Cold Dark Matter

Cold dark matter is the working hypothesis that most of the matter in the universe is made of particles with three defining properties at once. They are dark — no electric or colour charge, so they neither emit, absorb, nor scatter light, and to an excellent approximation collide with nothing, including themselves. They are cold — non-relativistic by the epoch when cosmological perturbations enter the horizon, so their thermal velocities are tiny and their free-streaming length is astrophysically negligible. And they are non-baryonic — not protons and neutrons, so they are invisible to Big Bang nucleosynthesis and counted separately from the baryons the CMB weighs. Of the three, cold is load-bearing: it is what lets dark matter cluster gravitationally on every scale from dwarf galaxies to superclusters. The evidence is old, gravitational, and stubbornly consistent. Zwicky's 1933 virial estimate for the Coma cluster demanded far more mass than its galaxies shine; flat rotation curves embed spirals in halos whose mass keeps growing where the light has run out; lensing traces mass, not luminosity; and in the Bullet Cluster the lensing mass sails straight through while the X-ray gas is shocked and left behind — collisionless matter caught in the act. On the largest scales, the CMB acoustic peaks and the matter power spectrum both demand a pressureless, non-baryonic component with $\Omega_{\mathrm{c}}h^2 \approx 0.12$ , about five times the baryon density. The deepest reason cosmologists cannot do without it is timing. Having no pressure and never tied to the photons, cold dark matter begins growing at matter–radiation equality and digs the potential wells into which baryons — released only at recombination — later fall. Without that head start, the tiny $\delta \sim 10^{-5}$ imprinted on the CMB could never have grown into galaxies by today. Cold dark matter is the C in $\Lambda$ CDM, and this article is the synthesis the density-perturbation, Jeans, and linear-growth articles built toward.

Key takeaways

Cold dark matter (CDM) is non-baryonic matter that is non-relativistic ("cold") when perturbations enter the horizon and interacts only gravitationally — effectively collisionless and pressureless.
Being cold, it has negligible free-streaming and thermal dispersion, so it clusters on all scales down to sub-galactic masses, unlike hot dark matter which erases small-scale structure.
CDM dominates the matter budget: $\Omega_{\mathrm{c}}h^2 \approx 0.12$ against $\Omega_{\mathrm{b}}h^2 \approx 0.022$ , outweighing baryons roughly five to one.
Having no pressure, CDM perturbations grow as $\delta_{\mathrm{c}} \propto a$ from matter–radiation equality, seeding the wells into which baryons fall after recombination — this is why structure formation works.
Its fingerprints span independent probes: rotation curves, cluster masses, lensing, the CMB acoustic peaks, and the $P(k)$ turnover at the equality scale $k_{\mathrm{eq}}$ .
The evidence is purely gravitational; the particle's identity (WIMP, axion, sterile neutrino, primordial black hole, …) is still unknown.
Key idea: cold dark matter is pressureless and collisionless, so its perturbations are not damped by pressure or free-streaming and begin growing at matter–radiation equality — seeding the potential wells into which baryons fall to form all observed structure.

Where this lives today

Cold dark matter is at once the most successful and the most frustrating idea in cosmology. On the largest scales it is triumphant: N-body simulations of collisionless CDM — Millennium, IllustrisTNG, and successors evolving billions of particles under the Vlasov–Poisson system — reproduce the observed cosmic web of filaments, walls, voids, and halos, and surveys such as DESI and Euclid now measure the power spectrum and its $k_{\mathrm{eq}}$ turnover precisely enough to weigh the CDM to a few percent. The open questions live on small scales and in the particle physics. Pure-CDM simulations predict cuspy halo centres ( $\rho \propto r^{-1}$ ) and more dwarf satellites than are seen — the "cusp–core", "missing-satellites", and "too-big-to-fail" tensions — though whether these signal new physics (self-interactions, warmth) or just baryonic feedback is unsettled. The particle's identity is wide open: LZ, XENONnT, and PandaX hunt for WIMP recoils, ADMX tunes cavities for axions, and astrophysical searches probe sterile neutrinos and primordial black holes. Every laboratory bound so far is a null result, and after decades the dark matter remains known only by its gravity.

The physical picture

Picture two crowds bunching up under their own gravity. Photons and baryons before recombination are one hot, tightly coupled fluid: squeeze it and radiation pressure pushes back, so the compression rebounds and travels off as a sound wave — no runaway collapse below the sound horizon. Cold dark matter is the other crowd, a swarm of slow, non-interacting particles: squeeze it and nothing pushes back, so self-gravity wins and the clump deepens. That single difference — pressure versus none — is why one component oscillates and the other collapses.

