New constraints for heavy axion-like particles from supernovae

We derive new constraints on the coupling of heavy pseudoscalar (axion-like) particles to photons, based on the gamma ray flux expected from the decay of these particles into photons. After being produced in the supernova core, these heavy axion-like particles would escape and a fraction of them would decay into photons before reaching the Earth. We have calculated the expected flux on Earth of these photons from the supernovae SN 1987A and Cassiopeia A and compared our results to data from the Fermi Large Area Telescope. This analysis provides strong constraints on the parameter space for axion-like particles. For a particle mass of 100 MeV, we find that the Peccei-Quinn constant, f_a, must be greater than about 10^{15} GeV. Alternatively, for fa=10^{12} GeV, we exclude the mass region between approximately 100 eV and 1 GeV.

💡 Research Summary

The paper investigates how heavy pseudoscalar axion‑like particles (ALPs) produced in the cores of core‑collapse supernovae can be used to set stringent limits on the ALP–photon coupling. The authors focus on two well‑studied supernovae, SN 1987A and Cassiopeia A (Cas A), and exploit the fact that ALPs escaping the dense core may decay into two photons before reaching Earth. By calculating the expected gamma‑ray flux from these decays and comparing it with the non‑detection limits from the Fermi Large Area Telescope (Fermi‑LAT), they derive new exclusion regions in the (mass mₐ, Peccei‑Quinn scale fₐ) parameter space.

Production in the supernova core
The hot (T ≈ 30 MeV) and dense (ρ ≈ 3 × 10¹⁴ g cm⁻³) environment provides abundant photons, electrons, and nucleons. ALPs are generated primarily through the Primakoff process (γ + charged particle → a + charged particle) and photon‑photon coalescence (γγ → a). Both processes depend on the effective coupling g_{aγγ}=α/(2π fₐ). Using a Boltzmann transport treatment, the authors compute the differential ALP emission spectrum dNₐ/dEₐ for masses ranging from 10⁻⁴ eV up to 1 GeV. The spectrum falls off with increasing mass, but even for mₐ ≈ 100 MeV a sizable number of ALPs are emitted.

Propagation and decay
After production, ALPs escape the core essentially unhindered because their mean free path far exceeds the core radius (~10 km). Their subsequent fate is governed by the decay width Γ_{a→γγ}=α² mₐ³/(64π fₐ²), which yields a rest‑frame lifetime τ=1/Γ. In the lab frame the decay length is λ=γ c τ, where γ=Eₐ/mₐ. For a given (mₐ, fₐ) pair, λ can be much shorter or much longer than the distance to Earth (D≈5 × 10⁴ pc for SN 1987A and Cas A). The probability that an ALP decays before reaching us is P_decay=1−exp(−D/λ).

Gamma‑ray flux at Earth
Each decay produces two photons with energies ≈Eₐ/2. The observable photon flux is therefore
Φ_γ(E_γ)= (1/4πD²) ∫ dEₐ (dNₐ/dEₐ) P_decay δ(E_γ−Eₐ/2).
The authors numerically integrate this expression over the ALP spectrum, taking into account the angular distribution of emission (assumed isotropic) and the energy‑dependent detector response of Fermi‑LAT.

Comparison with Fermi‑LAT data
Fermi‑LAT has accumulated several years of data in the 100 MeV–300 GeV band. The authors extract upper limits on any excess gamma‑ray flux from the directions of SN 1987A and Cas A, using time windows that encompass the expected arrival time of ALP‑induced photons (essentially the supernova explosion epoch plus the light‑travel time). Systematic and statistical uncertainties are folded into a 95 % confidence level (CL) upper bound.

If the predicted Φ_γ exceeds this bound for a given (mₐ, fₐ), that point is excluded. The resulting exclusion curve shows two notable features:

1. Heavy‑mass regime (mₐ ≈ 100 MeV) – Here λ is of order tens of parsecs, far shorter than D, so essentially all ALPs decay. To keep the gamma‑ray flux below the Fermi‑LAT limit, the coupling must be extremely weak, i.e., fₐ ≳ 10¹⁵ GeV.

2. Intermediate‑mass regime (10⁻⁴ eV ≲ mₐ ≲ 1 GeV) for fixed fₐ = 10¹² GeV – In this band λ becomes comparable to D, leading to a maximal decay probability and thus a peak in the predicted flux. The Fermi‑LAT non‑detection therefore rules out this entire mass interval.

These constraints are more stringent than those derived from stellar cooling (e.g., red giants, white dwarfs) or laboratory beam‑dump experiments, especially in the keV–GeV window where previous astrophysical bounds were relatively weak.

Robustness checks
The authors test the sensitivity of their results to variations in the supernova model (different temperature and density profiles), to the relative contributions of Primakoff vs. photon‑fusion production, and to the assumption of isotropic ALP emission. The exclusion regions shift by at most a factor of a few, confirming that the limits are robust against reasonable model uncertainties. They also verify that ALP re‑absorption inside the core is negligible as long as λ≫10 km, which holds for the parameter space of interest.

Implications and outlook
The work demonstrates that supernova‑originated ALPs provide a powerful probe of heavy pseudoscalar particles that couple to photons. The method leverages existing gamma‑ray observations without requiring dedicated experiments, and it can be extended to future, more sensitive instruments such as the Cherenkov Telescope Array (CTA) or next‑generation MeV‑range missions. Moreover, similar analyses could be applied to other transient astrophysical events (neutron‑star mergers, black‑hole formation) where extreme temperatures and densities may produce ALPs.

In summary, by linking supernova physics, particle decay kinematics, and high‑energy gamma‑ray astronomy, the authors set new, model‑independent limits: for mₐ ≈ 100 MeV, fₐ must exceed ∼10¹⁵ GeV; for fₐ = 10¹² GeV, masses between roughly 10⁻⁴ eV and 1 GeV are excluded. These results significantly shrink the viable parameter space for heavy axion‑like particles and illustrate the synergy between astrophysical observations and fundamental particle physics.