Astron. Astrophys. 342, 57-68 (1999)

3. Spatial analysis

3C 219 was observed with the ROSAT HRI for a total of 28.6 Ks between April 10 and May 8, 1997. Standard procedures have been employed within the MIDAS/EXSAS software. The source counts, estimated with a maximum likelyhood (ML) detection algorithm as described in Cruddace et al. (1988), are 84730 corresponding to a background subtracted count rate of 0.03 cts s^-1 in the 0.1-2.4 keV band. Since the ML algorithm has been conceived to study point-like sources and is not efficient for extended sources, the source flux has been computed measuring the counts in a circle of radius. We find evidence of emission outside from the nucleus accounting for some 15% of the total flux (see Sect. 4). The resulting total count rate (nuclear and extended emission) is 0.035 cts s^-1. The total HRI flux of 1.6 erg cm^-2 s^{- 1} is consistent, within about 10%, with that derived from the PSPC observation (see Sect. 2.1) assuming the spectral parameters reported in Table 1.

It is known that the intrinsic spatial resolution of the ROSAT HRI of FWHM is blurred by the errors due to the relatively poor knowledge of the pointing position as a function of time. We have tried to correct for the residual aspect solution errors following a procedure developed by Harris et al. (1998) which selects observing periods with the same roll angle and folds the data according to the wobbling period of 402 s. Several subimages are created according to the source statistics and the wobble period, then shifted to a common center and coadded.

The radial profile of the innermost region of the wobble corrected image is then compared with an inflight calibrated PSF, kindly provided by I. Lehmann (private communication), derived from the average profile of 21 bright stars (Fig. 2). From the analysis of the azimuthal distribution of the counts within ( 50 kpc) from the source peak intensity we find that the nuclear source is resolved in the north-south direction (Fig. 2). A Kolmogorov-Smirnov test shows that the observed distribution of the counts in the north-south direction differs from the PSF at the 99% confidence level, while the test is not conclusive in the case of the east-west direction.

Fig. 2. The wobble corrected ROSAT HRI radial profile in the 0.1-2.4 keV band. The counts have been taken from two symmetric sectors: one in the north-south direction (filled squares) from to and from to clockwise starting from the north, and one in the east-west direction from to and from to . The expected profiles of a point source according to the theoretical PSF (continuous line) and to the inflight calibrated PSF (dotted line) are reported normalized to the source intensity and background level.

In order to get more information on the structure an X-ray image of 3C 219, smoothed with a circular gaussian of FWHM= (Fig. 3), has been produced with the EXSAS package, while the further image processing has been performed with the Astronomical Image Processing System (AIPS) package. An extended structure up to in radius is indeed detected in the north-south direction (Fig. 3).

Fig. 3. The X-ray image of the bright central source coincident with 3C 219, smoothed with a Gaussian of FWHM = , is shown in the left panel. Contour levels are 0.15, 0.25, 0.4, 0.5, 0.7, 1, 2.6, 3.4, 4.3, 5.1, 6.8, 7.7 cts pixel-1, with the peak corresponding to 8.56 cts pixel-1 (1 pixel = ). The asymmetry of the central source (highest contour levels) can be attributed to a slight degradation of the PSF (FWHM= with a position angle of clockwise from N), while the lower level structure is likely to be due to real source extension. This is confirmed by the profile across the source (right panel), obtained at the position angle of (clockwise from N), which clearly shows that the excess emission at distance of from the peak cannot be due to the PSF.

The smoothed PSF appears to be elliptical (FWHM= ) with the major axis positioned at clockwise from the north, at variance with the elongation of the extended structure.

Of course, the small FWHM used in the previous smoothing procedure does not allow a detection of the outer extended low brightness features. The X-ray image of Fig. 4 has been produced with the EXSAS package by binning the photon event table in pixels of and by smoothing the map with a circular gaussian of . With these values the signal to noise ratio is good enough to image the low brightness extended emission significantly above the background. The ROSAT HRI image has been shifted by and in right ascension and declination, respectively, in order to align the X-ray peak with the radio core position. The X-ray map has been overlaid to the optical POSS-II digitized image of the field in Fig. 4. Since 3C 219 is in a cluster of galaxies, the optical image is crowded and a number of objects fall within the X-ray structure; no coincidence with relevant optical objects is found.

Fig. 4. The X-ray image of 3C 219 (contours) superposed on the red optical POSS-II digitized plate (gray-scale). The X-ray flux is dominated in the central region by a point-like source, but an extended component is detected up to from the nucleus. The X-ray image has been produced by binning the photons in pixels of and by smoothing the map with a circular gaussian of . The plotted contour levels are: 0.05 (dashed), 0.2, 0.22, 0.24, 0.26, 0.31, 0.65, 1.5, 3, 6, 9 cts pixel-1 (1 pixel = ), with the first two contours respectively at the 2- level below (dashed) and above the background. The background is 0.125 cts pixel-1. The optical image also shows a companion galaxy at S-E of 3C 219.

The structure of the X-ray brightness distribution between  from the core can be enhanced if the central unresolved component is subtracted from the HRI image. The model used in the subtraction is a circular gaussian with the peak coincident with the X-ray peak. This function represents the convolution between the instrument PSF and the smoothing function applied to the image. In fact, although the instrument PSF is not exactly a gaussian, the smoothing gaussian dominates the resulting convolved PSF.

In order to subtract the nuclear component from the total HRI image we have made use of the spectral analysis results of Sect. 2. The 0.1-2.4 keV flux of the nuclear emission has been estimated by assuming the best fit spectral parameters of the high energy absorbed component in the 0.1-10 keV fits. Taking into account the different responses of the HRI, PSPC and ASCA detectors for a given spectrum, as well as the errors on the best fit spectral parameters, the nuclear flux is estimated to account from 55% to 70% of the HRI counts. The subtraction procedure has been performed several times constraining the amplitude of the nuclear component within this interval. We also note that the largest contribution of the point-like nuclear source (leaving zero counts after the subtraction at the position of the peak) is % of the total net counts; this limit is implied from the HRI data alone. The residual map resulting from the subtraction of the nuclear source with a representative amplitude of 64% is shown in Fig. 5 superposed on the VLA radio image at 1.4 GHz (Clarke et al. 1992). It is evident that the residual X-ray isophotes are strongly correlated with the radio extension and appear to be elongated along the radio structure. We distinguish three main regions: one coincident with the low brightness part of the northern radio lobe (N), one positioned between the nucleus and the southern radio hot-spot (S) and the strongest one (C) centered on the radio-core. The C-component appears like a curious eight-shaped figure with the axis inclined by with respect to radio-axis, but the isophotes become more aligned with the radio-axis with increasing distance from the nucleus. This basic structure persists even by varying the amplitude of the point-like source within the allowed interval and/or by applying a slightly different in the subtraction procedure. Furthermore one can notice that the C-component seems to show a remarkable continuity with the extended structure of Fig. 3.

Fig. 5. The X-ray image of 3C 219 (contours) after subtraction of the central point-like source superposed on the VLA radio image at 1.4 GHz with resolution (gray-scale). The contours are: 0.22, 0.24, 0.26, 0.29, 0.33, 0.45, 0.65, 1 cts pixel-1 (1 pixel= ). The conversion from counts to X-ray brightness (erg cm-2 s-1 arcsec-2) is for a power law spectrum () or for a thermal spectrum ( keV). The C-component contains about 60% of the counts.

© European Southern Observatory (ESO) 1999

Online publication: December 22, 1998
helpdesk.link@springer.de