[ CATS home ] [ Back to CATS list ] [ ftp ]


1995PASJ...47..725Inoue+

(OCR+proof, and reformatting by H.Andernach 1/96)

Search for High Rotation Measures in Extragalactic Radio Sources I. Multi-Channel Observations at 10 GHz


Makoto INOUE

Nobeyama Radio Observatory,* Minamimaki, Minamisaku. Nagano, 384-13

Hiroto TABARA and Tatsuji KATO

Faculty of Education, Utsunomiya University, Mine, Utsunomiya, Tochigi 321

and Ko Aizu

Katahira S-24-3, Asao-ku, Kawasaki, Kanagawa 215

Abstract

Multi-channel polarimetry has been performed to detect high rotation measure (RM) at 3 cm using the Nobeyama 45-m telescope. The high RM candidates of 96 radio sources were selected to be observed, and RMs of 35 sources were derived from the observations. Since the four channels are set contiguously from 2.84 cm to 3.31 cm, |RM| can be derived uniquely up to 15000 rad/m^2 by this polarimeter. We found that there exist sources with RM of several thousands rad/m^2. In fact, 5 sources have |RM| > 1000 rad/m^2. On the other hand, all sources observed are well within this system limits, and therefore we suggest the observed upper limit of |RM| is around 5000 rad/m^2 for extragalactic radio sources, even taken into account the redshift of sources.

Key words: Galaxies: active - Galaxies: magnetic fields - Polarization - Radio Sources

1. Introduction

The Faraday rotation is one of the most fundamental propagation processes in magnetized plasma in the radio wave region. The Faraday rotation measure (RM) provides us information on the magnetic field and the electron density of the plasma between the radio source and,observer, in addition to the direction of the magnetic field. However, in general, it is very difflcult to derive the RM from the existing data, because the n*pi ambiguity (n = 0,+-1, +-2, ...) in observed polarization angles cannot be fully removed, as the number of observed wavelengths is very limited and they are sparsely distributed.

Tabara and Inoue (1980) compiled polarization data and derived 637 RMs for radio sources. In deriving RMs, they set an upper limit of |RM| to avoid the n*pi ambiguity, otherwise a large value of RM could be often fitted to the oserved data by chance coincidence due to the limited number of observed wavelengths and their finite errors. in fact, the upper limit of 200 rad/m^2 was set for high galactic latitude sources. Because of certain amount of Galactic contribution (Inoue, Tabara 1981), the upper limit was increased with decreasing galactic latitude. More than 70% of the sources have |RM| < 50 rad/m^2 (Tabara, Inoue 1980). However, it has been known that some radio sources certainly show |RM| > 200 rad/m^2 (e.g., Tabara, Inoue 1980; Simard-Normandin et al. 1981; Dreher et al. 1987). Therefore, in order to distinguish the real high RM from false value of high RM, a system that is free from the nPI ambiguity is greatly needed. With such system, we could see high RM and study if there exists an upper limit to RM or not.

We have therefore built a multichannel polarimeter at 10 GHz attached to the Nobeyama 45-m telescope. Gauss and Goldstein (1973) first made a multichannel polarimeter at around 1300 MHz. High frequency observations are essential to study high RM because (1) high degree of polarization is expected because of low depolarization, (2) the amount of Faraday rotation is small, which could reduce the n*pi ambiguity even for high RM, (3) by adding a new data point to a plot of polarization angle vs. wavelength square, the n*pi ambiguity is very much reduced, and (4) more active, dense regions such as hot spots and nuclear regions can be seen, where high RM is expected. In fact, this system eventually enabled us to find high RMs in the Galactic center region (Inoue et al. 1984; Tsuboi et al. 1986), and in four extragalactic radio sources of |RM|>1000 rad/m^2 (Kato et al. 1987). Since it is HOW obvious that at least several sources show intrinsically high RM, we have made a survey for high RM sources using the polarimeter. In this paper we describe the results of this survey. In section 2, source selection for the survey is described, and in section 3 the observation system including the multichannel polarimeter, the calibration method, and data reduction is given in detail. The results of the observations on 103 sources are given in section 4. The results are incorporated in the revised catalogue of Tabara and Inoue (1980). Using this new catalogue, we have derived much reliable RMs applying a new method on the Or ambiguity. The results will be given in Part 11 (Aizu et al. 1995 in preparation). Time variation of RM will be also published elsewhere.

2. Source selection

Based on the revised catalogue of Tabara and Inoue (1980), we selected candidates for high RM sources with the following three criteria:

(a) Sources that are not well fitted to small RMs on polarization angle vs. wavelength square plot by the existing data. These sources are referred to as CPA (Candidates based on the Polarization Angle).

(b) Sources with a strong depolarization at longer wavelengths. Although the strong depolarization is not always associated with high RM (e.g., Dreher et al. 1987), the sources deserve to be examined. These sources are re ferred to as CDP (Candidates based on the DePolarization).

