Conformal boundary rigidity from null geodesic travel times

Abstract: The gravitational field of a distant, isolated system is manifested by the conformally invariant Weyl tensor. Thus the conformal structure far from the system encodes the system's gravitational mass. It also encodes the causal structure, thereby linking it to the mass. For asymptotically anti-de Sitter (AdS) spacetimes, this link led to a novel positive mass theorem of Page, Surya, and the second author \cite{PSW} which did not rely on any traditional energy condition. Here we ask whether that theorem has a rigidity case. Specifically, we consider all null geodesics in an asymptotically AdS spacetime that depart from the Penrose conformal infinity, travel through spacetime, and return to conformal infinity. If all such geodesics from a given point refocus at an antipodal point at infinity, is the spacetime conformal to anti-de Sitter space? It is easy to answer the question if the asymptotically AdS spacetime either (i) obeys the null energy condition or (ii) is static, and we give simple proofs in those cases. We also answer the question in the case of globally stationary, asymptotically AdS spacetimes, by applying the theory of magnetic geodesics on the Riemannian manifold-with-boundary obtained by quotienting by the stationary Killing vector field. The question has an analogue for asymptotically flat spacetimes, which we also discuss.

• The paper demonstrates that if every null geodesic precisely refocuses at boundary antipodes, then the spacetime is conformally isometric to anti-de Sitter under the null energy condition.
• It employs a reduction to an optical manifold where magnetic geodesic analysis clarifies rigidity in static and stationary asymptotically AdS spacetimes, even without strict energy conditions.
• The approach extends to asymptotically flat spacetimes, showing that appropriate null geodesic travel data uniquely characterize Minkowski space under the NEC.

Conformal Boundary Rigidity from Null Geodesic Travel Times

Overview and Motivation

This work addresses a rigidity problem for asymptotically anti-de Sitter (AdS) spacetimes in Lorentzian geometry, specifically investigating whether knowledge of null geodesic travel times connecting boundary points determines the conformal geometry of the bulk. The main question: If every null geodesic in an asymptotically AdS spacetime, departing from a boundary point, precisely reconverges at its antipodal boundary point, is the spacetime necessarily (up to conformal isometry) anti-de Sitter? This is closely connected to positive mass results in AdS geometry, the structure of causal boundaries, and to longstanding inverse rigidity problems.

The context is inspired by a new positive mass theorem in asymptotically AdS spacetimes that does not rely on the classical energy conditions, but instead on a "holographic causality" property [PSW]. The central rigidity question is the counterpart to the classical rigidity statements about zero-mass solutions (e.g., Schwarzschild or hyperbolic space being uniquely determined in the Riemannian setting): Can AdS be characterized by the precise reconvergence of null geodesics at boundary antipodes, purely from conformal or causal data?

Main Results

The principal results fall into two categories, based on additional geometric assumptions:

1. Rigidity under the null energy condition (NEC): If the conformal class includes a representative satisfying the NEC and the aforementioned null geodesic refocusing property, then the spacetime metric is necessarily anti-de Sitter.
2. Rigidity for static and stationary asymptotically AdS spacetimes: If the spacetime admits a globally defined timelike Killing vector field (stationarity)—even without imposing the NEC—then the same rigidity result holds: the manifold is conformally isometric to AdS.

The strongest result is an affirmative answer to the rigidity question in the class of globally stationary, asymptotically AdS spacetimes, with a detailed geometric proof via quotienting out the stationary action and analyzing the induced (magnetic) geodesics on the optical manifold.

Furthermore, the authors discuss analogous questions for asymptotically flat spacetimes, showing that under NEC, such a null geodesic rigidity property also characterizes Minkowski space.

Technical Approach

Lorentzian Boundary Rigidity and Its Inverse Problem Structure

The question echoes both the Blaschke returning manifold problem and the boundary rigidity problem in Riemannian geometry, but in a Lorentzian, conformal boundary-at-infinity setup. The authors focus on spacetimes whose conformal infinities (Penrose boundaries) have precise causal structures—a property formalized via "holographic causality" and its violation via what they term the "timelike boundary Penrose property."

Null geodesics in AdS have the property that each departs a boundary point and returns precisely to the antipodal boundary point after traversing the bulk. The central question is whether this dynamical datum characterizes AdS uniquely up to conformal isometry.

