Lindblad Resonances in Orbital Dynamics

• Lindblad resonances are frequency commensurabilities between orbital motion and a rotating perturbation that trigger angular momentum exchange and shape disc and galaxy structures.
• They are identified through relations among mean motion, epicyclic frequency, and pattern speed, which govern phenomena such as eccentricity growth and orbit trapping.
• Analyses across contexts—from galactic bars to protoplanetary discs—highlight their role in resonant torque theory and the organization of complex orbital dynamics.

Lindblad resonances are frequency commensurabilities between orbital motion and a rotating non-axisymmetric perturbation. In practice they are identified through relations among mean motion or azimuthal frequency, radial or epicyclic frequency, and a pattern speed, and they govern angular-momentum exchange, eccentricity growth, orbit trapping, wave excitation, and the organization of structure in rings, discs, and galaxies. The same underlying idea appears in several notational conventions: sectoral resonances around rotating non-axisymmetric bodies, inner and outer Lindblad resonances of bars and spirals in galactic dynamics, eccentric Lindblad resonances in satellite problems, and torque-bearing Lindblad resonances in gaseous and relativistic discs (Sicardy, 2020, Sellwood, 2010, Hirata, 2010).

1. Formal definitions and notational conventions

The defining relation depends on the dynamical setting. Around a rotating non-axisymmetric body with spin rate $\Omega$, sectoral resonances satisfy

$j\kappa = m(n-\Omega),$

with $\kappa$ the epicyclic frequency, $n$ the mean motion, and integers $j>0$ and $m$; Lindblad resonances are the first-order sectoral cases $j=1$, so that

$\kappa = m(n-\Omega).$

In that notation, the corresponding location is $n/\Omega \approx m/(m-1)$ (Sicardy, 2020).

In galactic dynamics the same phenomenon is usually written in action-angle or frequency form as

$m(\Omega_p-\Omega_\phi)=l\Omega_R,$

where $j\kappa = m(n-\Omega),$0 is the pattern speed, $j\kappa = m(n-\Omega),$1 the azimuthal frequency, and $j\kappa = m(n-\Omega),$2 the radial frequency. Here $j\kappa = m(n-\Omega),$3 denotes the inner Lindblad resonance (ILR), $j\kappa = m(n-\Omega),$4 the outer Lindblad resonance (OLR), and $j\kappa = m(n-\Omega),$5 corotation (Sellwood, 2010). In rotation-curve analyses of barred galaxies this becomes

$j\kappa = m(n-\Omega),$6

for ILR and OLR, respectively (Ruiz-García et al., 2024).

Context	Resonance condition	Conventional identification
Rotating non-axisymmetric body	$j\kappa = m(n-\Omega),$7	Lindblad resonance: $j\kappa = m(n-\Omega),$8 (Sicardy, 2020)
Galactic bar or spiral	$j\kappa = m(n-\Omega),$9	ILR: $\kappa$0, OLR: $\kappa$1 (Sellwood, 2010)
Bar diagnostics from rotation curves	$\kappa$2	Minus sign: ILR; plus sign: OLR (Ruiz-García et al., 2024)
Relativistic disc around a perturber	$\kappa$3	Inner and outer Lindblad resonances in strong gravity (Hirata, 2010)

These forms are not different phenomena so much as different reductions of the same commensurability structure. In all cases, the resonance selects orbits for which the response to a rotating perturbation accumulates coherently over many cycles.

2. Periodic-orbit structure and phase-space geometry

For sectoral resonances in a rotating frame, the periodic-orbit kinematics depend only on the irreducible pair $\kappa$4, obtained after removing the greatest common divisor of $\kappa$5. The periodic streamline can be written as

$\kappa$6

and for Lindblad resonances this reduces to

$\kappa$7

The orbit then has $\kappa$8 braid, $\kappa$9 identical sectors, and

$n$0

self-crossing points. A central structural result is therefore that first-order Lindblad resonances are free of self-crossings, whereas higher-order sectoral resonances generically self-intersect (Sicardy, 2020).

