Galaxy Rotation Curves and Dark Matter

The dynamics of galaxies do not match their light. A disk galaxy's rotation speed, which should fall off beyond the luminous edge if the stars were all the mass, instead stays constant to the largest measurable radii. The same excess appears in the velocities of cluster galaxies, in the temperature of intracluster gas, and in the deflection of light. This lesson turns the flat rotation curve into a halo density profile, contrasts the isothermal and NFW models, adds the independent lensing and mass-to-light constraints, presents the Bullet Cluster as direct evidence that the dark mass is collisionless, and states where the MOND alternative succeeds and fails.

Flat rotation curves and the missing mass

A star or gas cloud on a circular orbit at galactocentric radius $r$ obeys

v_{c}^{2} (r) = \frac{G M ( < r )}{r},

for the mass $M (< r)$ interior to its orbit. Beyond the luminous disk, where the light and hence (if light traced mass) the enclosed mass would be nearly constant, the speed should decline as $v_{c} \propto r^{- 1/2}$ . Instead, the 21-cm line traces neutral hydrogen far past the optical edge and shows $v_{c} (r)$ flat — constant to tens of kpc, twice the optical radius or more. A flat curve inverts to

M (< r) = \frac{v _{c}^{2} r}{G} \propto r,

mass rising linearly with radius through a region containing negligible light. The implied density profile follows from $d M / d r = 4 π r^{2} ρ$ ,

ρ (r) = \frac{1}{4 π r ^{2}} \frac{d M}{d r} = \frac{v _{c}^{2}}{4 π G r ^{2}} \propto r^{- 2} .

An extended halo with $ρ \propto r^{- 2}$ produces exactly the flat curve. This is the dark halo: mass that gravitates but does not shine, dominating outside the stellar disk.

Decomposition of a flat rotation curve into the falling contributions of the luminous bulge and disk plus a rising dark-halo term; only the halo keeps the total flat to large radius.

Halo density profiles

Two profiles fit the data. The pseudo-isothermal sphere builds in a finite central density with a core radius $r_{c}$ ,

ρ_{iso} (r) = \frac{ρ _{0}}{1 + ( r / r _{c} ) ^{2}},

which is constant for $r ≪ r_{c}$ and falls as $r^{- 2}$ for $r ≫ r_{c}$ , giving a rotation curve that rises linearly through the core and flattens to $v_{\infty} = 4 π G ρ_{0} r_{c}^{2}$ at large radius. It is the phenomenological choice that reproduces observed curves with a central core.

Cosmological N-body simulations of collisionless cold dark matter instead produce the Navarro–Frenk–White (NFW) profile,

ρ_{NFW} (r) = \frac{ρ _{s}}{( r / r _{s} ) ( 1 + r / r _{s} ) ^{2}},

with a scale radius $r_{s}$ and characteristic density $ρ_{s}$ . It has a cusp, $ρ \propto r^{- 1}$ , in the center and steepens to $ρ \propto r^{- 3}$ in the outskirts, passing through the $r^{- 2}$ behavior near $r_{s}$ . Integrating gives the enclosed mass in closed form,

M_{NFW} (< r) = 4 π ρ_{s} r_{s}^{3} [ln (1 + \frac{r}{r _{s}}) - \frac{r / r _{s}}{1 + r / r _{s}}] .

The two profiles differ most in the center: the isothermal core is flat while the NFW cusp diverges. Whether real galaxies, especially dwarfs, have cores or cusps is the core–cusp problem, one of the standing tensions of the cold-dark-matter model.

The isothermal profile flattens to a constant-density core while the NFW profile cusps as the inverse of radius in the center and steepens to the inverse cube outside the scale radius; they agree near the scale radius where both fall as the inverse square.

Mass-to-light ratios

The dark mass shows up as a rising mass-to-light ratio $Υ = M / L$ , expressed in solar units $M_{⊙} / L_{⊙}$ . The stellar populations of galaxies have $Υ_{*} \sim 1$ – $5$ in the optical. Dynamical masses give systematically higher values as the aperture grows:

Inner disk (starlight): $Υ \sim 3$ , consistent with stars alone.
Full galaxy including the halo out to the last measured rotation point: $Υ \sim 10$ – $30$ .
Groups and clusters from the virial theorem: $Υ \sim 200$ – $400$ .

The ratio climbs by two orders of magnitude from the stellar cores of galaxies to the scale of clusters, tracking the increasing dark-matter fraction on larger scales. The cluster value implies a matter density parameter $Ω_{m} \approx 0.3$ , far above the $Ω_{b} \approx 0.05$ in baryons, so most of the matter is non-baryonic.

