Astron. Astrophys. 356, 73-82 (2000)

1. Introduction

Extensive HST emission-line imaging of Seyfert galaxies has for the first time resolved details of the structure of their Narrow Line Regions (NLR). In several cases cone-like morphologies have been revealed, similar in shape to - but of much smaller linear extent than - the Extended Narrow Line Regions (ENLR) seen in the lower resolution ground based images (Wilson & Tsvetanov, 1994 and references therein). In the standard model of the NLR, the UV emission of the nucleus is responsible of photoionizing the Interstellar Medium (ISM) of the host galaxy. These conical distributions of the ionized gas have been interpreted as a confirmation of the anisotropy of the nuclear radiation field which, in the framework of the unified scheme for Seyfert galaxies (e.g. Antonucci 1993) is caused by the shadowing of an obscuring circumnuclear torus. However, in galaxies with linear radio structures, the morphology of the emission-line region appears to be directly related to that of the radio emission. In particular, in Seyferts with radio jets (e.g. Mrk 3, Mrk 348, Mrk 6, Mrk 1066, ES0 428-G14), the NLR itself appears jet-like and is spatially coincident with the radio jet, while the emission-line region takes a different form when a radio lobe is present (e.g. Mrk 573, Mrk 78, NGC 3393): arc-like shells of emission, very reminiscent of bow-shocks, surround the leading edge of the lobes (Capetti et al. 1995a, 1995b, 1996; Falcke et al. 1996, 1998). This dichotomy in radio and emission-line morphology is reflected in their different scales: bow-shock structures cover several kiloparsecs, while the jet-like features extend only over a few hundred parsecs. The simplest interpretation of this radio-to-optical correspondence is that the radio emitting outflow creates an expanding and cooling gas halo. The compression induced by the outflow causes the line emission to be highly enhanced in the regions where the jet-cloud interactions occur. A clear confirmation of this scenario came recently from HST spectroscopy of Mkn 3 (Capetti et al. 1999): its NLR has velocity field characteristic of a cylindrical shell expanding at a rate of 1700 km/s. They interpreted this as the consequence of the rapid expansion of a hot gas cocoon surrounding the radio-jet, which compresses and accelerates the ambient gas.

HST observations also provided evidence for spatial variations in the NLR ionization structure. In NGC 1068 the material located along the radio jet is in a much higher ionization state than its surroundings. This might suggest the presence of a local source of ionization which dominates over the nuclear radiation field (Capetti, Axon and Macchetto 1997; Axon et al. 1998). In other sources, too distant for such a detailed analysis, the radial variations of the ionization parameter are generally much flatter than expected from pure nuclear photoionization on the basis of the measured density gradients (Capetti et al. 1996, Allen et al. 1999) requiring again a local source of ionizing photons. An appealing possibility of interpreting these data is to invoke the ionizing effects of shocks, originated by jet-cloud interactions: if these shocks are fast enough (velocities a few hundred km s^-1) the hot, shocked gas could produce a significant flux of ionizing photons (Sutherland, Bicknell & Dopita 1993; Dopita & Sutherland 1995, 1996). Direct evidence for this emission has been found by Axon et al. (1999) who showed that the radio-jets in the Seyfert 2 galaxies Mrk 348 and Mrk 3 are associated with an extended linear structure in UV and optical continua. In this picture, the radio-jet would not only determine the morphology of the NLR but is physically involved in its ionization. A radio imaging survey of the CfA sample of Seyfert (Kukula et al. 1995) shows that radio linear structures are present in a large fraction of sources (more than 50%) suggesting that such an interaction is likely to be a quite common phenomenon in this class of objects.

The jet interaction with the external medium is clearly a complex physical problem which involves both a hydrodynamical study of the jet propagation as well as a detailed understanding of the microphysics of the induced shocks, which might also be magnetized, and of the radiative processes.

In the framework of Seyfert galaxies this issue has been tackled by several authors (Dopita & Sutherland 1995, 1996, Evans et al. 1999, Wilson & Raymond 1999, Allen et al. 1999). Their focus is however mainly on the shocks properties with a very detailed treatment of the emission mechanisms, with simplifying assumptions about the hydrodynamics (e.g. plane parallel geometry, steady-state shock). The comparison with the observations is based on the emitted spectrum and in particular on diagnostic line-ratios, particularly with the aim of distinguishing the different signatures of nuclear versus local photoionization.

In this paper we follow a complementary, albeit different, approach by studying in detail the jet hydrodynamics, while adopting a simplified treatment of the radiative processes, as we employ an equilibrium cooling function in an optically thin approximation. This approach allows us to compare the results of simulations with the observed properties of NLR, in particular their morphology, the expansion velocities and the characteristic values of gas density and temperature. More precisely we consider the interaction of the jet with an inhomogeneity in the external medium (cloud) and our aim is that of constraining the jet and cloud physical parameters for which it is possible to reproduce the observed conditions. In this way, in addition of getting a better understanding of the NLR physics, we can also obtain information on the jet properties from the NLR data. Moreover, we can calculate the fraction of the jet power converted in radiation by shocks, resulting from the interaction of the jet with the environment. We then get from the global dynamics a conversion efficiency from kinetic to radiative power and we can determine whether the jet itself, via shocks, can provide an in situ photoionization source for the NLR emitting material, as discussed above.

Steffen et al. (1997a) have used a rather similar approach with the main difference that they considered the jet propagating into a uniform medium. It seems that in this situation it is impossible to reach the high densities typical of the NLR with jet-like emission (see discussion below) on which we will focus in the present paper. This is because, at low density, radiation is not efficient enough to give the needed compression factors. The conditions of the emitting material obtained by Steffen et al. seem to be appropriate for the case of the more extended (lobe-like) line emission structure.

Steffen et al. (1997b) considered also jet-cloud interactions mainly from an analytical point of view. They found that when a jet interacts with a large number of clouds the most relevant effects on the NLR structure are due to the most massive clouds located along the jet path. This lead us to our choice for the geometry of the simulation in which the jet hits a single dense cloud.

The paper is structured as follows: In the next section (Sect. 2), we describe the basic physical problem and the observational constraints, while the equations used and the method of solution are examined in Sects. 3 and 4; the results of simulations are discussed in Sect. 5; conclusions are drawn in Sect. 6.

© European Southern Observatory (ESO) 2000

Online publication: March 28, 2000
helpdesk.link@springer.de