 |  |
Astron. Astrophys. 321, 665-671 (1997)
3. Stability analysis
The ambiguity of the shock locations and
was first removed by Chakrabarti & Molteni
(1993) for adiabatic flows around a non-rotating black hole in the
context of Paczynski & Wiita (1980) potential. By checking the
variation of the total pressure of the post-shock flow as the shock is
perturbed, they argued that the shock at is
stable for accretion, and unstable for winds, but did not give a
definite conclusion about the shock at . More
recently, Yang & Kafatos (1995) made a fully relativistic analysis
on the stability of shocks in isothermal accretion flows onto a
Schwarzschild black hole, and found that the shock at
is unstable, while the shock at
is stable except when is
very close to the point at which . Here we
follow Yang & Kafatos' (1995) method, extending to the case of
adiabatic accretion and wind flows in Kerr geometry.
Across the shock, the momentum flux density, defined as (different
from only by a constant factor)
![[EQUATION]](img84.gif)
is conserved, resulting exactly in Eq.(2e). If due to some
perturbation, the shock location is moved from
to , the momentum flux density may not be in
balance, the resulting difference across the shock is
![[EQUATION]](img87.gif)
The stability of the shock depends on the sign of
If , for accretion flows
(i.e. along decreasing r), when (or
), the momentum flux just after the shock is
larger (or smaller) than that just before the shock, so the shock
should be put to shift towards further increasing (or further
decreasing) r, thus the shock is unstable; but for wind flows
(i.e. along increasing r) the shock is stable, because the
imbalance of the momentum flux due to would
always cause the shock to move back towards its unperturbed location.
On the contrary, implies that the shock in
accretion flows is stable, and the shock in winds is unstable.
The present case is mathematically more complex than that studied
by Yang & Kafatos (1995), so that it is not possible to obtain an
analytic criterion of the sign of , and the only
way to estimate is that through numerical
calculations. It is seen from Eqs (5) and (2) that to evaluate
, the values of and
are needed, which can be solved out by
combining the differential form of Eqs (2a) and (2c), i.e.
and ; similarly, the
values of and , needed
to evaluate , can be solved out by combining
the differential form of Eqs (2b) and (2d), i.e.
and . In this way, we
have performed calculations for the Schwarzschild black hole and the
Kerr hole, as well as for prograde flows and retrograde ones, and have
been able to reach the following conclusion: for accretion flows
the shock at is always stable, while the shock
at is unstable except when the value of
l is very close to that of ; for
wind flows the shock at is always unstable,
while the shock at is stable except when
l is very close to . It is seen
from Fig. 2 that when l is close to
the values of and
become close to each other. To show more
precisely the exceptional case when l is very close to
, we give a numerical example: for the black
hole's specific angular momentum and prograde
accretion flows (cf. Fig. 1b), when the
corresponding is slightly smaller than 2.0754,
in the numerical calculation when the parameter l is taken to
be 2.0754 both the shocks at and
are stable (in fact and
are so close to each other now, so that it is
practically not possible to distinguish them), but when
the shock at becomes
unstable, and the value 2.08 is still not the lower limit of l
to which the resulting keeps being an unstable
shock location.
© European Southern Observatory (ESO) 1997
Online publication: June 30, 1998
helpdesk.link@springer.de  |