An open question remains whether Ultraluminous X-ray Sources (ULXs) really contain intermediatemass black holes (IMBHs). We carefully investigated the XMM-Newton EPIC spectra of the four ULXs that were claimed to be strong candidates of IMBHs by several authors. We first tried fitting by the standard spectral model of disk blackbody (DBB) + power-law (PL), finding good fits to all of the data , in agreement with others. We, however, found that the PL component dominates the DBB component at ∼ 0.3 to 10 keV. Thus, the black hole parameters derived solely from the minor DBB component are questionable. Next, we tried to fit the same data by the "p-free disk model" without the PL component, assuming the effective temperature profile of T eff ∝ r −p where r is the disk radius. Interestingly, in spite of one less free model parameters, we obtained similarly good fits with much higher innermost disk temperatures, 1.8 < kT in < 3.2 keV. More importantly, we obtained p ∼ 0.5, just the value predicted by the slim (supercritical) disk theory, rather than p = 0.75 that is expected from the standard disk model. The estimated black hole masses from the p-free disk model are much smaller; M < ∼ 40M ⊙ . Furthermore, we applied a more sophisticated slim disk model by Kawaguchi (2003, ApJ, 593,69), and obtained good fits with roughly consistent black hole masses. We thus conclude that the central engines of these ULXs are super-critical accretion flows to stellar-mass black holes.

We use XMM–Newton and Swift data to study a spectral variability in the ultraluminous X‐ray source (ULX), Holmberg IX X−1. The source luminosity varies by a factor of 3–4, giving rise to corresponding spectral changes which are significant, but subtle, and not well tracked by a simple hardness ratio. Instead, we co‐add the Swift data in intensity bins and do full spectral fitting with disc plus thermal Comptonization models. All the data are well fitted by a low temperature, optically thick Comptonizing corona, and the variability can be roughly characterized by decreasing temperature and increasing optical depth as the source becomes brighter, as expected if the corona is becoming progressively mass loaded by material blown off the super‐Eddington inner disc. This variability behaviour is seen in other ULX which has similar spectra, but is opposite to the trend seen in the ULX with much softer spectra. This supports the idea that there are two distinct physical regimes in ULXs, where the spectra go from being dominated by a disc‐corona to being dominated by a wind.

We propose a methodology to derive a black-hole mass for super-critical accretion flow. Here, we use the extended disk blackbody (extended DBB) model, a fitting model in which the effective temperature profile obeys the relation $T_{\rm eff} \propto r^{-p}$, with $r$ being the disk radius and $p$ being treated as a fitting parameter. We first numerically calculate the theoretical flow structure and its spectra for a given black-hole mass, $M$, and accretion rate, $\dot{M}$. Through fitting to the theoretical spectra by the extended DBB model, we can estimate the black-hole mass, $M_{\rm x}$, assuming that the innermost disk radius is $r_{\rm in}=3r_{\rm g} (\propto M_{\rm x})$, where $r_{\rm g}$ is the Schwarzschild radius. We find, however, that the estimated mass deviates from that adopted in the spectral calculations, $M$, even for low-$\dot{M}$ cases. We also find that the deviations can be eliminated by introducing a new correction for the innermost radius. Using this correction, we calculate mass correction factors, $M/M_{\rm x}$, in the super-critical regimes for some sets of $M$ and $\dot M$, finding that a mass correction factor ranges between $M/M_{\rm x} \sim$1.2-1.6. The higher is $\dot{M}$, the larger does the mass correction factor tend to be. Since the correction is relatively small, we can safely conclude that the black holes in ULXs, which Vierdayanti et al. (2006, PASJ, 58, 915) analyzed, are stellar-mass black holes with the mass being $&lt;100M_{\odot}$.