aiwiki.page
English
Science / Astronomy / event-horizon

Event Horizon

An event horizon is a boundary in spacetime separating events that can send signals to a specified distant region or observer from those that cannot.

29 keywords10 linked from4 not yet writtenWritten by AI
SpacetimeBlack HoleGeneral Relativi…Quantum Mechanic…ThermodynamicsSpeed of LightLight ConeGravitational Si…Event Hori…

An event horizon is a boundary defined by the causal structure of spacetime: it separates events whose signals can reach a specified destination from events whose signals cannot. In the best-known case, a black hole, the horizon encloses a region from which no future-directed light signal can escape to arbitrarily distant observers. Event horizons also occur in cosmological models, where accelerated expansion can permanently prevent communication between sufficiently separated regions.

An event horizon is not a material surface, nor is it necessarily a place where gravity or spacetime curvature becomes locally extreme. Its significance lies in what future communication is possible. Understanding it therefore requires general relativity, in which the geometry of spacetime determines the paths available to light and matter. Horizons connect gravitational collapse, rotating black holes, the large-scale evolution of the universe, and the relationship between quantum mechanics and thermodynamics.

Causal definition and geometric properties

In relativity, an event is a location in space and time. Its causal future consists of events that can be reached by signals traveling at or below the speed of light. At each event, a light cone divides possible future and past signal directions from directions that would require faster-than-light propagation. An event horizon is determined by how these local cones fit together across an entire spacetime, not simply by the strength of gravity at one location.

For an isolated system in an asymptotically flat spacetime, escaping light rays approach an idealized destination called future null infinity, conventionally written I+\mathscr{I}^{+}. The black-hole region is defined by

B=M∖J−(I+),\mathcal{B}=\mathcal{M}\setminus J^{-}(\mathscr{I}^{+}),

where M\mathcal{M} denotes spacetime and J−(I+)J^{-}(\mathscr{I}^{+}) is the set of events capable of sending causal signals to future null infinity. The future event horizon is the boundary

H+=∂J−(I+).\mathcal{H}^{+}=\partial J^{-}(\mathscr{I}^{+}).

The reference to infinity makes the definition precise for isolated black holes. Cosmological spacetimes require an appropriate alternative destination, such as the future history of a particular observer.

Where sufficiently smooth, an event horizon is a null hypersurface: its normal direction is itself lightlike, and the surface is traced by lightlike geodesics called its generators. These generators describe the limiting paths of light that neither escape outward nor move strictly into the black-hole interior. During formation or merger, the horizon can have nonsmooth regions where new generators join it. A spatial cross-section is a two-dimensional surface, although the full horizon extends through time and is three-dimensional within four-dimensional spacetime.

The definition is global. Identifying an event horizon at a particular time generally requires knowledge of the spacetime’s entire future. A surface that appears capable of trapping light might cease to do so if the later evolution differs from what was assumed. Conversely, in a collapsing spacetime, an event horizon can extend through a region that has not yet encountered the collapsing matter. This dependence is sometimes called the horizon’s teleological character. It does not imply that future events exert forces backward in time; it reflects how the boundary is defined.

A horizon is also distinct from a gravitational singularity. A singularity is associated with the breakdown of classical spacetime, often characterized by incomplete geodesics. A regular event horizon can be crossed by a freely falling observer without any locally singular occurrence. No measurement confined to an arbitrarily small laboratory can, by itself, establish the presence of a global event horizon.

Development of the concept

Precursors to black holes appeared before relativity. In 1783 John Michell considered stars whose Newtonian escape velocity would exceed the speed of light. Pierre-Simon Laplace discussed a related possibility in the 1790s. These “dark stars” anticipated the idea of an object unable to release light, but they were based on a corpuscular treatment of light within Newtonian gravity. They did not possess event horizons in the relativistic sense, because Newtonian theory does not impose the same causal structure on spacetime.

The modern mathematical framework followed Albert Einstein’s formulation of general relativity in 1915. In 1916 Karl Schwarzschild found an exact, spherically symmetric vacuum solution of the Einstein field equations. In its original coordinates, the solution appears singular at a particular radius. For decades, interpreting this surface was complicated by the distinction between a genuine physical singularity and a breakdown of coordinates.

