The Cosmological Constant

The cosmological constant, $\Lambda$ , is the one loose thread Einstein's equations can never quite cut. Ask for the most general law of gravity that keeps general relativity's defining virtues — general covariance, second-order field equations, and automatic conservation of energy-momentum — and the answer is not $G_{\mu\nu}=8\pi G\,T_{\mu\nu}$ but $G_{\mu\nu}+\Lambda g_{\mu\nu}=8\pi G\,T_{\mu\nu}$ . Lovelock's theorem makes this sharp: in four dimensions the most general symmetric, divergence-free tensor built from the metric and its first two derivatives is a linear combination of $G_{\mu\nu}$ and $g_{\mu\nu}$ . So $\Lambda$ is not a decoration bolted on; it is a second fundamental constant of gravitation beside Newton's $G$ , and setting it to zero is a choice, not a theorem. What makes $\Lambda$ inexhaustible is that it wears two faces. On the left it is pure geometry — a constant curvature that empty spacetime carries even with no matter present. Slide it to the right and it becomes physics — a stress-energy $T^{(\Lambda)}_{\mu\nu}=-\rho_\Lambda g_{\mu\nu}$ , the unique energy content that looks identical to every freely-falling observer, which Lorentz invariance forces to have the strange equation of state $p=-\rho$ . That negative pressure is the whole story: because pressure gravitates in general relativity, a large enough tension flips gravity's sign and pushes the universe apart. Einstein reached for $\Lambda$ in 1917 to hold a static cosmos together; Hubble's expansion seemed to make it pointless; then in 1998 the accelerating universe brought it back as the leading face of dark energy.

Key takeaways

The cosmological constant $\Lambda$ (units of inverse length squared) is the unique term that can be added to the Einstein field equations, $G_{\mu\nu}+\Lambda g_{\mu\nu}=8\pi G\,T_{\mu\nu}$ , without breaking general covariance or local energy-momentum conservation.
Moved to the right-hand side it is a vacuum stress-energy $T^{(\Lambda)}_{\mu\nu}=-\rho_\Lambda g_{\mu\nu}$ with density $\rho_\Lambda=\Lambda/(8\pi G)$ and equation of state $p_\Lambda=-\rho_\Lambda$ , i.e. $w=-1$ .
Because pressure gravitates through the combination $\rho+3p$ , the vacuum's negative pressure makes $\Lambda>0$ repulsive: it drives accelerated expansion, $\ddot a>0$ , the observed behaviour of dark energy.
Matter dilutes as $a^{-3}$ and radiation as $a^{-4}$ , but vacuum energy is $a^0$ — constant — so however subdominant $\Lambda$ once was, it inevitably comes to dominate and expansion accelerates forever.
Observation pins $\Omega_\Lambda\approx0.69$ , giving $\Lambda\approx1.1\times10^{-52}\ \mathrm{m}^{-2}$ and $\rho_\Lambda\approx5.8\times10^{-27}\ \mathrm{kg/m^3}$ — while quantum field theory naively predicts a value about $10^{120}$ times larger, the cosmological constant problem.
Key idea: The cosmological constant is the unique term you can add to Einstein's equations without breaking general covariance or energy-momentum conservation; it is simultaneously a geometric constant of spacetime and a Lorentz-invariant vacuum energy with $p=-\rho$ , and a positive value sources accelerated cosmic expansion.

Where this lives today

The cosmological constant is now a measured quantity under active assault by the largest galaxy surveys ever built. The strategy is to test whether dark energy really has $w=-1$ exactly, as a true $\Lambda$ demands, or whether $w$ drifts — which would signal a dynamical field ("quintessence"). DESI maps millions of galaxies to trace the expansion history through baryon acoustic oscillations; Euclid and Rubin's LSST add weak-lensing maps of the dark matter. These datasets remain consistent with $w=-1$ to a few percent, though DESI's first results hinted, at modest significance, at a possible time-varying equation of state — a claim that, if it survived, would rival the 1998 discovery. Meanwhile the tension between early- and late-universe measurements of $H_0$ keeps pressure on the whole $\Lambda$ CDM framework, and on the theory side $\Lambda<0$ (anti-de Sitter) is the indispensable backdrop of the AdS/CFT correspondence. The one number Einstein added and retracted now anchors both observational cosmology and the deepest ideas in quantum gravity.

