Lindblad Resonance

The metronome that builds a galaxy

A disk galaxy does not rotate like a solid record. The inner stars sweep around fast, the outer ones lag, and any spiral arm drawn through them would wind up into a tight coil within a couple of rotations. Yet spiral arms persist for billions of years, bars hold their shape, and rings sit at fixed radii. The resolution is that the arm is not a fixed group of stars — it is a pattern, a wave, that turns rigidly at its own single angular speed Ω_p while the stars stream through it. The question that organizes all of galactic dynamics is then: where, and how, does that rigid pattern talk to the differentially rotating stars beneath it? The answer is almost entirely "at the resonances," and the most important of those are the Lindblad resonances.

The intuition is a metronome. A star on a near-circular orbit is not really moving in a circle; it executes a small radial oscillation — an epicycle — at a characteristic frequency κ, the epicyclic frequency, while its guiding center circles at the orbital frequency Ω. Now let a two-armed spiral sweep past at pattern speed Ω_p. In the frame that co-rotates with the pattern, the star feels a periodic tug every time an arm passes, at the forcing frequency m(Ω − Ω_p), where m = 2 counts the arms. When that forcing frequency happens to equal the star's own radial frequency κ, the tugs always arrive at the same phase of the epicycle. They add up coherently rather than averaging to nothing. That is a Lindblad resonance, and it is the only place in the disk where the pattern and the orbits can trade energy and angular momentum efficiently.

The resonance condition

Written out, the resonance condition is

m ( Ω(R) − Ω_p ) = ± κ(R)        ⇔        Ω_p = Ω(R) ± κ(R)/m

The minus sign defines the Inner Lindblad Resonance (ILR); the plus sign defines the Outer Lindblad Resonance (OLR). Between them sits the special radius where the forcing frequency vanishes entirely, m(Ω − Ω_p) = 0, that is corotation, where the stars orbit at exactly the pattern speed: Ω(R_CR) = Ω_p. Because Ω decreases outward in any realistic disk, corotation is a single radius, and the ILR lies inside it (Ω > Ω_p, stars overtaking the pattern) while the OLR lies outside it (Ω < Ω_p, pattern overtaking the stars). The three radii — ILR, corotation, OLR — are the skeleton on which a galaxy's visible structure hangs.

It is convenient to think in terms of the combination Ω ± κ/m. For the dominant m = 2 (a bar, or a grand-design two-armed spiral) the relevant curves are Ω − κ/2 and Ω + κ/2. The quantity Ω − κ/2 has a beautiful interpretation: it is the angular rate at which a near-circular orbit's long axis precesses. A pattern of pinched, aligned elliptical orbits all precessing at the same rate Ω_p is a spiral or a bar — which is exactly why density-wave theory and resonance theory are the same theory seen from two angles.

The epicyclic frequency, and why κ ≠ Ω

Everything hinges on κ differing from Ω. The epicyclic frequency is fixed by the local shape of the rotation curve through

κ²(R) = R dΩ²/dR + 4Ω²  =  (2Ω/R) d(R²Ω)/dR

Three reference cases bound the behaviour:

• Keplerian (point mass, Ω ∝ R⁻³ᐟ²): κ = Ω. Orbits close after one turn — they are fixed ellipses. With κ = Ω there is no precession and resonances degenerate; this is why planetary systems behave differently from galaxies.
• Flat rotation curve (v = const, Ω ∝ R⁻¹), the realistic case for galaxy disks: κ = √2 Ω ≈ 1.41 Ω. The orbit precesses, and Ω − κ/2 ≈ 0.29 Ω is a slowly varying, gently declining function of radius.
• Solid body (Ω = const, the deep interior): κ = 2Ω. The orbit closes after half a radial period.

For the realistic flat-rotation case, the function Ω − κ/2 is nearly constant across a wide range of radii. That near-constancy is the deep reason spiral patterns can be quasi-stationary: a single pattern speed Ω_p can stay close to the precession rate Ω − κ/2 over much of the disk, so the elliptical orbits do not shear apart and the spiral keeps its shape.

Worked example: locating the resonances in a flat-rotation disk

Take a disk with a flat rotation curve v = 220 km/s (a reasonable Milky Way value), so the orbital frequency is Ω(R) = v/R and κ(R) = √2 Ω(R). Convert units: 220 km/s ÷ 1 kpc = 220 km/s / (3.086 × 10¹⁶ km) ≈ 7.1 × 10⁻¹⁶ s⁻¹, which is the friendly figure Ω(1 kpc) = 220 km/s/kpc. So Ω(R) = 220/R km/s/kpc with R in kpc.

