To understand the origin of hard X-ray emissions from black hole X-ray binaries during their low/hard states, we calculate the X-ray spectra of black-hole accretion flow for the following three configurations of hot and cool media: (a) an inner hot flow and a cool outer disk (inner hot flow model), (b) a cool disk sandwiched by disk coronae (disk corona model), and (c) the combination of those two (hybrid model). The basic features we require for successful models are (i) significant hard X-ray emission whose luminosity exceeds that of soft X-rays, (ii) high hard X-ray luminosities in the range of (0.4 - 30) times 10^{37} erg s^{-1}, and (iii) the existence of two power-law components in the hard X-ray band with the photon indices of Gamma_s ~ 2 > Gamma_h, where Gamma_s and Gamma_h are the photon indices of the softer (<10 keV) and the harder (>10 keV) power-law components, respectively. Contribution by non-thermal electrons nor time-dependent evolution are not considered. We find that Models (a) and (b) can be ruled out, since the spectra are always dominated by the soft component, and since only one power-law component, at most, can be reproduced. Only Model (c) can account for sufficiently strong hard X-ray emissions, as well as the existence of the two power-law components, for a large ratio of the accretion rate in the corona to that in the thin disk. The outer disk corona (where the Compton y-parameter is smaller, y < 1) produces the softer power-law component with photon index of Gamma_s ~ 2, whereas the inner hot flow (where y gtrsim 1) generates the harder component with Gamma_h < 2. This model can also account for the observed relationship between the photon index and the reflection fraction.

We study mass flow rate through a disc resulting from a varying mass supply rate. Variable mass supply rate occurs, e.g., during disc state transitions, and in interacting eccentric binaries. It is, however, damped by the viscosity of the disc. Here, we calculate this damping in detail. We derive an analytical description of the propagation of the flow rate using the solution of Lynden-Bell & Pringle, in which the disc is assumed to extend to infinity. In particular, we derive the accretion-rate Green's function, and its Fourier transform, which gives the fractional damping at a given variability frequency. We then compare this model to that of a finite disc with the mass supply at its outer edge. We find significant differences with respect to the infinite disc solution, which we find to overestimate the viscous damping. In particular, the asymptotic form of the Green's function is power-law for the infinite disc and exponential for the finite one. We then find a simple fitting form for the latter, and also calculate its Fourier transform. In general, the damping becomes very strong when the viscous time at the outer edge of the disc becomes longer than the modulation time scale. We apply our results to a number of astrophysical systems. We find the effect is much stronger in low-mass X-ray binaries, where the disc size is comparable to that of the Roche lobe, than in high-mass binaries, where the wind-fed disc can have a much smaller size.

Radiatively inefficient accretion flows (RIAFs) are common feature of low-luminosity accretion flows, including quiescent states of X-ray binaries and low-lunimosity active galactic nuclei. Thermally driven winds are expected from such hot accretion flows. By assuming that the flow has self-similarity structure in the radial direction, we solve the vertical structure of the wind and accretion flows simultaneously and evaluate the mass loss rates by wind. We find that the ratio of the outflow rate to the accretion rate is approximately unity for a viscosity parameter, alpha lesssim 0.1, despite some uncertainties in the angular momentum and temperature distributions. That is, the accretion rate in the RIAFs is roughly proportional to the radius. Moreover, we elucidate the effect of cooling by wind on the underneath accretion flow, finding that this effect could be important for calculating energy spectrum of the RIAF. Observational implications are briefly discussed in the context of Sgr A*.

Hypercritical accretion flows onto stellar mass black holes (BHs) are commonly considered as a promising model of central engines of gamma-ray bursts (GRBs). In this model a certain fraction of gravitational binding energy of accreting matter is deposited to the energy of relativistic jets via neutrino annihilation and/or magnetic fields. However, some recent studies have indicated that the energy deposition rate by neutrino annihilation is somewhat smaller than that needed to power a GRB. To overcome this difficulty, Ramirez-Ruiz & Socrates (2005) proposed that high energy neutrinos from hot corona above the accretion disk might enhance the efficiency of energy deposition. We elucidate the disk corona model in the context of hypercritical accretion flows. From the energy balance in the disk and the corona, we can calculate the disk and coronal temperature, Td and Tc, and neutrino spectra, taking into account the neutrino cooling processes by neutrino-electron scatterings and neutrino pair productions. The calculated neutrino spectra consist of two peaks; one by the neutrino emission from the disk and the other by that from the corona. We find that the disk corona can enhance the efficiency of energy release but only by a factor of 1.5 or so, unless the height of the corona is very small, H<<r. This is because the neutrino emission is very sensitive to the temperature of emitting region, and then the ratio Tc/Td cannot be so large.