This absence of self-crossing is dynamically consequential. In the same analysis, higher-order self-crossings are linked to singularities in hydrodynamic equations for collisional disks or rings, whereas the $n$1 case avoids that complication. The resonance is first order in eccentricity, $n$2, and at such a resonance the particle executes one synodic cycle around the body per $n$3 radial oscillations, exchanging energy and angular momentum with the rotating mass distribution and exciting eccentricity in the process (Sicardy, 2020).

The near-circular epicyclic picture is not the only geometrical description. For eccentric galactic orbits, the apsidal precession frequency increases with eccentricity, so a given pattern speed does not correspond to a single resonant radius but to a continuous family of radii, each tied to a specific eccentricity. This is the basis of the “Lindblad Zone,” defined through

$n$4

in contrast to the epicyclic form $n$5. The p-ellipse approximation is used to track these eccentric resonant orbits and to show how ensembles of such orbits can support bars and spirals without winding up in the pattern frame (Struck, 2015). This suggests that the familiar isolated-resonance radius is the small-eccentricity limit of a broader eccentric-orbit construction.

3. Satellite, ring, and planetary-system applications

In the elliptic planar restricted three-body problem near a first-order mean-motion resonance $n$6, two resonant angles appear: $n$7 When the symmetry between them is broken in the restricted problem, they define the Lindblad Eccentric Resonance (LER) and Corotation Eccentric Resonance (CER). The LER primarily excites eccentricity with only modest change in semimajor axis, whereas the CER mainly alters semimajor axis. Their joint dynamics are summarized by the CoraLin model, controlled by a distance parameter $n$8 between the CER and LER centers and a forcing parameter $n$9. Three regimes are identified: $j>0$0 gives an integrable system, $j>0$1 of order unity produces prominent chaotic regions, and $j>0$2 yields regular behavior described qualitatively by adiabatic invariants. This framework is applied to Aegaeon, Methone, and Anthe, which are dynamically linked to Mimas (Moutamid et al., 2013).

Ring dynamics around a homogeneous triaxial ellipsoid provide a complementary Hamiltonian setting. Using epicyclic variables and resonant normal forms, a recent perturbative and bifurcation analysis verified the non-degeneracy of the normal form and hence the existence of invariant KAM tori, which confine trajectories in phase space near corotation and Lindblad resonances. In the nearly spherical and highly aspherical test cases studied there, the 1:3 resonance shows no evidence of bifurcations for relevant values of the eccentricity, and the analysis therefore supports a greater probability of selecting the $j>0$3 resonance than corotation or $j>0$4 (Celletti et al., 21 Jul 2025).

Lindblad resonances have also been invoked in satellite migration driven by planetary oscillation modes. For Neptune’s inner moons, direct numerical simulations indicate that the present Naiad–Thalassa 73:69 mean-motion resonance is unstable on Myr timescales under equilibrium tidal evolution because of perturbations from Despina. A proposed resolution is resonant-lock tides in which Thalassa and Despina are locked to Lindblad resonances with Neptune’s internal oscillation modes; low-order $j>0$5, $j>0$6 g-modes are identified as suitable candidates, with the mode frequencies evolving approximately in parallel and thereby allowing parallel orbital migration (Ćuk et al., 14 Apr 2026).

4. Galactic bars, spirals, and resonance rings

In stellar dynamics, Lindblad resonances are prominent tracers of non-axisymmetric structure. An action-angle analysis of the Geneva-Copenhagen Survey of $j>0$7 nearby F and G dwarfs found that stars in the Hyades stream occupy action-angle loci characteristic of scattering at an inner Lindblad resonance of a rotating disturbance potential. The resonance condition was tested both in action space and through the phase relation

$j>0$8

and the resulting signal was interpreted as a recent ILR event, with the sharpness of the angle correlation implying a timescale of $j>0$9 Myr. In that interpretation, the Hyades stream is a fossil of recent resonant scattering rather than a dissolved cluster remnant (Sellwood, 2010).

