Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.mrao.cam.ac.uk/yerac/mack/mack.ps Äàòà èçìåíåíèÿ: Wed Feb 22 23:05:03 1995 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 00:31:26 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: angular size

High­frequency Radio Continuum
Observations of Giant Radio Galaxies
By K.­H. Ma c ky, U. K l e i ny, L. S a r i p a l l iy, R. S t r omz
AND R. W i e l e b i n s k ix
e­mail: p643mac@astro.uni­bonn.de
y Radioastronomisches Institut der Universit¨at Bonn, Auf dem H¨ugel 71, 53121 Bonn,
GERMANY
z Sterrenwacht Dwingeloo, Postbus 2, 7900 AA Dwingeloo, THE NETHERLANDS
x Max­Planck­Institut f¨ur Radioastronomie, Auf dem H¨ugel 69, 53121 Bonn, GERMANY
Giant radio galaxies (GRGs) form the extreme end of the linear size distribution of radio galaxies,
with sizes in excess of 1 Mpc. Using the Effelsberg 100­m telescope we have performed a high­
frequency radio continuum survey of these objects. The extraordinary size and the high dynamic
range of these maps required the installation of a number of new data reduction procedures.
One of these -- a H¨ogbom­like CLEAN algorithm of single­dish data -- enables us to reach
dynamic ranges of 30 dB at present and an estimated dynamic range of order 40 dB after further
improvements in near future. Our maps provide the necessary data base complementing low­
frequency observations gained with the WSRT and the VLA. Together with the interferometer
data it is possible to determine for the first time important physical parameters (like spectral
indices, break frequencies, spectral ages, rotation measures, depolarization) over a very large
frequency range. Because of their immense sizes, these sources can also serve as unique probes
for the characteristics of the surrounding intergalactic medium (IGM).
1. Introduction
In this paper we present a project on high­frequency radio continuum observations
of giant radio galaxies. These are radio galaxies with linear sizes larger than 1 Mpc,
assuming H 0 = 75 km s \Gamma1 Mpc \Gamma1 and q 0 = 1.
Much work has already been done on giant radio galaxies by various groups. The
previous database for entire sources consisted mainly of low­frequency interferometric
data, which can be affected by ``missing­spacings''­problems. Therefore, we started a new
GRG­survey 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 cal­
culate source ages. From the linear polarization data we derive the magnetic field struc­
tures, rotation measures and depolarization. Asymmetries in spectra and polarization
characteristics of the sources will be investigated.
2. Observations
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 GHz­observations can be found in Klein et al. (1994). In addition, we
observed the eight sources 3C130, 4C39.04, 4C74.26, 1331­099, 1358+305, 0503­286,
1245+67, 4C34.47, and 8C0821+695, which are smaller in angular size. Some of them
have been discovered quite recently. The high­frequency observations have been done
1

2 K.­H. Mack et al.: High­frequency Observations on Giant Radio Galaxies
3C236 2.8 cm 'Dirty'
35 O 00'
35 O 10'
35 O 20'
10 h 01 m
10 h 02 m
10 h 03 m
10 h 04 m
10 h 05 m
3
3 3
3
3
3
3
3
3 3
3 3
3
3
3 3
3 3
3
3
3
3
3
3
3 3 3
3
3
3
3 3
3 3
3
3 3
3
3
3
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> >
>
>
>
>
15
15
15
15
15 15
1
5
20 20
20
20
20
>
>
>
>
>
>
>
>
>
>
>
>
40
40
40
>
> 80 160 >
Figure 1. Dirty 10.6 GHz map of 3C236
since 1990 using the 2.7, the 4.75, and the 10.55 GHz­receiver systems, all installed in
the secondary focus cabin of the Effelsberg 100­m telescope.
The low­frequency 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.
3. CLEANing of single dish data
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,
high­dynamic range maps are required. But since the first sidelobe of the 100­m 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.
In the 10.55 GHz map of 3C236 (Figure 1) one clearly sees the artifacts from dif­
fraction 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
so­called ``dirty­map''. 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 H¨ogbom­CLEAN algorithm (H¨ogbom (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 anti­
cipate observations with dynamic ranges up to 40 dB. Another advantage of CLEANing
single dish­data 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 low­frequency maps. For a more detailed description and further
examples of our CLEAN­procedure we refer to Klein & Mack (1994).

K.­H. Mack et al.: High­frequency Observations on Giant Radio Galaxies 3
3C236 10.55 GHz Total Intensity
35 O 00'
35 O 10'
35 O 20'
10 h 01 m
10 h 02 m
10 h 03 m
10 h 04 m
10 h 05 m
3
3 3
3
3
3
3
3
3 3
3 3
3
3 3
3
3
3
3 3
3
3 3
3 3
3
3
3
3 3
3
3
5
5
5
5
5
5
5
5
5
10
10
10
10
1 0
10
10
>
>
>
>
> >
>
>
>
>
>
>
>
>
> > >
>
3C236 10.55 GHz Linear Polarization
35 O 00'
35 O 10'
35 O 20'
10 h 01 m
10 h 02 m
10 h 03 m
10 h 04 m
10 h 05 m
1 .
5
1 .
5
1 . 5
1 .
5
1. 5
1 .
5
1.
5
1 .
5 1.
5
1
.5 1
.
5 1 .
5
1.
5
3
3
3
3
5
5
>
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.
4. An example: 3C236
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 south­eastern 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.
The north­western 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 ú 30% to 40%.

