arXiv:2606.21518·v2·Cosmology and Nongalactic Astrophysics
Are Cosmological Data Excluding Sterile Neutrinos or Only the Fully Thermalized Limit?
Artur Ladeira🇧🇷 · Rafael C. Nunes🇧🇷 · Eleonora Di Valentino🇬🇧 · Stefano Gariazzo🇪🇸
Abstract
We present a cosmological reassessment of light sterile-neutrino scenarios, examining whether current observations exclude sterile neutrinos as a class or primarily constrain the fully thermalized case. We consider three distinct realizations: (i) a fully thermalized sterile species (FTS), (ii) a different-temperature sterile-neutrino thermal relic (DTS) relative to the active neutrino background and (iii) a Dodelson--Widrow-like (DW) sterile neutrino with reduced phase-space normalization. Constraints are derived within both CDM and the CPL dynamical dark-energy framework using combinations of Planck-CMB data, DESI DR2 BAO measurements, and the PantheonPlus and Union3 Type Ia supernova samples. For baseline data combinations without a local prior, the FTS scenario is strongly disfavored in both cosmological models. Adding the local prior allows CDM+FTS to accommodate the high local value and become statistically competitive with standard CDM once SNIa data are included, although the sterile-neutrino mass remains consistent with zero. By contrast, partially populated sterile-neutrino scenarios remain viable: the DW realization is broadly compatible with current observations, while the DTS scenario yields the least cosmological pressure among the cases considered. Overall, cosmological data mainly require a strongly suppressed effective sterile abundance, leading to tight constraints on \textbf{} while allowing substantially weaker bounds on the physical sterile mass. We conclude that current observations do not generically exclude sterile neutrinos, but rather place strong pressure on fully thermalized or highly populated scenarios, highlighting the importance of production history and phase-space distribution when interpreting cosmological constraints.
Comments: 22 pages, 6 figures, and 6 tables. Match the version published in Physical Review D