The physical picture

Start from the fact that makes gravity in general relativity different from Newton's: it is not mass that gravitates but energy and pressure together, in the combination $\rho+3p$ . For ordinary matter this is positive, and a ball of test particles released at rest converges — attraction, as expected. But the vacuum energy of $\Lambda$ carries a pressure equal to minus its energy density, $p_\Lambda=-\rho_\Lambda$ , so the combination becomes $\rho_\Lambda+3p_\Lambda=-2\rho_\Lambda<0$ . The sign flips, the ball diverges, and a positive energy density produces repulsive gravity — the single counter-intuitive move on which everything rests, shown directly in the first figure.

Figure

Pressure gravitates in general relativity: the source in the acceleration equation ä/a = −(4πG/3)(ρ + 3p) is the combination ρ + 3p. Ordinary matter has ρ + 3p = ρ > 0 and pulls a cloud of test particles inward, but the vacuum has p = −ρ so ρ + 3p = −2ρ_Λ < 0 and pushes them apart — repulsive gravity from a positive energy density

Why must the vacuum have $p=-\rho$ ? Because it looks the same to every inertial observer — you cannot tell you are moving through empty space. A stress-energy with that property must be built from the only rank-two tensor invariant under all Lorentz boosts, the metric itself, so $T^{(\Lambda)}_{\mu\nu}\propto g_{\mu\nu}$ ; matching that to the perfect-fluid form forces $\rho+p=0$ . Lorentz invariance alone, before any cosmology, fixes the equation of state. Geometrically the same $\Lambda$ sets a curvature scale for empty space: with $\Lambda>0$ the vacuum is not flat Minkowski but de Sitter spacetime, expanding exponentially at rate $H=\sqrt{\Lambda/3}$ , while $\Lambda<0$ gives anti-de Sitter. The last piece is why $\Lambda$ , however tiny, always wins. As the universe expands, matter thins as $\rho_m\propto a^{-3}$ and radiation faster still as $\rho_r\propto a^{-4}$ , because expanding space both dilutes particles and stretches their wavelengths. Vacuum energy does neither: it is $\rho_\Lambda\propto a^0$ , a fixed energy per unit volume, so making more space simply makes more of it. The second figure tracks the three components across cosmic history and shows the inevitable crossover.

Figure

Energy density versus scale factor on logarithmic axes: radiation ∝ a⁻⁴ and matter ∝ a⁻³ dilute as the universe expands while the vacuum ρ_Λ stays constant, so with Ω_Λ = 0.685 today the cosmological constant overtakes matter near redshift z ≈ 0.3 and dominates the far future

The core equations

The starting point is Einstein's equation carrying the extra term. The left side is pure geometry, the right side is matter, and $\Lambda g_{\mu\nu}$ is the one geometric object that can join $G_{\mu\nu}$ without spoiling the conservation law the right side must obey. With $c=1$ ,

G_{\mu\nu} + \Lambda g_{\mu\nu} = 8\pi G\, T_{\mu\nu} \qquad (1)

where: $G_{\mu\nu}=R_{\mu\nu}-\tfrac12 R g_{\mu\nu}$ — Einstein tensor; $\Lambda$ — cosmological constant, units length $^{-2}$ ; $g_{\mu\nu}$ — metric; $T_{\mu\nu}$ — stress-energy of matter and radiation; $G$ — Newton's constant. Reading: $\Lambda$ curves spacetime even where $T_{\mu\nu}=0$ , so the vacuum is not featureless. Carry the term to the right and it becomes an energy content of the vacuum — nothing changes but the bookkeeping, which exposes its physical meaning.

T^{(\Lambda)}_{\mu\nu} = -\frac{\Lambda}{8\pi G}\, g_{\mu\nu} = -\rho_\Lambda\, g_{\mu\nu} \qquad (2)

where: $\rho_\Lambda=\Lambda/(8\pi G)$ — vacuum energy density; $g_{\mu\nu}$ — the only Lorentz-invariant rank-two tensor available. Reading: a positive $\Lambda$ is a positive vacuum energy, and the minus sign before $g_{\mu\nu}$ is what becomes a negative pressure. Expanding that vacuum stress-energy in perfect-fluid language gives its density, pressure, and equation-of-state parameter $w$ .