Now suppose a bar turns at Ω_p = 40 km/s/kpc, an m = 2 pattern. Corotation is where Ω = Ω_p:

220 / R_CR = 40   →   R_CR = 5.5 kpc

The OLR is where Ω + κ/2 = Ω_p... but Ω + κ/2 > Ω > Ω_p at corotation, so we instead solve Ω(1 + √2/2) = Ω_p outward of corotation. With κ/2 = (√2/2)Ω = 0.707 Ω:

OLR:  Ω − κ/2 ... no — use the outer branch  Ω(1 + 0.707) = ...
      Correct branch: Ω_p = Ω + κ/2  ⇒  40 = Ω(1 + 0.707) = 1.707 Ω
      Ω_OLR = 23.4 km/s/kpc  →  R_OLR = 220 / 23.4 = 9.4 kpc
ILR:  Ω_p = Ω − κ/2  ⇒  40 = Ω(1 − 0.707) = 0.293 Ω
      Ω_ILR = 136.5 km/s/kpc  →  R_ILR = 220 / 136.5 = 1.6 kpc

So this bar, turning at 40 km/s/kpc, has its ILR at about 1.6 kpc, corotation at 5.5 kpc, and OLR at 9.4 kpc. The ratio R_OLR / R_ILR ≈ 5.9, and the OLR sits at roughly 1.7 × corotation — the classic spacing. A nuclear ring of vigorous star formation would be expected near 1.6 kpc, the bar would terminate near 5.5 kpc, and a detached outer pseudo-ring would encircle the system near 9.4 kpc. Change Ω_p and the whole skeleton slides in or out: a faster pattern (larger Ω_p) pulls all three radii inward; a slower pattern pushes them out. Measuring any one ring therefore measures Ω_p, and measuring Ω_p predicts the other two rings.

Regimes: one ILR, two ILRs, or none

The number of inner Lindblad resonances is not fixed — it depends on the detailed run of Ω − κ/2 versus R, which in turn depends on the central mass concentration. In a galaxy with a steep central density rise (a massive bulge or a concentrated nuclear disk), the curve Ω − κ/2 rises to a peak in the inner kpc and then declines. A horizontal line at the pattern speed Ω_p can then cut that curve in two places, giving an inner ILR and an outer ILR with a "forbidden" annulus between them, or it can miss the peak entirely and give no ILR at all.

• No ILR. A trailing density wave traveling inward is not absorbed; it reflects off the center and returns as a leading wave, which can amplify by swing amplification and feed a growing bar. The absence of an ILR is one of the conditions that favors bar formation (Toomre's "no-ILR" criterion).
• Two ILRs. The wave is absorbed at the outer ILR; gas driven inward by the bar piles up between the two ILRs, building a nuclear ring and a secondary inner bar. Many double-barred galaxies live in this regime.
• One ILR (marginal case). The transitional configuration where the Ω_p line is tangent to the Ω − κ/2 peak.

This is why the inner structure of barred galaxies — whether they host nuclear rings, nuclear spirals, or inner bars — is read directly off the Ω − κ/2 diagram. The same diagram, with the Ω + κ/2 curve added, fixes the OLR and hence the outer ring.

From resonance to ring: how gas responds

Stars passing through a Lindblad resonance gain or lose angular momentum but keep streaming. Gas behaves differently: it is collisional, it shocks, it dissipates energy, and it must settle onto closed, non-self-intersecting orbits. Near each principal resonance the orientation of the stable closed-orbit family flips by 90°, and in the transition zone the orbits crowd together. Gas piling into those crowded zones forms rings:

• Nuclear ring — just inside the ILR. Bar-driven inflow stalls here, builds up surface density, and ignites starburst rings like the ~1 kpc ring of NGC 1097 or the ring in NGC 1300. This is also the gas reservoir that can ultimately feed the central supermassive black hole.
• Inner ring — encircling the bar, near corotation. These are the prominent oval rings (e.g. NGC 2523) that trace the bar's end.
• Outer ring / pseudo-ring — near the OLR, at roughly twice the bar radius. Often broken into two spiral arcs forming an "outer pseudo-ring" (NGC 1433, NGC 6782).

Ronald Buta's ring catalogs classify thousands of galaxies by exactly which resonance each ring traces, turning ring morphology into a galactic-scale tachometer: read the ring, get Ω_p, predict the bar length and the other rings.

The quantitative engine: angular-momentum transfer

Why do resonances matter dynamically, not just geometrically? Lynden-Bell and Kalnajs (1972) showed that for a steadily rotating pattern, the time-averaged torque between the pattern and the disk is nonzero only at the resonances. The pattern absorbs angular momentum at the ILR (stars there lose angular momentum and sink inward) and deposits it at the OLR (stars there gain angular momentum and move outward). Corotation can act either way depending on the disk's angular-momentum gradient. The net effect is an outward transport of angular momentum through the disk, mediated by the wave.

This transfer is what lets bars grow and lengthen by donating angular momentum to the outer disk and the dark-matter halo, and it is what drives gas inward of the ILR to fuel central activity. The same resonant bookkeeping governs planetary rings (Lindblad resonances with moons open the gaps and sharpen the edges in Saturn's rings — the Encke and Keeler gaps, the outer edge of the A ring at the Janus 7:6 resonance) and protoplanetary disks (a forming planet opens an annular gap at its Lindblad resonances). The Milky Way's Lindblad resonances leave kinematic fingerprints in the solar neighborhood: the Hercules stream in the local velocity distribution is widely attributed to stars trapped near the bar's outer Lindblad resonance or a corotation resonance, a signature now mapped in fine detail by Gaia.

