Dark matter

Look up at a galaxy and you see a swirl of stars and glowing gas. Weigh all of it, then watch how fast the galaxy spins, and something is badly wrong. The outer stars are racing so quickly that the gravity of everything we can see should not be able to hold them. They ought to be flung off into space. They are not.

The explanation is that there is far more matter there than meets the eye. We call it dark matter, not because it is black, but because it gives off no light of any kind. We cannot see it, photograph it, or touch it. It just sits there, unseen, adding its gravity and holding galaxies together.

And there is a lot of it. For every kilogram of ordinary stuff, the stars and planets and gas and you, there are about five kilograms of dark matter. It wraps every galaxy in a giant invisible cloud, far bigger than the bright part, and long ago its gravity pulled ordinary matter together into the first galaxies. The whole structure of the universe grew on a scaffold we cannot see.

So how do we know it is real? By its gravity, in three ways: galaxies spin too fast to hold together without it, its mass bends passing starlight like a lens, and the faint afterglow of the Big Bang carries its fingerprint. The evidence is overwhelming. Yet despite decades of searching, no one has caught a single particle of it, and no one knows what it actually is.

The rotation problem. For a star orbiting in a galaxy, gravity sets its speed. Once you are outside most of the visible mass, Newton predicts the speed should fall with distance, like the outer planets moving slower than the inner ones:

\[ v(r) = \sqrt{\frac{G\,M(

If the mass \(M(

Three more lines of evidence. Rotation curves are only the start. Mass bends light, so a clump of dark matter acts as a gravitational lens, smearing background galaxies into arcs; lensing maps reveal far more mass than the visible matter can account for. The clinching case is the Bullet Cluster, two clusters that collided: the hot gas (most of the ordinary matter) slammed together and lagged behind, while the lensing mass sailed straight through with the galaxies, so the mass and the visible matter ended up in different places. And the afterglow of the Big Bang, the cosmic microwave background, has a pattern of ripples that only fits if the cosmos is about 5% ordinary matter, 27% dark matter and 68% dark energy.

What it is not, and might be. It is not ordinary matter, atoms, gas or dust, because the Big Bang's recipe and the microwave background both cap how much normal matter exists, and there is not nearly enough. It is not antimatter, and it is not just dim things like faint stars or black holes, which searches have largely ruled out. The leading idea is a new kind of particle that barely interacts with anything: candidates include WIMPs (heavy, weakly interacting), axions (extremely light), and sterile neutrinos. Experiments hunt them in deep underground detectors shielded from all other particles, in the sky for signs of them colliding, and in atom smashers. So far, nothing.

Or is gravity wrong? A minority view says there is no missing matter, and instead our law of gravity is slightly off on galactic scales. This idea (MOND) reproduces flat rotation curves remarkably well with no dark matter at all. But it struggles to explain the Bullet Cluster, where mass and matter come apart, and the detailed pattern of the microwave background, so most physicists conclude the dark matter is real and the puzzle is simply what it is made of.

The mass discrepancy. For a circular orbit, \(v^2(r) = G\,M(

The concordance. Independent probes converge on the same numbers. Weak and strong gravitational lensing map the projected mass directly. In the Bullet Cluster (1E 0657-558) the lensing centroids are spatially offset from the X-ray gas, a near model-independent argument for collisionless dark matter and a bound on its self-interaction cross-section. The acoustic peaks of the CMB pin the baryon and total matter densities separately through their effect on the photon-baryon fluid; with baryon acoustic oscillations and supernovae this gives the \(\Lambda\)CDM parameters \(\Omega_m \approx 0.31\), \(\Omega_b \approx 0.049\), \(\Omega_\Lambda \approx 0.69\). Dark matter must be cold (non-relativistic at decoupling) for structure to form hierarchically; hot dark matter would have erased small-scale structure by free-streaming.

Candidates and constraints. A viable particle must be stable on cosmological timescales, non-baryonic, and effectively collisionless. WIMPs, with weak-scale masses and cross-sections, give roughly the observed relic abundance through thermal freeze-out, the so-called WIMP miracle, but direct-detection experiments (XENONnT, LZ, PandaX) have excluded much of the natural parameter space without a signal. Axions, born to solve the strong-CP problem, are sought by haloscopes like ADMX. Sterile neutrinos at the keV scale are constrained by X-ray searches, including the disputed 3.5 keV line. Primordial black holes survive only in narrow mass windows after microlensing surveys; ordinary MACHOs are excluded.

The standing problem. On large scales \(\Lambda\)CDM is a triumph, fitting the CMB, lensing and the cosmic web in detail. The open issues are at small scales (the core-cusp and missing-satellites tensions) and, more fundamentally, that the particle has never been found. Modified-gravity theories such as MOND and its relativistic completion TeVeS capture the tight baryonic Tully-Fisher relation that pure dark matter does not obviously predict, but fail to reproduce cluster masses and the CMB without effectively reintroducing a dark component. After ninety years the gravitational evidence is overwhelming and the identity is unknown, one of the sharpest open questions in physics. It sits right next to the Big Bang and the unfinished account of gravity itself.