Astron. Astrophys. 329, 809-820 (1998)

3. Results

The PPA- plots are shown in Fig. 2. The RMs and the statistics of the fit are listed in Table 2.

As mentioned earlier, RM studies, particularly involving low resolution observations in which the magnetic field structure is unresolved, have to consider the crucial question of whether or not the observed change of PPA with frequency is due to Faraday Rotation (or merely due to unresolved multiple components). In view of this, Table 2 also lists some parameters which quantify the quality of the straight line fit to the points in the PPA- plane. The parameters listed include the mean of the absolute value of the deviation from the fit, the largest deviation and the reduced along with the number of points in the fit.

The errors on the PPA values used here have been calculated by adding in quadrature the errors due to polarisation calibration (5 for 1.425 GHz bands and 3 for 4.8 and 8.4 GHz, the difference being due to the larger effect of the ionosphere at lower frequencies) and due to those on Stokes Q and U flux density.

The reduced is 1 in 20 of the 23 lobes (and much less in many) indicating that the straight line is indeed a good fit in most cases. The sources for which the values are large are 0156-252 foll. lobe (low resolution image), 0406-244 foll. lobe and 0349-211 prec. lobe. However, it must be noted that the is critically dependent on the formal error assigned to each point and in fact one very discrepant point could affect the statistic to a large extent. In fact, an examination of the residuals shows that the mean and the maximum residuals are only 3.8 and 6.6 in 0156-252 and 5.0 and 5.4 in 0406-244. This indicates that either the error has been underestimated by a few (1-3) degrees in the case of a few points or that those points have been contaminated at the same level (i.e. 1-3 ) by PPA changes due to multiple components. The low value of the and/or the small mean deviations of a few degrees seen in almost all sources justifies our assigning Faraday rotation as the cause for the change of PPA with frequency. However, the residuals and the value seen in the case of 0349-211 prec. lobe are too large to be assigned to random error in the PPA values and multiple components could be the explanation.

Some notes on the RM calculations of individual sources are given below :
0030-219: An unresolved source with no polarised flux detected.
0140-257: foll. lobe - No polarised flux detected.
0156-252: foll. lobe - RM values of both 7.5 and 82 rad m^-2 fit the data from low resolution images. At higher resolution, the lobe is well resolved and consists of the hotspot, parts of the jet and extended emission surrounding the two. The higher resolution image (see Fig. 3) shows a large variation in the RM_obs ranging from 18 to 160 rad m^-2. This source clearly shows the effect of low resolution on RM. The areas with low RM contribute most of the polarised flux and so the integrated RM from low resolution images is much smaller.
prec. lobe - No polarised flux detected.
0406-244: foll. lobe - It is difficult to resolve the 180 ambiguity and two values of the RM make for equally good fits.
prec. lobe - An increase of 180 in the PPA at 1.425 GHz resulted in an excellent fit. There is a good agreement between the PPA values in the high and low resolution images. There are 4 data points (several of them superposed in the plots) in both the 4.7 and 8.4 GHz bands. The four consistent PPA values in both 4.7 and 8.4 GHz bands justify the higher RM value and the 180 added to the PPA in the 1.4 GHz band.
0943-242: Polarised flux seen at only one frequency in each of the lobes.
1138-262: foll. lobe - Has the highest RM for the whole sample. The linear fit to the data indicates that the high RM value is genuine. The values obtained from low and high resolution images are consistent with each other.
prec. lobe - Consists of multiple components which have large polarised flux but appear to be consistent with RM of zero, though the errors are large.
1324-262: foll. lobe - This is another high RM source with a good linear fit to the data.
prec. lobe - No polarised flux detected.
2025-218: prec. lobe - consists of multiple components, all of which are consistent with zero RM.
2036-254: prec. lobe - Another high RM source with an excellent linear fit. One of the frequencies in the 8.21 GHz is very noisy and polarisation is barely detected. However, the PPA appears to be consistent.
2104-242: prec. lobe - Two values of the RM appear to fit the data though the lower value provides a better fit.

Fig. 3. Radio maps of 0156-252. The contour images at 8.44 GHz are superposed with (i) magnetic field vectors in the top image, (ii) grey scale of the fractional polarisation in the eastern lobe in the middle and the grey scale of the observed RM in the eastern lobe in the bottom image. The contour levels are 0.2 -2, -1, 1, 2, 4, 8, 16, mJy/bm. The beam size is 0:0065 0:005 at 40 PA. The first contour level is 3 noise in the image. The polarisation images have been blanked below 5 (0.3 mJy/bm at 8.44 GHz and 0.15 mJy/bm at 4.71 GHz).

© European Southern Observatory (ESO) 1998

Online publication: December 16, 1997
helpdesk.link@springer.de