Kerr and Electrovacuum Uniqueness

Kerr–Newman uniqueness has two logically separate parts. Rigidity and staticity first turn a stationary black hole into either a static problem or a stationary-axisymmetric one. The axisymmetric Einstein–Maxwell equations then become a harmonic-map boundary problem. A nonnegative divergence identity shows that two maps with the same singular boundary behavior and asymptotic charges cannot differ. The phrase “same mass, angular momentum, and charge” is therefore shorthand for a much richer matching statement that includes the orbit-space and horizon data.

Required background. Analysis for Mathematical Relativity supplies elliptic uniqueness and boundary estimates, while Spacetime Symmetries and Killing Fields supplies rigidity, axial actions, twists, and Killing horizons.

Helpful background. Einstein Equations provides the Einstein–Maxwell system, Causality and Global Structure provides the regular domain of outer communications, Partial Differential Equations and Green Functions supports the half-plane estimates, and Black-Hole Geometry and Mechanics supplies horizon charges and angular velocity.

The classification theorem and its hypotheses

A standard global theorem assumes a four-dimensional, analytic, stationary, asymptotically flat electrovacuum spacetime with a suitably regular domain of outer communications. Here regularity means the $I^+$-regularity package: the stationary Killing field is complete, the exterior is globally hyperbolic, and it contains a connected acausal spacelike hypersurface whose closure consists of a compact part and finitely many asymptotic ends. Its compact boundary lies on the future boundary of the exterior and meets each generator there exactly once; see Chruściel, Costa, and Heusler 2012, Definition 2.1, pp. 10–11 (PDF). Assume the event horizon is connected and either mean nondegenerate or rotating. Then its domain of outer communications is isometric to that of a Kerr–Newman spacetime. This is the precise content of Chruściel, Costa, and Heusler 2012, Theorem 3.3, pp. 15–17.

Every adjective has work to do. Global regularity supplies a well-behaved Cauchy hypersurface, complete stationary orbits, compact horizon sections, and the causal control needed by topology and elliptic reduction. Analyticity is used in the classical acquisition of global axisymmetry; it is not needed if the axial symmetry and corresponding orbit-space regularity are assumed at the start. Connectedness selects a single horizon rod. Nondegeneracy supports the standard compact-boundary geometry, while the theorem’s rotating branch allows the rigidity route stated there. Electrovacuum means no charged matter current is present in the exterior; the Maxwell field also inherits the stationary symmetry. Mean nondegeneracy is the nonzero average of the Gaussian-null horizon coefficient defined before the Killing generator has been constructed. Once rigidity gives a Killing horizon of constant surface gravity, it is the usual condition $\kappa\ne0$; see Chruściel, Costa, and Heusler 2012, §2.5.3, pp. 8–10 (PDF).

Within the already stationary-axisymmetric class, a more direct theorem says that a regular connected-horizon exterior is Kerr–Newman. The static branch gives Reissner–Nordström; the rotating branch enters the Ernst system. Electric and magnetic potentials may both be retained, with electromagnetic gauge constants fixed at infinity. If a duality convention suppresses magnetic charge, that is a choice of representative rather than an extra no-hair theorem.

Two electrovacuum maps and the Mazur identity

Reduction by the axial Killing field packages its norm, twist, and the electric and magnetic potentials into a positive Hermitian matrix $\mathbf\Phi$ representing the noncompact symmetric target $SU(2,1)/S(U(2)\times U(1))$. Define the current

$$J=\mathbf\Phi^{-1}d\mathbf\Phi.$$

For two candidate exteriors, choose the same Weyl half-plane and write their maps as $\mathbf\Phi_1$ and $\mathbf\Phi_2$. A relative difference is

$$\mathbf\Psi =\mathbf\Phi_2\mathbf\Phi_1^{-1}-\mathbf 1.$$

After factorizing the positive matrices, the difference of the currents can be conjugated to a matrix $\mathcal M$ for which the on-shell Mazur identity has a manifestly nonnegative right-hand side. In axisymmetric flat three-space it reads schematically

$$\operatorname{Tr}(\Delta_3\mathbf\Psi) =\operatorname{Tr}(\mathcal M\mathcal M^\dagger)\ge0.$$

