Galactic Dynamics and Dark Matter

Point a telescope at a spiral galaxy and it looks reassuringly like a scaled-up Solar System: a bright, crowded bulge in the middle, a thin disk of stars spiralling outward, and darkness beyond the visible edge. So we expect it to behave like a scaled-up Solar System. Mercury tears round the Sun at 48\ \text{km/s}; distant Neptune ambles along at only 5\ \text{km/s}. The farther out you go, the slower you orbit — that is Newtonian gravity working exactly as Kepler charted it. Do the same measurement on a galaxy — clock how fast stars and gas circle the centre as a function of their distance r — and Newton makes a clean, confident prediction: beyond the luminous disk the orbital speed should fall away.

The sky flatly refuses. When Vera Rubin and Kent Ford actually measured these rotation curves in the 1970s, the speed did not fall. It stayed stubbornly, almost boringly constant, kilometre for kilometre, out to distances far beyond where the last star fades to black. Something is holding those outer stars in orbit — something with a great deal of gravitational mass and no light at all. Chase that single discrepancy honestly and it drags you to one of the most extraordinary conclusions in modern science: roughly 85% of the matter in the universe is invisible. This page builds that argument the way nature forced it on us — from one equation and one graph that would not behave.

The one equation: what sets an orbital speed

A star on a (near-)circular orbit is doing the same thing as a ball on a string: something must supply the inward centripetal force that bends its straight-line motion into a circle. For a planet or a star, that something is gravity. Balance the two. A star of mass m, orbiting at radius r with speed v, feels an inward gravitational pull from all the mass lying interior to its orbit — call that enclosed mass M(r):

\underbrace{\frac{G\,M(r)\,m}{r^{2}}}_{\text{gravity in}} = \underbrace{\frac{m\,v^{2}}{r}}_{\text{centripetal}}.

The orbiting star's own mass m cancels, and one factor of r comes out. Rearranging for the speed gives the master relation of galactic dynamics:

The rotation speed. A star on a circular orbit at radius r moves at v(r) = \sqrt{\frac{G\,M(r)}{r}}.
Only the enclosed mass matters. By Newton's shell theorem, a spherically distributed mass pulls as though it were all concentrated at the centre, and mass lying outside the orbit exerts no net force. So v(r) probes M(r), the mass within radius r — nothing beyond.
A rotation curve is a mass-o-meter. Measure v at every r and you can invert the relation, M(r) = v^{2} r / G, to weigh the galaxy shell by shell — without ever seeing the mass.

That last point is the whole game. Light tells you where the stars are. The rotation curve tells you where the mass is. If the two disagree, the light is lying about the mass.

Newton's prediction: the curve should fall

Now feed in what we can see. In a spiral galaxy the starlight is concentrated in the bulge and disk and drops off steeply; by the time you reach the visible edge, essentially all the luminous matter is already interior to you. Go farther out, into the dark region beyond the last stars, and you enclose no significant new mass: M(r) \approx M_{\text{disk}} = \text{const}. Drop a constant M into the master relation and the radius dependence stands out bare:

v(r) = \sqrt{\frac{G\,M}{r}} \;\propto\; r^{-1/2}.

This is exactly the Keplerian fall-off of the planets — quadruple the radius and the orbital speed halves. Beyond the luminous disk of a galaxy, Newton plus the visible matter predicts a rotation curve that peaks near the edge of the light and then declines, tracing a gentle 1/\sqrt{r} tail off to the right. That is the honest, well-motivated prediction, and for decades no one had any reason to doubt it.

The observation: flat, and it will not fall

The graph below is the crux of the whole subject. The falling curve is Newton's prediction from the visible matter alone — rising through the disk, peaking near the edge of the light, then sliding down its Keplerian 1/\sqrt{r} tail. The flat curve is what Rubin and Ford, and radio astronomers tracing the 21\ \text{cm} line of neutral hydrogen far past the optical disk, actually measured, galaxy after galaxy: the speed climbs and then simply levels off, holding roughly constant out to the last point they could detect gas.

