Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 341, 44-57 (1999)

Previous Section Next Section Title Page Table of Contents

3. New observations

3.1. WENSS, NVSS and FIRST radio survey data

B1144+352 has been observed in three recent major sky surveys: WENSS (WSRT 325 MHz, Rengelink et al. 1997), NVSS (VLA D-array 1.4 GHz, Condon et al. 1998) and FIRST (VLA B-array 1.4 GHz, Becker et al. 1995). The WENSS radio map, with a (FWHM) beamsize of [FORMULA] at the declination of B1144+352, shows the bright unresolved GPS radio core and two diffuse extended radio structures on either side of it (see Fig. 2a). The eastern structure is connected to the GPS source by a faint `bridge' and has a morphology which resembles the lobes of edge-brightened (or FRII-type, Fanaroff & Riley 1974) double-lobed radio sources. The western structure is elongated in a direction perpendicular to the radio axis as defined by the core and the eastern lobe. It consists of three bright regions and an extended, diffuse southern tail.

[FIGURE] Fig. 2a and b. Radio maps of the source B1144+352 from the WENSS and the NVSS surveys. a  Contour plot from the 325-MHz WENSS survey. Contourlevels are at (-8, 8, 11.3, 16, 22.6, 32, 45.3, 64, 128, 256, 512) mJy beam-1. The grey scale represents the spectral index [FORMULA] between 1400 and 325 MHz, calculated using only pixels above a [FORMULA]-level in the WENSS and convolved NVSS maps. It ranges from -0.4 (white) to -1.1 (black). b  Contour plot from the 1.4-GHz NVSS survey. Contours are at (-1.2,1.2,1.7,2.4,3.4,4.8,6.8,9.6,19.2,38.4,76.8,307.2) mJy beam-1. Also plotted are the observed E-field polarization vectors, with a length proportional to the polarized intensity ([FORMULA] mJy beam-1).

The 1.4-GHz NVSS survey, which uses a beamsize of [FORMULA] (FWHM), has also detected the diffuse extended structures and the southern tail (see Fig. 2b). The total flux density of the extended radio structure in the NVSS maps is [FORMULA] mJy, which is much more than the 80 mJy discrepancy that S95 found between their core flux density and the flux density found in the 20-cm Green Bank survey. We believe that there are two reasons for this difference. First, the Green Bank survey was conducted in April 1983, when the flux density of the core was much lower than during the observation of S95 (see Sect. 4.1 and Table 2). Second, the flux density given in White & Becker (1992) is the peak flux density, which, for an extended source like B1144+352, is an underestimate of the total flux density.

We have used the NVSS and the WENSS maps to produce a spectral index map, which is shown in grey scale in Fig. 2a. To avoid artefacts resulting from combining data with different beamsizes, we have first convolved the NVSS map to the resolution of WENSS. The `bridge' that connects the central source to the eastern structure has a somewhat steeper radio spectrum than the rest of that structure. The surface brightness of the tail of the wetsern structure is rather low, so that the spectral index is poorly determined. We measure a value of [FORMULA], which is not much different from the other extended source components.

The NVSS has further detected linear polarization in both extended radio structures. Halfway through the eastern lobe, the position angles of the vectors of the E-field of the linear polarized emission change from perpendicular to more parallel with respect to the radio axis. Polarization has also been detected towards the core. The fractional polarization is [FORMULA]%. It will be shown in Sect. 3.2 that this polarized emission mainly originates in the eastern jet and not in the GPS radio core.

The radio source has also been mapped in the FIRST survey, which uses a beamsize of 5:004 (FWHM). Therefore it has a slighter lower resolution than the 1.4-GHz VLA B-array observations of Parma et al. (1986), but it has a comparable sensitivity. The FIRST radio map (see the inset in Fig. 3) confirms the presence of the jet and counter-jet like features. The position of the core, determined by fitting a Gaussian to the central component, is [FORMULA] in Right Ascension and [FORMULA] in Declination (B1950.0), with a positional uncertainty of [FORMULA]. The radio axis of the structure visible in FIRST and the map of Parma et al. (1986) has the same position angle ([FORMULA], measured CCW from the North) as that of the pc-scale structure mapped by Henstock et al. (1995). There is therefore no indication of a change in the jet axis between pc and kpc-scales. A summary of the properties of the central component, as measured in the FIRST radio map, is presented in Table 1.