\rho_\Lambda = \frac{\Lambda}{8\pi G},\qquad p_\Lambda = -\rho_\Lambda,\qquad w \equiv \frac{p_\Lambda}{\rho_\Lambda} = -1 \qquad (3)

where: $p_\Lambda$ — vacuum pressure, a tension; $w$ — equation-of-state parameter. Reading: $w=-1$ is the operational definition of a cosmological constant, and measuring $w$ is how observers test whether dark energy really is $\Lambda$ . Fed into cosmology, the second Friedmann (acceleration) equation shows the payoff: $\Lambda$ contributes a term of definite positive sign to $\ddot a$ .

\frac{\ddot a}{a} = -\frac{4\pi G}{3}(\rho + 3p) + \frac{\Lambda}{3} \qquad (4)

where: $a(t)$ — scale factor; $\rho,p$ — density and pressure of matter and radiation only. Reading: with matter negligible, $\ddot a/a=\Lambda/3>0$ — an empty universe with $\Lambda>0$ accelerates. The first Friedmann equation is the companion energy constraint, showing $\Lambda$ as a constant floor under the expansion rate that neither matter nor curvature can erode.

H^2 = \left(\frac{\dot a}{a}\right)^2 = \frac{8\pi G}{3}\rho - \frac{k}{a^2} + \frac{\Lambda}{3} \qquad (5)

where: $H$ — Hubble parameter; $k=+1,0,-1$ — spatial curvature; $\Lambda/3$ — constant vacuum term. Reading: as $\rho$ and $k/a^2$ redshift away, $H\to\sqrt{\Lambda/3}$ — the de Sitter attractor. Setting $T_{\mu\nu}=0$ and tracing reveals the geometry a vacuum $\Lambda$ produces: an Einstein space of constant curvature.

R_{\mu\nu} - \tfrac12 R g_{\mu\nu} + \Lambda g_{\mu\nu} = 0 \;\Rightarrow\; R = 4\Lambda,\;\; R_{\mu\nu} = \Lambda g_{\mu\nu} \qquad (6)

where: $R_{\mu\nu}$ — Ricci tensor; $R$ — Ricci scalar; the trace in $D=4$ uses $g^{\mu\nu}g_{\mu\nu}=4$ . Reading: the maximally symmetric solutions are de Sitter ( $\Lambda>0$ ) and anti-de Sitter ( $\Lambda<0$ ). To compare with observation, $\Lambda$ is packaged as a fraction of the critical density, the density parameter $\Omega_\Lambda$ .

\Omega_\Lambda \equiv \frac{\rho_\Lambda}{\rho_{\rm crit}} = \frac{\Lambda}{3H_0^2},\qquad \rho_{\rm crit}=\frac{3H_0^2}{8\pi G} \qquad (7)

where: $\Omega_\Lambda\approx0.69$ today; $H_0$ — present Hubble constant; $\rho_{\rm crit}$ — density for exact flatness. Reading: in a flat universe $\Omega_\Lambda+\Omega_m+\Omega_r=1$ , so $\Lambda$ supplies about two-thirds of the present energy budget. At the level of the action, $\Lambda$ is a constant added to the gravitational Lagrangian — which is why it is unavoidable and why quantum field theory insists on contributing to it.

S = \frac{1}{16\pi G}\int \mathrm{d}^4x\,\sqrt{-g}\,(R - 2\Lambda) + S_{\rm matter} \qquad (8)

where: $\sqrt{-g}$ — invariant volume element; $-2\Lambda$ — a constant vacuum-energy term. Reading: varying this action returns equation (1), and any constant in the Lagrangian is indistinguishable from $\Lambda$ .

Derivation AWhy the metric term is the only addition allowed

The goal is to prove the uniqueness that gives $\Lambda$ its status: among all ways to modify the left-hand side of Einstein's equations, adding a constant multiple of the metric is the only one that respects both the geometric structure of the theory and the conservation of energy-momentum. The argument follows from one fact — that the covariant divergence of the matter stress-energy vanishes — with Lovelock's theorem closing it in four dimensions.

