© CAMBRIDGE UNIVERSITY PRESS 1991

3.2. Direct evidence concerning extremely small jets

The VLBI data concerning pc-scale jets has been discussed and interpreted in Chapters 1 and 4 but a few particular points concerning what that data might say with respect to jet formation models should be made here. While current VLBI maps frequently do show clearly elongated features that are best classified as jets, rarely if ever is the resolution across the jet good enough to enable a good measurement of the jet opening angle to be made. The size of the opening angle and its variation with distance from the nucleus might be determined for some sources by RADIOASTRON, and could have a bearing on the choice of formation model. Models relying on winds from thin accretion discs are likely to produce rather wide beams; beams emerging from radiation supported thick discs are likely to have intermediate opening angles (3° 15°) while truly narrow beams can probably only be produced by models where magnetic forces dominate the focusing even on extremely small scales. In light of the very narrow jets often observed on multi-kpc scales, a great deal of intermediate scale focusing would be necessary for either of the first two general classes of models. While purely hydrodynamical processes in an appropriately flattened and dimpled interstellar medium make this possible to some degree (e.g., Wiita & Siah 1986), it is not likely to be good enough in most cases, and good focusing probably requires magnetic dominance at some distance (e.g., Mitteldorf 1987).

Explicit predictions of the cross-sectional shapes of beams emerging from different models of the powerhouse are usually not made, and when they are, they tend to be rather vague. In general, it is expected that beams from radiation supported discs or winds off thin discs would be somewhat denser and perhaps a bit faster towards the axis, probably implying a limb darkened structure. Several of the models involving magnetically dominated flows yield beams which are much denser towards the edges, so that one would naively predict an edge-brightened structure. However, there are many caveats and uncertainties involved here, and it would probably be worthwhile to direct some effort toward pinning down predictions in this regard. Although it does not appear that currently possible observations will be able to discriminate between these options, RADIOASTRON or a lunar-based VLBI system might have the necessary resolution for nearby sources.

Polarization measurements of VLBI sources are beginning to appear (Roberts et al. 1984; Wardle et al. 1986) and have been mentioned in Chapters 1 and 4. As they improve, they will give information on magnetic field orientation for nuclear jets similar to currently available data on extended jets, and will provide an extremely interesting tool. Strong polarizations and indications of predominantly toroidal magnetic fields would most easily fit models where the beams are formed and collimated by magnetic effects but it is not clear if they could distinguish between fields anchored or formed in large accretion discs or those threading a BH's horizon. If the magnetic field is apparently aligned with the flow at early stages it might argue in favour of models that advect and stretch the field and do not rely upon magnetic collimation, but some types of magnetically dominated flows could also exhibit this topology. Moreover, the effects of depolarization and Faraday rotation, as well as the real possibility of the physical separation between the strongest fields and the densest radio emitting plasma (e.g., Clarke et al. 1986) all make it very difficult to apply any such observations in a model-independent fashion. Still, the variability to be expected from sub-pc scale motions means that polarization measurements might allow viewing angles, and thus three-dimensional velocities, for nuclear flows to be determined (Phinney 1985).

Very firm evidence for highly relativistic bulk motion (_j 5) would cause severe difficulties for any method (wind or radiation supported disc funnel) that relied upon radiation pressure for acceleration, regardless of whether the beam were to be predominantly ordinary plasma or mostly leptonic. Conversely, strong evidence in favour of significant mass flows down nuclear jets, which might eventually emerge from multi-frequency VLBI polarization measurements, would be negative for the ion-supported thick disc or other models that depend upon the dominant energy flow being in the form of electromagnetic Poynting flux. Large mass fluxes at greater distances do not cause such a problem because of the probability of entrainment of ambient matter (cf. Sections 5.4 and 3.3). The possible future detection of extended collimated -rays would also probably argue in favour of dynamo type low density jets, but the implications of such observations have not yet really been explored.