Dark Matter
Five-sixths of the matter in the universe has never been seen, touched, or directly measured. We know it is there because galaxies would otherwise fly apart, the cosmic microwave background would not fit, and the cosmic web would never have formed. Nearly a century after Fritz Zwicky's 1933 Coma Cluster measurements first demanded it, we still cannot say what it is.
The case
The evidence stacks from four independent directions, and they agree on the same number.
Galaxy rotation curves. In the 1970s, Vera Rubin and Kent Ford measured stars at the edges of spiral galaxies and found them orbiting just as fast as stars near the center. Newton predicts a falloff like the outer planets. Every spiral shows the flat curve instead — each one sitting inside an invisible halo extending five to ten times the visible radius.
The Bullet Cluster (1E 0657-558). Two galaxy clusters collided 150 million years ago, 3.7 billion light-years away. X-ray imaging shows the hot gas — the bulk of the visible mass — slowed by electromagnetic drag and lagging behind. Gravitational lensing shows the mass itself passed through unimpeded, sitting ahead of the gas. The mass and the visible matter are spatially separated. This single image is the cleanest argument that dark matter is a thing, not a theory of gravity in disguise.
The CMB power spectrum. Planck 2018 fit the acoustic peaks of the cosmic microwave background to within fractions of a percent: 68.3% dark energy, 26.8% dark matter, 4.9% baryonic matter. The third peak in particular is the dark matter peak — its height directly encodes the non-baryonic component. The numbers match galaxy surveys, Big Bang nucleosynthesis, and weak lensing independently.
Structure formation. N-body simulations without dark matter cannot grow the filaments, voids, and clusters we see. With cold dark matter, they do — at the right scale, at the right epoch.
Key facts
- 27% of cosmic energy content; 85% of all matter.
- Cold (non-relativistic at structure-formation epoch). This is the C in ΛCDM.
- Collisionless to the limits of current data: self-interaction cross-section per mass < 1 cm²/g from Bullet Cluster analysis.
- Distribution: roughly spherical halos extending to ~200 kpc around the Milky Way, ten times the visible disk.
- Ruled out: ordinary gas (would radiate), Standard Model neutrinos (too hot — they wash out small-scale structure), MACHOs as the dominant component (EROS/MACHO microlensing surveys, late 1990s).
The candidates, ranked by how badly 2025 hurt them
WIMPs. Weakly Interacting Massive Particles, 1–1,000 GeV. For 40 years the favored candidate because of the "WIMP miracle": a thermally produced particle at the weak scale naturally gives the observed relic density without tuning. LUX-ZEPLIN — 10 tonnes of liquid xenon, 1,478 metres underground in the Sanford Lab — published its December 2025 result after 417 live-days. No signal. The exclusion now cuts so deep into the natural parameter space that the canonical WIMP miracle window between 3 and 9 GeV is essentially closed. The community is not yet conceding defeat, but the framing has shifted from "when" to "if."
Axions. Originally proposed in 1977 (Peccei-Quinn, then Weinberg and Wilczek) to fix the Strong CP problem in QCD — not dark matter. Mass between 10⁻³³ and 10⁻¹⁰ GeV, so light they behave as a coherent classical field rather than discrete particles. ADMX (Fermilab) probes the 3–5 μeV range with resonant microwave cavities. As WIMP space closes, axion experiments are pulling theorist attention and DOE budgets.
Primordial black holes. Black holes formed in the first second of the universe, before any star. Microlensing surveys (EROS, MACHO, Subaru HSC) have ruled them out as the dominant component across most masses — but a stubborn asteroid-mass window between 10⁻¹⁶ and 10⁻¹⁰ solar masses remains open. LIGO's heavy black hole mergers (around 30 solar masses each) revived a separate, larger-mass PBH hypothesis briefly; data has since softened that case.
Sterile neutrinos. A 3.5 keV X-ray line reported in galaxy cluster spectra in 2014 was the leading hint. Subsequent observations (Hitomi briefly, then XRISM) have not confirmed it. Contested.
Two-component models. A 2026 proposal (Bhatt et al., JCAP, April 2026) posits two dark matter species whose annihilation requires both to meet — explaining why the Fermi gamma-ray excess from the Galactic Centre is absent in dwarf spheroidal satellites, which are dominated by one species. Speculative but solves a real puzzle. To verify: independent follow-up.