Gaia-based studies extended this program to bar resonances. Using kinematically hot stars in Gaia EDR3, action-space ridges and angular-momentum peaks were identified with inner $m$0, corotation, outer $m$1, outer Lindblad, and outer $m$2 resonances for $m$3 km s$m$4 kpc$m$5, or with corotation, outer $m$6, outer Lindblad, outer $m$7, and outer $m$8 resonances for $m$9 km s$j=1$0 kpc$j=1$1. In both interpretations the OLR remains a sharp, prominent feature in the hot-star distribution (Kawata et al., 2020). A related chemodynamical analysis found weak evidence for associating the “hat” moving group with the OLR of a slow bar and “Hercules” with corotation, while emphasizing that future Gaia releases with greater azimuthal coverage should permit more reliable identification of bar resonances (Wheeler et al., 2021).

Morphological ring structure in barred spirals is likewise organized by Lindblad resonances. In a spectroscopic study of five spiral galaxies, circumnuclear rings were found between the inner and outer ILR when both resonances are present, or between the center and a single ILR otherwise; the inferred pattern speeds lie in the range 26–47 km s$j=1$2 kpc$j=1$3, and the dimensionless ratio $j=1$4 lies between 1.1 and 1.6 (Schmidt et al., 2019). A larger PHANGS analysis of 74 nearby star-forming galaxies determined corotation radii for 38 of 46 barred systems, found a mean $j=1$5 with standard deviation $j=1$6, and then used rotation curves to estimate ILR and OLR positions; 19 barred galaxies have a measurable ILR, and nuclear rings almost always lie inside the ILR radius (Ruiz-García et al., 2024). In Milky Way Ferrers-bar models, the bar OLR at $j=1$7 kpc for $j=1$8 km s$j=1$9 kpc$\kappa = m(n-\Omega).$0 supports outer resonance rings $\kappa = m(n-\Omega).$1 and $\kappa = m(n-\Omega).$2, with $\kappa = m(n-\Omega).$3 capturing about twice as many particles as $\kappa = m(n-\Omega).$4 because it is wider (Melnik, 2019).

5. Torque theory from protoplanetary to relativistic discs

In gaseous discs, Lindblad resonances are a torque mechanism rather than only a kinematic marker. For type-I migration of a low-mass planet, the standard Lindblad torque formula assumes local power-law surface-density and temperature profiles, but it fails near opacity transitions where these gradients change sharply. A generalized expression that remains simple enough for population-synthesis use is

$\kappa = m(n-\Omega).$5

where the $\kappa = m(n-\Omega).$6-derivatives are evaluated at the inner and outer wakes. The key physical input is the shift of the Lindblad resonances caused primarily by the third derivative of the temperature profile. In the same framework, the vortensity-related corotation torque is boosted at opacity transitions and can counteract migration (Masset, 2012).

Three-dimensional disc theory has also revealed resonances that are distinct from, but best understood relative to, the classical Lindblad case. In a simplified linear model for planet–disc interaction, buoyancy resonances lie along tilted planes rather than cylindrical resonance radii, their width depends on damping rather than sound speed, and they do not launch propagating waves. At fixed large azimuthal wavenumber $\kappa = m(n-\Omega).$7, the buoyancy resonance acts closer to corotation than the Lindblad resonance, whereas the classical Lindblad torque retains its role as the wave-launching epicyclic resonance (Lubow et al., 2014).

