3. Some theoretical considerations related to the morphological dichotomy
3.1. Jet deceleration
In light of the evidence for relativistic jet velocities persisting up to multi-kpc scales in FR II sources, coupled with the likelihood that the jets in both FR types start out with bulk relativistic speeds, many theoretical studies have stressed the need for deceleration of the jet flow in FR I sources. Begelman (1982) argued that viscous dissipation in jets can balance adiabatic heating and cause a rather rapid deceleration of weaker jets to transonic or subsonic speeds. These jets could remain undisrupted for substantial distances, thereby yielding typical FR I morphologies, provided the external pressure gradient is appropriately steep. De Young (1993) noted that the Owen-Ledlow transition from FR II to FR I at a fixed and increasing could correspond to a supersonic (perhaps relativistic) jet being severely decelerated in the inner kpc of the parent galaxy, where more gas is likely to be available for entrainment. Plausibly, enough of such gas could arise from stellar winds, or perhaps from cooling flows onto the cD galaxies which often host FR I sources.
Bicknell (1984, 1994, 1995) focused on the idea that turbulent entrainment of cool interstellar medium at the jet boundary could dramatically decelerate a jet. His original work (Bicknell 1984) assumed non-relativistic FR I jets throughout, and ran into some difficulties (e.g., Laing et al. 1999). But the later model (Bicknell 1994, 1995), which assumed initially relativistic jets which eventually come into pressure balance with the external medium, is quite successful in reproducing many aspects of the observations. Bicknell argued that the instability to Kelvin-Helmholtz modes that would produce jet flaring, and thus an FR I morphology, tended to occur at Mach numbers of or flow velocities of c. This result was shown to hold for wide ranges of initial relativistic velocities and of initial ratios of cold to relativistic matter in the jet (Bicknell 1995). By incorporating the known empirical relationships between the optical and X-ray properties of elliptical galaxies, Bicknell's (1995) model could effectively account for the observed slope (and approximate intercept) of the Owen-Ledlow boundary for the FR I/FR II transition in the - plane. Self-similar models for radio source growth by Kaiser & Alexander (1997) feature a turbulent shear layer that could disrupt weaker jets, turning them from FR II into FR I type morphologies if the external density gradient was rather shallow. Komissarov (1994) and Bowman et al. (1996) also considered entrainment as leading to jet deceleration in FR I sources. Bowman et al. (1996) argued that FR I plasma was initially hotter and stressed the importance of cool stellar matter directly swept up by the jets. They showed that this could produce substantial deceleration, even if the jets were highly relativistic initially, without causing a precipitous dissipation of the jet's kinematic power and the ensuing dramatic brightening, which is not observed (see, Scheuer 1983).
Observational support for the models invoking decelerating FR I jets comes from the anti-correlation found between the apparent brightness ratio and the width ratio of the twin jets in FR I sources, which is readily understood in terms of Doppler boosting of a centrally peaked velocity profile (Laing et al. 1999, and references therein). A wide variety of the observed characteristics of the radio jets in FR I sources, such as the emission gaps seen near the nucleus (Komissarov 1990), the asymmetries in apparent emission from the two jets, and their magnetic field patterns, are reasonably explained if the jets in these sources consist of a narrow "spine" of relativistic flow with a predominantly transverse magnetic field, surrounded by a slower moving "sheath" contaminated by entrained material (a shear layer) where the magnetic field is stretched into a predominantly longitudinal configuration (Laing 1993, 1996; Laing et al. 1999). This picture is in accord with Bicknell's (1995) transonic relativistic jet models which are confined by external pressure at large distances, and where the deceleration usually occurs within kpc of the core.