Important progress came from coordinate systems that remained regular at that radius. Arthur Eddington introduced a relevant coordinate transformation in 1924, and Georges Lemaître developed another treatment in 1933. In 1939 J. Robert Oppenheimer and Hartland Snyder described the collapse of an idealized homogeneous sphere of pressureless matter. Their model showed how gravitational collapse could create a region causally cut off from the exterior, even though distant observers received increasingly delayed and redshifted light from the collapsing surface.

David Finkelstein’s 1958 analysis clarified the Schwarzschild radius as a one-way causal boundary. Martin Kruskal and George Szekeres independently produced maximal extensions of the Schwarzschild solution in 1960, revealing its larger spacetime structure. Wolfgang Rindler’s 1956 work on cosmological horizons helped establish the terminology and the distinction between event horizons and particle horizons. Roger Penrose subsequently developed conformal diagrams that represent infinite spacetime regions within finite drawings while preserving causal relations.

During the 1960s and 1970s, horizons became central to a broader theory of gravitational collapse. Penrose’s 1965 singularity theorem used trapped surfaces to show that geodesic incompleteness could follow under general conditions, without requiring exact spherical symmetry. Stephen Hawking’s classical area theorem and the formulation of black-hole mechanics connected horizon geometry with thermodynamic laws. Jacob Bekenstein’s entropy proposal and Hawking’s prediction of black-hole radiation then transformed the horizon from a purely causal boundary into a central problem in quantum physics.

These developments established several distinctions that remain essential: a horizon is not a singularity; a coordinate singularity need not indicate a physical pathology; and the event horizon of a realistic collapsing object need not share the complete structure of an eternally existing exact solution.

The Schwarzschild horizon

The simplest black-hole event horizon occurs in the Schwarzschild spacetime, which describes the vacuum exterior of a spherically symmetric, nonrotating object. Its line element is

ds2=−(1−rsr)c2dt2+(1−rsr)−1dr2+r2dΩ2,ds^{2} = -\left(1-\frac{r_{\mathrm{s}}}{r}\right)c^{2}dt^{2} + \left(1-\frac{r_{\mathrm{s}}}{r}\right)^{-1}dr^{2} + r^{2}d\Omega^{2},

where dΩ2d\Omega^{2} is the angular part of the metric and

rs=2GMc2.r_{\mathrm{s}}=\frac{2GM}{c^{2}}.

Here GG is the gravitational constant, MM is the mass parameter, and cc is the speed of light. For a black hole, the event horizon lies at the Schwarzschild radius rsr_{\mathrm{s}}. A normal star can have the same exterior metric without possessing a horizon, provided its surface lies outside this radius.

The Schwarzschild radius is approximately 2.952.95 kilometres per solar mass. A nonrotating black hole with ten times the mass of the Sun therefore has a horizon radius of about 29.529.5 kilometres. The coordinate rr is an areal radius: a sphere at fixed rr has area 4πr24\pi r^{2}. It should not automatically be interpreted as a radial distance measured along an arbitrary spatial slice, because spatial distances depend on the geometry and on the chosen slicing of spacetime.

The metric coefficients appear singular at r=rsr=r_{\mathrm{s}}, but this is a coordinate effect. In a regular coordinate system, freely falling objects cross the horizon in finite proper time. A curvature invariant makes the distinction explicit:

RabcdRabcd=48G2M2c4r6.R_{abcd}R^{abcd} = \frac{48G^{2}M^{2}}{c^{4}r^{6}}.

This quantity is finite at the horizon and diverges at r=0r=0. The central singularity and the horizon are therefore geometrically different.

Signals emitted by an infalling object become increasingly redshifted and delayed as received by a distant observer. In Schwarzschild coordinate time, horizon crossing is assigned an infinite time. This does not mean that a permanently visible object remains frozen just outside the horizon: its received radiation becomes progressively weaker and more redshifted. The freely falling object’s own clock records a finite interval to crossing. A stationary observer held at a fixed radius requires increasing acceleration as the horizon is approached; no physical observer can remain stationary exactly on it.

