Gravitational Collapse and Black-Hole Dynamics

Black holes are not merely stationary metrics. They form from finite-amplitude initial data, acquire and shed mass and angular momentum, develop dynamical horizons and interiors, collide, merge, recoil, and—under controlled hypotheses—approach stationary end states. This volume develops that nonlinear history. Its central question is how the Einstein equations turn initial data into a black hole while preserving a clear distinction between geometry, matter modeling, numerical evidence, and theorem-level conclusions.

The emphasis is classical four-dimensional general relativity. Exact spherical models and rigorous small-data theorems provide anchors; numerical spacetimes supply access to fully three-dimensional mergers and rotating collapse; carefully dated comparisons mark questions that remain conditional or unsettled. When a higher-dimensional topology theorem or a five-dimensional holographic model is useful, the departure from the default is labeled and kept to that bounded comparison. The goal is not to attach one preferred picture to every dynamical black hole, but to learn which statement is invariant, which depends on a foliation or boundary condition, and which has been established only in a restricted model.

Helpful background. Familiarity with Black-Hole Geometry and Mechanics, Black-Hole Perturbation Theory, and Formulations of General Relativity makes the main route faster. Each chapter also points to the shortest specific repair path when it needs a sharper prerequisite.

From initial data to a remnant

The nonlinear story begins before a horizon exists. Initial data must satisfy the Einstein constraint equations; the evolution then decides whether matter settles into a material object, whether matter or gravitational radiation disperses, or whether energy concentrates strongly enough to create trapped surfaces. In spherical symmetry, the alternatives can be seen with unusual clarity. Oppenheimer–Snyder dust gives an exact matching problem, Lemaître–Tolman–Bondi models expose radial inhomogeneity, scalar fields display a clean dispersal threshold, and Vaidya geometries isolate the effect of null flux. These are controlled laboratories, not templates for every astrophysical collapse.

Rotation and nonspherical structure add genuinely new questions. Angular momentum can remain in the black hole, be stored in a disk or cloud, leave in radiation or ejecta, or participate in an instability. No single coordinate angular velocity measures all of these channels. The useful quantities are asymptotic charges, quasi-local horizon measures, matter currents, and radiative fluxes, compared in a common convention and closed as a budget.

Once a black hole forms, several notions of horizon become available. An event horizon is the causal boundary of the region that cannot communicate with the relevant future infinity and therefore depends on the entire future. A marginally outer trapped surface is defined on one slice by the vanishing of an outgoing null expansion. A dynamical horizon is a smooth spacelike hypersurface foliated by closed marginally trapped two-surfaces, each with zero outgoing and negative ingoing expansion, and it supports local balance laws. These objects can coincide in equilibrium and differ sharply during collapse or merger. Their distinctions are operational: they decide what can be located during an evolution, what is slicing-dependent, and which area statement is being used.

A first calculation with a round sphere

Begin with a spherical, collapsing spacetime with $\Lambda=0$. The areal radius $r$ is defined by the sphere’s area $A=4\pi r^2$; it is a geometric measurement, not an arbitrary radial coordinate. Define its Misner–Sharp mass $m$ by

$$g^{ab}\nabla_a r\nabla_b r=1-\frac{2Gm}{r}.$$

The dimensionless ratio $2Gm/r$ compares the enclosed mass scale with the sphere’s size. To understand why it matters, send light normally away from the sphere in both future directions. The expansions $\theta_{(\ell)}$ and $\theta_{(n)}$ measure the fractional rates of change of a small bundle’s area. In ordinary weak-field collapse, the outward bundle expands while the inward bundle contracts.

For future null normals with $\ell\cdot n=-1$, spherical symmetry gives

$$\theta_{(\ell)}=\frac{2\ell(r)}{r}, \qquad \theta_{(n)}=\frac{2n(r)}{r}, \qquad \theta_{(\ell)}\theta_{(n)} =-\frac{2}{r^2}\left(1-\frac{2Gm}{r}\right).$$

If the inward expansion is negative, $2Gm/r\gt1$ makes the outward expansion negative too: even the outward light bundle initially loses area. Such a sphere is future trapped. Equality gives a round marginal sphere. This local calculation does not locate the event horizon, because escape to infinity depends on the later spacetime as well. The dust-collapse example and null-shell example show that distinction exactly.

