Hawking Rigidity

Hawking rigidity is the bridge from time independence to the extra symmetry needed by black-hole classification. A stationary event horizon is invariant under time translations, but this fact alone neither makes the horizon a Killing horizon nor produces a periodic rotation. Under suitable field, causal, horizon, and regularity hypotheses, rigidity constructs a horizon- normal Killing field; on a rotating horizon it can then be combined with the stationary field to obtain an axial Killing field. The distinction between a local construction and a global axial action is essential.

Required background. Analysis for Mathematical Relativity supplies characteristic uniqueness and unique-continuation tools, while Spacetime Symmetries and Killing Fields supplies Killing horizons, flows, and commuting isometries.

Helpful background. Einstein Equations provides the vacuum and electrovacuum field systems, Causality and Global Structure provides event horizons and domains of outer communications, Partial Differential Equations and Green Functions supports the Killing-wave equation, and Black-Hole Geometry and Mechanics supplies surface gravity and horizon generators.

From a stationary horizon to an axial symmetry

Let $T$ be the complete stationary Killing field, timelike in the chosen asymptotically flat end. A useful benchmark statement assumes a four-dimensional, analytic, stationary vacuum or electrovacuum spacetime with a regular domain of outer communications and a smooth rotating event horizon. In the standard uniqueness route one also assumes the relevant global regularity, a connected horizon, and either mean nondegeneracy or rotation. Strong rigidity first says that the event horizon is a Killing horizon: there is a Killing field $\chi$ normal to its null generators. In the rotating case $\chi$ is not proportional to $T$. After globalizing the second symmetry and establishing closed orbits, one may normalize an axial field $\Phi$ to period $2\pi$ so that

$$\chi=T+\Omega_H\Phi,$$

with constant horizon angular velocity $\Omega_H\ne0$. Thus stationarity plus rigidity gives the commuting group $\mathbb R\times U(1)$ used by the two-dimensional uniqueness problem. The analytic route, including the separate globalization and periodicity steps, is summarized precisely in Chruściel, Costa, and Heusler 2012, §3.3.1, pp. 15–16.

This statement deliberately separates three conclusions. Invariance of the event horizon under $T$ is a causal fact. Existence of $\chi$ makes it a Killing horizon. Existence of a complete $Φ$ with closed orbits and an axis is global axisymmetry. None of these conclusions follows merely by renaming the stationary field.

Why the proof needs horizon and regularity input

The analytic proof architecture has four stages. First, the field equations and stationarity control the null geometry of the horizon and identify a horizon generator with constant surface gravity. Second, nondegeneracy, $\kappa\ne0$, permits an extension containing a bifurcation surface; data on the two intersecting null sheets then determine a candidate Killing field. Third, one proves that the deformation tensor

$$\pi_{\mu\nu}=\nabla_\mu\chi_\nu+\nabla_\nu\chi_\mu$$

vanishes. Analytic continuation is the classical device for extending the local field through the exterior. Finally, global results for the isometry group show that a suitable linear combination of $T$ and $\chi$ has periodic orbits and a nonempty rotation axis. The last step is not a local consequence of the Killing equation.

Analyticity is powerful because agreement of the metric with its pullback by the candidate flow propagates beyond a characteristic neighborhood. It is also a genuine hypothesis, not a consequence already available everywhere in the exterior: stationary electrovacuum metrics are analytic where $T$ is timelike, while the ergoregion is exactly where that shortcut is inadequate. The connectedness, causal regularity, and asymptotic hypotheses prevent a locally produced symmetry from acquiring pathological global behavior.

A local model for the bifurcation surface

In flat coordinates, the boost field $B=x\partial_t+t\partial_x$ has norm $t^2-x^2$. It is null on the two planes $t=\pm x$ and vanishes on their intersection $t=x=0$. In the wedge $x\gt|t|$, set $t=\xi\sinh\eta$, $x=\xi\cosh\eta$; then

$$-dt^2+dx^2=-\xi^2d\eta^2+d\xi^2, \qquad B=\partial_\eta.$$

The degeneration of these coordinates at $\xi=0$ does not make the flat metric singular. It explains why two intersecting null sheets and a field vanishing at their intersection naturally appear in a nondegenerate horizon proof. This is a local model, not a black-hole example: its bifurcation surface is noncompact and it has no asymptotically flat black-hole exterior with the stated global hypotheses.

