Work done to date on the SMBH mass function has brought to light some interesting puzzles, the best known of which is cosmic downsizing. The solution to this problem is based upon the methodology by which we determine SMBH masses, which comes mainly from studies of active galaxies. The most massive SMBHs cease to be active at earlier epochs than their less massive counterparts, thus leading to an apparent decrease in the SMBH mass function with time. The larger SMBHs should still be present in the local Universe, but they lie in quiescent galaxies. This lack of certainty reflects the biggest gap in the observational record – the mass of SMBHs in quiescent galaxies.

This review summarizes the important properties of active black holes (BHs) up to z ~ 2; their mass, accretion rate, and growth rate. At higher redshifts, such information is only available for small samples that do not represent the entire population of active galactic nuclei (AGNs). Black hole spin is still unknown; it is speculated to change with redshift, but with little experimental evidence. The available data sets also enable a direct comparison of BH accretion rates and host galaxy star-formation rates (SFRs). The ratio of the BH growth rate g(BH) and the bulge growth rate g(bulge), suggests that the two are proportional to each other. The local value of g(bulge)/g(BH) in low-luminosity AGNs is of order 100 and the corresponding ratio in high-luminosity, high-redshift AGNs is of order 10. This has important implications regarding the parallel evolution of active BHs and their hosts.

We present results from statistical analysis of the largest quasar surveys currently in existence: the Sloan Digital Sky Survey (SDSS), the 2dF QSO Redshift Survey (2QZ), and the 2dF-SDSS LRG and QSO (2SLAQ) Survey. The quasar luminosity function shows highly significant (>99.9%) evidence of downsizing: the number density of faint quasars peaks at lower redshift than bright quasars. We measure the distribution of quasar broad-line widths as a function of luminosity, particularly focusing on the Mg ii λ2799 line. There is a clear trend, such that more luminous quasars show less scatter in line width than fainter quasars. However, the scatter in line width for any given luminosity (particularly in the brightest objects) is so small that it challenges our current understanding of virial black hole mass estimates.

Over the last few years, the existence of mutual feedback effects between accreting supermassive black holes powering AGN and star formation in their host galaxies has become evident. This means that the formation and the evolution of AGN and galaxies should be considered as one and the same problem. As a consequence, the search for, and the characterization of the evolutionary and physical properties of AGN over a large redshift interval is a key topic of present research in the field of observational cosmology. Significant advances have been obtained in the last ten years thanks to the sizable number of XMM–Newton and Chandra surveys, complemented by multiwavelength follow-up programs. I will present some of the recent results and the ongoing efforts (mostly from the COSMOS and CDFS surveys) aimed at obtaining a complete census of accreting black holes in the universe, and a characterization of the properties of the host galaxies.

I present mass functions of actively accreting black holes detected in different quasar surveys which in concert cover a wide range of cosmic history. I briefly address what we learn from these mass functions. I summarize the motivation for such a study and the methods by which we determine black hole masses.

We revisit a basic model of quasar activation by major mergers of dark matter halos (with “galactic” masses of ≲1013M⊙h−1). This model usually consists of two main ingredients: the halo merger rate describing triggering, and a quasar light curve, which describes the evolution of individual quasars. We show how the matching between model predictions and a variety of new, independent data sets allows one to efficiently constrain several aspects of black hole growth and evolution that must be taken into account in future studies by more advanced models of galaxy formation. Our results can be summarized as follows: (1) A descending phase modelled such that quasars in more massive halos shut down faster than those in less massive ones allows a good description of the bright end of the AGN luminosity function at all epochs and is compatible with downsizing, with more massive galaxies shutting down star formation earlier. (2) We measure the average bias of type 2 AGNs in SDSS to be b = 1.233 ± 0.195, independent of luminosity in the range 42.5≤ log L(erg s−1) ≤ 45.5. Such a value of the bias implies that faint AGNs at z<0.3 are mainly hosted by halos more massive than ~1011.5–12M⊙h−1. The black hole mass function predicted by this model is flatter than previously found. (3) The high clustering signal measured at z>3 in SDSS forces successful models to be characterized by rather short delay times of tdelay≲108 yr between the triggering and the shining epochs, implying massive “seed” BHs ≳ 105M⊙h−1 and initial super-Eddington growth. (4) The low number counts of X-ray AGNs measured in recent deep surveys are better reproduced by models with a minimal post-peak phase and a higher minimum hosting halo mass at high redshifts. (5) Cross-correlating the feedback-constrained MBH–M relation, with the redshift-dependent Mstar–M relation obtained from the cumulative number-matching of the stellar and halo mass functions, we find a factor of ~2 larger BH-to-stellar mass ratio at high redshifts. We discuss the meaning of such trends in connection with the mild, positive evolution in the MBH-σstar relation, and the strong observed evolution in the sizes and velocity dispersions of their hosts.

