5.1. The link with synchrotron X-ray emission
Chandra found X-ray synchrotron emission to be common in the resolved kpc-scale jets of FRI radio sources . The X-ray jets are readily detected in sources covering the whole range of orientation in unified schemes. The several tens of detected sources range from beamed jets in BL Lac objects [21, 157, 173] to two-sided jets in radio galaxies [46, 97], with most X-ray jets corresponding to the brighter radio jet, (e.g. 217, 96, 104, 143, 68, 221). Several of the observations have been targeted at sources already known to have optical jets, from ground-based work or HST. However, it's proved easier to detect X-ray jets in modest Chandra exposures than to detect optical jets in HST snapshot surveys, because of better contrast with galaxy emission in the X-ray band than in the optical .
Inverse Compton models for any reasonable photon field suggest an uncomfortably large departure from a minimum-energy magnetic field in most low-power X-ray jets (e.g., 96), although the beamed iC-CMB model is a contender for the emission from some BL Lac objects (e.g., 173). Otherwise synchrotron mission from a single electron population, usually with a broken power law, is the model of choice to fit the radio, optical, and X-ray flux densities and the relatively steep X-ray spectra (e.g., 29, 96). Given Equation 6, X-ray synchrotron radiation at 1 keV requires electrons of energy ~ 1013 eV (Lorentz factor 2 × 107) if the magnetic field strength is of order 20 nT (200 µG; the electron energy scales as B-1/2). Averaging over pitch-angle distribution, the lifetime of synchrotron-emitting electrons is given by
where me is the electron mass, T is the Thomson cross section, and uB is the energy density in the magnetic field. We thus see that electrons emitting 1 keV synchrotron radiation in a 20 nT magnetic field have an energy-loss lifetime of about 30 years (lifetime scales as B-3/2). The electrons must therefore be accelerated in situ, since their lifetimes against synchrotron losses are less than the minimum transport times from the active nuclei, or even from side to side across the jet. (This should not be the case if proton synchrotron radiation is important , since lifetime scales as (mp / me)5/2.) Particle acceleration is generally discussed for the cases of a particle interacting with a distributed population of plasma waves or magnetohydrodynamic turbulence, or shock acceleration (see e.g., 25, 66, 108, 5).
For electrons, particle acceleration and energy losses are in competition (e.g., 106), no more so than in hotspots of FRIIs (e.g., 34), which mark the termination points of the beam. Hotspots display considerable complexity in the X-ray, with synchrotron components seen in the less powerful sources indicating that TeV electrons are present (e.g., 98, 126). It has been suggested that the low-energy radio spectral-slope change seen in hotspots may mark a transition between electrons that are accelerated through electron-proton cyclotron resonance and those (at higher energy) that are simply undergoing shock acceleration (e.g., 192, 91). If in FRIs the far-IR spectral break consistently maps electrons of a particular energy, it is possible that the break here is also more related to acceleration than loss processes .
Whether or not particle acceleration is required along the jets of quasars depends on the emission process at high energies. If the beamed iC-CMB model holds, then the electrons participating in radiation at wavelengths currently mapped are generally of low enough energy to reach the end of the jet without significant energy loss, except if a relatively high level of optical emission must be explained as synchrotron radiation. The knotty nature could then be understood as variable output in the jet (e.g., 191). However, in nearby FRII radio-galaxy jets, where synchrotron X-ray emission is seen (Section 3.4), the need for particle acceleration is secure, and similar underlying processes are expected in quasars even where the synchrotron X-rays might be outshone by beamed iC-CMB emission.
Details of the regions of particle acceleration are best studied in the closest sources. Cen A (Fig. 15) and NGC 315 (Fig. 1) are particularly good examples of FRI jets where the X-ray jet emission is resolved across as well as along the jet, and X-ray knots are embedded in more diffuse structure [97, 221, 100, 223]. The fact that the X-ray emission is not just confined to regions within energy-loss light travel distances of the knots shows that particle acceleration can occur also in diffuse regions. The relatively soft X-ray spectrum seen in the diffuse emission in Cen A has been used to argue that something other than shock acceleration (proposed for the knots) might be taking place in the diffuse regions , although no specific explanation is suggested, and the competition between energy losses and acceleration may be more important here.
5.2. Particle acceleration in knotty structures
The model of jet deceleration through entrainment (Section 4) leaves unanswered important questions about the origins of the bright knots that appear in many jets, particularly FRIs, and that are usually interpreted as the sites of strong shocks. Radio studies have searched for high-speed knot motions, with apparent speeds greater than the speed of light having been noted in M 87 . A proper-motion study of the knots in Cen A over a 10-year baseline found that some knots, and even some more diffuse emission, travel at about 0.5c, indicative of bulk motion rather than pattern speed . This motion, coupled with the jet-to-counter jet asymmetry, suggests considerable intrinsic differences in the two jets, to avoid the jets being at an implausibly small angle to the line of sight.
