In most cases, the kinematics of emission-line gas in low redshift radio galaxies appear to be predominantly gravitational in origin - the velocity amplitudes and velocity fields are consistent with those expected from the mass distributions of early-type galaxies (e.g., Tadhunter, Fosbury & Quinn 1989; Baum, Heckman & van Breugel 1992). However, a significant minority of galaxies show emission-line features clearly associated with the radio jets or lobes (e.g., Miley et al. 1981; van Breugel et al. 1984, 1985a, b, 1986; Heckman et al. 1982, 1984; Tadhunter et al. 1986; Tadhunter 1991). The most spectacular interactions between radio jets/lobes and thermal gas are seen in radio galaxies at high redshift, where the emission-line, optical continuum and radio continuum distributions are almost always closely aligned (e.g., McCarthy et al. 1987; Chambers et al. 1987; McCarthy et al. 1990; Chambers & McCarthy 1990). This alignment effect in high redshift radio galaxies seems to result from star formation triggered by the radio jet, although polarization measurements show that scattered light from the nucleus may also contribute to the extended optical emission. Because of space limitations, I shall discuss only low redshift galaxies in the present article. The images of 3C277.3 (Coma A) in Fig. 2 illustrate many of the relationships between radio intensity, radio polarization and emission-line gas.
Figure 2. (from van Breugel et al. 1985a). Various radio/optical overlays of 3C277.3 to illustrate the morphological relationships of the line-emitting gas and the radio continuum of this source. Panel 1: Gray scale plot represents H + [NII] 6548, 6583 emission with no continuum subtracted. The contours represent the 21 cm total intensity. Panel 2: Gray scale plot represents the continuum subtracted H+[NII]6548, 6583 image. Contours are the polarized intensity at 6 cm. Panel 3: Gray scales represent continuum subtracted H+[NII]6548, 6583 near the northern hotspot (H). The contours represent the total intensity at 6 cm. Panel 4: As in panel 3 for the nucleus (N) and the knots (K1, K2) in the jet. Panel 5: Gray scales represent the percentage polarization at 6 cm. The contours are the 6 cm total intensity. The bars represent the radio polarization directions. Panel 6: As in panel 5 for 21 cm. See original paper for full details.
While each galaxy has its own individual characteristics, a number of generalities can be made about the gas associated with jets and radio lobes, as I now summarize.
(i) The emission-line gas is local gas interacting with the jet/lobe, not material carried out by the jet. This conclusion is favored by a number of observations. First, the ratio [NII]6583 / H is sometimes lower for the extranuclear gas than the nucleus itself, consistent with a lower nitrogen abundance at large nuclear distances. Second, the radio continuum and optical line morphologies suggest the jet has collided with, and been deflected by, a dense cloud. Third, in some galaxies there is other evidence for cool (HI 21cm emission and/or absorption) or hot (extended soft X-ray emission) gas and dust (obscuration and reddening) enveloping the radio source. Fourth, the line-emitting gas tends to be within 50 kpc of the nucleus (i.e., a reasonable extent for an interstellar medium), even when the total extent of the radio source is much larger.
(ii) A jet may be deflected by an interstellar cloud. First, in the vicinity of an emission-line cloud, a jet may change its apparent direction of flow, the radio emission brightens, the sharpest radio intensity gradients are found on the upstream side of the cloud and the radio continuum and emission lines are not coincident. Second, there are strong kinematic effects - velocity jumps and broadened lines - at or near the cloud. Third, thrust arguments (based on the equation Mc Ttc / Vc, where Mc is the mass of the cloud and tc is the timescale for it to be accelerated to Vc) indicate that the clouds may be massive enough to deflect the jet. Typical masses inferred for the ionized component of the gas clouds are 106-107 M.
(iii) The gas is probably photoionized. When measured, the electron temperature is too low (Te < 20,000K) for collisional ionization in shock waves. Ionization by relativistic particles is energetically possible, but not favored, because the radio continuum and the line emission are usually displaced from each other. This displacement is, however, not a conclusive argument against cosmic ray ionization because such ionization would be dominated by low energy cosmic rays, which may be more widely distributed than the high energy electrons responsible for centimeter wavelength synchrotron radiation. In general, the line ratios are consistent with photoionization by a source with a power-law spectrum. However, photoionization by an isotropic nuclear source can often be ruled out, since extrapolation of the observed nonstellar optical or uv continuum to ionizing energies provides too few ionizing photons. There is increasing evidence that the nonstellar nuclear source is anisotropic and partially beamed in the direction of the jet (e.g., Wilson 1992). Alternatively, ionizing radiation could be produced "in situ" by the jet itself. About half a dozen jets have been detected in optical continuum light (Keel 1988). This emission is generally presumed to be synchrotron radiation (probably from electrons accelerated to 106 in shocks resulting from collisions with a cloud or from other non-steady phenomena in the jet) and may extend into the uv. However, extrapolation of the observed optical fluxes using the radio to optical spectral index will likely overestimate the ionizing flux, since the optical spectra tend to steepen towards higher frequencies (Keel 1988). Further discussion of the ionization of the gas is deferred to Section 4.
(iv) The radio emission is depolarized by an irregular "Faraday screen" on the outskirts of the radio source. The evidence here is the finding of emission-line gas preferentially in boundary layers around the radio lobes, rather than uniformly distributed within them. For some galaxies, there is a clear anti-correlation between the distributions of percentage radio polarization and H flux. In actuality, the depolarization may be dominated by a lower density, more diffusely spread gas than that detectable in emission lines.
(v) Jets may entrain surrounding gas. The large kinematic gradients observed along some jets are strong evidence for interaction with the surrounding gas. Second, the gas is usually found along the edges of jets. Last, there is often an approximate equality between relativistic and thermal pressures (B2 / 4 nkT), consistent with a mixing of the two gases.
(vi) The emission lines originate from dense clouds into which radiative shocks have been driven by the jets and lobes. In general, the density of the optical line-emitting gas cannot be reached through cooling at constant pressure of a uniform medium with density inferred from the thermal bremsstrahlung X-ray emission. This result suggests that the observed emission-lines result from collisions between the jets and clouds with above average density. Further, there is strong evidence that many powerful radio galaxies are recent mergers, as indicated by optical "tails" or "fans" (e.g., Hutchings 1987), unusually blue extranuclear colors (Smith & Heckman 1989), and both lower stellar velocity dispersions and more rapid rotation in the morphologically peculiar radio galaxies than in normal ellipticals (Smith, Heckman & Illingworth 1990). Mergers involving at least one gas-rich galaxy are probably the source of the dense interstellar clouds.
(vii) The total emission-line luminosity of FRII radio galaxies is comparable to the total radio luminosity. In general, most of this line emission is not spatially associated with a jet. However, if the kinetic energy of the jet is ultimately responsible for the ionization and excitation of gas which is associated it, upper limits to the efficiency with which the jet produces radio emission may be established. Alternatively (and more likely - see Section 4), the gas may be photoionized by the nucleus (e.g., Baum & Heckman 1989a, b) and the line luminosity probes the intensity of ionizing radiation emitted in the direction of the emission-line cloud (if the cloud is optically thick).
(viii) The gas velocities tend to be blueshifted with respect to systemic on the side with the brightest radio jet. This result indicates that the one-sidedness of jets on the kpc scale is a result of Doppler boosting.