A fully relativistic version of Lindblad torque theory exists for thin discs in Schwarzschild and Kerr spacetimes. The basic formalism gives a gauge-invariant torque density at resonance in terms of metric perturbations, and the companion calculation of resonance strengths shows that black-hole discs possess an $\kappa = m(n-\Omega).$8 inner Lindblad resonance with no Newtonian Keplerian analogue. Its strength is weak, in part because the gravitoelectric octupole and gravitomagnetic quadrupole contributions partially cancel. For Kerr holes, the $\kappa = m(n-\Omega).$9 ILR is enhanced for retrograde spins and suppressed for prograde spins, while the $n/\Omega \approx m/(m-1)$0 ILRs are enhanced relative to the nonrelativistic case; for a Schwarzschild hole the enhancement reaches a factor of 2 even when the perturber is at $n/\Omega \approx m/(m-1)$1 (Hirata, 2010, Hirata, 2010).

Retrograde circumbinary discs show that the existence of Lindblad torques can depend sensitively on the Fourier content of the forcing potential. For a circular binary, no Lindblad resonances lie inside a circular coplanar retrograde circumbinary disc, so no Lindblad torques are produced. If the binary is eccentric, however, components with $n/\Omega \approx m/(m-1)$2 appear in the potential expansion

$n/\Omega \approx m/(m-1)$3

and some rotate progradely with respect to the retrograde disc, enabling resonant torques. These torques are weaker than in the prograde case but can, at sufficiently high eccentricity, open a gap while still allowing gas inflow through streams (Nixon et al., 2015).

6. Symmetry constraints, exceptions, and current directions

Lindblad resonances are filtered by the symmetry of the perturbing potential. If the potential is invariant under a $n/\Omega \approx m/(m-1)$4 rotation, only harmonics with $n/\Omega \approx m/(m-1)$5 a multiple of $n/\Omega \approx m/(m-1)$6 are present. For a homogeneous triaxial ellipsoid, which has $n/\Omega \approx m/(m-1)$7-rotation symmetry, only even-$n/\Omega \approx m/(m-1)$8 Lindblad resonances survive; odd-$n/\Omega \approx m/(m-1)$9 Lindblad resonances disappear and are replaced by higher-order commensurabilities, such as the replacement of a $m(\Omega_p-\Omega_\phi)=l\Omega_R,$0 resonance by a $m(\Omega_p-\Omega_\phi)=l\Omega_R,$1 resonance. The same kinematic classification also yields “twins,” resonances with the same $m(\Omega_p-\Omega_\phi)=l\Omega_R,$2 and opposite $m(\Omega_p-\Omega_\phi)=l\Omega_R,$3, and “true twins,” resonances with the same $m(\Omega_p-\Omega_\phi)=l\Omega_R,$4 and opposite $m(\Omega_p-\Omega_\phi)=l\Omega_R,$5, sharing the same kinematics and, in the latter case, the same dynamics (Sicardy, 2020).

Several apparent contradictions in the literature are resolved once the dynamical context is specified. In the sectoral-resonance classification around a rotating body, there are no retrograde Lindblad resonances because retrograde resonances must have $m(\Omega_p-\Omega_\phi)=l\Omega_R,$6, whereas in eccentric retrograde circumbinary discs Lindblad resonances do occur through negative-frequency Fourier components of the binary potential (Sicardy, 2020, Nixon et al., 2015). This contrast indicates that “retrograde Lindblad resonance” is not a context-free statement: the allowed resonance set depends on the underlying decomposition of the perturbation and on which degrees of freedom are being averaged.

Current work increasingly treats Lindblad resonances as identifiable structures in large phase-space surveys and as ingredients in multiscale dynamical models. Chemodynamical diagnostics aim to separate OLR, corotation, and higher-order bar resonances in the Milky Way with wider azimuthal coverage (Wheeler et al., 2021). Large observational catalogs now tabulate CR, ILR, and OLR positions across nearby barred galaxies (Ruiz-García et al., 2024). In planetary systems, resonance-lock scenarios connect Lindblad resonances to planetary oscillation spectra rather than only to static figure asymmetries or companion satellites (Ćuk et al., 14 Apr 2026). A broader implication is that Lindblad resonances are best regarded not as a single specialized construction, but as a family of commensurability mechanisms whose detailed expression depends on symmetry, eccentricity, thermodynamics, and the degree of nonlinearity.