Dark matter

Dark matter is the single largest piece of the material universe that the Standard Model cannot name. Weigh a spiral galaxy by the light of its stars and you get one number; weigh it by how fast those stars orbit and you get a number five to ten times larger. The discrepancy does not shrink with better telescopes — it appears again in the velocities of galaxies swarming inside clusters, in the way clusters bend the light of objects behind them, and, most precisely of all, in the pattern of hot and cold spots frozen into the cosmic microwave background. Every one of these measurements says the same thing: most of the gravitating matter in the cosmos neither shines nor absorbs light, and it is not made of the protons and neutrons that make up everything we have ever touched. This article is where the astrophysical evidence meets particle physics. The Standard Model Lagrangian — Peskin's compact $\mathcal{L}=q\,i\!\not{\!D}\,q+\ell\,i\!\not{\!D}\,\ell-\tfrac14(F^a_{\mu\nu})^2+|D_\mu\phi|^2-V(\phi)-(\lambda_{ij}\text{ Yukawa}+\text{h.c.})$ — contains no particle that is at once stable, electrically neutral, colour-neutral, and heavy enough to be cold. The neutrinos are neutral but far too light and too fast; every other particle is charged, coloured, or short-lived. So dark matter is a genuine command to look beyond the Standard Model. The most-studied answer, the weakly interacting massive particle, ties the cosmic abundance to physics at the electroweak scale through a thermal calculation of startling economy — the "WIMP miracle" — and it is the reason dark matter, supersymmetry, and collider searches are one conversation and not three.

Key takeaways

Dark matter is the dominant non-luminous, non-baryonic matter of the universe: independent evidence from rotation curves, cluster dynamics, gravitational lensing, and the cosmic microwave background all converge on it.
It is about 27% of the cosmic energy budget — roughly five times the ordinary (baryonic) matter — with the measured density fixed to Ω_DM h² = 0.120.
The Standard Model has no candidate: no particle in the Lagrangian is simultaneously stable, neutral, colour-neutral, and cold, so dark matter is a concrete motivation for new physics.
A thermal relic freezes out of the early plasma when annihilation can no longer keep pace with expansion; the survivor abundance scales as Ω_χ h² ∝ 1/⟨σv⟩.
The WIMP miracle: a weak-scale mass (~100 GeV) and coupling automatically give ⟨σv⟩ ~ 3×10⁻²⁶ cm³/s, the exact cross section that reproduces the observed relic density.
Key idea: dark matter is the beyond-the-Standard-Model matter whose gravitational presence is proven six ways over, and whose most economical explanation — a thermal weak-scale relic — ties its cosmic abundance directly to particle physics at the electroweak scale.

The core insight

Dark matter is the beyond-the-Standard-Model matter whose gravitational presence is proven six independent ways, and whose leading explanation is a stable, neutral, weak-scale particle whose annihilation cross section — through thermal freeze-out — fixes its cosmic abundance to Ω_χ h² ∝ 1/⟨σv⟩.

Where this lives today

The hunt for dark matter is pursued on three fronts. Direct detection watches for a dark-matter particle recoiling off an atomic nucleus in an ultra-quiet underground detector: multi-tonne liquid-xenon time-projection chambers — LZ in South Dakota, XENONnT under Gran Sasso, PandaX in China — now probe spin-independent WIMP–nucleon cross sections below $10^{-47}\ \mathrm{cm^2}$ for masses near 30–50 GeV, and are closing in on the "neutrino fog," where coherent scattering of solar and atmospheric neutrinos mimics a signal. Indirect detection looks for annihilation products in dense regions: the Fermi-LAT gamma-ray telescope stares at dwarf spheroidal galaxies (dark-matter-dominated and star-poor), the persistent Galactic-centre gamma-ray excess remains tantalising but ambiguous, and IceCube searches for neutrinos from annihilation in the Sun. Colliders hunt for dark matter produced and escaping unseen: the LHC looks for events with large missing transverse momentum recoiling against a jet or photon (mono-jet, mono-photon signatures). So far all three have returned null results across most of the classic WIMP parameter space, and the field has broadened accordingly — resonant-cavity haloscopes like ADMX search for ultralight axions (which also solve the strong-CP problem), keV sterile neutrinos would be warm dark matter, and primordial black holes and feebly interacting freeze-in relics are under active study. The most important open problem in the subject is unchanged since 1933: what is it made of?

