We investigate the relationship between the optical and radio emission of active galactic nuclei (AGN) by analysing optical and 15+22+43 GHz Very Long Baseline Array (VLBA) polarization observations simultaneous to within a day for 11 BL Lacertae (BL Lac) objects and the blazar 3C279. We have determined and corrected for the Faraday rotation measures in the very long baseline interferometry (VLBI) cores, enabling us to compare the intrinsic (zero-wavelength) VLBI-core polarization angles and the optical polarization angles chi(opt). A clear alignment between these two angles emerges in the transition toward higher radio frequencies, and a prominent peak at 0 degrees is visible in the distribution of |chi(opt) - chi(43 GHz)|. This correlation implies that the magnetic-field orientations in the regions giving rise to the optical and radio polarization are the same, and can be easily understood if the radio and optical polarization are roughly cospatial. It is difficult to rule out the possibility that they arise in different regions in a straight jet with a uniform magnetic-field structure, but this seems less likely, since the VLBI jets of AGN are often bent on parsec-scales. This may suggest that much or all of the strong optical polarization in these sources arises in the inner radio jets, possibly associated with the formation and emergence of compact new VLBI components

We present the results of parsec-scale circular-polarization measurements based on Very Long Baseline Array data for a number of radio-bright, core-dominated active galactic nuclei obtained simultaneously at 15, 22 and 43 GHz. The degrees of circular polarization m(c) for the Very Long Baseline Interferometry (VLBI) core region at 15 GHz are similar to the values reported earlier at this wavelength, with typical values of a few tenths of a per cent. We find that mc as often rises as falls with increasing frequency between 15 and 22 GHz, while the degree of circular polarization at 43 GHz is in all cases higher than at 22 and 15 GHz. This behaviour seems contrary to expectations, since the degree of circular polarization from both synchrotron radiation and the Faraday conversion of linear to circular polarization - the two main mechanisms considered thus far in the literature - should decrease towards higher frequencies if the source is homogeneous. The increase in mc at 43 GHz may be due to the presence of regions of both positive and negative circular polarization with different frequency dependences ( but decreasing with increasing frequency) on small scales within the core region; alternatively, it may be associated with the intrinsic inhomogeneity of a Blandford-Konigl like jet. In several objects, the detected circular polarization appears to be near, but not coincident with, the core, although further observations are needed to confirm this. We find several cases of changes in sign with frequency, most often between 22 and 43 GHz. We find tentative evidence for transverse structure in the circular polarization of 1055+018 and 1334-127, that is consistent with their being generated by either the synchrotron mechanism or the Faraday conversion in a helical magnetic field. Our results confirm the earlier finding that the sign of the circular polarization at a given observing frequency is generally consistent across epochs separated by several years or more, suggesting stability of the magnetic-field orientation in the innermost jets

We observed the blazar 0716+714 with the VLBA during its active state in [2003][2004]. In this paper we discuss multi-frequency analysis of the inner jet (first 1 mas) kinematics. The unprecedentedly dense time sampling allows us to trace jet components without misidentification and to calculate the component speeds with good accuracy. In the smooth superluminal jet we were able to identify and track three components over time moving outwards with relatively high apparent superluminal speeds (8.5-19.4 c), which contradicts the hypothesis of a stationary oscillating jet in this source. Component ejections occur at a relatively high rate (once in two months), and they are accompanied by mm-continuum outbursts. Superluminal jet components move along wiggling trajectories, which is an indication of actual helical motion. Fast proper motion and rapid decay of the components suggest that this source should be observed with the VLBI at a rate of at least once in one or two months in order to trace superluminal jet components without confusion.

Context. We study the jet kinematics of the blazar S5 0716+714 during its active state in 2003-2004 with multi-epoch VLBI observations.Aims. We present a kinematic analysis of the large-scale (0-12 mas) jet of 0716+714, based on the results of six epochs of VLBA monitoring at 5 GHz. Additionally, we compare kinematic results obtained with two imaging methods based on different deconvolution algorithms. Methods. The blazar 0716+714 has a diffuse large-scale jet, which is very faint compared with the bright compact core. Experiments with simulated data showed that the conventional data reduction procedure based on the CLEAN deconvolution algorithm does not perform well in restoring this type of structure. This might be the reason why previous kinematic studies of this source yielded ambiguous results. In order to obtain accurate kinematics of this source, we independently applied two imaging techniques to the raw data: the conventional method, based on difference mapping, which uses CLEAN deconvolution, and the generalized maximum entropy method (GMEM) realized in the VLBImager package developed at the Pulkovo Observatory in Russia. Results. The results of both methods give us a consistent kinematic scenario: the large-scale jet of 0716+714 is diffuse and stationary. Differences between the inner (0-1 mas) and outer (1-12 mas) regions of the jet in brightness and velocity of the components could be explained by the bending of the jet, which causes the angle between the jet direction and the line of sight to change from ∼5 • to ∼11 • . Conclusions. We tested the performance of the two imaging methods on real data and found that they yield similar kinematic results, but determination of the jet component positions by the conventional method was less precise. The method based on the GMEM algorithm is suitable for kinematic studies. It is especially effective for dim diffuse sources with the average brightness of several mJy with bright point-like features. For the source 0716+714, both methods worked at the limit of their capability.