What smooth rigidity proves

There are important smooth theorems, but their domains must not be conflated with the analytic global statement. For a smooth vacuum spacetime containing a local, regular, bifurcate, nonexpanding horizon, one can construct a nontrivial Killing field tangent to both families of null generators in a full neighborhood of the bifurcation sphere. If another Killing field $T$ is tangent to the horizon and nonzero on that sphere, a local rotational Killing field commuting with $T$ follows. These are the exact conclusions of Alexakis, Ionescu, and Klainerman 2010, Theorems 1.1–1.2, pp. 1–3.

The smooth proof replaces analytic continuation by PDE uniqueness. One first solves the covariant wave equation obeyed by a Killing field from characteristic data on the bifurcate horizon. The deformation tensor vanishes in the characteristic domain of dependence. Beyond it, where the characteristic problem is not well posed, the candidate flow produces another vacuum metric agreeing on the horizon; a coupled wave–transport uniqueness argument and Carleman estimate force equality in a full neighborhood. Earlier smooth electrovacuum work obtains a horizon-normal Killing field on a one-sided neighborhood but explicitly notes the exterior-extension limitation; see Friedrich, Rácz, and Wald 1999, §§3–4 and Proposition B.1, pp. 5–15.

Local smooth rigidity therefore does not by itself prove global axisymmetry of an arbitrary stationary exterior. Global smooth extensions are known under additional geometric control, for example perturbative near-Kerr hypotheses. Dropping both analyticity and such continuation input leaves a real gap rather than a harmless technical omission.

Branches and limitations

If the horizon is nonrotating, the equation above has $\Omega_H=0$ and does not manufacture an independent $Φ$. The classification argument instead uses a maximal hypersurface and a staticity theorem, then applies static uniqueness. Degenerate horizons lack the bifurcation surface central to the smooth local construction, so extremal rigidity requires different methods. Multiple horizon components create further global compatibility questions.

Matter also matters. Vacuum and Einstein–Maxwell equations provide the propagation identities used in the proof; a generic stress tensor need not inherit the spacetime symmetry or preserve the same unique-continuation system. Consequently, a stationary hairy black hole is not a counterexample to the theorem unless it satisfies the theorem’s exact matter and regularity hypotheses. Likewise, a local Killing field near a horizon is not sufficient evidence for a global $U(1)$ action.

Check the rotating branch

Suppose rigidity supplies $\chi$ on a connected horizon and the global group argument supplies a period-$2\pi$ axial field. Explain why $\Omega_H\ne0$ makes $Φ=(\chi-T)/\Omega_H$ independent of $T$, and why this calculation does not prove axisymmetry for a nonrotating horizon.

Solution

If $Φ$ were proportional to $T$, then $\chi=T+\Omega_H\Phi$ would also be proportional to $T$, contrary to the definition of the rotating branch: its horizon generator is distinct from the stationary generator. Division by $\Omega_H$ is legitimate only when $\Omega_H\ne0$. In the nonrotating branch $\chi$ is generated by $T$ itself, so the displayed formula is singular and contains no second direction. Staticity, not this algebraic rearrangement, is the next theorem needed there. Periodicity and the existence of an axis also come from the global rigidity argument, not from the formula alone.

References

• Alexakis, Spyros, Alexandru D. Ionescu, and Sergiu Klainerman. “Hawking’s Local Rigidity Theorem Without Analyticity.” Geometric and Functional Analysis 20 (2010): 845–869. DOI.
• Chruściel, Piotr T., João Lopes Costa, and Markus Heusler. “Stationary Black Holes: Uniqueness and Beyond.” Living Reviews in Relativity 15 (2012): 7. DOI. Open PDF.
• Friedrich, Helmut, István Rácz, and Robert M. Wald. “On the Rigidity Theorem for Space-Times with a Stationary Event Horizon or a Compact Cauchy Horizon.” Communications in Mathematical Physics 204 (1999): 691–707. DOI. Open PDF.