13.2. Some Observations: Variability, Small-Scale Structure
There are a few observations of Cygnus A which may be relevant to our understanding of the central engine. First is the simple fact that the jet is collimated and directed on scales 0.5 h-1 pc, and remains so out to 40 h-1 kpc.
A second important observable for the AGN in Cygnus A is variability. The radio nucleus has been observed over many years at many frequencies. Interferometric images at 90 GHz with arc-second resolution set an upper limit to core variability of < 10% on time-scales of both 10 months and 10 years (Wright and Sault 1993, Wright and Birkinshaw 1984). Likewise, imaging at 15 GHz with arc-second resolution at a number of epochs has detected no variability at the 10% level over 15 years (Carilli 1989, Alexander et al. 1984). Lastly, VLBI imaging at 5 GHz with 1 mas resolution reveals no variation of the inverted-spectrum nucleus, nor of the integrated flux from the core-jet, to a limit of 10% over 6 years, although individual knots in the jet may have varied by up to 40% (Carilli et al. 1994a).
Arnaud et al. (1987) mention possible variability of the hard X-ray (non-thermal) component in the nucleus of Cygnus A at about the 20% level over 6 months. Ueno et al. (1994) have reconsidered this possibility, and find no evidence for variability in hard, medium, or soft X-rays (20-10keV, 10-4 keV, and 4-2 keV, respectively) at the level of 10% on time-scales of 1 day to 1 month. Comparing their results with the previous hard X-ray results, Ueno et al. set a limit of less than a factor two variability on a time-scale of a few years.
Lastly, observations of structures on pc-scales in Cygnus A provide model dependent probes of the pressure in the nucleus of Cygnus A. The nucleus has a steeply rising spectrum of index > 0.9 between 5 GHz and 43 GHz (Carilli et al. 1994a, Krichbaum et al. 1993). If this rising spectrum is the result of synchrotron self absorption (SSA), a self-consistent model can be constructed from the SSA equation (Kellermann and Pauliny-Toth 1981), and the minimum energy equation. We assume max = 25 GHz, and use the observed flux density of the unresolved component of 0.7 Jy at 43 GHz (Krichbaum et al. 1993). The result is a nuclear size of 0.15 mas, and a core magnetic field of 0.16 G. The pressure in fields and particles is 10-3 dyn cm-2. For comparison, the pressure in clouds in broad line regions of quasars is also thought to be of order 10-3 dyn cm-2, or higher (Ulrich 1983). Hence the nuclear jet could be collimated by external pressure if a medium similar to that proposed for confining broad-line clouds in quasars exists in the core of Cygnus A. It should be noted that we have not considered relativistic beaming in the above calculations.
Overall, any model for the central engine in Cygnus A requires efficient conversion of energy into collimated outflow at a fair fraction of the speed of light starting on scales < 0.5 h-1 pc, while avoiding rapid variability of the non-thermal radio and X-ray luminosity of the AGN itself on time-scales up to 15 years. Also, the beam formation mechanism must maintain the jet direction to an accuracy of better than 20° over time-scales the radio source lifetime (of order 107 years). As pointed out by Wiita (1991) and Begelman (1993), extremely narrow beams on small scales suggest the fundamental involvement of magnetic forces as focusing agents, while the stability of the spin axis of a massive black hole is the logical origin for this long-term `memory' of jet direction. The efficient model of Rees et al. (1982) for channeling the rotational energy of the black hole directly into collimated outflow on small scales via magnetic fields without generating tremendous waste heat in the process seems attractive in this regard.
We are still a long way from understanding the fundamental origin of the jets in Cygnus A. One glimmer of hope on the horizon is the advent of high frequency space VLBI, which will allow us to probe structure on scales 1017 cm in Cygnus A. While this is still about three orders of magnitude larger than the fundamental scale of the Schwarzschild radius, it approaches the scale of the hypothetical thick accretion disk and its associated funnels (Abramowitz and Piran 1980), and the general scale of the broad line region in quasars (Rees 1986).