Starving accretion disks in the center of galaxies: the examples of Sagittarius A* and M 31.

Observations of the Galactic Center source SgrA* suggest the presence of a central hot accretion disk around a supermassive black hole with an accretion rate a factor 10^5^ below the Eddington limit. We argue that purely spherical accretion models have problems with obervational constraints. The low accretion rate inferred for SgrA* requires a distinct type of accretion disk - a starving disk. We investigate the structure of this type of disk within the framework of thin Keplerian disk models taking into account relativistic effects, detailed opacity tables and turbulent heat transport. It appears as if starving disks are the perfect realization of geometrical thin, optically thick disks with well behaved structure functions. Although in general gas-pressure dominated, radiation pressure may become important in the inner region of the disk. All in all starving disks are very stable structures with high surface density, high optical depth and long viscous time scales. However, S-curves in the {SIGMA}-˙(M) diagram occur, indicating a thermal limit-cycle behaviour observable in the IR-NIR with timescales ranging between three years and some thousand years. We propose that, at least in SgrA*, the main mass supply for the disk might be a wind from the surrounding star cluster. As the disk is a very stable and slowly evolving structure, it will inhibit spherical accretion and incorporate the inflowing matter in the general accretion flow. On the other hand self-gravitation and fragmentation of the outermost part of the disk could have resulted in star formation at 0.1 pc and hence even the wind of the star cluster can already have substantial angular momentum. The situation in M 31 may be very similar but with a more extreme set of parameters.