Radio synthesis maps have shown jets in hundreds of AGN, on scales from subparsec to megaparsecs. The continuity of jets in direction indicates that the central generator has a memory over millions of years, and disk structures provide a natural way to control the direction of the jets. There is a vast literature on the collimation and production of jets; I will mention only a few points here.
How fast are they? Structures of jets can indicate their Mach numbers (with respect to the external medium), but not immediately their absolute velocities. Some sources look as if the jets are rather slow and flexible, while others look like highly relativistic blowtorches. We do not even know for sure whether we are seeing a phase or group velocity when motions can be measured. Strong evidence for relativistic bulk motions comes from superluminal sources, in which the projected speed of motion (always outward from the core) of distinct blobs is 1-10c. A natural explanation is (backwards) time dilation in material approaching us at ~0.9c; the Doppler boosting of this material would make these objects bright, so that the boosted sources were observed first. The emitter stays only slightly behind its earlier radiated wavefronts, so the projected motion is quite rapid (see Superluminal Radio Sources, ed. Zensus and Pearson, Cambridge 1987). The governing equations reflect the relativistic Doppler dilation and boost effects. If we consider the projected separation between a stationary core and a blob moving away from it at a rate c b at an angle q to our line of sight, the apparent transverse velocity will be
which has a maximum value vmax ~ gc. The apparent jet/counterjet ratio R (for physically identical jets) becomes
where xis related to the spectral index a by 3-a for a confined blob and 2-a for a continuous jet. The relativistic g factor does not appear in the ratio because it is identical for both components, so that the geometric factors alone are left. The data show v = 1-10 c for superluminal sources (and subluminals also exist, mostly for nearby and fairly low-power objects like M87 and Cen A).
This material is not always the conventional jet in an early stage; Barthel has shown that radio galaxies of large projected size (i.e. presumably viewed 90° to the jet axes) can have superluminal motions, and proposed an intermediate model in which material is initially ejected over a broad cone angle. Only the tiny fraction coming near our line of sight is boosted enough to see at high angular resolution, and it is this fraction that would exhibit superluminal motion. On larger scales, structures in the M87 jet a kiloparsec from the core have been found to show transverse velocities of 0.3 - 0.5c (Reid et al 1989 ApJ 336, 112; Biretta et al. 1989 ApJ 342, 128) - so far the only direct evidence that something in large-scale jets is moving at high speeds. In M87, HST imaging shows individual features with apparent transverse speeds anywhere from 0.6-6c (Biretta et al. 1999 ApJ 520, 621). These make sense for G~6 and a jet oriented within 19° to the line of sight, rather different than the visual impression shown in the image below (a rotated section of the Hubble Heritage picture). Light-time effects work to make planar features within a relativistic jet appear more edge-on than they really are, probably important for the region around knot A in the M87 jet which is often thought to be an internal shock front.
This issue - relativistic motion producing apparent superluminal motion - is not unique to radio galaxies and quasars. There is a class of galactic superluminal sources, associated with strong gamma-ray emission and evidently generated by accretion onto compact objects in a genuinely small-scale counterpart to the extragalactic cases (Levinson and Blandford 1996 A&AS 120, 129). In these instances, the distance to the source is not in serious dispute, using the galactic rotation curve and velocities of foreground H I clouds to estimate where they must lie. For the best-studied ones, GRS 1915+105, and GRO J1655-40, we have both the apparent separation velocity and the core-lobe separation on each side, giving extra data to fit the velocities. In both cases, the intrinsic velocity is close to 0.9c, with angles to the line of sight of 8-20°. These are in turn reminiscent of SS433 with velocities of 0.26c in jets which are highly collimated, precessing in a binary system, and cool enough to emit optical line radiation.
Work on radio galaxies with jets and ionization cones shows that there really are two different levels of collimation - a broad ionization cone, perhaps produced by an obscuring torus, and the much narrower collimated jets inside this cone (see the ESO Extranuclear Activity workshop).
