Cosmological constraints on a light non-thermal sterile neutrino

TL;DR

This work investigates cosmological constraints on a light, possibly non-thermal sterile neutrino within an extended ΛCDM framework, aiming to reconcile oscillation anomalies with observations of the early and late Universe. By parameterizing the sterile sector with observable quantities $\Delta N_{\rm eff}$, $\omega_s$, and $\langle v_s \rangle$ (and mapping to thermal or Dodelson–Widrow production), the authors derive model-independent bounds on the sterile fraction $f_s$ and velocity dispersion and translate these into mass-temperature (thermal) or mixing-parameter (DW) constraints using a comprehensive data set (WMAP5, small-scale CMB, SDSS LRG, SNLS, VHS Lyman-α). They find that a $2$ eV sterile neutrino can be accommodated only if it is thermally distributed with $T_s/T_ν^{\rm id} < 0.8$ or non-resonantly produced with $ΔN_{\rm eff} < 0.5$, with bounds tightening rapidly for larger masses; for $m_s < 0.9$ eV, a standard thermalized scenario remains compatible. The results offer practical guidance for model-building and emphasize the potential synergy between cosmology and neutrino oscillation experiments in testing light sterile neutrino scenarios.

Abstract

Although the MiniBooNE experiment has severely restricted the possible existence of light sterile neutrinos, a few anomalies persist in oscillation data, and the possibility of extra light species contributing as a subdominant hot (or warm) component is still interesting. In many models, this species would be in thermal equilibrium in the early universe and share the same temperature as active neutrinos, but this is not necessarily the case. In this work, we fit up-to-date cosmological data with an extended LambdaCDM model, including light relics with a mass typically in the range 0.1 -10 eV. We provide, first, some nearly model-independent constraints on their current density and velocity dispersion, and second, some constraints on their mass, assuming that they consist either in early decoupled thermal relics, or in non-resonantly produced sterile neutrinos. Our results can be used for constraining most particle-physics-motivated models with three active neutrinos and one extra light species. For instance, we find that at the 3 sigma confidence level, a sterile neutrino with mass m_s = 2 eV can be accommodated with the data provided that it is thermally distributed with (T_s/T_nu) < 0.8, or non-resonantly produced with (Delta N_eff) < 0.5. The bounds become dramatically tighter when the mass increases. For m_s < 0.9 eV and at the same confidence level, the data is still compatible with a standard thermalized neutrino.

Figures (6)

• Figure 1: (Top) the parameter space ($f_s$,$\langle v_s \rangle$) chosen in our general analysis. The thin bands delimited by red/solid lines show regions of equal $\Delta N_{\rm eff}$ (assuming $\omega_{\rm dm} = 0.11 \pm 0.01$); these bands are fully model-independent. We also show the model-dependent regions of equal mass, delimited by blue/dotted lines for the case of early decoupled thermal relics, and consisting in horizontal green/dashed lines for Dodelson-Widrow sterile neutrinos. (Bottom) same with, in addition, the regions allowed at the 68.3% (1$\sigma$), 95.4% (2$\sigma$) and 99.7% (3$\sigma$) C.L. by our cosmological data set, in a Bayesian analysis with flat priors on $f_s$ and $\log_{10}\langle v_s \rangle$ within the displayed range.
• Figure 2: 1$\sigma$, 2$\sigma$ and 3$\sigma$ contours of the marginalized likelihood for the two parameters ($f_s$, $\langle v_s \rangle$), with different priors than in previous figures. As explained in the text, this plot shows the region where the sterile neutrino is heavy and behaves like warm dark matter, in complement to Figure 1, which is based on a different range/prior for $\langle v_s \rangle$ adapted to the case of a light, hot sterile neutrino.
• Figure 3: (Top) the parameter space ($m_s$,$T_s/T_\nu^{\rm id}$) used in the particular case of early decoupled thermal relics of temperature $T_s$ (with $T_\nu^{\rm id} \equiv (4/11)^{1/3} T_{\gamma}$). The thin bands delimited by blue/dot-dashed lines show regions of equal $f_s$ (assuming $\omega_{\rm dm} = 0.11 \pm 0.01$); the magenta/dotted lines correspond to fixed values of the velocity dispersion today; horizontal red/solid lines to fixed $\Delta N_{\rm eff}$. (Bottom) same with, in addition, the regions allowed at the 68.3% (1$\sigma$), 95.4% (2$\sigma$) and 99.7% (3$\sigma$) C.L. by our cosmological data set, in a Bayesian analysis with flat priors on $\log_{10}(m_s)$ and $T_s/T_\nu^{\rm id}$ within the displayed range.
• Figure 4: (Top) the parameter space ($m_s$,$\chi$) used in the particular case of DW relics. The thin bands delimited by blue/dot-dashed lines show regions of equal $f_s$ (assuming $\omega_{\rm dm} = 0.11 \pm 0.01$); the magenta/dotted lines correspond to fixed values of the velocity dispersion today; horizontal red/solid lines to fixed $\Delta N_{\rm eff}$. (Bottom) same with, in addition, the regions allowed at the 68.3% (1$\sigma$), 95.4% (2$\sigma$) and 99.7% (3$\sigma$) C.L. by our cosmological data set, in a Bayesian analysis with flat priors on $\log_{10}(m_s)$ and $\log_{10}(\chi)$ within the displayed range.
• Figure 5: (Top) the parameter space ($\Delta N_{\rm eff}$,$m_s$) used for comparison with Cirelli & Strumia in the particular case of DW relics. The thin bands delimited by blue/dot-dashed lines show regions of equal $f_s$ (assuming $\omega_{\rm dm} = 0.11 \pm 0.01$); the magenta/dotted lines correspond to fixed values of the velocity dispersion today; horizontal red/solid lines to fixed $\Delta N_{\rm eff}$. (Bottom) same with, in addition, the regions allowed at the 90%, 99% and 99.9% C.L. by our cosmological data set, in a Bayesian analysis with flat priors on $\log_{10}(\Delta N_{\rm eff})$ and $\log_{10}(m_s)$ within the displayed range.