Read the two curves side by side and the problem is inescapable. Where there is no more light, there should be no more mass, and the speed should drop — yet it does not drop. Stars and gas out in the dark are orbiting far too fast for the gravity of the matter we can see; at the visible edge they are often moving at twice the speed the luminous mass could sustain. Left as is, they would fly off into intergalactic space and the galaxy would unravel. They stay bound only if there is a great deal of extra, unseen mass out there gripping them. Move the slider and watch the falling prediction lift and flatten as you pour in halo mass: that is the amount of darkness the flat curve is demanding.

What a flat curve is telling us: an invisible halo

A flat rotation curve is not just a curiosity — it is a precise instruction about how the mass is arranged. Set v(r) = \text{const} in the master relation and solve for the enclosed mass:

v^{2} = \frac{G\,M(r)}{r} = \text{const} \quad\Longrightarrow\quad M(r) = \frac{v^{2}}{G}\,r \;\propto\; r.

The enclosed mass keeps growing linearly with radius — piling on more and more mass in every outer shell — in exactly the region where the starlight has run out and there is nothing to see. And we can go one step further and ask what density profile does that. The mass in a thin shell is dM = 4\pi r^{2}\rho(r)\,dr, so if M(r)\propto r then dM/dr = 4\pi r^{2}\rho(r) = \text{const}, which forces

\rho(r) \;\propto\; \frac{1}{r^{2}}.

Mass grows without light. A flat rotation curve requires M(r)\propto r — the total mass keeps increasing far beyond the luminous disk.
A dark halo. The extra mass follows \rho(r)\propto 1/r^{2}, an extended, roughly spherical halo that envelops the thin luminous disk and reaches many times farther out.
It dominates. Adding up the halo, most of a galaxy's mass — and of the universe's matter — is this unseen stuff. We call it dark matter: it has gravity but emits, absorbs, and reflects no detectable light.

So the luminous spiral we admire in photographs is only the bright, dense core of something far bigger: a vast, invisible, roughly spherical cloud of dark matter, within which the pretty disk of stars is merely the sediment that happened to shine.

Not one galaxy's quirk: five independent witnesses

If flat rotation curves were the only evidence, a cautious physicist would be right to wonder whether we simply misunderstand gravity on galactic scales. The reason dark matter is the consensus is that a whole jury of completely independent observations, on wildly different scales, all point the same way — and they agree on how much dark matter there is.

Galaxy clusters (Zwicky, 1933). Fritz Zwicky measured how fast galaxies fly around inside the Coma cluster and applied the virial theorem (for a bound system, 2\langle T\rangle + \langle U\rangle = 0, which ties the velocity spread to the total mass). The galaxies were moving so fast that the cluster's own visible mass could not possibly hold them together — it should have evaporated. He inferred a huge amount of unseen matter and coined the term dunkle Materie, "dark matter." He was forty years ahead of everyone.
Gravitational lensing. Mass bends light (general relativity). Weigh a cluster by how strongly it distorts the images of galaxies behind it, and the lensing mass consistently exceeds the luminous mass by a large factor — an independent scale, an independent physics, the same missing mass.
The Bullet Cluster. Two galaxy clusters have collided head-on. The ordinary matter is mostly hot X-ray gas, which drags and piles up in the middle; the galaxies (and any dark matter) sail on through. Lensing shows the bulk of the mass has passed clean through and sits out ahead of the gas, spatially offset from where nearly all the ordinary matter is. The gravity and the visible matter are in different places — a smoking gun that the mass is something other than the gas.
The cosmic microwave background. The relative heights of the acoustic peaks in the CMB power spectrum measure the mix of ordinary and dark matter in the infant universe. They require about five times as much dark matter as ordinary matter — beautifully consistent with galaxies and clusters, and set long before either had formed.
Structure formation. The cosmic web of galaxies grew from tiny density ripples. Simulations only reproduce today's web if gravity had extra, non-luminous mass — cold dark matter — to seed the collapse early, before ordinary matter (still coupled to light) could clump.

So what is it? Whatever the dark matter is, it must have mass, feel gravity, and barely interact with light or ordinary matter. It is not ordinary faint stuff — cold gas, dust, dead stars or black holes (collectively baryonic matter and MACHOs): those are made of protons and neutrons, and Big Bang nucleosynthesis plus the CMB tightly cap how much baryonic matter can exist, well below what is needed. The leading candidates are therefore new particles — WIMPs (weakly interacting massive particles) and axions — that lie outside the Standard Model and have so far escaped every direct-detection experiment.

