Galaxy Bar

What a bar actually is

The word "spiral galaxy" hides a fact that startled astronomers when the first systematic surveys ran the numbers: most spiral galaxies are not pure spirals. They are barred spirals. Threaded through the centre of the rotating disk, stretching from one side of the bulge to the other, is a linear, elongated assembly of stars — the bar. It is not a rigid rod of matter. It is a rotating density pattern: at any instant the bar contains a specific population of stars, but the orbits that make it up rearrange themselves continuously to maintain the elongated shape. The bar pattern circulates around the galactic centre at its own angular speed Ω_b, distinct from the orbital frequency of individual stars.

Bars are dynamically the most important non-axisymmetric feature of disk galaxies. They store angular momentum, transport gas, sculpt nuclear rings, fuel central black holes, and over billions of years they restructure the entire inner disk — a process now called secular evolution. They are the principal way an isolated disk galaxy reshapes itself when no merger or close encounter is forcing the issue.

How common bars are

Hubble's original classification scheme distinguished three families: SA (unbarred), SAB (intermediate), and SB (strongly barred). For a long time the bar fraction was estimated in the optical at around one-third — which made bars feel like an interesting minority. Two large surveys revised this dramatically. Sloan Digital Sky Survey classifications in the optical (the GalaxyZoo and Nair–Abraham–Lintott samples) found strong-plus-weak bar fractions of 50–70 percent. The Spitzer Survey of Stellar Structure in Galaxies (S4G), imaging at 3.6 μm where dust extinction is minimal and old-star structure is best traced, settled on a bar fraction of roughly 65 percent in nearby disk galaxies.

The fraction depends on what you measure:

Population	Wavelength	Bar fraction	Notes
Local disks (S4G)	3.6 μm (IR)	~65%	SB + SAB; dust-penetrating
Local disks (SDSS optical)	0.3–0.9 μm	~50%	Dust-attenuated; misses weak bars
Luminous red disks	IR	70–80%	Gas-poor, dynamically settled
Blue gas-rich disks	IR	40–50%	Turbulent, gas-rich, hotter
Dwarfs (M_B > −18)	Optical	~20%	Dark matter–dominated, halo stable
z = 0.5	HST optical	~50%	Comparable to local
z = 1	HST optical	~20%	Disks still settling
z > 2 (JWST early data)	NIR rest-frame	10–20%	More than pre-JWST expected

The pattern is consistent: bars are common, and the bar fraction grows toward the present as disks cool, settle, and become more rotation-dominated.

Our own bar

The Milky Way is itself a barred spiral. This was hard to establish from inside the disk and was only firmly confirmed in the 1990s, primarily from infrared photometry. Several lines of evidence converge on the same structure:

• COBE/DIRBE near-infrared maps show a clear left–right asymmetry in the inner disk — the bar's near end is tipped toward us on one side.
• Red clump giants (standard candles among old stars) are systematically more distant on one side of the Galactic centre than the other — a kpc-scale offset matching a bar of semi-major axis ~3 kpc.
• Gas kinematics in the inner Galaxy show non-circular streaming consistent with motion along the leading edges of a rotating bar.
• OGLE microlensing surveys show enhanced lensing rates toward the centre that match an elongated star distribution.

The best-fit geometry is a primary bar with semi-major axis a ≈ 3 kpc, oriented at roughly 30° to the Sun–Galactic-centre line of sight (the near side is on the positive-longitude side). Some analyses also report a more extended "long bar" reaching ~5 kpc — debate persists about whether this is a separate component or simply the outer envelope of the same structure. Either way, the Milky Way is solidly in the SB or SAB category, not unbarred.

How a bar forms

The defining property of a disk that can form a bar is that it is dynamically cold and rotationally supported. The criterion is most easily stated in terms of the Toomre Q parameter — gravitational stability of a rotating disk to local axisymmetric perturbations — and in terms of the ratio of disk to total enclosed mass. If a disk is too hot (high random velocity, high Q) it resists collective deformation. If it is dominated by a stabilising halo or bulge, the disk's self-gravity is too feeble to drive a global instability. But if a disk is cold and self-gravitating relative to the halo, a generic m = 2 (bisymmetric) instability spontaneously grows on a few rotation periods.

