An active galactic nucleus (AGN) is a compact region at the center of a galaxy that has a much higher than normal luminosity over at least some portion—and possibly all—of the electromagnetic spectrum, with characteristics indicating that the excess luminosity is not produced by stars. Such excess non-stellar emission has been observed in the radio, microwave, infrared, optical, ultra-violet, X-ray and gamma ray wavebands. A galaxy hosting an AGN is called an active galaxy. The radiation from an AGN is believed to be a result of accretion of matter by a supermassive black hole at the center of its host galaxy. AGN are the most luminous persistent sources of electromagnetic radiation in the universe, and as such can be used as a means of discovering distant objects; their evolution as a function of cosmic time also puts constraints on models of the cosmos. The observed characteristics of an AGN depend on several properties such as the mass of the central black hole, the rate of gas accretion onto the black hole, the orientation of the accretion disk, the degree of obscuration of the nucleus by dust, and presence or absence of jets. Numerous subclasses of AGN have been defined based on their observed characteristics; the most powerful AGN are classified as quasars.
Early photographic observations of nearby galaxies detected some characteristic signatures of AGN emission, although there was not yet a physical understanding of the nature of the AGN phenomenon. Some early observations included the first spectroscopic detection of emission lines from the nuclei of NGC 1068 and Messier 81 by Edward Fath (published in 1909), and the discovery of the jet in Messier 87 by Heber Curtis (published in 1918). Further spectroscopic studies by astronomers including Vesto Slipher, Milton Humason, and Nicholas Mayall noted the presence of unusual emission lines in some galaxy nuclei. In 1943, Carl Seyfert published a paper in which he described observations of nearby galaxies having bright nuclei that were sources of unusually broad emission lines. Galaxies observed as part of this study included NGC 1068, NGC 4151, NGC 3516, and NGC 7469. Active galaxies such as these are known as Seyfert galaxies in honor of Seyfert's pioneering work.
The development of radio astronomy was a major catalyst to understanding AGN. Some of the earliest detected radio sources are nearby active elliptical galaxies such as Messier 87 and Centaurus A. Another radio source, Cygnus A, was identified by Walter Baade and Rudolph Minkowski as a tidally distorted galaxy with an unusual emission-line spectrum, having a recessional velocity of 16,700 kilometers per second. The 3C radio survey led to further progress in discovery of new radio sources as well as identifying the visible-light sources associated with the radio emission. In photographic images, some of these objects were nearly point-like or quasi-stellar in appearance, and were classified as quasi-stellar radio sources (later abbreviated as "quasars"). A major breakthrough was the measurement of the redshift of the quasar 3C 273 by Maarten Schmidt, published in 1963. Schmidt noted that if this object was extragalactic (outside the Milky Way, at a cosmological distance) then its large redshift of 0.158 implied that it was the nuclear region of a galaxy about 100 times more powerful than other radio galaxies that had been identified. Shortly afterward, optical spectra were used to measured the redshifts of a growing number of quasars including 3C 48, even more distant at redshift 0.37. The enormous luminosities of these quasars as well as their unusual spectral properties indicated that their power source could not be ordinary stars. Accretion of gas onto a supermassive black hole was suggested as the source of quasars' power in papers by Edwin Salpeter and Yakov Zel'Dovich in 1964. In 1969 Donald Lynden-Bell proposed that nearby galaxies contain supermassive black holes at their centers as relics of "dead" quasars, and that black hole accretion was the power source for the non-stellar emission in nearby Seyfert galaxies. In the 1960s and 1970s, early X-ray astronomy observations demonstrated that Seyfert galaxies and quasars are powerful sources of X-ray emission, which originates from the inner regions of black hole accretion disks.
Today, AGN are a major topic of astrophysical research, both observational and theoretical. AGN research encompasses observational surveys to find AGN over broad ranges of luminosity and redshift, examination of the cosmic evolution and growth of black holes, studies of the physics of black hole accretion and the emission of electromagnetic radiation from AGN, examination of the properties of jets and outflows of matter from AGN, and the impact of black hole accretion and quasar activity on galaxy evolution.
For a long time it has been argued that an AGN must be powered by accretion of mass onto massive black holes (106 to 1010 times the Solar mass). AGN are both compact and persistently extremely luminous. Accretion can potentially give very efficient conversion of potential and kinetic energy to radiation, and a massive black hole has a high Eddington luminosity, and as a result, it can provide the observed high persistent luminosity. Supermassive black holes are now believed to exist in the centres of most if not all massive galaxies since the mass of the black hole correlates well with the velocity dispersion of the galactic bulge (the M-sigma relation) or with bulge luminosity. Thus AGN-like characteristics are expected whenever a supply of material for accretion comes within the sphere of influence of the central black hole.
