Cosmic rays, gamma rays and neutrinos from discrete black hole X-ray binary ejecta

TL;DR

This paper assesses whether discrete ejecta from black hole X-ray binaries can contribute to Galactic cosmic rays, using MAXI J1820+070 as a case study and grounding the analysis in an analytic acceleration/expansion framework. It finds that ejecta can accelerate protons up to around $E_{ ext{max}} oughly 1.6\times10^{16}\,\mathrm{eV}$, with the value scaling as $E_{ ext{max}} \propto \mu^{-1/2}$, and estimates a Galactic CR contribution at the few percent level near the knee, though propagation would reduce the effective knee contribution. Gamma-ray and neutrino outputs from steady-state ejecta are predicted to be undetectable with current facilities, but brief, early-time γ-ray bursts could be observable with sensitive instruments like CTAO if ejecta are sufficiently compact. Overall, BH-XRB ejecta represent a sub-dominant but potentially non-negligible component of the knee region, motivating further propagation modeling and coordinated multi-wavelength observations to test connections with the latest TeV γ-ray detections.

Abstract

The origin of cosmic rays from outside the Solar system are unknown, as they are deflected by the interstellar magnetic field. Supernova remnants are the main candidate for cosmic rays up to PeV energies but due to lack of evidence, they cannot be concluded as the sources of the most energetic Galactic CRs. We investigate discrete ejecta produced in state transitions of black hole X-ray binary systems as a potential source of cosmic rays, motivated by recent $>100$ TeV $γ$-ray detections by LHAASO. Starting from MAXI J1820+070, we examine the multi-wavelength observations and find that efficient particle acceleration may take place (i.e. into a robust power-law), up to $\sim2\times 10^{16}μ^{-1/2}$ eV, where $μ$ is the ratio of particle energy to magnetic energy. From these calculations, we estimate the global contribution of ejecta to the entire Galactic spectrum to be $\sim1\%$, with the cosmic ray contribution rising to $\sim5\%$ at PeV energies, assuming roughly equal energy in non-thermal protons, non-thermal electrons and magnetic fields. In addition, we calculate associated $γ$-ray and neutrino spectra of the MAXI J1820+070 ejecta to investigate new detection methods with CTAO, which provide strong constraints on initial ejecta size of order $10^7$ Schwarzschild radii ($10^{-5}$ pc) assuming a period of adiabatic expansion.

Figures (7)

• Figure 1: Internal energy over time from equipartition for the MAXI J1820+070 ejection, as inferred from radio observations. An adiabatic expansion fits the data, with the purple dashed line highlighting the need for a break in expansion speed (best-fit at 80 days). Data plotted for $R(90 \text{ days})=3.3 \times 10^3$ AU.
• Figure 2: The proton energy spectrum of the MAXI J1820+070 ejection at $t=90$ days after inferred ejection, multiplied by volume. The shaded area represents departures from equipartition ($0.1<\mu<100$) for the mean size. Maximum energy scales as $E_\text{max}\propto \mu^{-1/2}$ and in general size uncertainty of MAXI J1820+070 at 90 days dominates over potential deviations from equipartition.
• Figure 3: Fitted log-normal distribution of synchrotron flare energies of 30 candidate BH-XRBs (Cowie et al., in prep.), with the MAXI J1820+070 flare energy indicated. The mean flare energy is $2\times10^{38\pm1.1}$ erg. Assuming a proportional relation between flare and ejecta energies gives an internal energy distribution for ejecta.
• Figure 4: Blue: predicted contribution to the CR spectrum with $p=2.2$, $E_{\text{max}} = 10^{16}$ eV and $\kappa = 1.3\times10^4$; red: a rough estimate (see discussion below) for the effect of propagation, with $p=2.6$, $E_{\text{max}} = 10^{16}$ eV and $\kappa = 1.3\times10^4$, normalised to have the same energy in CRs; data: CR spectrum Maurin2023 and protons LHAASO2025.
• Figure 5: $\gamma$-rays (from $p+p$ interactions in MAXI J1820+070) could have been detectable with CTAO Observatory2021 for ejecta radii $\lesssim 1.5\times10^7 R_S$.