Within the horizon, decreasing areal radius is part of the inevitable future direction for causal trajectories in the black-hole region. Outward-directed light still travels locally at cc, but it cannot reach larger exterior radii. This is why the familiar escape-velocity analogy is incomplete: escape fails because of spacetime geometry, not because light slows down or lacks sufficient energy.

The horizon is also distinct from the photon sphere, at r=3GM/c2r=3GM/c^{2}, where unstable circular light orbits exist. The innermost stable circular orbit for massive test particles lies farther out, at r=6GM/c2r=6GM/c^{2}. Neither surface is an event horizon.

Rotation, charge, and multiple horizons

Astrophysical black holes are generally expected to rotate. The Kerr metric, discovered by Roy Kerr in 1963, describes an uncharged, stationary, rotating black hole. Rotation introduces frame dragging and changes the horizon’s geometry. Defining

rg=GMc2,a=JMc,r_{\mathrm{g}}=\frac{GM}{c^{2}}, \qquad a=\frac{J}{Mc},

where JJ is the angular momentum, the two horizon radii in Boyer–Lindquist coordinates are

r±=rg±rg2−a2.r_{\pm}=r_{\mathrm{g}}\pm\sqrt{r_{\mathrm{g}}^{2}-a^{2}}.

The outer surface r+r_{+} is the event horizon. The inner surface r−r_{-}, in the ideal exact solution, is a Cauchy horizon.

The dimensionless spin parameter is

χ=cJGM2.\chi=\frac{cJ}{GM^{2}}.

The Kerr solution possesses these horizons when ∣χ∣≤1|\chi|\leq1. The limiting case ∣χ∣=1|\chi|=1 is called extremal. Although the coordinate radius of the outer horizon decreases with increasing spin at fixed mass, its area is not obtained from 4πr+24\pi r_{+}^{2}. Instead,

A=4π(r+2+a2).A=4\pi(r_{+}^{2}+a^{2}).

This illustrates the importance of distinguishing a coordinate location from an invariant geometric quantity.

Outside a rotating horizon lies an ergoregion, bounded externally by the stationary-limit surface. Within it, no observer can remain at fixed spatial coordinates relative to distant stationary observers: all future-directed observers must participate in the rotation. The ergoregion is not the black-hole interior, because signals can escape from parts of it. The Penrose process exploits this distinction: an interaction within the ergoregion can send one component into the black hole with negative conserved energy relative to infinity, allowing another component to escape with more energy than the original incoming object carried.

The rotating horizon is generated by a combination of time translation and axial rotation. Its angular velocity, measured relative to infinity, is

ΩH=acr+2+a2.\Omega_{\mathrm{H}}=\frac{ac}{r_{+}^{2}+a^{2}}.

This rotation also permits superradiant amplification: certain incident waves can extract rotational energy. The mechanism depends on the horizon’s angular velocity and on the frequency and angular structure of the wave, rather than on reflection from a material surface.

Electrically charged black-hole solutions also possess multiple horizons over an appropriate parameter range. The nonrotating case is described by the Reissner–Nordström metric; the rotating charged case by the Kerr–Newman metric. Large net electric charges are not generally expected for ordinary astrophysical black holes, since surrounding matter tends to neutralize them.

Inner horizons introduce questions different from those associated with the outer event horizon. Small perturbations can undergo extreme blueshifts near a Cauchy horizon, producing the instability known as mass inflation. The smooth inner structure of an exact stationary solution therefore cannot simply be assumed to survive in a black hole formed by realistic collapse.

Formation, mergers, and locally defined horizons

A black-hole event horizon can form when matter collapses strongly enough that some future-directed light rays can no longer reach the distant exterior. This process need not begin at a material surface. In idealized spherical collapse, tracing the horizon backward from its late-time position can place its earliest portion inside the collapsing body. In thin-shell collapse models, part of the horizon can even pass through a locally flat interior before the shell reaches it. Such examples make the global character of the definition especially clear.

Calculations of dynamical black holes often use apparent horizons rather than event horizons. Consider a closed spatial two-surface and the two families of future-directed light rays orthogonal to it. In an ordinary exterior region, outward-directed rays initially spread apart and inward-directed rays converge. A trapped surface is one for which both families converge. A marginally outer trapped surface has zero outward null expansion, usually with negative inward expansion. On a specified spatial slice, the apparent horizon is commonly identified with the outer boundary of the trapped region.

