Light to Heavy, Brief to Eternal: An Axion for Every Occasion (in the Early Universe)

TL;DR

The paper investigates how the early universe can probe axions across a broad mass–lifetime range by classifying scenarios into four regimes: long-lived axions as dark radiation, long-lived axions as dark matter, metastable axions imprinting on the global $21$ cm signal, and very short-lived axions acting as portals to hidden sectors. It uses a momentum-space Boltzmann formalism to track thermal production and decay timing, using the lifetime scaling $\tau_{a \rightarrow \gamma\gamma} \propto ( f_a / C_\gamma )^2 m_a^{-3}$ to map into three phenomenological regions. For dark radiation, the study computes $\Delta N_{\rm eff}$ as a function of $f_a$ and SM fermion couplings, obtaining robust CMB constraints for leptonic and top-quark couplings while noting IR challenges for gauge-boson couplings in thermal field theory. In the dark-matter and portal sectors, freeze-in production yields a warm-DM-like bound of $m_\chi^{\min} = 22\,\mathrm{keV} \left( \frac{m_{\rm WDM}^{\min}}{6\,\mathrm{keV}} \right)^{4/3} \left( \frac{\sigma_q}{3.6} \right) \left( \frac{106.75}{g_{\star s}(T_P)} \right)^{1/3}$ and axion-mediated portals with non-Abelian $\mathbb{Z}_3$ stabilizers produce distinctive indirect-detection spectra via semi-annihilation, with relic density controlled by dark-sector couplings rather than SM couplings.

Abstract

The early universe grants access to energy scales far beyond those achievable in terrestrial experiments and allows unstable Standard Model particles to play an active dynamical role. In this contribution, we focus on recent studies aimed at quantifying the potential of the early universe to probe the properties and interactions of axions. The discussion is organized around four classes of axion scenarios, ordered from long to short lifetimes: (i) stable or long-lived axions contributing to dark radiation; (ii) stable or long-lived axions produced out-of-equilibrium and constituting dark matter; (iii) metastable axions whose decays inject energy into the primordial plasma and leave observable signatures in the global 21 cm signal; and (iv) very short-lived axions that act only as portals to additional degrees of freedom. Together, these scenarios highlight the interplay between axion phenomenology and early universe cosmology and demonstrate the potential of cosmological data to probe axions over a broad range of masses and lifetimes.

Figures (6)

• Figure 1: Schematic illustration of the axion parameter space in the plane spanned by the axion mass $m_a$ and the inverse coupling $f_a / C_X$ to a generic visible-sector particle $X$. Lines of constant axion lifetime for the decay $a \rightarrow XX$ partition the parameter space into three qualitatively distinct regions. From right to left, these correspond to short-lived axions that decay before BBN, shown in green; axions with intermediate lifetimes whose decays inject energy and can affect cosmological observables, shown in red, with the solid line indicating the current age of the universe; and effectively stable axions that survive to late times and may contribute to dark matter or dark radiation, shown in blue.
• Figure 2: Axion dark radiation arising from couplings to SM fermions. The predicted contributions to $\Delta N_{\rm eff}$ are obtained from a full phase space analysis that consistently tracks the momentum distribution of thermally produced axions. The results shown here are taken from Ref. DEramo:2024jhn.
• Figure 3: Compact and model-independent visualization of mass bounds on freeze-in dark matter. The horizontal and vertical axes correspond to the second moment of the momentum distribution, $\sigma_q$, and the characteristic production temperature $T_P$. Black isocontours indicate the minimum allowed dark matter mass from small-scale structure constraints, obtained by rescaling to a reference warm dark matter mass of $6.8\,\mathrm{keV}$. Superimposed on these model-independent contours are benchmark realizations, with $T_P$ chosen to reflect the characteristic mass or energy scale of the dominant production process. Figure from Ref. DEramo:2025jsb.
• Figure 4: Prospects for constraining the parameter space of an axion coupled to photons in the plane spanned by the axion mass $m_a$ and the axion–photon coupling $g_{a\gamma\gamma}$, as defined in Eq. \ref{['eq:gaggg']}. Figures from Ref. Cima:2025zmc.
• Figure 5: Phenomenology of axion portal scenarios in which the dark matter candidate is either a fermion stabilized by a $\mathbb{Z}_2$ symmetry or a scalar stabilized by a $\mathbb{Z}_3$ symmetry. While scenarios with fermionic dark matter have been extensively explored in the literature, the case of scalar dark matter stabilized by a non-Abelian symmetry exhibits qualitatively new features. Figure from Ref. DEramo:2025xef.