Resonances at a glance

Resonance	Condition (m=2)	Position	Stars vs. pattern	Typical feature	Example radius (Ω_p=40)
Inner Lindblad (ILR)	Ω_p = Ω − κ/2	Inside corotation	Stars overtake pattern	Nuclear ring, wave absorption	~1.6 kpc
Inner Lindblad (inner branch)	Ω_p = Ω − κ/2	Deep interior	Stars overtake pattern	Inner bar, nuclear spiral	~0.5–1 kpc (if present)
Ultraharmonic (4:1)	Ω_p = Ω − κ/4	Inside corotation	Stars overtake pattern	Spiral-arm bifurcation	~3 kpc
Corotation (CR)	Ω = Ω_p	Center of bracket	Co-moving	Bar end, inner ring	5.5 kpc
Outer Lindblad (OLR)	Ω_p = Ω + κ/2	Outside corotation	Pattern overtakes stars	Outer pseudo-ring	~9.4 kpc
Outer 4:1	Ω_p = Ω + κ/4	Just outside CR	Pattern overtakes stars	Spiral-arm truncation	~7 kpc

The pattern is invariant: every resonance is a rational ratio between the forcing frequency m(Ω − Ω_p) and the radial frequency κ. The Lindblad resonances (1:1 with κ) are the strongest; the ultraharmonics (κ/2, κ/4, the 4:1 resonances) are weaker but visible as places where spiral arms branch or truncate.

Who worked it out

The Swedish astronomer Bertil Lindblad recognized in the 1940s and early 1950s that galactic spiral structure must be understood through orbital resonances in a differentially rotating disk, introducing the resonance condition that now bears his name. The picture was incomplete until C. C. Lin and Frank Shu (1964) reframed spiral arms as a quasi-stationary density wave — a self-consistent gravitational pattern rather than a material structure — and Alar Toomre supplied the stability analysis (the Toomre Q parameter) and the swing-amplification mechanism. Lynden-Bell and Kalnajs (1972) proved that angular-momentum transfer is concentrated at the resonances, closing the loop between Lindblad's resonance bookkeeping and the energetics of wave-driven evolution. The framework now underpins everything from bar-driven galaxy fueling to the gaps in Saturn's rings and the carving of protoplanetary disks.

Common pitfalls and misconceptions

• Confusing pattern speed with orbital speed. Ω_p is one rigid number for the whole pattern; Ω(R) is a steeply falling function of radius. At the solar circle the stars orbit at Ω ≈ 28 km/s/kpc while the Milky Way bar turns at Ω_p ≈ 38 km/s/kpc — the bar is faster than the local stars, which is exactly why the Sun lies outside corotation, near the OLR.
• Assuming there is always an ILR. Whether a galaxy has zero, one, or two ILRs depends on the central concentration via the Ω − κ/2 curve. The "no-ILR" case actually favors bar growth, because waves reflect and amplify instead of being absorbed.
• Thinking resonances are points where orbits collide. A resonance is a frequency-matching condition, not a collision. The coherent forcing pumps the orbits gradually; the dramatic structure (rings, gaps) emerges from the dissipative gas response and the orbit-family transitions, not from any catastrophe at a single radius.
• Treating κ = Ω. That is only true for a Keplerian point-mass potential. In a galaxy disk κ ≈ √2 Ω, and it is precisely the difference κ ≠ Ω (the orbital precession) that makes Lindblad resonances exist at all.
• Equating corotation with the bar's physical end exactly. Corotation is near the bar end, but the ratio R_CR / a_bar is typically 1.0–1.4, not exactly 1. "Fast" bars have R_CR/a_bar ≈ 1.0–1.4; "slow" bars exceed that, and the distinction probes the dark-matter content via dynamical friction on the bar.

Observational status and applications

• Milky Way bar. Gaia DR2/DR3 astrometry plus APOGEE spectroscopy place the bar pattern speed at roughly Ω_p ≈ 35–40 km/s/kpc, putting corotation near R ≈ 6 kpc and the OLR not far outside the solar circle — consistent with the Hercules moving group being an OLR/corotation feature.
• External barred galaxies. Integral-field surveys (MaNGA, CALIFA) and the PHANGS-MUSE/ALMA program have applied the Tremaine–Weinberg method to dozens of bars, finding most are "fast" (R_CR/a_bar ≲ 1.4), which constrains how much dark matter sits within the bar region.
• Resonance rings as tachometers. Buta's de Vaucouleurs-Atlas ring classifications let observers infer Ω_p purely from ring morphology and the bar length, cross-checking the kinematic measurements.
• Planetary rings. The same physics, scaled down: Lindblad resonances with Saturn's moons open the Encke and Keeler gaps and define the sharp outer edge of the A ring (the 7:6 resonance with the co-orbital moons Janus and Epimetheus).
• Protoplanetary disks. A growing planet exerts Lindblad torques on its disk, opening an annular gap — the rings and gaps imaged by ALMA in systems like HL Tauri are widely modeled as resonance-carved structures.