3.2. Jet composition
Total energy and synchrotron radiation constraints led Celotti & Fabian (1993) to conclude that FR II jets were made of -p plasma, since they argued that - plasma of the required density would yield too much annihilation radiation. On the other hand, Reynolds et al. (1996b) used similar energetic and radiation constraints to conclude that the jet in the FR I source M87 was likely to be made of - plasma. A similar argument favors an electron-positron jet in the Optically Violently Variable Quasar 3C 279 (Hirotani et al. 1999). If all of these arguments are taken at face value, one might infer that the main difference between FR I and FR II sources lies in the composition of the jet plasma, and this would imply the existence of a qualitative difference between their central engines. However, evidence for the presence of pair plasma jets, even in FR II sources, comes from the interpretation of the radio power-linear-size (P-D) diagram in terms of a model for self-similar growth of double radio sources (Kaiser et al. 1997). Furthermore, there are viable alternatives to the annihilation constraint invoked by Celotti & Fabian (1993) to argue against pair plasmas in FR II jets; for example, the earliest stage of the energy transport could be predominantly via Poynting flux (Reynolds 1999, private communication), in which case the radiating relativistic matter in all jets could indeed be essentially an - plasma.
The "spine/sheath" model (Sect. 3.1) is broadly reminiscent of the two-fluid-type configuration for jets, put forward by Pelletier & Roland (1989). (See also Sol et al. 1989.) They suggest that the spine of the jet is relativistic, at least on parsec scales, and is composed of a pair plasma, while the outer sheath is made of -p plasma, and carries the bulk of the energy to the outer lobes.
3.3. Galactic mergers
Substantial isophotal distortions are observed in the ellipticals hosting both FR I and FR II sources, strongly implying that galactic encounters/mergers have occurred (Heckman et al. 1986; Colina & de Juan 1995). However, the distinctive sharpness of the distortions seen in the FR II hosts (Smith & Heckman 1989), combined with the presence of strong optical emission lines and significantly higher MFIR emission (Heckman et al. 1994) suggests the occurrence of a starburst due to merger of a gas rich spiral with the elliptical host (Smith & Heckman 1989; Colina & de Juan 1995). In contrast, elliptical-elliptical mergers have been invoked in the case of FR I galaxies (Colina & de Juan 1995).
Consecutive mergers of galaxies containing central supermassive black holes (SMBHs) could produce multiple SMBH systems. Such triple systems usually become unstable and eject a single black hole in one direction, and the recoil sends the surviving binary black hole system in the opposite direction; this is the core idea of the gravitational slingshot model for radio source production (e.g., Saslaw et al. 1974). In this scenario, FR I sources correspond to SMBHs ejected at less than the escape velocity from the merged host galaxy, while FR II sources arise from SMBHs that do escape (Valtonen & Heinämäki 2000). Since this picture naturally produces different velocities for SMBHs of different masses, it could both produce HYMORS, and even make actual predictions as to their frequency. Valtonen & Heinämäki (2000) also argue that the slingshot model can roughly account for the dependence of on , as well as for many other properties of radio galaxies.
3.4. Spin of the central engine
Some hydromagnetic process is now widely believed to be responsible for launching relativistic jets from the vicinity of the accretion disk/supermassive black hole combination which is believed to constitute the central engine in all AGN, although the details remain highly contentious (e.g., Scheuer 1996; Wiita 1996). A possible hint that the angular momentum of the central engine is important in launching powerful FR II jets comes from the rotational kinematics of the (presumably accreted) ionized gas observed in FR II host galaxies (Baum et al. 1992). The idea that a merger of two SMBHs belonging to a merged pair of elliptical galaxies could yield a single rapidly spinning SMBH, which propels powerful relativistic jets, was advocated by Wilson & Colbert (1995). While the black hole spin may well be an important ingredient for ejection of powerful jets, the existence of HYMORS (Sect. 4) disfavors differences in the SMBH spin as the principle mechanism for the FR dichotomy.
One basic class of scenarios involves variants of the Blandford-Znajek (1977, B-Z) mechanism, which could tap the SMBH's rotational energy via magnetic field lines threading the SMBH horizon. While extremely efficient in principle, and capable of providing powerful radio jets with minimal optical thermal emission if an ion-supported torus forms in the innermost region (Rees et al. 1982), the viability of this mechanism has recently been questioned on several grounds. Ghosh & Abramowicz (1997) argued that the strength of the magnetic field that could actually thread the SMBH horizon may have been substantially overestimated. Even if the B-Z mechanism works, Livio et al. (1999) have claimed that the emitted power is dominated by energy output from the inner disk regions, at least for the standard thin accretion disks, and therefore the efficiency is much reduced. This last limitation may be overcome if the accretion disk is actually thick in the inner regions (Armitage & Natarajan 1999). An additional potential problem for the B-Z mechanism was recently pointed out by Li (2000); he argued that the plasma screw instability must set in and this implies that even if the B-Z mechanism does work locally, any jet it launches would be severely limited in its overall length.
