3. X-ray data
We have observed the cluster with the ROSAT HRI and the ASCA GIS and SIS detectors. The HRI has a high spatial resolution of , which provides a high resolution X-ray surface brightness profile, but it has no energy resolution. ASCA on the other hand has a low spatial resolution () but relatively high energy resolution and high sensitivity in the energy range 1-10 keV, which provides a reliable gas temperature measurement for clusters of galaxies.
3.1. Spectral analysis
The cluster was observed with ASCA using both detectors of the Gas Scintillation Imaging Spectrometers (GIS) and Solid-state Imaging Spectrometers (SIS) in February 1996. The SIS detectors were operated in 1-CCD mode. The data was screened and cleaned according to the standard procedures recommended (Day et al. 1995). The spectra were extracted from the central radius from the GIS2 and GIS3 detectors, excluding one discrete source. Similarly, spectra were extracted from the central radius from the SIS0 and SIS1 detectors. A standard blank-sky exposure screened and cleaned in the same way as the cluster field was used for background subtraction by extracting a background spectra from the same region on the detector as the cluster spectra. The spectra were grouped into energy bins such that the minimum number of counts before background subtraction was above 40, which ensures that statistics would still be valid. The 4 spectra from each detector were simultaneously fitted with a Raymond-Smith thermal spectra (Raymond & Smith 1977) with photoelectric absorption (Morrison & McCammon 1983) from the XSPEC package (Fig. 7). We adopted the abundance table with the relative abundance of the various elements from Feldman (1992). The free parameters were the gas temperature (), Galactic neutral hydrogen absorption column density (N(H)), metal abundance (abund) and the emission integral. All 4 spectra were to have the same value for the free parameters except for the emission integral, since the GIS and SIS PSF were different and the extraction regions were smaller for the SIS spectra compared to that of the GIS. The two GIS spectra were assumed to have the same emission integral but different from the SIS emission integrals. Results of the best simultaneous fit to the 4 spectra along with fits to the individual spectra are tabulated in Table 1. Only data in the energy range where the effective area of the detectors are cm2 were used for the spectral fitting, i.e. 0.6-7.5 keV for SIS data and 0.85-10.0 keV for GIS data.
Table 3. (continued).
The neutral hydrogen column density derived from the ASCA data were 2 times larger than the N(H) ( cm2) measured from radio data by Starck et al. (1992). If we try to fix N(H) to the value determined by Starck et al. (1992), then there is obvious discrepancy between the model spectrum and the SIS data below 1 keV. Unfortunately, there is no PSPC data available for this cluster to place definitive constraints on the N(H) value. It is possible that there is a local over-density of absorbing neutral gas along the line-of-sight to the cluster, though it is more likely to be a calibration error for the SIS detector. Calibration of the low-energy part of the SIS detector is known to produce erroneous results such that it favours a high N(H) inconsistent with PSPC results (Schindler et al. 1998; Liang et al. 2000). In view of the possible calibration error for the SIS, the data were also fitted with the above models with a fixed N(H) given by Starck et al. (1992) by excluding the SIS data below 1 keV. The temperature thus deduced was significantly higher than before. In the following studies, we will adopt these parameters deduced from a simultaneous fit of data from the GIS detectors in the energy range 0.85 to 10 keV and the SIS detectors between 1 and 7.5 keV.
3.2. ROSAT HRI data
The cluster was observed by the ROSAT HRI in February (7.6ksec) and August (36ksec) 1996. The X-ray centroid was found to be 15:40:08.1 -03:18:17, which is from the position of the cD galaxy 15:40:07.96 -03:18:16.7. The positional error for the X-ray centroid is , hence the small apparent displacement between the cD position and the X-ray centroid is insignificant. The X-ray surface brightness was obtained by extracting the photons in a radius of and the background was extracted from an annulus of radius from the X-ray peak. Discrete X-ray sources were excluded from the extraction. There were 8 discrete X-ray sources in the HRI image. Fig. 1 shows the X-ray contours overlaid on the optical image of the cluster field. The X-ray image show significant substructure in the centre with an overall elliptical appearance. The discrete X-ray sources at 15:40:07.2 -03:19:53 is embedded in the cluster emission. The relative astrometry between X-ray and optical was checked using 4 of the discrete X-ray sources that had clear optical identification. The X-ray positions had a maximum displacement of relative to the optical coordinates. The X-ray contours were adjusted to the optical coordinate system using the 4 discrete X-ray sources, which gave a relative astrometric accuracy of between optical and X-ray coordinates.
convolved with the instrument PSF is shown as a solid curve superimposed on the data. The best fit gave and . The uncertainties quoted are . The total X-ray luminosity within the central radius is ergs s-1 in the ROSAT band of 0.1-2.4 keV, assuming cm2, keV (or K), and an abundance of 0.22. The X-ray luminosity thus deduced is consistent with that estimated from the ROSAT all-sky survey (Pierre et al. 1997). The corresponding bolometric X-ray luminosity is ergs s-1. The central electron density was thus derived to be cm-3. The central cooling time for this cluster is yr, greater than a Hubble time.
© European Southern Observatory (ESO) 2000
Online publication: December 11, 2000