3.1. Evidence for the magnetic `seeding` of the intergalactic medium due to normal' activity of nearby galaxies
A subset of spiral galaxies possess a synchrotron-emitting halo - best seen if the galaxy happens to be edge on. Two recently studied examples are NGC 4631, an edge-on spiral galaxy (figure 7), which has widespread enhanced star formation activity, and M82, the nearest, hence `prototypical' starburst galaxy which has pronounced outflow (figure 8). M82 has recently been found to possess a large scale poloidal-like magnetic field. These systems, especially M82, may provide important clues to the strength and origin of extragalactic magnetic fields. This is because halo magnetic fields are associated with outflowing winds, which are the consequence of supernovae and related stellar activity in galaxy disks, and thus can be anticipated on a priori energetic grounds, as Heiles (1987) has argued.
Figure 7. A radio-optical overlay of the galaxy NGC 4631, shows the optical image, and lines which indicate the projected magnetic field orientation in the galaxy's halo (source: Golla and Hummel 1994).
Figure 8. The projected magnetic field lines within the outflow region of M82, from de-Faraday-rotated polarization images whose sensitivity was 200 µJy/beam and angular resolution 15" (reproduced from Reuter et al 1994).
The observations illustrated in figures 7 and 8 show a number of interesting facts which need to be accounted for in a theoretical field regeneration model. They show that, in distinction to large scale field ordering parallel to galaxy disks, the out-of-plane magnetic field lines, at least in some zones, are directed out of the plane. Using assumptions of equipartition, Hummel et a1 (1991) estimate 8 µG for the magnetic field strength in the disk of NGC 4631 and 5 µG in its halo. The same authors estimate 8µG in the (smaller) halo of another galaxy, NGC 891, compared with 13 µG in the disk due to its higher synchrotron luminosity. Recent analysis of M31 (the Andromeda galaxy), which is seen partly face-on, reveals zones where the field appears to turn up out of the galaxy's plane (Beck et a1 1989). It is apparent that these halo fields originate in the galaxy disks, and propagate somehow into the halo.
A more dramatic example of halo magnetic field structure is in the 'starburst galaxy', M82. A small zone within 400 pc of M82's nucleus is undergoing intense star formation and is generating about 4 × 1010 L (Kronberg et al 1985). The resulting luminosity density makes it inevitable that outflow will occur in such a system. This active region emits H [O II]), CO, etc line emission, as well as x-rays, IR continuum radiation due to heated dust, and synchrotron radiation. Figure 8 shows one of the first magnetic field patterns seen in the outflow zone around a starburst galaxy, obtained from recent VLA observations of high sensitivity, and at a resolution of only 15". An extensive magnetized halo has also been detected around the active spiral galaxy NGC 253 (Carilli el a1 1992).
The vigorous stellar activity within the inner 1 kpc of M82 (Kronberg el al 1985), and its associated outflow of cosmic rays (Seaquist and Odegard 1991, Reuter et al 1992), x-ray emission (Kronberg et a1 1985, Schaaf et al 1989) and associated strong magnetic field (Reuter et al 1993) have made M82 a key testbed for explaining the magnetic field generation process in this and other galaxies with starburst-like nuclear activity. The 2-dynamo (section 2.3.4) proposed for the inner corotating zones of 'quiescent' galaxies, clearly does not apply here. Observations now make it clear that magnetic fields are expelled along with plasma at greater than M82's escape velocity, and that if Parker's reconnecting, detached magnetic loops from a spiral disk are correct, we may be seeing (figure 8) a more powerful version of a similar physical process.
M82 appears thus to be a nearby-universe testbed for the generation and outward transport of magnetic fields. The starburst produces a high velocity outflow, which is at most 108 years old. By contrast, the dynamo amplification time of the out-of-plane fields illustrated in figure 5 would need to be of order 10" years in the slow-acting mean-field dynamo model. Thus, the magnetic fields around M82 are unambiguously produced by the outflow phenomenon, which happens on a much shorter timescale than the dynamical age of the galaxy. This interesting discovery puts constraints on processes of magnetic field regeneration. It demonstrates that mechanisms other than the slow galactic scale dynamo are able to amplify and, as discussed below, perhaps create magnetic fields.
Lesch et al (1989a) have proposed an ingenious model for the generation as well as the amplification of large scale magnetic fields in the nucleus of M82 and similar systems. In their model, M82 can create a kpc-scale magnetic field by a galactic scale version of Biermann's battery, caused by the interaction of the galactic plane component of the nuclear starburst wind with the rotating, dense molecular gas torus which surrounds M82's nucleus. The latter has been revealed in both 21 cm H I observations (Weliachew et al 1984), and in 2.7 mm CO (1-0) observations (Lo et al 1985, Nakai et al 1987, Loiseau et al 1988). A hot, subsonic outflowing plasma from within the molecular ring excites sound waves propagating at cs = (5/3)1/2(kTe / mp)1/2, (k is Boltzmann's constant, Te the electron temperature, and Mp the proton mass). In the model of Lesch et al (1989a), these become compressional waves which, on arrival at the rotating torus, push the radially streaming plasma into the azimuthally rotating molecular gas. Because the electrons couple more easily to the gas than the protons, the differential stopping time (tsp = (mp / me)tse) causes a current to build up in the ring, which generates a poloidal magnetic field. This initial magnetic field is amplified by a combination of v × B compression and turbulence, a process described in detail by Alfvén and Falthämmar (1963), Parker (1979), and others. Lesch et al (1989a) point out that, in M82, the wavelength, s, of the sound waves is comparable with the dimensions of the system. The waves will reach the radius of the torus, RT 225 pc before being damped, they are then reflected back from the inner edge of the torus on a time scale tR 2RT / cs which is 107 yr. This is the time scale over which the battery current breaks down, so that as long as tR > the diffusion time tD(= L2 / 4 s), a current is maintained. (L is the characteristic scale). A consequence is that the ring current will prevail over the tendency of the plasma to cancel it. Lesch et al (1989) estimate tD 106 yr in M82's inner molecular ring, in which case the current, and hence a poloidal field (which is observed - see figure 8) will be built up.
Given the creation of the nuclear poloidal field by a galactic battery, the molecular ring, which marks the end of the co-rotation zone can probably `feed' field outward into the differentially rotating disk so that, over longer times, an - dynamo might be able to propagate and amplify the galaxy's disk field, as discussed in the preceding sections. A significant implication of the scenario proposed by Lesch et al (1989a), if correct, is that a galaxy can generate, and propagate its own, self-created field. The additional fact, just demonstrated, that starburst galaxies and their cousins are able to eject magnetized plasma into the ambient intergalactic medium presents, in principle, the possibility that seed fields in the early universe could originate in galaxies (cf section 5.4.1 for a related discussion).