Simulation-Based Prediction of Black Hole X-ray Spectra and Spectral Variability

TL;DR

This paper uses advanced simulations combining relativistic MHD and radiation transfer to predict black hole X-ray spectra and variability, matching observed features and revealing how turbulence affects spectral properties.

Contribution

It introduces a novel simulation-based method that integrates relativistic MHD with detailed radiation transfer to study spectral formation and variability in black hole accretion disks.

Findings

Simulated spectra resemble observed hard state X-ray binaries.

Turbulence causes significant luminosity and spectral slope variations.

Coronal temperature distribution critically influences spectral shape.

Abstract

Data derived from general relativistic magnetohydrodynamic simulations of accretion onto black holes can be used as input to a postprocessing scheme that predicts the radiated spectrum. Combining a relativistic Compton scattering radiation transfer solution in the corona with detailed local atmosphere solutions incorporating local ionization and thermal balance within the disk photosphere, it is possible to study both spectral formation and intrinsic spectral variability in the radiation from relativistic accretion disks. With this method, we find that radiatively efficient systems with black holes of $10M_\odot$ accreting at $\approx 0.01$ in Eddington units produce spectra very similar to those observed in the hard states of X-ray binaries. The spectral shape above 10keV is well described by a power law with an exponential cutoff. Intrinsic turbulent variations lead to order-unity changes in bolometric luminosity, variations in the logarithmic spectral slope $\sim 0.1$, and factor of 2 alterations in the cutoff energy on timescales $\sim 50(M_{\rm BH}/10 M_\odot)$ms. Within the corona, the range of gas temperature spans more than 1 order of magnitude. The wide distribution of temperatures is central to defining the spectrum: the logarithmic spectral slope is harder by $\sim 0.3$ and the cutoff energy larger by a factor $\sim 10 - 30$ than if the coronal temperature everywhere were its mass-weighted mean.