Universal lower bound on the axion decay constant from free streaming effects

TL;DR

This paper derives universal lower bounds on the axion decay constant based on free-streaming effects of warm dark matter behavior, independent of specific models or cosmological histories.

Contribution

It establishes model-independent constraints on axion parameters from free-streaming effects, applicable across various scenarios enhancing relic abundance.

Findings

Lower bound on $f_a$ from Lyman-$eta$ forest data: $oxed{10^{15} ext{ GeV}}$ for gradient axion modes.

Weaker bound: $f_a oxed{ ext{greater than } 10^{14} ext{ GeV}}$ for delayed oscillation models.

Provides scalable parameters for stronger constraints with future observations.

Abstract

We show that enhancement of the axion relic abundance compared to the standard misalignment contribution generically leads to the production of nonzero momentum axion modes, resulting in warm dark matter behavior and enhanced isocurvature perturbations. It leads to universal constraints on the axion parameter space that are independent of detailed model assumptions and cosmological history. For models enhancing relic abundance with gradient axion modes, observations of the Lyman-$\alpha$ forest impose a lower bound on the axion decay constant, $f_a \gtrsim 10^{15} {\rm GeV}\,(10^{-18}{\rm eV}/m_a)$, from the free-streaming effect. For models relying on the delay of coherent axion oscillations, we obtain a slightly weaker bound, $f_a \gtrsim 10^{14} {\rm GeV}\,(10^{-18}{\rm eV}/m_a)$. We make relatively conservative choices to establish these universal bounds but also provide scaling parameters that can be calibrated for stronger constraints in concrete models and updated as observations improve.