3 Summary

We have presented a multi-layer model for initiating ion-jets in AGNs. The model agrees with a previous numerical study which confirmed the formation of thermally-induced outflows in the transition layer (Hujeirat & Camenzind 2000). Here we have shown that incorporating the effects of MFs manifests their formation and dramatically strengthen their dynamics. Three ingredients for initiating winds have been detected: 1) a highly diffusive plasma dominated by virial-hot ions, 2) large scale magnetic fields that efficiently transport angular momentum from the disk into the TL, where the plasma rotates super-Keplerian, and 3) an underlying advection-dominated accretion disk.

Taking into account that the corona is dynamically unstable, adopting a large scale magnetic topology, and allowing ion-electron thermal decoupling appear to force accretion flows to undergo a global energy re-distribution: confined inflows (negative Bernoulli number) in the equatorial region and in the corona, and thermally and magneto-centrifugally-driven outflows in the TL characterized through a positive Bernoulli number. This feature may survive under different conditions: strong MFs suppress turbulence, weakening thereby the effect of the turbulent-viscosity and dominate the transport of angular momentum. On the other hand weak MFs in rotating stratified flows would be amplified via dynamo-action and reach equipartition, beyond which turbulence is again suppressed. This interplay between MFs and turbulent-viscosity, Balbus-Hawley and Parker instabilities may settle into an equilibrium state, in which inflows are simultaneously associated with low-cooling out-flowing ion-plasma.

Figure 4: 40 equally-spaced isolines of the poloidal-component $B_{\rm P}$ (red lines) and 30 equally-spaced isolines of $B_{\rm T}$ (black lines) superposed on the density-distribution (yellow color corresponds to very high density-values, blue to middle and violet to low values). $B_{\rm T}$ attains it maximum values in the innermost part of the transition layer and diminishes in the corona.

Figure 5: A snap-shot of the distribution of the Bernoulli number in two-dimensional time-dependent calculations. The decrease from large to low positive values is represented via yellow, green and red colors. The blue color corresponds to negative values. The figure shows two ejected blobs of large positive energies in the TL. Ejection starts from the innermost region of the TL and evolves non-linearly.

We note that in the absence of thermal conduction and adopting the one-temperature description, low-viscosity radiatively inefficient HD and MHD accretion flows become inevitably convection-dominated. Therefore, in the early phases of jet-initiation, CDAFs may play an important role in powering the jets in AGNs and microquasars (Abramowicz et al. 2002).

The multi-layer model presented here accommodates some elements of BP82. In particular, we agree with BP82 about the necessity for a super-Keplerian rotation of the plasma overlying the accretion disk. However, the plasma here is dominated by highly-diffusive and virial-hot ions; it does not require a special ${B_{\rm P}}-$alignment with respect to the disk-normal to enable jet-launching, as ideal-MHD treatment requires. While our results agrees with the ion-torus model with respect to the necessity of 2T-plasma to maintain the ions hot for a significant time of their propagation-life in the ISM, no signatures for the formation of ion-supported tori have been detected (Rees et al. 1982).

Our results differ from ADAF and ADIOs in several issues, and in particular with respect to 1) the existence of a layer adjusting to the disk, where the plasma is found to rotate super-Keplerian, 2) the configurations of the in- and the out-flows, 3) stability of the corona in the vicinity of the BH, 4) the transition from SS-disk to advection-dominated disks and 5) with respect to the essence of Bernoulli number in dissipative flows, i.e., a positive Bernoulli number is necessary but not sufficient for outflows (Abramowicz et al. 2000).

Finally, we note that since the flow in the TL is highly dissipative (strengthen thermal and rotational coupling with the central nucleus), the innermost region of the disk rotates synchronously with Kerr black holes (due to the frame dragging effect), and since ${\tau_{\rm rem}}$ decreases with radius and depends inversely on  $B_{\rm T}$ and  $B_{\rm P}$, we think that the plasma attached to the poloidal magnetic field would be forced to deposit its angular momentum to the plasma in the TL, thereby considerably enhancing the centrifugal power and ejecting the ion-plasma into space with relativistic speeds.