4.5.2. FR I vs. FR II radio galaxies: the ADAF model

The ADAF model

It is believed that most gE galaxies possess a nuclear BH with a mass in excess of 10⁸ M. Bondi accretion from the interstellar medium might then be expected to produce quasar-like luminosities from the nuclei of most gE galaxies. It is a puzzle that such luminosities are not observed ([114]). Motivated by this problem, Fabian & Rees (1995) have suggested that the final stages of accretion in these objects occur in an advection dominated accretion flow (ADAF) with a small radiative efficiency.

ADAFs occur when the local cooling time-scale becomes longer than the accretion time-scale so that most of the dissipatively liberated energy is advected inward with the accreting gas and lost into the BH. The viscous heating is assumed to affect mainly the ions, while the radiation is produced primarily by the electrons; as the ions transfer only a small fraction of their energy to the electrons via Coulomb collisions, the radiative efficiency of an ADAF is much less than the total energy released during accretion. The result is that the ion temperature becomes almost virial (T ~ 10¹² K) which causes the flow to have a nearly spherical morphology; the electron temperature is determined by a balance between heating due to collisions with the ions and inverse Compton cooling and reaches T_e ~ 10⁹-10^9.5 K. This happens for accretion rates below M_crit ~ ² _Edd, where is the viscosity parameter and _Edd is the accretion rate corresponding to the Eddington limit. If = 0.1 then advection can dominate the flow if the accretion proceeds at less than 1% of the Eddington rate. For such low accretion rates, the expected luminosity scales as ² rather than M and the luminosity of the system falls well below the luminosity which is normally expected from accretion ([316]; [275]). It is often assumed that there is, in these objects, an outer region extending outward from about 3000 Schwarzschild radii where the accretion flow is in the form of a standard geometrically thin disk ([316]).

The entire spectrum of an ADAF is completely determined by the mass of the BH and by the value of the transition radius out of which the disk is thin; it is very characteristic: a ^1/3 slope in the radio regime, a submillimeter-to-X-ray Compton spectrum, and a hard-X-ray-to--ray Bremsstrahlung spectrum ([274]).

The radio emission is due to cyclo-synchrotron radiation from hot electrons in the equipartition magnetic field; it should be isotropic. In the absence of a radio jet, the expected core radio luminosity is relatively low and roughly proportional to the mass of the central BH; the 5 GHz ADAF luminosity is L (= L) < 10³⁸ erg s^-1 for M_BH < 10⁹ M; however, some level of jet activity not directly associated with the ADAF seems to be present in most sources and the observed radio emission can be considerably larger ([115]; Yi & Boughn 1999). [103] have shown that the high frequency radio emission of the nucleus of the gE galaxies for which the mass of the central BH has been estimated is significantly lower than predicted by the ADAF model, suggesting not only that there is no radio jet in these objects but also that that the radio emission due to the ADAF is being suppressed.

FR I vs. FR II radio galaxies

The differences in the large-scale properties of FR Is and FR IIs radio sources are most probably related to both intrinsically different core properties (Lorentz factor, accretion rate etc) and differences in the environment outside the nuclear region of the associated optical galaxy ([442]; [332]; [146]). This would explain why there is no exact correspondence between high- and low-ionization objects (governed by the properties of the core) and their radio morphology which depends both on the ``strength'' of the jet and the density of the environment.

Radio loud AGNs are basically similar to their radio quiet counterparts, except for the presence in the former of a radio jet which is generally attributed to the effect of a rapidly rotating BH. The rotation rate of the BH seems therefore to have no influence on the other observable properties of the AGNs. If FR Is and FR IIs were differing only by the spin rate of their BH, they would differ by the properties of their jets, not by those of their optical nuclei.

It is customary to assume that the accretion power is radiated away with an efficiency 0.1 close to the surface of a BH; in other words, it is assumed that most of the gravitational energy released through viscous dissipation is radiated away locally from the accretion disk; this condition is very well satisfied for a geometrically thin accretion disk.

It seems now well established (see above) that FR Is and BLLs have jets which are relativistic near to the nucleus but with Lorentz factors systematically smaller than in FR IIs and HPQs. According to [12] and [47], this difference could be due to the fact that FR II jets arise from a nucleus with a rapidly rotating BH, and FR I jets from a nucleus with a slowly rotating BH. Alternatively, it could be the result of different jet collimation mechanisms associated with an accretion flow of a different nature ([28]).

While supplying large power to their radio emitting regions, the nuclei of FR I radio galaxies emit little detectable radiation; however, the total energy content of the extended radio components implies that the galaxy nucleus contains a BH of mass 10⁷-10⁸ M; this suggests that their center contains a spinning BH surrounded by an ADAF; the thick disk anchors magnetic fields which extract rotational energy from the hole in the form of two collimated beams of relativistic particles and magnetic fields; these, in turn, drive the observed radio jets ([346]).

The main difference between FR Is and FR IIs (more specifically between low and high-excitation galaxies) would be that the accretion rate is small in the former and that, consequently, the accretion flow is advection dominated ([28]; [348]).

It has been suggested that the FR I galaxy M 87 and the weak radio galaxy NGC4649 which, despite possessing evidences for a supermassive BH, have a very low core luminosity, contains an ADAF ([348]; [102]).