Gravitational lensing

Lensing weighs mass through its gravity alone, with no assumption of dynamical equilibrium. A point mass $M$ deflects a light ray with impact parameter $b$ by the general-relativistic angle $α = 4 GM / (c^{2} b)$ , twice the Newtonian value. A source directly behind a lens is imaged into an Einstein ring of angular radius

θ_{E} = \frac{4 GM}{c ^{2}} \frac{D _{L S}}{D _{L} D _{S}},

where $D_{L}$ , $D_{S}$ , and $D_{L S}$ are the angular-diameter distances to the lens, to the source, and between them. Measuring $θ_{E}$ inverts directly to the mass inside the ring,

M (< θ_{E}) = \frac{c ^{2}}{4 G} θ_{E}^{2} \frac{D _{L} D _{S}}{D _{L S}} .

Strong lensing — multiple images, arcs, and rings around cluster cores — probes the mass in the dense center; weak lensing — the coherent few-percent tangential distortion of thousands of faint background galaxies — maps the mass in the outskirts statistically. Lensing masses agree with the dynamical and X-ray masses and exceed the luminous mass by the same large factor, an independent confirmation that does not rely on orbital motion.

A foreground mass bends the light of a background galaxy into arcs and, for perfect alignment, a full Einstein ring whose angular radius fixes the projected mass inside it.

The Bullet Cluster

The Bullet Cluster (1E 0657-56) is a collision of two galaxy clusters caught in the act, and it separates the mass from the gas. The hot X-ray-emitting intracluster gas, which holds most of the baryonic mass, is collisional: as the clusters passed through one another the two gas clouds rammed together, shocked, slowed, and lagged behind at the center. The galaxies themselves, effectively collisionless points, sailed through and now sit ahead of the gas on each side.

Weak-lensing reconstruction locates the gravitational mass. It coincides with the galaxies — the collisionless component — and is offset from the gas, where most of the baryons actually reside. If the excess gravity came from a modification of the gravitational law sourced by the visible (mostly gaseous) matter, the lensing signal would center on the gas. That it centers on the galaxies instead is direct evidence for a dominant, collisionless, non-baryonic mass component that passed through the collision unimpeded.

In the Bullet Cluster the collisional X-ray gas is decelerated at the center of the collision while the lensing mass tracks the collisionless galaxies that flew ahead, placing the gravitating mass away from most of the baryons.

The MOND alternative

Modified Newtonian Dynamics (MOND) proposes that the discrepancy is not missing mass but a departure from Newton's law at very low accelerations. Below a threshold $a_{0} \approx 1.2 \times 1 0^{- 10} m s^{- 2}$ the effective acceleration becomes $a = a_{N} a_{0}$ rather than the Newtonian $a_{N} = GM / r^{2}$ . For a star far out in a galaxy,

\frac{v _{c}^{2}}{r} = \frac{GM}{r ^{2}} a_{0} ⟹ v_{c}^{4} = GM a_{0},

a flat rotation curve with $v_{c}$ independent of $r$ , and a built-in $L \propto v_{c}^{4}$ Tully–Fisher relation with the correct normalization. MOND fits the rotation curves of individual disk galaxies with a single universal parameter and no per-galaxy halo. Its failures are on larger scales: it cannot fully account for the mass of galaxy clusters without residual dark matter, it has no natural relativistic extension that matches the CMB and structure formation, and — most directly — it does not explain the Bullet Cluster, where the gravitating mass is displaced from the baryons that a modified-gravity law would have to follow. The evidence favors dark matter as a substance, with MOND capturing a still-unexplained regularity of galactic dynamics.

Summary

Flat rotation curves imply $M (< r) \propto r$ and a halo density $ρ \propto r^{- 2}$ extending well beyond the light. The pseudo-isothermal profile has a constant-density core, while the cosmologically motivated NFW profile cusps as $r^{- 1}$ inside $r_{s}$ and steepens to $r^{- 3}$ outside, with the core–cusp question still open. The mass-to-light ratio rises from $Υ \sim 3$ in stellar cores to $\sim 300$ in clusters, giving $Ω_{m} \approx 0.3$ against $Ω_{b} \approx 0.05$ . Gravitational lensing weighs the same mass without dynamics through the Einstein radius $θ_{E}$ , and the Bullet Cluster displaces the lensing mass from the collisional X-ray gas, evidence for a collisionless non-baryonic component. MOND reproduces individual galaxy rotation curves and Tully–Fisher through a low-acceleration law $v_{c}^{4} = GM a_{0}$ but fails on clusters, in the CMB, and at the Bullet Cluster.¹²³

Carroll & Ostlie, §24.3, §25.3 — the flat rotation curve of the Galaxy and other spirals, halo density profiles, and the mass-to-light ratio as evidence for dark matter. ↩
Ryden, Ch. 7 — dark matter: rotation curves, cluster masses, gravitational lensing, and the Bullet Cluster. ↩
Maoz, Ch. 6, Ch. 10 — dark-matter halos, the isothermal and NFW profiles, and the MOND alternative. ↩