The canonical numerical demonstration is Frank Hohl's 1971 N-body experiment. Hohl set up an axisymmetric, rotationally supported disk of 100,000 particles, gave them no special initial perturbation beyond Poisson noise, and let them evolve. Within ~2 rotation periods the disk spontaneously formed a strong bar. The instability is robust: it appears in every cold disk simulation unless explicitly stabilised by a hot halo or a dynamically hot bulge.

Evangelie Athanassoula's analytic and numerical work (1980s–2000s) clarified how bars exchange angular momentum with surrounding components. Bar growth is governed by the rate at which the bar can shed angular momentum — primarily to the dark matter halo, secondarily to the outer disk. Haloes that can accept more angular momentum (cuspy, dense) let stronger bars develop; rigid halo "shells" suppress bar growth. This is now the standard framework: bars are not just kinematic decorations of the disk — they are angular-momentum exchanging engines coupled to every component of the galaxy.

Pattern speed and resonances

Once formed, the bar rotates as an approximate solid body at pattern speed Ω_b. Individual stars in the bar move on closed orbits in the rotating frame (the dominant family is called x1 orbits, elongated along the bar axis). Their positions sweep around at Ω_b — much slower than their own orbital frequency Ω(r) at most radii.

The interplay between Ω(r) and Ω_b sets up a system of resonances at specific radii:

Corotation:      Ω(R_CR) = Ω_b
Inner Lindblad:  Ω(R_ILR) − κ(R_ILR)/2 = Ω_b
Outer Lindblad:  Ω(R_OLR) + κ(R_OLR)/2 = Ω_b

where κ(r) is the local epicyclic frequency. The corotation radius R_CR is where the bar's solid-body rotation matches the local circular speed; stars there drift slowly with respect to the bar. Inside R_CR the disk overtakes the bar; outside, it lags. The inner Lindblad resonance (ILR) is the key inflow choke point — gas streamlines change orientation across the ILR, and the bar's torque on gas vanishes there. The outer Lindblad resonance (OLR) sets the outer reach of the bar's influence.

The dimensionless ratio R = R_CR / R_bar classifies bars:

R = R_CR / R_bar	Class	Inferred halo
1.0 – 1.4	Fast bar	Low-density halo (slowed bar little)
1.4 – 1.7	Intermediate	Moderate halo coupling
> 1.7	Slow bar	Dense halo strongly braked the bar

Most observed bars are fast (R < 1.4) — a constraint that has been used to argue that real galaxy haloes are less dense at small radii than naive Λ-CDM N-body predictions, contributing to the "cusp–core" debate.

Gas inflow, nuclear rings, AGN fuelling

The bar's most observationally consequential effect is on gas. Gas, unlike stars, is dissipative — it loses energy in shocks. As gas in the disk crosses the rotating bar potential, it cannot follow the elongated stellar x1 orbits without colliding with itself. Strong shocks form along the leading edges of the bar — visible in many barred galaxies as straight, narrow dust lanes that bracket the bar axis.

These shocks do three things at once: they dissipate kinetic energy, they radiate it away, and they extract angular momentum from the gas. The result is steady inflow of gas along the bar at typical rates of 0.1 to several solar masses per year. The gas streams inward until it crosses the inner Lindblad resonance — at which point the bar's torque effectively turns off and the gas piles up into a dense, rotation-supported nuclear ring with radius typically 100–1000 parsecs.

Nuclear rings are spectacular. They are among the most molecule-rich, star-formation-luminous structures in disk galaxies; many host hundreds of young massive star clusters in a ring of furious starburst activity. NGC 1097 and NGC 1300 are textbook examples. Residual gas that drifts further inward — past the ring, through some secondary mechanism such as a nuclear bar or a circumnuclear disk instability — can ultimately feed the central supermassive black hole. Numerically, bar-driven inflow is the single most efficient large-scale mechanism for delivering gas to the central kpc of an isolated disk galaxy. Statistical studies link bar presence to enhanced central star formation and, modestly, to higher AGN fractions — though the AGN connection is messy because AGN duty cycles are short compared with bar lifetimes.

