3.2. Injection of magnetic fields into the intergalactic medium by extragalactic radio source jets and lobes

The supra-galaxy scale morphology of double radio sources, their associated large energies (Burbidge 1956), and the high collimation of their radiating jets make them a remarkable phenomenon of the extragalactic universe. Large scale jets and associated radio source structure (reviewed recently by Bridle 1991) also have their analogues on stellar scales, e.g. in binary star systems and protostellar objects, where a magnetized accretion disk produces a highly collimated outflow (cf Mestel 1972, Lovelace et al 1991, Stone and Norman 1992a, b, Norman and Heyvaerts 1990 for a review). Recent observations (e.g. Yusef-Zadeh et al 1990) reveal associated synchrotron emission in stellar jet systems, thereby providing direct evidence of magnetic fields associated with these stellar jets. These reinforce the idea that the physical mechanisms operating in these stellar systems are generically similar to those in extragalactic jets. Figure 9 shows, from VLA observations, the highly collimated nature of an extragalactic radio source jet, and its coherent magnetic field structure as illustrated by its linearly polarized synchrotron emission.

Figure 9. Radio image of the radio galaxy NGC 6251, showing the strong, and ordered variations of magnetic field direction. The lines show the projected magnetic field orientation (from Perley et al 1984b).

3.2.1. Magnetic fields in galaxy-scale jets. Significant impetus to our understanding of extragalactic radio jets was given by a model proposed by Blandford and Rees (1974), which could reproduce the collimation and large energy transport into the outer lobes of extragalactic radio sources (> 10⁶⁰ erg in 10⁸ yr) by means of a jet containing a light, relativistic gas. In Blandford and Rees' model, the highly shocked termination point of the jet forms a 'working surface' from which energy is dissipated into the inflated radio lobes surrounding the jet. The jet may well have bulk relativistic motion, although the Lorenz factor of the bulk motion is a topic of current investigation (cf Begelman 1992 for a recent discussion of this point). However the large magnetic energy associated with the radio lobes, and the details of the polarization and filamentary structure revealed over the last dozen years (see e.g. Carilli et a1 1988) cannot be explained in detail without incorporating magnetic fields and magnetic field regeneration into the models. To achieve a physical understanding of these non-linear systems, detailed computational MHD modeling of all stages of the phenomenon is required, from the sub-pc scale accretion disk to the 100 kpc-scale radio lobes (cf Balbus and Hawley 1991, Begelman el al 1984, Begelman 1993, Lovelace 1976, Priest 1985, Pudritz and Norman 1986, Asseo and Sol 1987, Benford 1987, Camenzind 1987, Hawley and Balbus 1991, Königl 1989, Fraix-Burnet and Peltier 1991, Matthews and Scheuer 1990, Uchida 1990, Norman et a1 1991, Clarke et al 1992, Romanova and Lovelace 1992).

A current, idealized MHD model of a jet works roughly as follows (see figure 10(a)). A beam current, which generates a toroidal magnetic field, flows close to the jet axis. Magnetic field zones occur within nested `surfaces', and it is along these surfaces that the plasma actually flows. A return current, which balances the `beam current', flows along the outer magnetic sheath (figure 10(a)). The rotation of the field lines gives rise to electric fields which are perpendicular to the magnetic surfaces (Lesch et al 1989b). The beam terminates in a hot spot, or `working surface' (figure 10(b)), beyond which is the contact discontinuity separating the jet material from the ambient IGM. The outermost discontinuity is the bow shock due to the supersonic motion relative to the IGM.

Figure 10. (a) Idealized Structure of relativistic MHD jet, showing the magnetic sheaths and current flow. (b) The termination and outer shock zone where a relativistic, magnetized jet terminates against an outer shock zone in the ambient IGM (adapted from Lesch et al 1989b).

The evidence is now persuasive that magnetic fields are integral to the functioning of extragalactic jets. Their function is twofold; (i) The interaction between magnetic and gas dynamical forces, described by solutions to the MHD equations (equations (2.1)-(2.6)), can be shown to make the jet function as a highly collimated energy flow conduit over distances approaching intergalactic dimensions. The magnetic field distribution is largely toroidal relative to the jet axis, as illustrated in figure 10(a). (ii) Magnetized jets may also serve as accelerators of cosmic rays which, by synchrotron radiation, illuminate both the jets and the large, inflated radio lobes which are fed by the jet. The radio spectra in jets give direct evidence that relativistic electrons must be either accelerated in the jet, or else be transported in a very low B-field environment from the galaxy nucleus, since their synchrotron radiation lifetimes are often much shorter than the transport time from the base of the jet (i.e. the galaxy nucleus, accretion disk, etc).

Various possibilities exist for relativistic particle acceleration in jets: acceleration can occur by magnetic reconnection (a collisionless process) in the electric field of the neutral layer close to the jet axis (see figure 10(a)), as a result of non-axisymmetric flows, and turbulence - cf Romanova and Lovelace (1992) for a recent discussion of reconnection acceleration. A magnetic reconnection mechanism for accretion disks has also been recently proposed by Lesch (1991). Particle acceleration has also been proposed to be due to a combination of magnetic forces and Kelvin-Helmholz type velocity shear instabilities (Romanova and Lovelace 1992, see also Königl and Choudhuri 1985, Eilek and Hughes 1990). In addition, Fermi acceleration due to particle-Alfvén wave scattering in shocks and regions of plasma turbulence might be operative (cf Axford 1977, Blandford and Ostriker 1978), provided that the particles' average gyroradius r_g << r_s, the scale of the shock discontinuity, or r_g << _A the minimum Alfvén wave wavelength. Fermi-type mechanisms might be efficient for cosmic ray acceleration in strong shocks in radio hotspots (Meisenheimer et al 1989). However their relative effectiveness for electrons in extragalactic jets is unclear.