We performed an analysis of the mid-infrared properties of the 12 μm Seyfert sample, a complete unbiased 12 μm flux limited sample of local Seyfert galaxies selected from the IRAS Faint Source Catalog based on low-resolution spectra obtained with the Infrared Spectrograph (IRS) on-board Spitzer Space Telescope. A detailed presentation of this analysis is discussed by Wu et al. (2009). We find that, on average, the 15–30 μm slope of the continuum is 〈 α15–30〉 = −0.85 ± 0.61 for Seyfert 1s and −1.53 ± 0.84 for Seyfert 2s, and there is substantial scatter in each type. Moreover, nearly 32% of Seyfert 1s, and 9% of Seyfert 2s, display a peak in the mid-infrared spectrum at 20 μm, which is attributed to an additional hot dust component. The polycyclic aromatic hydrocarbon (PAH) equivalent width decreases with increasing dust temperature, as indicated by the global infrared color of the host galaxies. However, no statistical difference in PAH equivalent width is detected between the two Seyfert types of the same bolometric luminosity. Finally, we propose a new method to estimate the AGN contribution to the integrated 12 μm galaxy emission, by subtracting the “star formation” component in the Seyfert galaxies, making use of the tight correlation between PAH 11.2 μm luminosity and 12 μm luminosity for star forming galaxies.

In this work, we use the low-redshift, narrow-line Seyfert 1 (NLS1) galaxies Ark 564 and I Zw 1 as laboratories for modeling the Fe ii emission features in the wavelength range 1200–6700 Å. We utilize data acquired with the Hubble Space Telescope and the 2.7-m telescope at McDonald Observatory.

We present clustering results for AGN jointly selected to have optical and X-ray data at z ~ 3 in the ECDF-S field. Using Chandra X-ray imaging and UVR optical colors from MUSYC photometry, we selected a sample of 58 z ~ 3 AGN candidates. From the optical data we also selected 1385 Lyman break galaxies (LBG) in the range 2.8 < z < 3.8 with R < 25.5 mag. From optical spectroscopy, we have also determined redshifts and AGN types and estimated the UV and X-ray luminosities for the set of active galaxies selected by this method. We performed autocorrelation and cross-correlation analyses and present here our results for the clustering amplitudes and dark matter (DM) halo masses of each sample. The clustering analysis yields median DM halo masses of log M/M⊙ = 11.9±0.2 for the LBG sample, and log M/M⊙ = 12.9+0.4−0.9 for the sample of AGN. This shows that the active supermassive black holes (SMBHs) targeted by this selection method tend to sit in galaxies more massive than the typical population of UV–continuum selected galaxies, although showing much less clustering than SDSS quasars at the same redshift. Additionally, the expected evolution of the DM halos in ΛCDM cosmology implies that today these z ~ 3 SMBHs are hosted by group-sized halos, with masses log M/M⊙ = 14.1+0.3−0.2, at the high-mass end of the distribution of LBG DM halo descendants at z = 0.

The coalescence of black hole binaries is a significant source of gravitational wave radiation. The typically asymmetric nature of this emission, which carries linear momentum, can result in the recoil of the black hole remnant with velocities in the range 100 < Vrecoil < 3750 km s−1. The detectability of recoiling massive black holes (MBH) as off-nuclear QSOs is tightly connected with the properties of the host galaxy, which determine the MBH's orbit and fuel reservoir. We present the results of N-body simulations of recoiling MBHs in high-resolution, non-axisymmetric potentials. We find that if the recoil velocities are high enough to reach regions of the galaxy dominated by the generally triaxial dark matter distribution, the return time is significantly extended when compared to a spherical distribution. We also perform simulations of recoiling MBHs traveling in gas merger remnants, where large amounts of gas have been funneled to the central regions, In this case, the MBHs remain within R<1 kpc from the center of the host even for high recoil velocities (Vrecoil = 1200 km s−1) due to the compactness of the remnant galaxy's nuclear disk. We discuss the implications of both scenarios for detectability.