The physical picture

The founding observation is kinematic. A star far out in a galaxy's disk feels the gravity of everything interior to its orbit; balancing that pull against the centripetal demand of circular motion gives an orbital speed $v(r)=\sqrt{GM(r)/r}$ . Outside the visible starlight the enclosed luminous mass stops growing, so the speed should fall away as $1/\sqrt{r}$ , exactly as the planets slow with distance from the Sun. It does not. Measured rotation curves stay flat, out to radii far beyond the last visible star, and a flat curve means the enclosed mass keeps rising in lockstep with radius where there is no light to account for it. The first figure is that confrontation: the Keplerian decline the stars predict, against the flat curve the galaxy actually shows.

Figure

Galactic rotation curve — the orbital speed stays flat at v ≈ 220 km/s far beyond the visible disk, instead of falling off as 1/√r the way the starlight alone predicts. A flat curve forces M(r) ∝ r and a halo density ρ(r) = v²/(4πGr²) ∝ 1/r², the shaded dark matter that dominates the outer galaxy

The same missing mass shows up wherever gravity is weighed against light, and one image settles the case that it is real matter and not a failure of gravity. In the Bullet Cluster two galaxy clusters have passed through each other; the hot X-ray gas — the bulk of the baryons — was slowed by ram pressure and piled up in the middle, while the total mass, mapped independently by gravitational lensing, sailed straight through with the nearly collisionless galaxies. The mass and the baryons ended up in different places. The second figure is the cosmic ledger that all of this feeds into: dark matter is not a rounding error but the second-largest term in the entire energy budget of the universe, dwarfing everything the Standard Model describes.

Figure

The cosmic energy budget from Planck — dark energy 68.5%, dark matter 26.5%, ordinary baryonic matter only 4.9%. Dark matter outweighs everything the Standard Model describes by a factor of about 5.4, with its density pinned to Ω_DM h² = 0.120 from the acoustic peaks of the microwave background

The core equations

The evidence begins with Newton. A star on a circular orbit is held by the gravity of the mass interior to it, so its speed measures that enclosed mass directly — and the way the speed behaves at large radius diagnoses how the mass is distributed.

v(r)=\sqrt{\frac{G\,M(r)}{r}}\ \xrightarrow{\ v=\text{const}\ }\ M(r)\propto r,\qquad \rho(r)=\frac{v^2}{4\pi G\,r^2}\propto\frac{1}{r^2} \qquad (1)

where: $v(r)$ — circular orbital speed at radius $r$ ; $M(r)$ — total mass enclosed within $r$ ; $G$ — Newton's constant; $\rho(r)$ — mass density of the halo. A flat $v(r)$ is the whole story: it forces the enclosed mass to grow linearly with radius, which requires an extended halo whose density falls only as $1/r^2$ — the "isothermal sphere" — long after the starlight has died away. Cosmology weighs the universe as a whole. Densities are quoted as fractions $\Omega_i$ of the critical density that separates an ever-expanding from a recollapsing universe, and the microwave background fixes these fractions to per-cent precision.

\Omega_i=\frac{\rho_i}{\rho_c},\qquad \rho_c=\frac{3H_0^2}{8\pi G};\qquad \Omega_{\rm DM}h^2=0.120,\quad \Omega_{\rm DM}\approx0.265,\quad \frac{\Omega_{\rm DM}}{\Omega_{\rm b}}\approx5.4 \qquad (2)

where: $\rho_c$ — critical density; $H_0=100\,h\ \mathrm{km\,s^{-1}\,Mpc^{-1}}$ with $h\approx0.674$ ; $\Omega_{\rm DM}$ , $\Omega_{\rm b}$ — dark-matter and baryon density fractions. Writing $\Omega h^2$ removes the uncertainty in $h$ : the combination $\Omega_{\rm DM}h^2=0.120$ is what the acoustic peaks measure, and the ratio to baryons, $\Omega_{\rm b}h^2=0.0224$ , is independently checked by Big Bang nucleosynthesis — the two agree, which is why the dark matter must be non-baryonic. To turn that abundance into particle physics we track the number density $n$ of a would-be dark-matter species $\chi$ as the early universe expands and cools. Its evolution is a competition between dilution, annihilation, and creation — the Boltzmann equation.

