Supermassive Black Hole Evolution

In the past decade, studies of the local Universe have established the presence of supermassive black holes (SMBH) in the nuclei of virtually all galaxies with a bulge/spheroidal component, dramatically changing our perception of this class of objects, and implying a clear relationship between the growth of SMBH and that of the galaxy. Because the cosmological growth of SMBHs is mostly due to accretion of matter during active phases and the energy released in the process of accretion can be higher than the total binding energy of a massive galaxy, active galactic nuclei (AGN) can in principle have a profound effect on the galaxy formation and evolution processes. X-ray emission offers a unique signpost of accretion of matter onto the supermassive black holes in AGN, being able to penetrate through obscuring material and overcome light from stellar processes. Investigating whether and how nuclear black holes influence their host galaxies, and vice versa, has been and remains to be a major focus in the activity of the MPE HE Group.

From X-ray surveys to AGN demographics

Fig. 1. An example of one of the X-ray surveys to which we contribute to. A false colour XMM-Newton  X-ray image of the COSMOS field (2 sq deg.)

It is well known that, for the study of AGN, X-rays have merits over other selection techniques, primarily a uniform and quantifiable selection function at all redshifts, relatively little attenuation by absorbing material along the line of sight, and minimal host galaxy light dilution. The MPE HE group has a leading role in the science exploitation of nearly all current high-profile extragalactic surveys carried out by major high energy missions, including Chandra, XMM-Newton, Swift and Integral. Members of the group held the leadership in the XMM-COSMOS survey (see Fig. 1), the XMM deep observation of the Lockman Hole, the Chandra AEGIS-wide and AEGIS-deep programmes. The group also has heavy involvements in the Chandra Deep Field South and Chandra COSMOS projects. We also use wide area serendipitous surveys such as XMM-SDSS and 2XMM, the latter via our involvement in the XMM Science Survey Centre. For the analysis and science exploitation of those observations the group is developing novel methods with emphasis on the Bayesian approach, e.g. for the detection of sources and the characterisation of their properties, the construction of sensitivity maps, the association of X-ray sources with counterparts at other wavelenghts. The MPE HE group is also launching the ambitious eROSITA X-ray mission, which will map the entire sky at energies 0.5-10 keV and will yield an unprecedented sample of around 3 million AGN. This will be a uniquely resource for the study of AGN evolution and demographics, but also in terms of discovery space, e.g. finding rare or new classes of sources.

Fig. 2. An example of photometric vs spectroscopic redshifts for the X-ray point sources in the AEGIS-X deep Chandra field, where accuracy and fraction of outliers are also indicated. It is only recently that X-ray source detections, identifications and redshift determinations have become adequate for proper demographic studies, thanks in large part to efforts by memebers of our group.

In addition to extensive use of high energy observatories the MPE HE is following a multiwavelenth approach for the study of AGN populations. This includes conducting imaging observations across the electomagnetic spectrum and involvement in extensive spectroscopic campaigns to measure source distances. As part of preparations for the eROSITA follow-up program we are heavily involved in the design of two wide-area spectroscopic surveys, SPIDERS and 4MOST.

Besides their intrinsic scientific value, these spectroscopic data have been used to create an ad-hoc library of hybrid SED templates (combining normal galaxies and AGN) that have allowed us to obtain reliable photometric redshifts even for faint and high-redshift AGN with an accuracy never reached before at these regimes, limited only by the number of bands available in a given field (Fig. 2.).

Fig. 3. The AGN number density as a function of redshift shows the growth of SMBH. It is well known that the accretion power in the Universe declines from z=1-0, but the exact behaviour at all redshifts depends on a precise understanding of the selection functions, redshift determinations and obscuration properties. 

The accretion history of the Universe

Previous studies of the X-ray Luminosity Function (XLF) have revealed that AGN are a strongly evolving population, with the overall distribution shifting to lower luminosities between z~1 and the present day. Such evolution is broadly characterized by a shift in the peak of the space density of AGN towards lower redshifts for objects of lower luminosities, but there are remaining uncertainties as to the exact form of the evolution, especially at high redshifts. In part, this is due to the difficulties of accurately measuring the faint end and the high redshift part of the XLF, where incompleteness in X-ray detection and identification, and uncertain redshift information are all serious issue. The High energy group at MPE is currently improving on this, compiling highly redshift-complete samples of AGN selected in different bands, using Bayesian techniques to compute the best fitting XLF, properly accounting for selection effects, thus providing a meaningful census of both (mildly) obscured and unobscured AGN activity in the Universe out to z~4 (see Fig. 3). eROSITA will improve greatly our understanding of highly luminous and distant AGN with its unique capability to detect these rare objects with its full sky coverage.