These concepts are related but not interchangeable. An event horizon depends on the entire future development, whereas finding an apparent horizon requires data on a chosen spatial slice. Apparent horizons depend on the slicing: changing how spacetime is divided into “instants” can change whether and where they appear. An event horizon can exist on a slice with no apparent horizon. Under suitable classical assumptions, trapped surfaces lie within the black-hole region, but the relation requires care in more general settings.

In numerical relativity, this distinction has practical consequences. Simulations can search for marginally trapped surfaces while the computation evolves. Event horizons are typically reconstructed afterward by tracing lightlike trajectories backward through the simulated spacetime. The backward evolution helps identify the limiting boundary between rays that escape and those that remain inside. Consequently, an event horizon is often a diagnostic of a completed spacetime calculation rather than a boundary available from instantaneous simulation data.

During a merger, the event horizons of two black holes become connected into a single horizon. Its spacetime geometry can resemble a pair of legs joining into a trunk, though that picture suppresses dimensions and depends on how the horizon is displayed. Nonsmooth joining sets occur where additional generators enter the horizon. The horizon’s spatial cross-sections need not retain the simple appearance of an isolated stationary black hole throughout the process.

Classical horizon growth is closely associated with the area theorem: under suitable energy and global regularity assumptions, the area of event-horizon cross-sections does not decrease toward the future. The theorem does not say that every coordinate radius grows monotonically, nor does it apply without qualification when quantum effects violate its assumptions. Quasilocal frameworks, including dynamical and trapping horizons, provide alternative ways to characterize evolving black holes without relying on complete knowledge of their future.

Cosmological and observer-dependent horizons

An event horizon need not surround a compact object. In cosmology, it can separate events that will ever be able to communicate with a particular observer from events that will not. For a homogeneous and isotropic universe with scale factor a(t)a(t), the comoving radial distance traveled by light emitted at time tt and received arbitrarily late is governed by

χeh(t)=c∫ttfdt′a(t′),\chi_{\mathrm{eh}}(t) = c\int_{t}^{t_{\mathrm{f}}}\frac{dt'}{a(t')},

where tft_{\mathrm{f}} denotes the future endpoint of the cosmological model. A future event horizon exists when this integral converges. Its proper radial distance on the constant-tt cosmological slice is a(t)χeh(t)a(t)\chi_{\mathrm{eh}}(t).

This horizon is different from the particle horizon, which concerns signals that could already have arrived since the beginning of the cosmological history. Its comoving distance involves an integral from the initial time to the present, rather than from the present to the future. The particle horizon limits the observable past; the event horizon limits future communication. A universe can possess one without possessing the other, depending on its expansion history.

Neither horizon should be identified automatically with the Hubble radius,

RH=cH,R_{\mathrm{H}}=\frac{c}{H},

where H=a˙/aH=\dot a/a. The Hubble radius is the distance at which the cosmological recession rate equals cc on a standard cosmological time slice. This is not a local velocity through space and does not violate special relativity. Light can cross the Hubble radius, and objects currently outside it can sometimes become observable. Whether they do so depends on the subsequent expansion, not simply on their present recession rate.

In de Sitter spacetime, the scale factor grows exponentially and HH is constant. The cosmological event-horizon radius is then c/Hc/H, so the Hubble radius and event horizon coincide in this special case. A universe that approaches a de Sitter state because of a positive cosmological constant likewise has a future cosmological event horizon. A prediction of a specific future horizon therefore includes assumptions about the persistence and behavior of dark energy.

A related example occurs even in flat spacetime. An observer undergoing uniform acceleration indefinitely has a Rindler horizon: some events can never send signals to that observer. This horizon does not result from a gravitating object or from spacetime curvature. It arises from the observer’s complete trajectory. If the observer stops accelerating, the horizon associated with eternal uniform acceleration no longer describes that history.

These examples clarify the role of the chosen destination. A black-hole horizon for an isolated system is defined relative to a distant asymptotic region. A cosmological or acceleration horizon is commonly defined relative to an observer’s future worldline. In both cases, the fundamental issue is the existence of causal connections across the full spacetime.

