Dark Energy Survey finds a ~3σ hint that dark energy may change over time, challenging the cosmological constant
The Dark Energy Survey (DES) has reported a notable tension with the standard cosmological model known as ΛCDM. By combining two geometric probes that measure the Universe’s expansion — type Ia supernovae (SNe) and baryon acoustic oscillations (BAO) — with early-Universe measurements from the Planck satellite of the cosmic microwave background (CMB), DES finds a preference for a model in which the density of dark energy changes with time. Interpreted in a common two‑parameter model for evolving dark energy (the “w0–wa” model), the DES combination gives w0 = −0.673 +0.098 −0.097 and wa = −1.37 +0.51 −0.50, a result the collaboration reports as about a 3.2σ deviation from a constant dark energy (the cosmological constant), where w0 = −1 and wa = 0.
DES used two main late‑time probes. Type Ia supernovae act like standardized candles: their measured brightness tells us their distance and so the expansion history. BAO are a preferred spacing imprinted in the distribution of galaxies from sound waves in the early Universe; measuring the angle of that scale on the sky gives a distance at a given redshift. DES surveyed about 5,000 square degrees of sky from 2013 to 2019 with the Dark Energy Camera (DECam). Its BAO measurement at redshift z ≈ 0.85 reached 2.1% precision and was found to be about 4.3% below the prediction of the Planck ΛCDM model (roughly a 2σ difference). DES’s final supernova analysis by itself showed about a 2σ deviation that also pulls the fit away from the cosmological constant.
To allow dark energy to vary with time, researchers use the w0–wa parametrization, where w(a) = w0 + wa(1 − a) and a is the cosmic scale factor. This model reduces to the cosmological constant when w0 = −1 and wa = 0. DES finds that combining its BAO and supernova measurements with Planck CMB data makes the four datasets consistent within the w0–wa model and that the fit is substantially better than for ΛCDM: a change in chi‑squared of Δχ2 = 11.6 for just two extra parameters. Other groups and surveys give related hints. The Dark Energy Spectroscopic Instrument (DESI), when combined with Planck and different supernova compilations, reports a preference for evolving dark energy in the range 2.5σ to 3.9σ depending on the supernova sample; a combined DES and DESI treatment could reach still higher nominal significance if small correlations between their measurements are neglected.