4.1. Photoionization by the Nucleus: Wide Angle Collimation of Ionizing Photons
Emission-line gas is detected in some 85% of a radio flux-limited sample of radio galaxies, most of which are FRIIs (Baum & Heckman 1989a). The emission-line luminosity is remarkably well correlated with both the total radio and the radio core power (the latter having more scatter) over 4 orders of magnitude in luminosity (Baum & Heckman 1989b; Rawlings & Saunders 1991). This correlation indicates that the ultimate power source of both synchrotron and line radiation is the active nucleus. Baum & Heckman (1989b) find the luminosity ratio Ltotal narrow line/Lradio ranges between 0.1 and 3 in their sample. Therefore, if the gas is photoionized, there is at least as much ionizing continuum as radio luminosity in powerful radio galaxies. There are no steep spectrum, powerful radio galaxies with Lionizing rad'n << Lradio. The evidence that the emission-line gas in radio galaxies is photoionized by the nucleus is as follows:
(a) A strong correlation is found between the number of ionizing photons, Nion, emitted by the nucleus (as estimated by extrapolation of the nuclear non-stellar optical continuum) and the number of narrow-line, Balmer recombination photons, Nrec, emitted by the nucleus plus nebulosity (Baum & Heckman 1989a). A similar correlation is well known for Seyfert galaxies and quasars (e.g., Shuder 1981).
(b) Extended emission-line regions have similar emission-line spectra to nuclear narrow line regions, ranging from low excitation LINER-type ([OIII]5007 / H < 1) to high excitation Seyfert type ([OIII]5007 / H 3-10; cf. Robinson 1989).
(c) Photoionization models with power-law and hot ([1 - 2] x 105 K) black body continua provide equally good fits to the line ratios (Robinson 1989).
If the ionizing photons escape from the nucleus isotropically, then Nrec / Nion = C, where C is the covering factor of optically thick gas. For "reasonable" extrapolations of the featureless continuum, covering factors C 1 are required. However, the nebulosities are very patchy which, together with the high gas densities (Ne 102 - 103 cm-3) in the emission-line gas, strongly suggests C << 1. The implication is that the gas "sees" a much more luminous ionizing source than is inferred by extrapolation of the optical continuum seen at Earth. The extended gas in both radio galaxies and radio-loud steep-spectrum quasars is found preferentially, but not exclusively, in the two radio quadrants (Baum & Heckman 1989b; Fosbury 1989). Considered by itself, this last trend can be accounted for by excess gas in the radio quadrants, or greater gaseous compression by the radio jets and lobes. However, a more plausible explanation for both the energetic and morphological results is that the nuclear ionizing uv source is collimated along and around the radio axis. The ionizing radiation probably escapes in oppositely directed wide-angle cones, as increasingly found through direct observation of conically-shaped emission-line regions in radio-quiet AGN (e.g., Storchi-Bergmann, Wilson & Baldwin 1992). In Seyferts, the cone axes show a strong tendency to coincide with the axes of the "linear" radio sources, indicating that the collimating entities for the radio jets and the ionizing radiation are quite well aligned. These "ionization cones" may represent the intrinsic emission pattern of the central accretion disk (e.g., Madau 1988) or result from "shadowing" by a much larger (pc to hundreds of pc) scale dusty torus (e.g., Krolik & Begelman 1986). Probably the best method of discriminating between these two mechanisms is to search for the strong, thermal-infrared signal expected from a dusty torus which absorbs the radiation incident on it from a smaller, intrinsically isotropic, optical-uv-X-ray source. This energetic test has recently been performed for some 9 Seyfert 2 galaxies with "ionization cones"; the results favor the dusty absorber model of the collimation of the ionizing photons (Storchi-Bergmann, Mulchaey, & Wilson 1992). A similar procedure could probe the putative wide-angle collimation process in radio galaxies. Searches for dense molecular gas in the cores of radio galaxies would also be valuable.