In the standard model of AGN, cold material close to a black hole forms an accretion disc. Dissipative processes in the accretion disc transport matter inwards and angular momentum outwards, while causing the accretion disc to heat up. The expected spectrum of an accretion disc peaks in the optical-ultraviolet waveband; in addition, a corona of hot material forms above the accretion disc and can inverse-Compton scatter photons up to X-ray energies. The radiation from the accretion disc excites cold atomic material close to the black hole and this in turn radiates at particular emission lines. A large fraction of the AGN's radiation may be obscured by interstellar gas and dust close to the accretion disc, but (in a steady-state situation) this will be re-radiated at some other waveband, most likely the infrared.
Some accretion discs produce jets of twin, highly collimated, and fast outflows that emerge in opposite directions from close to the disc. The direction of the jet ejection is determined either by the angular momentum axis of the accretion disc or the spin axis of the black hole. The jet production mechanism and indeed the jet composition on very small scales are not understood at present due to the resolution of astronomical instruments being too low. The jets have their most obvious observational effects in the radio waveband, where very-long-baseline interferometry can be used to study the synchrotron radiation they emit at resolutions of sub-parsec scales. However, they radiate in all wavebands from the radio through to the gamma-ray range via the synchrotron and the inverse-Compton scattering process, and so AGN jets are a second potential source of any observed continuum radiation.
There exists a class of 'radiatively inefficient' solutions to the equations that govern accretion. The most widely known of these is the Advection Dominated Accretion Flow (ADAF), but other theories exist. In this type of accretion, which is important for accretion rates well below the Eddington limit, the accreting matter does not form a thin disc and consequently does not efficiently radiate away the energy that it acquired as it moved close to the black hole. Radiatively inefficient accretion has been used to explain the lack of strong AGN-type radiation from massive black holes at the centres of elliptical galaxies in clusters, where otherwise we might expect high accretion rates and correspondingly high luminosities. Radiatively inefficient AGN would be expected to lack many of the characteristic features of standard AGN with an accretion disc.
There is no single observational signature of an AGN. The list below covers some of the features that have allowed systems to be identified as AGN.
It is convenient to divide AGN into two classes, conventionally called radio-quiet and radio-loud. Radio-loud objects have emission contributions from both the jet(s) and the lobes that the jets inflate. These emission contributions dominate the luminosity of the AGN at radio wavelengths and possibly at some or all other wavelengths. Radio-quiet objects are simpler since jet and any jet-related emission can be neglected at all wavelengths.
AGN terminology is often confusing, since the distinctions between different types of AGN sometimes reflect historical differences in how the objects were discovered or initially classified, rather than real physical differences.
See main article Radio galaxy for a discussion of the large-scale behaviour of the jets. Here, only the active nuclei are discussed.
|Emission lines||X-rays||Excess of||Strong
|OVV||yes||no||stronger than BL Lac||yes||yes||no||yes||yes||yes||yes|
Unified models propose that different observational classes of AGN are a single type of physical object observed under different conditions. The currently favoured unified models are 'orientation-based unified models' meaning that they propose that the apparent differences between different types of objects arise simply because of their different orientations to the observer. However, they are debated (see below).
At low luminosities, the objects to be unified are Seyfert galaxies. The unification models propose that in Seyfert 1s the observer has a direct view of the active nucleus. In Seyfert 2s the nucleus is observed through an obscuring structure which prevents a direct view of the optical continuum, broad-line region or (soft) X-ray emission. The key insight of orientation-dependent accretion models is that the two types of object can be the same if only certain angles to the line of sight are observed. The standard picture is of a torus of obscuring material surrounding the accretion disc. It must be large enough to obscure the broad-line region but not large enough to obscure the narrow-line region, which is seen in both classes of object. Seyfert 2s are seen through the torus. Outside the torus there is material that can scatter some of the nuclear emission into our line of sight, allowing us to see some optical and X-ray continuum and, in some cases, broad emission lines—which are strongly polarized, showing that they have been scattered and proving that some Seyfert 2s really do contain hidden Seyfert 1s. Infrared observations of the nuclei of Seyfert 2s also support this picture.
At higher luminosities, quasars take the place of Seyfert 1s, but, as already mentioned, the corresponding 'quasar 2s' are elusive at present. If they do not have the scattering component of Seyfert 2s they would be hard to detect except through their luminous narrow-line and hard X-ray emission.