Integrating over the orbit half-plane $S$ gives

$$\int_{\partial S}\rho\,*d\operatorname{Tr}\mathbf\Psi =\int_S\rho\, \operatorname{Tr}(\mathcal M\mathcal M^\dagger)\,d\rho\,dz.$$

The coset construction and the identity are derived in Chruściel, Costa, and Heusler 2012, §7.1, pp. 36–37, and the original electrovacuum application appears in Mazur 1982, §§2–4, pp. 3173–3180.

Why vanishing boundary flux forces equality

At infinity, equality of mass, angular momentum, electric charge, and magnetic charge makes the relative map approach the identity rapidly enough for its flux to vanish. On a regular axis rod, elementary flatness and matched potential constants control the apparent singularity at $\rho=0$. On the horizon rod, the two solutions must have the same interval structure and compatible horizon potentials. Endpoint estimates handle the corners where axis and horizon meet. These are boundary hypotheses, not consequences of the nonnegative bulk integrand.

Once the total boundary integral is zero, nonnegativity implies $\mathcal M=0$ pointwise. Hence $J_1=J_2$. The relative matrix is constant, and its value at infinity is the identity, so $\mathbf\Phi_1=\mathbf\Phi_2$. The remaining conformal metric function obeys first-order quadrature equations with the same normalization, and therefore the four-metrics coincide. Electromagnetic potentials coincide up to the gauge constants already fixed at infinity.

There is an analytic subtlety here. The factor $ρ$ vanishes on the axis and horizon, while individual potential expressions can look singular there. The formal statement “$ρ=0$, so the boundary term vanishes” is not a proof. One needs the precise asymptotic and corner estimates. The review notes a gap in a completely general derivative-boundary analysis and explains why the alternative harmonic-map distance method avoids it; see Chruściel, Costa, and Heusler 2012, §§8.4.1–8.4.2, pp. 45–47. For maps into the negatively curved target, their distance is subharmonic; a maximum principle with the same singularity matching then gives equality without differentiating that distance on the rods.

Scope and discriminating failures

The conclusion is not that every stationary object with the same global charges is Kerr–Newman. A second disconnected horizon changes the rod data; formal multi-black-hole maps may also carry conical singularities. Degenerate components require different endpoint analysis. If asymptotic flatness is replaced by anti-de Sitter boundary conditions, infinity contributes different data. None is a counterexample to the stated theorem.

Non-Abelian matter gives a sharper limitation. The reduced equations need not form the same nonpositively curved symmetric-space model, and asymptotic gauge charges need not encode the interior gauge field. Hairy Einstein–Yang–Mills black holes therefore invalidate a naive extension of the proof, not the Einstein–Maxwell classification. Higher dimensions likewise admit additional horizon topology and rod structures; four-dimensional topology is doing real work here.

Check the identity’s logical step

Suppose two regular maps have matching boundary behavior and satisfy

$$0=\int_S\rho\, \operatorname{Tr}(\mathcal M\mathcal M^\dagger)\,d\rho\,dz.$$

Explain why they coincide. Then identify why equality of asymptotic charges alone is not sufficient to start this conclusion.

Solution

The integrand is continuous and nonnegative, and $ρ>0$ in the interior of $S$. A zero integral forces $\mathcal M=0$ there. By construction this makes the two sigma-model currents equal, so the relative map is constant. Matching its limit at infinity fixes that constant to the identity. Continuity extends equality to the rods, and the quadrature equations then give the same metric.

Charges alone control only the asymptotic part of the boundary integral. They do not guarantee the same number and lengths of horizon rods, matching axis constants, regularity at corners, or absence of conical defects. Without those conditions the left-hand boundary flux has not been shown to vanish, so the nonnegative identity cannot yet be used.

References

• 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.
• Mazur, Paweł O. “Proof of Uniqueness of the Kerr–Newman Black Hole Solution.” Journal of Physics A: Mathematical and General 15 (1982): 3173–3180. DOI.