Step 1. Physical matter obeys $\nabla^\mu T_{\mu\nu}=0$ ; since the field equation ties $T_{\mu\nu}$ to a geometric tensor, whatever sits on the left must itself be identically divergence-free. Why: otherwise the field equation would silently impose an extra, unphysical constraint on the matter.

Step 2. The Einstein tensor already passes this test through the contracted Bianchi identity, $\nabla^\mu G_{\mu\nu}=0$ , which holds for any metric as an identity. Why: this is exactly why Einstein used $G_{\mu\nu}$ , not the Ricci tensor, whose divergence does not vanish.

Step 3. Metric compatibility gives $\nabla_\alpha g_{\mu\nu}=0$ , so $\nabla^\mu(\Lambda g_{\mu\nu})=0$ for any constant $\Lambda$ . Why: adding $\Lambda g_{\mu\nu}$ preserves conservation for free — and fails the moment $\Lambda$ varies, which is why it must be a true constant, not a field.

Step 4. Lovelock's theorem closes it: in four dimensions the most general symmetric, divergence-free tensor built from the metric and its first two derivatives is $\alpha G_{\mu\nu}+\beta g_{\mu\nu}$ . Why: no other term is admissible at this level in 4D, so $G_{\mu\nu}+\Lambda g_{\mu\nu}$ is the whole space of options. *Together, $\nabla^\mu G_{\mu\nu}=0$ (the contracted Bianchi identity) and $\nabla^\mu g_{\mu\nu}=0$ (metric compatibility) guarantee that $G_{\mu\nu}+\Lambda g_{\mu\nu}$ is identically divergence-free for any constant $\Lambda$ ; Lovelock's theorem then establishes that this combination is the unique such tensor in four dimensions.*

Go deeper · Derivation B · 1 minFrom Lorentz invariance to the vacuum equation of state

Derivation BFrom Lorentz invariance to the vacuum equation of state

The aim is to show that $p=-\rho$ is not an assumption imposed on the vacuum but a theorem forced by symmetry: requiring the vacuum stress-energy to look the same to every freely-falling observer fixes it proportional to the metric, and matching to the perfect-fluid form leaves only the overall scale, $\rho_\Lambda=\Lambda/(8\pi G)$ . This is the argument that welds the geometric $\Lambda$ to the physical idea of a vacuum energy, and it is why $w=-1$ is the defining signature of a cosmological constant.

Step 1. Write the general perfect fluid, $T_{\mu\nu}=(\rho+p)u_\mu u_\nu + p\,g_{\mu\nu}$ , with $u^\mu u_\mu=-1$ . Why: this is the most general isotropic stress-energy, and it carries a preferred rest frame through $u^\mu$ — exactly the structure the vacuum must lack.

Step 2. Demand invariance under all Lorentz boosts; the only rank-two tensor with that property is a multiple of the metric, so $T^{(\Lambda)}_{\mu\nu}\propto g_{\mu\nu}$ . Why: any $u_\mu u_\nu$ piece would single out a rest frame and let an observer detect motion through the vacuum, which is forbidden.

Step 3. For the fluid form to collapse onto a pure $g_{\mu\nu}$ term, the coefficient of $u_\mu u_\nu$ must vanish: $\rho+p=0$ , i.e. $p=-\rho$ . Why: this is the whole content of $w=-1$ — a tension exactly equal in magnitude to the energy density.

Step 4. Then $T^{(\Lambda)}_{\mu\nu}=-\rho\,g_{\mu\nu}$ ; comparing with the term $-\tfrac{\Lambda}{8\pi G}g_{\mu\nu}$ carried over from geometry gives $\rho_\Lambda=\Lambda/(8\pi G)$ . Why: this is the dictionary between the two faces — a length $^{-2}$ curvature becomes an energy density, differing by $8\pi G$ . *The perfect-fluid stress-energy reduces to $-\rho g_{\mu\nu}$ precisely when $\rho+p=0$ , and matching that to the geometric term gives $\rho_\Lambda=\Lambda/(8\pi G)$ .*

Go deeper · Derivation C · 1 minEinstein's static universe and why it collapses

Derivation CEinstein's static universe and why it collapses

This derivation reconstructs the 1917 motivation for $\Lambda$ and its fatal flaw: a positive $\Lambda$ can be tuned to cancel the pull of matter and hold a closed universe static, but the equilibrium is unstable, so the slightest disturbance sends it expanding or collapsing without return. It is a clean application of the two Friedmann equations to pressureless dust, explaining both why Einstein introduced the constant and why the static solution could not survive.

