The Growth of Supermassive Black Holes Across Cosmic Time

A major recent development in astrophysics has been the discovery of the relationship in the local Universe between the properties of galaxy bulges, and dormant black holes in their centers (Magorrian et al. 1998;Ferrarese & Merrit 2000;Gebhardt et al. 2000). These tight relationships represent definitive evidence for the co-evolution of galaxies and Active Galactic Nuclei (AGN). The remarkable implication of this is that some consequence of the accretion process on the scale of the black hole event horizon is able to influence the evolution of an entire galaxy. The main idea is that radiative and mechanical energy from the AGN regulates both star formation and accretion during periods of galaxy growth. This kind of black-hole driven feedback is thought to be essential in shaping the first galaxies. Current models propose that mergers of small gas-rich protogalaxies in deep potential wells at high redshift drive star formation and black hole growth (in proto-quasar active galaxies) until a luminous quasar forms. At this point, a black-hole driven wind evacuates gas from the nascent galaxy, limiting additional star formation and further black hole growth (Silk & Rees 1998; Fig. 1). Further episodes of merger-driven star formation, accretion, and feedback are expected to proceed through cosmic time. This provides a plausible origin for the M-! relation (e.g. King 2003), and explains many outstanding problems in galaxy evolution (e.g. Croton et al. 2005;Hopkins et al. 2006). Despite the intense current interest in this topic, and its great importance, direct evidence for widespread AGN feedback at high redshift is scarce and the details of the physical processes are unclear. It is thus certain to remain one of the most active topics in astrophysics during the next decade and beyond.

The very first stars formed in primordial structures where gravity was able to overpower the pressure of the ambient baryons, some hundred million years after the Big Bang. The first seed black holes (100 M ! ) are left behind as remnants of the most massive stars. The first galaxies, hosting these first black holes in their cores, are responsible for reionizing the Universe by z10, as shown by WMAP. Still, the highest redshift galaxies and quasars currently known are all in the range z=6-7. To understand the inner workings of the first luminous sources we need to bridge the gap between the few known sources at this redshift, and the information we can extract from the microwave background.

Figure 1: Formation of a high-redshift quasar from hierarchical galaxy mergers as simulated by Li et al. (2007). Color shows gas temperature, and intensity shows gas density. Black dots represent black holes. Small, gas-rich galaxies merge in the deepest potential wells at high redshift, promoting star formation and black hole growth. At z ~ 7 to z ~ 5 a luminous quasar forms, associated with the most massive black hole. It drives a wind (yellow) that evacuates gas from the nascent galaxy.

The known AGN population at z=6-7 consists of luminous optical quasars (e.g. Fan et al. 2003). Growing the extremely massive black holes required in <1 Gyr represents a challenge for theoretical models, because it requires Eddingtonlimited accretion over many folding times. Recent gas-dynamical cosmological simulations are nevertheless able to produce quasars with ~10 9 M ! at z=6.5 through a rapid sequence of mergers in small groups of proto-galaxies (Li et al. 2007; Fig. 1). The growth is likely to proceed in a self-regulated manner owing to feedback with the progenitor host, with a period of intense star formation and obscured accretion preceding the optically bright quasar phase. The complex physics involved in such a scenario is, however, poorly understood.

It must also be borne in mind that these luminous QSOs, hosting among the most massive black holes (>10 9 M ! ) in the Universe, a

…(Full text truncated)…