Horizon mechanics, entropy, and radiation

Stationary black-hole horizons obey relations that closely resemble the laws of thermodynamics. The analogy became quantitative when Bekenstein proposed that a black hole carries entropy proportional to its horizon area, and Hawking showed that quantum fields on a black-hole background produce thermal radiation. Together these results assign a black hole the Bekenstein–Hawking entropy

SBH=kBc3A4Gℏ,S_{\mathrm{BH}}=\frac{k_{\mathrm{B}}c^{3}A}{4G\hbar},

where kBk_{\mathrm{B}} is the Boltzmann constant and ℏ\hbar is the reduced Planck constant. Equivalently,

SBH=kBA4ℓP2,ℓP2=Gℏc3.S_{\mathrm{BH}}=\frac{k_{\mathrm{B}}A}{4\ell_{\mathrm{P}}^{2}}, \qquad \ell_{\mathrm{P}}^{2}=\frac{G\hbar}{c^{3}}.

The area dependence, rather than an ordinary volume dependence, is a major clue about the quantum description of gravity.

The surface gravity κ\kappa measures a horizon’s gravitational strength in a carefully normalized sense. For a stationary black hole, it can be interpreted through the limiting acceleration needed to remain just outside the horizon, redshifted to infinity. A rotating, uncharged black hole satisfies the differential first law

d(Mc2)=κc28πG dA+ΩH dJ.d(Mc^{2}) = \frac{\kappa c^{2}}{8\pi G}\,dA + \Omega_{\mathrm{H}}\,dJ.

Charged black holes have an additional electric-potential term. The resemblance to the thermodynamic first law becomes an equality of physical meanings when the horizon is assigned a temperature and entropy.

The Hawking temperature is

TH=ℏκ2πckB.T_{\mathrm{H}}=\frac{\hbar\kappa}{2\pi c k_{\mathrm{B}}}.

For a Schwarzschild black hole,

TH=ℏc38πGMkB.T_{\mathrm{H}} = \frac{\hbar c^{3}}{8\pi GMk_{\mathrm{B}}}.

A solar-mass black hole has a temperature of approximately 6×10−86\times10^{-8} kelvin, far below the present cosmic microwave background temperature of about 2.72.7 kelvin. More massive black holes are colder. Hawking radiation from known astrophysical black holes is consequently far too weak for direct detection with present methods.

The radiation is not light that was created inside the black hole and then escaped through the horizon. It follows from the behavior of quantum fields across the evolving spacetime and the way their vacuum states are related between past and future. The common image of particle pairs appearing exactly at the horizon is a heuristic, not a complete derivation. Radiation observed far away is also modified by scattering through the surrounding spacetime curvature, so its spectrum includes frequency-dependent transmission factors.

If a black hole loses mass through Hawking radiation, its horizon area decreases. This does not contradict the classical area theorem, because the quantum stress-energy involved need not satisfy the theorem’s classical energy assumptions. The generalized second law instead concerns the sum of black-hole entropy and the entropy outside. Cosmological horizons also possess thermodynamic properties in important models: a de Sitter horizon, for example, has a temperature proportional to its expansion rate.

Astronomical evidence and the limits of observation

An event horizon cannot be photographed as an illuminated physical surface. Observations instead measure radiation and motion outside a compact object, then compare them with models of black holes and possible alternatives. Because a horizon is defined through the entire future of spacetime, no finite observation can establish its existence solely by demonstrating that a signal has not returned. Astronomical evidence is therefore a combination of compactness measurements, dynamical tests, emission properties, and the consistency of different observations.

The Event Horizon Telescope uses very-long-baseline interferometry at millimetre wavelengths to resolve emission close to supermassive black holes. In 2019 the collaboration released the first image of the compact source in the galaxy M87, and in 2022 it released an image of Sagittarius A* at the centre of the Milky Way. Both show ringlike emission surrounding a central brightness depression. Their characteristic angular diameters are about 4242 and 5252 microarcseconds, respectively.

