Killing Horizons and Bifurcate Geometry

A Killing horizon is a null hypersurface generated by a spacetime symmetry. That compact definition links local null geometry to global extension questions: the Killing field becomes both normal and tangent, its non-affinity defines surface gravity, nonzero surface gravity permits boost-like bifurcation geometry under additional hypotheses, and zero surface gravity changes the affine and near-horizon structure. This chapter develops those connections without assuming that every Killing horizon is an event horizon or importing a quantum thermal interpretation.

Helpful background. Exact Solutions of Einstein’s Equations supplies regular Schwarzschild, Kerr, and Reissner–Nordström charts; Causality and Global Structure supplies achronal boundaries, extensions, and completeness; The Einstein Equation clarifies which later mechanics statements use dynamics rather than kinematics.

Enter the horizon geometry

Fix a smooth null hypersurface $\mathcal H$ and a nontrivial Killing field $\chi^a$ normal to it. Because a null normal is also tangent, the Killing orbits are the null generators. Their non-affine geodesic equation is

$$\chi^b\nabla_b\chi^a=\kappa\chi^a,$$

or, equivalently,

$$\nabla_a(\chi^2)=-2\kappa\chi_a \quad\text{on }\mathcal H.$$

These identities are purely geometric. The numerical value of $\kappa$ is not: a constant rescaling $\chi^a\mapsto c\chi^a$ gives $\kappa\mapsto c\kappa$. In an asymptotically flat stationary black hole, the usual scale comes from unit time translation at the selected infinity and $2\pi$-periodic rotations. A local acceleration horizon has no such preferred scale. The normal-generator relation and its rescaling are derived in Gourgoulhon and Jaramillo 2006, arXiv v2, §2.5, Eqs. (2.21)–(2.26), Open PDF, pp. 21–23.

Throughout this chapter, $\chi^a$ is future-directed on a future branch, and any quoted $\kappa$ is tied to an explicit generator normalization. The global metric signature and curvature conventions are inherited. Detailed quantum field theory and thermal-state arguments belong to the curved-spacetime QFT volume; the subject here is classical horizon geometry.

Check your preparation

The shortest diagnostic has three parts. Can you verify $\nabla_{(a}\chi_{b)}=0$ for a candidate field? Can you distinguish a null normal from a unit normal, which does not exist for a null hypersurface? Can you distinguish a complete vector-field flow from affine completeness of a geodesic?

If the Killing equation or symmetry normalization is uncertain, review Killing Fields and Isometry Groups. For null boundaries and generator endpoints, use Causality and Global Structure. For exact checks, it is enough to know Schwarzschild in ingoing Eddington–Finkelstein coordinates and the root structure of Reissner–Nordström; Exact Solutions of Einstein’s Equations supplies both.

The most important preparation is conceptual: a Killing horizon is defined by symmetry, while an event horizon is defined by escape to a future receiver. Rindler horizons in Minkowski spacetime satisfy the first definition without being black-hole event horizons. A dynamical black-hole event horizon may satisfy the second without the first.

Three routes through the chapter

The local-geometry route follows the first, second, and fourth pages. It derives the generator equation, compares definitions of surface gravity, and constructs Gaussian null coordinates. This is the efficient route for calculations on an already identified stationary horizon.

The extension and theorem route reads the second, third, and fifth pages. It asks when a nondegenerate branch can be embedded into a spacetime with a bifurcation surface and which causal assumptions control generator endpoints. The Rácz–Wald theorem is stated with global hyperbolicity, stationarity, compact cross-sections, and no-white-hole hypotheses rather than reduced to a Kruskal picture.

The extremal route reads the second, fourth, and sixth pages. It tracks $\kappa$ from normalization to the order of the Killing norm’s transverse zero, then compares affine parameters, proper-distance throats, and near-horizon scaling. Reissner–Nordström supplies a discriminating family in which a simple root becomes double.

The six pages in working order

1. Normal Killing Fields and Null Generators proves that a normal Killing symmetry follows the null generators and distinguishes a norm-zero set from a Killing horizon.
2. Surface Gravity and Normalization derives the non-affinity, norm-gradient, invariant-square, static-lapse, and simple-root formulas and tracks their common normalization.
3. Bifurcation Surfaces and Extensions identifies the fixed surface of a bifurcate horizon and states exactly when a stationary future branch admits a global bifurcate enlargement.
4. Gaussian Null and Horizon-Adapted Coordinates constructs a regular chart and reads off expansion, shear, normal-bundle connection, surface gravity, and transverse data.
5. Generator Completeness and Horizon Components separates Killing-flow completeness, affine completeness, and endpoint structure, including crease sets and merging components.
6. Nondegenerate and Degenerate Horizon Geometry compares simple and double norm zeros, ordinary bifurcation geometry, extremal throats, and near-horizon limits. The general degenerate scaling form is developed in Kunduri and Lucietti 2013, §2.1, Eqs. (5)–(7), pp. 11–12.