A fair and serious question. The rival idea is MOND (Modified Newtonian Dynamics): maybe there is no dark matter, and instead gravity gets slightly stronger than 1/r^{2} at the very low accelerations found in a galaxy's outskirts. Tune that tweak and — impressively — flat rotation curves fall right out, with fewer free knobs than a dark-matter fit for many individual galaxies. So MOND is not a crank idea; it captures a real regularity.

Where it struggles is precisely where dark matter shines. In the Bullet Cluster, the gravitational mass (from lensing) sits offset from essentially all the visible matter. If gravity were merely a modified response to the seen matter, the extra pull would have to centre on that matter — the X-ray gas — but it doesn't; it centres on the empty-looking region the galaxies flew into. A lump of unseen mass that moved separately from the gas explains this immediately; a law that only amplifies visible matter cannot. Add the CMB peak ratios, which MOND has no natural account of, and the weight of evidence sits firmly with real, particulate dark matter.

Worked examples

Example 1 — weighing a galaxy from its flat curve. A spiral galaxy has a flat rotation speed v = 200\ \text{km/s} that holds out to r = 10\ \text{kpc}. Invert the master relation to find the enclosed mass. Using the handy unit combination G = 4.30\times 10^{-6}\ \text{kpc}\,(\text{km/s})^{2}\,M_\odot^{-1}:

M(r) = \frac{v^{2}\,r}{G} = \frac{(200)^{2}\,(10)}{4.30\times 10^{-6}}\ M_\odot = \frac{4.0\times 10^{5}}{4.30\times 10^{-6}}\ M_\odot \approx 9.3\times 10^{10}\ M_\odot.

Nearly 10^{11} solar masses inside 10\ \text{kpc} — yet the starlight there accounts for only a few times 10^{10}\ M_\odot. The rotation curve weighs in several times heavier than the light. The difference is dark.

Example 2 — the Milky Way at the Sun. Our own Sun orbits the Galactic centre at r \approx 8\ \text{kpc} with v \approx 220\ \text{km/s}. How much mass lies interior to the Sun's orbit?

M(<8\ \text{kpc}) = \frac{(220)^{2}\,(8)}{4.30\times 10^{-6}}\ M_\odot = \frac{3.87\times 10^{5}}{4.30\times 10^{-6}}\ M_\odot \approx 9\times 10^{10}\ M_\odot \approx 10^{11}\ M_\odot.

About a hundred billion solar masses — and because the curve stays flat well beyond the Sun, the mass keeps climbing outward, so the Milky Way's total mass (halo included) is more like 10^{12}\ M_\odot, most of it dark.

Example 3 — Keplerian versus flat, quantified. Compare a star at radius r with one four times farther out, at 4r. In the Keplerian (visible-only) prediction, v\propto r^{-1/2}, so

\frac{v(4r)}{v(r)} = \sqrt{\frac{r}{4r}} = \sqrt{\tfrac{1}{4}} = \frac{1}{2}.

The outer star should crawl along at half the speed. But the observed curve is flat, so in reality v(4r)/v(r) \approx 1 — the outer star moves just as fast as the inner one. That factor-of-two gap between "should be" and "is", repeated at every radius, is the fingerprint of the dark halo.

Two traps catch almost everyone. First: dark matter is not dark energy. They share an unfortunate adjective and nothing else. Dark matter clumps — it gathers into halos around galaxies and clusters and adds gravity, pulling things together and holding galaxies from flying apart. Dark energy is smooth — spread uniformly through space, it does the opposite, driving the accelerating expansion of the universe as a whole. One is extra pulling matter inside galaxies; the other is a repulsive pressure of empty space on cosmic scales. Do not let the vocabulary fool you into thinking they are the same thing wearing two hats.

Second: dark matter is not just faint ordinary matter — not cold gas, not dust, not dim dead stars or stray black holes. All of those are baryonic: made of the same protons and neutrons as you, and so they still emit, absorb or block light and would show up in infrared, radio or as absorption shadows. More decisively, the amount of baryonic matter is pinned down independently by Big Bang nucleosynthesis and the CMB, and it falls far short of the missing mass. The dark matter has to be something genuinely non-baryonic — a new kind of particle — not merely the ordinary stuff with the lights off.