Reconnection-driven Flares in M87*: Proton-Synchrotron-powered GeV Emission

TL;DR

The paper addresses the origin of GeV–TeV gamma-ray emission from M87* during magnetically arrested disk–driven flux eruptions. It combines analytic estimates of proton (ion) synchrotron emission with 3D radiative PIC simulations of ion–pair reconnection to quantify energy partition and spectra. The authors show that proton synchrotron can account for roughly 5–20% of the dissipated power, producing a GeV component peaking near ~40 GeV, while MeV emission from pair synchrotron dominates and TeV photons arise from inverse Compton scattering by pairs, yielding a consistent MeV–TeV picture during flares. The results indicate proton feedback does not change the reconnection rate (∼0.1–0.15) but does shape the ion energy distribution, and the proton–pair reconnection framework offers a robust mechanism for the GeV component with broader implications for reconnection-driven high-energy emission in accreting black holes.

Abstract

Magnetic reconnection in current layers that form intermittently in radiatively inefficient accretion flows onto black holes is a promising mechanism for particle acceleration and high-energy emission. It has been recently proposed that such layers, arising during flux eruption events, can power the rapid TeV flares observed from the core of M87. In this scenario, inverse Compton scattering of soft radiation from the accretion flow by energetic electron-positron pairs produced near the reconnection layer was suggested as the primary emission mechanism. However, detailed calculations show that radiation from pairs alone cannot account for the GeV emission detected by the Fermi observatory. In this work, we combine analytic estimates with 3D radiative particle-in-cell simulations of pair-proton plasmas to show that the GeV emission can be naturally explained by synchrotron radiation from protons accelerated in the current sheet. Although the exact proton content of the layer is uncertain, our model remains robust across a broad range of proton-to-pair number density ratios. While protons are subdominant in number compared to pairs, our simulations demonstrate that they can be accelerated more efficiently, leading to a self-regulated steady state in which protons dominate the energy budget. Ultimately, proton synchrotron emission accounts for approximately 5%-20% of the total dissipation power. The majority is radiated as MeV photons via pair synchrotron emission, with a smaller fraction emitted as TeV photons through inverse Compton scattering.

Figures (6)

• Figure 1: Schematic illustration of the simulation box (the black rectangle) near the exposed equatorial current layer around the black hole. Toroidal magnetic field lines of different polarities are sketched with the red and blue arrows. The colored $x$-, $y$-, and $z$-axes correspond to the actual orientations of the axes in our simulation. Note that the geometry of the layer implies the absence of a guide magnetic field (in this case, a component along the $y$-axis).
• Figure 2: Time evolutions of space-averaged quantities for two simulations with marginal (red; $\gamma_{\rm syn}^\pm/\sigma_\circ= 0.5$) and strong (blue; $\gamma_{\rm syn}^\pm/\sigma_\circ=0.15$) synchrotron cooling of pairs. (a): Width of the layer along $z$ measured as the FWHM of the energy density (averaged in $x$ and $y$) of pairs (dashed) and protons (solid); the shaded region corresponds to about $25\%$ variation in the FWHM value, indicating how steep the gradient of $w(z)$ is. (b): Mean Lorentz factor of pairs (dashed) and protons (solid), that have participated in reconnection. (c): Space-averaged dimensionless reconnection rate measured as $(\bm{E}\times\bm{B})_z/B^2$ in the region $0.1<z<0.2$. Note that in the case with stronger cooling (blue), the pressure inside the plasmoids is partially provided by the out-of-plane magnetic field. Thus, the width of the pair-dominated region in (a) at later times is similar for both strong and weak cooling.
• Figure 3: Snapshots from two different simulations, with the columns corresponding to different values for the synchrotron cooling strengths of pairs, as indicated in the headings. Each panel is split into two: on the left, we show all the quantities related to pairs, while on the right, we show the quantities related to protons. The first row shows the number densities of pairs and protons (the number density of protons is compensated by their upstream ratio). In the second row we show the cold (left half) and the hot (right half) magnetization parameters, where the role of ions is clearly emphasized. The final row shows the mean Lorentz factors. These plots demonstrate a clear separation between the cooled pairs confined within the plasmoids and the hot uncooled protons that supply most of the pressure farther upstream.
• Figure 4: Time- and $x$-$y$-averaged forces plotted against the $z$-coordinate (perpendicular to the layer) for the $\gamma_{\rm syn}^\pm/\sigma_\circ=0.15$ simulation. The green line shows the magnetic tension force acting toward $z=0$ attempting to compress the sheet. The blue and red lines show the pressure gradients of pairs and protons, respectively, acting in opposition to the $\bm{j}\times\bm{B}$ force. Outside $|z|\gtrsim 0.02$, the magnetic tension is balanced primarily by the pressure of ions.
• Figure 5: Time-averaged energy distributions for ions (solid) and pairs (dashed) for two runs with different pair-cooling strengths. The averaging is done in a steady state, over a period of $6\lesssim ct/L\lesssim8$. The black arrows indicate $\sigma_\circ$ and $\sigma_i$ (the same for both runs), while the colored arrows indicate the average Lorentz factors of ions (solid arrows) and pairs (dashed arrows).