Second-Order Perturbations and Mode Coupling

Second-order perturbation theory keeps the first nonlinear backreaction of a small disturbance while retaining a linear differential operator for the new unknown. It predicts harmonic mixing, parameter drift, quadratic radiation, and possible secular growth. Control requires an appropriate field norm, compatible constraints, and meaningful invariant diagnostics; seeing a sum frequency in one metric component is not, by itself, evidence for a physical nonlinear mode.

Helpful background. Black-Hole Geometry and Mechanics fixes the background and horizon data used in applications; PDEs and Green Functions for Gravitation supplies the differential-operator viewpoint; Formulations of General Relativity separates covariant equations, constraints, and evolution; Presymplectic Potentials and Currents prepares the variational language used later in this chapter.

The sourced second-order equation

Fix the normalization

$$g_{\mu\nu}(\varepsilon) =\bar g_{\mu\nu}+\varepsilon h_{\mu\nu} +\frac12\varepsilon^2 k_{\mu\nu}+O(\varepsilon^3),$$

and expand the complete metric–matter equations $E[g,\psi]=0$ as follows. In the next two formulas, $h$ denotes the complete first-order field variation $(h_{\mu\nu},\delta\psi)$ and $k$ the corresponding second-order pair; $E^{(1)}$ includes the matter equations as well as the metric equation:

$$E=\bar E+\varepsilon E^{(1)}[h] +\frac12\varepsilon^2 \left(E^{(1)}[k]+E^{(2)}[h,h]\right)+O(\varepsilon^3).$$

On a background and first-order solution, the second-order field satisfies

$$E^{(1)}[k]=-E^{(2)}[h,h].$$

The left side is the same linearized operator encountered at first order. The right side is a quadratic effective source built from $h$ and its derivatives, plus matter terms when present. It is not a freely specifiable stress–energy tensor. The second variation of the Bianchi identity imposes an integrability condition that the source satisfies only when the first-order equations and constraints hold.

Regularity is a discriminating check. A source that diverges at the future horizon or infinity may reflect a poor gauge or master-variable definition, not a physical divergence. For radiative Schwarzschild sectors, a complete gauge-invariant reduction, source regularization, reconstruction, and flux at null infinity are given in Brizuela, Martín-García, and Tiglio 2009, §§II–V. The construction assumes radiative $\ell\ge2$ sectors; low multipoles and parameter shifts require separate equations.

Gauge transformations beyond first order

For the pullback convention used in this chapter,

$$h' = h+\mathcal L_{\xi_1}\bar g,$$ $$k' = k+\mathcal L_{\xi_2}\bar g +2\mathcal L_{\xi_1}h +\mathcal L_{\xi_1}^2\bar g.$$

The mixed and squared Lie derivatives mean that a first-order gauge choice feeds the second-order field. A quantity invariant at first order can acquire a second-order gauge term unless its nonlinear completion is constructed. Changing the background parameters at order $\varepsilon^2$ also redistributes content between $k$ and a shifted background; for ringdown this includes the mass and spin absorbed through the horizon.

The composition law and the dependence on identification maps are derived in Bruni et al. 1997, arXiv v2, p. 6, Eq. (2.13), and pp. 13–14, Eqs. (4.4)–(4.12), PDF. In practice, a second-order observable must state its first- and second-order gauge vectors, parameter-renormalization convention, tetrad or harmonic gauge, and boundary frame. Agreement of two coordinate components is not the relevant check; agreement of a reconstructed invariant flux or relational quantity is.

Harmonic selection and resonant forcing

Products of first-order spherical harmonics generate constrained daughter sectors. Two parent modes with labels $(\ell_1,m_1)$ and $(\ell_2,m_2)$ can couple only when the angular-momentum triangle condition is satisfied, $m=m_1+m_2$ for the complex product, and the parity product matches the daughter. Because a real field includes complex-conjugate modes, both sum and difference azimuthal numbers and frequencies occur.

If first-order time dependence is $e^{-i\omega_1t}$ and $e^{-i\omega_2t}$, the quadratic source contains $e^{-i(\omega_1+\omega_2)t}$. The product of the first mode with the complex conjugate of the second instead gives $e^{-i(\omega_1-\omega_2^*)t}$, not $e^{-i(\omega_1-\omega_2)t}$ for complex frequencies. For a single damped parent $\omega=\omega_R-i\gamma$ with $\gamma\gt0$, this conjugate product is the nonoscillatory but decaying factor $e^{-2\gamma t}$, not a constant.

One angular example is completely explicit. For unit-normalized complex spherical harmonics with the standard Condon–Shortley convention,

$$Y_{22}^2=\sqrt{\frac5{14\pi}}\,Y_{44}.$$

This follows by inserting their $\sin^2\theta\,e^{2i\phi}$ and $\sin^4\theta\,e^{4i\phi}$ forms. It exhibits an allowed daughter whose first-order amplitude may be exactly zero. The coefficient shown is for the scalar harmonic product; derivative and tensor contractions in Einstein’s equation determine the actual gravitational coupling.

