Anti-de Sitter Confinement and Reflections

Anti-de Sitter confinement is a causal and boundary-value statement, not a Newtonian force pulling radiation inward. The conformal boundary is timelike, signals reach it in finite time, and an evolution is not determined until admissible boundary data are supplied. With reflecting data the bulk behaves like a gravitational cavity: energy cannot escape, so successive refocusing events can accumulate nonlinear effects that asymptotically flat dispersal would carry away.

Helpful background. Black-Hole Geometry and Mechanics supplies horizon notions; Black-Hole Perturbation Theory supplies mode and flux methods; Formulations of General Relativity supplies initial-boundary formulations; Exact Solutions of Einstein’s Equations supplies global AdS coordinates; Causality and Global Structure supplies conformal completion; and Computational Relativity supplies long-time evolution checks.

The timelike conformal boundary

In four spacetime dimensions, set $\Lambda=-3/L^2$. Global anti-de Sitter spacetime can be written

$$ds^2=\frac{L^2}{\cos^2x} \left(-dt^2+dx^2+\sin^2x\,d\Omega^2\right), \qquad 0\le x\lt \frac{\pi}{2},$$

where $t$ and $x$ are dimensionless. Multiplying by the conformal factor $\cos^2x/L^2$ places the boundary at $x=\pi/2$. Its induced conformal metric contains $-dt^2+d\Omega^2$, so the boundary is timelike.

Radial null rays satisfy $dt=\pm dx$. A ray travels from the center to the boundary in coordinate time $\pi/2$ and returns after $\pi$; the proper time of the central observer is $\pi L$. This finite return time is the basic kinematic fact behind repeated focusing.

The same fact explains why AdS is not globally hyperbolic: causal curves can enter from its conformal boundary without meeting a chosen earlier spacelike slice. Boundary conditions do not turn that slice into a Cauchy surface of the boundaryless AdS spacetime; they complete an initial-boundary-value problem. No statement that “AdS is stable” or “AdS collapses” is complete without the field content, boundary condition, dimension, and norm in which perturbations are small.

Boundary conditions complete the dynamics

For a linear wave reduced on a static background, time evolution can be organized through a spatial operator $A$. A positive self-adjoint extension of $A$ gives a deterministic, energy-conserving evolution. Distinct extensions correspond, when they exist, to different boundary conditions. Ishibashi and Wald derive this principle and reduce scalar, electromagnetic, and gravitational perturbations of AdS to the relevant one-dimensional operators in Ishibashi and Wald 2004, §1, pp. 2–5, and §§2.3–2.5, pp. 13–24, PDF.

The familiar labels require care.

• Reflecting Dirichlet data fix the leading boundary value and give zero normal energy flux for the standard massless-scalar problem.
• Neumann or Robin data relate leading and subleading falloffs. They are admissible only in the mass, dimension, and positivity ranges for which the corresponding extension exists.
• Dissipative data allow outward flux. They define an open system rather than the conservative cavity used in the classic instability experiment.
• Driven data inject energy through a prescribed time-dependent boundary source. Their conserved quantities and censorship question differ from those of isolated reflecting data.

For electromagnetic and gravitational sectors, the admissible choices depend on dimension and mode type; a scalar-field slogan cannot be transferred unchanged. The detailed classifications and their positivity qualifications are given in Ishibashi and Wald 2004, §3.3, Theorem 3.2, pp. 32–38, PDF.

If $E(t)$ denotes the energy in a bulk region, its balance law has the schematic form

$$\frac{dE}{dt}=-\int_{\partial\Sigma_t}\mathcal F\,dS.$$

Reflecting conditions impose vanishing integrated flux $\mathcal F$ at conformal infinity. Dissipative conditions make the right-hand side negative, while a source can give either sign. This one equation separates confinement from geometry alone: the timelike boundary makes return possible, but the boundary condition decides whether the signal actually returns with conserved energy.

