Science › Big Bang & cosmos
Dark matter
1933· Open question
Well supportedStrong evidence or math, but not fully tested.
Invisible matter that outweighs ordinary matter about five to one, detected only through its gravity.
See it on the Science map →How it works
Galaxies spin too fast and clusters hold together too tightly for the visible matter they contain. Something unseen adds gravity. It does not glow, absorb light or collide much, so it passes through everything. Most physicists think it is an undiscovered type of particle. It makes up about 27% of the universe's contents, against about 5% for everything made of atoms. Fritz Zwicky first noticed it in a galaxy cluster in 1933.
Why it matters
Without it, galaxies would not have formed in the time available. It shapes the entire cosmic web, yet no known particle can be it.
Think of it like
Like discovering, from how a building settles, that it carries five times the load shown on the drawings.
How we know
Strong, independent lines: galaxy rotation (Rubin, 1970s), gravitational lensing, the 2006 Bullet Cluster collision, and microwave-background patterns. Direct detectors and the Large Hadron Collider have found no particle yet.
Still a mystery
What is it? Heavy particles, light 'axions', or something else? A minority argue gravity itself behaves differently instead.
Connections (13)
▸Vera Rubin found evidence for Dark matterProven
The relationship
Rubin measured orbital speeds across spiral galaxies. In a galaxy whose mass sits mostly in its bright center, outer stars should orbit slower. Instead, speeds stayed flat far out. Within known gravity, that requires much more mass than is visible, spread in a halo well beyond the stars. Her data made unseen mass impossible to dismiss.
How the link was found
Vera Rubin with Kent Ford, starting with Andromeda in 1970 and extending to many spiral galaxies by 1978-80. Fritz Zwicky had seen similar signs in a galaxy cluster in 1933.
Why it's proven
Radio measurements of hydrogen gas showed flat rotation far beyond the stars. Gravitational lensing and the microwave background independently confirm the missing mass.
Open questions and mysteries
Is the missing mass a particle, or does gravity change at low accelerations? Rotation curves alone cannot decide.
▸Dark matter is the hidden scaffolding of Galaxies and the cosmic webWell supported
The relationship
Dark matter does not feel light's pressure, so in the early universe it began clumping first, while ordinary gas was still being pushed around by radiation. Its clumps became gravitational wells. Ordinary gas later fell in, forming galaxies at the densest points. The filaments of the cosmic web trace the underlying dark matter skeleton.
How the link was found
Developed in the 1980s; the 1985 'cold dark matter' simulations of Davis, Efstathiou, Frenk and White showed web-like structure. The 2005 Millennium Simulation made it vivid.
Is it proven?
Simulations with dark matter match galaxy surveys; without it, galaxies would not have formed in time. Gravitational-lensing maps show dark matter along filaments.
Open questions and mysteries
On small scales, simulations predict more tiny satellite galaxies and denser galaxy centers than are seen. Is this dark matter's nature or messy gas physics?
▸Cosmic microwave background: the Big Bang's afterglow measures the amount of Dark matterWell supported
The relationship
Before atoms formed, ordinary matter and light sloshed back and forth like sound waves, while dark matter sat still and just pulled. The balance between the two sets the relative heights of the ripples in the microwave background. Reading those heights gives an inventory: dark matter is about five times ordinary matter.
How the link was found
Worked out by Jim Peebles, Rashid Sunyaev and others in the 1970s-80s; measured precisely by WMAP (2003 onward) and Planck (2013-2018).
Is it proven?
The ordinary-matter amount read from the background matches the amount implied by Big Bang element abundances, two independent methods agreeing.
Open questions and mysteries
The background weighs dark matter but cannot say what it is made of.
▸The Standard Model cannot explain Dark matterOpen question
The relationship
Dark matter must be stable, heavy enough to clump, and almost invisible to light. No Standard Model particle fits. Neutrinos come closest but are far too light and fast-moving; they would smear out galaxy formation. So if dark matter is a particle, it lies outside the Standard Model, a clear sign the model is incomplete.
How the link was found
Fast neutrinos were ruled out as the main dark matter by structure-formation simulations in the early 1980s.
Is it proven?
Not a proof that new particles exist, but a strong inference. Large Hadron Collider searches and underground detectors such as LZ have found no candidate so far.
Open questions and mysteries
Is dark matter one particle or a whole hidden sector? Could it interact with ordinary matter only through gravity, making it undetectable?