\frac{dn}{dt}+3Hn=-\langle\sigma v\rangle\left(n^2-n_{\rm eq}^2\right) \qquad (3)

where: $H$ — Hubble expansion rate; $3Hn$ — dilution by the expanding volume; $\langle\sigma v\rangle$ — thermally averaged annihilation cross section times relative velocity; $n_{\rm eq}$ — the equilibrium number density. When $\chi\chi\to\text{SM}$ annihilation is fast the density tracks its equilibrium value; when expansion outruns annihilation the comoving number "freezes out" and is preserved as the relic we see today. Solving equation (3) across freeze-out gives the payoff — the surviving abundance is set almost entirely by the annihilation cross section, and it is inversely proportional to it.

\Omega_\chi h^2\approx 0.12\times\frac{3\times10^{-26}\ \mathrm{cm^3\,s^{-1}}}{\langle\sigma v\rangle},\qquad \langle\sigma v\rangle_{\rm therm}\approx 3\times10^{-26}\ \mathrm{cm^3\,s^{-1}}\approx 2.6\times10^{-9}\ \mathrm{GeV^{-2}} \qquad (4)

where: $\Omega_\chi h^2$ — relic density of $\chi$ ; $\langle\sigma v\rangle$ — its annihilation cross section. The reading is the crux of the whole subject: a particle that annihilates more efficiently leaves less of itself behind, and the specific cross section that reproduces the measured $\Omega_{\rm DM}h^2=0.120$ is the "thermal relic" value $\approx3\times10^{-26}\ \mathrm{cm^3\,s^{-1}}$ — which, remarkably, is what weak-scale physics naturally supplies.

Derivation AFlat rotation curves imply an extended dark halo

The goal is to show, from Newtonian mechanics alone, that the flat rotation curves observed in spiral galaxies are not merely surprising but logically demand a distribution of unseen mass whose density falls as $1/r^2$ and whose total keeps growing well past the visible edge. We start from the balance of gravity and circular motion, read off what the starlight predicts, confront it with observation, and then invert the measured speed into a halo density profile — turning a kinematic anomaly into a statement about where the mass is.

Step 1. For a star on a circular orbit of radius $r$ , the gravitational pull of the interior mass supplies the centripetal acceleration: $v^2/r=GM(r)/r^2$ , hence $v(r)=\sqrt{GM(r)/r}$ . Why: only the mass enclosed within the orbit matters (Newton's shell theorem), so the orbital speed is a direct scale that weighs everything interior to $r$ .

Step 2. Outside the luminous disk the starlight — and with it the luminous mass — falls off steeply, so $M(r)\to M_{\rm lum}=\text{const}$ and the prediction is a Keplerian decline, $v(r)\propto1/\sqrt{r}$ . Why: with no more mass being added, a wider orbit feels weaker gravity, exactly as Neptune orbits the Sun more slowly than Mercury.

Step 3. Observed curves are instead flat, $v(r)\approx v_0=\text{const}$ , out to many disk radii. Setting $v_0=\sqrt{GM(r)/r}$ constant forces $M(r)=v_0^2\,r/G\propto r$ — the enclosed mass grows linearly with radius in a region where there is essentially no light. Why: a constant speed can only be maintained if the extra gravity from extra enclosed mass keeps exactly compensating the $1/r$ weakening, which pins $M(r)\propto r$ .

Step 4. Differentiate the enclosed mass to get the local density of a spherical halo: $\rho(r)=\dfrac{1}{4\pi r^2}\dfrac{dM}{dr}=\dfrac{v_0^2}{4\pi G r^2}\propto\dfrac{1}{r^2}$ . Why: this "singular isothermal sphere" is the simplest halo consistent with a flat curve, and integrating it shows the mass diverges linearly — the halo dominates the galaxy's total budget even though it never shines.

Go deeper · Derivation B · 1 minThermal freeze-out and the relic abundance

Derivation BThermal freeze-out and the relic abundance

The aim is to convert the Boltzmann equation (3) into the relic-density formula (4), showing that a species which was once in thermal equilibrium leaves behind a fossil abundance fixed by one number — its annihilation cross section — and that the abundance is inversely proportional to that cross section. We follow the number density from the hot early universe, through the moment expansion overtakes annihilation, to the frozen relic, and we pin the numerical coefficient against the measured density.

