5. THE INNER JET

Several speculative models have been proposed to connect the inner jet with the radio jet observed with VLBI. Here I summarize the most popular ones and relate them to past and future observations.

As discussed by Begelman, Blandford, & Rees (1984), much of the collimation of the jet must occur close to the central engine, since external pressure gradients are insufficient to focus a flow to within an opening angle of 1° to 3°. The collimation most likely involves the twisting of magnetic fields caused by the differential rotation of the accretion disk (see Blandford's contribution to these proceedings). If there is only partial focusing, pressure gradients can do the rest. In this case, the internal energy of the relativistic plasma is converted to bulk flow energy as long as the jet is confined and the internal energy of the plasma significantly exceeds the rest mass energy. The geometry of such a jet is shown in the top panel of Figure 6. Another possibility, depicted in the bottom panel of the same figure, occurs if a magnetic dynamo mechanism generates a highly relativistic (Lorentz factors of 10² to 10⁴) beam of electrons and positrons streaming along the polar magnetic field lines of the accretion disk (Lovelace 1976; Blandford & Znajek 1977). Phinney (1987) and Melia & Königl (1989) have shown that up-scattering of the uv photons emitting by the accretion disk decelerates the electron-positron stream to a terminal bulk Lorentz factor ~ 10. The scattering, along with plasma instabilities, can also randomize the pitch angles such that the beam becomes a flowing plasma by this point, which corresponds to the core of the radio jet.

Figure 6. Two basic models for the inner jet that connects the compact radio jet with the central engine, here depicted as a massive black hole with an accretion disk. Presumably, a second jet extends to the left of the central engine as well. The figures are schematic cross sections and not drawn to scale. The arrows inside the jet correspond to the magnitude of the bulk Lorentz factor of the jet flow on a logarithmic scale. The right-hand side of the jet corresponds to the left-hand side of the radio jet depicted in Fig. 3. The primary emission mechanism of each region is indicated above the jet, with "SSC" corresponding to synchrotron self-Compton emission. The main frequency bands of the nonthermal emission are indicated below the jet, with synchrotron or Compton reflection on top and self-Compton on bottom. Upper panel: Jet flow accelerates as internal energy is converted to bulk flow energy (e.g., Marscher 1980; Maraschi et al. 1992). Lower panel: A highly relativistic stream of electrons and positrons reflects ultraviolet radiation from the accretion disk, up-scattering the uv photons to X-ray and -ray energies. The interaction causes the beam to decelerate to a final Lorentz factor ~ 10 (Melia & Königl 1989; Phinney 1987).

The radiation expected from an accelerating inner jet has been explored by Marscher (1980), Reynolds (1982), Ghisellini, Maraschi, & Treves (1985), Celotti, Maraschi, & Treves (1991), and Maraschi, Ghisellini, & Celotti (1992). If the acceleration of the relativistic electrons occurs only in the region closest to the central engine, the synchrotron emission at uv, optical, and IR frequencies is confined to this region as well, which is opaque to radio emission. Outside of this UVOIR region, the highest energy electrons emit only at lower frequencies, having suffered from radiative and adiabatic losses. However, if the inner jet does not open too abruptly, the maximum radio emission occurs where the Lorentz factor, and hence the Doppler boosting, is strongest. This region is then identified as the radio core, with the radio jet visible on the downstream side. Substantial self-Compton -ray and X-ray emission can occur either in the UVOIR region or the radio core (see Maraschi et al. 1992). In addition, (inverse) Compton relection of optical and uv photons from the accretion disk can take place in the UVOIR region, producing X-rays and rays (Dermer, Schlickeiser, & Mastichiadis 1992).

Begelman & Sikora (1987) and Melia & Königl (1989) have discussed Compton reflection of the optical and uv photon from the accretion disk off a highly relativistic stream of electrons. The process emits rays from the deepest part of the inner jet and X-rays somewhat downstream of this, with the nonthermal optical to radio emission occurring farther out. An interesting aspect of Compton reflection is that photons approaching a stream of electrons from behind are scattered primarily into a hollow cone of opening angle ~ , so that the emission is quite low if one views directly down the axis of the jet (Melia & Königl 1989; Dermer et al. 1992). This conflicts with the model described in Section 3 for the intraday flux variations in the BL Lac object 0716+714. The model requires that we view the jet directly down the axis. This implies that either the jet is bent through an angle ~ or the hard rays observed from this object (Michelson et al. 1992) do not arise from Compton reflection.

In both of these basic models the emission regions at different wavebands are connected but lie at different distances from the central engine. One can therefore potentially use multifrequency observations to discriminate among the models. Variations occurring in shocked regions near the radio core should have the characteristics described in Section 3: simultaneous brightness and polarization fluctuations at higher frequencies and slightly time-delayed and less pronounced variations at lower frequencies (but still above the self-absorption turnover). For the two inner jet models, however, disturbances propagating down the jet are time-delayed at different wavebands, depending on the location of the primary emission region at each frequency. For example, in the decelerating jet model, uv flares should occur before -ray and X-ray flares, which should later lead into radio-infrared outbursts. It is also possible in the context of this model for the flare to be caused by an enhanced flow of electrons in the stream, in which case a -ray and X-ray flare would precede a radio-infrared outburst with no uv flare. In the specific accelerating jet model of Maraschi et al. (1992), the uv and -ray emission varies simultaneously, as does the IR and X-ray emission, but the latter fluctuations follow the former by a day or so, with the submillimeter outburst occurring later still. There are a number of other possible combinations, since precisely where the emission regions at each waveband lie depends on the details of the jet geometry and particle acceleration.