Figure

Two crowds under gravity: cold dark matter has no pressure, so it falls into the well and its contrast grows as δ ∝ a, while the photon–baryon fluid pushes back and oscillates as a sound wave — and dark matter outweighs baryons about five to one, Ω_c h² ≈ 0.12 versus Ω_b h² ≈ 0.022

"Cold" is a statement about a particle's memory of its own velocity. Hot particles — light neutrinos are the textbook case — move near the speed of light and stream freely out of small perturbations before they can collapse, wiping out structure below their free-streaming scale and forcing a top-down history. Cold particles barely drift relative to the Hubble flow, so even the smallest perturbations survive and structure grows bottom-up: small halos form first and merge into larger ones. Between these limits sits warm dark matter, a keV-scale particle that free-streams just enough to smooth the smallest scales. Timing is the last piece, and the quantitative one. While radiation dominates, even cold matter cannot grow: the expansion is so fast that Hubble drag overwhelms the feeble self-gravity of a subdominant component, and sub-horizon CDM perturbations are nearly frozen — the Mészáros effect, at most a factor $5/2$ through the whole radiation era. The instant radiation redshifts away at matter–radiation equality, the brakes release and $\delta_{\mathrm{c}}$ grows linearly with the scale factor. The second figure traces this whole biography of a perturbation.

Figure

Growth of the density contrast versus scale factor a on log–log axes: cold dark matter is nearly frozen before matter–radiation equality at z ≈ 3400 by the Mészáros effect, then grows as δ_c ∝ a; baryons stay locked to the photons and oscillate until recombination at z ≈ 1100, then fall into the pre-built wells and catch up as δ_b → δ_c times the factor 1 − a_rec/a

The reason this is called "concordance" and not "a fudge" is that a single number carries it. The same $\Omega_{\mathrm{c}}h^2 \approx 0.12$ that sets the height of the third CMB acoustic peak also sets the redshift of equality, the turnover scale of the matter power spectrum, the masses of dark halos, and the flat rotation curves of galaxies. One parameter, fixed by microwave-background photons, predicts the dynamics of objects across ten orders of magnitude in scale — that web of agreement is the real case for cold dark matter.

The core equations

The starting point is the census. Independent measurements — CMB acoustic peaks, BBN light-element abundances, cluster baryon fractions — fix how much of each kind of matter the universe holds, and they disagree by a factor of five, which is the whole problem.

\Omega_{\mathrm{c}} h^2 \approx 0.12, \qquad \Omega_{\mathrm{b}} h^2 \approx 0.022, \qquad \Omega_{\mathrm{m}} \approx 0.31 \qquad (1)

where: $\Omega_{\mathrm{c}}$ — cold dark matter density parameter (fraction of critical density); $\Omega_{\mathrm{b}}$ — baryon density parameter; $\Omega_{\mathrm{m}}$ — total matter, $\Omega_{\mathrm{c}}+\Omega_{\mathrm{b}}$ plus a small neutrino piece; $h$ — Hubble constant in units of $100\ \mathrm{km\,s^{-1}\,Mpc^{-1}}$ . Reading it: the dark component is $0.12/0.022 \approx 5.5$ times the baryons, so most matter is not atoms. Whatever the particle is, its gravitational bookkeeping is that of pressureless dust: a fixed number of particles thinned only by the expanding volume. This is what "matter" means in the Friedmann equations.

\bar\rho_{\mathrm{m}}(a) = \frac{3H_0^2}{8\pi G}\,\Omega_{\mathrm{m}}\,a^{-3} \qquad (2)

where: $\bar\rho_{\mathrm{m}}$ — mean matter density, dominated by CDM; $a$ — scale factor, $a_0=1$ today; $H_0$ — present Hubble constant; $G$ — Newton's constant. Reading it: the $a^{-3}$ scaling is the signature of pressureless matter, versus radiation's $a^{-4}$ — the extra power of $a$ being the redshift radiation suffers and matter does not. Now the dynamics. Perturbing this pressureless fluid in an expanding background and combining mass conservation, momentum balance, and gravity, the density contrast obeys a damped, self-gravitating oscillator equation — except that with no pressure there is no restoring force, only drive and drag.