(c) Sources with compact steep spectrum (CSS). Two of four sources listed by Kato et al. (1987) were CSS sources, and they suggested that the dense medium around the nucleus, which stops the jets and hence prevents the formation of extended radio structure, is responsible for the high RM. CSS lists of Perley (1982), Peacock and Wall (1982), and Pearson et al. (1985) were surveyed.

These three criteria for the high RM candidates are not exclusive with each other. We further imposed the following three conditions to these candidates:
(i) Due to the sensitivity and different polarization natures of each criteria, we set limits in different way; the polarized flux density at 5 GHz is above 30 mJy for CPA, and for CDP the expected polarized flux density at 10 GHz is above 20 mJy. For CSS, the total flux density S at 5 GHz is S > 750 mJy and the spectral index alpha < -0.2 (S ~ freq^alpha).
(ii) The galactic latitude is above 20d so that the galactic RM due to the interstellar plasma should be negligibly small (moue, Tabara 1981), and also alpha < -0.2.
(iii) The polarized flux density changes no more than 30% from the average, or the polarization angle is no longer than 20d, irrespective of the time of observation. The variation in polarization was taken from Berge & Seielstad (1972), Altschuler & Wardle (1976), and Aller et al. (1985).
The total number of the candidates thus selected were 96: 28 CPAs, 19 CDPs, and 64 CSSs.

3. Observations

The system of polarization observations and the observing procedures are described. Then the polarization performances of the system derived from some calibration sources are shown in detail.

3.1. The System

Linear polarization observations were made from 1984 to 1988 using the multichannel polarimeter attached to the Nobeyama 45-m telescope at 10 GHz. The polarizer has a rotatable half-wave plate in a circular waveguide followed by an orthogonal polarization divider. The two orthogonal polarizations were switched by a diode switch matrix at 200 Hz to derive differential polarization. The matrix also worked as a Dicke switch for the total intensity measurements. The receiver was a cooled paranietric amplifier with a 2-GHz bandwidth from 8.8 GHz to 10.8 GHz, the system temperature being around 150 K.

The total bandwidth of 2 GHz was divided into four contiguous channels with 500-MHz bandwidth each at IF stage of 5 7 GHz. Each channel was sampled simultaneously, and converted by a 12-bit A/D converter in the receiver room. Figure 1 shows the block diagram of this system.

The HPBW at 10 GHz was 2.7', and the position of the telescope was switched by 5' in the azimuthal direction every 15 s to remove the sky fluctuation (i.e., ON-OFF switching). The pointing error was less thank 30", which did not cause effective error in polarization measurements. The direction of the received polarization was rotated by 22.5d from 0d to 180d for every 2 min. or four ON-OFF cycles, so that one sequence took 18 min. Just before or aver this sequence, the total intensity was measured by the Dicke switching mode. The differential polarizations from 0d to 180d were fitted by a sinusoidal curve, which gives the polarization angle and the polarized intensity. The one-sigma detection limit for the polarization is 10 mJy.

3.2. Calibration

A large part of the instrumental polarization came from the main surface of the telescope, and a minor part from the phase shifter. The calibration was made mainly by observing 3C 84, assumed to have no polarization. The instrumental polarization thus derived was essen tially proportional to the total intensity. The amount was less than 1% of the total intensity, which decreased linearly with the elevation angle of the telescope, its position angle being constant at the horizontal direction with respect to the main surface. Figure 2 shows the amplitude of the instrumental polarization against the elevation angle. The small amplitude is reasonable because this telescope operates well at short mm wavelengths, and the surface accuracy was around 100 pm. The calibration error was less than 0.1%, and hence this instrurnental polarization gave essentially no effect particularly for strongly polarized sources. For weakly polarized sources, the amount of correction was comparable or less than the error of the polarized intensity.

The instrumental polarization due to the phase shifter had a half-period component. This had a fixed phase with respect to the coordinate of the phase shifter against one period rotations of the polarizations angle (180d, or 90d rotation of the half-wavelength phase shifter) for a linearly polarized input. The residual after calibration of this component was less than 0.1% of the input intensity. The offset of the polarization angle was calibrated using 3C 286, the assumed RM and intrinsic polarization angle being 0 rad/m^2 and 33.0d +/- 0.1d, respectively. The amount of the correction was 3.7d. These calibrations were made at each observing run for each channel by observing 3C 84 with a wide range of elevation angles. Since the phase shifter was sometimes removed and re set, the correction parameters change from time to time. The flux densities of the calibrators are based on Baars et al. (1977), and they are summarized in table 1. The receiving band was so wide that the observations sometimes suffered from interference. In particular, the second lowest channel was affected seriously. In that case, we only give results of the remaining channels in the next section.