Historically, work on radio-loud unification has concentrated on high-luminosity radio-loud quasars. These can be unified with narrow-line radio galaxies in a manner directly analogous to the Seyfert 1/2 unification (but without the complication of much in the way of a reflection component: narrow-line radio galaxies show no nuclear optical continuum or reflected X-ray component, although they do occasionally show polarized broad-line emission). The large-scale radio structures of these objects provide compelling evidence that the orientation-based unified models really are true. X-ray evidence, where available, supports the unified picture: radio galaxies show evidence of obscuration from a torus, while quasars do not, although care must be taken since radio-loud objects also have a soft unabsorbed jet-related component, and high resolution is necessary to separate out thermal emission from the sources' large-scale hot-gas environment. At very small angles to the line of sight, relativistic beaming dominates, and we see a blazar of some variety.
However, the population of radio galaxies is completely dominated by low-luminosity, low-excitation objects. These do not show strong nuclear emission lines — broad or narrow — they have optical continua which appear to be entirely jet-related, and their X-ray emission is also consistent with coming purely from a jet, with no heavily absorbed nuclear component in general. These objects cannot be unified with quasars, even though they include some high-luminosity objects when looking at radio emission, since the torus can never hide the narrow-line region to the required extent, and since infrared studies show that they have no hidden nuclear component: in fact there is no evidence for a torus in these objects at all. Most likely, they form a separate class in which only jet-related emission is important. At small angles to the line of sight, they will appear as BL Lac objects.
In the recent literature on AGN, being subject to an intense debate, an increasing set of observations appear to be in conflict with some of the key predictions of the Unified Model, e.g. that each Seyfert 2 has an obscured Seyfert 1 nucleus (a hidden broad-line region).
Therefore, one cannot know whether the gas in all Seyfert 2 galaxies is ionized due to photoionization from a single, non-stellar continuum source in the center or due to shock-ionization from e.g. intense, nuclear starbursts. Spectropolarimetric studies reveal that only 50% of Seyfert 2s show a hidden broad-line region and thus split Seyfert 2 galaxies into two populations. The two classes of populations appear to differ by their luminosity, where the Seyfert 2s without a hidden broad-line region are generally less luminous. This suggests absence of broad-line region is connected to low Eddington ratio, and not to obscuration.
The covering factor of the torus might play an important role. Some torus models predict how Seyfert 1s and Seyfert 2s can obtain different covering factors from a luminosity- and accretion rate- dependence of the torus covering factor, something supported by studies in the x-ray of AGN. The models also suggest an accretion-rate dependence of the broad-line region and provide a natural evolution from more active engines in Seyfert 1s to more “dead” Seyfert 2s and can explain the observed break-down of the unified model at low luminosities and the evolution of the broad-line region.
While studies of single AGN show important deviations from the expectations of the unified model, results from statistical tests have been contradictory. The most important short-coming of statistical tests by direct comparisons of statistical samples of Seyfert 1s and Seyfert 2s is the introduction of selection biases due to anisotropic selection criteria.
Studying neighbour galaxies rather than the AGN themselves first suggested the numbers of neighbours were larger for Seyfert 2s than for Seyfert 1s, in contradiction with the Unified Model. Today, having overcome the previous limitations of small sample sizes and anisotropic selection, studies of neighbours of hundreds to thousands of AGN have shown that the neighbours of Seyfert 2s are intrinsically dustier and more star-forming than Seyfert 1s and a connection between AGN type, host galaxy morphology and collision history. Moreover, angular clustering studies of the two AGN types confirm that they reside in different environments and show that they reside within dark matter halos of different masses. The AGN environment studies are in line with evolution-based unification models where Seyfert 2s transform into Seyfert 1s during merger, supporting earlier models of merger-driven activation of Seyfert 1 nuclei.
While controversy about the soundness of each individual study still prevails, they all agree on that the simplest viewing-angle based models of AGN Unification are incomplete. Seyfert-1 and Seyfert-2 seem to differ in star formation and AGN engine power. 
While it still might be valid that an obscured Seyfert 1 can appear as a Seyfert 2, not all Seyfert 2s must host an obscured Seyfert 1. Understanding whether it is the same engine driving all Seyfert 2s, the connection to radio-loud AGN, the mechanisms of the variability of some AGN that vary between the two types at very short time scales, and the connection of the AGN type to small- and large-scale environment remain important issues to incorporate into any unified model of active galactic nuclei.
For a long time, active galaxies held all the records for the highest-redshift objects known either in the optical or the radio spectrum, because of their high luminosity. They still have a role to play in studies of the early universe, but it is now recognised that an AGN gives a highly biased picture of the "typical" high-redshift galaxy.
Most luminous classes of AGN (radio-loud and radio-quiet) seem to have been much more numerous in the early universe. This suggests that massive black holes formed early on and that the conditions for the formation of luminous AGN were more common in the early universe, such as a much higher availability of cold gas near the centre of galaxies than at present. It also implies that many objects that were once luminous quasars are now much less luminous, or entirely quiescent. The evolution of the low-luminosity AGN population is much less well understood due to the difficulty of observing these objects at high redshifts.