\ddot{\delta}_{\mathrm{c}} + 2H\dot{\delta}_{\mathrm{c}} = 4\pi G\,\bar\rho_{\mathrm{m}}\,\delta_{\mathrm{c}} \qquad (3)

where: $\delta_{\mathrm{c}}$ — CDM density contrast $(\rho_{\mathrm{c}}-\bar\rho_{\mathrm{c}})/\bar\rho_{\mathrm{c}}$ ; $H=\dot a/a$ — Hubble rate; $2H\dot\delta_{\mathrm{c}}$ — Hubble friction opposing collapse; $4\pi G\bar\rho_{\mathrm{m}}\delta_{\mathrm{c}}$ — self-gravity driving it. Reading it: the missing term is the $c_{\mathrm{s}}^2 k^2 \delta / a^2$ pressure term of a normal fluid — deleting it is exactly what "cold" buys, and why CDM grows where the photon–baryon fluid only rings. Solving that equation on a matter-dominated background gives the growing mode, and comparing it with the radiation era exposes the Mészáros stagnation that shapes the power spectrum.

\delta_{\mathrm{c}} \propto a \quad (\text{matter domination}), \qquad \delta_{\mathrm{c}} \propto \ln a \quad (\text{radiation domination, sub-horizon}) \qquad (4)

where: $a$ — scale factor; $\propto \ln a$ — the Mészáros effect, only logarithmic growth of sub-horizon modes while radiation dominates. Reading it: modes that entered the horizon before equality were held back, so small scales carry less power than the primordial spectrum laid down — a suppression that turns into a measurable bend. That bend sits at the comoving scale that entered the horizon exactly at equality, and it is the most direct cosmological fingerprint of the CDM density.

L_{\mathrm{eq}} \simeq c\!\int_0^{t_{\mathrm{eq}}}\!\frac{\mathrm{d}t}{a} \;\Rightarrow\; k_{\mathrm{eq}}^{-1}\ \text{sets the turnover of}\ P(k) \qquad (5)

where: $t_{\mathrm{eq}}$ — time of matter–radiation equality; $k_{\mathrm{eq}}$ — comoving wavenumber entering the horizon at equality; $P(k)$ — matter power spectrum, peaking near $k_{\mathrm{eq}}$ . Reading it: measuring where $P(k)$ turns over measures $z_{\mathrm{eq}}$ , and hence $\Omega_{\mathrm{m}}$ . The equality redshift itself is fixed by the ratio of matter to radiation densities, since each dilutes at its own rate.

1+z_{\mathrm{eq}} = \frac{\Omega_{\mathrm{m}}}{\Omega_{\mathrm{r}}} \approx 3400 \qquad (6)

where: $z_{\mathrm{eq}}$ — redshift of equality; $\Omega_{\mathrm{r}}$ — present radiation density parameter (photons plus relativistic neutrinos); $\Omega_{\mathrm{m}}$ — present matter density parameter. Reading it: raise the CDM density and equality happens earlier, giving perturbations more time to grow. On galactic scales the same matter announces itself kinematically: a star on a circular orbit reports the mass interior to it, and that mass keeps rising where the light does not.

v_{\mathrm{circ}}^2(r) = \frac{G M(<r)}{r} = \text{const} \;\Rightarrow\; \rho(r)\propto r^{-2},\quad M(<r)\propto r \qquad (7)

where: $v_{\mathrm{circ}}$ — circular orbital speed; $M(<r)$ — mass enclosed within radius $r$ ; $\rho(r)$ — halo density profile implied by a flat curve. Reading it: visible matter alone predicts a Keplerian falloff $v\propto r^{-1/2}$ , so a flat curve demands an extended halo with $\rho \propto r^{-2}$ . Finally, because CDM is collisionless, its full description is not a fluid but a phase-space flow: the distribution function is advected along orbits in the self-consistent potential.

\frac{\partial f}{\partial t} + \mathbf{v}\!\cdot\!\nabla_{\mathbf x} f - \nabla_{\mathbf x}\phi\!\cdot\!\nabla_{\mathbf v} f = 0 \qquad (8)

where: $f(\mathbf x,\mathbf v,t)$ — phase-space distribution function; $\phi$ — gravitational potential, sourced self-consistently by $f$ through Poisson's equation. Reading it: with no collision term, phase-space density is conserved along trajectories — the content of "collisionless", and the equation every cosmological N-body simulation solves.

Derivation AWhy cold matter grows as δ ∝ a in matter domination

The goal is to solve the pressureless growth equation (3) on a flat, matter-dominated Einstein–de Sitter background and show that the growing solution is exactly $\delta_{\mathrm{c}} \propto a$ — the linear-in-scale-factor growth that is the dynamical signature of cold dark matter. The calculation also yields a decaying mode, so a physical initial condition selects the growing branch and fixes the whole later history. Everything downstream — the transfer function, the linear power spectrum, the amplitude $\sigma_8$ — inherits this one power law.