Step 1. Impose $\dot a=\ddot a=0$ on dust ( $p=0$ ) with $k=+1$ . From (4), $0=-\tfrac{4\pi G}{3}\rho_m+\tfrac{\Lambda}{3}$ , so $\Lambda=4\pi G\rho_m$ . Why: this is the tuning — repulsion set exactly equal to attraction so the universe neither expands nor contracts.

Step 2. Use the constraint (5) with $\dot a=0$ : $0=\tfrac{8\pi G}{3}\rho_m-\tfrac{1}{a^2}+\tfrac{\Lambda}{3}$ . Substituting $\tfrac{\Lambda}{3}=\tfrac{4\pi G}{3}\rho_m$ gives $\tfrac{1}{a^2}=4\pi G\rho_m=\Lambda$ , so $a=\Lambda^{-1/2}$ . Why: the size of Einstein's universe is not free — it is fixed by $\Lambda$ alone.

Step 3. Perturb $a\to a(1+\epsilon)$ . Growing $a$ dilutes matter as $a^{-3}$ while $\Lambda$ stays fixed, so repulsion wins and expansion runs away; shrinking $a$ enhances matter, attraction wins, and it collapses. Why: the equilibrium sits at a maximum of the effective potential — unstable, not restoring.

Step 4. So the static universe is doubly doomed: it needs fine-tuning, and even tuned it cannot persist. Why: Hubble's 1929 expansion then removed the motivation, and Einstein reportedly called $\Lambda$ his greatest blunder — prematurely. *The balance $\Lambda=4\pi G\rho_m=a^{-2}$ is the standard Einstein static result, and the $a^{-3}$ dilution of matter against a constant $\Lambda$ is precisely what confirms the instability.*

Limiting cases

The behaviour of $\Lambda$ is governed by its sign and by what else shares the spacetime.

Case	Result	Meaning
Vacuum, $\Lambda>0$ (de Sitter)	$a(t)\propto e^{Ht},\ H=\sqrt{\Lambda/3}$	empty space with positive $\Lambda$ expands exponentially at constant rate — the asymptotic fate of our universe
Vacuum, $\Lambda<0$ (anti-de Sitter)	$R_{\mu\nu}=\Lambda g_{\mu\nu},\ \ell=\sqrt{-3/\Lambda}$	constant negative curvature with a timelike boundary; the arena of the AdS/CFT correspondence
Weak-field limit	$\nabla^2\Phi = 4\pi G\rho - \Lambda$	$\Lambda$ adds a constant repulsive source; around a point mass an outward force $+\tfrac13\Lambda c^2 r$ growing with distance
$\Lambda\to 0$	$G_{\mu\nu}=8\pi G\,T_{\mu\nu}$	recovers ordinary general relativity; asymptotically flat spacetimes require $\Lambda=0$
Einstein static universe	$\Lambda = 4\pi G(\rho+3p),\ a^2=\Lambda^{-1}$	fine-tuned balance of dust and $\Lambda$ , unstable to any perturbation

Common trapA positive vacuum energy should pull the universe together, not push it apart

This is the trap that snares almost everyone: positive energy density gravitates attractively for ordinary matter, so surely a positive vacuum energy $\rho_\Lambda>0$ should decelerate the expansion. It does the opposite, because in general relativity pressure gravitates too. The source of acceleration is not $\rho$ but $\rho+3p$ , and the vacuum's $p_\Lambda=-\rho_\Lambda$ makes $\rho_\Lambda+3p_\Lambda=-2\rho_\Lambda<0$ : the negative pressure overwhelms the positive energy density and the net effect is repulsive — which is why $\Lambda>0$ violates the strong energy condition ( $\rho+3p\ge0$ ). A related slip is to confuse $\Lambda$ itself (length $^{-2}$ ) with the energy density $\rho_\Lambda=\Lambda/(8\pi G)$ (energy per volume); they differ dimensionally by $8\pi G$ .