Resonant normal modes and refocusing

For a massless scalar with the standard reflecting falloff in global AdS$_4$, the physical frequencies, conjugate to central proper time $T=Lt$, have

$$\omega_{j\ell}L=3+\ell+2j, \qquad j=0,1,2,\ldots .$$

Equivalently, frequencies conjugate to dimensionless $t$ are $\widehat\omega_{j\ell}=L\omega_{j\ell}$. For spherical fields ($\ell=0$), the spatial eigenvalue problem is

$$\mathcal L e_j=-\frac{1}{\tan^2x} \partial_x\!\left(\tan^2x\,\partial_xe_j\right) =\widehat\omega_j^2 e_j, \qquad \widehat\omega_j=3+2j,$$

with a regular center, $e_j=O(\cos^3x)$ at the boundary, and inner product $\langle f,g\rangle=\int_0^{\pi/2}\overline f g\tan^2x\,dx$. For example, direct substitution gives $\mathcal L(\cos^3x)=9\cos^3x$, verifying the lowest frequency. The equally spaced spectrum exactly satisfies many resonance conditions of the form

$$\omega_i+\omega_j=\omega_k+\omega_l$$

Linear stability still follows: every frequency is real and positive. The issue is nonlinear. Weak mode couplings can add coherently over many reflections, moving energy between modes on a time much longer than one crossing time.

The standard spherical Einstein–massless-scalar model makes this structure explicit. Its metric ansatz, hyperbolic wave equations, radial constraints, regular-center expansion, and reflecting asymptotics appear in Bizoń and Rostworowski 2011, Eqs. (1)–(8), pp. 1–2. Their boundary expansion fixes a no-flux evolution for that particular smooth finite-mass problem. It does not establish uniqueness of boundary conditions for every field or mass.

Repeated focusing is also not synonymous with black-hole formation. Perturbative and numerical constructions of time-periodic solutions, and long-lived neighborhoods around them, support regular alternatives Maliborski and Rostworowski 2013, perturbative construction and numerical results, pp. 2–5, PDF. Their finite-time stability tests do not prove global regularity of every nearby datum. The confinement mechanism says why nonlinear interactions get repeated opportunities, not what every initial datum must eventually do; Turbulent Instability of Anti-de Sitter separates the relevant claims.

Independent checks and limitations

Three checks keep the comparison honest.

1. The conformal metric must make $x=\pi/2$ timelike; changing the sign of $dt^2$ destroys the causal conclusion.
2. A claimed reflecting evolution must close its energy budget at the boundary, not merely keep the field numerically small there.
3. A mode-cascade claim must be compared with an altered spectrum or altered boundary flux. Otherwise one has not isolated resonance from generic numerical drift.

The discussion is four-dimensional and uses global AdS. Planar AdS has a different spatial geometry, and an AdS black hole absorbs radiation at its horizon even when infinity reflects it. Those systems need their own flux budgets.

Exercise: compute the return time

For a radial light ray emitted at $t=0$, $x=0$, find the first time it returns to the center after one reflection at conformal infinity. Restore dimensions.

Solution

Radial null propagation gives $dt=dx$ on the outgoing leg. Reaching $x=\pi/2$ therefore takes $\Delta t=\pi/2$. The ingoing leg has the same duration, so the first return occurs at dimensionless time $t=\pi$. At $x=0$, the metric is $ds^2=-L^2dt^2$, hence the central observer measures proper time $\Delta\tau=\pi L$. The result depends only on the global AdS causal geometry; whether the ray returns with the same amplitude depends on the boundary condition.

References

• Bizoń, Piotr, and Andrzej Rostworowski. “On Weakly Turbulent Instability of Anti-de Sitter Space.” Physical Review Letters 107 (2011): 031102. doi:10.1103/PhysRevLett.107.031102. Open PDF.
• Ishibashi, Akihiro, and Robert M. Wald. “Dynamics in Non-Globally-Hyperbolic Static Spacetimes III: Anti-de Sitter Spacetime.” Classical and Quantum Gravity 21 (2004): 2981–3014. doi:10.1088/0264-9381/21/12/012. Open PDF.
• Maliborski, Maciej, and Andrzej Rostworowski. “Time-Periodic Solutions in an Einstein AdS–Massless-Scalar-Field System.” Physical Review Letters 111 (2013): 051102. doi:10.1103/PhysRevLett.111.051102. Open PDF.