3.2.2. Convergence of observation and theory of radio jets and lobes.     Extended radio jet sources have been broadly classified into two groups by Fanaroff and Riley (1974). `FR I' sources have lower radio power (< 10²⁵ W Hz^-1 at 1.4 GHz), are center- rather than edge-brightened, and have two-sided, more diffuse jets in which the line-of-sight averaged magnetic field is predominantly transverse to the jet axis. These merge into diffuse, edge-fading outer lobes. The higher power FR II sources have strongly emitting outer lobes which are much more abruptly bounded than the FR I radio sources, and have higher surface brightness radio knots (cf Bridle 1986). When they have luminous jets between the quasar nucleus and (usually just one) outer lobe, the magnetic field orientation is along the local jet direction. Within the past 5 years or so, MHD simulations have been able to at least qualitatively explain these characteristics, thanks to leading edge computing power (see e.g. Clarke et a1 1989). Figure 11 shows the result of a recent 3-D jet-radio lobe MHD simulation (Clarke 1993). Simulations of this sort are able for the first time to reproduce the magnetic field morphology in reasonable detail within jets and lobes, including the filamentary structure (cf also Rees 1987). These have recently emerged in high dynamic range and high resolution VLA images (e.g. Perley et a1 1984a for the Cygnus A radio galaxy, and Dreher and Feigelson 1984 for the radio galaxy 3C353).

Figure 11. Simulation of a jet-cloud combination (from D. A. Clarke 1993). (a) shows a sketch of the salient features, (b) shows the total emissivity (Stokes parameter I) . (c) shows a grey scale of the magnetic pressure ( |B2|), and (d) shows the velocity shear modulus integrated along the line of sight.

3.2.3. Radio jets and lobes as sources of intergalactic magnetic fields?     The answer to this important question probably lies in the results of MHD simulations such as in figure 11, which can successfully reproduce the observed emission filaments and polarization structure of the radio lobes of some sources. The simulations confirm that significant magnetic field amplification occurs when Kelvin-Helmholtz instabilities in a shear layer are amplified by the kinematic (fast) dynamo. This is a direct consequence of the induction equation (equation (2.7)). As the field is amplified - by a few to 100 times in the simulation in figure 11 - the shear layer is transformed into filaments, as required by the solution to the magnetic induction equation. In a flat shear layer where shear is in the x-direction, and y is the direction normal to the shear, equation (2.7) (with = 0) reduces to

(3.1)

where l is the length of the filamentary ('roller') eddy, and t is its lifetime. Transsonic flow in the lobe is assumed, i.e. v_x ~ c_s, the speed sound. Then, since t ~ / c_s, (3.1) can be rewritten as

(3.2)

(Clarke 1993). This illustrates how the kinematic dynamo amplification of the magnetic field can be directly inferred from the observed width-to-length ratio ( / l) of the filaments (figure 11). These ratios range typically up to a few tens. This fairly unambiguous, and remarkable result confirms that extended radio lobes amplify magnetic fields (also predicted by De Young 1980 and Ruzmaikin et a1 1989), thus explaining their strong synchrotron radiation over large volumes of intergalactic space. Given the large size of extended lobes of radio galaxies and quasars, which range from a few kpc to ~ 1 Mpc, it is demonstrated how the combination of strong radio source jets and lobes can magnetize the intergalactic medium, as Rees (l987), Ruzmaikin et al (1989) and Jafelice and Opher (1990, 1992) have suggested.

The substantial contrast in |B| mentioned above in the radio lobe filaments has two further implications; it suggests that, since such filaments are dynamic, some particle acceleration likely occurs throughout the lobes (i.e. not just in the jets and hotspots). Also, the relativistic electron lifetimes ( B^-2 E^-1) will vary significantly within the lobes, which also makes it likely that the cosmic ray diffusion rates will be different within and outside of the filaments, but within the lobes.

Due to our limited knowledge of the cosmic occurrence rate of extended radio sources at earlier cosmological epochs, it is difficult to calculate the global effect of radio sources on the general IGM field. Daly and Loeb (1990) have produced calculations which yield B_IGM 10^-11 G. If the seed fields of galaxies came from previous generations of extended EGRS, then `EGRS-seeding' of the IGM must have been efficient at least by redshifts of 3 . Very extended (required for efficient seeding) radio sources are generally not found beyond z 2. It is not yet clear whether they are a phenomenon of only the `mature' universe, or whether they did exist in the past, but their synchrotron-emitting electrons were `snuffed out' by inverse-Compton scattering off the much denser photon field in the earlier universe. In that case, their fields would remain, and would only be illuminated by synchrotron emission at very low frequencies, which would be redshifted to still lower frequencies at our epoch of observation. Another potential galaxy-generated source of an IGM field, namely starburst galaxies and their analogues, is discussed in section 3. The existence and generation of magnetic fields in the IGM in galaxy clusters (ICM) is discussed in section 4.