Stochastic galactic supernova flux of semi-relativistic particles

TL;DR

The paper challenges the standard assumption that semi-relativistic particles (SRPs) from galactic core-collapse SNe produce a smooth, stationary flux at Earth, arguing that the short observational window relative to the SRP packet duration yields only narrow energy slices from individual SNe and a flux that depends sensitively on the SN history. It introduces a stochastic numerical framework that simulates the Milky Way SN history over $9\times 10^{5}$ years, computing arrival times with $t_{\rm arr}(E^{\mathrm{Earth}})=t_{k}+t_{\rm SRP}(E^{\mathrm{Earth}})$ and aggregating contributions to obtain a stochastic flux $\mathcal{F}(E^{\mathrm{Earth}})$, thereby revealing substantial spectral fluctuations. The framework is applied to MeV ALPs and fermionic DM, showing that smooth-diffuse bounds can be overstated and that stochastic realizations can produce spectral peaks and weaker constraints, especially for sub-MeV masses ($m_{a}\lesssim 1$ MeV, $m_{\chi}\lesssim 10$ MeV). This stochastic approach provides a robust tool for interpreting terrestrial searches, extends analyses to sub-MeV regimes, and delivers publicly useful predictions and uncertainties across different SN histories.

Abstract

New exotic particles with MeV masses, such as axion-like particles or light dark matter, can be emitted from core-collapse supernovae (SNe) with semi-relativistic velocities. Due to their speed dispersion, they would arrive at Earth as an extended packet with a time spread that can be as large as tens of millennia for typical detectors. It has been argued in the literature that the superposition of packets from all galactic SNe would give rise to a smooth and stationary diffuse flux that could be observable on terrestrial experiments. In this article, we critically examine this hypothesis by carrying out a numerical simulation of the galactic history of SN explosions. We show that, although the particle packets do overlap, due to the short observational time window, each of them only contributes with a very narrow range of energies and with an intensity that depends on the SN distance. As a consequence, the energy dependence of the resulting flux is extremely sensitive to the stochastic nature of the SN population and far from smooth. This has profound implications for the expected signature in terrestrial experiments, which displays a spectral shape that is not properly described by the smooth approximation. We develop a numerical tool to compute this stochastic galactic flux for generic semi-relativistic particles, which also allows us to explore sub-MeV particles, where the smooth diffuse flux approach does not hold. To test this framework, we revisit existing bounds on axion-like particles and fermionic dark matter, finding weaker constraints than previously reported.

Figures (8)

• Figure 1: Example of the location in galactic coordinates of the last 10000 SNe in a simulation. The colour bar shows the time of each event.
• Figure 2: Left: fluence factor $\mathcal{F}_{k}$ for each SN as a function of the energy at Earth in a simulation example (red lines). The weight value and the energy range of each event depend on the SRP mass, as well as on both the location and the time of the SN. The blue histogram represents the number of SNe that contribute to each energy range. Right: smooth $\bar{\mathcal{F}}$ in green, and the stochastic fluence factor integrated in 2 MeV energy bins and averaged over the 20 SN galactic histories $\left\langle \mathcal{F} \right\rangle_{\rm sim}$ as defined in \ref{['eq:medE']} in purple. For the latter, the line corresponds to the mean value while the dashed lines are the 1$\sigma$ band, taking 20 simulated galactic SN histories. In both panels we consider a SRP of mass $m=30$ MeV and an observation time window of $\Delta t_{\rm obs}=20$ yr.
• Figure 3: Left: total fluence of ALPs at Earth computed with either the smooth (green line) or the stochastic (purple and red lines) approach for an observation time window of $\Delta t_{\rm obs}=20$ yr and a single SN history. Right: expected event rate in SK phase IV via $a\ p\rightarrow p\ \gamma$, considering the smooth diffuse approximation (green line) and the stochastic approach (purple, orange and blue lines) for 3 different galactic SN history simulations. Expected events from SM processes (black dashed line) and observed events (black points with error bars) are taken from Ref. Super-Kamiokande:2021jaq. Both panels consider $m_a=30$ MeV and $g_{ap}=9.4\times 10^{-5}$.
• Figure 4: Current (left) and projected (right) bounds on the ALP parameter space, using the smooth diffuse approximation (gray), and the stochastic approach (purple), including a 1$\sigma$ uncertainty band from 20 SN galactic history simulations. Complementary bounds are also shown from expected solar axion events in SNO Bhusal:2020bvx (red) and from SN 1987A cooling Lella:2023bfb (light blue).
• Figure 5: Left: total fluence of $\chi$ at Earth computed with both the smooth and the stochastic approach for a single galactic SN history. Right: expected event rate in LZ with a final exposure of 15 tonne-years, considering the smooth diffuse approximation (green line) and the stochastic approach (purple, orange and blue lines) for 3 different galactic SN history simulations. Both panels consider $m_{\chi}=26$ MeV, $\log(y)=-15.3$ and an observational time window of $\Delta t_{\rm obs}=2.74$ yr. The vertical dotted lines delimit the LZ energy region.