Limits on Neutrino-Neutrino Scattering in the Early Universe

TL;DR

This paper investigates how nonstandard neutrino self-interactions modify early-Universe evolution and the CMB. By modeling a four-fermion neutrino self-interaction with an effective coupling $G_{eff}$, it tracks transitions between tightly coupled and free-streaming neutrinos and their imprint on the CMB power spectra. The analysis reveals a multimodal posterior: a Standard Cosmological Mode with strong bounds on $G_{eff}$ and an Interacting Neutrino Mode with $G_{eff} \sim 10^{-2}$ MeV$^{-2}$, offering an alternate cosmology where neutrinos stay coupled until around $z \sim 10^4$; the viability of this mode depends on priors and may be constrained by BBN considerations. Overall, the results show that cosmology can accommodate neutrino self-interactions far stronger than the Fermi interaction, highlighting the potential of CMB data to reveal or constrain new neutrino physics; upcoming polarization and lensing measurements are expected to sharpen these limits.

Abstract

In the standard model neutrinos are assumed to have streamed across the Universe since they last scattered at the weak decoupling epoch when the temperature of the standard-model plasma was ~MeV. The shear stress of free-streaming neutrinos imprints itself gravitationally on the Cosmic Microwave Background (CMB) and makes the CMB a sensitive probe of neutrino scattering. Yet, the presence of nonstandard physics in the neutrino sector may alter this standard chronology and delay neutrino free-streaming until a much later epoch. We use observations of the CMB to constrain the strength of neutrino self-interactions G_eff and put limits on new physics in the neutrino sector from the early Universe. Recent measurements of the CMB at large multipoles made by the Planck satellite and high-l experiments are critical for probing this physics. Within the context of conventional LambdaCDM parameters cosmological data are compatible with G_eff < 1/(56 MeV)^2 and neutrino free-streaming might be delayed until their temperature has cooled to as low as ~25 eV. Intriguingly, we also find an alternative cosmology compatible with cosmological data in which neutrinos scatter off each other until z~10^4 with a preferred interaction strength in a narrow region around $G_{\rm eff} \simeq 1/({\rm 10 \, MeV})^{2} \simeq 8.6\times10^8 G_{\rm F}$, where $G_{\rm F}$ is the Fermi constant. This distinct self-interacting neutrino cosmology is characterized by somewhat lower values of both the scalar spectral index and the amplitude of primordial fluctuations. While we phrase our discussion here in terms of a specific scenario in which a late onset of neutrino free-streaming could occur, our constraints on the neutrino visibility function are very general.

Figures (4)

• Figure 1: Visibility function for self-interacting neutrinos for different values of the effective coupling constant $G_{\rm eff}$ as a function of redshift. Here, we assume three neutrinos species. Note that we have divided out each case by their maximum visibility in order to show them on the same scale.
• Figure 2: Snapshot of neutrino and photon density fluctuations in configuration space at a fixed redshift. The black dot-dashed line shows the standard free-streaming neutrino fluctuation while the green dashed line displays the corresponding photon density fluctuation. The solid blue and red dotted lines show the density fluctuation of self-interacting neutrinos and the corresponding photon perturbation, respectively. These two lines lie on top of one another since both neutrinos and photons behave as tightly-coupled fluids at the epoch shown here. The difference between the green dashed and the red dotted lines readily illustrates the phase shift and amplitude suppression of the photon fluctuation associated with free-streaming neutrinos. Here we have adopted a Planck cosmology planckXVI.
• Figure 3: CMB temperature power spectra for different values of $G_{\rm eff}$. The upper panel shows the temperature spectra themselves together with recent measurements, while the lower panel displays the relative difference between the interacting neutrino models and the best-fit Planck $\Lambda$CDM cosmology planckXVI with three neutrinos. The dashed red lines illustrate the interacting neutrino cosmology given in Table \ref{['table1']}.
• Figure 4: Top panel: Marginalized posterior distribution of $\log_{10}({\rm G_{\rm eff}MeV}^2)$ for different combinations of datasets and priors. Bottom panel: 2D marginalized constraints in the $n_{\rm s}$ and $\log_{10}({\rm G_{\rm eff}MeV}^2)$ plane.