Consider an isothermal galaxy in which a significant fraction fc of cooled gas remains as cold dusty clouds instead of rapidly forming stars. At the centre a black hole grows by accretion from the surrounding cold (and hot) gas. Assume that the nucleus also blows a wind of velocity vw which has a power Lw = LEdd. Eventually the wind becomes powerful enough to blow away the surrounding gas and so shut off the accretion and further growth to the black hole and spheroid. The Magorrian et al (1998) black-hole - spheroid mass relation can then be obtained (Silk & Ress 1998; Fabian 1999; Blandford 1999).
The kinetic power of a wind at which it ejects cold gas of column density NH from a spheroid is given by
or
where is the velocity
dispersion within the spheroid. (I have used a force argument here,
see Fabian 1999;
Silk & Rees 1998
use an energy argument to obtain a
limit of 5 /
G, which is a factor
/ vw smaller
than the above Lw.) Ejection occurs when
Using
the Faber-Jackson relation for spheroids (Msph
4)
then yields, if vw/c
fc/ ~ 1
close to the
Magorrian et al
(1998) relation.
At that point the column density in to the accretion radius
NH ~
NT = T-1, so the growth is (just)
Compton thick. The growth
of massive black holes is radiatively efficient, highly obscured and
gives rise to much of the XRB. It is also intimately linked with the
growth of galaxy spheroids, the main evolution of which is terminated
by a quasar wind. X-ray observations probe best the underlying
obscured nucleus at (rest frame) ebergies of about 30 keV. Indeed
X-rays are the best diagnostic of the black hole accretion history of
the Universe.
The optically bright quasar phase (from an outside observer's point of
view) follows over the next few million years as the accretion disc
around the black hole empties. The early phase as the wind clears the
gas away can be identified with BAL quasars. The central engine is
only revived after the quasar phase if a merger or other event brings
in sufficient low angular momentum gas to fuel it.
Figure 3. The observed 2-10 keV flux as a
function of redshift from a
source of intrinsic (unabsorbed) 2-10 keV luminosity of
1045 erg s-1 with a column density of
1024.5 cm-2.
Scattering fractions (by thin ionized gas) of 5, 1, 0.1 and 0.01 per
cent are included (top to bottom). Note that the negatice K-correction
means that sources at z ~ 0.1, 0.8 and 7 can have the same observed
2-10 keV flux. From
Wilman & Fabian
(1999).
The prospects of testing the above scenario and absorption models of
the XRB are close at hand, with Chandra and XMM. They should detect
large numbers of faint, but powerful absorbed sources in the 3-10 keV
band, due to the negative K correction involved (see
Fig. 3) and
identify them with luminous FIR/sub-mm-emitting young galaxy spheroids.
I am grateful to Kazushi Iwasawa, Paul Nulsen and Richard Wilman for
continued collaboration and the organisers for an interesting
conference. The Royal Society is thanked for support.
Acknowledgements