Black-Hole Perturbation Theory

This calculation-led path is for graduate students and researchers entering black-hole perturbation theory. It develops the chain from perturbative gauge freedom and fixed-background fields to master equations, Kerr separation, retarded Green functions, quasinormal resonances, tails, stability, superradiance, and tidal response.

The central discipline is to define the problem before solving the equation: background, field variable, gauge, operator domain, boundary conditions, normalization, observable, and approximation order are all part of the result.

Helpful background. Black-hole geometry and mechanics supplies event and Killing horizons, charges, mechanics, rigidity, and energy extraction. Partial differential equations and Green functions supplies hyperbolic and boundary-value problems, distributions, Fourier methods, and retarded response. Exact spacetimes supplies Schwarzschild and Kerr coordinates, invariants, extensions, limits, and physical domains. Frames, coframes, and anholonomy supplies tetrad gauge and noncoordinate bases. Presymplectic potentials and currents supplies the variational language used in response and flux calculations.

Response begins with a well-posed problem

A separated radial equation is not yet a physical prediction. The same differential expression can represent retarded evolution, scattering, a bound-state problem, a quasinormal resonance, or a static response problem depending on its domain and boundary conditions. Likewise, a metric component can contain physical variation, a gauge transformation, a shift within the background family, or a constraint violation.

This route keeps those distinctions visible and requires every analytic result to survive a limiting, flux, convergence, or independent-method check.

What you will be able to do

After completing the route and its assessment work, you should be able to:

1. Choose variables, gauge, boundary conditions, and observables for a black-hole perturbation problem.
2. Derive and analyze spherical and Kerr master equations using Green-function and spectral methods.
3. Interpret quasinormal modes, tails, response, absorption, and stability with appropriate caveats.
4. Validate an analytic or numerical perturbation result through limiting, convergence, and independent checks.

Check your starting point

1. Can you distinguish a coordinate horizon from a boundary of the exterior scattering domain?
2. For a wave equation, can you state the difference among retarded data, outgoing boundary conditions, and an eigenvalue condition?
3. Can you explain why a perturbation of the metric is not automatically gauge invariant?
4. Can you translate a metric into a null or orthonormal tetrad while keeping tetrad gauge distinct from coordinate gauge?
5. Can you extract a presymplectic potential from an action variation and identify which boundary term controls the current?

Answers and review

1. A coordinate chart may fail at a regular horizon, while the exterior problem still has a geometric null boundary with a physically chosen flux condition. Review horizon and background geometry at Black-Hole Physics.
2. Retarded response is selected by causal support and initial/source data; scattering specifies incident and outgoing bases; a resonance selects complex frequencies admitting homogeneous outgoing/ingoing conditions. If the initial, boundary, singular-point, and spectral distinction is not yet secure, begin with Differential Equations for Relativity, then review causal response at PDEs and Green Functions for Gravitation.
3. Under an infinitesimal diffeomorphism generated by $\xi^\mu$, $h_{\mu\nu}$ shifts by $\nabla_\mu\xi_\nu+\nabla_\nu\xi_\mu$. Review the exact background and geometric perturbation language through the Graduate General Relativity Self-Study Path.
4. A tetrad has its own local Lorentz or null-rotation freedom in addition to coordinate freedom. Review this distinction at Frames, Coframes, and Anholonomy, then return before the Kerr phase.
5. Writing $\delta L=E\cdot\delta\phi+d\theta$ separates the equations of motion from the boundary potential; the antisymmetrized second variation of $\theta$ gives the presymplectic current. Review this construction at Presymplectic Potentials and Currents, then return before the tidal-response phase.

Route from gauge to response

The broader chapter maps are Perturbative General Relativity and Gauge Invariance, Test Fields on Black-Hole Backgrounds, Spherical Black-Hole Perturbations, Regge–Wheeler, Zerilli, and Gauge-Invariant Master Variables, Kerr Perturbations and Teukolsky Theory, Green Functions, Scattering, and Retarded Response, Quasinormal Modes, Resonances, and Pseudospectra, Late-Time Tails, Branch Cuts, and Spectral Representations, Mode, Linear, and Nonlinear Stability Interfaces, Superradiance and Environmental Instabilities, and Tidal Response, Love Numbers, Absorption, and Heating. These are optional orientation maps; they do not replace the required readings or problems below.