This is a useful starting route: calculate a geometric quantity, identify what it proves, then ask what further information would answer the global question. The later chapters follow the same pattern when symmetry no longer reduces the problem to one radius.

The recurring physical tests

Five tests organize the volume.

1. Causal test. Does the statement concern escape to infinity, a trapped region on one slice, or a quasi-local boundary? Event, apparent, trapping, isolated, and dynamical horizons answer different questions.
2. Invariant test. Is the reported quantity independent of coordinates and, where necessary, of an asymptotic frame? Coordinate tracks and coordinate merger times are useful diagnostics but are not observables by themselves.
3. Regime test. Is the result exact, perturbative, small-data, symmetry-reduced, or fully nonlinear? A successful linear fit does not establish that the entire spacetime has entered a linear regime.
4. Evidence test. Does a computational claim survive constraint, convergence, boundary, gauge, extraction, and finder checks? Agreement at one resolution is not enough.
5. Endpoint test. Is an endpoint a stationary solution that exists, an attractor demonstrated in a specified evolution, or a conjectured generic final state? Existence, stability, and dynamical selection are separate claims.

These tests recur in critical collapse, mass inflation, superradiance, mergers, recoil, and cosmic censorship. They also explain why a theorem can be both decisive and limited. For example, a current preprint proves nonlinear stability for small perturbations of every fixed subextremal Kerr background under explicit vacuum and asymptotic hypotheses. It does not give estimates uniform through the extremal limit, formation from arbitrary large data, or stability with matter. Combined with an interior theorem, it does support a narrower conclusion about continuous extension at the Cauchy horizon for suitable small rotating vacuum data; that conclusion does not assert bounded curvature or classify every interior endpoint Hintz 2026, abstract, consequence (1), p. 3, and §§ 1.1–1.6, PDF. The interior-regularity chapter explains the distinction.

Conventions that matter here

The site uses metric signature $(-,+,+,+)$, retains $G$ explicitly by default, and uses time dependence $e^{-i\omega t}$. Sustained black-hole calculations may declare $G=c=1$, but a representative dimensional observable is then restored. Future-directed timelike vectors have norm $-1$.

Null expansions require a pair of future null normals. Rescaling $\ell^a\mapsto f\ell^a$ and $n^a\mapsto f^{-1}n^a$ with $f>0$ changes the individual normalization but not the statement that an expansion vanishes. Every consequential use of extrinsic curvature declares its sign. Angular momentum specifies whether $a=J/M$, $a_*=J/(GM^2)$, or a quasi-local construction is meant. A recoil velocity requires a declared asymptotic Bondi frame, and a quasinormal-mode statement uses retarded boundary conditions consistent with $e^{-i\omega t}$.

Chapter architecture

The chapters follow a physical progression. Each link leads to a substantive orientation page and a six-page route.

1. Spherical Gravitational Collapse establishes exact models, matter-dependent thresholds, and causal ordering.
2. Rotating and Nonspherical Collapse introduces angular-momentum budgets, mode competition, and validation beyond one dimension.
3. Trapped Surfaces and Horizon Formation develops null expansions, stability operators, formation criteria, and horizon location.
4. Critical Phenomena in Gravitational Collapse explains type I and type II behavior through self-similarity and unstable modes.
5. Dynamical Horizons and Flux Laws relates changing horizon geometry to matter, gravitational radiation, and angular momentum.
6. Black-Hole Interiors, Cauchy Horizons, and Mass Inflation separates blueshift, mass inflation, weak null singularities, and regularity-dependent extension questions.
7. Nonlinear Stability and Instability moves from a precise stability statement to gauge control, modulation, decay estimates, and bootstrap closure.
8. Superradiant Evolution, Saturation, and Hair Formation follows bosonic bound states from linear growth into model-dependent nonlinear endpoints.
9. Finite-Amplitude Tidal Dynamics and Strong-Field Encounters connects perturbative response to scattering, capture, disruption, and nonlinear matching.
10. Black-Hole Mergers and Common Horizons organizes merger by spacetime geometry, marginal-surface branches, and closed budgets.
11. Recoil, Remnants, and Final-State Formation develops kick accumulation, independent remnant estimators, Kerrness, and final-state qualifications.
12. Dynamical Tests of Cosmic Censorship turns weak and strong censorship into statements with explicit data classes, asymptotics, genericity, and regularity.
13. Collapse and Dynamics in de Sitter and Anti-de Sitter Spacetimes shows how cosmological horizons, timelike boundaries, reflections, and boundary conditions alter the evolution.