▸Gravitational lensing maps the hidden mass of Dark matterProven
The relationship
Lensing measures all the mass in a galaxy cluster, visible or not, by how much it bends light from galaxies behind it. Comparing that with the visible stars and gas shows roughly five times more mass than we can see. Weak-lensing surveys turn the subtle distortions of millions of galaxies into maps of where the hidden mass sits.
How the link was found
Cluster-lensing studies grew through the 1990s. Douglas Clowe's team mapped the Bullet Cluster in 2006, showing mass separated from the visible gas.
Why it's proven
In the Bullet Cluster, lensing mass sits where the galaxies are, not where most normal matter (hot gas) is, which strongly favors unseen dark matter.
Open questions and mysteries
Lensing shows where dark matter is but not what it is. No dark-matter particle has been detected in any laboratory.
▸Dark matter gathers gas into clumps during The cosmic dark agesWell supported
The relationship
Dark matter had been clumping since before atoms formed, because light's pressure never held it back. When gas went neutral at 380,000 years, it was freed from light too and fell into dark matter's existing gravity wells. That head start was essential: the ordinary-matter ripples seen in the microwave background, about 1 part in 100,000, could not alone have grown into galaxies by now.
How the link was found
Jim Peebles showed in 1982 that cold, slow dark matter could reconcile the smooth microwave background with today's galaxies; George Blumenthal, Sandra Faber, Joel Primack and Martin Rees developed the picture in 1984.
Is it proven?
Simulations with cold dark matter reproduce galaxy timing and the cosmic web; without it, galaxies form far too late. Direct observations of the dark ages are still missing.
Open questions and mysteries
The 21-centimeter signal could show whether dark matter feels anything besides gravity. EDGES' 2018 report of an unexpectedly deep dip at cosmic dawn, about 180 million years, prompted such ideas; SARAS 3 (2022) found no sign of it.
▸Fritz Zwicky first proposed Dark matterProven
The relationship
Zwicky measured how fast galaxies in the Coma Cluster move relative to one another, roughly 1,000 km/s. In a cluster that holds together, that motion must be balanced by gravity, so the speeds reveal the total mass. Comparing it with the cluster's light, he found far too little visible matter, estimating at least 400 times more mass than glowed, and called the rest dark matter.
How the link was found
Fritz Zwicky, 1933, in the Swiss journal Helvetica Physica Acta, expanded in 1937. Sinclair Smith found the same problem in the Virgo Cluster in 1936.
Why it's proven
His factor was too large: the expansion rate he used was far off, and clusters' hot X-ray gas, unknown then, holds much ordinary matter. Even including that gas, modern lensing and X-ray data leave clusters about 85% dark matter.
Open questions and mysteries
Clusters are where MOND still needs unseen mass, so Coma-type clusters remain a key battleground. What the missing mass actually is remains unknown.
▸The Bullet Cluster: dark matter caught in the act shows, separated from gas, Dark matterProven
The relationship
In a cluster collision, the hot gas clouds slam together and slow, like two water balloons meeting. Galaxies, being far apart, pass through like two crowds walking through each other. Dark matter that barely interacts should keep going with the galaxies. That is what lensing shows: most mass is with the galaxies, though the gas outweighs the stars several times. So the extra mass is neither gas nor stars.
How the link was found
Clowe and colleagues framed it this way in 2006. Markevitch's team (2004) and Scott Randall's (2008) used the same collision to measure how much dark matter can bump into itself.
Why it's proven
The mass is offset from the gas at 8σ. The dark clumps were barely slowed, which caps how strongly dark matter can collide with itself.
Open questions and mysteries
Could dark matter collide with itself weakly, just under that cap? Some astronomers favor such 'self-interacting' dark matter to explain why small galaxies' centers are less dense than simulations predict.
▸Dark matter may be made of WIMPs: heavy, shy particlesSpeculative
The relationship
In the hot early universe, WIMPs and their antiparticles would be constantly created and destroyed in pairs. As space cooled and thinned, they stopped finding each other, and a fixed amount 'froze out'. For a particle with weak-force-strength interactions and a mass near 100 protons, the leftover matches today's dark matter. Physicists called this coincidence the 'WIMP miracle'.
How the link was found
Benjamin Lee and Steven Weinberg (1977), among others, calculated such leftovers for heavy neutrinos. Mark Goodman and Edward Witten proposed in 1985 detecting them by their rare bumps into atomic nuclei.