4 K.­H. Mack et al.: High­frequency Observations on Giant Radio Galaxies
3C236 Spectral index distribution 327 MHz ­ 608.5 MHz
Levs = 0.1, ­0.1, ... ­2.1, ­2.3
DECLINATION
(B1950)
RIGHT ASCENSION (B1950)
10 05 00 04 30 00 03 30 00 02 30 00 01 30 00
35 25
20
15
10
05
00
34 55
50
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
3C236 Spectral index distribution 608.5 MHz ­ 4750 MHz
Levs = 0.1, ­0.1, ... ­1.3, ­1.5
DECLINATION
(B1950)
RIGHT ASCENSION (B1950)
10 05 00 04 30 00 03 30 00 02 30 00 01 30 00
35 25
20
15
10
05
00
34 55
50
0.0
0.2
0.4
0.6
0.8
1.0
1.2
3C236 Spectral index distribution 4750 MHz ­ 10550 MHz
Levs = 0.1, ­0.1, ... ­1.9, ­2.1
DECLINATION
(B1950)
RIGHT ASCENSION (B1950)
10 05 00 04 30 00 03 30 00 02 30 00 01 30 00
35 25
20
15
10
05
00
34 55
50
0.0
0.5
1.0
1.5
Figure 3. Spectral index distributions of 3C236.

K.­H. Mack et al.: High­frequency Observations on Giant Radio Galaxies 5
3C236 Break frequency distribution
Levs = 3, 6, 10, 15, 20, 30, 60, 100, 150 GHz
DECLINATION
(B1950)
RIGHT ASCENSION (B1950)
10 05 00 04 30 00 03 30 00 02 30 00 01 30 00
35 25
20
15
10
05
00
34 55
50
20
40
60
80
100
120
140
Figure 4. Break frequency distribution of 3C236
5. Spectral index investigations on 3C236
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 Effelsberg­beam at 4.75 GHz, 153 00 . The low­frequency 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 high­frequency spectral index range
(Figure 3). Since the south­eastern 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 north­western lobe region. The low­frequency 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 north­western lobe to be generally steeper than that of the south­eastern one. This
asymmetry becomes more obvious looking at the high­frequency spectral index. The
mean values are \Gamma0:7 in the south­eastern and \Gamma1:2 in the north­western lobe.
These values indicate a typical synchrotron spectrum where the intensity is propor­
tional to š ff . 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 fre­
quency šB 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 šB at each pixel. Looking at the break­frequency map
(Figure 4) we get extreme values in the north­western 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
t R ¸
p
B
B 2 +B 2
IC
1
p
šB \Delta (1 + z)
;
where B = 1:2 ¯G corresponds to the source magnetic field derived from equipartition
(Strom & Willis (1980)), B IC = 3:25 \Delta (1 + z) 2 ¯G is the magnetic field equivalent to the

6 K.­H. Mack et al.: High­frequency Observations on Giant Radio Galaxies
microwave background and z = 0:0988 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 north­western lobe of ú 10 4 km s \Gamma1 . This confirms the
value derived by Strom et al. (1981).
6. Conclusions
The on­going project of high­frequency 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 single­dish 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.
We would like to thank Drs. C. O'Dea and A. Willis for the use of their data. The
Westerbork Radio Observatory is operated by the Netherlands Foundation for Radio
Astronomy. This work was supported by the Deutsche Forschungsgemeinschaft, grant
KL533/4­1.
REFERENCES
Alexander P. & Leahy J.P., 1987, MNRAS, 225, 1.
Carilli C., Perley R.A., Dreher J.M. & Leahy J.P., 1991, ApJ, 383, 554.
H¨ogbom J., 1974, A&AS, 15, 417.
Klein U. & Mack K.­H., 1994, in Proceedings Workshop on Multi­Feed Systems for Radio Tele­
scopes, ed. Emerson, D.T., (Astronomical Society of the Pacific Conference Series).
Klein U., Mack K.­H., Strom R., Wielebinski R. & Achatz U., 1994, A&A, 283, 729.
Strom R.G., Baker J.R. & Willis A.G., 1981, A&A, 100, 220.
Strom R. & Willis A., 1980, A&A, 85, 36.
van der Laan H. & Perola G.C., 1969, A&A, 3, 468.
Willis A. & O'Dea C., 1990, in IAU Symposium, 140, 455.