Simulation-Based Prediction of Black Hole Spectra: From $10M_\odot$ to $10^8 M_\odot$

TL;DR

This work demonstrates that coupling GRMHD disk simulations with comprehensive, self-consistent radiation transfer can reproduce observed black hole spectra across eight orders of magnitude in mass, at two sub-Eddington accretion rates. By using Pandurata for the corona and PTransX for the disk, the authors enforce energy, ionization, and thermal balance, generating broadband spectra that align with hard and steep power-law states in X-ray binaries and typical AGN slopes around $\Gamma\approx2$, including a mass- and accretion-rate–dependent soft excess linked to warm Comptonization. The results highlight the importance of a realistic, multi-region thermal structure, inhomogeneous coronae, and disk illumination in shaping emergent spectra, and they establish a framework for connecting GRMHD dynamics directly to observables without free spectral parameters. The study also identifies limitations, such as the single spin value, the boundary-condition sensitivity of the soft excess, and the need to extend to optical/UV for AGN, guiding future improvements in modeling accretion physics from stellar to supermassive black holes.

Abstract

It has long been thought that black hole accretion flows are driven by magnetohydrodynamic (MHD) turbulence, and there are now many general relativistic global simulations illustrating the dynamics of this process. However, many challenges must be overcome in order to predict observed spectra from luminous systems. Ensuring energy conservation, local thermal balance, and local ionization equilibrium, our post-processing method incorporates all the most relevant radiation mechanisms: relativistic Compton scattering, bremsstrahlung, and lines and edges for 30 elements and all their ions. Previous work with this method was restricted to black holes of $10 M_\odot$; here, for the first time, we extend it to $10^8 M_\odot$ and present results for two sub-Eddington accretion rates and black hole spin parameter 0.9. The spectral shape predicted for stellar-mass black holes matches the low-hard state for the lower accretion rate and the steep power law state for the higher accretion rate. For high black hole mass, both accretion rates yield power-law continua from $\sim 0.5 - 50$~keV whose X-ray slopes agree well with observations. For intermediate mass black holes, we find a soft X-ray excess created by inverse Compton scattering of low-energy photons produced in the thermal part of the disk; this mechanism may be relevant to the soft X-ray excess commonly seen in massive black holes. Thus, our results show that standard radiation physics applied to GRMHD simulation data can yield spectra reproducing a number of the observed properties of accreting black holes across the mass spectrum.

Figures (15)

• Figure 1: Optical depth (integrated over $\theta$ from pole to midplane) minus one for each accretion rate case. The white region corresponds to the region where no $\tau=1$ photosphere exists or the area interior to the computational domain. The red overlay shows the PTransX grid.
• Figure 2: Several temperatures as a function of radius for each mass (color; see legend) and accretion rate (left: $\dot{m} = 0.01$, right: $\dot{m} = 0.1$). The dashed lines show the azimuthally averaged thermal core effective temperature. The solid lines with triangles show the radiation temperature one cell outside the thermal core (that is, at the base of the atmosphere), while the solid lines with circles show the gas temperature at this location. The dotted lines show the photospheric gas temperature. The solid lines show the mass-weighted coronal temperature, averaged over azimuthal and polar angle.
• Figure 3: Temperature solutions for an azimuthal slice where each panel shows one model. The black dotted line shows the photosphere in this azimuthal slice, while the black dashed line shows the azimuthally averaged thermal core boundary of PTransX. The region outside the photosphere is the Pandurata solution (note that Fig. \ref{['fig:maps']} has a different temperature scale) and the region inside the photosphere is the PTransX solution. White regions within the photosphere are either the thermal core (where it exists) or slabs which fail to converge.
• Figure 4: Several example temperature PTransX temperature solutions for $\phi=0$ in the $\dot{m}=0.01$, $M=10\;M_\odot$ model (corresponding to the upper left panel in Fig. \ref{['fig:T']}).
• Figure 5: Pandurata azimuthal slices of temperature (left) and Compton power (center) and HARM3D dissipation rate (right) for the $M=10^8\;M_\odot$, $\dot{m} = 0.01$ model (upper) and $\dot{m} = 0.1$ model (lower). The red dashed line shows the ISCO while the black dotted line is the photosphere.