Step 1. At temperatures $T\gg m_\chi$ the species is relativistic and abundant, $n_{\rm eq}\sim T^3$ ; once $T$ drops below $m_\chi$ its equilibrium density is Boltzmann-suppressed, $n_{\rm eq}\propto(m_\chi T)^{3/2}e^{-m_\chi/T}$ . Why: making a heavy $\chi\bar\chi$ pair costs energy $\sim2m_\chi$ , so as the plasma cools, creation shuts off exponentially while annihilation into light particles continues.

Step 2. Freeze-out occurs when the annihilation rate per particle, $\Gamma=n\langle\sigma v\rangle$ , falls to the expansion rate $H$ . Below this point a $\chi$ can no longer find a partner to annihilate before the universe doubles in size, so its comoving number is frozen. Defining $x=m_\chi/T$ , this happens at $x_f\approx20\text{–}25$ , only mildly dependent on the mass. Why: the relic is set not by how much $\chi$ there was, but by when annihilation could no longer keep up with dilution.

Step 3. Solving (3) near $x_f$ gives the frozen yield $Y_\infty=n/s\propto x_f/(M_{\rm Pl}\,m_\chi\,\langle\sigma v\rangle)$ , and converting the number density today into a density fraction gives $\;\Omega_\chi h^2\approx\dfrac{1.07\times10^{9}\ \mathrm{GeV^{-1}}\,x_f}{\sqrt{g_*}\,M_{\rm Pl}\,\langle\sigma v\rangle}$ , with $M_{\rm Pl}=1.22\times10^{19}$ GeV and $g_*$ the effective degrees of freedom at freeze-out. Why: a larger cross section keeps $\chi$ in equilibrium longer, to lower density, so fewer survive — the relic scales as $1/\langle\sigma v\rangle$ , with no explicit dependence on $m_\chi$ except through $x_f$ and $g_*$ .

Step 4. Insert $x_f=20$ , $g_*\approx80$ : $\Omega_\chi h^2\approx1.96\times10^{-10}/\langle\sigma v\rangle[\mathrm{GeV^{-2}}]$ . Demanding $\Omega_\chi h^2=0.12$ gives $\langle\sigma v\rangle\approx1.6\times10^{-9}\ \mathrm{GeV^{-2}}\approx1.9\times10^{-26}\ \mathrm{cm^3\,s^{-1}}$ — the thermal relic cross section of equation (4). Why: the abstract solution lands on a concrete, weak-strength number, and reproducing the Planck value fixes the coefficient once and for all.

Go deeper · Derivation C · 1 minThe WIMP miracle — why the weak scale gives the right abundance

Derivation CThe WIMP miracle — why the weak scale gives the right abundance

Here the goal is to explain why the thermal relic cross section that reproduces the dark-matter density is not an arbitrary input but emerges automatically from a particle at the electroweak scale, an argument compelling enough to have driven three decades of experiment. We estimate the annihilation cross section of a weakly coupled particle by dimensional analysis, plug in the electroweak mass and coupling, convert to physical units, and compare with the thermal target — finding agreement to within the order-of-magnitude honesty such an estimate allows.

Step 1. A particle annihilating through a weak-strength interaction has two coupling vertices, so its cross section scales as $\sigma v\sim\alpha^2/m^2$ , with $\alpha$ the coupling and $m$ the mass setting the only dimensionful scale. Why: two vertices give a factor $\alpha^2$ in the amplitude squared, and dimensional analysis fixes the rest, since $[\sigma v]=[\text{mass}]^{-2}$ in natural units.

Step 2. Take electroweak values — mass $m\sim100$ GeV and coupling $\alpha_w\approx1/30$ : $\sigma v\sim(1/30)^2/(100\ \mathrm{GeV})^2\approx1.1\times10^{-7}\ \mathrm{GeV^{-2}}$ . Why: the electroweak scale is the natural home of any new physics tied to the Higgs, and it supplies both numbers at once.

Step 3. Convert with $1\ \mathrm{GeV^{-2}}=1.17\times10^{-17}\ \mathrm{cm^3\,s^{-1}}$ : the geometric estimate is $\sim1.3\times10^{-24}\ \mathrm{cm^3\,s^{-1}}$ , and folding in the generic phase-space and coupling factors of order $1/8\pi$ together with the freeze-out velocity $v\sim0.3\,c$ brings it down to a few $\times10^{-26}\ \mathrm{cm^3\,s^{-1}}$ . Why: the raw dimensional number overshoots by the expected $\mathcal{O}(10\text{–}50)$ , and once the standard prefactors are restored it lands squarely on the thermal relic value.

