amplitude decays. They are physically distinct regimes governed by different equations.