[FIGURE] Fig. 3. Radio map of the source B1144+352 from our 1.4-GHz WSRT observations. The contours are at (-0.3, 0.3, 0.42, 0.6, 0.85, 1.2, 1.7, 2.4, 4.8, 9.6, 19.2, 76.8, 307.2) mJy beam-1. The greyscale ranges from -0.3 to 2.0 mJy beam-1 and clearly shows the ringlike artefacts in the radio map. The inset is a radio map of the central source from the FIRST survey. Contours are at (-0.45, 0.45, 0.64, 0.90, 1.27, 1.8, 3.6, 7.2, 14.4, 28.8, 115.2, 460.8) mJy beam-1.


Table 1. Properties of the central component of B1144+352 as measured in the FIRST survey map.
The jet is the eastern component, the counter-jet the western. The flux density at 1435 Mhz, [FORMULA], of the core is the peak flux determined using a Gaussian fit to this component on the map. The date of the FIRST observations is 3 July 1994. The flux densities of the jet and the counter-jet have been measured on the map after subtraction of the core. [FORMULA] is the projected distance between the most outer peak in flux density of the jet/counter-jet and the core. The position angles, PA, are measured counter-clockwise from the North.


Table 2. 1.4-GHz flux densities of the core of B1144+353.
Column 1 refers to the epoch of observation. Column 2 gives the flux density at 1.4 GHz. Column 3 gives the reference for the flux density measurement (C75: Colla et al. 1975a; P86: Parma et al. 1986; F87: Fanti et al. 1987; S95: Snellen et al. 1995; F: FIRST survey; W: This paper).

3.2. 1.4-GHz WSRT observations

To better understand the nature of the extended structures around the GPS source, we observed B1144+352 for 12 hr with the WSRT at 1.4 GHz on August 31, 1997. We used a total bandwidth of 50 MHz, divided into five channels of 10 MHz each and centered on 1395 MHz. 11 out of 14 antennas were operational during these observations. The sources 3C 286 and 3C 147 were used as primary calibrators. The data have been reduced and mapped using the NFRA data reduction package NEWSTAR . We used the flux density scale of Baars et al. (1977) for absolute gain calibration. During the reduction process, we encountered a problem in the data, manifested as a system of ringlike structures around the bright central source (see Fig. 3). Our best guess is that it is due to some kind of time-variable baseline-based errors. The radial decrease of this effect ensures that the observed structure of the diffuse lobes is not influenced much by it.

The final total power map has an RMS noise of 0.053 mJy, although this value increases substantially towards the center of the map. The FWHM beamsize of the restoring beam is [FORMULA]. The Stokes' Q and U parameter maps have the same RMS noise as the total power map. Only the Q-map has similar ringlike structures to those seen in the total power map. Both diffuse structures have been detected and mapped. The FRII-type radio lobe morphology of the eastern structure (see also Fig. 4) is obvious, and we also detect a leading hotspot. Surprisingly, the western structure (see also Fig. 5) appears to be a superposition of a separate FRII-type radio source and a radio lobe with a tail towards the south. The position angles of the E-field vectors of the linear polarized emission in the two extended structures (Fig. 4 and Fig. 5) are consistent with those measured in the NVSS (Fig. 2b). Near the core, we only measure significant polarization at the position of the most eastern component of the FIRST radio map. The polarized intensity is [FORMULA] mJy, and the fractional polarization of this component is therefore [FORMULA]%. Integrated over the whole central component, the fractional polarization is [FORMULA]%, which is equal to what we found in the NVSS data. Towards the radio core no significant polarized emission is detected. The position angle of the E-field vector of the linear polarized emission is [FORMULA], and so the observed projected E-field is parallel with the radio axis.

[FIGURE] Fig. 4. Contourplot of the eastern radio structure from our 1.4-GHz WSRT observations. Contours are at (-0.3, 0.3, 0.42, 0.6, 0.85, 1.2, 1.7, 2.4) mJy beam-1. The vectors are the observed E-field, and their length corresponds to the polarized intensity ([FORMULA] mJy beam-1). The two crosses give the position of sources detected in the FIRST survey; their size is arbitrary and not related to any physical quantity or the positional accuracy of these sources.

