Ultra-luminous X-ray Sources as Supercritical Accretion Disks: Spectral Energy Distributions

TL;DR

This paper presents a spectral energy distribution model for supercritical accretion disks applied to ULXs, distinguishing it from standard models by predicting flat spectra in X-ray and smaller black hole masses.

Contribution

The paper introduces a supercritical accretion disk model that explains ULX spectra and predicts stellar-mass black holes, differing from standard models that suggest larger black hole masses.

Findings

The SCAD model fits ULX spectra with stellar-mass black holes (~10 M_sun).

The model predicts a flat nu F_nu spectrum in the 0.3-1 keV X-ray range.

Distinguishing features between SCAD and DISKIR models are in the X-ray spectrum and inferred black hole masses.

Abstract

We describe a model of spectral energy distribution in supercritical accretion disks (SCAD) based on the conception by Shakura and Sunyaev. We apply this model to five ultra-luminous X-ray sources (ULXs). In this approach, the disk becomes thick at distances to the center less than the spherization radius, and the temperature dependence is T \propto r^{-1/2}. In this region the disk luminosity is L_bol ~ L_Edd ln(Mdot/Mdot_Edd), and strong wind arises forming a wind funnel above the disk. Outside the spherization radius, the disk is thin and its total luminosity is Eddington, L_Edd. The thin disk heats the wind from below. From the inner side of the funnel the wind is heated by the supercritical disk. In this paper we do not consider Comptonization in the inner hot winds which must cover the deep supercritical disk regions. Our model is technically similar to the DISKIR model of Gierlinski et al. The models differ in disk type (standard - supercritical) and irradiation (disk - wind). We propose to distinguish between these two models in the X-ray region, ~0.3 - 1 keV, where the SCAD model has a flat nu F_nu spectrum, and the DISKIR model never has a flat part, as it is based on the standard alpha-disk. An important difference between the models can be found in their resulting black hole masses. In application to the ULX spectra, the DISKIR model yields black hole masses of a few hundred solar masses, whereas the SCAD model produces stellar-mass black holes ~10 M_sun.