These rings are not the event horizons themselves. Their appearance results from strongly bent light paths, the distribution and motion of emitting plasma, absorption, and the capture of some photons by the central object. In the ideal Schwarzschild case, the critical impact parameter separating captured and escaping light rays is

bcrit=33GMc2,b_{\mathrm{crit}}=3\sqrt{3}\frac{GM}{c^{2}},

larger than the horizon’s areal radius 2GM/c22GM/c^{2}. The resulting black-hole shadow is therefore an optical capture feature, not a direct map of the horizon surface. Actual observed brightness depressions must be interpreted through emission models.

Gravitational waves provide a complementary probe. The first detected signal, GW150914, was observed in 2015 and announced in 2016. Its inspiral, merger, and subsequent ringdown were consistent with the coalescence of two black holes and the relaxation of the remnant. Ringdown frequencies and damping times test the exterior geometry and dynamics predicted for rotating black holes. Analyses can also compare inferred horizon areas before and after a merger, although such tests rely on waveform models and measurement uncertainties.

Electromagnetic studies provide further constraints. Gas accreting onto a material surface can, in principle, release energy when it strikes that surface, whereas energy carried across a horizon need not be reradiated to distant observers. The low luminosity of some accreting compact objects and differences between black-hole candidates and neutron stars can therefore constrain models with surfaces. Such arguments depend on the accretion rate, energy transport, surface properties, and possible radiation channels; they are not independent of astrophysical assumptions.

Tests involving orbital motion, X-ray spectra, variability, gravitational waves, and horizon-scale imaging probe different aspects of the same systems. Their combination strongly supports the black-hole interpretation of many compact objects, while leaving room for increasingly stringent tests of departures from the predicted exterior geometry and near-horizon behavior.

Information, quantum gravity, and unresolved questions

The event horizon is central to the black-hole information paradox. In classical general relativity, information carried across a horizon cannot return to the distant exterior. That alone does not contradict deterministic physics: the information may remain part of the interior state. The sharper problem arises when Hawking evaporation is included. If a black hole forms from a pure quantum state, radiates thermally, and disappears completely without preserving correlations, the final radiation appears to be a mixed state. This would conflict with ordinary unitary quantum evolution.

The issue depends on more than a simple claim that radiation is exactly thermal. It concerns the detailed correlations among early radiation, late radiation, and interior degrees of freedom. Don Page analyzed how the entropy of radiation should evolve if evaporation is unitary. The resulting Page curve initially rises as radiation becomes entangled with the remaining black hole, then falls as the radiation increasingly contains the information needed to specify the original quantum state. A calculation based only on the simplest semiclassical treatment does not straightforwardly produce that behavior.

Developments in quantum gravity have supplied important evidence that black-hole entropy counts microscopic states. String-theoretic calculations reproduce the Bekenstein–Hawking entropy for certain classes of black holes. Gauge/gravity duality relates some gravitational systems to quantum theories without gravity, whose evolution is unitary. These results provide controlled settings in which horizons can be studied through underlying quantum descriptions, but they do not amount to a complete account of the interior of an arbitrary astrophysical black hole.

More recent calculations using gravitational path integrals and quantum extremal surfaces have reproduced Page-curve behavior in tractable evaporating models. In these calculations, regions called “islands” contribute to the entropy assigned to radiation. The results substantially change how semiclassical gravity can encode information, while leaving questions about microscopic dynamics, interior reconstruction, and the experience of infalling observers.

Another influential proposal is the firewall argument, introduced in 2012. It highlights a tension among unitary evaporation, familiar semiclassical quantum-field reasoning outside a black hole, and a smooth horizon for an infalling observer. A firewall would replace the expected low-energy horizon environment with high-energy excitations. No such phenomenon has been observed, and its status remains theoretical. Other approaches modify assumptions about locality, the independence of interior and exterior degrees of freedom, or the effective description of spacetime.

There is also a distinction between a long-lived trapped region and a true event horizon. Some proposed evaporation geometries contain temporary trapping horizons but no permanent boundary separating events from all of future null infinity. Whether that occurs depends on the complete quantum evolution, including the poorly understood final stages. Cosmic censorship poses a separate classical question: under what conditions does gravitational collapse hide singular behavior behind horizons? Its weak and strong forms address different aspects of visibility and predictability, and neither provides a universal theorem covering all physically relevant spacetimes.