Other knots in Cen A appear to be stationary, which might suggest that they result from intruders in the flow, such as gas clouds or high-mass stars (e.g., 75, 97). Some of these have emission profiles in the X-ray and radio that are unexpected from a simple toy model where the electrons are accelerated and then advect down the jet, losing energy from synchrotron radiation. Instead the bulk of the radio emission peaks downstream from the X-ray within these knots, leading to suggestions that both radio and X-ray-emitting electrons are accelerated in the standing shock of a stationary obstacle, and a wake downstream causes further acceleration of the low-energy, radio-emitting, electrons . The resulting radio-X-ray offsets, averaged over several knots, could give the radio-X-ray offsets commonly seen in more distant jets (e.g., 96, 219, 63).
The knots of Cen A are not highly variable in observations to date , but dramatic variability on a timescale of months is seen in a knot in the jet of M 87, and the X-ray, optical and radio light curves are broadly consistent with shock acceleration, expansion, and energy losses, although the timeline is currently too short for strong conclusions to be drawn .
It is important to study the location of jet knots within the flow, to see if that can provide a clue as to their nature. A particularly interesting example is NGC 315 . Here the diffuse emission contains a knotty structure in the radio and X-ray that appears to describe an oscillatory filament (Fig. 1). Although the structure could be the result of a chance superposition of non-axisymmetric knots, the level of coherence led to suggestions that the knots might be predominantly a surface feature residing in the shear layer between the fast spine and slower, outer, sheath plasma. If this interpretation is correct, we might expect the X-ray spectra of the knots to be similar across the transverse width of the jet. However, the distinct knotty emission is only about 10% of the total in X-rays and radio along the ~ 2.5 kpc of projected jet length over which it is detected, and with a source distance of ~ 70 Mpc the observations did not allow the spectra of the knot and diffuse emission to be separated.
At 3.7 Mpc, Centaurus A is a much closer example of an FRI radio galaxy whose knots and diffuse emission are seen over a similar projected linear distance to that of NGC 315. An X-ray spectral study of Cen A's knots found a spectral steepening with increasing lateral distance from the jet axis, disfavouring these knots all residing in a shear layer . A flatter X-ray spectrum is seen more central to the flow, and an alternative explanation to acceleration in stationary shocks is that the knots here might be formed by stronger turbulent cascades with more efficient particle acceleration. Knot migration under the influence of the shear flow might then be expected, and proper-motion studies might then distinguish between this interpretation and stationary shocks from stellar or gaseous intruders entering the flow .
5.3. Incorporating polarization data
There are no current X-ray missions with polarization capabilities. However, the radio and optical bands probe electron populations responsible for the X-ray emission, albeit at different electron energies. If the emission is synchrotron, polarization data provide our best handle on the direction and relative degree of alignment of the magnetic field. Radio observations show that the fields are relatively well ordered, although there is much complexity. Broadly, the magnetic fields in FRII jets tend to be parallel to the jet axis, whereas in FRI jets they are either predominantly perpendicular, or perpendicular at the jet centre and parallel near the edges, with the mixed configurations pointing to perpendicular fields associated with shocks and parallel fields from shear or oblique shocks .
Optical polarization measurements of resolved jet structures have been made with HST. So far these have mostly concentrated on nearby FRI radio galaxies, where the optical features are brighter and the emission mechanism is synchrotron radiation (for an atlas of polarization images see ). Work is under way to explore optical polarization in the jets of FRII radio galaxies and quasars. As mentioned in Section 3.4, the optical emission should be essentially unpolarized if it is an extension of a beamed iC-CMB X-ray component, in contrast to being of synchrotron origin.
The first jet to be studied in detail in both its radio and optical polarized emission was M 87, where there is evidence for strong shock acceleration in compressed transverse magnetic fields at the base of bright emitting regions, although the polarization fraction becomes low at the flux maxima . Significant differences between the polarization structures seen in the optical and radio suggest that the sites of acceleration are different for different electron energies, with the strongest shocks, that provide acceleration to the highest energies, appearing in the most central parts of the jet . Detailed work on 3C 15 shows a jet that narrows from the radio to the optical to the X-ray, showing that acceleration to the highest energies occurs more centrally to the flow, and a mixture of strong shocks and stratified flows can account for the broad features seen in the optical and radio polarization .
A third source for which optical and radio polarization data have been important is 3C 346 (Fig. 16). Here X-ray emission is associated with a bright radio and optical knot where the jet bends by 70° in projection (the X-ray emission peaks somewhat upstream of the radio, as seen in other sources), leading to a suggestion that the bending and X-ray brightening are the result of a strong oblique shock located in the wake of a companion galaxy . Polarization data has supported the model by revealing a compressed and amplified magnetic field in a direction consistent with that of the proposed shock, in both the radio and optical ( and see Fig. 16).
Figure 16. 3C 346. Upper: Schematic showing an oblique shock formed in the wake of the passage of a companion galaxy to 3C 346, and how it affects the radio jet, from . Circles are the galaxies and red marks the path of the radio jet. Lower: Radio intensity contours and polarization vectors (rotated through 90° roughly to represent the magnetic-field direction) on a smoothed Chandra X-ray image, indicating compressed field lines aligned with the proposed shock, from .