Restoring cosmological concordance with early dark energy and massive neutrinos?

The early dark energy (EDE) solution to the Hubble tension comes at the cost of an increased clustering amplitude that has been argued to worsen the fit to galaxy clustering data. We explore whether freeing the total neutrino mass M-nu, which can suppress small-scale structure growth, improves EDE's fit to galaxy clustering. Using Planck Cosmic Microwave Background and BOSS galaxy clustering data, a Bayesian analysis shows that freeing M-nu does not appreciably increase the inferred EDE fraction f(EDE): we find the 95 per cent C.L. upper limits f(EDE) < 0.092 and M-nu < 0.15eV. Similarly, in a frequentist profile likelihood setting (where our results support previous findings that prior volume effects are important), we find that the baseline EDE model (with M-nu = 0.06 eV) provides the overall best fit. For instance, compared to baseline EDE, a model with M-nu = 0.24 eV maintains the same H-0(km/s/Mpc) = (70.08, 70.11, respectively) whilst decreasing S-8 = (0.837, 0.826) to the lambda CDM level, but worsening the fit significantly by delta chi(2) = 7.5. For the datasets used, these results are driven not by the clustering amplitude, but by background modifications to the late-time expansion rate due to massive neutrinos, which worsen the fit to measurements of the BAO scale.

Automated summary

The study investigates whether freeing the total neutrino mass can improve the fit of the early dark energy (EDE) model to galaxy clustering data. Bayesian and frequentist profile likelihood analyses using Planck CMB and BOSS galaxy clustering data show that freeing the total neutrino mass does not significantly improve the fit to galaxy clustering. The worsening fit is mainly caused by background modifications to the late-time expansion rate due to massive neutrinos which affect the measurements of the BAO scale.