While none of the above critical arguments can be considered to be watertight, they tend to support the alternative basic scenario, which involves hydromagnetic launching of the jets from the accretion disk, rather than from the immediate vicinity of the SMBH (e.g., Blandford 1994). Most of these disk-origin models of jets (e.g. Appl & Camenzind 1993; Chiueh et al. 1991) can be considered to be variants of the Blandford-Payne (1982) scheme. However, it should be noted that if the screw-instability argument of Li (2000) turns out to be valid, it probably also applies to disk-launched jets and would cause difficulties for any MHD dominated jet formation process.
One possible approach for producing asymmetric jets by a single central engine was proposed by Wang et al. (1992). They took a semi-analytical approach to the force-free Grad-Shafronov equation and found solutions in which the bulk of the power was carried by the Poynting flux, while most of the angular momentum in the jet was carried by the magnetic fields emerging from the accretion disk. Wang et al. (1992) found that substantially more thrust could flow off of one side of the disk than the other if sufficiently large asymmetries could be maintained in the magnetic field within the disk.
The idea that the accretion disk corona can generate two fundamentally different types of jet has been proposed recently (Meier et al. 1997; Meier 1999). Fast (highly relativistic, FR II) jets are ejected when the coronal plasma is unbound by the magnetic field, while slower (transrelativistic, FR I) jets, moving at roughly the disk's escape velocity, are produced when the corona is inertially bound to the SMBH. The original version of this "magnetic switch" mechanism (Meier et al. 1997) could explain how jets of very different speeds arise from otherwise similar sources, but this version fails to explain how these differences can persist over the extended lifetimes of FR I sources (Meier 1999). In his revised scenario, Meier (1999) argued that the difference in radio jet power among galaxies of the same mass arises from different speeds of rotation of the magnetic field lines associated with their central engines, which are in turn produced by different spin rates of their SMBHs. (The idea that the SMBH spin was critical had been put forward already by Baum et al. (1992, 1995) based on observational inferences about the merger of the host galaxy with a high angular momentum, gas-rich, disk galaxy in the case of FR II sources.) The transition occurs at a critical power when the MHD luminosity, , (where is the poloidal magnetic field, is the size of the magnetic rotator, and its angular velocity) exceeds a critical luminosity, defined as the liberation of an escape energy in a free-fall time:
In Meier's scenario, this magnetic switch luminosity plays the same role in MHD acceleration as does the Eddington luminosity in radiative acceleration. Note that this magnetic switch model relies on extracting substantial power from a portion of the accretion disk extending within the ergosphere, thereby avoiding the problem highlighted by Livio et al. (1999). This model is capable of yielding a decent match to the Owen-Ledlow variation in the value of with in terms of a critical SMBH spin (Meier 1999).
The possibility that Advection Dominated Accretion Flows (ADAFs; e.g., Narayan & Yi 1995), which are inefficient radiators, are present in FR I radio galaxies was first proposed by Reynolds et al. (1996a). In this picture, which is an intriguing option (e.g., Jackson 1999), standard thin accretion disks, which radiate more efficiently, yield FR II radio sources (Reynolds et al. 1996a). A good fit to the low-frequency radio and X-ray emission of M87 (Reynolds et al. 1996a) as well as to that of several quiescent ellipticals (Di Matteo & Fabian 1997; Mahadevan 1997) could be attained using ADAF models. However, the ADAF models grossly overpredict the high-frequency radio and sub-mm emission from these quiescent galaxies unless: the magnetic fields are very much below equipartition; or there is enough cold material for free-free absorption of the synchrotron absorption to be very important; or powerful winds remove much of the energy, angular momentum and mass from the inner part of the accretion flow (Di Matteo et al. 1999).
© European Southern Observatory (ESO) 2000
Online publication: December 11, 2000