Interacting dark energy after DESI DR2: a challenge for $Λ$CDM paradigm?

TL;DR

This study investigates an interacting dark energy (IDE) scenario in which the dark matter density evolves as ρ_dm ∝ a^{-3+δ0} with a constant coupling δ0, and the interaction is described by Q = -δ0 H ρ_dm, while dark energy is treated as vacuum (w_de = -1). Using Planck 2018 CMB, DESI DR2 BAO, Cosmic Chronometers, and multiple SNIa compilations (PantheonPlus, Union3, DES-Y5), the authors perform a seven-parameter MCMC analysis with CAMB/Cobaya to constrain δ0 and standard ΛCDM parameters. They find a mild preference for nonzero δ0 (≳1σ) but no strong evidence, and a mixed model comparison: Δχ^2_min generally favors IDE for several datasets, whereas AIC and Bayesian evidence often favor ΛCDM. The results show that the inferred direction of energy transfer between dark sectors depends on dataset (DM→DE or DE→DM) and that including CC shifts constraints on H0 and Ω_m0 in a manner tied to δ0’s sign, indicating nuanced implications for the standard cosmological model. Overall, the paper highlights that DESI DR2 can mildly hint at IDE, motivating further exploration with time-varying δ(a) or alternative interaction forms to clarify the dark-sector dynamics.

Abstract

We investigate the scenario of interacting dark energy through a detailed confrontation with various observational datasets. We quantify the interaction in a general way, through the deviation from the standard scaling of the dark matter energy density. We use the cosmic microwave background (CMB) data from Planck 2018, data from Baryon Acoustic Oscillations (BAO) from the recently released DESI DR2, observational Hubble Data from Cosmic Chronometers (CC), and finally various Supernova Type Ia (SNIa) datasets (PantheonPlus, Union3 and DESY5). For the basic and simplest interacting model we find a mild preference of the interaction at slightly more than $1σ$ however still within $2σ$, and thus no strong evidence of interaction is found. However, comparison with $Λ$CDM scenario through $Δχ^2_{\rm min}$, AIC and Bayesian analysis, reveals a mixed picture, namely according to $Δχ^2_{\rm min}$ the interaction is mildly favored by most of the datasets, while the remaining statistical measures are inclined toward $Λ$CDM.

Figures (6)

• Figure 1: One-dimensional marginalized posterior distributions and two-dimensional joint contours of the model parameters considering CMB, CMB+DESI, CMB+DESI+SNIa, where SNIa is either PantheonPlus or Union3 or DESY5.
• Figure 2: One-dimensional marginalized posterior distributions and two-dimensional joint contours of the model parameters considering CMB+CC+DESI and CMB+CC+DESI+SNIa where SNIa is either PantheonPlus or Union3 or DESY5.
• Figure 3: Whisker graphs of $\delta_0$ (upper plot), $H_0$ (middle plot) and $\Omega_{m0}$ (lower plot), with their 68% CL constraints, for various datasets used in our analysis. The horizontal bar in the middle panel (red) corresponds to $H_0 = 67.36\, \pm \, 0.54$ km/s/Mpc at 68% CL Planck:2018vyg and the horizontal bar in the lower plot (violet) corresponds to $\Omega_{m0} = 0.3153\, \pm \, 0.0073$ at 68% CL Planck:2018vyg.
• Figure 4: Redshift evolution of $\Omega_m/\Omega_r$ where $\Omega_m = \Omega_{\rm dm} +\Omega_b$, for different values of the interaction parameter $\delta_0$. The horizontal solid line corresponds to the matter radiation equality (i.e. $\Omega_m = \Omega_r$). In order to draw the curves we fix the other parameters from the constraints obtained from CMB+CC+DESI+PantheonPlus.
• Figure 5: The CMB TT power spectrum (upper graph) and the matter power spectrum (lower graph) for different values of $\delta_0$. In order to draw the curves, we fix the other parameters from the constraints obtained from CMB+CC+DESI+PantheonPlus.