IceCube's Sensitivity Prospects to MeV-Scale Axion-Like Particles from Core-Collapse Supernovae

We present a novel framework to estimate the sensitivity and discovery potential of IceCube to axion-like particles (ALPs) produced in core-collapse supernovae (CCSNe), covering ALP masses from 1 MeV to several hundred MeV. A key feature of this work is the explicit handling of the final-state leptons produced in ALP interactions with $^{16}$O nuclei and protons, which can generate Cherenkov light detectable in IceCube. These processes are being fully integrated into a detector-level simulation chain, enabling realistic detector signal modeling beyond existing estimates. The framework enables sensitivity forecasts for both direct detection and constraints based on time delays relative to the neutrino burst, across a range of ALP emission models. This approach may also extend to other MeV-scale dark sector particles. Preliminary sensitivity estimates are in progress and will be presented.

💡 Research Summary

This paper introduces a comprehensive simulation framework to evaluate IceCube’s sensitivity to axion‑like particles (ALPs) produced in core‑collapse supernovae (CCSNe) with masses in the MeV to few‑hundred‑MeV range. The authors focus on ALPs that couple to nucleons (characterized by the coupling constant (g_{aN})) and explore two direct detection channels inside the IceCube detector: radiative scattering on free protons ((a + p \rightarrow p + \gamma)) and inelastic scattering on oxygen nuclei ((a + {}^{16}!O \rightarrow {}^{16!*}!O \rightarrow {}^{16}!O + \gamma)). The proton channel yields a continuous gamma‑ray spectrum, while the oxygen channel excites discrete nuclear levels that de‑excite via characteristic gamma‑ray cascades.

The paper begins with a review of ALP production in the proto‑neutron star (PNS) phase of a CCSN. At temperatures of 30–50 MeV and nuclear‑saturation densities, ALPs are generated primarily through nucleon‑nucleon bremsstrahlung ((N+N\rightarrow N+N+a)) and pion‑induced processes ((\pi^-+p\rightarrow n+a)). Depending on the strength of (g_{aN}), ALPs either free‑stream out of the PNS, carrying away energy and potentially shortening the observed neutrino burst (as constrained by SN 1987A), become partially trapped, or become fully thermalized. For masses above ∼1 MeV the kinematics shift the emitted spectrum toward lower velocities, leading to measurable arrival‑time delays relative to the prompt neutrino signal.

To translate the astrophysical flux into an observable IceCube signal, the authors integrate several components: (1) state‑of‑the‑art ALP production models that provide energy‑dependent fluxes at Earth; (2) cross‑section calculations for ALP‑proton and ALP‑oxygen interactions, including nuclear excitation probabilities and branching ratios; (3) a dedicated Monte‑Carlo cascade code (based on the ASTERIA package) that propagates the primary gamma rays through Antarctic ice, handling photo‑electric absorption, Compton scattering, pair production, and subsequent electron/positron bremsstrahlung. The resulting electron/positron spectra constitute the source of Cherenkov photons.

The detector‑response layer converts the electron/positron energy and geometry into expected digital optical module (DOM) hit rates. Because IceCube’s supernova detection operates as a collective rate increase over a ∼10 s window rather than event‑by‑event reconstruction, the observable is the excess in total DOM hit counts coincident with the supernova. The authors implement a fast response model that incorporates Cherenkov photon yield, DOM angular acceptance, dark noise, and trigger thresholds. This model yields a predicted rate increase for a given ALP flux, allowing the construction of a significance metric (e.g., a 5σ cut corresponding to (\xi=5)).

Preliminary results are presented for a benchmark scenario with (g_{ap}=2.3\times10^{-5}) and (m_a\ll1) MeV, based on the flux shown in Figure 1 of the paper. The simulation predicts a detectable signal (≥5σ) for supernovae located anywhere within the Milky Way and even out to the Large and Small Magellanic Clouds (∼50 kpc). The authors emphasize that this direct detection channel is complementary to indirect constraints from supernova cooling and from ALP‑photon conversion in Galactic magnetic fields.

A novel aspect of the work is the exploitation of arrival‑time delays. For heavier ALPs the velocity distribution becomes non‑relativistic, producing a spread of arrival times that can be measured against the sharply peaked neutrino burst. By fitting the time profile of the DOM rate excess, IceCube could set timing‑based limits on the ALP mass‑coupling parameter space, providing an additional handle that is largely independent of the overall flux normalization.

The paper concludes by outlining future directions: extending the framework to ALPs that couple to photons or electrons, applying the same methodology to other large‑volume Cherenkov detectors (e.g., KM3NeT, Hyper‑Kamiokande), and performing a full systematic study of uncertainties (nuclear cross sections, supernova model variations, ice optical properties). The authors anticipate that, once the full sensitivity study is completed, IceCube will be able to probe a sizable region of MeV‑scale dark sector parameter space that is currently unconstrained, turning any future Galactic CCSN into a powerful laboratory for new physics.