Pseudoscalar - sterile neutrino interactions: reconciling the cosmos with neutrino oscillations

TL;DR

This work addresses the conflict between short-baseline hints for an eV-scale sterile neutrino and cosmological constraints by introducing a hidden-sector pseudoscalar mediator that couples exclusively to the sterile state. The model suppresses early thermalisation and drives a late-time, collisional neutrino-pseudoscalar fluid, effectively increasing $N_{\textrm{eff}}$ without erasing active neutrino free-streaming and yielding a higher $H_0$ in agreement with local measurements. Cosmological data, including Planck and BAO, can be reconciled with SBL results within this pseudoscalar framework, with joint analyses showing a preference for $\Delta m^2_{41} \gtrsim 1\ \mathrm{eV}^2$ and $m_s$ values compatible with the SBL range. The approach thus provides a natural, testable way to bridge laboratory neutrino anomalies and cosmological observations, informing future short-baseline experiments and cosmological surveys.

Abstract

The Short BaseLine (SBL) neutrino oscillation anomalies hint at the presence of a sterile neutrino with a mass of around 1 eV. However, such a neutrino is incompatible with cosmological data, in particular observations of the Cosmic Microwave Background (CMB) anisotropies. However, this conclusion can change by invoking new physics. One possibility is to introduce a secret interaction in the sterile neutrino sector mediated by a light pseudoscalar. In this pseudoscalar model, CMB data prefer a sterile neutrino mass that is fully compatible with the mass ranges suggested by SBL anomalies. In addition, this model predicts a value of the Hubble parameter which is completely consistent with local measurements.

Figures (7)

• Figure 1: Regions in the planes of the effective amplitudes $\sin^22\vartheta_{e\mu}$, $\sin^22\vartheta_{ee}$ and $\sin^22\vartheta_{\mu\mu}$ versus $\Delta{m}^2_{41}$ which are allowed by the Bayesian global fit of short-baseline (SBL) neutrino oscillation data. Also shown are the $3\sigma$ bounds obtained from the separate fits of appearance (APP; the regions inside the two blue contours are allowed) and disappearance (DIS; the regions on the left of the red lines are allowed) data. $1\sigma$, $2\sigma$ and $3\sigma$ correspond, respectively, to 68.27%, 95.45% and 99.73% posterior probability.
• Figure 2: Marginalised constraints in the $m_s$--$\Delta N_{\textrm{eff}}$ plane, obtained from the analyses of the cosmological data in the context of the $\Lambda$CDM $+\,N_{\textrm{eff}} \,+\,m_s$ model, at 1$\sigma$ and 2$\sigma$ confidence level. The excluded regions are on the right of each line. The points obtained with the TT dataset are colour-coded by their $H_0$ value.
• Figure 3: Comparison of the one-dimensional marginalised posterior distribution of the $H_0$ (left panel) and $m_s$ (right panel) parameters as obtained from the analyses of the TT data in the $\Lambda$CDM+$1\nu_s$ model and in the pseudoscalar model. In the left panel, we also report the local measure $H_0=73.00\pm1.75\,\,\text{Km s}^{-1}\text{ Mpc}^{-1}$Riess:2016jrr and the constraints obtained in the $\Lambda$CDM $+\,N_{\textrm{eff}} \,+\,m_s$ model (see table \ref{['tab:lsnBounds']}). In the panel on the right, the posteriors are in logarithmic scale and are normalized such that they integrate to one. We also show the posterior for $m_s$ obtained in the SBL analysis described in section \ref{['sec:sbl']}. It is visible the secondary peak at $m_s\simeq2.4$, corresponding to the regions at $\Delta{m}^2_{41}\simeq6$ in figure \ref{['fig:sbl']}.
• Figure 4: Marginalised $1$ and $2\,\sigma$ contours in the plane $m_s$--$N_{\rm fluid}$ for the pseudoscalar model. The points inside the contours are coloured according to the value of $H_0$ obtained by fitting only TT data.
• Figure 5: Marginalised $1$ and $2\,\sigma$ contours in the plane $\sigma_8$--$\Omega_m$ of the $\Lambda$CDM model (left panel) and of the pseudoscalar model (right panel). The points inside the contours are coloured according to the value of $H_0$ (see the scale on the right hand side of each plot). The red and gray shaded areas show the constraints from Planck CMB lensing Ade:2015zua and from CFHTLenS weak lensing data Heymans:2013fya, respectively.