Worked examples

1 · From the sky to the number: $\Lambda$ and the density of the vacuum (fully worked)

Given $H_0=67.4\ \mathrm{km\,s^{-1}\,Mpc^{-1}}$ and $\Omega_\Lambda=0.685$ , find $\Lambda$ , the vacuum energy density $\rho_\Lambda$ , and express it as hydrogen atoms per cubic metre.

Step 1. Put $H_0$ in SI. With $1\ \mathrm{Mpc}=3.086\times10^{19}\ \mathrm{km}$ , $H_0=67.4/(3.086\times10^{19})=2.184\times10^{-18}\ \mathrm{s^{-1}}$ .

Step 2. From (7) with $c$ restored, $\Lambda=3\Omega_\Lambda H_0^2/c^2=3(0.685)(2.184\times10^{-18})^2/(2.998\times10^8)^2$ , giving $\boxed{\Lambda\approx1.1\times10^{-52}\ \mathrm{m^{-2}}}$ — a curvature radius $\Lambda^{-1/2}\approx10\ \mathrm{Gly}$ , of order the observable universe, which is why $\Lambda$ matters now and not in the lab.

Step 3. Convert with $\rho_\Lambda=\Lambda c^2/(8\pi G)=(1.1\times10^{-52})(2.998\times10^8)^2/(8\pi\cdot6.674\times10^{-11})\approx\mathbf{5.8\times10^{-27}\ kg\,m^{-3}}$ , an energy density $\rho_\Lambda c^2\approx5.3\times10^{-10}\ \mathrm{J\,m^{-3}}$ .

Step 4. Interpret. Dividing by the hydrogen mass $1.67\times10^{-27}\ \mathrm{kg}$ gives about $\mathbf{3.5}$ hydrogen atoms per cubic metre — all of dark energy is a few atoms' worth in a small room. In particle units $\rho_\Lambda\approx2.5\times10^{-47}\ \mathrm{GeV^4}$ , a scale $\rho_\Lambda^{1/4}\approx2.3\ \mathrm{meV}$ . Dark energy lives at a milli-electron-volt scale — and that tiny number is the crux of Example 2.

2 · The cosmological constant problem in one estimate

Quantum field theory fills the vacuum with zero-point energy. Estimate its density with a Planck-energy cutoff $E_{\rm Pl}=1.22\times10^{19}\ \mathrm{GeV}$ and compare with the observed $\rho_\Lambda\approx2.5\times10^{-47}\ \mathrm{GeV^4}$ .

Step 1. A vacuum energy density has dimensions (energy) $^4$ , and the only scale is the cutoff, so $\rho_{\rm vac}^{\rm QFT}\sim E_{\rm Pl}^4$ .

Step 2. Evaluate: $E_{\rm Pl}^4=(1.22\times10^{19})^4\approx\mathbf{2.2\times10^{76}\ GeV^4}$ .

Step 3. Take the ratio: $2.2\times10^{76}/2.5\times10^{-47}\approx\mathbf{9\times10^{122}}$ — theory exceeds observation by about 120 orders of magnitude, the largest mismatch in physics.

Step 4. The moral: either an unknown mechanism cancels the vacuum energy to 120 decimal places and leaves exactly the milli-eV residue of Example 1, or our reasoning about vacuum energy is incomplete. This is the cosmological constant problem, and no accepted solution exists.

3 · The de Sitter future

If matter and radiation eventually dilute away, what expansion does the leftover $\Lambda$ impose?

Step 1. As $\rho\to0$ , equation (5) leaves $H\to H_\infty=\sqrt{\Lambda/3}=H_0\sqrt{\Omega_\Lambda}=67.4\sqrt{0.685}\approx\mathbf{55.8\ km\,s^{-1}\,Mpc^{-1}}$ — de Sitter space.

Step 2. Then $a\propto e^{H_\infty t}$ with e-folding time $1/H_\infty\approx\mathbf{17.5\ Gyr}$ and event horizon $c/H_\infty\approx\mathbf{17.5\ Gly}$ . A universe with $\Lambda>0$ ends cold, empty, and exponentially expanding, each observer walled inside a de Sitter horizon.

Physics of the portrait

The cosmological constant is inseparable from one man, who invented it, disowned it, and never guessed it would return as the dominant substance of the universe.