The daughter solves an inhomogeneous equation, so its response contains both the source frequencies and the homogeneous spectrum allowed by its boundary conditions. A forcing frequency near a homogeneous resonance can produce a large response. In the elementary model

$$\ddot q+\omega_0^2q=F\cos(\omega_0t),\qquad \omega_0\gt0,$$

the solution with $q(0)=\dot q(0)=0$ is

$$q(t)=\frac{F}{2\omega_0}\,t\sin(\omega_0t).$$

Twice differentiating verifies the forcing and both initial conditions. If the perturbation is $\varepsilon A\cos(\omega_0t)+\varepsilon^2q(t)/2$ with $A\ne0$, the ratio of the second-order envelope to the first is $\varepsilon|F|t/(4\omega_0|A|)$. It is an envelope comparison, not division by the first-order signal at its zeros. Once this ratio is order unity the naive fixed-background expansion is no longer uniformly ordered. This secular failure is not necessarily a physical instability: multiple-scale evolution, a shifted frequency, or a slowly varying background may resum the growing term.

Quasinormal modes add a subtlety: they are resonances with radiative boundary conditions, not an ordinary orthonormal basis on a Cauchy slice. Assigning a daughter amplitude depends on a projection or residue prescription, source regularization, background parameters, and the time interval over which a mode description is useful.

Second-order results through 2026-08-12

The equation hierarchy and second-order gauge law are established. The unresolved frontier is how universally particular nonlinear components and amplitude ratios describe generic black-hole merger ringdown. The evidence available through 2026-08-12 depends on background, observable, initial-data class, extraction surface, and fitting norm.

Claim	Demonstrated regime	Evidence	Limitation
Gauge-invariant quadratic source	Schwarzschild, radiative ℓ≥2 sectors	Covariant second-order master equations and reconstruction	Low multipoles, generic Kerr, and global gauges need separate treatment
Standard linear frequencies dominate studied daughters	Selected Schwarzschild parity couplings and time-domain initial data	High-accuracy first- and second-order evolutions	Does not exclude quadratic source frequencies in other observables or windows
Quadratic modes are visible	Head-on-merger horizon shear in selected multipoles	Amplitudes scale quadratically with parent amplitudes	Horizon evidence is not yet a universal asymptotic waveform prescription
Some nonlinearity ratios can be approximated accurately	Selected Schwarzschild channels at the horizon and infinity	Matched asymptotic and steepest-descent estimates compared with simulations	Accuracy is strongly channel-dependent and fails in a documented axisymmetric case
Kerr quadratic source can be formulated	Vacuum second-order Teukolsky equation in specified gauges	Explicit nonlinear source and regular-gauge analysis	Precision reconstruction, completion, and generic amplitude extraction remain active work

The second row records the contrast found by Pazos et al. 2010, §§II–V: for the Schwarzschild couplings and initial data studied, fitted second-order decay frequencies matched the linear spectrum rather than earlier doubled- frequency expectations. The third row reflects nonlinear horizon simulations that identify quadratic components with quadratic amplitude scaling Khera et al. 2023, pp. 2–5. These results are credible, but they answer different questions and use different observables.

A revised 2026 preprint finds good agreement for the $(2,2)\times(2,2)\to(4,4)$ Schwarzschild nonlinearity ratio and relative insensitivity to several source regularizations, but only a crude estimate in the $(2,0)\times(2,0)\to(2,0)$ channel because its approximation and source- support assumptions fail Singh and Suneeta 2026, arXiv v3, §§9–12. At this evidence cutoff it is a version-3 preprint marked “to appear” in Physical Review D, not a journal article with final bibliographic data. Its channel dependence is evidence against treating one reported amplitude ratio as universal.

For rotating black holes, an explicit second-order Teukolsky source and a comparison of formulations provide essential infrastructure Spiers, Pound, and Moxon 2023, §§II–V. They do not settle a unique nonlinear ringdown basis or a generic numerical-relativity amplitude map. Competing descriptions can emphasize forced sum frequencies, homogeneous linear resonances, or background drift; which dominates depends on the parameter regime, observable, projection, and time window. A review is triggered by a gauge-controlled generic-Kerr calculation matched across second-order perturbation theory and convergent numerical relativity, or by a new theorem fixing the relevant resonance expansion and norm.