Phase 1 — Linearized geometry and a scalar prototype

1. Linearizing Geometry and the Einstein Tensor — derive the first variation before imposing gauge or symmetry.
2. Infinitesimal Diffeomorphisms and Gauge Transformations — distinguish physical variation from infinitesimal coordinate freedom.
3. Gauge-Invariant Observables and Master Variables — identify what the reduced variable represents and what must be reconstructed.
4. Scalar Fields and Effective Potentials — establish separation, tortoise coordinates, flux conditions, and effective-potential reasoning in the cleanest sector.

Phase 2 — Spherical decomposition and radiative master equations

5. Tensor Harmonics and Parity Decomposition — decompose by harmonic number and parity without losing low-mode exceptions.
6. Low Multipoles, Parameter Shifts, and Gauge Modes — separate mass, angular-momentum, and gauge shifts from radiation.
7. The Regge–Wheeler Equation — derive the odd-parity master problem and its physical boundary data.
8. The Zerilli Equation — derive the even-parity invariant and compare its potential with the axial sector.
9. Metric Reconstruction and Radiative Observables — translate master variables into metric and asymptotic radiation while retaining reconstruction limits.

Phase 3 — Kerr and Teukolsky separation

10. Kinnersley Tetrads and Newman–Penrose Curvature — fix tetrad normalization, spin weight, curvature scalar, and residual gauge.
11. Derivation of the Teukolsky Equation — expose the decoupling assumptions and the operator being separated.
12. Radial Teukolsky Equation and Physical Solutions — derive horizon and infinity bases with their flux conventions.
13. Angular Teukolsky Equation and Separation Constants — determine the angular eigenproblem and track the spherical limit.

Phase 4 — Green functions and scattering

14. Mode Green Functions and Wronskians — assemble the radial inverse from two homogeneous bases and their Wronskian.
15. In, Up, Out, and Down Scattering Bases — keep basis choice, normalization, and physical incidence distinct.
16. Reflection, Transmission, and Absorption — derive flux-weighted coefficients rather than comparing raw amplitudes.
17. Inverse Transforms and Causal Time-Domain Response — recover causal evolution from contours, poles, cuts, and prompt response.

Phase 5 — Resonances, numerical spectra, and tails

18. Quasinormal Boundary Conditions and Resonances — define resonances by the operator domain rather than by a damped sinusoid alone.
19. Frobenius Series and Continued Fractions — turn endpoint behavior into a convergent spectral condition.
20. Spectral Collocation and Direct Eigenvalue Methods — compute spectra with resolution, residual, and spurious-mode checks.
21. Pseudospectra and Nonnormal Sensitivity — distinguish eigenvalue stability from resolvent amplification and norm dependence.
22. Completeness, Excitation, and Ringdown Content — separate pole locations from excitation and finite-time signal content.
23. Price-Law Tails in Schwarzschild — connect low-frequency scattering to late-time power-law decay.
24. Branch-Cut Contributions to Retarded Response — evaluate the non-pole part of the inverse transform and its domain of validity.

Phase 6 — Stability and rotational amplification

25. Whiting Mode Stability of Kerr — state the separated mode result with its field sector and boundary hypotheses.
26. Boundedness and Decay of Linear Fields — move from mode exclusion to initial-data estimates in explicit norms.
27. From Linear Decay to Nonlinear Stability — identify the additional gauge, constraint, bootstrap, and asymptotic control a nonlinear theorem needs.
28. The Superradiant Frequency Condition — derive the sign of horizon energy flux from the horizon generator.
29. Scattering Amplification Factors — normalize and check amplification without confusing amplitude and energy gain.

Phase 7 — Conservative and dissipative response

30. Relativistic Love Numbers and Gauge Invariance — define the tidal and response multipoles together with gauge and matching choices.
31. Static Love Numbers of Schwarzschild and Kerr in Four-Dimensional Vacuum GR — state the vanishing result with its dimensional, field-content, and static-limit scope.
32. Horizon Absorption and Tidal Heating — connect dissipative response to horizon flux and restore the physical scaling.

A representative reasoning pattern: flux before interpretation

For a real-frequency massless field on Schwarzschild, a separated master variable often satisfies

$$\frac{d^2\psi}{dr_*^2}+ \left[\omega^2-V(r)\right]\psi=0,$$

with real $V$ and tortoise coordinate $r_*$. Normalize a unit incident wave at infinity:

$$\psi\sim e^{-i\omega r_*}+\mathcal R e^{+i\omega r_*} \quad (r_*\to+\infty),$$

and a future-horizon ingoing wave

$$\psi\sim \mathcal T e^{-i\omega r_*} \quad (r_*\to-\infty).$$

For real $V$, the Wronskian $W=\psi^*\psi'-\psi\psi'^*$ is constant. Evaluating it at both ends gives

$$\lvert\mathcal R\rvert^2+\lvert\mathcal T\rvert^2=1$$

when the asymptotic real wavenumbers and flux normalizations agree. This is a strong numerical and analytic check: a missing sign, inaccurate integration, or inconsistent amplitude convention breaks the budget.

