With the results obtained above from the behavior of the broad-line
components we will try to derive the structure of the BLR in some detail, under
the assumption that the gas producing these components is gravitationally
bound. Then the mass of the central object (the black hole in the standard
theory) can be calculated, as usual, from , where v is the
gas velocity and r its distance to the object. We find the distance can be
estimated from the delays obtained with the correlation functions in
combination with the distances deduced from the central photoionization
models. To determine the velocity we assume that the main mechanism of
line-broadening is macroscopic turbulent movement, implying that
the mean velocity is
.
For the central component this gives
. Its
distance to the
central object is more difficult to establish, due to the fact that the one
deduced from the CCF,
, is twice as large as
that from the photoionization models of
. This
could indicate some special geometry with respect to the observer, which will
be discussed below. For the current discussion we will use a distance of
, obtaining
, equal to the estimated one by Clavel et al. (1989). The
similarity in these results is not too surprising since we find a smaller
distance but the turbulent velocity of the individual component is compensating
for this.
If we now use this value of M and the FWHM for the blue and the red
components, distances of 75 and 77 light-days, respectively, are obtained.
This distance, equal for both components, is
significantly different from the delay calculated from the correlation
functions,
and
, respectively. This can be
explained if infalling motion of the gas at a distance of
is considered, because then the red flux does not lag behind the
continuum variations and the blue flux will be delayed by twice the light travel
time (Koratkar &
Gaskell 1991). Koratkar & Gaskell, applying cross correlations to
the wings of the Fairall-9 lines, obtain for CIV
and
for the red and the blue flux, respectively, concluding
that the gas producing these fluxes is at a distance of
and has an infall movement, compatible with our result. However,
they
obtain for
that both wings are delayed with respect to the continuum
, suggesting that the movement is chaotic or circular. We
have not been able to analyze the wings of
because the correlation
functions are not significant due to the small variability amplitude of the
flux with the continuum.
Supposing that the red and the blue components of
Ly, CIV and SiIV are produced in the same zone of gas at
and is infalling, there is a
problem if this region is spherically symmetric.
In that case we should obtain central gas at the same distance,
while we find that the central component originates further away. The
absence of this central gas indicates a non-spherically symmetric
or anisotropic continuum emission, so that the majority of the gas
responsible for the red and the blue flux is along the line of
sight of the observer (the angular extend of course limited by the time resolution of
our data). Besides, the central component is produced
at a larger distance,
, explaining
the derived twice larger delay if this region of gas is
behind the source, suggesting that either the symmetry is non-spherical or the
continuum emission is anisotropic, or both. The absence of central gas between the
source and
the observer could, in this context, be explained as due to the
presence of dust in the non-illuminated or neutral part of the
clouds, which would absorb its emission. Then the dust emitting
in the near IR at
(Clavel et al.\
1989) should extend between
and larger
distances.