[FIGURE] Fig. 5. Contourplot of the western radio structure from our 1.4-GHz WSRT observations. Contours are at (-0.3, 0.3, 0.42, 0.6, 0.85, 1.2, 1.7, 2.4, 4.8) mJy beam-1. The vectors represent the observed E-field of the linear polarization; their length corresponds to the polarized intensity ([FORMULA] mJy beam-1).

3.3. ROSAT observations

B1144+352 has also been observed in X-rays by ROSAT on 10 June 1993. We retrieved the X-ray data, obtained with the High Resolution Imager (HRI, David et al. 1993), from the ROSAT archive. The X-ray image available from the archive, which has been binned with 8 arcsec pixels (the FWHM of the HRI is [FORMULA]), shows a rather strong point-like source, whose position, within the ROSAT pointing errors, agrees well with that of the central radio source in the FIRST radio map ([FORMULA], [FORMULA]). In this 3000 s exposure, [FORMULA] net counts have been detected from this source. Two other sources are visible near the edge of the field, but there are no other significant detections within the extent of the radio source.

To investigate the structure of the X-ray source in more detail we rebinned the original data on a grid with 2 arcsec large cells. The result is shown in Fig. 6a, which is a contourplot of the raw datacounts. We have set the coordinate frame such that after convolution with an [FORMULA] (FWHM) Gaussian the position of the peak corresponds to the peak of the radio emission from the FIRST survey. For comparison, we have also plotted the ROSAT HRI point response function (PRF  1) after scaling it to match the peak intensity in the image. The X-ray source appears to be extended in a direction close to the radio axis. This extension did not disappear if we shifted our binning grid by 1 arcsec in either direction. Further, the exposure was continuous, and therefore could not be hampered by pointing errors between different exposures. In Fig. 6b we have plotted the radial profile of the counts. The dashed line gives the PRF of the HRI. There is some evidence for extended structure in the inner 8 arcsec. However, a much longer exposure would be needed to firmly establish it.

[FIGURE] Fig. 6. a  Contourplot representation of the raw counts in the ROSAT HRI data in the area surrounding the X-ray source associated with B1144+352. The counts have been binned on a grid with 2 arcsec large cells. Contourlevels are at [FORMULA] counts arcsec-2. The coordinates are relative to the peak of the central radio component visible in the FIRST survey. The inset in the lower left corner shows the HRI point response function, scaled to the peak flux of the image and using the same contourlevels. b  Plot of the radial profile of the raw ROSAT counts. The dashed line is the HRI point response function, normalized to the counts in the innermost bin. The dotted line is the background level, measured in an annulus from 200 to 800 arcsec from the central source.

To convert the X-ray counts to a received flux, we used the PIMMS mission simulation program (version 2.4b; Mukai 1993). The X-ray flux can either be due to the AGN, in which case a powerlaw spectrum is expected, or to thermal bremsstrahlung in a shocked ISM. In case of B1144+352, the reality is probably a mixture of these. Therefore we calculate the X-ray luminosity using both a powerlaw and a thermal bremsstrahlung model and regard these as two limiting cases. For the powerlaw model we use a photon index of 1.8, which has been found to be the average for high-luminosity AGN (Williams et al. 1992). To correct for galactic extinction, a galactic atomic hydrogen column density of [FORMULA] cm-2 towards B1144+352 has been used, as determined from the Leiden-Dwingeloo HI -survey (Hartmann 1994). We have no information on the amount of extinction in or near the source, and so we neglect this contribution in our calculations. We find a total X-ray flux density of [FORMULA] erg cm-2 s-1 in the [FORMULA] keV band. This translates to a total X-ray luminosity of [FORMULA] erg s-1 between 0.1 and 2.4 keV. For the thermal bremsstrahlung model we use a gas temperature of [FORMULA] K ([FORMULA] keV) and a metal abundance of 0.25 solar (model RSQ70 in PIMMS ). Using the same galactic HI column density as before, we find a total flux of [FORMULA] erg cm-2 s-1 in the [FORMULA] keV band. The total emitted power in this band then yields [FORMULA] erg s-1.

