Glossary · Cosmology
Dark Energy
Dark energy is the dominant component of the Universe (~68%) driving its accelerating expansion over the past ~6 billion years. Its nature remains one of physics' deepest mysteries.
Categorie
Cosmologie · Composante accélérant l'expansion
Proportion Univers
≈ 68 % du contenu énergétique total (Planck 2018)
Parametre Etat
w ≈ -1 (proche d'une constante cosmologique Λ)
Densite Energie
≈ 6 × 10⁻²⁷ kg/m³ équivalent
Decouverte
1998 — Perlmutter, Schmidt, Riess (supernovae Ia)
Prix Nobel
2011 (Perlmutter + Schmidt + Riess)
Full definition
For eighty years, the working assumption was that gravity had to slow cosmic expansion — since all matter attracts. The open question was simply: is the deceleration strong enough to halt expansion and trigger a Big Crunch, or will the Universe expand forever while slowing? In 1998, two competing teams measuring distant type Ia supernovae to answer that question discovered something else entirely: expansion is accelerating rather than slowing. An unknown energy component must therefore push space to stretch faster and faster.
This component — named dark energy — has unsettling properties. It is uniformly distributed (unlike dark matter, which clumps into halos), nearly constant in time (its density doesn't dilute as the Universe expands, or barely), and exerts negative pressure. This last trait is counter-intuitive: in general relativity, negative pressure produces repulsive gravity. That's what pushes cosmic distances to grow ever faster.
The simplest hypothesis to describe it is the cosmological constant Λ (lambda), introduced by Einstein in 1917 to obtain a static Universe, then dismissed by himself as 'the biggest blunder of my life' when Hubble proved expansion in 1929. It returned in spectacular fashion in 1998. In the standard ΛCDM model (Lambda Cold Dark Matter), the Universe consists of ~68 % dark energy, ~27 % dark matter, ~5 % ordinary matter. Planck measurements (2018) pin the equation of state w = p/(ρc²) to -1 ± 0.03 — consistent with a pure cosmological constant, but leaving the door open to dynamical models (quintessence).
Numbers and scale
Dark energy is almost nothing, and yet it dominates.
Its energy density is about ρ_Λ ≈ 6 × 10⁻²⁷ kg/m³ (mass-energy equivalent via E = mc²). On the scale of a room in your house (say 30 m³), that's the energetic equivalent of one billionth of a billionth of a gram — negligible. But multiplied by the volume of the observable Universe (~ 4 × 10⁸⁰ m³), it becomes the largest energy reservoir in the cosmos.
The dark-energy equation of state reads:
p = w · ρ · c²
with p the pressure, ρ the density, and w a dimensionless number. For a pure cosmological constant, w = -1 exactly. Combining Planck + Ia supernovae + baryon acoustic oscillations (BAO) yields w = -1.03 ± 0.03 (2024), extremely close to the expected value.
Expansion acceleration is calibrated by the deceleration parameter q₀ ≈ -0.55 (negative = accelerating). Historically, the Universe first slowed down (matter-dominated era), then dark energy took over around z ≈ 0.7, roughly 6 billion years after the Big Bang.
The cosmological constant problem. A naive calculation of the quantum vacuum energy predicts a value ~10¹²⁰ times too large. This is 'the worst theoretical prediction in the history of physics', in Steven Weinberg's words. No accepted solution to date.
Theoretical hypotheses
Four major families of explanation are on the table.
Cosmological constant Λ. The simplest hypothesis: quantum vacuum energy has a non-zero density, constant in space and time. Consistent with all current observations at available precision. Problem: the value predicted by quantum field theory is absurdly too large. Why did nature fine-tune it to 10⁻¹²⁰? The anthropic principle answers, but not everyone is satisfied.
Quintessence. A dynamical scalar field, analogous to the Higgs field but slowly rolling on its potential, making dark energy evolve over cosmic time. Predicts w slightly different from -1 and potentially time-dependent. DESI data (2024-2026) begin to hint at a slight evolution, but this is debated.
Modified gravity. Rather than adding a new component, modify general relativity at large scales. f(R) models, DGP brane-worlds, cosmological MOND. Struggle to simultaneously fit supernovae, CMB and BAO.
Scale illusion. A minority view: acceleration wouldn't exist, it would be an artefact of averaging over an inhomogeneous Universe (LTB models, backreaction). Poorly supported by data.
More exotic models — phantom energy (w < -1), Big Rip, Steinhardt's cyclic universe, braneworlds — remain speculative but have the merit of being testable.
How do we measure it?
