Mass–Angular-Momentum–Charge Inequalities

Sharp inequalities involving mass, angular momentum, and charge are not consequences of the positive mass theorem with symbols appended. Angular momentum needs an axial symmetry or another conserved definition; charge needs Maxwell constraints and a fixed normalization; equality identifies an extremal stationary geometry only within a carefully specified data class. For the asymptotically flat end of a two-ended, maximal, axisymmetric electrovacuum data set, the characteristic Kerr–Newman bound is

$G^2M^2\ge \tfrac12\bigl(Q^2+\sqrt{Q^4+4G^2J^2}\bigr)$ in the charge normalization stated below. Its proof is a harmonic-map minimization problem, not a dimensional estimate.

Required background. Analysis for Mathematical Relativity supplies variational estimates, Hypersurfaces and Foliations supplies initial data, Curvature supplies constraint geometry, and Einstein Equations fixes the matter coupling.

Helpful background. Causality and Global Structure clarifies end and horizon assumptions, while Partial Differential Equations and Green Functions supports the elliptic reduction.

Charges and conventions

Let $(\Sigma,h,K,E,B)$ be smooth three-dimensional Einstein–Maxwell initial data with one chosen asymptotically flat end and a second end that is asymptotically flat or asymptotically cylindrical. There is no inner boundary in the standard global theorem. Use

$$K_{ij}=-h_i{}^\mu h_j{}^\nu\nabla_\mu n_\nu,$$

and assume the source-free Maxwell constraints hold on the initial manifold. Define the geometrized charge $Q$ so that the Reissner–Nordström horizon condition is $G^2M^2\ge Q^2$; equivalently, $Q$ has units of length. Let $J$ be the ADM angular momentum associated with a complete axial Killing field $\eta$ whose orbits have period $2\pi$. Physical $J$ then has units of action, and $GJ$ has units of length squared in $c=1$ units.

For maximal data, $\operatorname{tr}_hK=0$. Assume a complete $U(1)$ action preserving $h,K,E,B$, the topology needed for global twist and electromagnetic potentials, the specified end decay, and a simply connected orbit space. The axisymmetric constraints then reduce the mass functional to the energy of a harmonic map, with fixed $(J,Q)$ as boundary data on the rotation axis. The original two-asymptotically-flat-end theorem is Chruściel and Costa 2009, Theorem 1.1 and §§2–3, printed pp. 1–7. The version allowing an asymptotically cylindrical second end, together with the strict equality statement, is summarized in Dain and Gabach-Clement 2018, Theorem 2 and §3.2.1 and rests on the gap estimate of Schoen and Zhou 2013, Theorem 6.1 and §6.1, printed pp. 21–25.

The vacuum mass–angular-momentum bound

For smooth, complete, maximal, asymptotically flat, axisymmetric vacuum initial data in the required Brill class, with its end and topology assumptions, one obtains

$$G^2M^2\ge |GJ|, \qquad\text{or}\qquad M^2\ge\frac{|J|}{G}.$$

Extreme Kerr saturates the inequality. The proof writes the ADM mass as a functional bounded below by the energy of a map from the orbit half-plane to a negatively curved target. Angular momentum fixes the jump of the twist potential between axis components. Extreme Kerr is the minimizer in that boundary class, yielding the lower bound. Dain’s theorem and variational characterization are stated in Dain 2008, Theorem 1.1 and §§2–3, printed pp. 34–54.

Axisymmetry is consequential: without it, no global twist potential supplies this reduction and a general conserved angular momentum need not exist. Maximality is likewise used in the mass functional. Removing either hypothesis does not create a known universal theorem of the same form.

Adding electromagnetic charge

In axisymmetric Einstein–Maxwell data, electric and magnetic potentials enlarge the harmonic-map target. If $Q^2=Q_E^2+Q_B^2$, the sharp Kerr–Newman relation can be written

$$G^2M^2 \ge \frac12\left(Q^2+\sqrt{Q^4+4G^2J^2}\right).$$

Setting $J=0$ gives $GM\ge|Q|$, the extreme Reissner–Nordström limit. Setting $Q=0$ gives $G^2M^2\ge|GJ|$, the extreme Kerr limit. Both checks fix the normalization and show why adding separate bounds would not reproduce the coupled square root. Equality requires the extreme Kerr–Newman data within the two-ended, maximal, electrovacuum, axisymmetric theorem class. This conclusion uses the later gap and rigidity analysis: the 2009 two-asymptotically-flat-end theorem did not contain the extremal cylindrical-end data and therefore did not itself prove the equality classification.

