Horizon Topology Theorems in Higher Dimensions

The dimension-independent horizon restriction is not “the cross-section is a sphere.” Under stability and an effective dominant energy condition, a smooth compact marginally outer trapped cross-section must be of positive Yamabe type, apart from a sharply rigid equality case. Four dimensions turn that conclusion into $S^2$ through Gauss–Bonnet and surface classification; higher dimensions do not. Symmetry can narrow the possibilities further, but only after its own hypotheses are added.

Helpful background. Exact Solutions of Einstein’s Equations supplies examples against which the theorem is tested; Global Structure supplies event horizons and marginal surfaces; the Einstein equation relates energy to curvature; Black-Hole Perturbation Theory supplies stability-operator techniques; and Black-Hole Geometry and Mechanics supplies null expansions. A more global treatment appears in Horizon Cross-Section Topology.

Geometric setting and hypotheses

Let $(M^{n+1},g)$ be a smooth, time-oriented spacetime with $n\geq3$, and let $V^n$ be a smooth spacelike hypersurface. Take a smooth, closed, connected, two-sided surface $\Sigma^{n-1}\subset V$. If $U$ is the future unit normal to $V$ and $N$ the chosen outward unit normal to $\Sigma$ within $V$, then

$$K=U+N$$

is an outward future null normal. Its null second fundamental form and expansion are

$$\chi(X,Y)=g(\nabla_XK,Y), \qquad \theta_+=\operatorname{tr}_{\Sigma}\chi.$$

A marginally outer trapped surface, or MOTS, has $\theta_+=0$. It is stably outermost when the principal eigenvalue of the outward MOTS stability operator is nonnegative, and strictly stable when that eigenvalue is positive. In the terminology used by Galloway and Schoen, an outer apparent horizon has the stronger property that no homologous surface outside it has $\theta_+\lt0$; modern treatments often call this a weakly outermost MOTS. This stronger condition supplies stability: a negative principal eigenvalue would generate an outer-trapped surface just outside, contradicting weak outermostness. Merely assuming that no second MOTS lies outside is not the same argument.

Write the field equation as

$$G_{ab}=8\pi G\,T^{\mathrm{eff}}_{ab}.$$

The tensor $T^{\mathrm{eff}}_{ab}$ is assumed to obey the dominant energy condition near $\Sigma$. If the cosmological term is written on the geometric side, then

$$T^{\mathrm{eff}}_{ab}=T_{ab}-\frac{\Lambda}{8\pi G}g_{ab}.$$

The sign of $\Lambda$ therefore matters. The pure negative-$\Lambda$ term does not satisfy the effective condition used in the asymptotically flat theorem; additional matter would have to make the combined $T^{\mathrm{eff}}_{ab}$ obey it. Flat or higher-genus AdS horizons are therefore not counterexamples to the stated theorem.

No stationarity, asymptotic flatness, or global hyperbolicity is needed for the local stable-MOTS statement itself. Those assumptions enter when one proves that a cross-section selected from an event horizon has the stated marginal and outermost properties. The surface must be compact, without boundary, and two-sided; corners, intersections with timelike boundaries, or nonsmooth event-horizon cuts require different results.

Positive-Yamabe conclusion

A closed manifold is of positive Yamabe type if it admits at least one Riemannian metric with positive scalar curvature. This is a property of the differentiable topology. It does not assert that the physical induced metric on the horizon has positive scalar curvature point by point.

Stable-MOTS topology theorem. Under the hypotheses above, $\Sigma$ is of positive Yamabe type unless the induced metric is Ricci flat, the outward null second fundamental form vanishes, and the effective flux $T^{\mathrm{eff}}(U,K)$ vanishes along $\Sigma$. Strict stability excludes this exceptional alternative. Galloway and Schoen state the outer-apparent-horizon result, prove the scalar-curvature conclusion, and explain the stability extension in Galloway and Schoen 2006, Theorem 2.1 and proof, pp. 573–574; concluding remark 1, p. 576.

The equality case is simultaneous saturation of several nonnegative terms, not a broad second class of horizons. Additional strict stability, generic flux, or suitable outer area-minimizing information removes it. Omitting the exception from a merely non-strict theorem would turn a nonnegative statement into an unjustified positive one.