Reader routes

For a first route through nonlinear black-hole physics, read spherical collapse, trapped surfaces, critical phenomena, dynamical horizons, mergers, and remnants. This route builds the concepts in the order in which they enter an evolution.

For geometric analysis, begin with trapped surfaces, then study interiors, nonlinear stability, and censorship. The key habit is to write every theorem with its dimension, regularity, asymptotics, matter model, data topology, and conclusion before interpreting it.

For numerical relativity, take rotating collapse, horizon location, strong-field encounters, mergers, and remnants. The mathematical target is always a continuum and preferably invariant quantity; discretization details and executable workflows belong in Computational Relativity.

For nonlinear field dynamics, take critical collapse, superradiant evolution, mode coupling in encounters, and AdS dynamics. Compare the phase-space mechanisms: a codimension-one intermediate attractor, a confined negative-horizon-energy mode, nonlinear resonant transfer, and saturation through changing background parameters.

Interfaces with neighboring subjects

This volume follows the finite-amplitude spacetime and its horizons once a matter model or binary configuration has been specified. Equations of state, transport, and compact-star microphysics are developed in Relativistic Matter and Compact Objects. Discretizations, evolution infrastructure, and executable studies are developed in Computational Relativity. The pages here use their outputs only with declared continuum equations, diagnostics, convergence tests, and error budgets.

Linear spectra, response functions, and quasinormal-mode theory are developed in Black-Hole Perturbation Theory; inspiral expansions and effective orbital descriptions in Relativistic Two-Body Dynamics; and waveform construction and inference in Gravitational Radiation and Waveforms. Population synthesis, environmental demographics, and observational interpretation remain in Black-Hole Astrophysics. Causality and Global Structure gives the general definitions of infinity, maximal development, and cosmic censorship. This volume applies those inputs to nonlinear collapse, interiors, encounters, mergers, and remnant formation without replacing their general treatments.

What the volume does not claim

Spherical examples do not establish genericity without a topology on unrestricted initial data. A marginal surface does not by itself identify the event horizon. Continuous extension of an interior metric does not imply bounded curvature or deterministic continuation. A stationary black hole with hair need not be a dynamically selected endpoint. A quasinormal-mode fit beginning near a waveform peak does not uniquely define the onset of linear ringdown. A numerical absence of a horizon is not evidence for visible singularity until resolution, slicing, boundary conditions, and the horizon search have been tested.

The same discipline applies to positive results. The dynamical-horizon balance laws are exact under their hypotheses and make quasi-local flux meaningful Ashtekar and Krishnan 2004, §§3–5. Critical mass scaling follows from one unstable mode when the assumed phase-space picture holds Gundlach and Martín-García 2007, §2, pp. 4–11. These statements become more useful, not less, when their domains are explicit.

A successful exit

After an appropriate route, a reader should be able to classify a collapse calculation by its data, matter, symmetry, asymptotics, and diagnostics; distinguish causal and quasi-local horizons; derive the basic scaling and flux relations; formulate an interior or stability claim at the correct regularity; evaluate a merger budget with meaningful uncertainties; and identify exactly what would be needed to strengthen a conditional endpoint or censorship statement.

That collection of skills is the real subject of nonlinear black-hole dynamics: not a catalog of dramatic outcomes, but a controlled way to decide what the Einstein evolution has actually shown.

References

• Ashtekar, Abhay, and Badri Krishnan. “Isolated and Dynamical Horizons and Their Applications.” Living Reviews in Relativity 7 (2004): 10. doi:10.12942/lrr-2004-10.
• Gundlach, Carsten, and José M. Martín-García. “Critical Phenomena in Gravitational Collapse.” Living Reviews in Relativity 10 (2007): 5. doi:10.12942/lrr-2007-5.
• Hintz, Peter. “Nonlinear Stability of Subextremal Kerr Black Holes.” arXiv:2606.28253v2 (3 August 2026). arXiv.