1999ApJ...514..691K

Time-dependent, axisymmetric hydrodynamic simulations have been used to study accretion disks consisting of counterrotating components with an intervening shear layer(s). Configurations of this type can arise from the accretion of newly supplied counterrotating matter onto an existing corotating disk. The grid-dependent numerical viscosity of our hydrocode is used to simulate the influence of a turbulent viscosity of the disk. First, we consider the case where the gas well above the disk midplane (z>0) rotates with angular rate+Ω(r) and that well below (z<0) has the same properties but rotates with rate-Ω(r). We find that there is angular momentum annihilation in a narrow equatorial boundary layer in which matter accretes supersonically with a velocity that approaches the free-fall velocity. This is in accord with the recent analytic model of Lovelace & Chou. The average accretion speed of the disk can be enormously larger than that for a conventional α-disk rotating in one direction. Under some conditions the interface between the corotating and counterrotating components shows significant warping. Second, we consider the case of a corotating accretion disk for r<r_t and a counterrotating disk for r>r_t. In this case we observed, that matter from the annihilation layer lost its stability and propagated inward, pushing matter of inner regions of the disk to accrete. Third, we investigated the case where counterrotating matter inflowing from large radial distances encounters an existing corotating disk. Friction between the inflowing matter and the existing disk is found to lead to fast boundary layer accretion along the disk surfaces and to enhanced accretion in the main disk.

These models are pertinent to the formation of counterrotating disks in galaxies and possibly in active galactic nuclei and in X-ray pulsars in binary systems. For galaxies, the high accretion speed allows counterrotating gas to be transported into the central regions of a galaxy in a time much less than the Hubble time.