Step 4. The conclusion is the miracle: a stable, neutral particle with electroweak mass and coupling automatically freezes out with roughly the observed dark-matter density, no tuning required. The archetype is the neutralino — the lightest supersymmetric particle, kept stable by R-parity — which is why solving the Higgs hierarchy problem and explaining dark matter can be the same TeV-scale story. Why: two independent motivations, one from the Higgs and one from cosmology, point to the same energy scale, which is the strongest circumstantial case the WIMP paradigm has.

Limiting cases

Case	Result	Meaning
Deep inside the disk	$v(r)$ rises, $M(r)\approx M_{\rm lum}(r)$	luminous matter dominates; ordinary starlight tracks the mass
Far outside the disk	$v(r)\to$ const, $\rho\propto1/r^2$	the dark halo takes over; enclosed mass grows without light
Hot early universe, $T\gg m_\chi$	$n=n_{\rm eq}\sim T^3$	$\chi$ in full thermal equilibrium, created and destroyed freely
Freeze-out, $x_f=m_\chi/T\approx20$ – $25$	$\Gamma=n\langle\sigma v\rangle=H$	annihilation can no longer keep up with expansion; abundance frozen
Large $\langle\sigma v\rangle$	$\Omega_\chi h^2$ small	over-efficient annihilator; under-abundant, cannot be all the dark matter
Small $\langle\sigma v\rangle$	$\Omega_\chi h^2$ large	feeble annihilator; over-closes the universe unless very light or non-thermal
Relativistic (hot) relic	free-streaming erases small structure	ruled out — dark matter must be cold to grow galaxies first

Common trap“Dark matter is just ordinary matter — faint stars, gas, or black holes — that we haven't seen yet.”

It cannot be, and the reason is quantitative. Two entirely separate measurements fix the amount of baryonic matter — protons and neutrons — in the universe: the primordial abundances of deuterium and helium from Big Bang nucleosynthesis, and the relative heights of the acoustic peaks in the cosmic microwave background. Both give $\Omega_{\rm b}h^2\approx0.0224$ , roughly one-fifth of the total matter density $\Omega_{\rm m}h^2\approx0.14$ . So even if we hid every baryon in dim stars, cold gas, or stellar-mass black holes (collectively "MACHOs"), we would still be missing about 80% of the gravitating matter. Microlensing surveys toward the Magellanic Clouds independently rule out MACHOs as the bulk of the halo. The dark matter is not faint baryons — it is a substance the Standard Model does not contain, which is exactly why it points beyond it. (A separate trap: dark matter is not dark energy — one clumps and gravitates like matter, the other is smooth and drives acceleration.)

Worked examples

1 · The local dark matter density from the Milky Way rotation curve (fully worked)

The Sun orbits the Galactic centre at $r=8$ kpc with speed $v\approx220$ km/s, on a rotation curve that is flat out to this radius and beyond. Model the halo as a singular isothermal sphere and estimate the dark-matter density at the Sun's location. Compare with the value used in direct-detection experiments.

Step 1. For a flat curve the isothermal-sphere density is $\rho(r)=v^2/(4\pi G r^2)$ (Derivation A, Step 4). Convert to SI: $v=2.2\times10^5$ m/s so $v^2=4.84\times10^{10}\ \mathrm{m^2/s^2}$ , and $r=8\times3.086\times10^{19}\ \mathrm{m}=2.47\times10^{20}$ m so $r^2=6.10\times10^{40}\ \mathrm{m^2}$ .

Step 2. Assemble the denominator: $4\pi G r^2=4\pi(6.674\times10^{-11})(6.10\times10^{40})=5.12\times10^{31}$ . Then $\rho=4.84\times10^{10}/5.12\times10^{31}=\mathbf{9.5\times10^{-22}\ kg/m^3}$ .

Step 3. Convert to particle-physics units using $1\ \mathrm{GeV/cm^3}=1.783\times10^{-21}\ \mathrm{kg/m^3}$ : $\rho=9.5\times10^{-22}/1.783\times10^{-21}=\mathbf{0.53\ GeV/cm^3}$ .