Step 1. Fix the background. In a flat matter-dominated universe $a \propto t^{2/3}$ , so $H = 2/(3t)$ ; and since $\bar\rho_{\mathrm{m}} = 3H^2/8\pi G$ , the source coefficient is $4\pi G\bar\rho_{\mathrm{m}} = \tfrac{3}{2}H^2 = 2/(3t^2)$ . Why: in terms of $t$ the growth equation becomes a pure power-law ODE.

Step 2. Substitute $2H = 4/(3t)$ and $4\pi G\bar\rho_{\mathrm{m}} = 2/(3t^2)$ into equation (3): $\ddot\delta + \tfrac{4}{3t}\dot\delta - \tfrac{2}{3t^2}\delta = 0$ . Why: this Cauchy–Euler form is scale-free, so it admits power-law solutions $\delta \propto t^n$ .

Step 3. Insert $\delta \propto t^n$ . Every term scales as $t^{n-2}$ , and cancelling it leaves the indicial equation $n(n-1) + \tfrac{4}{3}n - \tfrac{2}{3} = 0$ , i.e. $n^2 + \tfrac{1}{3}n - \tfrac{2}{3} = 0$ . Why: a purely algebraic condition on the exponent.

Step 4. Solve: $n = \tfrac{1}{2}(-\tfrac{1}{3} \pm \tfrac{5}{3})$ , giving $n = \tfrac{2}{3}$ or $n = -1$ . Why: two roots means two independent modes, one growing and one decaying, as a second-order system must have.

\delta_{\mathrm{c}}(t) = A\,t^{2/3} + B\,t^{-1} \;\Longrightarrow\; \delta_{\mathrm{c}} \propto a \quad (\text{growing mode})

Step 5. Interpret. Since $a \propto t^{2/3}$ , the growing mode is $\delta \propto a$ while the $t^{-1}$ mode decays away. With no pressure, self-gravity always beats Hubble friction on the growing branch, so the contrast climbs steadily and linearly — never oscillating. Why: this monotone growth, seeded at equality, is what lets $\delta \sim 10^{-5}$ reach order unity and collapse within a Hubble time.

Go deeper · Derivation B · 1 minThe Mészáros effect and why growth stalls in the radiation era

Derivation BThe Mészáros effect and why growth stalls in the radiation era

The goal is to show quantitatively that before matter–radiation equality, sub-horizon CDM perturbations barely grow — held nearly frozen by the fast, radiation-driven expansion — and to derive the exact growing mode in the natural variable $y = a/a_{\mathrm{eq}}$ . This "Mészáros stagnation" suppresses small-scale power relative to the primordial spectrum and carves the turnover of $P(k)$ at $k_{\mathrm{eq}}$ . The clean result: through the entire radiation era a mode grows by at most a factor of $5/2$ , so structure must wait for matter domination to assemble.

Step 1. Change the clock from $t$ to $y = a/a_{\mathrm{eq}}$ , using the Friedmann equation for a matter-plus-radiation mix. During this epoch the CDM is gravitationally subdominant, so its self-gravity is a weak source against the fast expansion. Why: $y$ is the only dimensionless variable, so the two eras separate cleanly at $y = 1$ .

Step 2. The growth equation becomes the Mészáros equation, $\dfrac{\mathrm{d}^2\delta_{\mathrm{c}}}{\mathrm{d}y^2} + \dfrac{2+3y}{2y(1+y)}\dfrac{\mathrm{d}\delta_{\mathrm{c}}}{\mathrm{d}y} - \dfrac{3}{2y(1+y)}\,\delta_{\mathrm{c}} = 0$ . Why: its coefficients carry the full expansion history, spanning both sides of equality exactly.

Step 3. Verify by substitution that the growing solution is $\delta_{\mathrm{c}} = 1 + \tfrac{3}{2}y$ : with $\delta' = 3/2$ and $\delta'' = 0$ , the friction and source terms both reduce to $\tfrac{3(2+3y)}{4y(1+y)}$ and cancel. Why: an exact closed form makes the two limits transparent.

\delta_{\mathrm{c}}(y) = 1 + \tfrac{3}{2}\,y, \qquad y \equiv \frac{a}{a_{\mathrm{eq}}}

Step 4. Read the limits. For $y \ll 1$ (radiation era) $\delta_{\mathrm{c}} \to 1$ , frozen; across the whole era $y: 0 \to 1$ so $\delta_{\mathrm{c}}$ grows only from $1$ to $5/2$ . For $y \gg 1$ it recovers $\delta_{\mathrm{c}} \to \tfrac{3}{2}y \propto a$ . Why: the earlier a mode entered the horizon, the more growth it lost — suppressing small-scale power.