Worked example: pattern speed and gas inflow rate

Consider a Milky-Way-like galaxy with circular speed v_c = 220 km/s and a bar of semi-major axis R_bar = 3 kpc. Suppose the bar is fast: R_CR / R_bar = 1.2, so R_CR = 3.6 kpc. The pattern speed follows from the corotation condition Ω_b = v_c / R_CR:

Ω_b = 220 km/s ÷ 3.6 kpc
    ≈ 61 km/s/kpc
    ≈ 60 km/s/kpc

Equivalently, the bar's rotation period is T_bar = 2π / Ω_b ≈ 100 million years. Over a Hubble time (~13.7 Gyr) the bar completes roughly 130 revolutions — assuming it doesn't slow down (it does).

For the gas inflow, a typical strong bar drives 0.3–1 M☉/yr from the disk into the central kpc. Integrating that over a Gyr delivers a few × 10⁸ M☉ of gas to the nuclear region — enough to grow the bulge by tens of percent or to feed a central black hole at Eddington rates for ~10⁶ years (the bottleneck is not the bar but the final-parsec transport).

Why bars slow down

A bar embedded in a live (responsive) dark matter halo loses angular momentum to that halo over Gyr. The mechanism, worked out analytically by Martin Weinberg in 1985 and demonstrated numerically by Victor Debattista and Jerry Sellwood (1998, 2000) and Athanassoula (2002), is dynamical friction in the halo. Halo particles on orbits in resonance with the bar — particularly at the corotation and OLR — exchange angular momentum with it through gravitational coupling. The net flow is from bar to halo: the bar slows down, the halo speeds up (modestly).

Sellwood and Debattista pointed out that this prediction is in tension with the observation that most bars are fast. A bar embedded in a Λ-CDM-cuspy halo for ~10 Gyr should have slowed to R_CR / R_bar > 1.7, becoming a slow bar. Most observed bars are fast. The resolution lies somewhere in: lower-density haloes at small radii (cusp–core debate), feedback baryonic processes that flatten the inner halo, or — less popular — that observed bars are young. The bar slowdown problem remains a useful pressure on the small-scale structure of dark matter.

Bar morphology and Hubble classification

Hubble's tuning-fork diagram splits disk galaxies into two parallel tracks distinguished by bar presence:

• SA — no bar. Pure spirals (M31, M51, M101). About one-third of local disks.
• SAB — intermediate / weak bar. Often only visible in IR or in unsharp-masked imaging.
• SB — strong bar. The bar is unmistakable in optical (NGC 1300, NGC 1097, NGC 1365).

Within the barred families, the secondary suffix (a, b, c, d) tracks the same bulge-to-disk, arm tightness, and gas content axis as for unbarred spirals. SBc galaxies, for example, are gas-rich, loosely wound barred spirals; SBa galaxies are gas-poor, tightly wound, bulge-dominated barred spirals.

Bars also come in nested form. Roughly 30% of strongly barred galaxies host a secondary "nuclear bar" — a smaller, kpc-scale bar oriented differently from the primary, residing inside the nuclear ring. Nuclear bars are thought to be a continuation of the inflow problem: gas piled up at the ILR can itself go bar-unstable on small scales and drive a second round of inflow toward the central few hundred parsecs.

Bar fraction across cosmic time

The bar fraction is not constant with redshift. Local disks: ~65 percent. At z ≈ 1 (about 8 Gyr ago): ~20 percent. At z > 2: ≤10 percent in most pre-JWST studies, perhaps 10–20 percent in early JWST samples. The growth of the bar fraction since z ≈ 1 is interpreted as gradual disk settling. At high redshift, disks were:

• Geometrically thicker and dynamically hotter (σ_z / v_circ closer to unity);
• Gas-rich (gas fractions 30–70 percent versus ~10 percent today);
• Turbulent on tens of km/s scales from cold accretion and frequent minor mergers;
• Frequently disturbed by tidal encounters.

