2. Shock vs. AGN photoionization in HZRG ( 1)
2.1. Evidence of AGN illumination in HZRG
In the popular anisotropic illumination model (e.g. Fosbury 1989), it is postulated that quasars hidden in the cores of powerful radio galaxies illuminate the ambient ISM with intense cones of UV/X-ray radiation, with the radiation photoionizing the extended gas within the cones, leading to line emission. This model appears to be consistent with the observed properties of the majority of low-redshift radio galaxies. Most importantly, the line ratios measured in the nuclear regions and EELR of radio galaxies, and the trends in these line ratios, generally agree well with photoionization models (Robinson et al. 1987; Binette et al. 1996). The (weak) alignment of the EELR with the radio axes (Baum & Heckman 1989) and relatively undisturbed kinematics (Tadhunter et al. 1989) are also consistent with anisotropic illumination by the broad radiation cones predicted by the unified schemes. Although there are no radio galaxy EELR which show the clear cone-like morphologies seen in some Seyfert galaxies, this is likely to be due to relatively sparse and inhomogeneous distribution in the early-type host galaxies (Tadhunter 1990).
Anisotropic AGN-photoionization of the ambient ISM is also a viable model for high-redshift radio galaxies. The detection of scattered light from a hidden quasar (and also broad line components) in the polarized flux of several HZRG is strong evidence for the existence of luminous quasars which illuminate the ISM (e.g. Cimatti et al. 1996, Dey et al. 1996). Moreover, the presence of large diffuse halos of ionized gas in some HZRG, which extend far beyond the radio structures strongly suggests the existence of a quiescent ISM ionized by the central AGN (e.g. van Ojik 1995).
McCarthy (1993) constructed a composite radio galaxy spectrum from observations of galaxies with 0.1 3. Photoionization calculations reproduce the radio galaxy spectrum rather well and this was presented as an argument in favour of the AGN photoionization as the main ionization mechanism in radio galaxies. However, the composite was built from the spectra of very different objects covering a wide range in redshift, and it is dominated by one or two of the most highly ionized objects. Thus the comparison of this composite spectrum with the models does not resolve the issue of the ionization mechanism for the general population of HZRG.
Villar-Martín et al. (hereafter VMBF96) concluded that photoionization by the AGN can explain the positions of a large sample of HZRG in the CIV 1550/Ly vs. CIV 1550/CIII] 1909 diagnostic diagram. The sequence defined by the data can be parametrized in terms of the so-called ionization parameter, that is, the ratio of the density of ionizing photons impinging on the slab to the density of the outermost gas layer of the slab:
where c is the speed of light, is the density of the gas in the front layer and is the Lyman limit frequency. is the monochromatic ionizing energy flux impinging on the slab. By varying the ionization parameter it is possible to produce the variety observed in the UV line ratios of HZRG. A similar result is obtained at low z: the sequences defined by the optical line ratios of powerful radio galaxies are explained in terms of a sequence defined by U (Robinson et al. 1987).
2.2. Evidence of shocks in HZRG
A possible alternative ionization mechanism in these sources is the ionization by fast shocks produced by violent interactions between the advancing radio jet and the ambient gas: there is already clear evidence for such jet-cloud interactions in HZRG. Firstly, whereas the gas kinematics in most nearby radio galaxies are consistent with gravitational motions (Tadhunter et al. 1989; Baum et al. 1992), extreme non-gravitational motions are observed along the radio axes in the majority of HZRG (van Ojik 1995; McCarthy et al. 1996). Secondly, the extended emission line regions (EELR) are often not only aligned with the radio axis, but closely correlated in detail with the radio emission (Chambers et al. 1990; Miley et al. 1992; Rigler et al. 1992; van Ojik 1995). Even when there are no direct radio/optical associations, the degree of collimation seen in the narrow jet-like EELR along the radio axes of sources like 3C 368 and 3C 324 (Longair et al. 1995) is difficult to explain without invoking interactions between the line-emitting gas and the radio jets. The radio-optical asymmetries (McCarthy et al. 1991), the relationship between optical structure and radio size (Best et al. 1996), and the fact that the extent of the line-emitting gas is almost always smaller than that of the associated radio source (van Ojik 1995), provide further evidence for a close association between the radio plasma and the warm emission line gas.
Recently, we have made a detailed study of the EELR in a sample of low-intermediate redshift radio galaxies which show clear morphological evidence for jet-cloud interactions (Tadhunter et al. 1994; Clark & Tadhunter 1996; Clark 1996; Clark et al. 1996). This study provides clear evidence that jet-induced shocks determine the distribution, kinematics and physical conditions of the EELR along the radio axes of the objects in the sample. There is also evidence that shocks have an ionizing effect in these sources: in particular, the high temperatures indicated by the the [OIII]4363/(5007+4959) ratio, and the low HeII(4686)/H ratio measured in the extended gas, are more consistent with shock-ionization than AGN-photoionization. Even complex multi-phase photoionization models such as those presented recently by Binette et al. (1996) and Simpson & Ward (1996) cannot reproduce the measured values of these two line ratios.
With the above discussion in mind, it is imperative that the UV line ratios of HZRG be compared in detail with the predictions of both shock-ionization and AGN-photoionization models.
© European Southern Observatory (ESO) 1997
Online publication: June 5, 1998