Step 5. Connect to $P(k)$ . Modes crossing before equality (large $k$ ) are held back while super-horizon modes are untouched, so the processed spectrum bends at $k_{\mathrm{eq}}$ : it keeps the primordial $k^{n_{\mathrm s}}$ slope on large scales and steepens toward $k^{n_{\mathrm s}-4}$ on small scales. Why: the turnover is a clean ruler for $z_{\mathrm{eq}}$ , hence for $\Omega_{\mathrm{m}}h^2$ .

Go deeper · Derivation C · 1 minThe dark matter head start over the baryons

Derivation CThe dark matter head start over the baryons

The goal is to make quantitative the claim that cold dark matter builds its potential wells before the baryons are free to fall, and that this head start rescues structure formation given the tiny $\delta \sim 10^{-5}$ the CMB reveals. Before recombination the baryons are locked to the photons and cannot cluster on sub-horizon scales, while the CDM has been growing as $\delta_{\mathrm{c}} \propto a$ since equality. After recombination the baryons are released into pre-dug wells and rush to catch up, giving a simple catch-up law.

Step 1. After recombination both species feel gravity with negligible pressure: $\ddot\delta_{\mathrm b} + 2H\dot\delta_{\mathrm b} = 4\pi G(\bar\rho_{\mathrm c}\delta_{\mathrm c} + \bar\rho_{\mathrm b}\delta_{\mathrm b})$ . Since $\bar\rho_{\mathrm c} \gg \bar\rho_{\mathrm b}$ , the CDM term dominates the source as an external, slowly varying drive. Why: this reduces a coupled problem to a driven equation for $\delta_{\mathrm b}$ .

Step 2. Impose the initial condition: the pressure-supported baryons carry almost no contrast at release, $\delta_{\mathrm b}(a_{\mathrm{rec}}) \approx 0$ , while $\delta_{\mathrm c}$ is already substantial. Why: this asymmetry is the entire content of the head start.

Step 3. Solve on the matter-dominated background with $\delta_{\mathrm c} \propto a$ as driver: the particular solution $\delta_{\mathrm c}(a)$ plus homogeneous modes, with the constants fixed by $\delta_{\mathrm b} \approx 0$ at $a_{\mathrm{rec}}$ . Why: matching release ties the baryon history to recombination.

\delta_{\mathrm b}(a) \;\longrightarrow\; \delta_{\mathrm c}(a)\left[\,1 - \frac{a_{\mathrm{rec}}}{a}\,\right]

Step 4. Read the consequence. The bracket rushes to unity: by $a = 10\,a_{\mathrm{rec}}$ the baryons share $90\%$ of the CDM contrast and thereafter move as one. Without the pre-existing wells, baryons growing from $\delta \sim 10^{-5}$ by $(1+z_{\mathrm{rec}}) \approx 1100$ reach only $\delta \sim 10^{-2}$ — far short of collapse. Why: this is why a baryon-only universe cannot make galaxies, and why CDM is required. *The baryon-only estimate $10^{-5}\times 1100 \approx 10^{-2}$ confirms that baryons alone fall far short of the collapse threshold, underscoring why CDM is required.*

Limiting cases

Cold dark matter is defined by three idealizations, and each becomes exact or breaks down in a recognizable limit; the fourth row marks where "cold" shades into "warm".

Case	Result	Meaning
Coldness (pressureless)	$w_{\mathrm{c}} = \dfrac{p_{\mathrm{c}}}{\rho_{\mathrm{c}}c^2} \to 0,\quad c_{\mathrm{s}}^2 \to 0$	Equation of state $w \to 0$ and sound speed $\to 0$ : negligible Jeans length, so perturbations grow on all scales — the opposite of hot dark matter, whose relativistic speeds give a huge free-streaming length.
Collisionless (dark)	$\sigma_{\mathrm{SIDM}}/m \lesssim 1\ \mathrm{cm^2\,g^{-1}}$	The self-interaction cross-section per unit mass is bounded small by cluster observations (the Bullet Cluster, halo shapes), so CDM behaves as a collisionless fluid obeying the Vlasov equation.
Non-baryonic	$\Omega_{\mathrm{c}}h^2 \approx 0.12 \gg \Omega_{\mathrm{b}}h^2 \approx 0.022$	BBN and the CMB fix the baryon density independently, and the dark density is about five times larger — the gap cannot be atoms in any form.
Warm dark matter boundary	$\lambda_{\mathrm{fs}} \propto \dfrac{\langle v\rangle}{m},\quad m_{\mathrm{WDM}} \sim \mathrm{keV}$	A keV-mass particle free-streams just enough to smooth sub-galactic power — the warm limit that interpolates between cold ( $m \gg \mathrm{keV}$ ) and hot ( $m \sim \mathrm{eV}$ ) dark matter.