Three observational pillars, independent and convergent.
Type Ia supernovae (standard candle). These thermonuclear explosions of white dwarfs at the Chandrasekhar limit nearly all reach the same intrinsic luminosity. Comparing apparent magnitude with redshift reconstructs the expansion history. The High-Z Supernova Search Team (Brian Schmidt, Adam Riess) and the Supernova Cosmology Project (Saul Perlmutter) independently announced acceleration in 1998 — 2011 Nobel Prize in Physics to the three researchers.
Baryon acoustic oscillations (BAO). In the early Universe, pressure waves propagating in the photon-baryon plasma left a ~150 Mpc imprint — a cosmic 'standard ruler'. Measuring it at various redshifts in galaxy distributions (SDSS, BOSS, eBOSS, DESI) maps expansion. DESI (Dark Energy Spectroscopic Instrument, Arizona) released its first results in April 2024 on ~6 million galaxies, hinting at a mild tension with pure Λ.
Cosmic Microwave Background (CMB). The Planck satellite (2009-2013) measures CMB anisotropies with ultimate precision. The position of the acoustic peaks fixes spatial geometry and, combined with other probes, tightly constrains dark energy.
Current and upcoming missions. Euclid (ESA, launched July 2023, first science images 2024) maps 3D structure over a third of the sky out to z = 2. Nancy Grace Roman Space Telescope (NASA, 2027 planned) will take over on supernovae and weak lensing. Vera C. Rubin Observatory / LSST (Chile, first data 2025) will multiply detected type Ia supernovae by a factor of a hundred.
Not to be confused with
Dark energy is routinely confused with several cousins.
Dark matter. The textbook confusion. Dark matter = invisible mass that clumps and slows expansion by attraction. Dark energy = uniform component that accelerates expansion by negative pressure. They share the adjective 'dark' by symmetry of ignorance (we understand neither), but play completely different cosmological roles.
Zero-point / quantum vacuum energy. The quantum vacuum carries real, measurable energy (Casimir effect in the lab). It is the natural candidate for the cosmological constant — except its calculated value is 10¹²⁰ times too large. So vacuum energy, perhaps, but we don't know how to match the number.
Dark fluid. A speculative unified model attempting to describe dark matter and dark energy as two facets of a single substance. Not widely supported by data. Don't assume they are the same thing.
Cosmic inflation. A phase of extreme exponential expansion in the first 10⁻³² seconds after the Big Bang, driven by a scalar field (inflaton) since vanished. Today's dark energy is infinitely gentler and has acted for only ~6 billion years. The two phenomena may be theoretical cousins (same negative-pressure mechanism) but don't coexist.
Cosmic pressure force. Dark energy is not a 'force' in the sense of the four fundamental interactions — it's an energy component whose gravity, via general relativity, produces a repulsive effect at large scales.
Frequently asked
Will dark energy tear the Universe apart?
If it is a pure cosmological constant (w = -1), no: expansion will keep accelerating exponentially, distant galaxies will vanish beyond our horizon, but the Milky Way and its gravitationally bound neighbours will remain tightly held together. If dark energy is instead phantom energy (w < -1), the Big Rip scenario becomes possible: in tens of billions of years, expansion would tear apart galaxies, then stars, then atoms. Current data strongly favour w ≈ -1.
Was Einstein right about the cosmological constant?
Paradoxically yes, even though he rejected it. Einstein introduced it in 1917 to maintain a static Universe, then abandoned it after Hubble discovered expansion in 1929. When Perlmutter, Schmidt and Riess discovered cosmic acceleration in 1998, the cosmological constant came back as the best explanation. Einstein was right about the mathematics — not about the static Universe he wanted. The Λ in ΛCDM stands as a tribute.
When was dark energy discovered?
January 8, 1998 at an AAS meeting in Washington, followed by Riess et al. (September 1998) and Perlmutter et al. (1999) papers. Two rival teams were studying distant Ia supernovae to measure cosmic deceleration. They found acceleration instead, with about 70% of the energy content in a negative-pressure component. The 2011 Nobel Prize in Physics was awarded jointly to Saul Perlmutter, Brian Schmidt and Adam Riess for this discovery.
Can we use dark energy as a power source?
No, totally impossible. Its local density is minuscule (~10⁻²⁷ kg/m³ equivalent, about 6 protons per m³). It is uniformly distributed, so no gradient to exploit like a waterfall or a radiation source. And we have no way to couple ordinary matter to this component other than through gravity. It is a cosmological energy — it shapes the Universe at the largest scales without providing any usable power locally.