All four conditions suppress the bar instability — they keep the disk too hot, too turbulent, or too disturbed for a coherent m = 2 mode to grow. As disks cooled, drained their cold gas, and quieted, more of them crossed the bar threshold. JWST observations of bars at z > 2 are now testing whether the transition is gradual or sharp, and how the very earliest disks managed to form bars despite the chaos.

Variants and related structures

• Boxy / peanut-shaped bulges (B/P bulges). The vertical extension of a mature bar — old bar stars develop vertical oscillations through the buckling instability and produce a characteristic boxy or X-shaped bulge when viewed edge-on. The Milky Way's own central X-shape, identified in WISE 2MASS infrared maps, is the side-on view of the bar buckled vertically.
• Nuclear bars (double-barred galaxies). Secondary inner bars, ~100–1000 pc, nested inside the primary bar. Drive a second tier of inflow toward the central black hole.
• Lenses and ovals. Smoother elongations — less sharply defined than bars, but dynamically related. Sometimes are former bars that have weakened or merged into the bulge.
• Tidal bars. Bar-like distortions in galaxies that have recently been tidally perturbed by a companion — induced rather than spontaneously instability-grown.
• Pseudobulges. Disky, rotation-supported "bulges" built up by bar-driven gas inflow — secular products of the bar rather than primordial spheroidal bulges built by mergers.

Famous bars in nearby galaxies

• NGC 1300 — The textbook barred spiral. A near face-on SBbc galaxy with a long, narrow, well-defined bar; the dust lanes along the bar leading edges and the inner spiral arms emerging from the bar ends are visible in HST imaging.
• NGC 1365 — A spectacular SBb in the Fornax cluster; the bar drives gas into a luminous nuclear ring with multiple star-forming clusters, and the galaxy hosts a Seyfert 1 AGN at the centre.
• NGC 1097 — A bar-with-nuclear-ring archetype; the ring is a closely studied starburst with a Seyfert 1 LINER nucleus inside it. ALMA has mapped the molecular gas flow along the bar in detail.
• M83 (NGC 5236) — A nearby SBc grand-design barred spiral; the inner bar is short but well-developed, and the galaxy is one of the most actively star-forming in the local universe.
• Milky Way — Our own SB/SAB. Best evidence: bar in IR red-clump distance maps, central X-shape buckled bar (DIRBE, WISE), inner gas kinematics. Bar a ≈ 3 kpc, oriented ~30° to the Sun line.

Common pitfalls

• Treating "spiral" and "barred spiral" as separate families. They are not. Bars are the rule, not the exception, among disk galaxies; about two-thirds of nearby disks are barred. SA is the special case.
• Confusing the bar with the bulge. The bulge is a roughly spheroidal old-star concentration; the bar is an elongated, rotation-supported pattern. The two coexist (the bar threads through the inner bulge) but are dynamically different. Photometric decompositions must include a separate bar component or the bulge fit ends up overweight.
• Identifying bars from optical alone. Optical images suffer dust extinction; many bars are easily missed at λ < 1 μm and show up clearly at 3.6 μm. Bar fractions from the S4G IR survey are systematically higher than from SDSS optical for the same galaxies.
• Confusing pattern speed with star orbital speed. The bar's pattern speed Ω_b is the rate at which the elongated shape rotates around the centre — not the orbital speed of any individual star. Stars in the bar have their own (much faster) frequencies and execute closed orbits in the rotating bar frame.
• Forgetting that gas, not stars, is what the bar transports. Bar inflow is a gas effect — gas dissipates in shocks at the leading edges. The stellar bar itself does not migrate; its constituent stars are exchanged but the pattern persists.
• Treating bar slowdown as universal. The expectation that all bars must slow strongly is a prediction of cuspy live haloes; observations show many fast bars, which is a constraint on halo structure, not a violation of dynamical friction.