When the expansion is controlled

Choose a domain $D$, an admissible gauge with a regular frame, and a positive field norm $\|\cdot\|_D$ that controls the derivatives required by the equations. For $\mathcal F=\mathcal F^{(0)}+\varepsilon\mathcal F^{(1)} +\varepsilon^2\mathcal F^{(2)}/2+\cdots$, a useful sufficient ordering test, when the first-order norm is nonzero, is

$$\frac{\varepsilon^2}{2}\|\mathcal F^{(2)}\|_D \ll \varepsilon\|\mathcal F^{(1)}\|_D \ll \mathcal F_{\mathrm{ref}},$$

where $\varepsilon\gt0$ and $\mathcal F_{\mathrm{ref}}$ is a nonzero physical scale with the same units as the norm. It need not be a pointwise background component. The norm and the time interval are part of the claim: small fields without control of the required derivatives need not produce small nonlinear terms.

Invariant diagnostics should corroborate this control, but they need not all start at first order. If an observable’s first nonzero contribution is order $q$, compare its higher-order corrections with that leading term in an appropriate norm, or with an independently specified physical scale. Radiative energy flux can start at second order; the $(4,4)$ daughter above can likewise be nonzero although its first-order coefficient vanishes. Neither fact violates perturbation theory. A demand that every second-order observable be smaller than its identically zero first-order coefficient would incorrectly exclude both examples.

Also require small constraint residuals at each order, finite horizon and infinity fluxes, stability under admissible gauge and extraction choices, and convergence when truncation or numerical resolution is varied. These are consistency and accuracy diagnostics, not by themselves a remainder theorem comparing the truncated series with an exact nonlinear solution.

Breakdown is signaled by loss of this hierarchy, a secular term becoming as large as the first-order contribution, a resonance denominator invalidating the assumed scaling, source singularities that cannot be removed within the admitted gauge class, or sensitivity to neglected higher orders. None of these signs alone determines the nonlinear endpoint.

Exact background construction remains in Exact Solutions of Einstein’s Equations. Full finite-amplitude evolution, merger dynamics, saturation, and endpoints remain in Black-Hole Dynamics. Second-order theory is a controlled bridge only within its stated data class, boundary conditions, frequency range, gauge, and time interval.

Exercise: diagnose a claimed quadratic mode

A simulation reports a component proportional to $A^2e^{-2i\omega t}$ after seeding a first-order mode of amplitude $A$. List the checks required before interpreting it as a physical second-order mode.

Solution

First replace the coordinate component by a second-order gauge-invariant curvature, flux, or relational observable and specify the tetrad and boundary frame. Fix whether the order-$A^2$ mass and spin changes are included in the background or in the perturbation. Verify that the daughter harmonic obeys the angular and parity selection rules and that its amplitude scales as $A^2$ over a genuine small-amplitude sequence.

Next solve or evaluate the sourced second-order equation with regular horizon and infinity behavior, checking constraints and source regularization. Separate a response at the forcing frequency $2\omega$ from homogeneous resonances by varying initial data and the fitting window; state the projection or residue prescription. Demonstrate numerical resolution and extraction-radius convergence, and compare with an independent gauge or formulation. Finally, check field-norm ordering and compare each invariant with its first nonzero order or a nonzero physical scale throughout the claimed interval. Do not compare a newly excited daughter with an identically zero first-order amplitude in that same daughter sector. Without these steps, the observation may be a gauge term, background drift, transient forcing, or fit degeneracy rather than a robust nonlinear mode.

References

• Brizuela, David, José M. Martín-García, and Manuel Tiglio. “A Complete Gauge-Invariant Formalism for Arbitrary Second-Order Perturbations of a Schwarzschild Black Hole.” Physical Review D 80 (2009): 024021. DOI. Open PDF.
• Bruni, Marco, Sabino Matarrese, Silvia Mollerach, and Sebastiano Sonego. “Perturbations of Spacetime: Gauge Transformations and Gauge Invariance at Second Order and Beyond.” Classical and Quantum Gravity 14 (1997): 2585–2606. DOI. Open PDF, arXiv:gr-qc/9609040v2.
• Khera, Neev, Ariadna Ribes Metidieri, Béatrice Bonga, Xisco Jiménez Forteza, Badri Krishnan, Eric Poisson, Daniel Pook-Kolb, Erik Schnetter, and Huan Yang. “Nonlinear Ringdown at the Black Hole Horizon.” Physical Review Letters 131 (2023): 231401. DOI. Open PDF.
• Pazos, Enrique, David Brizuela, José M. Martín-García, and Manuel Tiglio. “Mode Coupling of Schwarzschild Perturbations: Ringdown Frequencies.” Physical Review D 82 (2010): 104028. DOI. Open PDF.
• Singh, Jasveer, and Vardarajan Suneeta. “Computing Nonlinearity Ratios Using Second Order Black Hole Perturbation Theory.” arXiv preprint, version 3 (2026), marked to appear in Physical Review D. arXiv:2512.00943v3.
• Spiers, Andrew, Adam Pound, and Jordan Moxon. “Second-Order Teukolsky Formalism in Kerr Spacetime: Formulation and Nonlinear Source.” Physical Review D 108 (2023): 064002. DOI. Open PDF.