The detection frontier
| Experiment | Target | 2026 status |
|---|---|---|
| LUX-ZEPLIN | WIMPs, 5 GeV–1 TeV | No signal; running to 2028 |
| XENONnT | WIMPs | No signal; complementary limits |
| SuperCDMS SNOLAB | sub-GeV light DM | First physics data 2026 |
| ADMX | Axions, μeV range | Running |
| XLZD | 50-tonne xenon, post-LZ | Proposed; will reach the neutrino fog |
The neutrino fog is the floor: at sufficiently large detectors, solar and atmospheric neutrinos produce nuclear recoils indistinguishable from a WIMP. XLZD is designed to live in that fog and characterize it. Past that point, scaling up xenon stops helping.
What's contested
The deepest unresolved question is whether dark matter is one thing or many. Every candidate above solves part of the picture; none solves all of it cleanly. The WIMP miracle was a coincidence story, and after LZ that story is weaker than it has been in a generation.
Modified gravity is the minority alternative. MOND (Milgrom 1983) fits galaxy rotation curves with a single acceleration scale a₀ ≈ 1.2 × 10⁻¹⁰ m/s². TeVeS (Bekenstein 2004) is its relativistic extension. The Bullet Cluster is the standard counterargument — spatially separating mass from visible matter is hard to do without a real particle. But MOND's success at galactic scales is unreasonably good for a theory most cosmologists treat as wrong. Something about it is capturing a real regularity, and nobody quite agrees what.
There is also a possibility almost nobody publishes about: that 85% of the matter in the universe is permanently inaccessible to non-gravitational probes, and the next decade of null results will simply be the final word.
Why this has to do with other realms
Dark matter is the precondition for the rest of the wiki. Without the gravitational scaffold its halos provided in the first billion years, no galaxies — no Sun, no Earth, no concept great oxygenation event, no oxygen, no us. The same invisible substance that cosmologists chase in xenon tanks is also the reason there are atoms arranged into people doing the chasing. The 13.8-billion-year causal chain from a non-electromagnetic particle to a biosphere is one of the strangest dependencies in physics. It also makes concept bootes void interesting: the void is empty of galaxies because it was empty of dark matter first. Structure follows the invisible.
An open question
If LZ and XLZD reach the neutrino fog and find nothing, and ADMX exhausts its axion window without a peak — what is the experimental program for decade three? Is there one, or do we admit gravity has been telling us about something we cannot reach?
Key sources
- Rubin, V. & Ford, W.K. (1970) — the canonical rotation-curve paper, Astrophysical Journal.
- Clowe, D. et al. (2006) — "A Direct Empirical Proof of the Existence of Dark Matter," Bullet Cluster analysis, ApJL.
- Planck Collaboration (2020) — Planck 2018 results. VI. Cosmological parameters, A&A 641, A6.
- LUX-ZEPLIN Collaboration (December 2025) — most recent WIMP exclusion result. To verify: arXiv ID on publication.
- Bhatt et al. (April 2026) — two-component DM proposal, JCAP. To verify: full citation.
- Milgrom, M. (1983) — original MOND papers, ApJ 270.
Further reading
- The 4 Percent Universe by Richard Panek (2011) — the social history of how dark matter and dark energy entered the standard model, told through the people.
- Katherine Freese, The Cosmic Cocktail (2014) — a participant's account from one of the architects of WIMP detection.
- ADMX collaboration website — accessible technical briefings on axion search status.
- The Joe Walker Podcast episode with David Deutsch on what counts as evidence in cosmology — for the epistemology, not the physics.
- concept fermi paradox — if dark matter is PBHs, the distribution of habitable systems changes.
See Also
- concept fermi paradox — dark matter shapes where life can plausibly arise.
- concept bootes void — supervoids are where the scaffold is missing.
- concept cosmic strings — the competing structure-seeding mechanism, mostly ruled out but instructive.
- concept holographic principle — entropy bounds and the question of whether gravity is the only handle.
- concept great oxygenation event (cross-realm: biology depends on dark matter via galaxy formation) — the long causal chain from invisible halos to breathing animals.
- tech kugelblitz drive — engineered black holes, sibling to the primordial ones.
- concept arrow of time — cosmological initial conditions and the structure-growth direction.