11.1. Interplay between super-massive black holes and their host galaxy
The central super-massive black holes are an essential ingredient of galaxies despite their relatively small mass compared to the total galactic mass. Indeed, the total energy radiated during the AGN phases may be comparable to the total energy radiated by stars, and thus may deeply perturb the whole galaxy. It is now well established that many - and perhaps all - massive luminous nearby galaxies contain central supermassive black holes at their centres with masses ~ 106 - 109 M, with tight relations between the black-hole mass and the spheroid which harbours it (e.g. Kormendy & Richstone 1995, Magorrian et al. 1998, Kormendy & Gebhardt 2000, Ferrarese et al. 2006). While the tightest correlation is with the central velocity dispersion, , of its host bulge, with MBH 4 (e.g. Tremaine et al. 2002, Ferrarese 2002), it also results that MBH is approximately proportional to the spheroid mass, MBH 2.5 × 10-3 Mbulge (see e.g. Merritt & Ferrarese 2001, McLure & Dunlop 2004). Although the origin of the MBH- relation is not yet fully established, various suggestions have been made (see e.g. a list in Di Matteo et al. 2003, 2005 and Begelman & Nath 2005). Most of them, following e.g. Silk & Rees (1998), have stressed the probable dominant influence of strong feedback on the black hole growth due to the action of quasar flows on the galactic gas reservoir.
It is therefore not surprising that there is some correlation between the phases when the galaxy and the black hole build most of their respective masses, and thus between AGN and starburst activities. An essential factor for explaining such a correlation is certainly the fact that both the black-hole accretion and the starburst are fed from the interstellar gas which must be strongly perturbed in both cases, either to be transported to the galactic centre, or to be compressed to initiate star formation. In both cases the gas must also eventually be dense. It is thus natural that the association of molecules with AGN is important in various respects and at different scales. We will distinguish three scales: i) on parsec scale, the accretion molecular disk and accompanying jets and outflows, and the much thicker molecular torus (Krolik & Begelman 1988, Elitzur & Shlosman 2006 for recent references); ii) on hundred parsec scale, the dense gas with complex structure, location of nuclear starbursts and essential for transporting the gas to the very centre; iii) the scale of a whole galaxy, or at least its bulge for spirals, where the relations with the black-hole mass hold. This large scale is the theatre where the AGN feedback eventually plays in dispersing the interstellar gas. This is also the case of galaxy mergers. They generate major molecular starbursts, both extended and nuclear, and affect the AGN by transporting large amounts of gas to the centre. Molecular clouds also contribute to the dissipation which eventually allows black-hole mergers (e.g. Makino & Funato 2004 and references therein).
We will not develop much further the peculiarities of the molecular medium of AGN host galaxies at such large galactic scales, because much of this have been discussed in other sections: about the general relation between starbursts and AGN in local LIRGs and ULIRGs (Section 5), in high-z ULIRGs (SMGs) (Section 9.2), in high-z bright QSOs and radio galaxies (Section 9.3); as well as about H2 emission in AGN host galaxies and at very large distance in shocks triggered by AGN jets (Section 10).
Let us stress nevertheless that the main effect of AGN on the molecular medium of galaxies at all scales is its eventual destruction in the final feedback, and when this effect is total, the AGN emission has also stopped. However, there are interesting cases to consider, especially for QSOs, when there is significant and even strong AGN emission without much molecular gas. This occurs especially at relatively low redshift, 1. In the feedback model, it could take place just before complete feedback stops the accretion, or at the occasion of minor late mergings. Let us also recall the case of cooling flows in massive cluster galaxies which are powerful AGN with radio jets, strong resulting feedback, and often CO emission (Section 4.2). In this context the CO and cold dust emission of QSOs and radio galaxies is interesting for tracing the presence of starbursts. It has been seen that very strong CO and dust millimetre emission is frequent in z 2 QSOs and radio galaxies. However, it is quite different at z 1. Both systematic surveys of CO in local (z < 0.15) radio galaxies (Evans et al. 2005), and 1.2 mm dust emission in 3CR radio galaxies and quasars (Haas et al. 2003) have shown that starburst ULIRGs in the host galaxy of powerful AGN are much rarer locally than at high z. Nevertheless, CO studies show that a strong starburst activity, typical of LIRGs and nuclear starbursts, is present in a large fraction of prominent local AGN, including: luminous QSOs (Palomar Green, PG; Scoville et al. 2003; however, see Bertram et al. 2006), hard X-ray selected Seyferts (Rigopoulou et al. 1997), and galaxies hosting F-R I and compact radio jets (Evans et al. 2005). See also the large fraction of PAH emission observed by Schweitzer et al. (2006) in a sample of 27 PG QSOs at z 0.3. Even ULIRG starbursts are present in local infrared ultraluminous QSOs (e.g. Hao et al. 2005) with large FIR excess, and CO has been detected in a number of them (Solomon et al. 1997, Evans et al. 2001, 2006, etc.). Detailed interferometric studies of prominent objects such as 3C 48 (Krips et al. 2005a) and PDS 456 (Yun et al. 2004) confirm the existence of rich merger structures and give clues about the joint AGN-starburst evolution. However, it is shown by Ho (2005) that the evidence of molecular gas is not enough to infer the presence of a strong starburst, since the star formation efficiency in otherwise gas-rich host galaxies may be suppressed in the presence of strong AGN feedback.
