Dark Radiation from Neutrino Mixing after Big Bang Nucleosynthesis

TL;DR

This paper explores how light dark fermions mixing with neutrinos can equilibrate in the early universe, affecting dark radiation and potentially explaining cosmological tensions, with observable signatures in the CMB and large-scale structure.

Contribution

It demonstrates that dark sector equilibration can occur at very small mixing angles and describes the resulting cosmological implications, including potential signals detectable by future experiments.

Findings

Dark fermions can equilibrate with neutrinos at small mixing angles.

Dark radiation can transition from non-interacting to interacting, creating observable 'steps'.

Minimal models predict detectable CMB signatures above upcoming experiment sensitivities.

Abstract

A light ($m_{\nu d} \lesssim $ MeV) dark fermion mixing with the Standard Model neutrinos can naturally equilibrate with the neutrinos via oscillations and scattering. In the presence of dark sector interactions, production of dark fermions is generically suppressed above BBN, but then enhanced at later times. Over much of the parameter space, we find that the dark sector equilibrates, even for mixing angles $\theta_0$ as small as $10^{-13}$, and equilibration occurs at $T_{\rm equil} \simeq m_{\nu d} \left(\theta_0^2 M_{Pl}/ m_{\nu d} \right)^{1/5} $ which is naturally at most a few orders of magnitude above the dark fermion mass. The implications of this are twofold: one, that light states are often only constrained by the CMB and LSS without leaving an imprint on BBN, and two, that sectors which equilibrate before recombination will typically have a mass threshold before recombination, as well. This can result in dark radiation abruptly transitioning from non-interacting to interacting, or vice-versa, a ''step'' in the amount of dark radiation, and dark matter with similar transitions in its interactions, all of which can leave important signals in the CMB and LSS, and may be relevant for cosmological tensions in observables such as $H_0$ or $S_8$. Minimal models leave an unambiguous imprint on the CMB above the sensitivity of upcoming experiments.