4 Discussion

In the following discussion, it is assumed that the redshifts, z, are cosmological in origin with $H\rm _0=100~h~km^{\rm -1}\ Mpc^{\rm -1}$and q₀=0.5. The extended cosmological information on the sources can be found in Table19.

   Table 19: Cosmological information for the observed sources. z, is the measured redshift, r, is the linear size in cm corresponding to 10 $\mu$as at the distance of the source, v, is the velocity in units of c for a proper motion of 100$~\mu$as/year for a component, and $\mu$, is the observed proper motion of this source from the 1990 and 1993 epochs

We have presented the maps from the 1990 April and 1993 April observations of active galactic nuclei at $\lambda$3mm with 50 $\mu$as resolution. These maps are among the highest resolution images of AGNs available to astronomers at this date and are vital to our understanding of the processes taking place in the central region of AGNs, especially the physical procesesses behind the rapid structural changes. In spite of the limited dynamic range and long time between the epochs we are able to draw the following conclusions:

1.

The core is very dominant at $\lambda$3mm, any extended jet is in many cases resolved out or to weak compared with the core to be seen with our limited dynamic range. We can divide the observed sources into three groups:

(a)

Jet dominated: Our maps of sources that are dominated by a jet at lower frequencies (3C84, 3C273, and, 3C279) or have an extended halo (2145+067) contain only a small fraction (less than 30%) of the single dish flux density. This is due to the effect that the interferometric array resolves out the large scale structures (i.e. jet or halo) that contain the major part of the flux density. We do not exclude the possibility that parts of the missing flux density also can be attributed to calibration and UV-coverage problems.

(b)

Core dominated sources: A few of the sources (0735+178, 1510-089, and BLLac), typically those that are unresolved or have a simple structure with core and a single component close to the core, have all the single dish flux density present in the map. Most of their emission at $\lambda$3mm is emitted from very compact ($\mu$as) components.

(c)

There are intermediate cases (1055+018, 3C345, 3C446, and OJ287) where some flux density has been lost ($\sim$50%) due to resolution effects, calibration inaccuracies, and the poor UV-coverage at short baselines.

2.

It is clear from the maps that rapid structural changes take place. The process of determining proper motions of components from one epoch to another is fraught with several complications.

• In the cases where we have attempted to identify components then the difficulty of identification will function as a selection effect against high proper motions as the components with high proper motions will have moved too much to be unambigously identified.
• In some cases we are able to determine only lower limits to the proper motion, based on conservative assumptions of component motion. All this combined will give low values to the proper motions detected with the infrequent $\lambda$3mm VLBI observations.

More frequent observations are needed to determine proper motions. We estimate that observations every month are required to obtain adequate time sampling to follow the ejection of components.

We see much lower proper motions at $\mu$as scales (Table19) than the proper motions seen at mas scales (Table3). It is important to state that this is not clear evidence for acceleration of components as they move out from the core, but rather a selection effect as discussed above.

3.

The curvature and twisting of the jets seen at larger scales can also be seen at $\lambda$3mm, but is even more pronounced (3C84, OJ287, 3C345, BLLac, CTA102, and 3C446). The PA of the components varies dramatically between $\mu$as and mas scales. The difference between $\mu$as and mas scale jet, $\Delta$PA, is in these cases $\sim$90$\hbox{$^\circ$}$. Some sources, on the other hand, exhibit jets that are straight from $\mu$as to mas scales, i.e. 3C111, 0735+178, 0748+126, 1055+018, 3C273, and 3C279. This modality in $\Delta$PA has been observed previously looking at the difference in PA between VLBI and VLA images (Pearson & Readhead 1988; Wehrle et al. 1992). Conway & Murphy (1993) explained this modality by having two different populations of AGNs. The misaligned population has parsec scale jets in the form of low pitch helices and the aligned population has straight jets with small changes in PA due to intrinsic bends. The modality seen in $\Delta$PA between mmVLBI and on VLBI then implies that the curvature in the helical jets increases as we get closer to the core. Note all the misaligned sources are either core dominated (point 1.b above) or intermediate cases (point 1.c above) strengthening the case for applying the Conway & Murphy model to our $\lambda$3mm VLBI observations, and the grouping of objects into two groups does not address which actual physical process causes the observed helicity.

4.

The majority of the sources exhibit an unresolved core even with 50$\mu$as resolution. The diameters of the core seen in our maps range from 10¹⁶-10¹⁷cm, still magnitudes larger than the expected Schwarzschild radius of the MBH, $R_{\rm s}=10^{13}-10^{14}$cm (Band & Malkan 1989), but close to the expected size of the accretion disc or torus, 40$\times$$R_{\rm s}$.

With $\lambda$3mm VLBI a intricate picture of the inner regions of AGN's emerges. Our unique observations show some sources with components moving along twisted paths, other sources where components move along a straight jet from $\mu$as to mas scales, weak superluminal motion, and tentative support for acceleration (although the last feature may be a selection effect as discussed above).

One model proposed to explain the helical structure seen in many sources is the helical jet model proposed by Hardee (1987). In this model the observed components are interference points of Kelvin-Helmholtz instabilities traveling down a conical, adiabatically expanding flow. Numerical simulations of slabs propagating down such a flow (Hardee et al. 1992) show that the jet can remain collimated over large distances, before the oscillations become large enough to disrupt it. The simulations also show that the oscillation wavelength could scale with the distance from the core.

Another model which also has been successful in modeling the helical structure of 3C345 (Steffen et al. 1995) is the jet model by Camenzind & Krockenberger (1992). This model uses bulk relativistic plasma moving along helical magnetic field lines, driven by magnetized accretion disc winds.

Our data are not of sufficient quality to discern which of these two main models is best suited to explain the observed helicity.

The $\Delta$PA observed by us strongly suggests that any model trying to explain the appearance of the jets of the misaligned jet population, needs to scale with distance from the core, thus models that are scale-free face severe problems.

We conclude that $\lambda$3mm VLBI is a very promising tool for investigating the central regions of AGN's. Our observations are limited by dynamic range and the long time between epochs but we have newer epochs in the process of being reduced that will improve both of these weak points.

Acknowledgements

We would like to thank the people at the individual telescopes for helping in making the observations, without their support these observations would not have been possible. We would also like to thank the people at the Haystack correlator for their help in correlating these experiments. Without the help of D. Graham and A. Greve VLBI observations at Pico Veleta would not have been possible. We further would like to thank K. Standke for his help with the observation and calibration.

Fredrik T. Rantakyrö acknowledges support for his research by the European Union under contract ERBCHGECT920011. Part of this research was carried out at the Onsala Space Observatory. Onsala Space Observatory at Chalmers University of Technology is the Swedish National Facility for Radio Astronomy. The Hat Creek array is operated by the Berkeley-Illinois-Maryland Association with funding from the National Science Foundation, Grant AST 93-20238 to the University of California. The NRAO is a facility of the National Science foundation, operated under cooperative agreement by Associated Universities, Incorporated. FCRAO is operated with support of the National Science Foundation under grant AST 91-15721 and with permission of the Metropolitan District Commission of the Commonwealth of Massachusetts. This is contribution number 819 of the Five College Astronomy Department. The work of T.P.K. was supported in part by the German Verbundforschung. The hydrogen maser and the VLBA terminal at Pico Veleta were financed by the Stiftung Volkswagenwerk. This research has made use of the NASA/IPAC Extra-galactic database (NED), which is operated by the Jet Propulsion Laboratory, Caltech, under contract with the National Aeronautics and Space Administration.