Dark matter

Dark matter is a theorised form of matter that accounts for most of the mass of the universe. It is called "dark" because it does not interact with the electromagnetic spectrum: it neither absorbs, reflects nor emits light, so it cannot be seen directly. Its existence is inferred from its gravitational effects on visible matter, radiation and the structure of the universe. According to the standard model of cosmology, the universe consists of about 5% ordinary (baryonic) matter, about 27% dark matter and about 68% dark energy.

Artist's illustration of the composition of the universe: roughly 68% dark energy, 27% dark matter and 5% ordinary matter. Dark matter outweighs ordinary matter by a factor of about five. Credit: NOIRLab/NSF/AURA/P. Marenfeld (CC BY 4.0).

Dark matter does not emit, absorb or reflect light, nor does it interact with ordinary matter through any force other than gravity in any way yet observed. Because it has mass, it takes up space and responds to gravity, and it is through those gravitational effects that astronomers "see" it: dark matter shapes the cosmos, organising galaxies and larger structures that would otherwise fly apart under their own rotation or expansion. It is thought to be the invisible glue that holds galaxies and galaxy clusters together, outweighing the visible stars and gas in most systems.

Discovery: the case for missing mass

Zwicky and the Coma cluster (1933)

The modern concept of dark matter took shape in the early 1930s. In 1933, the Swiss astronomer Fritz Zwicky applied the virial theorem to the Coma cluster, a large cluster of galaxies. By measuring how fast the individual galaxies move within the cluster, he estimated its total mass and compared it with the mass of the visible galaxies.

Zwicky found that the galaxies in the Coma cluster were moving far too quickly for the gravity of the visible matter to hold the cluster together: the galaxies should have been escaping, yet the cluster remained bound. He concluded that a vast amount of invisible matter must be present to supply the missing gravity, and he dubbed this material dunkle Materie, German for "dark matter". His estimate implied a mass-to-light ratio of several hundred solar masses per solar luminosity, far above what the stars alone could provide. At the time, however, the evidence rested on only a handful of galaxies and was regarded as a fringe idea.

Rotation curves of spiral galaxies

The evidence became compelling in the 1970s, when the American astronomer Vera Rubin and her collaborator W. Kent Ford Jr. measured how fast stars and gas orbit the centres of spiral galaxies. In a galaxy whose mass is concentrated where its light is concentrated, the orbital speed of a star should decline with distance from the centre, following a Keplerian falloff much as the planets orbit the Sun. Newtonian gravity fixes that speed through the mass enclosed within the orbit:

v_c(r) = \sqrt{\frac{G M(r)}{r}}

where G is the gravitational constant, M(r) the mass enclosed within radius r, and v_c(r) the circular orbital speed at that radius. If the mass traced the light, the rotation curve would rise in the inner galaxy and then fall off as the stars thinned out. What Rubin and Ford observed instead was that the rotation curves stayed flat: stars at the outskirts of the galaxies orbited just as fast as stars near the centre, far beyond the point where the visible light faded.

Expected (A) and observed (B) rotation curve of a spiral galaxy. The flat observed curve implies that the galaxy is embedded in a massive, extended halo of unseen matter. Credit: PhilHibbs (public domain).

In their landmark 1980 study of 21 Sc spiral galaxies, from the small NGC 4605 (radius 4 kpc) to the giant UGC 2885 (radius 122 kpc), Vera Rubin, W. Kent Ford Jr. and Norbert Thonnard found that the rotation curves remained flat to the last measured point in every galaxy. The only way to reconcile the flat curves with Newtonian gravity was to suppose that each spiral galaxy sits inside a large, roughly spherical halo of invisible matter whose mass continues to grow well beyond the visible disk. Rubin's work provided such strong evidence that the existence of dark matter became widely accepted.

Vera Rubin measuring spectra in 1972 at the Department of Terrestrial Magnetism of the Carnegie Institution. Her measurements of galaxy rotation curves provided some of the strongest evidence for dark matter. Credit: NOIRLab/NSF/AURA (CC BY 4.0).