Tracking the features within small-scale jets has revealed interesting complications. The paths are not always radial to the nucleus, usually taken as the source with the flattest spectrum in ambiguous cases. This is based on the general principle that synchrotron spectra are flattened at lower frequencies by self-absorption, so the densest plasma will have a flat or inverted spectrum. An interesting case is 3C 345, in which emission features repeatedly appear off the core and follow fairly consistent nonradial paths.This can be seen in Fig. 1 of Zensus et al. (1995 ApJ 443, 35, courtesy of the AAS) in which a new component appears in late 1985, brightens, and moves outward changing its relative position angle in the process:
Such motions, and the pronounced wrapped filamentary structure seen in nearby jets such as M87 and Centaurus A, suggest an important role for motion in helical patterns. This is easier to understand if much of what we see isn't physical blobs but enhancements in particle emission (perhaps linked to injection of particles) coupled with relativistic beaming. These objects are the ones most often detected in gamma rays. Survey with VLBA and incidence of superluminal/compact structures.
An important theme in jet studies has been the notion that BL Lacertae objects are dominated by the relativistically beamed emission from jets seen almost exactly end-on. This makes sense from viewpoints of energetics and host-galaxy properties, and predicts well-defined relations between the observed BL Lac counts and the luminosity function of the parent population which make sense if the parent population consists of the numerous FR I radio galaxies. In these cases, the jet emission is so strongly boosted that it becomes quite difficult to learn anything else about the source. Extended emission around blazars, while requiring high dynamic range to see in the presence of the strong core, is typically of about the luminosity and extent we'd see from FR I lobes seen end-on, giving some additional credence to this picture (Antonucci et al. 1986 AJ 92, 1). Antonucci reviewed these issues extensively in the14th Texas Symposium, 1989 Ann. NY Acad. Sci. 571, 180).
The magnetic field is well-ordered in many jets, as shown by polarization measurements. Synchrotron radiation can be very highly polarized (50%) if the field is globally ordered, and some sources apprach this level. The electric vectors show clear structure and alignment; an especially common pattern is for the field lines to be along the jet in the inner portions and transition to an azimuthal configuration farther out. This is seen in M87, as well as in PKS0521-36, for which I'll show my own 2cm observations (1986 ApJ 302, 296).
Many objects with jets, especially the powerful FR II radio sources with long and highly collimated jets, show hot spots - compact enhancements in brightness of the lobes. Cygnus A is a prime example. These may in turn have internal structure, and often have the flattest spectra (thus most energetic particle populations) in the extended lobes. They have been pictures as encounter surfaces between the jet flows and a mostly unseen surrounding medium, with compression of the magnetic field occurring and thus vastly increased emissivity. Some (such as Pictor A) have such high-energy electon populations that sychrotron emission continues through the optical into the X-ray regime.
At this point, there are only two important things we really don't understand about jets - how they get accelerated to begin with and how they manage to stay so well collimated. An overpressured, freely expanding jet would have much larger cone angles than we see even for highly relativistic motion, so such notions as magnetic confinement are attractive (especially since we know there's a significant field - because we see synchrotron radiation). A rought estimate (and fairly robust minimum value) for the field strength is the often-used equipartition value, which has equal energy density in particles and fields (which also almost exactly minimizes the energy density for a particular observed luminosity). Kellerman and Owen (in Galactic and Extragalactic Radio Astronomy, 2nd edition) give minimum-energy field Bmin = 1.5 × 10-4 q9/7 z-2/7S2/7 where the field is in gauss, the angular size q is in arcseconds, and flux density S in Jy. Typical values are 10-3-10-4 gauss.
Many of the same considerations applicable to relativistic jets in AGN also seem to apply to gamma-ray bursts and their afterglows. The energy requirements become much more tractable if the luminosity is enhanced by beaming, and some of the afterglow light curves suggest that indeed the beaming is within a fairly narrow solid angle (instead of the isotropic but beamed emission we'd see from a relativistic fireball).
