The Dark Energy Survey: Cosmology Results With ~1500 New High-redshift Type Ia Supernovae Using The Full 5-year Dataset

TL;DR

This study presents cosmological constraints from the DES five-year SN dataset, leveraging photometric SN Ia classification, BEAMS treatment of contamination, SALT3 light-curve fitting, and BBC bias corrections to build a robust Hubble diagram. By combining 1635 DES SNe (plus 194 low-z SNe) with Planck CMB and BAO/3×2pt data, the authors tighten constraints on ΩM, ΩK, and the dark-energy equation of state across flat ΛCDM, ΛCDM with curvature, and evolving-EoS models, finding results consistent with a cosmological constant to ~2σ and no strong evidence for new physics. The analysis demonstrates that systematic uncertainties are subdominant to statistics in this dataset and argues that photometric classification can unlock much larger SN samples for future cosmology, provided calibration, low-z anchors, and dust/intrinsic-scatter models are well controlled. Overall, the work confirms cosmic acceleration, yields competitive late-universe constraints, and outlines a path forward for next-generation SN surveys and joint analyses with other DES probes.

Abstract

We present cosmological constraints from the sample of Type Ia supernovae (SN Ia) discovered during the full five years of the Dark Energy Survey (DES) Supernova Program. In contrast to most previous cosmological samples, in which SN are classified based on their spectra, we classify the DES SNe using a machine learning algorithm applied to their light curves in four photometric bands. Spectroscopic redshifts are acquired from a dedicated follow-up survey of the host galaxies. After accounting for the likelihood of each SN being a SN Ia, we find 1635 DES SNe in the redshift range $0.10<z<1.13$ that pass quality selection criteria sufficient to constrain cosmological parameters. This quintuples the number of high-quality $z>0.5$ SNe compared to the previous leading compilation of Pantheon+, and results in the tightest cosmological constraints achieved by any SN data set to date. To derive cosmological constraints we combine the DES supernova data with a high-quality external low-redshift sample consisting of 194 SNe Ia spanning $0.025<z<0.10$. Using SN data alone and including systematic uncertainties we find $Ω_{\rm M}=0.352\pm 0.017$ in flat $Λ$CDM. Supernova data alone now require acceleration ($q_0<0$ in $Λ$CDM) with over $5σ$ confidence. We find $(Ω_{\rm M},w)=(0.264^{+0.074}_{-0.096},-0.80^{+0.14}_{-0.16})$ in flat $w$CDM. For flat $w_0w_a$CDM, we find $(Ω_{\rm M},w_0,w_a)=(0.495^{+0.033}_{-0.043},-0.36^{+0.36}_{-0.30},-8.8^{+3.7}_{-4.5})$. Including Planck CMB data, SDSS BAO data, and DES $3\times2$-point data gives $(Ω_{\rm M},w)=(0.321\pm0.007,-0.941\pm0.026)$. In all cases dark energy is consistent with a cosmological constant to within $\sim2σ$. In our analysis, systematic errors on cosmological parameters are subdominant compared to statistical errors; paving the way for future photometrically classified supernova analyses.

Figures (12)

• Figure 1: Overview of supporting papers for DES-SN5YR cosmological results.
• Figure 2: All DES light curves, showing observed magnitudes in $g$, $r$, $i$, and $z$ bands (left to right respectively) normalized by the maximum brightness of each light curve, and with the time-axis de-redshifted to the rest-frame. Each light curve has been arbitrarily offset by their redshift, with higher-redshift objects higher on the plot (as labeled on vertical axis). Lines show best-fit SALT3 light-curve fits. The $g$-band and $r$-band light curves are not used above $z\sim0.4$ and $z\sim0.85$ respectively because that corresponds to the redshifts at which the lower-wavelength limit of the SALT3 model ($3500\AA$ in the rest frame) passes out of their observed wavelength ranges.
• Figure 3: Histogram showing the redshift distribution of the DES-SN5YR sample, with new DES SNe in blue and our low-$z$ sample in red. For comparison the distribution of redshifts in the existing Pantheon+ sample is shown in grey brout22_pantheon, which also includes the DES SNe from the DES-SN3YR analysis (blue dashed line). The five-year DES sample contains $\sim4\times$ more supernovae above $z\sim0.4$ than the Pantheon+ compilation.
• Figure 4: Hubble diagram of DES-SN5YR. We show both the single SN events and the redshift-binned SN distance moduli. Redshift bins are adjusted so that each bin has the same number of SNe ($\sim50$). The 1635 new DES supernovae are in blue, and in the upper panel they are shaded by their probability of being a Type Ia; most outliers are likely contaminants (pale blue). The inset shows the number of SNe as a function of redshift (same $z$-range as the main plot). The lower panel shows the difference between the data and the best fit Flat-$w$CDM model from DES-SN5YR alone (third result in Table \ref{['tab:cosmo_results']}), and overplots three other best fit cosmological models --- Flat-$\Lambda$CDM model from DES-SN5YR alone (magenta line, first result in Table \ref{['tab:cosmo_results']}), Flat-$w_0w_a$CDM model from DES-SN5YR alone (green line, fourth result in Table \ref{['tab:cosmo_results']}), and Planck 2020 Flat-$\Lambda$CDM model without SN data (dashed line, $\Omega_{\rm M}^{\rm Planck}=$$0.317\pm0.008$).
• Figure 5: Constraints on matter density in the Flat-$\Lambda$CDM model from DES-SN5YR only (cyan), DES-SN5YR combined with CMB constraints from planck18_VI (blue), and DES-SN5YR combined with BAO+3$\times$2pt (orange), and all probes combined (DES-SN5YR+BAO+3$\times$2pt and CMB constraints, red). CMB constraints only and BAO+3$\times$2pt constraints alone are also shown for comparison (dashed and dotted-dashed respectively).