Crawford & Fabian (1995) find similar values for the four narrow-line radio galaxies at [FORMULA] in their sample of powerful radio galaxies from the 3CR sample. Worrall & Birkinshaw (1994) find an order of magnitude lower values for low-power radio galaxies. The X-ray luminosity of B1144+352 is therefore more comparable to that of powerful radio galaxies. The detection of such a strong X-ray source in a GPS galaxy is surprising. O'Dea et al. (1996b) observed two GPS radio sources with the ROSAT PSPC and no X-ray emission was detected. Therefore the X-ray luminosity of these sources must be [FORMULA] erg s-1 in the 0.2-2 keV band (3[FORMULA] upper limit), which is a factor of three below the luminosity of B1144+352. O'Dea et al. explain the non-detections by stating that these sources either must be intrinsically weak in X-rays and/or they must be highly obscured by cold gas. To obscure a source with an X-ray luminosity of [FORMULA] erg s-1 in their observations, a column density of a few times [FORMULA] cm-2 is needed. The high X-ray luminosity of B1144+352 is therefore suggestive of a much lower column density towards this source: for the powerlaw model, a column density of a few [FORMULA] cm-2 would imply an intrinsic X-ray luminosity of [FORMULA] erg s-1. However, a more modest column density of [FORMULA] cm-2 yields an emitted power of [FORMULA] erg s-1, which is only 35% higher than the power that we find assuming no intrinsic absorption.

As will be shown in Sect. 4.2, we believe that the extended radio components are associated with the GPS source. This means that some medium must be present to confine the radio lobes. Such a medium may reveal itself through the emission of X-rays. However, the HRI data show no sign of a large extended X-ray emitting halo. To find an upper limit for the luminosity of such a halo we have summed all counts in a circular area of radius 250 kpc centered on the host galaxy of B1144+352. This is the typical core radius of low-luminosity clusters (Mulchaey et al. 1996; Jones et al. 1998) and therefore most of the X-ray halo emission is expected to come from this region. To estimate the background contribution we have summed all counts in a 600 arcsec wide annulus with an inner radius of 200 arcsec (330 kpc). In the inner circular area we measure 391 counts. The expected number of background counts in this region is [FORMULA], and the inner [FORMULA], or 15 kpc, contains [FORMULA] counts. Therefore the number of counts from a possible X-ray halo in the HRI data, integrated from 15 kpc outwards to 250 kpc, is [FORMULA] and thus not significant. To calculate an upper limit for the luminosity of the halo we have again used the PIMMS mission simulation program (version 2.4b). We further assume that the density profile, [FORMULA], of the halo follows a modified King model, [FORMULA], with core radius [FORMULA] kpc and [FORMULA] (e.g. Forman & Jones 1991). The radial surface brightness profile then behaves as [FORMULA], which implies that 10% of the X-ray luminosity emitted innerhalf of 250 kpc originates from within a radius of 15 kpc. For the cluster gas, we have assumed a Raymond-Smith spectral model (model RSQ70 ) with a temperature of [FORMULA] K (0.86 keV) and a metal abundance of 0.25 solar. These are reasonable values for the medium in groups of galaxies (Mulchaey et al. 1996). Using the 1[FORMULA] flux-level as an upper limit, we find that the total received flux in the [FORMULA] keV band of the ROSAT HRI detector must be [FORMULA] erg s-1 cm-2. To compare this value with the literature we calculate the expected flux in the [FORMULA]) keV band, which results in an upper limit of [FORMULA] erg s-1 cm-2. At the redshift of B1144+352, this translates into a 1[FORMULA] upper limit of the X-ray luminosity of [FORMULA] erg s-1 in the inner 15 to 250 kpc. Corrected for the inner 15 kpc, this implies an upper limit of the X-ray halo luminosity within a radius of 250 kpc of [FORMULA] erg s-1. Extrapolated to much larger radii, assuming the modified King profile, this would yield [FORMULA] erg s-1.

The luminosity we find is much lower than those for luminous X-ray clusters ([FORMULA] erg s-1) which is consistent with the lack of a rich optical cluster on the DSS. Mulchaey et al. (1996) and Mulchaey & Zabludoff (1998) have investigated the X-ray properties of a sample of nearby poor groups of galaxies. In the majority of cases, they find extended halo components with a typical luminosity range of [FORMULA] erg s-1. They attribute this to a low-mass version of the intra-cluster medium found in X-ray luminous clusters. The upper limit we determined from ROSAT HRI data for B1144+352 therefore does not exclude a low-mass intra-cluster medium as found in poor groups of galaxies.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1999

Online publication: November 26, 1998