Proof Under Energy Conditions

If the NEC holds for some representative in the conformal class, the Raychaudhuri equation for null congruences is employed alongside the classical avoidance of conjugate points under Ricci non-negative curvature. The complete refocusing at the antipode together with NEC forces the vanishing of both the Ricci contraction and the Weyl tensor on all null directions. A detailed polarization lemma—showing that any (0,2)-tensor vanishing along all null directions must be proportional to the metric—leads to the conclusion that the geometry is (locally) conformally flat, with the appropriate negative cosmological constant, i.e., (locally) AdS.

Rigidity for Static and Stationary Spacetimes

For static and stationary geometries—crucial in relativity—the analysis proceeds by reducing the problem to an optical (Riemannian) geometry via quotienting out the stationary Killing flow. Null geodesics project to magnetic geodesics on the so-called Fermat (optical) manifold, and the temporal separation between boundary points matches the "Mañé action" (length functional with a magnetic field term) between their projections.

If all boundary-to-boundary magnetic geodesics have the same Mañé action (equal to $\pi$), rigidity for hemispherical Riemannian manifolds (Bangert's theorem [Bangert]) and an original "magnetic rigidity" theorem proved here imply that the metric is isometric to a standard hemisphere. The geometric reduction relies on magnetic boundary rigidity, the constancy of action for magnetic geodesics, and variational properties of the action.

The stationary—but not necessarily static—case is handled by proving a magnetic analogue of Bangert's rigidity: If all boundary magnetic geodesics have fixed Mañé action, the "magnetic field" (2-form) must vanish identically, and the optical manifold must be a hemisphere.

Key technical ingredients include:

• First-variation formulas for Mañé action and analysis of the canonical scattering relation.
• Symplectic and variational arguments demonstrating that the return map for magnetic geodesics matches the canonical hemisphere's.
• Use of Santaló's formula and the structure of the unit tangent/cotangent bundle to establish global minimization properties and prevent geodesic trapping.

The Asymptotically Flat Case

In the asymptotically flat context, where conformal infinity is null rather than timelike, a similar line of reasoning under the NEC leads to the conclusion that the only such spacetime with the null geodesic antipodal property is conformally Minkowskian.

Implications and Theoretical Significance

The results are significant for several reasons:

• Conformal and Causal Determination: They demonstrate that bulk conformal geometry can be rigidly encoded in the null travel-time data of the conformal boundary, at least in rich geometric classes (NEC or stationarity). This has implications for holographic settings (e.g., AdS/CFT), where boundary data are conjectured to encode bulk geometry.
• Non-Energy Condition Rigidity: The stationary case result stands out, as it does not require classical energy conditions, only geometric stationarity and the null geodesic refocusing property, expanding the class of spacetimes for which holographic data suffices to determine the bulk.
• Inverse Problems and Integrability: The reduction to Riemannian (and magnetic) boundary rigidity deepens connections between Lorentzian inverse problems, Riemannian rigidity, and spectral geometry.
• Minkowski Characterization: The result in the asymptotically flat setting, although limited to the NEC case, points toward a broader principle tying boundary causal data to Lorentzian metric rigidity.

Future Directions

The results present several lines of exploration:

• Removal or weakening of restrictive hypotheses outside NEC or stationarity, seeking metric rigidity under weaker (e.g., purely causal) assumptions.
• Extension to settings with horizons or ergoregions, where the analysis of trapped geodesics and the Mañé action may be nontrivial.
• Investigation of the quantum or semiclassical analogues, where holographic causal data may be accessible via correlation functions.
• Development of explicit algorithms or integral geometry methods reconstructing conformal bulk geometry directly from boundary null geodesic data.
• Applications to AdS holography and the formulation of refined bulk reconstruction conjectures.

Conclusion

This work confirms that, within significant geometric classes (stationary or NEC-satisfying metrics), the precise knowledge of null geodesic crossing data at conformal infinity rigidly determines the conformal bulk geometry as anti-de Sitter. The approach uses an overview of causal geometry, variational calculus of magnetic geodesics, and rigidity theory, pushing forward the understanding of how causal boundary data constrain the interior spacetime. In both theoretical relativity and holographic contexts, these results underline the deep interplay between asymptotic causal data and global geometric rigidity.

Reference: "Conformal boundary rigidity from null geodesic travel times" (2606.28689)