We present the first estimate of the black hole mass function (BHMF) of broad-line quasars (BLQSOs) that self-consistently corrects for incompleteness and the statistical uncertainty in the mass estimates, based on a sample of 9886 quasars at 1 < z < 4.5 drawn from the Sloan Digital Sky Survey. We find evidence for “cosmic downsizing” of black holes in BLQSOs, where the peak in their number density shifts to higher redshift with increasing black hole mass. We estimate the lifetime of the BLQSO phase to be 70 ± 5 Myr for supermassive black holes (SMBHs) at z = 1 with a mass of MBH = 109M⊙, and we constrain the maximum mass of a black hole in a BLQSO to be ~ 1010M⊙. We find that most BLQSOs are not radiating at or near the Eddington limit. Our results are consistent with models for self-regulated black hole growth, where the BLQSO phase occurs at the end of a fueling event when black hole feedback unbinds the accreting gas.

The recently completed AT20G survey provides the largest and most complete sample of high-frequency selected radio sources yet obtained, and offers new insights into the nature of the high-frequency active galaxy population. Here we focus on the optical properties of this survey which highlights the difference of the AT20G source population compared to other radio-selected AGN samples.

We demonstrate that quantifying the intrinsic variability of quasars by fitting individual structure function data pairs with a 2-parameter power law model separates quasars from contaminating variable and non-variable point sources with a completeness of 93% and a purity of 99%. This approach can be used to select quasar samples in surveys like that being performed by Pan-STARRS1, where the usual color selection of quasars is not possible due to a filter system that is too red.

The observed relations between the black hole mass and the properties of the spheroidal galaxy component imply a close connection between the growth of supermassive black holes and the evolution of their host galaxies. An effective approach to study black hole growth is to measure black hole masses and Eddington ratios of well-defined type 1 AGN samples and determine the underlying distribution functions.

We present the results of a cross-correlation analysis of the projected positions of AGNs and galaxies at redshifts from 0.3 to 3.0. It is widely accepted that the origin of AGN activity is accretion of matter onto a massive black hole at the center of a galaxy (e.g., Lynden-Bell 1969). To explain the activity of AGNs, a large fraction of matter in the galaxy must be delivered to the inner region on a short timescale (Hopkins et al. 2008). One possible mechanism for causing rapid gas inflows into the central region is a major galaxy merger between gas-rich galaxies (e.g., Kauffmann & Haehnelt 2000). If this is the case, AGNs are expected to be found in an environment with higher galaxy density than that of typical galaxies. We investigated environments of ~ 750 AGNs, which is about a ten times larger sample than used in previous studies, and we find a significant excess of galaxies around the AGNs in the redshift range of 0.3 to 1.8. We used the Japanese Virtual Observatory (JVO) to obtain the Subaru Suprime-Cam images and UKIDSS data around known AGNs. The datasets accessed through the JVO are: Catalog of Quasars and Active Galactic Nuclei by Veron-Cetty et al. (2006), SDSS DR-5 Quasar Catalog by Schneider et al. (2007), Subaru Suprime-Cam Reduced Image Archive of JVO, and UKIDSS DR2 catalog by Warren et al. (2007). We divided all the AGN samples into four redshift groups, 0.3≲z ≲ 0.8, 0.8≲ z ≲ 1.5, 1.5≲ z ≲ 1.8, and 1.8≲ z ≲ 3.0. For each redshift group, the dataset was further divided into a fainter group (MV ≥ −25 mag) and a brighter group (MV < −25 mag). We found that the correlation length of the high-redshift bright sample (1.5≲ z ≲ 1.8) was larger than that of the low-redshift faint sample (0.3≲ z ≲ 0.8). We also found that the correlation length was larger for the faint group at redshift range 0.8≲ z ≲1.5. These results can be explained by downsizing of mass assembly. More details can be found in the paper by Shirasaki et al. (2009). Our result implies that the Japanese Virtual Observatory can be a powerful tool to investigate the co-evolution of central black holes and galaxies at the intermediate redshift universe.

