Cosmological Constraints on Neutrino Masses in a Second-Order CPL Dark Energy Model

TL;DR

The paper examines how the sum of neutrino masses, $M_ ext{tot}$, is constrained in cosmology under three dark energy parameterizations—ΛCDM, CPL, and EXP—across four neutrino hierarchies, using Planck CMB, DESI DR2 BAO, and SN data (PantheonPlus and DESY5). By employing both Bayesian and frequentist analyses with oscillation-based lower limits, it shows that CPL yields tighter 95% upper bounds on $M_ ext{tot}$ than EXP, while ΛCDM often provides the strongest constraints; hierarchies have only mild effects. The results reveal a strong dependence of neutrino-mass bounds on the DE parameterization and dataset choice, with no statistically significant evidence for nonzero neutrino mass within the imposed lower limits. The work highlights the degeneracy between the dark energy EOS and neutrino mass and underscores the role of upcoming high-precision data in breaking these degeneracies and pinning down the true neutrino mass scale.

Abstract

Recent DESI results indicate a strong preference for dynamical dark energy (DE) when baryon acoustic oscillation (BAO) measurements are combined with supernovae (SNe) and cosmic microwave background (CMB) data using the Chevallier-Polarski-Linder (CPL) parameterization. We analyze the exponential (EXP) parameterization, which introduces a second-order correction to CPL. We determine and compare the 95% upper bounds on the sum of neutrino masses for three dark energy (DE) models -- $Λ$CDM, CPL, and EXP -- across four neutrino mass hierarchies (1 massive/2 massless, degenerate, normal, inverted) and multiple dataset combinations (CMB$+$BAO, CMB$+$BAO$+$PantheonPlus, CMB$+$BAO$+$DESY5), employing both Bayesian and frequentist frameworks with physical lower limits from oscillation experiments (0.059 eV and 0.11 eV). Our results show that CPL yields tighter ($\lesssim10$%) bounds compared to EXP. We further confirm earlier findings that neutrino mass constraints are only mildly sensitive to the assumed hierarchy and that the frequentist bounds are tighter than Bayesian ones. Furthermore, the imposed oscillation lower limits, the datasets used and the DE parameterizations play a crucial role in the inferred cosmological neutrino mass bounds. For the datasets, hierarchies, and DE parameterizations considered, we find no statistically significant evidence for nonzero neutrino mass consistent with oscillation lower limits.

Figures (11)

• Figure 1: Bayesian posteriors and frequentist PL for Planck$+$DESI $\Lambda$CDM comparing different hierarchies. Fig. \ref{['1a']}: Bayesian 1D marginalized posterior for different mass hierarchies. Fig. \ref{['1b']}: Frequentist PL for $M_\text{tot}$ extrapolated to unphysical region. The vertical dashed lines represent the minima of the parabolic curves. Fig. \ref{['1c']}: Frequentist PL while considering $M_\mathrm{tot}>0$. Intersection of the curves with the dotted lines represent the 95% upper bounds on $M_\mathrm{tot}$. $M_\mathrm{tot}$ has units of eV.
• Figure 2: Same as Fig. \ref{['fig1']} but for Planck$+$DESI CPL. $M_\mathrm{tot}$ has units of eV.
• Figure 3: Same as Fig. \ref{['fig1']} but for Planck$+$DESI EXP. $M_\mathrm{tot}$ has units of eV.
• Figure 4: Same as Fig. \ref{['fig1']} but for Planck$+$DESI$+$PantheonPlus $\Lambda$CDM. $M_\mathrm{tot}$ has units of eV.
• Figure 5: Same as Fig. \ref{['fig1']} but for Planck$+$DESI$+$PantheonPlus CPL. $M_\mathrm{tot}$ has units of eV.