Revised constraints and Belle II sensitivity for visible and invisible axion-like particles

TL;DR

<3-5 sentence high-level summary> This paper revisits axion-like particle (ALP) parameter space for masses in the MeV–GeV range, focusing on couplings to SM gauge bosons and how they can be probed by Belle II. It provides updated bounds from collider, beam-dump, and astrophysical data, and presents a detailed sensitivity study for Belle II in both visibly and invisibly decaying ALP scenarios, including ALP-DM interactions that can yield resonant thermal freeze-out. A key result is that Belle II can substantially improve existing limits, probe otherwise inaccessible regions of parameter space, and test predictive DM-relic scenarios via single-photon searches. The work also outlines complementarities with SHiP and LHC searches and identifies areas requiring full detector simulations for displaced or merged-photon signatures.

Abstract

Light pseudoscalars interacting pre-dominantly with Standard Model gauge bosons (so-called axion-like particles or ALPs) occur frequently in extensions of the Standard Model. In this work we review and update existing constraints on ALPs in the keV to GeV mass region from colliders, beam dump experiments and astrophysics. We furthermore provide a detailed calculation of the expected sensitivity of Belle II, which can search for visibly and invisibly decaying ALPs, as well as long-lived ALPs. The Belle II sensitivity is found to be substantially better than previously estimated, covering wide ranges of relevant parameter space. In particular, Belle II can explore an interesting class of dark matter models, in which ALPs mediate the interactions between the Standard Model and dark matter. In these models, the relic abundance can be set via resonant freeze-out, leading to a highly predictive scenario consistent with all existing constraints but testable with single-photon searches at Belle II in the near future.

Figures (8)

• Figure 1: Feynman diagrams for ALP production in $e^+ e^-$ collisions via ALP-strahlung (left) and photon fusion (right) and the subsequent decay of the ALP into two photons.
• Figure 2: Existing constraints on ALPs with photon coupling (left) and hypercharge coupling (right). Proton beam dump constraints are taken from ref. Dobrich:2015jyk, LEP constraints on $e^+ e^- \to \gamma \gamma$ from ref. Jaeckel:2015jla, CDF constraints on $Z \to \gamma\gamma$ from ref. Bauer:2017ris, bounds from horizontal branch stars from ref. Cadamuro:2011fd, bounds from visible decays of ALPs produced in SN 1987A from ref Jaeckel:2017tud and bounds from heavy-ion collisions from ref. Knapen:2017ebd. All other constraints have been revisited and updated in the present work.
• Figure 3: Comparison of the thermally averaged annihilation cross section $\langle \sigma v \rangle$ (orange, solid) and the annihilation cross section in the limit $v \to 0$ (blue, dotted) as a function of the mass ratio $r = m_\chi / m_a$ for $m_a = 1 \: \mathrm{GeV}$ and $g_{a\gamma\gamma} = 10^{-4}\:\mathrm{GeV}^{-1}$. The red dashed line indicates the simplified expression from eq. (\ref{['eq:sigsimple']}), which is valid close to the resonance. Away from the resonance the cross section depends also on the ALP-DM coupling, which has been set to $g_{a\chi\chi} = 10^{-3}\:\mathrm{GeV}^{-1}$.
• Figure 4: Present and future constraints on ALPs decaying into DM compared to the parameter region where one can reproduce the observed DM relic abundance via resonant annihilation of DM into photons. Note that this process is efficient only if $m_\chi$ is slightly smaller than $m_a / 2$ (see figure \ref{['fig:sigmav']}).
• Figure 5: Illustration of the different kinematic regimes relevant for ALP decays into two photons with Belle II.