Astron. Astrophys. 321, 64-70 (1997)

2. X-ray observations

A 2717 was observed with the Rosat PSPC in June 1993 with a total integration time of 13.7 ksec. The PSPC has a field of view of and is sensitive to photons with energy between 0.1 and 2.4 keV. The pointing centre was 00 03 19.2 -35 57 00 (J2000.0). The resolution of the PSPC is at the centre of the detector at keV. The resolution degrades with decreasing energies and increasing distance from the detector centre.

2.1. Analysis of the X-ray data

The Rosat PSPC data was analysed using the EXSAS package (Zimmermann et al. 1994) in MIDAS. Discrete sources were detected first and subtracted from the data, so that they can be removed from the intracluster gas emission.

2.1.1. Discrete sources

The discrete sources were detected from a broad band image (including photons of energy from to keV) of pixels excluding the regions affected by the support ribs of the detector. The sources were detected using a sliding window source detection technique whereby a pixels window was slid over the image. The source counts were obtained from this sliding window and background counts were determined from the surrounding 16 pixels. This method is best suited for the detection of point or slightly extended sources. Finally a maximum likelihood method (Cruddace et al. 1988) was applied to the sources detected to ascertain the reality and the extension of the sources. Table 3 lists all the sources found with maximum likelihood parameter . These sources were also scanned by eye and found to be consistent with real sources. None of the sources in Table 3 can be considered as extended. The positions and count rates for the sources are listed in Table 3. Correlations of the X-ray point sources with the SIMBAD and NED on-line catalogues were made, but no identification was found. A search through the COSMOS source list with an error radius of produced a number of identifications. The optical position and magnitudes of the identified objects within the error radius are tabulated in Table 4.

Table 3. Discrete X-ray sources in the field of A 2717

Table 4. Optical I.D. of the discrete X-ray sources

In order to separate the contributions from the discrete sources and that of the cluster, we subtract the discrete sources found above, from the photon events file before making images of the field in a soft (0.1-0.4 keV) and a hard (0.4-2 kev) band (Fig. 1). The peak of the hard image was identified with the central D-galaxy within uncertainty limits. The hard image peaks at 00 03 12.5 -35 56 02 which is north of the optical position of the D-galaxy. The majority of the discrete sources falling within the central panel of the detector are systematically found to be between - north of their optical counter part, suggesting a possible systematic error of in the Rosat positions. A similar discrepancy of between Rosat PSPC positions and optical positions was found by Pinkney et al. (1994). The peak of the soft image was displaced from that of the hard image and the D-galaxy, but coincides with a faint galaxy (see Fig. 1). This displacement is consistent with the galaxy being an X-ray source with a spectrum considerably softer than the cluster (e.g. an AGN). This soft source was not detected with the above detection algorithm in the broad-band because of the soft source spectrum and the proximity of the source to the centre of the cluster emission which was predominately composed of hard photons. The position of the soft source was 00 03 16.2 -35 56 12 (J2000.0).

Fig. 1. The X-ray soft and hard contour images are superimposed on the digitised sky-survey (SERC J plates) image of the field. The axes are in decimal degrees. The X-ray maps have been smoothed with a Gaussian filter of FWHM . The hard-band image (0.4-2keV) is represented by solid contours of levels (1.2, 4.0, 6.8, 10.8) counts s-1 arcsec -2 ; the soft-band image (0.1-0.4keV) is represented by the dotted contours of levels (0.8, 1.6, 2.4, 3.2) counts s-1 arcsec -2 ; and the optical image is in grey-scales. The central D-galaxy is identified with the peak of the hard image, whereas the peak of the soft image is identified with a faint galaxy and is displaced from the peak of the hard image.

There were 2 Abell clusters, A4074 and S1170, and a galaxy group ESO 349-26 that were identified with X-ray sources near the edge of the detector. These sources were not detected using the formal detecting methods described above, owing to the proximity of the sources to the edge of the detector.

2.1.2. The cluster

The Rosat PSPC has limited sensitivity in the high energy part of the X-ray spectra, namely keV, thus it can only constrain the temperature of relatively cold clusters with high precision.

The total cluster spectrum was obtained by including all the photons within a radius of (i.e. 0.65 Mpc) but excluding all the regions contaminated by the discrete sources, i.e. a circular area with a radius of the FWHM of the PSF at the position of the discrete source. The background spectrum was calculated from an annulus centred at the cluster centre and between a radius of and .

The X-ray total flux and temperature were derived by fitting a Raymond-Smith spectrum with galactic absorption to the spectrum, using a program kindly provided by M. Arnaud. A fit was made by allowing the X-ray emission measure, temperature and the neutral hydrogen column density to vary but fixing the abundance. A change in abundance between 0.3 to 0.5 only produced a 0.1 keV change in the best fit temperature. For an abundance of 0.3, the best fit found for the temperature is shown in Table 5 which is rather low for a cluster. The neutral hydrogen density was found to be cm² as compared to cm² from the radio surveys (Stark et al. 1992). The X-ray temperature was very well constrained and relatively independent of .

Table 5. Results from X-ray analysis of A 2717

The cluster surface brightness was obtained by first subtracting the point X-ray sources from the X-ray photon events file and producing a sky subtracted azimuthally averaged count rate profile (see Fig. 2). Since the PSPC has a much lower background level in the hard band and the cluster emission tends to dominate the hard band, we have chosen only the data between channels 42 to 201 ( 0.4-2 keV) to produce the surface brightness profile.

Fig. 2. The cluster X-ray surface brightness profile in the energy range of 0.4-2 kev. The data points with error bars are shown along with the cluster model X-ray surface brightness profile (including background) after convolution with the PSF (solid line).

The radially averaged X-ray surface brightness profile after the deconvolution of the PSF, was well fitted by a function of the form :

with and core radius or 0.047 Mpc.

© European Southern Observatory (ESO) 1997

Online publication: June 30, 1998
helpdesk.link@springer.de