The sequence moves from a local definition to global qualifications and then returns to a controlled family comparison. Each page supplies a solved diagnostic because the most common errors are transfers of a valid statement across an invalid boundary.

Recurring checks and non-implications

Schwarzschild normalization. With $\chi^a=(\partial_v)^a$ normalized at infinity, $\kappa=1/(4GM)$. Any formula for $\kappa$ must reproduce this value and its inverse-length dimension.

Kerr generator. The stationary field $t^a$ becoming null defines the stationary-limit surface, not generally the horizon. The horizon generator is $\chi^a=t^a+\Omega_H\phi^a$; normality, not norm zero alone, is decisive.

Affine parameter. For constant nonzero $\kappa$, $\lambda-\lambda_0\propto e^{\kappa v}$. A complete Killing parameter on one future branch therefore reaches a finite past affine value. For $\kappa=0$, $\lambda$ is linear in $v$.

Extension scope. A regular local horizon chart does not prove a global bifurcation surface, compact cross-section, complete generator, or event- horizon interpretation. Rácz and Wald’s exact global hypotheses and patching argument are in Rácz and Wald 1996, §§2–4, pp. 543–550. Event-horizon generators can have past endpoints at crease sets, while future completeness remains an additional hypothesis or consequence of suitable conformal-causal control; the alternatives are made explicit by Chruściel, Delay, Galloway, and Howard 2001, Theorem 1.1, pp. 109–112.

Review the chapter

Near a connected Killing horizon, a regular chart has

$$ds^2=2\,dv\,dr-2\kappa r\,dv^2+R_H^2d\Omega_2^2+O(r).$$

Assume $\partial_v$ is normalized and its orbits have $v\in\mathbb R$. For $\kappa\gt0$, identify the generator, its affine parameter range, and the extra hypotheses needed before asserting a global bifurcation surface. Then state what changes when $\kappa=0$.

Solution

The horizon is $r=0$ and its normal Killing generator is $\chi^a=(\partial_v)^a$. For $\kappa\gt0$, an affine parameter satisfies $\lambda-\lambda_0=C e^{\kappa v}/\kappa$, so the future branch is complete to the future but reaches finite $\lambda_0$ as $v\to-∞$. A global bifurcate extension does not follow from this local expansion: one needs an extension theorem’s stationarity, causal, compact-cross-section, constancy, and orbit- space hypotheses. In the Rácz–Wald final-state class, those conditions provide the enlargement and fixed surface.

For $\kappa=0$, the linear term in $g_{vv}$ vanishes and smoothness gives $g_{vv}=O(r^2)$. The Killing parameter is affine up to a linear change. An ordinary bifurcation surface for the same nontrivial generator cannot occur, and the natural near-horizon scaling is $v\mapsto v/\varepsilon$, $r\mapsto\varepsilon r$.

After this chapter, a reader should be able to identify a Killing horizon, compute and normalize its surface gravity in more than one way, construct a regular adapted chart, state a bifurcate-extension theorem without dropping its hypotheses, and diagnose the discontinuities of the degenerate limit. Continue to Conserved Charges, Smarr Relations, and First Laws to place these normalized horizon data in global integral and variational relations.

References

• Chruściel, Piotr T., Erwann Delay, Gregory J. Galloway, and Ralph Howard. “Regularity of Horizons and the Area Theorem.” Annales Henri Poincaré 2 (2001): 109–178. DOI. Open PDF.
• Gourgoulhon, Éric, and José Luis Jaramillo. “A 3+1 Perspective on Null Hypersurfaces and Isolated Horizons.” Physics Reports 423 (2006): 159–294. DOI. Open PDF.
• Kunduri, Hari K., and James Lucietti. “Classification of Near-Horizon Geometries of Extremal Black Holes.” Living Reviews in Relativity 16 (2013): 8. DOI.
• Rácz, István, and Robert M. Wald. “Global Extensions of Spacetimes Describing Asymptotic Final States of Black Holes.” Classical and Quantum Gravity 13 (1996): 539–553. DOI. Open PDF.