Proof architecture

An outward normal deformation with speed $\phi$ changes the expansion through an elliptic operator

$$\left.\frac{\partial\theta_+}{\partial t}\right|_{t=0}=L\phi,$$

where one convenient normalization gives

$$L\phi=-\Delta\phi+2\langle X,\nabla\phi\rangle +\bigl(Q+\operatorname{div}X-\lvert X\rvert^2\bigr)\phi,$$ $$Q=\frac12R_{\Sigma} -8\pi G\,T^{\mathrm{eff}}(U,K) -\frac12\lvert\chi\rvert^2.$$

Although $L$ is generally not self-adjoint, its principal eigenvalue is real and has a strictly positive eigenfunction. Stability makes that eigenvalue nonnegative. Substituting the positive eigenfunction and completing a square yields, for every smooth test function $\psi$,

$$\int_{\Sigma} \left(\lvert\nabla\psi\rvert^2+Q\psi^2\right)d\mu_{\Sigma} \geq0.$$

The dominant energy condition fixes the effective-flux term’s sign, and the shear term already has a definite sign. A conformal rescaling based on the lowest eigenfunction then constructs a metric with nonnegative scalar curvature. Unless every relevant contribution saturates, a further conformal argument makes the scalar curvature strictly positive. Complete saturation gives the Ricci-flat, shear-free, zero-flux exception.

The logic has three distinct steps: horizon stability supplies the elliptic inequality; the energy condition supplies sign control; conformal geometry turns that inequality into a Yamabe statement. Topological censorship and null completeness are not hidden inputs to this proof.

Four dimensions versus higher dimensions

In a four-dimensional spacetime, an orientable $\Sigma$ is a closed connected two-manifold. Once the exceptional case is excluded, choose a metric with positive scalar curvature. Since $R_{\Sigma}=2K$,

$$\int_{\Sigma}R_{\Sigma}\,d\mu =4\pi\chi(\Sigma)\gt0.$$

For an orientable surface of genus $g$, $\chi=2-2g$, so $g=0$ and $\Sigma$ is diffeomorphic to $S^2$.

When $\dim\Sigma\geq3$, positive scalar curvature no longer classifies the manifold as a sphere. The product $S^2\times S^1$ admits a positive-scalar-curvature metric and is realized by the five-dimensional black ring. Lens spaces and many other manifolds are also of positive Yamabe type. Conversely, scalar-curvature obstructions still exclude broad classes. The theorem is a nontrivial filter, not a complete list.

Additional symmetry produces a different, narrower statement. Consider an analytic, stationary, asymptotically flat five-dimensional vacuum black hole with globally hyperbolic exterior, nondegenerate horizon, and compact connected horizon cross-section. Suppose two $U(1)$ axial Killing fields commute with each other and with the stationary field, and the $U(1)^2$ action has no points with a discrete isotropy subgroup. The orbit-space data then restrict the cross-section to $S^3$, $S^1\times S^2$, or a lens space. Hollands and Yazadjiev derive this biaxisymmetric classification and show why rod data are needed beyond mass and angular momenta in Hollands and Yazadjiev 2008, §2, pp. 750–751; §3 and Proposition 2, pp. 753, 757–758. This conclusion cannot be quoted after dropping the symmetry action, nondegeneracy, regularity, or the stated asymptotics.

Limits and counterexamples to stronger claims

The black ring is the simplest counterexample to “positive Yamabe means spherical”: $S^1\times S^2$ satisfies the curvature condition and supports a regular vacuum horizon. AdS black holes with flat or higher-genus sections show why the effective energy and boundary assumptions matter. An arbitrary slicing of a nonsmooth event horizon can also produce cuts to which the stable-MOTS theorem does not apply.

Nor does an allowed topology guarantee a solution. Positive Yamabe type is necessary under the theorem’s hypotheses, not sufficient for existence. A candidate lens-space horizon may be compatible with the topological filter yet fail axis regularity, develop closed timelike curves, or lack a global vacuum exterior. Known exact families and theorem-permitted manifolds are different sets.

Check your understanding

Assume strict stability and the effective dominant energy condition. Explain why an orientable horizon section must be $S^2$ in four spacetime dimensions but may be $S^2\times S^1$ in five.

Solution

Strict stability gives positive Yamabe type. In four dimensions, $\Sigma$ is two-dimensional. Choose a metric with positive scalar curvature and integrate it. Gauss–Bonnet gives positive Euler characteristic, and a closed connected orientable surface has $\chi=2-2g$. Hence $g=0$ and $\Sigma\simeq S^2$.

In five dimensions, $\Sigma$ is three-dimensional, so the two-dimensional Gauss–Bonnet classification step is unavailable. Give $S^2\times S^1$ the product of a round sphere and a circle. Its scalar curvature is the positive curvature of the sphere plus zero from the circle. It is therefore positive Yamabe while remaining nonspherical, exactly as the theorem permits.

The final page combines the permitted topologies with exact, perturbative, approximate, and numerical solution branches: Phase Diagrams, Nonuniqueness, and Stability Boundaries.

References

• Galloway, Gregory J., and Richard Schoen, “A Generalization of Hawking’s Black Hole Topology Theorem to Higher Dimensions,” Communications in Mathematical Physics 266, 571–576 (2006). DOI. Open PDF.
• Hollands, Stefan, and Stoytcho Yazadjiev, “Uniqueness Theorem for 5-Dimensional Black Holes with Two Axial Killing Fields,” Communications in Mathematical Physics 283, 749–768 (2008). DOI. Open PDF.