8. Viscous evolution of advective disk
As we know, the structure of an accretion disk in the vertical direction, the relation between the viscous tensor and the surface density in particular, defines the type of its temporal evolution. In advective disks, which are the low-radiative accretion flows, the relations between their characteristic physical parameters differ significantly from those in standard disks. In this section, we discuss the results of Sect. 2 as applied to the disks which radial structure was presented by Spruit et al. (1987) and Narayan & Yi (1994, 1995, hereafter NY).
where is the averaged kinematic coefficient of turbulent viscosity. Then the relation between and is given by
Recall that is the real specific angular momentum and h is the Keplerian one. It can be seen that and define what class of solutions Eq. (5) will have.
If one adopts for the structure of advection-dominated accretion flow (ADAF) the self-similar solution by NY, it can be easily inferred that such disks exhibit the exponential with time behaviour. The solution of NY is given by:
The exponential temporal behaviour of NY flow is evident. Generally speaking, any disk possessing such properties of as constancy in time would have such exponential behaviour because its evolution would be described by a linear equation (like (52)).
The question is, would the confined NY disk keep such properties or it would not. The fact is that NY solution describes the infinite disk. Either the boundary conditions destroy the linearity of (52) or just the characteristic decay time changes, this problem requires further accurate numerical investigation. For instance, Narayan et al. (1997) calculated numerically the global structure of stationary advection-dominated flow with consistent boundary conditions; they noted that although the self-similar solution (48) makes significant errors close to the boundaries, it gives the reasonable description of the overall properties of the flow.
Further we assume that exponential trend of solution persists. Generally speaking, the equation determining will differ from (54). This difference may be not very significant. One can see that Eq. (54) is a particular case of (22) where and and, hence, the solution can be found according to (24) and (23). Besides, the solution of (54) can be found in terms of Airy functions (Bessel functions of order ).
The value of accretion rate can be determined if an initial condition is imposed at some t. Mahadevan (1997) showed that ADAF luminosity or according to whether the electron heating is dominated by the Coulomb interactions or by the viscous friction. Subsequently, the luminosity has an exponential decay too.
We use the expressions for and from NY. Here is the ratio of specific heats; measures the efficiency of radiative cooling. In the limit of no radiative cooling, we have while in the opposite limit of very efficient cooling . NY solution is degenerate if because the angular velocity of the flow is zero in this case.
We can see that the time-dependent advection-dominated disk is quickly depleted if is not small. For example, consider the light curves of X-ray novae which have the exponential decay time scales . To obtain of such order, should be . However, the advection-dominated solution ceases to exist if the accretion rate is greater than the critical value (Narayan & Yi 1995, Mahadevan 1997), where . Hence, yields the critical accretion rate g s.
© European Southern Observatory (ESO) 2000
Online publication: March 28, 2000