To briefly review properties of gamma-ray bursts:
A good set of overview reviews is included in the December 1995 PASP. Cosmic gamma-ray bursts were discovered serendipitously in 1965, while searching for terrestrial bursts which would indicate violations of the nuclear test-ban treaty. This happened when the Vela satellites were orbited; one might suspect that there was a comparable Soviet program, but no public information seems to have been forthcoming. The directional accuracy of a single detector was (and remains) poor; the best positions for bursts use time-of-flight "triangulation" from multiple detections including interplanetary spacecraft (for the stronger bursts, since only small hitch-hiker detectors can ride on probes designed for other purposes). Bursts last from a couple of seconds to two hours; there is a wide variety of temporal structure, from smooth decays to highly structured quasiperiodic bursts. Before the launch of CGRO, it was widely believed that the bursts come from galactic neutron stars (from accretion events, starting with instabilities in accretion up to and including comet and asteroid impacts into the surface). However, with the surprising isotropy in distribution, cosmological models have gotten a new look. The major schemes here are of merging neutron-star binaries (note this is a guaranteed non-repeating event) or of some relative of a supernova outburst ("hypernova").
At cosmological distances, the energy release must be of order 1052 ergs (Paczynski 1995 PASP 107, 1167). The arguments for a distant origin were originally quite general - isotropy plus log N -log S behavior, which to galaxy people fairly shout ``Cosmologically distant!". In this case, the behavior with flux implies that these objects occur at a rate increasing with cosmic time (which makes some sense given that the number of neutron stars and the number of coalescing binaries should grow with cosmic time). For these models, you do not expect repeating bursts, since the source is destroyed.
The major development was, of course, detection of optical afterglows around a few GRBs with well-determined positions. These turn out to be in galaxies at redshifts up to z=4 (starting with the report in IAU Circular 6588 on GRB 970228 and now a mainstay of the literature). Chandra data show Fe line emission for GRB 991216 at z~1, with line widths indicating that this surrounding material is expanding at about 0.1c. The implied abudances fit for recent supernova ejecta, reinforcing a connection between (some?) SN and GRBs.
It is virtually unavoidable that such powerful sources of high-energy radiation entail relativistic expansion of any material unfortunate enough to be involved. Even isotropic (spherical) expansion will involve beamed radiation, such that we would see radiation utterly dominated by material within a small angle ~g on the spherical surface, with its apparent flux boosted by the same relativistic beaming factor as above. It is important to know whether the expanding is really isotropic or jetlike because this vastly changes the energetics of the whole explosion (by numbers of order 2g²) as well as the physical picture. One often-discussed signature of jetlike structure would be a break in the fading of the afterglow. In general, as a jet cools and becomes less relativistic, one expects a fading enhanced by a decline in the beaming factor, which would undergo a slope transition when the beaming factor is comparable to the jet's cone angle. A spherical fireball would have no such abrupt transition. One way to see this is to note that initially, for a highly relativistic expansion, the jet and sphere will produce identical observed properties. Some afterglows (such as GRB 991216) do appear to show such breaks in their light curves, although others show inconsistent breaks at various frequencies or none at all during the few weeks they can typically be followed. Sari (in the 5th GRB Symposium volume from Huntsville, 1999, p. 504) points out that the best evidence for jets is in the bursts with the greatest calculated (isotropic) luminosity, another bit of support for tightly beamed radiation being important.
This year's most popular picture for GRBs involves some class of supernovae which produce relativistic jets. If a neutron star has just been formed, the energetics are appropriate, and in fact there were some puzzling speckle data on SN 1987A that might, in hindsight, have been showing us high-speed blobs leaving the scene. A promising interpretation has a supernova producing either a black hole or a "hot" neutron star surrounded by a very dense (and short-lived) accretion disk, so that some material escapes relativistically at its poles.
Last changes: 3/2003 © 2000-3