4. Results

Results of polarization observations are shown in table 2. Columns 1 and 2 give IAU name and other familiar name of the source, respectively. Columns 3 and 4 give the source identification and sample criterion, column 5 gives center wavelength of available channels (not affected by the interference; see the previous section), and column 6 the total intensity and its error. Columns 7 and 8 give the corresponding percentage polarization and the the polarization angle with errors. The polarization angle is given by the angle that was used to derive RM, so that it is sometimes shifted by 180d in order to give the difference < 90d from the next channel. In table 3, RMs of 35 sources out of 96 are given. The RMs listed are for sources whose errors of polarization degree were smaller than half its polarizations degree, arid the resultant error in RM is less than 700 rad/m^2.

Columns 1-4 repeat those in table 2, and column 5 gives RM arid its error. The percentage of CSS decreases from 66% in the original sample (table 2) to 43% in table 3, as polarization of CSS is, in general, weak. Figure 3 shows the histogram of RM distribution. More than 30% of the sources have |RM| > 500 rad/m^2. Among them, 5 sources (14%) show |RM| > 1000 rad/m^2. Although the error of RM is large, this distribution of large ratio of high RM is remarkably different from those of usual radio sources (e.g., Tabara, Inoue 1980). The number of sources, however, decreases rapidly and becomes rare above |RM| > 1000 rad/m^2. No clear trend of concentration for any source criteria can bee seen between RM distributions above and below 500 rad/m^2. However, it should be noted that more than 40% of CSS sources show |RM| > 500 rad/m^2.

5. Discussion

The upper limit of |RM| detectable by this system is determined by the receiving bandwidth of each channel. As Faraday rotation of more than 180d within the bandwidth causes the bandwidth depolarization, we cannot observe such high RM. Hence, if polarized flux is actually [...]

REFERENCES:

Aller H.D., Aller M.F., Latimer G.E., Hodge P.E. 1985, ApJS 59, 513
Altschuler D.R., Wardle J.F.C. 1976, MmRAS 82, 1
Baars J.W.M., Genzel R., Pauliny-Toth 1.1 K., Witzel A. 1977, A&A 61, 99
Berge G.L., Seielstad G.A. 1972, AJ 77, 810
Dreher J.W., Carilli C.L., Perley R.A. 1987, ApJ 316, 611
Gauss F.S., Goldstein S.J. Jr 1973, AJ 179, 439
Hewitt A., Burbidge G. 1987, ApJS 63, 1
Inoue M., Tabara H. 1981, PASJ 33, 603
Inoue M., Takahashi T., Tabara H., Kato T., Tsuboi M. 1984, PASJ 36, 633
Kato T., Tabara H., Inoue M., Aizu K. 1987, Nature 329, 223
Pearson T.J., Perley R.A., Readhead A.C.S. 1985, AJ 90, 738
Peacock J.A., Wall J.V. 1982, MNRAS 198, 843
Perley R.A. 1982, AJ 87, 859
Perley R.A., Taylor G.B. 1991, AJ 101, 1623
Simard-Normandin M., Kronberg P.P., Button S. 1981, ApJS 45, 97
Tabara H., Inoue M. 1980, A&AS 39, 379
Taylor G.B, Inoue M., Tabara H. 1992a, A&A 264, 415
Taylor G.B., Inoue M., Tabara H. 1992b, A&A 264, 421
Taylor G.B., Perley R.A., Inoue M., Kato T., Tabara H., Aizu K. 1990, ApJ 360, 41
Tsuboi M., Inoue M., Handa T., Tabara H., Kato T., Sofue Y., Kaifu N. 1986, AJ 92, 818
			 TABLE

Codes for selection criteria for sources:
SS = compact steep spectrum sources
DP = candidates based on depolarization
PA = candidates based on polarization angle

Table 2. Results of the four-channel polarimetry
                               |----- 2.84 cm / 10.56 GHz ---|----- 2.98 cm / 10.07 GHz ----|----- 3.14 cm / 9.55 GHz -----|----- 3.31 cm / 9.06 GHz -----|  RM  dRM
IAUname   Others   ID Criteria  S10.56/mJy p% +-%  PA/d  +-d | S10.07/mJy p% +-%  PA/d  +-d | S9.55/mJy  p% +-%  PA/d  +-d |  S9.06/mJy p% +-%  PA/d  +-d |  rad/m^2
----+----1----+----2----+----3----+----4----+----5----+----6----+----7----+----8----+----9----+----0----+----1----+----2----+----3----+----4----+----5----+----6----
0019-000  4C+00.02  G SS         535  26  1.7  1.0 114.5 17.3   483  25  0.8  1.6 168.5 54.6   542  28  2.0  0.7  56.6  9.3   540  27  1.7  1.0 106.5 16.2
0023-263              SS        1712  75  0.5  0.4 162.1 26.3  1798  79  0.4  0.3 121.1 20.9  1928  84  0.6  0.2  68.9 10.7  2082  89  0.3  0.3  17.8 33.1