Common trapThe word "cold" does not mean "slow today"

The commonest misconception is to read "cold" as "moving slowly today". It means non-relativistic at the epoch the relevant perturbations entered the horizon, so the comoving free-streaming length was already tiny then — that early history, not the present speed, decides whether small-scale structure survives. A second trap is to imagine the dark matter is ordinary matter that simply does not shine: cold gas, faint stars, black holes, or other MACHOs. It cannot be. BBN and the CMB fix the total baryon density at $\Omega_{\mathrm{b}}h^2 \approx 0.022$ from physics indifferent to whether the baryons are luminous, and that ceiling is about five times too low to be the dark matter. The missing mass is a new, non-baryonic species — full stop.

Worked examples

1 · When did matter overtake radiation? (fully worked)

Given the concordance densities $\Omega_{\mathrm{m}}h^2 = 0.143$ and present radiation density $\Omega_{\mathrm{r}}h^2 = 4.15\times10^{-5}$ (photons plus three neutrino species), find the redshift of matter–radiation equality and say why the CDM density controls it.

Step 1. Matter dilutes as $a^{-3}$ and radiation as $a^{-4}$ , so their ratio scales as $a$ . Setting $\bar\rho_{\mathrm{m}} = \bar\rho_{\mathrm{r}}$ gives $1+z_{\mathrm{eq}} = \Omega_{\mathrm{m}}/\Omega_{\mathrm{r}} = \Omega_{\mathrm{m}}h^2 / \Omega_{\mathrm{r}}h^2$ , the physical densities $\Omega h^2$ cancelling the poorly known $h$ .

Step 2. Put in the numbers: $1+z_{\mathrm{eq}} = 0.143 / (4.15\times10^{-5}) = 3446$ , so $z_{\mathrm{eq}} \approx 3400$ — matter domination began roughly $50{,}000$ years after the Big Bang.

Step 3. Since $\Omega_{\mathrm{m}}h^2 = \Omega_{\mathrm{c}}h^2 + \Omega_{\mathrm{b}}h^2 = 0.12 + 0.022$ , the CDM supplies $84\%$ of the matter and essentially sets $z_{\mathrm{eq}}$ . With baryons alone, $1+z_{\mathrm{eq}}$ would fall to $\approx 530$ , near recombination, leaving almost no matter-dominated era for growth. The CDM density advances equality by about a factor of six, buying the growth time that makes galaxies possible.

2 · The dark halo of the Milky Way from a flat rotation curve (fully worked)

The Milky Way's rotation curve is flat at $v_{\mathrm{circ}} \approx 220\ \mathrm{km\,s^{-1}}$ out to at least $r = 50\ \mathrm{kpc}$ . Find the mass enclosed and compare it with the luminous mass of about $6\times10^{10}\,M_\odot$ .

Step 1. For a circular orbit, gravity supplies the centripetal force: $v_{\mathrm{circ}}^2 = GM(<r)/r$ , so $M(<r) = v_{\mathrm{circ}}^2\,r / G$ . A flat curve means this enclosed mass grows linearly with radius — the hallmark of an extended halo.

Step 2. Convert to SI: $v = 2.20\times10^5\ \mathrm{m\,s^{-1}}$ , so $v^2 = 4.84\times10^{10}\ \mathrm{m^2\,s^{-2}}$ ; and $r = 50\ \mathrm{kpc} = 50\times3.086\times10^{19}\ \mathrm{m} = 1.543\times10^{21}\ \mathrm{m}$ .

Step 3. Evaluate: $M = (4.84\times10^{10})(1.543\times10^{21}) / (6.674\times10^{-11}) = 1.12\times10^{42}\ \mathrm{kg}$ . In solar masses ( $M_\odot = 1.989\times10^{30}\ \mathrm{kg}$ ), $M(<50\ \mathrm{kpc}) \approx 5.6\times10^{11}\,M_\odot$ .

Step 4. The enclosed dynamical mass is about $\mathbf{9}$ times the luminous mass of $\sim6\times10^{10}\,M_\odot$ , and it keeps rising with radius while the light has faded — the discrepancy a dark halo with $\rho \propto r^{-2}$ (refined to the NFW profile) resolves. Coldness is what lets such a halo form on a galactic scale; hot dark matter would have free-streamed it away.