Origins · 1 minAlbert Einstein and how this was discovered

Albert Einstein

1879–1955 · general relativity and relativistic cosmology

Einstein completed general relativity in November 1915 and within two years turned it on the universe as a whole. His 1917 paper "Cosmological Considerations in the General Theory of Relativity" faced a problem: the field equations as first written could not produce the static, eternal cosmos astronomers then believed in. To force a static solution he added a term — the cosmological constant $\Lambda$ — tuned so its repulsion balanced the gravitation of matter. The resulting Einstein static universe was closed and, as de Sitter and Eddington showed, dynamically unstable. When Hubble's 1929 distance-velocity relation established that the universe is expanding, the motivation vanished; Gamow later reported that Einstein called $\Lambda$ his "biggest blunder." Yet Einstein's deeper instinct — that the equations permit the term whether or not one wants it — proved sounder than his retraction. The constant never left the mathematics, because nothing forbids it, and in 1998 the discovery of cosmic acceleration resurrected $\Lambda$ as the leading description of dark energy. The blunder became the concordance model — and the same de Sitter whose empty-universe solution first embarrassed the static model gave his name to the exponentially expanding spacetime a positive $\Lambda$ now points us toward.

History

1915

— Einstein completes the general theory of relativity

1917

— Einstein introduces Λ for a static universe; de Sitter finds the empty expanding solution

1922

— Friedmann derives expanding and contracting solutions with and without Λ

1929

— Hubble establishes the distance-velocity relation, evidence for cosmic expansion

1998

— supernova surveys reveal accelerating expansion, reviving Λ as dark energy

2011

— Nobel Prize in Physics for the discovery of accelerating expansion

Going further

Texts. Misner, Thorne & Wheeler, Gravitation, §16.2 and §27.11 · Wald, General Relativity, §4.3 and §5.2 · Carroll, Spacetime and Geometry, §4.5 and §8.5 · Hartle, Gravity, Ch. 18–19 · Schutz, A First Course in General Relativity, §8.3 and Ch. 12 · Weinberg, "The Cosmological Constant Problem," Rev. Mod. Phys. 61, 1 (1989) · Planck 2018 Results VI, A&A 641, A6 (2020). Key results used throughout: the trace of the vacuum field equations gives $R=4\Lambda$ and $R_{\mu\nu}=\Lambda g_{\mu\nu}$ in $D=4$ ; the perfect-fluid reduction yields $\rho_\Lambda=\Lambda/(8\pi G)$ and $w=-1$ ; the dilution law is $\rho\propto a^{-3(1+w)}$ and the de Sitter rate is $H=\sqrt{\Lambda/3}$ ; the Einstein static balance is $\Lambda=4\pi G\rho_m=a^{-2}$ . Numbers from $H_0=67.4\ \mathrm{km\,s^{-1}\,Mpc^{-1}}$ , $\Omega_\Lambda=0.685$ : $\Lambda\approx1.1\times10^{-52}\ \mathrm{m^{-2}}$ , $\rho_\Lambda\approx5.8\times10^{-27}\ \mathrm{kg\,m^{-3}}\approx2.5\times10^{-47}\ \mathrm{GeV^4}$ ( $\approx3.5$ H atoms per m³); QFT estimate $E_{\rm Pl}^4\approx2.2\times10^{76}\ \mathrm{GeV^4}$ , a $\sim10^{123}$ discrepancy; de Sitter fate $H_\infty\approx55.8\ \mathrm{km\,s^{-1}\,Mpc^{-1}}$ , e-folding $\approx17.5\ \mathrm{Gyr}$ ; concordance $\Omega_\Lambda\approx0.69$ , $\Omega_m\approx0.31$ .

Builds on: The Einstein Field Equations · The Einstein Tensor · The Stress-Energy Tensor · Perfect Fluid Stress-Energy · Local Conservation of Energy-Momentum. Leads to: The Energy Conditions · The Einstein-Hilbert Action · Variational Derivation of the Field Equations.

Ask the tutor: Ask me to trace $\rho+3p$ from the acceleration equation to the accelerating sky, and show why the vacuum's negative pressure — not its energy density — is what flips gravity from pulling to pushing.

Is the cosmological constant the same thing as dark energy?