11.2. Molecules in the central regions and fueling the AGN
The region in the vicinity of the `molecular torus' is fundamental for the diagnostic of AGN processes. It is complex as regards its structure, dynamics, radiative processes, outflows, turbulence, shocks, magnetic field, etc. Besides, this region has the interesting property to be strongly irradiated by the X-rays of the AGN and allows one to study their physical and chemical effects on the surrounding molecular medium (Maloney et al. 1996).
As discussed in Section 10.2, H2 near-infrared emission lines are widely observed in nuclear and circumnuclear regions of AGN (see e.g Riffel et al. 2006b for a near-infrared atlas of 47 AGN including lines of H2, and Roussel et al. 2007 for H2 rotational lines in nearby Seyferts and LINERs). The lines of H2 are commonly used as diagnostic of the conditions in these regions. Comparing the ratios in several H2 lines allows one to probe the physical conditions in warm molecular clouds which are irradiated by ultraviolet (or X-ray) radiation or heated by shocks. With the ongoing increase in angular resolution they will become more important to probe the complex structure, dynamics and star formation of the immediate vicinity of the molecular torus, and their relation with feeding the black-hole accretion.
The current sensitivity and angular resolution of millimetre interferometers is well adapted to map one or several CO lines in the central kpc around nearby AGN. The processes which removes the angular momentum of the gas, brings it to the centre and feeds the AGN are complex and not very well understood (e.g. Combes 2005). Several CO surveys have thus addressed the central structure and dynamics of the molecular gas in a number of nearby AGN (Jogee et al. 2001, García-Burillo et al. 2003 [NUGA], Helfer et al. 2003). The most striking result is the large variety of circumnuclear disk morphologies found, especially in the most detailed survey, NUGA, of a dozen of low luminosity AGN with the IRAM interferometer. The results allow a detailed study of each individual object (e.g. García-Burillo et al. 2005, 2006c and references therein, Combes et al. 2004, Krips et al. 2005b). They propose an interpretation of the various observed dynamical states as the different epochs of the evolution cycle driven by bars: from the formation of a bar through gravitational instability of a cold disk, to destruction of the bar by the gas flow driven by the bar, and replenishment of the gas disk through accretion, the various stages of secular evolution concur to fuel the AGN and assemble bulges at the same time. However, the variety observed is rather challenging and urges to refine current dynamical models (García-Burillo et al. 2004, 2005, 2006c).
The field of observations of molecules other than CO in the central regions of AGN is active at various wavelengths. However, understanding the chemistry remains difficult, especially the direct effect of the AGN on the various layers through X-ray irradiation and heating. For instance, Evans et al. (2006) have carried out a systematic observation of HCN in QSOs to study their dense molecular gas and the role of star formation in their host galaxies. The observed enhanced ratio of infrared to HCN luminosities compared to cool IRAS galaxies may appear to be an indication that the AGN contributes significantly to heating the dust, but other interpretations are possible. The influence of buried AGNs at the cores of ULIRGs have been invoked for explaining the observed intensities of HCO+ or HNC (Imanishi et al. 2006, Aalto et al. 2006, Yamada et al. 2007), but the case remains unclear.