An elementary check recovers exactly this coupled expression from the Kerr–Newman horizon condition. Write $\mu=GM$ and $a=J/M$, so that $\Delta(r)=r^2-2\mu r+a^2+Q^2$. Real horizon roots require $\mu^2\ge a^2+Q^2$. Multiplication by $\mu^2$ gives

$$X^2-Q^2X-G^2J^2\ge0, \qquad X=G^2M^2.$$

The positive root is the bound above. This verifies the constants and shows why the extremal stationary family is a plausible minimizer. It is not a proof for general initial data: those data have no Kerr–Newman radial polynomial. The harmonic-map argument is what transports this stationary relation to the stated dynamical data class.

The matter assumption matters. Charged sources, non-Abelian fields, scalar hair, or modified gravitational theories change both the constraints and the target geometry. Their stationary solutions are not counterexamples to an Einstein–Maxwell theorem because they lie outside its matter model.

Horizon area inequalities

Related quasi-local statements bound the area of a smooth, closed, stable axisymmetric marginally outer trapped surface. In compatible geometrized normalization, a typical electrovacuum inequality is

$$A^2\ge16\pi^2\left(4G^2J^2+Q^4\right),$$

with equality at an extreme Kerr–Newman horizon cross-section. This is not the same theorem as the ADM mass inequality. It uses stability of the marginal surface, an axial symmetry intrinsic to the surface, the dominant energy condition, a nonnegative cosmological constant, and charge flux through that surface. Combining an area inequality with a Penrose inequality requires that both sets of hypotheses hold; one cannot chain them by analogy alone. The precise alternatives—a smooth stable axisymmetric MOTS in a spacetime obeying the dominant energy condition with $\Lambda\ge0$, or a smooth stable axisymmetric minimal surface in maximal data obeying the same conditions—and the equality classification are Gabach Clement, Jaramillo, and Reiris 2013, Theorem 1.1 and §3, printed pp. 2–13.

Proof architecture and equality

The proof first derives global scalar potentials from symmetry and topology. It rewrites the ADM mass lower bound as a reduced functional, fixes the singular axis boundary data determined by $J$ and $Q$, and compares with the explicit extreme Kerr–Newman harmonic map. Negative curvature of the target supplies convexity or a distance estimate. Cutoffs near the axis and ends justify the comparison despite singular boundary behavior. Equality forces coincidence with the extremal map, after which reconstruction identifies the initial data.

This architecture explains the theorem’s regularity, topology, maximality, axisymmetry, two-end, and electrovacuum hypotheses. Equality also requires saturation of the preliminary mass lower bound, not merely uniqueness of the reduced harmonic map. Extreme Kerr–Newman is an independent dimensional and equality check. A nonaxisymmetric data set is a counterexample to treating $J$ as automatically defined, not to the inequality under axial symmetry.

Scope and limitations

The bound does not say that every body with total angular momentum $J$ obeys the black-hole inequality, nor that every multiple-black-hole configuration has equality only at a single extreme Kerr–Newman slice. A one-ended complete manifold without boundary cannot carry the nonzero $J$ and source-free Maxwell charges used in this two-ended theorem. More than two ends alter the axis data and may introduce interaction terms, while an inner horizon boundary requires a different boundary-value theorem. Nonmaximal extensions require separate arguments.

The result also does not prove dynamical formation or stability of an extremal black hole. Equality is a rigidity classification inside the variational data class. Near equality may motivate a stability question, but a quantitative distance-to-Kerr–Newman theorem needs its own norm and hypotheses.

Check the limiting cases

Reduce the coupled mass inequality first to $J=0$ and then to $Q=0$. What do the equality cases represent?

Solution

For $J=0$, the square root is $Q^2$, so $G^2M^2\ge Q^2$, with equality for extreme Reissner–Nordström data. For $Q=0$, it becomes $G^2M^2\ge|GJ|$, with equality for extreme Kerr data. The two reductions verify dimensions and charge normalization. They do not remove the theorem’s completeness, two-end, no-inner-boundary, maximality, topology, symmetry, and electrovacuum hypotheses.

References

• Chruściel, Piotr T., and João Lopes Costa. “Mass, Angular-Momentum, and Charge Inequalities for Axisymmetric Initial Data.” Classical and Quantum Gravity 26 (2009): 235013. DOI. Open PDF.
• Dain, Sergio, and María Eugenia Gabach-Clement. “Geometrical Inequalities Bounding Angular Momentum and Charges in General Relativity.” Living Reviews in Relativity 21 (2018): 5. arXiv.
• Dain, Sergio. “Proof of the Angular Momentum–Mass Inequality for Axisymmetric Black Holes.” Journal of Differential Geometry 79 (2008): 33–67. Open PDF.
• Gabach Clement, María E., José Luis Jaramillo, and Martín Reiris. “Proof of the Area–Angular-Momentum–Charge Inequality for Axisymmetric Black Holes.” Classical and Quantum Gravity 30 (2013): 065017. DOI. Open PDF.
• Schoen, Richard, and Xin Zhou. “Convexity of Reduced Energy and Mass Angular Momentum Inequalities.” Annales Henri Poincaré 14 (2013): 1747–1773. DOI. Open PDF.