Please Note: the e-mail address(es) and any external links in this paper were correct when it was written in 1995, but may no longer be valid.
Radioastronomisches Institut der
Universität Bonn, Auf dem Hügel 71,
53121 Bonn, GERMANY
Sterrenwacht Dwingeloo, Postbus 2, 7900 AA Dwingeloo, THE NETHERLANDS
Max-Planck-Institut für Radioastronomie, Auf dem Hügel 69, 53121 Bonn, GERMANY
In this paper we present a project on high-frequency radio continuum observations of giant radio galaxies. These are radio galaxies with linear sizes larger than 1 Mpc, assuming and .
Much work has already been done on giant radio galaxies by various groups. The previous database for entire sources consisted mainly of low-frequency interferometric data, which can be affected by ``missing-spacings''-problems. Therefore, we started a new GRG-survey with a single dish in order to avoid these problems. Full linear polarization data were simultaneously obtained at high frequencies (2 - 10 GHz). Especially at 10.6 GHz Faraday effects can be neglected and we are able to directly map magnetic fields.
The aims of this project are to determine high frequency spectra, which allow to calculate source ages. From the linear polarization data we derive the magnetic field structures, rotation measures and depolarization. Asymmetries in spectra and polarization characteristics of the sources will be investigated.
Our sample so far consists of the six ``classical'' GRGs, i.e. 3C236, 3C326, DA240, NGC6251, NGC315, 4C73.08 which are the largest in angular size. A detailed discussion of the 10.6 GHz-observations can be found in Klein et al. (1994). In addition, we observed the eight sources 3C130, 4C39.04, 4C74.26, 1331-099, 1358+305, 0503-286, 1245+67, 4C34.47, and 8C0821+695, which are smaller in angular size. Some of them have been discovered quite recently. The high-frequency observations have been done since 1990 using the 2.7, the 4.75, and the 10.55 GHz-receiver systems, all installed in the secondary focus cabin of the Effelsberg 100-m telescope.
The low-frequency interferometer data were obtained using the WSRT at 327 MHz and 610 MHz, and the VLA at 1.4 GHz, with many data kindly contributed by C. O`Dea, R. Fanti, P. Parma, A. Singal, and A. Willis.
The aim of our project was to detect the diffuse emission of these objects. At the same time these sources also have very bright components (like cores or hot spots). Therefore, high-dynamic range maps are required. But since the first sidelobe of the 100-m telescope lies around 17 dB below the peak of the main beam, one finds the antenna pattern to generally cover the diffuse emission around bright components of these sources.
Figure 1: Dirty 10.6 GHz map of 3C236
In the 10.55 GHz map of 3C236 (Figure 1) one clearly sees the artifacts from diffraction by the support legs of the primary focus cabin. To get rid of this structure we have developed a program which iteratively subtracts the antenna pattern from the so-called ``dirty-map''. In contrast to interferometers we cannot calculate the antenna pattern but have to measure it by mapping a bright and isolated point source under the best conditions possible. We used 3C84 as point source. To clean our maps we work with a procedure similar to the Högbom-CLEAN algorithm (Högbom (1974)) which was already used for the analysis of interferometric data for a long time. The antenna pattern is iteratively subtracted from all map pixels above a certain level and substituted by a Gaussian. The result is stored in the cleaned map (Figure 2). Using this method we can increase the dynamic range up to 30 dB at the moment, and with a refined method which takes changes of the antenna patterns at different elevations into account, we anticipate observations with dynamic ranges up to 40 dB. Another advantage of CLEANing single dish-data is given by the substitution of the telescope beam by a Gaussian beam equivalent to interferometer maps. This results in a better adjustment of the beams for comparisons with the low-frequency maps. For a more detailed description and further examples of our CLEAN-procedure we refer to Klein & Mack (1994).
Figure 2 shows the final maps of 3C236 in total intensity and linear polarization. The radio core has a flux density of 900 mJy at this frequency. Both radio lobes extend out to about 1 Mpc. The south-eastern one is dominated by two bright sources. One is a background source, the other is the hot spot which forms the termination of the lobe.
Figure 2: Final maps of 3C236. The upper map shows total intensity as contours superimposed by vectors oriented parallel to the magnetic field with their lengths proportional to the polarized intensity. The lower map displays the polarized intensity as contours with vectors proportional to the percentage polarization.
Figure 3: Spectral index distributions of 3C236.
The north-western lobe is resolved and does not show any compact hot spot, but rather a broad relaxed plateau. As already mentioned there are a lot of background sources in the field. The magnetic field lines are oriented parallel to the global jet direction and appear to confine the lobes at their edges. The magnetic field structure resembles that deduced by Strom & Willis (1980) from rotation measure analyses. The polarized emission of the core region is strongly affected by instrumental polarization. The outer lobes are strongly polarized, with maximum degrees of to .
In order to derive e. g. spectral indices we have to compare these data with maps observed at different frequencies. All of them have been smoothed to the HPBW of the Effelsberg-beam at 4.75 GHz, 153. The low-frequency data were obtained with the WSRT at 327 MHz and 609 MHz by Willis & O'Dea (1990). We have calculated the spectral indices in the low-, intermediate-, and high-frequency spectral index range (Figure 3). Since the south-eastern lobe is more complex due to the background source and the bright hot spot, especially at this resolution, we restrict ourselves in the following to the north-western lobe region. The low-frequency spectral index (between 327 MHz and 610 MHz) does not show much change in the spectral index along the source's main axis. A more gradual steepening is already striking in the intermediate spectral index map (between 610 MHz and 4.75 GHz). In addition, there is a tendency for the spectrum in the north-western lobe to be generally steeper than that of the south-eastern one. This asymmetry becomes more obvious looking at the high-frequency spectral index. The mean values are in the south-eastern and in the north-western lobe.
Figure 4: Break frequency distribution of 3C236
These values indicate a typical synchrotron spectrum where the intensity is proportional to . If we now assume that only one single injection at the hot spot takes place without further replenishment of energy, we will expect synchrotron losses, which should first affect the highest energies. This results in a steepening of the spectrum towards higher frequencies. Thus, with increasing age of the radiating particles, the break frequency shifts towards lower frequencies. Similar to the work by Carilli et al. (1991) and Alexander & Leahy (1987) we tried to fit a theoretical spectrum to the data points to derive the break frequency at each pixel. Looking at the break-frequency map (Figure 4) we get extreme values in the north-western lobe between 200 GHz close to the injection point and 3 GHz towards the core position. The break frequency shifts towards lower frequencies with increasing age. The dependence (van der Laan & Perola (1969)) is given by
where corresponds to the source magnetic field derived from equipartition (Strom & Willis (1980)), is the magnetic field equivalent to the microwave background and is the redshift of the source. Relating the ages at the edges of the north western lobe to its linear size we finally get an average advance speed of the leading edge of the north-western lobe of . This confirms the value derived by Strom et al. (1981).
The on-going project of high-frequency observations of giant radio galaxies provides us with data of this intriguing class of objects in a wavelength range which could not be accessed hitherto for sources of this size. The required dynamic ranges of the high-frequency maps are provided by a new procedure to CLEAN single-dish data, yielding dynamic ranges of more than 30 dB.
Using 3C236 as an example, we indicate one way to study these sources by comparing the data at several frequencies to derive spectral indices, break frequencies, and ages. Investigations of the magnetic field structure will be carried out via the polarization data to study rotation measures and depolarization.
Besides the investigation of individual sources, studies of the whole sample will shed more light on the nature of the intergalactic medium which is probed by these sources because of their immense sizes.
Studies of this subject are still continuing. Results will be reported in forthcoming papers.