Abstract
The cosmological evolution of supermassive black holes (SMBHs) seems to be intimately linked to their host galaxies. Active galactic nuclei (AGN) can be probed by deep X-ray surveys. We review results from large X-ray selected samples including first results from the XMM-Newton COSMOS survey. A new picture arises from the fact that high-luminosity AGN grow earlier than low-luminosity AGN. In particular, the space density of low-luminosity AGN exhibits a significant decline for redshifts above z=1. This 'anti-hierarchical' growth scenario of SMBHs can be explained by two modes of accretion with different efficiency. The population of Compton-thick sources plays a key role in our understanding of the BH growth history. Their space density and redshift distribution is relevant to estimate the SMBH mass function. A comparison with the relic SMBH mass distribution in the local Universe constrains the average radiative efficiency and Eddington ratio of the accretion. We discuss a new synthesis model of Compton-thin and Compton-thick sources that is in concordance with deep X-ray observations and in particular predicts the right level of contribution of the Compton-thick source population observed in the Chandra Deep Field South observations as well as the first INTEGRAL and Swift catalogues of AGN. Currently, one of the most important problems is the evolution of obscuration with redshift.
Introduction: Mapping the Growth of Giants
Understanding the life cycle of Supermassive Black Holes (SMBHs) and their co-evolution with host galaxies is a cornerstone of modern extragalactic astronomy. While "relic" black holes are observable in the local universe, tracing their growth phase requires observing Active Galactic Nuclei (AGN) across cosmic time. This review by Andreas Müller and Günther Hasinger synthesizes data from major X-ray observatories to map this evolution, addressing challenges such as the "spectroscopic desert" at redshifts $1 < z < 2$ and the prevalence of heavily obscured sources.
To overcome these hurdles, researchers utilize deep X-ray surveys from missions like XMM-Newton and Chandra. By employing the Hardness Ratio (HR)—the ratio of hard to soft X-rays—AGN can be classified as Type-1 (unobscured) or Type-2 (obscured) even when optical spectra are unavailable. These classifications, combined with Luminosity-Dependent Density Evolution (LDDE) models, allow for the construction of luminosity functions that track how AGN populations change over billions of years.
Anti-Hierarchical Growth and Cosmic Downsizing
A primary finding of recent X-ray surveys is the "anti-hierarchical" growth of black holes, often termed "cosmic downsizing." In contrast to standard dark matter models where small structures merge into larger ones, the most massive black holes reached their peak activity much earlier ($z \sim 2$) than their lower-luminosity counterparts, which peak around $z \sim 0.7$.
This trend is clearly visible in the evolution of the X-ray luminosity function. High-luminosity sources were already well-established by $z=3$, while the space density of low-luminosity AGN declines significantly above $z=1$. This suggests two distinct modes of accretion or different efficiencies governing the growth of SMBHs at different epochs.
The Role of Obscuration and the "Cleanout Effect"
A significant portion of the AGN population is "hidden" behind thick columns of gas and dust. Accounting for these obscured sources is essential for a complete census of black hole growth and for explaining the total energy observed in the Cosmic X-ray Background (CXB).
One key insight is the "cleanout effect": the fraction of obscured (Type-2) AGN decreases as luminosity increases. This physical phenomenon suggests that more powerful black holes generate sufficient radiation pressure to clear their immediate environments of obscuring material.
While obscuration is primarily a function of luminosity, there is also evidence suggesting a slight increase in the obscured fraction at higher redshifts. Population synthesis models, such as those by Gilli et al. (2007), now include a significant population of "Compton-thick" sources—the most heavily obscured AGN—to successfully reproduce the peak of the CXB at 30 keV.
Radiative Efficiency and Black Hole Spin
By comparing the energy output of the AGN population to the mass of relic black holes in the local universe, researchers can constrain the average radiative efficiency ($\epsilon$). A value of approximately 10% ($\epsilon \sim 0.1$) is required to match observations. This relatively high efficiency implies that most supermassive black holes are rapidly rotating (Kerr black holes), as higher spin allows for more efficient conversion of accreted mass into radiation.
Future Outlook
Despite significant progress, current surveys remain flux-limited, potentially missing the faintest obscured objects. The debate regarding the evolution of obscuration with redshift continues, requiring more complete and deeper samples. Future missions like eRosita and Athena are poised to probe deeper into the high-redshift universe and harder X-ray bands, potentially revealing the "seeds" of these giants at $z > 10$.
Summary updated