The Sloan Digital Sky Survey has made it possible to identify the first samples of active galaxies with estimated black hole masses below ~ 106M⊙. We have obtained Spitzer IRS low-resolution spectra, covering 5–38 μm, of a sample of 41 Seyfert galaxies with low-mass black holes. Our sample includes SDSS-selected objects from the low-mass Seyfert 1 sample of Greene & Ho (2004) and the low-mass Seyfert 2 sample of Barth et al. (2008), as well as NGC 4395 and POX 52. The goals of this work are to examine the dust emission properties of these objects and investigate the relationship between type 1 and type 2 AGNs at low luminosities and low masses, to search for evidence of star formation, and to use emission-line diagnostics to constrain physical conditions within the narrow-line regions. Here we present preliminary results from this project.

Our research addresses the observed evolution of MBH and L/LEdd of type 1 radio-quiet AGNs. Measurements of MBH and L/LEdd in a sample of sources at z<0.75 using the “Hβ method” (Netzer & Trakkhtenbrot 2007) as well as new measurements of these properties at 0.75<z<2 using the “Mg ii λ2798 method”) show significant evolution of L/LEdd for any value of MBH up to z = 2 (see Figure 1). In a dedicated near-IR project we obtained the most reliable sample of MBH and L/LEdd estimates for z ~ 2.3 and z ~ 3.4 AGN (Netzer et al. 2007; Shemmer et al. 2004); the distribution of L/LEdd is very broad (Figure 1) and ~ 1/2 of the sources have L/LEdd<0.2, implying tgrowth>tuniverse and hence an earlier epoch of fast growth. Our on-going Gemini–VLT campaign focuses on z ~ 4.8 sources, with the Mg ii line observed in the H-band (Figure 2). The analysis of ~ 1/3 of the spectra uncovers, again, a broad distribution of L/LEdd, with several sources having L/LEdd ≫ 1, suggesting an early episode of fast BH growth.

Galaxy mergers play an important role in many astrophysical processes, such as growth of massive galaxies, triggering AGN, formation of supermassive black hole (SMBH) binaries, and gravitational wave (GW) radiation. Merger rate is one of key quantities for these studies. Previous studies show that the pair fraction varies in a range of 1%–10% in the redshift range of z = 0.2–1.2. These merger rates are usually calculated from projected close pairs, and very few previous authors have carefully checked the merging fraction of a large sample of pairs.

I review accretion and outflow in active galactic nuclei. Accreti4on appears to occur in a series of very small-scale, chaotic events, whose gas flows have no correlation with the large-scale structure of the galaxy or with each other. The accreting gas has extremely low specific angular momentum and probably represents only a small fraction of the gas involved in a galaxy merger, which may be the underlying driver.

Eddington accretion episodes in AGN must be common in order for the supermassive black holes to grow. I show that they produce winds with velocities v ~ 0.1c and ionization parameters implying the presence of resonance lines of helium-like and hydrogen-like iron. The wind creates a strong cooling shock as it interacts with the interstellar medium of the host galaxy, and this cooling region may be observable in an inverse Compton continuum and lower-excitation emission lines associated with lower velocities. The shell of matter swept up by the shocked wind stalls unless the black hole mass has reached the value Mσ implied by the M–σ relation. Once this mass is reached, further black hole growth is prevented. If the shocked gas did not cool as asserted above, the resulting (“energy-driven”) outflow would imply a far smaller SMBH mass than actually observed. Minor accretion events with small gas fractions can produce galaxy-wide outflows, including fossil outflows in galaxies where there is little current AGN activity.

Our adaptive optics observations of nearby AGN at spatial resolutions as small as 0.″085 show strong evidence for recent, but no longer active, nuclear star formation. We begin by describing observations that highlight two contrasting methods by which gas can flow into the central tens of parsecs. Gas accumulation in this region will inevitably lead to a starburst, and we discuss the evidence for such events. We then turn to the impact of stellar evolution on the further inflow of gas by combining a phenomenological approach with analytical modelling and hydrodynamic simulations. These complementary perspectives paint a picture in which all the processes are ultimately regulated by the mass accretion rate into the central hundred parsecs, and the ensuing starburst that occurs there. The resulting supernovae delay accretion by generating a starburst wind, which leaves behind a clumpy interstellar medium. This provides an ideal environment for slower stellar outflows to accrete inwards and form a dense turbulent disk on scales of a few parsecs. Such a scenario may resolve the discrepancy between the larger scale structure seen with adaptive optics and the small-scale structure seen with VLTI.