What is dark matter? The invisible mass holding galaxies

Dark matter is the name for mass in the universe that we can detect only by its gravity: it doesn't shine, absorb or reflect light, so telescopes can't see it. Galaxies spin too fast, light bends too much, and the early universe clumped too quickly for the visible matter alone to explain. The best measurements say there is about five times as much dark matter as ordinary matter. What dark matter actually is remains unknown.

In short

• Dark matter is inferred, not seen: several independent measurements all need extra mass that gives off no light.
• The leading idea is a new kind of particle that barely interacts with anything except through gravity; decades of searches haven't confirmed one.
• A minority of physicists argue the evidence points instead to gravity behaving differently at large scales.

Why do scientists think dark matter exists?

Because gravity keeps adding up wrong. In 1933, the astronomer Fritz Zwicky measured how fast galaxies move inside the Coma cluster and found them moving so fast that the cluster should fly apart unless it held far more mass than its starlight suggested. In the 1970s, Vera Rubin and Kent Ford measured how fast stars and gas orbit in spiral galaxies, starting with Andromeda, and found the same problem inside single galaxies. Since then, at least three independent kinds of evidence have pointed the same way.

Evidence	What is measured	What it shows
Galaxy rotation	How fast stars and gas orbit at different distances from a galaxy's center	Outer stars orbit too fast for the visible mass to hold them
Gravitational lensing	How much a galaxy cluster bends light from galaxies behind it	Clusters hold several times more mass than their stars and hot gas
The cosmic microwave background	The pattern of hot and cold spots in the oldest light, from about 380,000 years after the [[what-is-the-big-bang	Big Bang]]

The third is the most precise. In its final results, the Planck satellite team measured the density of dark matter and of ordinary matter separately, as 0.120 and 0.0224 in the units cosmologists use. Divide one by the other and dark matter comes out about 5.4 times ordinary matter. In shares of the whole universe, that is roughly 27% dark matter and 5% ordinary matter, with dark energy making up the rest, as the universe note sets out.

How do spinning galaxies reveal it?

Start where the answer is known: the solar system. Almost all its mass is in the Sun, so a planet's orbital speed falls the farther out it is, in a predictable way. Speed goes as one over the square root of distance. Earth, 1 astronomical unit (AU) from the Sun, orbits at about 29.8 km per second. Neptune is 30.1 AU out, so its speed should be 29.8 divided by the square root of 30.1, which is 5.4 km per second. That is what Neptune does.

Now apply the same rule to the Milky Way. Most of its stars, and most of its light, are concentrated toward the center. If that visible matter were all the mass, stars far out should orbit more slowly, the way Neptune does. Here is what is measured, from a 2019 study by Anna-Christina Eilers and colleagues that used more than 23,000 stars.

1. At the Sun's distance, about 8.1 kiloparsecs (26,500 light-years) from the center, stars orbit at 229 km per second.
2. Predict farther out. Suppose, to keep it simple, that almost all the mass is inside the Sun's orbit. At 25 kiloparsecs, about three times farther, speed should drop by the square root of about 3.1: from 229 to about 130 km per second.
3. Measure. The study found the speed declines only gently, by 1.7 km per second for each kiloparsec, so at 25 kiloparsecs it is still about 200 km per second.
4. Work out the mass. The mass needed inside an orbit is speed squared times distance, divided by the gravitational constant. That gives about 100 billion Suns' worth inside the Sun's orbit and about 230 billion inside 25 kiloparsecs.

So the mass inside more than doubles between the Sun's orbit and 25 kiloparsecs, though most of the light is already inside the first. The simple formula treats the galaxy's mass as if it were a ball, and a flat disk changes the numbers by tens of percent, but no reasonable shape removes the gap. Something with mass, but without light, surrounds the galaxy in a vast halo.

What did the Bullet Cluster show?

It separated the ordinary matter from the mass. The Bullet Cluster is two clusters of galaxies that collided about 150 million years before the light now reaching us left them. Most of a cluster's ordinary matter is hot gas between the galaxies, which glows in X-rays. When the clusters passed through each other, the gas clouds collided and slowed in the middle, while the galaxies sailed on.