The strong infrared continuum sources of AGN could look ideal for tracing absorption features such as ice, CO or other molecules, in deeply obscured sources. However, ISO and Spitzer infrared spectroscopic surveys of AGN (e.g. Spoon et al. 2002, 2005, 2007, Lutz et al. 2004) have found very few sources with absorption features. See nevertheless the case of NGC 6240 observed by Armus et al. (2006), and the strong molecular features in the infrared L-band (3-4 µm) and M-band (4-5 µm) observed by Sani et al. (2007) with VLT-ISAAC spectroscopy. Aromatic infrared bands of PAHs are generally weak in AGN (Section 8), but they are present in many of them (e.g. Schweitzer et al. 2006).
The sensitivity and angular resolution of ALMA will be essential to disentangle the complex structures and use the diagnostic of other molecules than CO and H2 in the central regions of AGN (e.g. Baker 2005).
11.3. H2O mega-masers and AGN molecular disks
Similarly to OH (Section 5), extragalactic H2O masers in the 61,6-52,3 transition (0 = 22.235 GHz) were first searched in nearby galaxies on model of Galactic H2O masers, mostly in nuclear starbursts such as M 82, NGC 253, NGC 2146 (see references in Lo 2005; see also Brunthaler et al. 2006 for the Local Group). They showed up there with a luminosity of up to a few L, comparable to the most luminous Galactic masers and are similarly likely mostly associated with star formation. However, it was soon discovered (Dos Santos & Lépine 1979, Gardner & Whiteoak 1982, Claussen et al. 1984, Claussen & Lo 1986) a class of much more luminous H2O masers, now called mega-masers with luminosity of up to 104 L (such maximum luminosities may be intermittent, and overestimated because of anisotropic emission). They were soon proven to be associated with active nuclei. Such powerful H2O mega-masers (see again the detailed recent review by Lo 2005, to which we refer for more detailed discussion and references) are certainly the molecular sources the most specifically associated with AGN. While OH mega-masers typically arise in LIRG extreme starburst regions that are distributed over a 100-pc scale (Section 5), H2O mega-masers are found mainly within the central parsec of an AGN (although a few less powerful ones are also found in starburst galaxies). Either in the circumnuclear accretion disk, or in the gas associated with jets or outflows, they are uniquely tracing the dense molecular gas directly exposed to the power of the AGN. Either direct irradiation by penetrating X-rays, or (jet-)induced shocks produce the high temperatures (several hundred Kelvin) and the large H2O abundance required for mega-maser emission. However, they are relatively rare; about 80 are known out of more than 1000 galaxies (including 450 AGN) searched following Lo (2005) and Kondratko et al. (2006a, b). With luminosities up to 104 L, they are found mainly in Seyferts 2 or LINERS (`Low-ionization nebular emission regions'). Such objects are probably the best for providing the long amplification paths and the protection of molecules which are required for large amplification. Among the known H2O mega-masers, 50% arise from Compton-thick and 85% from heavily obscured (> 1023 cm-2) active galactic nuclei (Zhang et al. 2006). Note that the detection has been recently extended to a single case of a Type 2 QSO at z = 0.66 (Barvainis and Antonucci 2005) with a luminosity of ~ 23000 L.
In NGC 4258, mapping the H2O mega-maser emission has provided the first direct evidence in an AGN for the existence of a thin Keplerian accretion disk with turbulence, as well as highly compelling evidence for the existence of a massive black hole. Such a splendid case study (see references in Lo 2005) has shown the unique power of H2O mega-masers for probing molecular circumnuclear disks, their dynamics and their black-hole masses. The very high surface brightness of such maser emission allows VLBI observations to achieve milliarcsecond resolution, which corresponds to subpc resolution of the emitting molecular medium within 1 pc of the nucleus. The NGC 4258 mega-maser has also provided a geometric distance determination of extremely high precision. The current case of the use of H2O masers to determine the mass of super-massive black holes is discussed in detail in Ferrarese & Ford (2005). It is clear that with the advent of powerful new facilities such as EVLA (see e.g. Menten 2007) and especially SKA, studies of mega-masers will be useful high-resolution probes of AGN and will provide accurate determinations of black-hole masses and cosmological high-z distance determinations.
The detection at IRAM of a new H2O megamaser in the 31,3-22,0 line (0 = 183.310 GHz) was recently reported in the local ULIRG Arp 220 by Cernicharo et al. (2006). This result opens up the possibility of using the 183 GHz H2O line as an additional tool to explore the physical conditions in LIRGs and ULIRGs, with a potential interest for high angular resolution observations with ALMA.