Probing the disk-jet coupling in M87

TL;DR

This work investigates the origin of the larger horizon-scale ring observed at 86 GHz in M87 compared with 230 GHz by combining 3D GRMHD simulations of magnetically arrested disks with radiative transfer that includes a mixed thermal+nonthermal electron distribution via a κ-distribution. Synthetic visibilities are generated and fitted in the Fourier domain with an azimuthally asymmetric thick $m$-ring model to emulate VLBI data, revealing that 86 GHz rings are systematically larger due to enhanced synchrotron self-absorption and emission from both the disk and jet footpoints, with MAD-driven magnetic-flux eruptions further increasing the ring size. A key finding is that nonthermal electron acceleration is necessary to reproduce the observed 86 GHz morphology, and that the ring diameter is sensitive to MAD variability and angular scale ($GM/Dc^2$); when adopting $GM/Dc^2=4.2 m{ ext{ } extmu as}$, the simulated 86 GHz diameter approaches the observed value. The results underscore the importance of including nonthermal populations and MAD dynamics in horizon-scale models of jet–disk coupling and inform future high-resolution observations with next-generation VLBI arrays.

Abstract

Context. Recent GMVA observations of M 87 at event horizon scales revealed a ring-like structure which is 50% larger at 86 GHz than the ring observed by the Event Horizon Telescope at 230 GHz. Aims. In this paper, we study a possible origin of the increased ring size at 86 GHz. We specifically aim to study the role the nonthermal electron population plays in the observed event horizon scales. Methods. We carry out 3D general relativistic magnetohydrodynamic simulations followed by radiative transfer calculations. We incorporate into the latter synchrotron emission from both thermal and nonthermal electrons. To better compare our results to observations, we generate synthetic interferometric data adjusted to the properties of the observing arrays. We fit geometrical models to this data in Fourier space through Bayesian analysis to monitor the variable ring size and width over the simulated time span of years. Results. We find that the 86 GHz ring is always larger than the 230 GHz ring, which can be explained by the increased synchrotron self-absorption at 86 GHz and the mixed emission from both the accretion disk and the jet footpoints, as well as flux arcs ejected from a magnetized disk. We find agreement with the observations, particularly within the error range of the observational value of M/D for M 87. Conclusions. We show that state-of-the art 3D GRMHD simulations combined with thermal and nonthermal emitting particles can explain the observed frequency-dependent ring size in M 87. Importantly we found that MAD events triggered in the accretion disk can significantly increase the lower frequency ring sizes.

Figures (21)

• Figure 1: Evolution of our GRMHD simulation over time. The top and middle panels show the logarithm of the density in the equatorial and a meridional plane, respectively, for three different times. The bottom panel shows the temporal evolution of the mass accretion rate ($\dot{M}$) and the MAD parameter, $\Phi=\phi_{\rm BH}/\sqrt{\dot{M}}$. The vertical dashed blue lines mark the time frames of the density plots.
• Figure 2: Results of the GRRT calculations at $t=26000\,M$. The images show the flux density at 86 GHz and 230 GHz. Notice that the images are convolved with half of the nominal resolution of the GMVA (27$\,\mu$as) and the EHT (10$\,\mu$as), indicated by the white circle at the bottom right. The contours begin at $1/1000$ of the maximum flux density and increase by factors of two. In the bottom panel we display the broadband radio spectrum of our model. The observational flux values are taken from prieto2016, perlman2001, whysong2004, hada2017, and lister2018.
• Figure 3: Structural decomposition of the horizon scale structure in our model at $t=26000\,M$. The panels show from left to right the total, disk, and jet plus wind contributions for 86 GHz (top row) and 230 GHz (bottom row).
• Figure 4: Fourier transforms of the central regions of our simulated data at $t=24500\,M$. Plotted on top is the $u-v$ coverage at 86 GHz and 230 GHz used for the generation of synthetic observations, using the array configurations presented in lu2023 and EHTa2019, respectively. Note the different scales in the two plots.
• Figure 5: Visibility amplitudes and closure phases over uv distance for the time step $t=28600\,M$ at 86 GHz. Overlaid in red is the result of fitting a thick $m$-ring model to the data. The $\chi^2$ values have been calculated taking an additional 10% error into account for ALMA visibilities.