Are cosmological neutrinos free-streaming?

TL;DR

The paper investigates whether cosmological neutrinos were free-streaming around recombination or could have recoupled to a light majoron-like field at a redshift $z_i$. It uses a 10-parameter cosmological model, including $z_i$ and $N_ u$, and Bayesian inference with data from WMAP-5, LSS, SN Ia, BAO, and HST to bound $z_i$ and translate this into a bound on the neutrino–majoron coupling $g$ (via $ au$). The analysis finds $z_i < 1500$ (95% C.L.) with a linear prior and $z_i < 795$ (95% C.L.) with a log prior, implying neutrinos must have been freely streaming near recombination; the corresponding coupling bound is $g < 1.05\times10^{-7}$ and a lifetime bound of $\tau > 1.0\times10^{10}$ s (for $m=50$ meV). Overall, strongly interacting neutrinos around recombination are disfavored, supporting a relativistic, anisotropic-stress-carrying neutrino background, with future CMB data likely to tighten the limits further.

Abstract

Precision data from cosmology suggest neutrinos stream freely and hence interact very weakly around the epoch of recombination. We study this issue in a simple framework where neutrinos recouple instantaneously and stop streaming freely at a redshift z_i. The latest cosmological data imply z_i < 1500, the exact constraint depending somewhat on the assumed prior on z_i. This bound translates into a limit on the coupling strength between neutrinos and majoron-like particles phi, implying tau > 1 x 10^10 s (m_2/50 meV)^3 for the decay nu_2 -> nu_1+phi.

Figures (2)

• Figure 1: 2D marginal 68% and 95% contours for $z_i$ and $N_\nu$. Top: Linear prior for $z_i$. Bottom: Logarithmic prior, i.e., linear in $\log(1+z_i)$.
• Figure 2: 2D marginal 68% and 95% contours for $z_i$ and $\Omega_{\rm M} h^2$, using the linear prior for $z_i$ and using $N_\nu$ as a fit parameter. Top: Full data set. Bottom: WMAP-5 data only.