Astron. Astrophys. 344, 402-408 (1999)

4. Discussion

4.1. Faraday rotation measures

There are essentially two possible origins for the high rotation measure of 3C 119: the presence of a dense Galactic cloud along the line of sight to the source, or the presence of thermal plasma near the source, possibly associated with 3C 119 itself. Although typical Galactic rotation measures are tens of rad/m², 3C 119 is very close to the Galactic plane (its Galactic latitude is ), making a high Galactic rotation measure more plausible.

If the rotation measure of 3C 119 were Galactic in origin, however, we would expect to see a nearly uniform rotation measure distribution over the source, with any gradients not having any apparent relation to the source structure. Flatters (1998) obtained VLBA observations similar to ours at 5 GHz, which suggested that the rotation measure of 3C 119 was concentrated near component C; however, the resolution and sensitivity provided by those observations were insufficient to draw definitive conclusions about the rotation-measure distribution across C. The plot in Fig. 4 and 8-GHz rotation-measure distribution in Fig. 5 also demonstrate that the region of high rotation measure is confined to the area near component C; although the uncertainty in the inferred rotation measure for component B is relatively large, it is clear that B's rotation measure is substantially lower than C's. In addition, the resolution provided by our observations has enabled detection of the rotation-measure gradient visible in Fig. 5, which shows a clear relation to the source morphology. This gradient is roughly along the inferred direction of the flow from the place where the western extension of component B joins the northern part of component C to the peak of C.

The rotation-measure gradient across component C is quite large,  rad/m² per mas, and increases fairly smoothly from the northeast to southwest, reaching values  rad/m² at the southern edge of C. This suggests the presence of a clump of thermal plasma in the intergalactic medium at the leading edge of C. In this case, the rotation-measure gradient in component C in 3C 119 is similar to rotation-measure features associated with the hot spots in Cyg A (Dreher et al. 1987), but on a much smaller scale. In addition, the inference of the presence of a dense cloud at the leading edge of C supports the suggestion by Nan et al. (1991) that the VLBI jet travels from the core to component C, then is deflected at component C and continues south toward component D. In this picture, the brightness of C is associated with compression due to the collision with the dense intergalactic medium.

4.2. Magnetic field

Our observations have, thus, resolved the Faraday screen on milliarcsecond scales. This enables us to "derotate" the observed distribution to determine the intrinsic magnetic-field directions at all points in the mas-scale structure where significant polarization was detected. The B vectors reconstructed in this way for component C, shown in Fig. 6, are well aligned with the direction of the rotation-measure gradient in Fig. 5. This suggests to us that both the rotation-measure gradient and this B field direction reflect the direction of the underlying flow of material in the northern part of component C; i.e., the B field there is longitudinal.

Just south of the peak of component C, the B field has swung towards the south, toward the direction of component D. It is a quite natural interpretation that this, likewise, reflects the local jet flow direction. As shown in Fig. 3, the degree of polarization in component C is maximum to the west of the peak of C. If, indeed, component C represents the place where the VLBI jet is deflected by a dense clump in the external medium, we would expect the magnetic field to be enhanced by compression. This may provide an explanation for the fact that the maximum degree of polarization in C is located to the west of the I peak: it is displaced toward the region of maximum compression.

4.3. Depolarization

The bandwidth depolarization within each of our 8-GHz frequency channels is estimated to be much less than one percent, so that it is negligible. The relatively high degree of polarization for component C (; see Table 3) and the ordered appearance of the distributions for each frequency also suggest that beam depolarization is not important. Together with the fact that our 8-GHz VLBI observations detected a high fraction of the integrated polarization measured by the VLA during our VLBI experiment, this is consistent with the relatively high degree of polarization () in the 3 cm integrated measurements using the Nobeyama 45-m telescope by Inoue et al. (1995). However, the large rotation-measure gradient in component C will give rise to large differential rotations of the vectors associated with the polarized emission in different regions in C at longer wavelengths. This can probably explain the low degree of polarization in integrated measurements at 6 cm ().

The degrees of polarization for component C derived from our P maps are 14.9, 16.1 and 17.0%, respectively, at our three different frequency channels, in order of increasing frequency. The two redundant measurements for 8.52 GHz (for IF34 and IF56) differ by only 0.2%, suggesting that this apparent decrease in degree of polarization with decrease in frequency may be real. In contrast, the corresponding three m values for component B are 16.6, 16.2, and 16.4%, so that they do not show any dependence on frequency. If the 2% drop in m across our entire bandwidth for component C is real, this could represent depolarization due to thermal plasma in the emitting region of the synchrotron radiation. This could reflect mixing of the jet material with the thermal plasma of the cloud with which C is colliding.

© European Southern Observatory (ESO) 1999

Online publication: March 18, 1999
helpdesk.link@springer.de