3 · Could baryons alone have made galaxies? (order-of-magnitude)

The CMB shows $\delta \sim 10^{-5}$ at recombination, $z_{\mathrm{rec}} \approx 1100$ . In a baryon-only universe, estimate the contrast today and compare with the collapse threshold.

Step 1. In matter domination the linear contrast grows as $\delta \propto a = 1/(1+z)$ , so from recombination to now it grows by $(1+z_{\mathrm{rec}}) \approx 1100$ .

Step 2. A baryon perturbation starting at $\delta \sim 10^{-5}$ reaches only $\delta \sim 10^{-5} \times 1100 \approx \mathbf{10^{-2}}$ today — a hundred times below the $\delta \sim 1$ needed to collapse. A baryon-only universe would be nearly featureless: no galaxies, no us.

Step 3. Cold dark matter breaks the impasse: never coupled to the photons, its contrast was not held at $10^{-5}$ but had been growing since equality ( $z \approx 3400$ ), already building wells. The baryons, released at recombination, fell in and caught up as $\delta_{\mathrm b} \to \delta_{\mathrm c}(1 - a_{\mathrm{rec}}/a)$ , reaching order unity on schedule. The head start is not a detail — it is the reason structure exists.

Physics of the portrait

The idea of unseen mass has one unmistakable origin: a combative Caltech astronomer who, weighing a cluster of galaxies in 1933, found the arithmetic impossible unless most of the mass was invisible — and was ignored for forty years.

Origins · 1 minFritz Zwicky and how this was discovered

Fritz Zwicky

1898–1974 · extragalactic astronomy, cluster masses, and the discovery of missing mass

Fritz Zwicky was a Swiss-Bulgarian-American astrophysicist at Caltech whose instincts ran decades ahead of his evidence and whose personality ensured he was rarely thanked for it. In 1933, applying the virial theorem to the velocity dispersion of galaxies in the Coma cluster, he found that the visible galaxies supplied only a small fraction of the mass needed to bind it — the rest he called "dunkle Materie", dark matter. The estimate was crude and, for the era's Hubble constant, numerically off, but the conclusion was exactly right and is the founding measurement of the field. In the same fertile decade he and Walter Baade coined "supernova", proposed that supernovae are powered by collapse to a neutron star, and predicted that galaxy clusters would act as gravitational lenses — all of which came true. A relentless, abrasive figure who feuded freely with colleagues, Zwicky saw his dark-matter claim lie dormant until the 1970s, when Rubin and Ford's flat rotation curves made the missing mass impossible to ignore and the cold dark matter paradigm — named only in 1984 — vindicated the number he had wrung from Coma half a century before.

History

1933

— Zwicky infers unseen mass in the Coma cluster from galaxy velocity dispersions

1970

— Rubin and Ford measure flat rotation curves in the Andromeda galaxy

1984

— Blumenthal, Faber, Primack and Rees articulate the cold dark matter scenario

2006

— the Bullet Cluster separates lensing mass from the shocked X-ray gas

2013

— Planck fixes the cold dark matter density at Ω_c h² ≈ 0.12

Going further

Texts. Ryden, Introduction to Cosmology, 2nd ed., Ch. 8 (Dark Matter) and Ch. 12 (Structure Formation) · Dodelson, Modern Cosmology, 2nd ed., Ch. 3–4, 7 (Boltzmann equations, growth of perturbations, the Mészáros equation) · Mo, van den Bosch & White, Galaxy Formation and Evolution, Ch. 4 (linear growth and transfer functions) · Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., Ch. 24 and 30 (galaxy dynamics, cosmology) · Binney & Tremaine, Galactic Dynamics, 2nd ed., Ch. 1, 4 (collisionless systems, dark halos). The density parameters ( $\Omega_{\mathrm{c}}h^2 \approx 0.12$ , $\Omega_{\mathrm{b}}h^2 \approx 0.022$ , $\Omega_{\mathrm{m}} \approx 0.31$ ) and $z_{\mathrm{eq}} \approx 3400$ are standard Planck-era concordance values. Key numerical checks: $1+z_{\mathrm{eq}} \approx 3446$ , $M(<50\ \mathrm{kpc}) \approx 5.6\times10^{11}\,M_\odot$ , the EdS indicial roots $n = 2/3$ and $n = -1$ , and the Mészáros growth factor of $5/2$ across the radiation era.

