It is generally assumed that the first objects to form in the universe were stars. However this is by no means assured. There is general agreement among theorists that, if cosmic structures form hierarchically (`bottom up'), as in cold dark matter (CDM) models, the first baryonic clouds have masses in the range and that the characteristic mass subsequently rises. But it is actually not obvious that these clouds would undergo fragmentation into stars. Conditions in primordial clouds differ from those prevailing in conventional star-forming clouds that the fate of nearby molecular clouds is not a reliable guide. For example, in the absence of magnetic flux, cloud collapse may have been far more catastrophic than is the case at the present epoch. A massive disk could have rapidly shed its angular momentum via non-axisymmetric gravitational instabilities, and become so dense and opaque that it continued to evolve as a single unit. At very high redshifts, the inefficiency of atomic and molecular cooling via H and H2 excitations is compensated by Compton cooling; Compton drag provides an additional mechanism for transferring angular momentum and allowing collapse.
This outcome seems no less likely, a priori, than the alternative evolutionary pathways found in the literature, according to which primordial clouds fragment into stars with an initial mass function that varies between being bottom-heavy, top-heavy or even normal (that is, solar neighbourhood-like), depending on the observations that are being interpreted. The quasar distribution tells us directly that at least some massive black holes form early. Indeed, the quasar comoving density peaks at , and only declines at (Shaver et al. 1996); on the other hand, the peak of galaxy formation occurs at (Madau et al. 1996; Connolly et al. 1997), although there is uncertainty about the effects of extinction in leading to an underestimate of the galaxy luminosity at high redshift.
In fact, the known quasars, or their dead counterparts, are likely to be within the cores of at least 30 percent of these galaxies. To see this, note that combining the integrated density in quasar light with the assumption that quasars radiate at or near the Eddington limit yields an estimate of typical dead quasar (or black hole) mass (Soltan 1982; Chokshi & Turner 1992) as Whether one actually could observe an AGN component in high redshift galaxies depends sensitively on the adopted lifetime of the active phase: higher redshift helps. The observed correlation between massive black holes and dynamically hot galaxies then suggests that most hot galaxies, amounting to of order a third of all galaxies in terms of stellar content, could contain such a massive black hole (Faber et al. 1996). One note of caution would therefore be that dynamically hot galaxies probably form systematically earlier than most galaxies, and that if starbursts characterize their birth, existing high redshift samples of such objects may be incomplete.
Nevertheless, while the case remains ambiguous, we are sufficiently motivated by the possible implications of a causal connection between quasars and galaxy formation to explore in this note the consequences of a cosmogonical scenario in which the first objects to form, at some highly uncertain efficiency, are supermassive black holes. We discuss how, during subsequent mergers, the holes could grow, and exert a feedback on star formation.
© European Southern Observatory (ESO) 1998
Online publication: February 16, 1998