Step 4. Interpret. The pure isothermal estimate slightly overshoots because it attributes the entire rotation speed at 8 kpc to the halo, whereas the baryonic disk and bulge also contribute there; subtracting the baryons brings the local dark-matter density down to the standard ρ_⊙ ≈ 0.3–0.4 GeV/cm³ used to set direct-detection rates. Either way the number is astonishingly small — about one WIMP of mass 100 GeV in every three to five litres of space near the Sun — which is exactly why detecting them takes tonnes of target and years of exposure.

2 · The relic abundance of a thermal WIMP

A hypothetical WIMP annihilates with thermally averaged cross section $\langle\sigma v\rangle=3\times10^{-26}\ \mathrm{cm^3\,s^{-1}}$ . What relic density does it leave? What happens if instead $\langle\sigma v\rangle$ is ten times larger, or ten times smaller?

Step 1. Apply equation (4): $\Omega_\chi h^2\approx0.12\times(3\times10^{-26})/\langle\sigma v\rangle$ . For $\langle\sigma v\rangle=3\times10^{-26}\ \mathrm{cm^3\,s^{-1}}$ the ratio is 1, so $\Omega_\chi h^2\approx\mathbf{0.12}$ — this is the observed dark-matter density. The WIMP is the whole of the dark matter.

Step 2. Ten times stronger annihilation, $\langle\sigma v\rangle=3\times10^{-25}\ \mathrm{cm^3\,s^{-1}}$ : $\Omega_\chi h^2\approx0.12/10=\mathbf{0.012}$ . The particle annihilated too efficiently and only a tenth of the required density survives — it could be a component of dark matter but not all of it.

Step 3. Ten times weaker, $\langle\sigma v\rangle=3\times10^{-27}\ \mathrm{cm^3\,s^{-1}}$ : $\Omega_\chi h^2\approx\mathbf{1.2}$ . This over-closes the universe — ten times too much matter, which would have recollapsed or drastically altered its expansion history. Such a particle is excluded as a thermal relic.

Step 4. The lesson is the anti-correlation baked into freeze-out: the relic abundance is inversely proportional to the annihilation cross section. The observed density selects one narrow band of cross sections — the thermal relic value — and that band coincides with weak-scale physics. That coincidence is the WIMP miracle, and it is what direct-detection experiments have spent two decades trying to confirm or exclude.

3 · Zwicky's Coma cluster and the birth of dark matter

The Coma cluster has a velocity dispersion of about $\sigma_v\approx1000$ km/s among its galaxies and a radius $R\approx1$ Mpc. Use the virial theorem to estimate its mass, and compare with the mass of its visible stars.

Step 1. The virial theorem for a self-gravitating system, $2\langle T\rangle+\langle U\rangle=0$ , gives a mass scale $M\sim\sigma_v^2 R/G$ . With $\sigma_v^2=10^{12}\ \mathrm{m^2/s^2}$ , $R=3.09\times10^{22}$ m, $G=6.67\times10^{-11}$ : $M\sim(10^{12})(3.09\times10^{22})/(6.67\times10^{-11})=4.6\times10^{44}$ kg.

Step 2. In solar masses ( $M_\odot=1.99\times10^{30}$ kg): $M\sim2.3\times10^{14}\,M_\odot$ , and restoring the order-unity virial coefficient ( $\sim5$ for a realistic profile) gives $M\sim\mathbf{10^{15}\,M_\odot}$ .

Step 3. The luminous stars of Coma total only a few $\times10^{12}\,M_\odot$ . The dynamical mass exceeds the stellar mass by a factor of several hundred — the mass-to-light ratio $M/L$ that so startled Zwicky in 1933 that he coined the term dunkle Materie, dark matter.

Step 4. Later work showed hot intracluster gas contributes more baryons than the stars, but even the full baryonic budget falls short of the virial mass by roughly a factor of five — the same universal ratio $\Omega_{\rm DM}/\Omega_{\rm b}\approx5.4$ that the microwave background would confirm seventy years later. Zwicky's cluster and Planck's sky measure the same missing matter.

Physics of the portrait

Dark matter was not discovered in a laboratory; it was inferred, twice, from the motions of things too big to hold in the hand — a cluster of galaxies and the outskirts of a spiral. The man who first put a number on it did so decades before anyone would take it seriously.