Not quite — the cosmological constant is the simplest candidate for dark energy, not a synonym. "Dark energy" is the observational name for whatever drives cosmic acceleration: a smooth component with pressure negative enough to give $\ddot a>0$ . A cosmological constant is the specific hypothesis that this component is exactly Lorentz-invariant vacuum energy with $w=-1$ and a density that never changes. Other candidates, collectively "quintessence," are dynamical fields with $w$ near but not equal to $-1$ . So far every measurement is consistent with $\Lambda$ , but distinguishing a true constant from a slowly rolling field is the central goal of surveys like DESI and Euclid.

Was it really Einstein's "biggest blunder"?

The phrase comes secondhand, from George Gamow's memoir, and Einstein never wrote it himself, so treat it as folklore. What is documented is that he introduced $\Lambda$ in 1917 for a static universe and abandoned it after Hubble's expansion made a static cosmos untenable. The irony is layered: the real misstep, if any, was removing $\Lambda$ , since nothing in the theory forbids the term, and in 1998 it returned as the leading account of dark energy. The lesson usually drawn is that once a term is allowed by a theory's symmetries, you generally do not get to set it to zero by decree.

How can empty space have energy — isn't the vacuum "nothing"?

"Empty" means the lowest-energy state, not a state of zero energy. In quantum field theory every field still fluctuates in its ground state — the zero-point energy of countless oscillators — present even with no particles. Because it is the same everywhere and for every observer, it takes the Lorentz-invariant form $-\rho_\Lambda g_{\mu\nu}$ and behaves exactly like a cosmological constant. The trouble is quantitative, not conceptual: the naive estimate is about $10^{120}$ times too large, so while the existence of vacuum energy is not in doubt, its observed smallness is one of the deepest puzzles in physics.

Why does positive Λ push things apart when gravity is attractive?

Because in general relativity the source of gravity is $\rho+3p$ , not $\rho$ alone — pressure gravitates. Ordinary matter has $p\approx0$ , so its source is $\rho>0$ and it attracts. The vacuum has $p=-\rho$ , so its source is $\rho+3p=-2\rho<0$ : the negative pressure (a tension) more than cancels the positive energy density, and the net effect is repulsive. This is the same $\rho+3p$ that appears in the acceleration equation and in the strong energy condition, which $\Lambda>0$ deliberately violates. Remember one thing: it is the pressure, not the energy density, that flips gravity's sign for the vacuum.

Does the cosmological constant change over time?

By definition, no — a cosmological constant is constant in space and time, exactly what the conservation argument demands: allowing $\Lambda$ to vary would break $\nabla^\mu(\Lambda g_{\mu\nu})=0$ . Its energy density stays fixed while matter and radiation dilute, which is why it comes to dominate the late universe. Observers do not assume constancy, though; they test it by measuring $w$ at different redshifts. If $w$ departed from $-1$ or evolved, the accelerating agent would not be a cosmological constant but a dynamical dark-energy field — which current data disfavour, with only tentative hints otherwise.

If our universe has Λ > 0, why does anyone care about anti-de Sitter?

Because anti-de Sitter space is the most useful theoretical laboratory in modern gravity, even though it is not our universe. A negative $\Lambda$ gives a spacetime with a timelike boundary — a box with reflecting walls — and that boundary makes the AdS/CFT correspondence precise: gravity in $(d+1)$ -dimensional anti-de Sitter is exactly dual to a conformal field theory on its $d$ -dimensional edge. This holographic duality is the sharpest known handle on quantum gravity and on strongly coupled systems, from quark-gluon plasmas to condensed matter. So $\Lambda<0$ is prized not for cosmology but as the setting where gravity and quantum field theory become two descriptions of one thing.

The Cosmological Constant

Where this lives today

The physical picture

The core equations

Limiting cases

Worked examples

1 · From the sky to the number: $\Lambda$ and the density of the vacuum (fully worked)

2 · The cosmological constant problem in one estimate

3 · The de Sitter future

Physics of the portrait

Going further

Before this

Where this leads

The Cosmological Constant

Where this lives today

The physical picture

The core equations

Limiting cases

Worked examples

1 · From the sky to the number: Λ\LambdaΛ and the density of the vacuum (fully worked)

2 · The cosmological constant problem in one estimate

3 · The de Sitter future

Physics of the portrait

Going further

Before this

Where this leads

1 · From the sky to the number: $\Lambda$ and the density of the vacuum (fully worked)