The physics of dark matter

Dark matter reveals itself through dynamics: wherever astronomers can weigh a bound system from its internal motions, they find more mass than the luminous matter can explain. The same logic that Zwicky applied to the Coma cluster, and Rubin and Ford to spiral galaxies, is at work in galaxy clusters, where the random velocities of hundreds of galaxies can be used, through the virial theorem, to weigh the whole cluster. In every case the dynamically measured mass greatly exceeds the mass of the visible stars and gas.

The abundance of dark matter is normally expressed relative to the critical density, the density that would just halt the expansion of a flat universe:

\rho_c = \frac{3 H^2}{8 \pi G}

where H is the Hubble parameter and G the gravitational constant. Each cosmic component is then described by its density parameter:

\Omega_m = \frac{\rho_m}{\rho_c} = \frac{8 \pi G \rho_m}{3 H^2}

with the matter density parameter Ω_m comprising the baryonic and the dark matter contributions. Cosmological observations, most precisely from the cosmic microwave background, give a total matter density of about 31% of the critical density, of which roughly 27% of the universe (about 85% of all matter) is dark matter and only about 5% is ordinary baryonic matter.

Whatever dark matter is, it must be very different from the matter of the periodic table. It does not clump into stars or gas clouds, it does not collide and dissipate energy as ordinary gas does, and it interacts with itself and with ordinary matter at most very weakly. Simulations of cosmic structure formation show that only "cold" dark matter, made of slow-moving particles that gathered into clumps in the early universe, produces the web of galaxies and clusters actually observed; "warm" or "hot" dark matter made of faster particles would blur the smallest structures out of existence. In these simulations, the dark matter forms roughly spherical halos around every galaxy and cluster, with a characteristic density profile:

\rho(r) = \frac{\rho_s}{\frac{r}{r_s}\left(1+\frac{r}{r_s}\right)^2}

in which ρ_s is a scale density and r_s a scale radius. This profile, named after Julio Navarro, Carlos Frenk and Simon White, is the standard description of dark matter halos produced by cold dark matter simulations.

Modern evidence for dark matter

Gravitational lensing and the Bullet Cluster

A further line of evidence comes from gravitational lensing: the gravity of a massive object bends the light of background galaxies, and the amount of distortion reveals the total mass doing the bending, luminous or not. Because dark matter does not emit light, lensing offers a way to map its distribution directly.

Composite of the Bullet Cluster (1E 0657-56). Pink shows the X-ray-emitting hot gas of the colliding clusters; blue shows the total mass distribution traced by gravitational lensing. The bulk of the mass is separated from the gas and follows the galaxies, direct evidence for dark matter. Credit: NASA/CXC/M. Weiss (public domain).

The most dramatic demonstration came in 2006 from the Bullet Cluster (1E 0657-56), the aftermath of the collision of two galaxy clusters about 3.8 billion light-years away. When the two clusters crashed into each other, the hot gas of the two clusters, which is ordinary matter, collided, slowed down and was left behind in the middle of the collision site. The galaxies themselves, and the bulk of the mass traced by gravitational lensing, sailed through almost unaffected. The lensing maps show that most of the mass lies not with the gas but with the galaxies.

This separation of mass from ordinary matter is what makes the Bullet Cluster a direct empirical proof of the existence of dark matter: if the extra gravity were due to modified gravity rather than to a new kind of matter, it would have to follow the baryons, whereas the observations show it following the collisionless galaxies instead.

The cosmic microwave background

All-sky map of the cosmic microwave background from nine years of NASA's WMAP data; the temperature fluctuations are of ±200 microkelvin. The pattern of fluctuations encodes the composition of the universe. Credit: NASA/WMAP Science Team (public domain).

Independent and very precise evidence comes from the cosmic microwave background (CMB), the faint afterglow of the Big Bang. The pattern of hot and cold spots in the CMB depends on the amounts of ordinary matter, dark matter and dark energy in the universe, because each component affects how the primordial fluctuations grew. The final full-mission Planck measurements, published in 2020, find that ordinary baryonic matter makes up about 5% of the universe and dark matter about 27%, with the remaining roughly 68% being dark energy.