Origins · 1 minFritz Zwicky and how this was discovered

Fritz Zwicky

1898–1974 · Swiss-American astrophysicist at Caltech

Fritz Zwicky was the most prophetic and least diplomatic astronomer of his generation. In 1933, applying the virial theorem to the Coma cluster, he found that the galaxies were moving so fast that the cluster's visible stars could supply only a few hundredths of the gravity needed to keep it bound — the rest had to be dunkle Materie, dark matter. The claim sat almost ignored for forty years, both because Zwicky's abrasive personality cost him allies and because the case rested on a single cluster with an uncertain distance scale. It was Vera Rubin's flat rotation curves of spiral galaxies in the 1970s that made the missing mass impossible to dismiss, turning Zwicky's outlier into a universal fact. His other predictions were equally ahead of their time: with Walter Baade he proposed in 1934 that supernovae are the collapse of ordinary stars into neutron stars — coining both "supernova" and "neutron star" in the same breath — and he predicted that galaxy clusters would act as gravitational lenses, the very effect that, in the Bullet Cluster, would one day give dark matter its most direct proof. Dark matter, neutron stars, and cluster lensing — three pillars of modern astrophysics — trace to one contrarian who was rarely believed in his lifetime.

History

1933

— Zwicky infers unseen mass in the Coma cluster from galaxy velocities and names it dunkle Materie

1970

— Rubin and Ford measure the flat rotation curve of the Andromeda galaxy

1980

— Rubin, Ford and Thonnard establish flat rotation curves across dozens of spirals

2006

— the Bullet Cluster shows the lensing mass offset from the X-ray gas, favouring collisionless dark matter

2018

— Planck fixes the cold dark matter density at Ω_c h² = 0.120

Going further

Texts. Peskin & Schroeder, An Introduction to Quantum Field Theory, Ch. 20 (beyond the Standard Model context) · Peskin, "Beyond the Standard Model," SLAC-PUB-7479 / hep-ph/9705479 · Griffiths, Introduction to Elementary Particles, 2nd ed., Ch. 12 · Kolb & Turner, The Early Universe, Ch. 5 (freeze-out and relic abundances) · Dodelson, Modern Cosmology · Bertone, Hooper & Silk, "Particle Dark Matter: Evidence, Candidates and Constraints," Phys. Rep. 405 (2005); Bertone & Hooper, Rev. Mod. Phys. 90 (2018) 045002 (history).

Provenance. The Standard Model Lagrangian and the "no candidate" framing follow Peskin's Beyond the Standard Model lectures (§2.1); the thermal relic machinery (Boltzmann equation, freeze-out, $\Omega h^2\propto1/\langle\sigma v\rangle$ ) is standard Kolb–Turner cosmology. A verifier should confirm: (1) $\rho=v^2/(4\pi Gr^2)\Rightarrow0.53\ \mathrm{GeV/cm^3}$ at $v=220$ km/s, $r=8$ kpc; (2) $\Omega_{\rm DM}=0.120/h^2\approx0.265$ and $\Omega_{\rm DM}/\Omega_{\rm b}\approx5.4$ at $h=0.674$ ; (3) the detailed relic formula $\Omega_\chi h^2\approx1.07\times10^9\,x_f\,\mathrm{GeV^{-1}}/(\sqrt{g_*}M_{\rm Pl}\langle\sigma v\rangle)$ gives $\langle\sigma v\rangle\approx1.9\times10^{-26}\ \mathrm{cm^3\,s^{-1}}$ for $\Omega_\chi h^2=0.12$ ; (4) the WIMP estimate $\alpha_w^2/m^2$ at $m=100$ GeV lands within an order of magnitude of the thermal value. Planck 2018 densities and the local halo density $\rho_\odot\approx0.3$ – $0.4\ \mathrm{GeV/cm^3}$ are standard tabulated values.

Prerequisites. Standard Model Summary · Electroweak Gauge Group · Higgs Mechanism. Leads to. Supersymmetry · Open Questions · Cosmological Connections.

Ask the tutor: Ask me to walk through the freeze-out calculation line by line, or to show exactly why the Bullet Cluster defeats modified gravity — pick a thread and we will pull it.

Could dark matter just be ordinary matter that happens to be too faint to see?

No — and this is settled quantitatively, not by lack of imagination. The total amount of baryonic matter (protons and neutrons) is fixed independently by two clocks: the primordial deuterium and helium abundances from Big Bang nucleosynthesis, and the heights of the acoustic peaks in the microwave background. Both give $\Omega_{\rm b}h^2\approx0.0224$ , only about one-fifth of the total matter density. So even hiding every baryon in dim stars, cold gas clouds, or stellar-mass black holes leaves roughly 80% of the gravitating matter unexplained. Microlensing surveys have separately searched for such compact baryonic objects in the halo and found far too few. Dark matter must be non-baryonic — a substance outside the Standard Model's roster.