AGN-type Liners and ADAFs

It is natural to assume that the basic difference between Liners and Seyferts is the same as between FR Is and FR IIs. Indeed, [250] suggested that the parameter that distinguishes Liners from Seyferts is their small L / L_Edd and that their accretion flow is advection dominated. For [485], Liners with both X-ray and high-frequency radio fluxes have ratios of radio to X-ray luminosities compatible with being due to an ADAF.

Sambruna et al. (1999) have found in five out of six LERGs a hard X-ray component of low luminosity (L_{2 - 10keV} ~ 10⁴⁰-10⁴² erg s^-1) suggesting that indeed they are low accretion rate AGNs powered by an ADAF.

Liners tend to show little or no significant short-term X-ray variability; this is a marked break from the trend of increased variability in Seyfert 1s with lower luminosity; this difference could be due to the presence of an ADAF resulting in a larger characteristic size for the X-ray producing region than in the case of Seyferts ([341]).

We have seen above (sec. 2.1.2) that a small number of AGNs have double-peaked broad line profile and that these objects have a Liner-like narrow line spectrum. [81] showed that, although a Keplerian disk fits the line profiles of these objects, a cool thin accretion disk does not account for the line fluxes; they propose that, in these cases, a thick, hot ion torus occupies the inner disk and that inverse Compton scattered X-rays from the torus illuminate a thin outer disk. In the spectra of these objects, the fraction of starlight is significantly larger than in the spectra of ``classical'' BLRGs; in BLRGs, the thermal blue/UV component which is commonly attributed to the inner parts of an optically thick accretion disk is easily observed; if the inner disk is replaced by an ion torus, the blue/UV bump should disappear which seems to be the case in the double-peaked objects ([109]). This is additional evidence for the presence of an ADAF in Liners; but it has not yet been shown that the continuum emitted by ADAFs could produce a Liner-like spectrum.

The presence of water megamasers in Seyfert 2s proves the existence in these objects of a molecular torus. The H₂O emission lines in NGC1052, a FR I radio galaxy, are relatively broad and smooth, unlike the narrow maser spikes seen in Seyfert 2s; the masers lie along, rather than perpendicular, to the jet and there is no indication of the existence in this object of an accretion disk ([83]).

NGC1052 is an FR I; a broad H component (2120 km s^-1 FWHM) is visible in total light; in polarized light, there is a significantly broader (4920 km s^-1 FWHM) component which is most probably due to scattering by electrons in a medium with T_e ~ 10⁵ K. It seems therefore that the nucleus is seen directly ([24]). However, [174] and [469] have observed NGC1052 in X-rays; they found that the most convincing model is a nuclear source obscured by a screen of matter with column density ~ 10²³ cm^-2. Such a high column density of cold matter would completely obscure the optical nucleus.

Barth et al. (1999b) have recently discovered the presence of broad polarized H emission in the spectrum of two additional Liners, both hosted by a FR I galaxy (NGC315 = B20055+30, and NGC4261 = 3C270.0) ([271]; [299]).

We have seen that the presence of double-peaked broad emission lines in Liners has been tentatively interpreted as emission from an external thin disk heated by the radiation of the internal hot thick torus; the weak broad lines observed in a number of Liners could be due to the same mechanism.

NGC4258 has often been classified as a Liner; it is however a Seyfert 2 ([198]); moreover, the presence of a molecular disk ([295]) is unexpected in Liners. The mass of the central BH in NGC4258 is 3.6 10⁷ M ([295]); the Eddington luminosity corresponding to this mass is L_Edd = 4.5 10⁴⁵ erg s^-1; the bolometric luminosity has been estimated to be L_bol ~ 3.4 10⁴² erg s^-1, implying a sub-Eddington luminosity L ~ 0.75 10^-3 L_Edd; it has therefore been suggested that the nucleus of NGC4258 contains an ADAF (Lasota et al. 1996; [137]; [80]). The low radio luminosity of the central engine however call into question the ADAF mechanism in this object ([191]), although radio luminosities lower than predicted may not be exceptional (Di Matteo et al. 1999).

To conclude, it seems that the difference between radio loud and radio quiet AGNs is that the BH in the former have a high spin rate. The difference between high- and low-ionization radio galaxies and between Seyferts and Liners is the low accretion rate in low-excitation objects resulting in the formation of these objects of a hot, optically thin, geometrically thick disk, while the Seyferts contain a cool, optically thick, geometrically thin disk. However although the ADAF model looks very promising, it should be stressed that no really convincing evidence exists for the presence of an ADAF in any AGN.

It appears that the main spectroscopic characteristic of the Liners is the low-ionization spectrum due to ionization by the emission of a very hot plasma, the thick inner torus, while Seyferts are ionized by the thermal UVX emission of a cool thin torus. Liners may have broad emission lines which however are weak compared to the narrow emission lines. This leads us to revise our classification of blazars: it is not the presence of broad emission lines which puts a blazar in the HPQ class, but the presence of broad lines associated with high-excitation narrow lines; unfortunately, as the EW of the lines in these objects is small, such a classification is rarely possible. However, the similarity we have found between the properties of BL/HPQs and HPQs suggest that only a small number of BL/HPQs could really be BLLs with a low-excitation narrow line spectrum. This may be the case for BL Lacertae itself as the spectrum published by [93] has strong [O I] and [N II] lines relative to the narrow component of H. Moreover, BL Lacertae has all the characteristics of the genuine BLLs: small radio brightness temperature, small amplitude of intraday optical variability and of -ray variability, small Doppler and Lorentz factors.