In 2006, Douglas Clowe and colleagues mapped where the mass is from how it bends light from galaxies behind. The mass wasn't with the gas, where most of the ordinary matter is. It was with the galaxies, out ahead, moving like something that passed through the collision without hitting anything. That is what a cloud of weakly interacting particles would do, and it is hard to explain by changing the law of gravity alone.

What could dark matter be?

Nobody knows, but the candidates have to pass strict tests. Dark matter must have mass, barely interact with light or ordinary matter, and move slowly enough to clump into the halos galaxies form in. That last property is why the standard model of cosmology calls it cold dark matter.

• WIMPs, weakly interacting massive particles, were the favorite for decades: new particles perhaps tens to thousands of times a proton's mass, which would occasionally bump into an atomic nucleus.
• Axions are very light hypothetical particles first proposed to fix a puzzle in the theory of the strong nuclear force; they might turn into faint radio waves in a strong magnetic field.
• Primordial black holes, formed in the first instants after the Big Bang rather than from dead stars, could act like dark matter, though observations have ruled them out over most mass ranges.
• Neutrinos, known particles with tiny masses, do count as a little dark matter, but they move too fast to build galaxies, so they can't be most of it.

Searches go three ways: wait for dark matter to hit a detector, make it in a collider, or catch the light from dark matter particles destroying each other in space. The Large Hadron Collider at CERN has found no new particle of the right kind. Underground detectors have grown huge to catch a rare collision. The LUX-ZEPLIN (LZ) experiment, which watches about 7 tonnes of liquid xenon almost a mile underground in South Dakota, reported in 2024 no excess over known backgrounds, the tightest limit yet on many WIMPs.

On 1 September 2026, LZ reported a single unexplained event in a search for a wider range of WIMP interactions. According to Brown University's announcement, its significance is 2.6 sigma, about a 0.5% chance that known backgrounds would produce it, far short of the 5 sigma physicists require for a discovery. The collaboration's spokesperson said: "We are not claiming to have seen dark matter." More data will show whether it grows or fades.

Could gravity be wrong instead?

Still argued

In 1983, the physicist Mordehai Milgrom proposed Modified Newtonian Dynamics (MOND): that gravity weakens more slowly than Newton's law says when accelerations are extremely small, as at the edges of galaxies. With a single extra constant, MOND predicts the rotation of many galaxies from their visible matter alone, often more neatly than dark matter models do, and its supporters see that as evidence that galaxies are telling us something about gravity. Most cosmologists favor dark matter because MOND struggles with galaxy clusters, where it still needs some unseen mass, with the Bullet Cluster, and with the pattern in the cosmic microwave background, which dark matter fits precisely. Newer relativistic versions of MOND try to fix those problems, and tests using widely separated pairs of stars have produced conflicting results. The question is open in the sense that no dark matter particle has been found; as of 2026, most of the evidence favors some form of dark matter.

The idea that dark matter is just ordinary matter we can't see, cold gas or dim stars, has largely been ruled out. Surveys for dark objects in the Milky Way's halo found too few, and the cosmic microwave background limits the total amount of ordinary matter. Dark matter is also not dark energy: dark matter pulls things together, while dark energy is the name for whatever is pushing the expansion of the universe to speed up. Both describe effects of gravity that the known contents of the universe don't account for.

Check yourself

• Without looking back: why should outer stars orbit more slowly if only visible matter pulled on them, and what is measured instead?
• What did the Bullet Cluster show that rotation curves alone could not?
• Why can't neutrinos, which do have mass, be most of the dark matter?

Sources

• Planck 2018 results. VI. Cosmological parameters — Planck Collaboration, Astronomy and Astrophysics, 2020
• The circular velocity curve of the Milky Way from 5 to 25 kpc — Eilers and others, The Astrophysical Journal, 2019
• A direct empirical proof of the existence of dark matter — Clowe and others, The Astrophysical Journal Letters, 2006
• Dark matter search results from 4.2 tonne-years of exposure of the LUX-ZEPLIN (LZ) experiment — LZ Collaboration, 2024
• LZ experiment sees surprising result in search for dark matter — Brown University, 2026
• Dark matter — CERN