Is dark matter the same thing as dark energy?

They share the word "dark" and nothing else. Dark matter is matter: it clumps under gravity, forms halos around galaxies, and its density thins out as the universe expands. Dark energy is a smooth, nearly constant energy density filling all of space that drives the accelerating expansion; it does not clump and it does not gravitate attractively. In the budget, dark energy is about 68.5% of the total and dark matter about 26.5%. A galaxy's rotation curve is a dark-matter effect; the acceleration of distant supernovae is a dark-energy effect. Confusing the two is one of the most common errors in the subject.

Why not just modify gravity (MOND) instead of adding an invisible particle?

Modified Newtonian dynamics can fit galactic rotation curves impressively with a single parameter, and it deserves respect for that. But dark matter is not one phenomenon — it is a consistent thread through galactic rotation, cluster dynamics, gravitational lensing, the growth of cosmic structure, and the relative heights of the microwave-background acoustic peaks, which specifically require a non-baryonic, gravitating-but-not-pressure-supported component. The Bullet Cluster is the sharpest test: there, the gravitational-lensing mass is physically offset from the baryonic gas, which a theory that ties gravity to visible matter struggles to reproduce. No modified-gravity theory yet matches the full set of observations that particle dark matter fits with one added ingredient.

What exactly is a WIMP, and why was it the favourite candidate?

A WIMP is a Weakly Interacting Massive Particle: a hypothetical particle with a mass around the electroweak scale (tens of GeV to a few TeV) and interactions of roughly weak strength — neutral, colourless, and stable. It became the leading candidate because of the WIMP miracle: if such a particle was in thermal equilibrium in the hot early universe and then froze out, its relic abundance comes out automatically near the observed dark-matter density, with no tuning. The same weak-scale physics motivated by the Higgs hierarchy problem — supersymmetry in particular, whose lightest neutralino is the classic WIMP — would supply exactly such a particle. Two independent puzzles pointing to the same energy scale is a powerful hint, which is why WIMPs drove the experimental programme for thirty years.

If WIMPs keep not showing up, is the whole idea dead?

Not dead, but humbled, and the field has broadened. Direct-detection experiments have excluded large swathes of the simplest WIMP models without a confirmed signal, which has cooled the assumption that dark matter must live at the weak scale. Attention has spread to axions — extremely light particles that also elegantly solve the strong-CP problem of QCD, hunted with resonant microwave cavities; to keV-scale sterile neutrinos as warm dark matter; to primordial black holes formed in the early universe; and to "feebly interacting" particles produced by freeze-in rather than freeze-out. The core evidence for dark matter is untouched by any of this — what is under revision is only our guess about its identity. The WIMP remains viable in its unexcluded corners, but it is now one hypothesis among several.

Does the Standard Model really contain nothing that could be dark matter — what about neutrinos?

Neutrinos are the Standard Model's only neutral, stable, weakly interacting particles, so they are the natural first guess — and they are a form of dark matter, just a hopelessly insufficient one. Two problems. First, they are far too light: with masses below about 0.1 eV, the cosmic neutrino background contributes at most a per cent or so of the dark-matter density. Second, they are "hot" — they were relativistic when structure began to form, so they free-stream out of small overdensities and erase them, whereas the universe demonstrably built small galaxies first and merged them into large ones. That requires cold (non-relativistic) dark matter. Every other Standard Model particle is charged, coloured, or unstable. Hence the whole point: dark matter is beyond the Standard Model.

How do we know dark matter is “cold” rather than fast-moving?

From the pattern of cosmic structure. If dark matter were "hot" — moving near light speed in the early universe, as light neutrinos would — it would stream freely out of small clumps before they could collapse, wiping out galaxy-scale structure and leaving only the largest superclusters to form first (top-down). What we observe is the opposite: small galaxies formed early and merged into larger structures (bottom-up), and the statistics of galaxy clustering and the microwave-background fluctuations match this beautifully only if the dark matter was slow — "cold" — when structure began. This is why the standard cosmological model is called ΛCDM, for a cosmological constant Λ plus Cold Dark Matter, and it is one more constraint any candidate particle must satisfy.