Fig. 4. Histogram of hardness ratios (H-S/H+S, where H is the flux in 2-10 keV band, and S is the flux in the 0.5-2 keV band) for a sample of local AGN with high quality XMM-Newton data. In blue are all AGN, in green sources with NH>1023 cm-2 and in red Compton Thick sources. All CT sources show soft hardness ratios due to scattered emission, and hence would be missed without more complex spectral fitting. Inset is an example spectrum of a Compton thick source in the AEGIS-XD survey identified through spectral fitting. With a hardness ratio of -0.05, it would have been missed using hardness ratios alone. 

A key additional uncertainty in determining the evolution of SMBH accretion is the effect of obscuration. In the search for a complete census of AGN, the missing population that systematically evades our accounting is the so-called Compton Thick (CT) AGN, i.e. those obscured by column densities so high to escape detection in the energy bands where X-ray imaging telescopes are most sensitive (up to ~10 keV). Until a few years ago, the only handle on such population came from the modelling of the cosmic X-ray background radiation. Here at MPE we have conducted work to better understand this population of obscured AGN and its evolution, through a combination of X-ray and infrared surveys, and through improved understanding of the X-ray spectral properties of this population (e.g. Fig. 4)

Co-evolution: the physical link between black holes and their host galaxies

The M-σ relationship seen in the local Universe suggests a direct link between SMBH and their host galaxies. The demographics of AGN, which signpost accretion events onto SMBHs, is powerful approach to shed light into this relationship. Our group has performed much of the pioneering work in this area relating X-ray AGN activity to host colours, morphology and environment.

Fig. 5. The specific Star Formation Rate distributions divided into three redshift bins and three mass ranges for X-ray selected type 2 AGN host galaxies (red histogram) and normal galaxies (grey histogram) in COSMOS. The dot-dashed vertical blue lines are used to distinguish quiescent from starforming galaxies. The upper panel of each histogram shows the observed AGN fractions. We do not see any clear difference in the SFR properties of AGN hosts and normal galaxies.

A major observational challenge in any comprehensive study of AGN–galaxy co-evolution, however, is the accurate separation of the AGN and galaxy emission components at optical–IR wavelengths. Members of the MPE HE group have worked on the development of a procedure to separate these components, yielding interesting results, such as showing that AGN hosts do follow quite closely the overall evolution of the stellar build-up of the parent galaxy population and populate both star-forming and more passive galaxies (see Fig. 5.)

Relating AGN to large-scale structure

The clustering properties of AGN, i.e. their distribution in dark matter haloes, are a diagnostic of the physical conditions under which super-massive black holes at the centres of galaxies grow their mass, e.g. fuelling mode, triggering mechanism. Fig. 6 shows a compilation of measurements of the typical (average) dark matter halo mass of AGN at different redshifts, including many data-points from MPE HE led studies. X-ray AGN live in halos with average masses in the range logM=12.5-13.5 Msol. This interval is inconsistent with major galaxy mergers as the only channel for triggering AGN. A limitation of the measurements in Fig. 6 is that they provide no information on the underlying distribution of AGN in dark matter haloes. Members of MPE HE group are making progress in addressing this limitation. Figure 7 for example shows the first direct measurement of the halo occupation of AGN. A striking result from that figure is the large fraction of satellites among AGN, which argues against major galaxy mergers as the only mechanism for triggering accretion onto supermassive black holes.

Fig. 6. Compilation of measurementsof X-ray AGN bias – a proxy for dark matter halo mass - many of which originate from the MPE HE group or our collaborators. X-ray AGN tend to be associated with massive haloes, log M=12.5-13.5 Msol

Fig. 7. AGN halo occupation as a function of halo mass split into satellite and central galaxies. A large fraction of the AGN population is associated with satellites, i.e. galaxies that fall into a larger halo.

Go to Editor View