Large-scale structure

Dark matter is also required to explain how the large-scale structure of the universe formed. After the Big Bang, the universe was almost perfectly smooth, with only tiny density fluctuations. Ordinary matter alone could not have grown into today's galaxies and clusters: before recombination it was coupled to radiation and could not collapse, and after recombination there was too little time for gravity to gather it into galaxies. Dark matter, which decoupled from radiation early and interacts only gravitationally, could begin collapsing immediately, providing the gravitational seeds into which ordinary matter later fell to form galaxies. On the largest scales, galaxies, clusters, filaments and voids arrange themselves into the cosmic web, whose structure reflects the underlying dark matter scaffolding.

Candidate explanations

Since the gravitational evidence for dark matter is strong but no dark matter particle has ever been detected, its true nature remains one of the deepest open questions in physics. The leading idea is that dark matter is made of one or more new kinds of elementary particles, but modified theories of gravity that change how gravity behaves on large scales are also studied as an alternative to invoking new matter.

WIMPs

The most widely studied candidate is the weakly interacting massive particle, or WIMP: a heavy, slow-moving particle that interacts with ordinary matter only through gravity and, very rarely, through the weak force. WIMPs arise naturally in several extensions of the standard model of particle physics, and if they were produced thermally in the hot early universe, their present density would naturally fall near the observed dark matter abundance, a coincidence known as the "WIMP miracle".

Axions

Another leading candidate is the axion, a very light, low-energy particle first proposed in 1977 to solve the strong CP problem of particle physics. Axions would have been produced in abundance in the early universe and could collectively make up the dark matter; their interactions would be extremely feeble, making them difficult to detect directly, although telescopes such as NASA's Fermi, Chandra and NuSTAR have searched for the signatures of axion oscillations.

Other particle candidates

Many other candidates have been proposed, and it is possible that dark matter is made of more than one kind of particle. Sterile neutrinos, hypothetical heavier cousins of the three known neutrinos, are a popular "warm" dark matter candidate. Massive compact halo objects (MACHOs), such as dim stars, brown dwarfs and black holes, were once considered, but gravitational microlensing surveys have shown that such ordinary-matter objects cannot account for more than a small fraction of the dark matter. Primordial black holes, formed not from collapsing stars but from density fluctuations in the first instants after the Big Bang, remain a candidate that recent studies continue to test.

Modified gravity

Because all the evidence for dark matter is gravitational, an alternative explanation is that gravity itself behaves differently on the scales of galaxies and clusters than general relativity predicts. The most developed such theory is modified Newtonian dynamics (MOND), proposed by Mordehai Milgrom in 1983, in which the acceleration of a test particle becomes stronger than Newtonian expectation below a characteristic acceleration. MOND reproduces galaxy rotation curves remarkably well with no dark matter at all. However, the Bullet Cluster observations, in which the gravitational mass is demonstrably separated from the baryonic mass, are very difficult to reconcile with theories in which the anomalous gravity tracks the baryons; lensing, cluster and cosmological data therefore favour particle dark matter within the standard model of cosmology.

Current and future probes

As of September 2026, no experiment has detected a dark matter particle, and its nature remains an active area of research. Direct-detection experiments, mostly underground, search for the faint recoil of a dark matter particle striking an atomic nucleus; indirect searches look for the gamma rays, neutrinos or cosmic rays that dark matter particles might produce when they annihilate or decay; and particle accelerators attempt to produce dark matter in the laboratory. Astronomers meanwhile map the distribution of dark matter across the sky with gravitational lensing, using surveys from ground-based observatories and space missions such as NASA's Nancy Grace Roman Space Telescope, launched in August 2026, which will chart how dark matter is distributed and help answer how the universe is organised.

Further reading

• Dark Matter, NASA Science
• The Universe, NASA Science
• Rotational Properties of 21 SC Galaxies..., Rubin, Ford & Thonnard (1980)
• A Direct Empirical Proof of the Existence of Dark Matter, Clowe et al. (2006)