Axiverse Baryogenesis

Abstract

The QCD axion may offer a unified origin for the baryon asymmetry and dark matter through axiogenesis. However, in the minimal QCD axion scenario, axiogenesis either underproduces baryons or overproduces dark matter, and the required kinetic misalignment initial conditions are in tension with axion quality. In this \textit{Letter}, we demonstrate that the axiverse naturally resolves these tensions: the QCD axion emerges as a linear combination of multiple axion-like fields, evading the overclosure problem thanks to new dissipation channels , while introducing additional Peccei--Quinn symmetries that ensure a high quality QCD axion. We illustrate these points in a toy model with two axions. This framework predicts a rich phenomenology within experimental reach, including dark matter detection prospects, astrophysical signals, and collider signatures.

Figures (4)

• Figure 1: Chirality-flipping process through a Yukawa interaction with a Higgs ($H$) in the thermal bath of the massless gauge boson $g$ (left) or via the mass of the fermion, $m_f$ (right). $f_L$ ($f_R$) denotes the left- (right-)handed fermion. These processes appear as wash-out terms in Boltzmann equations governing fermion asymmetries.
• Figure 2: Friction felt by an axion and normalized by the Hubble rate as a function of the temperature in the SM (black curve), a pure Yang-Mills theory (purple curve), and a confining SU(2) dark gauge group with a vector-like fermion of mass 100 GeV (orange curve) or 1 GeV (blue curve) and the same confining scale as QCD. We fix the axion decay constant $f_a$ to the minimum value compatible with experimental bounds to maximize the friction; the overclosure problem is only exacerbated for higher values of $f_a$. The vertical lines illustrate relevant temperatures: the colored lines labeled by $T_\Upsilon$ denote the temperature at which the friction in a vector-like confining SU(2) dominates over Hubble, while the lines labeled $T_{\rm EW}$ and $\Lambda_{\rm QCD}$ indicate the electroweak and QCD phase transitions, respectively. The rapid increase of the sphaleron friction in the pure Yang–Mills (and heavy vectorlike fermion) case arises from the fast growth of the gauge coupling as the temperature approaches confinement, combined with the steep $\alpha$-scaling of the thermal sphaleron rate $\propto \alpha^5 T^4$Arnold:1996dy.
• Figure 3: Schematic presentation of the evolution of the two axions in their potentials below $T_\mathrm{EW}$ (from left to right). The friction to the dark sector takes away enough energy out of the system to evade the overclosure - see the text for details.
• Figure 4: Parameter space (mass $m_i$ vs. PQ-breaking scale $f_i$) for $\theta_1$ (colored lines) and $\theta_2$ (blue regions) consistent with avoiding overclosure. The left boundary of the blue region depends on the decay channel: decay to dark photons or to SM via a bifundamental fermion. In the former case, the left boundary is set by requiring the dark confinement scale to exceed the QCD scale. In the latter case, it varies with the $\xi_d$ mass (600 GeV, 2.4 TeV, and 9.6 TeV in the figure) due to glueball lifetime constraints. 600 GeV is roughly the present LHC bound. The edges labeled (①,②) demarcate where $\theta_2$ constitutes the entirety of DM today, and ② is determined by the requirement that the friction on $\theta_2$ is not dominant over Hubble friction at temperatures above $T_\mathrm{EW}$. The bounds in dark gray rely on the axion-gluon coupling Hook:2017psmGomez-Banon:2024ouxKumamoto:2024wjdBlum:2014vsaSpringmann:2024retBalkin:2022qerCaloni:2022uya (the "Thermal Overproduction" bound depends only on the gluon coupling). The bounds in light gray assume a coupling of $\theta_2$ to photons given by $\frac{\alpha_{\rm EM}}{2\pi}\theta_2 F\tilde{F}$Yin:2024llaRegis:2020fhwTodarello:2023hdkSaha:2025anyPinetti:2025owqGrin:2006awJanish:2023kviWadekar:2021qaeBlout:2000ucCarenza:2023qxhNakayama:2022jzaTodarello:2024qciWang:2023imiCadamuro:2011fdCapozzi:2023xieLiu:2023nct and that it constitutes 100% of the dark matter. These are therefore conservative bounds, rather than true constraints on the model.