The formula is not universal. Kerr horizon flux depends on $\omega-m\Omega_H$ and can be negative in the exterior energy balance, producing superradiance. Massive fields or other asymptotics can have unequal wavenumbers and require velocity or flux weights. Conservation survives; the unweighted expression does not.

Check your understanding

Checkpoint 1. Same equation, different problem

Explain how one radial differential equation can define scattering coefficients, a retarded Green function, and quasinormal frequencies without contradiction.

Conceptual hint

The operator domain is part of the problem.

Setup hint

Compare source and initial data with homogeneous boundary conditions.

Next-step hint

Complex-frequency poles arise after analytic continuation.

Solution

Scattering uses a real-frequency incident basis and measures reflected and transmitted flux. A retarded Green function combines two homogeneous solutions with causal support and a Wronskian denominator to respond to sources or initial data. Quasinormal frequencies are complex values at which a homogeneous solution satisfies ingoing-horizon and outgoing-infinity conditions simultaneously, appearing as poles of an analytically continued response under suitable assumptions. The differential expression is shared, but domain, data, normalization, and requested observable differ.

Checkpoint 2. What “stable” means

A calculation finds no modes with $\operatorname{Im}\omega\gt0$ using $e^{-i\omega t}$. What has been shown, and what remains open?

Conceptual hint

Mode stability is one rung in a hierarchy.

Setup hint

Ask about completeness, nonmodal growth, norms, constraints, and nonlinear terms.

Next-step hint

A pseudospectrum concerns sensitivity even when eigenvalues are stable.

Solution

Within the searched field sector, operator domain, and boundary conditions, no exponentially growing separated mode was found. This does not prove boundedness or decay for general initial data, completeness of a mode expansion, control of transient nonnormal amplification, linearized Einstein stability after gauge and reconstruction, or nonlinear orbital stability. Each stronger conclusion needs its own estimates, data class, and norms.

Transfer task — Change the outer boundary

Take the Schwarzschild scattering problem and replace asymptotically flat infinity by a reflecting anti-de Sitter boundary. Predict which ingredients of the calculation remain and which physical conclusions change.

Solution outline

The local master equation, horizon regularity, Wronskian methods, and need for a specified field variable remain. The outer domain and boundary condition change: there is no asymptotic plane-wave incident/outgoing basis of the same kind, so the flat-space reflection/transmission interpretation and its unit-flux formula do not transfer unchanged. Reflecting boundaries can return energy, alter the spectrum from radiative resonances toward boundary-condition-dependent normal or quasinormal modes, and combine with rotational amplification to create instabilities. Causal response and stability must therefore be rebuilt for the selected AdS boundary condition.

Capstone — A reproducible response calculation

Choose one problem: Schwarzschild scalar scattering, a Regge–Wheeler quasinormal mode, a Kerr angular eigenvalue, or a static tidal response. Produce a compact research-style record that:

• defines background, field, harmonic sector, gauge or invariant variable, domain, and Fourier convention;
• states initial or boundary data and the target observable;
• derives the reduced equation and asymptotic behavior;
• computes the result with a convergence or approximation sequence;
• verifies one conserved flux or Wronskian relation, one controlled limit, and one independent formulation or method;
• states a failure regime and a conclusion no stronger than the evidence.

Criterion	What a complete solution shows
Problem definition	Variable, gauge, domain, data, normalization, and observable are complete.
Analysis	Reduction and asymptotics are derived rather than merely quoted.
Verification	Convergence, flux or residual, limit, and independent check agree.
Interpretation	Resonance, response, stability, and physical signal are not conflated.

Completion and onward routes

Complete all thirty-two core readings in order, both checkpoints, the AdS transfer task, prediction-and-check records for the calculations, and a capstone satisfying the rubric. The optional chapter maps and any preparation reviews support the route but do not add completion steps.

Continue to Compact Binaries and Gravitational Waves to insert response into orbital and waveform problems, Computational Relativity for production numerical methods, or Mathematical Relativity for theorem-level stability and decay.