Prerequisites. Density Perturbations and the Density Contrast · Jeans Instability and Cosmological Perturbation Regimes · Linear Growth of Structure. Leads to. The $\Lambda$ CDM Concordance Model · Transfer Function and the Matter Power Spectrum · The CMB Anisotropy Power Spectrum · Dark Energy and the Cosmological Constant.

Ask the tutor: ask me to walk through the Mészáros equation step by step, or to show numerically how the $P(k)$ turnover at $k_{\mathrm{eq}}$ moves when you change $\Omega_{\mathrm{c}}h^2$ .

Is dark matter just ordinary matter that happens not to shine?

No — this is settled by physics independent of luminosity. Big Bang nucleosynthesis predicts the primordial deuterium, helium, and lithium abundances as functions of the baryon density, and the CMB acoustic peaks weigh the baryons a second, independent way; both give $\Omega_{\mathrm{b}}h^2 \approx 0.022$ . That total counts every proton and neutron whether it sits in a star, a gas cloud, a planet, or a black hole made from ordinary matter. The dynamical and lensing mass is about five times larger, so the surplus cannot be baryons in any disguise. Dark matter must be a new, non-baryonic species.

Why must the dark matter be cold rather than hot?

Because hot (relativistic) dark matter free-streams. A fast particle drifts out of any perturbation smaller than its free-streaming length before it can collapse, erasing small-scale structure and forcing a top-down history in which superclusters form first and fragment. Observation shows the opposite: small galaxies are old and structure is built bottom-up. Only matter that was non-relativistic when the relevant scales entered the horizon — cold matter — has a negligible free-streaming length and can seed structure on every scale from dwarfs upward. Light neutrinos are the classic hot candidate, and their failure to make small-scale structure is exactly why they cannot be the dark matter.

What is the Mészáros effect, in plain terms?

It is the near-freezing of dark matter perturbations during the radiation-dominated era. While radiation controls the expansion, the universe expands so fast that Hubble drag overwhelms the weak self-gravity of the sub-dominant dark matter, so a perturbation inside the horizon can barely grow. The growing mode is $\delta_{\mathrm{c}} = 1 + \tfrac32 y$ with $y = a/a_{\mathrm{eq}}$ , so across the whole radiation era a mode grows by only a factor of $5/2$ ; growth then resumes as $\delta \propto a$ once matter dominates. Because small-scale modes entered the horizon earliest, they lost the most growth — which is why the matter power spectrum turns over at $k_{\mathrm{eq}}$ .

How does cold dark matter give the baryons a "head start"?

Before recombination the baryons are tied to the photons, so radiation pressure keeps their small-scale contrast oscillating near zero instead of growing. Cold dark matter feels none of this — decoupled from the outset, it has grown as $\delta_{\mathrm{c}} \propto a$ since equality, digging ever-deeper wells. When the baryons are released at recombination they find those wells waiting and fall in, catching up as $\delta_{\mathrm b} \to \delta_{\mathrm c}(1 - a_{\mathrm{rec}}/a)$ within a factor of a few in scale factor. Without the head start, baryons starting from the observed $\delta \sim 10^{-5}$ would reach only $\delta \sim 10^{-2}$ today — nowhere near collapse.

What actually sets the density profile of a dark matter halo?

A flat rotation curve gives only the coarse fact $M(<r) \propto r$ , implying $\rho \propto r^{-2}$ in the outer parts. The detailed shape comes from the collisionless dynamics of hierarchical merging, which N-body simulations solve by integrating the Vlasov–Poisson system. The generic outcome is the Navarro–Frenk–White profile, $\rho \propto 1/[(r/r_{\mathrm{s}})(1+r/r_{\mathrm{s}})^2]$ , cuspy ( $\rho \propto r^{-1}$ ) at the centre and $\propto r^{-3}$ far out. Whether real galaxy cores are as cuspy as pure-CDM simulations predict — the cusp–core problem — is a live small-scale tension of the paradigm.

If we have never detected the particle, how confident should we be it exists?

Very confident that something with its gravitational properties exists, and genuinely uncertain what it is. The case is not one measurement but a web: rotation curves, cluster virial masses, lensing, the Bullet Cluster's mass–gas offset, the CMB acoustic-peak heights, the power-spectrum turnover, and the very existence of galaxies all demand it and all agree on $\Omega_{\mathrm{c}}h^2 \approx 0.12$ . Modified-gravity alternatives can fit rotation curves but struggle with clusters, the Bullet, and the CMB together. What remains open is the microphysics — WIMP, axion, sterile neutrino, primordial black hole — which is what the direct, indirect, and collider searches chase.