It is expected that, in most cases, the rate of advance of the head of the jet is supersonic, so a bow shock wave will be present. However, the visibility of the bow shock is strongly dependent on the physical parameters. Conditions favorable to the visibility of the bow shock are: (i) a jet of sufficiently low power that the bow shock is slow enough for the postshock gas to cool to 104 K in less than the dynamical time, (ii) a high density interstellar or intergalactic medium (which leads to fast cooling and a dense, post-shock shell), and (iii) the presence of a luminous nuclear source of ionizing photons which will prevent the post-shock gas from cooling below 104 K and recombining. All of these conditions are found in Seyfert galaxies. Extended, high excitation emission-line gas in Seyferts tends to be aligned, and co-spatial, with the "linear" radio sources (e.g., Haniff, Wilson & Ward 1988). Further, the emission-line gas is found to be kinematically disturbed by the radio components (e.g., Whittle et al. 1988). Unfortunately, these nebulosities are usually too compact ( few arc secs or few hundred pc scale) for ground-based optical observations to separate emission from physically different components (such as the bow shock or a jet). In a few cases, however, bow shock-shaped morphologies are seen in both the radio synchrotron and optical emission-line distributions. Examples include M51 (Crane & van der Hulst 1992; Cecil 1988), as shown in Fig. 1, NGC 1068 (Wilson & Ulvestad 1987; Cecil, Bland & Tully 1990), and NGC 3516 (Miyaji, Wilson & Pérez-Fournon 1992). Pedlar et al. (1985) and Taylor et al. (1992) have developed models in which the expanding or outwardly moving radio components sweep up shells of gas which are photoionized by the uv continuum from the Seyfert nucleus. In related work, Wilson & Ulvestad (1987) modeled the radio emission of the NE lobe of NGC 1068 as synchrotron radiation from ambient magnetic fields and cosmic rays which have been swept up and compressed by a radiative bow shock driven into the interstellar medium by a jet. Emission-line gas is also found associated with this bow shock (Cecil, Bland & Tully 1990). The essential difference with the model of Pedlar et al. concerns the origin of the radio emission: in their model the radio emission comes from a classical radio lobe or jet, rather than compressed interstellar material. I conclude this Section with a summary of some current questions regarding bow shock models for the kpc-scale line emission of AGN.
Figure 1. (from Cecil 1988). Images of the circumnuclear region of M51. The top grayscale image represents the emission in [NII]6583 at +100 km s-1 with respect to systemic. Overlaid on this image are radio contours at 6 cm. At the bottom, the radio image smoothed to the same resolution is shown as a grayscale. The insert shows the unsmoothed radio map, together with contours of the [NII] flux from the +580 km s-1 (relative to systemic) channel map. Note the similar optical and radio morphologies, but the displacement of the most intense radio emission to the inner boundary of the optical. The sinuous radio jet extends from the nucleus and coincides with the axis of the high-velocity outflow.
Observations (a) What are the precise relative locations of the radio synchrotron and emission-line gases? A combination of images in emission lines from HST and high resolution radio maps can probably answer this question. Such observations should resolve the issue of whether the radio emission comes from close to the shock front itself (implying active particle acceleration at the shock, with the emission-line gas displaced a cooling length downstream), from the post-shock cooling zone, or from a classical radio lobe or jet (with the shocked interstellar gas around it). (b) What kind of physical conditions pertain in the post-shock gas? Emission line ratios should be measured at high spatial resolution (to separate the bow shock related emission from other components) to determine the gaseous density, electron temperature and excitation.
Modelling (c) Is "passive" compression (i.e., without Fermi acceleration of relativistic electrons at the shock) of ambient disk magnetic fields and cosmic rays sufficient to account for the synchrotron powers observed? (d) What are the key emission-line diagnostics of gas which has passed through a radiative shock, but is continuously illuminated by an external, hard, uv continuum (from the nucleus)? Intuitively, one might expect emission-line ratios similar to those of a classical "Seyfert 2" spectrum (for which ionizing radiation is considered to be the sole energy input), but with an elevated electron temperature. (e) Is there evidence for super-solar abundances, as might be expected for ambient gas in the central regions of a spiral galaxy?
Theory (f) Is it feasible to study the evolution of the jet and bow shock with 3d radiative hydrocode calculations, which cover a wide range of Mach numbers and density ratios and also include the important effects of nuclear photoionization? (g) Under what conditions are the dense post-radiative shock shells unstable (cf. Innes et al. 1987; Gaetz et al. 1988; Blondin, Königl & Fryxell 1989)?