Lyα forest bounds on sterile neutrino production via neutrino self-interactions

TL;DR

KeV-scale sterile neutrinos produced via active-neutrino self-interactions are constrained using a self-consistent calculation of the linear matter power spectrum that accounts for both sterile free streaming and delayed active-neutrino decoupling. A neural-network emulator maps $(m_4, m_\phi, \lambda_\phi, \theta)$ to the required mixing angle under a fixed relic density, and the resulting $P(k)$ is confronted with Ly$\alpha$ forest and CMB data using EFT full-shape and PRIYA likelihoods. The analysis yields the strongest current observational bounds on the production mechanism, with the Ly$\alpha$ constraints dominating relic-density and BBN limits and approaching laboratory bounds; together with X-ray and MW-satellite observations, a large portion of parameter space is excluded. The work establishes a robust framework for testing non-thermal DM production scenarios with small-scale structure data and highlights prospects for future Ly$\alpha$ and DESI-era analyses.

Abstract

Sterile neutrinos in the keV mass range have long been considered a well-motivated dark matter (DM) candidate. In this work, we explore a sterile neutrino production mechanism through active neutrino self-interactions in the early universe, assuming that they constitute the full DM abundance. We implement a self-consistent treatment of the sterile-neutrino free streaming and the active-neutrino self-interactions on structure formation, which yield a unique scale-dependent modification to the linear matter power spectrum. We then set bounds on this scenario using a combination of the cosmic microwave background and Ly$α$ forest constraints. Specifically, we utilize the two recent likelihoods derived from eBOSS data: (i) an effective field theory (EFT) based full-shape likelihood and (ii) a compressed likelihood obtained from the PRIYA-simulation emulator. We produce some of the most stringent observational constraints to date on sterile neutrino DM, comparable to the bounds from the most stringent laboratory constraints.

Figures (7)

• Figure 1: Self-interaction strength $\lambda_\phi$ required to produce the sterile neutrino DM consistent with observed DM relic density, $\Omega_c h^2 = 0.12$Planck:2018vyg, as a function of mediator mass, for several sterile-neutrino masses and mixing angles (listed in the legend). The characteristic $\mathcal{S}$-shaped curves shown by solid, dashed, and dot-dashed lines correspond to particle masses of $m_4= 100$, $40$, and $16$ keV, respectively; the curves are shown for two different values of $\sin^2(2\theta)$, as labeled in each set of curves. The gray solid lines separate the parameters space corresponding to the three production channels: Cases A, B, and C, as discussed in the text. The blue shaded regions are excluded by neutrino experiments involving the invisible $Z$ decay (Inv. $\Gamma_Z$) and flavor-dependent constraints ($ee$, $\mu\mu$, and $\tau\tau$ denote the type of flavor-specific coupling, corresponding to $\nu_e$, $\nu_\mu$, and $\nu_\tau$, respectively) 2018PhRvD..97g5030BBlinov:2019gcjBerryman:2022hdsEsteban:2021tub. The orange shaded region is inconsistent with BBN and the light-element-abundance measurements Huang:2017eglBlinov:2019gcjEscudero:2019gvw2020PhRvL.124h1802D.
• Figure 2: Ratio of the linear matter power spectrum in a cosmological model with sterile-neutrino DM and self-interacting neutrinos to the standard CDM case is shown for a range of DM particle masses $m_4$ and effective couplings $G_\mathrm{eff}$, shown in the legend. The power spectra exhibit two key features: a bump, which depends on effective self-interaction coupling constant, and a sharp cutoff, related to the mass of sterile neutrinos. In each case, other model parameters are set such that neutrino DM has the observed DM relic abundance. The gray shaded region represents the scales probed by Ly$\alpha$ forest observations from the eBOSS survey eBOSS:2018qyjHe:2023okeHe:2025jwp. The black dashed line denotes the pivot scale at which the reduced likelihood from the PRIYA simulation is evaluated Bird:2023evbFernandez:2023grgHe:2023okeHe:2025jwp.
• Figure 3: 1-D and 2-D marginalized posterior probability distributions for the active-sterile mixing angle, DM relic density, and the effective coupling constant between active neutrinos, assuming a sterile neutrino mass of $m_4 = 100$ keV. The two colors correspond to the results obtained using two different likelihoods based on eBOSS data: PRIYA-simulation and EFT-based likelihoods. The dark and light shaded regions denote the $68\%$ and $95\%$ C. L. contours, respectively.
• Figure 4: Constraints on sterile-neutrino DM from our analysis with Ly$\alpha$ forest observations. The two results of the current work are shown as the $95\%$ C. L. lower bounds on the mixing angle between active and sterile neutrino, as a function of sterile-neutrino mass, obtained from the EFT and on PRIYA models and likelihoods, respectively. We also show the lower bound on the mixing angle from a combine analysis of BBN data Huang:2017eglBlinov:2019gcjEscudero:2019gvw2020PhRvL.124h1802D with laboratory bounds on neutrino self-interactions 2018PhRvD..97g5030BBlinov:2019gcjBerryman:2022hdsEsteban:2021tub (pink dashed line). The combined bound from MW satellite abundance and invisible $Z$ decay from Ref. An:2023mkf is shown as the dashed green line; this bound is only valid under conditions discussed in the text. The black dashed line is lower bound on the mixing angle required to produce the observed DM abundance as measured by PlanckPlanck:2018vyg. The shaded gray region is ruled out by X-ray observations Boyarsky:2005us2012JCAP...03..018WHoriuchi:2013noaPerez:2016tcq2020Sci...367.1465D. Overall, we note that the results of this work provide the most stringent observational bound in this parameter space, surpassing the underproduction limits from the CMB by several orders of magnitude. Note that the result is comparable to the most stringent laboratory constraints in the full sterile-neutrino mass range.
• Figure 5: Schematic depiction of the $m_4$--$G_{\rm eff}$ parameter space where constraints on sterile-neutrino DM obtained while neglecting self-interaction effects on the matter power spectrum ("bump feature") are not applicable (green shaded region). The blue and orange dashed lines indicate the upper limit on the $G_{\rm eff}$ (at $95\%$ C. L. ) from the Ly$\alpha$ forest analysis, and the gray dashed line represent the lower bound on the sterile neutrino DM obtained from MW satellite abundances utilizing WDM simulations An:2023mkf. This bound can be applied to our model only where the effective coupling strength lies outside the green shaded region; within the shaded region, the power spectrum contains a feature that is not strictly modeled in the MW analysis. Gray shaded region represent the parameters space excluded by these MW bounds for our model.