Is it proven?
Not proven. The freeze-out arithmetic is solid, but the 'miracle' may be coincidence. Every direct, indirect and collider search so far has come up empty, ruling out the simplest WIMP models.
Open questions and mysteries
Are WIMPs ruled out, or just shyer than hoped? Detectors face a floor of neutrino noise. Heavier or lighter dark particles, outside the classic range, remain possible.
▸Dark matter may be made of Axions: ultralight particlesSpeculative
The relationship
Axions would not come from heat. As the young universe cooled, the axion field was left displaced from its resting point and began to oscillate everywhere at once. That oscillation behaves exactly like vast numbers of cold, slow particles. Though each axion is tiny, together they could supply all of dark matter, if the mass sits in the right range.
How the link was found
Three groups in 1983: John Preskill, Mark Wise and Frank Wilczek; Larry Abbott and Pierre Sikivie; Michael Dine and Willy Fischler. Sikivie also proposed the magnetic-field detection method that year.
Is it proven?
Not proven. The production mechanism follows from established physics if axions exist. No axion has been detected; experiments have excluded only slices of the possible mass range.
Open questions and mysteries
Did axions form before or after inflation? The answer changes their predicted mass and whether they clump into tiny 'miniclusters' that lensing might find.
▸Dark matter may be made of Primordial black holes as dark matterSpeculative
The relationship
Primordial black holes would behave just like cold dark matter: dark, heavy, slow, and gravitating without colliding. Crucially, they formed within the first second, before nuclei were cooked, so they do not count against the strict budget of ordinary matter set by hydrogen and helium abundances. Their mass would come from the radiation and particles swallowed in the collapse.
How the link was found
George Chapline proposed them as dark matter in 1975. Interest surged in 2016, when Simeon Bird and colleagues asked whether LIGO's first black holes could be dark matter.
Is it proven?
Not proven, and mostly ruled out: lensing, gamma rays, the afterglow and gravitational waves exclude them as all of dark matter across most masses, except roughly asteroid masses, where no test yet reaches.
Open questions and mysteries
Could dark matter be a mix of primordial black holes and particles? A black hole lighter than any star could make, found by lensing or gravitational waves, would be a smoking gun.
▸Dark matter is replaced by modified gravity in MOND: modified gravity instead of dark matterSpeculative
The relationship
Galaxy rotation reveals a gap between visible mass and the gravity required. Dark matter fills the gap with unseen mass; MOND instead changes the force law, so the gap is not missing mass but missing gravity. MOND's strength is galaxies, where it predicts rotation from visible matter alone. Dark matter's strength is everything larger: clusters, the afterglow and the cosmic web.
How the link was found
Milgrom (1983) framed it as an alternative. Stacy McGaugh and Erwin de Blok showed in 1998 that MOND had correctly predicted rotation in dim, diffuse galaxies before it was measured.
Is it proven?
Neither is proven by direct detection. Most physicists favor dark matter because one ingredient explains many independent observations; MOND cannot fit the afterglow ripples without adding unseen matter.
Open questions and mysteries
Why does visible mass alone predict galaxy rotation so tightly, as McGaugh, Federico Lelli and James Schombert showed across 153 galaxies in 2016? Dark-matter models must reproduce that regularity.
▸ΛCDM: the standard model of cosmology includes cold Dark matterWell supported
The relationship
ΛCDM treats dark matter as cold, meaning slow-moving early on, and as feeling only gravity. Cold matter clumps on every scale, so small clumps form first and merge into larger ones: galaxies before clusters. Fast-moving 'hot' dark matter, like neutrinos, would have smoothed out small clumps and built giant sheets first. Observations match cold, bottom-up growth. It makes up about 85% of all matter.
How the link was found
Simon White, Carlos Frenk and Marc Davis showed in 1983 that neutrino ('hot') dark matter cannot match how galaxies cluster. In 1990 George Efstathiou, Will Sutherland and Steve Maddox showed cold dark matter fits galaxy surveys better when paired with a cosmological constant.
Is it proven?
Galaxy clustering, the afterglow ripples and gas clouds seen against distant quasars all fit cold dark matter; the quasar data limit how 'warm' it can be.
Open questions and mysteries
Is it perfectly cold and collisionless, or slightly warm, self-interacting or wave-like? Each variant changes the smallest galaxies, where puzzles remain.