Astron. Astrophys. 332, 449-458 (1998)

2. Observations and data reduction

The galaxies observed by us with the ROSAT HRI are listed in Table 1, together with the observing times of the individual pointings and a few basic parameters. The total amount of time allocated for the current project was ca. 90 ks.

Table 1. ROSAT HRI observations

The data reduction was performed in the IRAF/PROS software package by NOAO and SAO. Since there are no indications of bad data due to high background, no further event selection than that by the ROSAT Standard Analysis Software System (SASS) was applied. From the event lists we created images by summing up detector pixels, thus creating image pixels. The background was determined in several source free regions of each image. Since we found no evidence for systematic background differences across the field of view, mean values of the levels measured in the source free regions were subtracted as constant offsets. The source free areas for the background determination were chosen to be distributed around the target galaxies randomly. Only in the case of NGC 4921 the emission of the nearby Coma cluster of galaxies was avoided by systematically measuring the background only eastward of the target. After this step, the images were divided by the net integration time (Table 1) in order to obtain count rates.

In those cases in which more than one pointing was performed in order to obtain the full amount of requested observing time, we added up the individual images before normalising them in order to increase the signal-to-noise ratios. The 1994 Aug. 08 data of NGC 6329 were not used because of the limited number of photons received from the galaxy and nearby sources and the resulting difficulties in determining the correct attitude solution compared to the other two runs.

In the final images the angular resolution of the ROSAT X-ray telescope, including pointing jitter degradation of the PSF, was measured by fitting gaussian profiles to the emission distribution of the brightest point sources and deconvolving these by the smoothing function that had been applied before. The resulting values were subsequently used to deconvolve the observed source profiles and thereby determine their true extent.

In three cases, NGC 533, NGC 4104 and NGC 4921, the alignment of the optical and X-ray images was achieved by determining the mean offset between both frames based on optical identifications of serendipitously detected X-ray sources in the field of view. For the different galaxies, 7 to 16 sources were used.

In the cases of NGC 2832 and NGC 2885 we could not find suitable X-ray sources in the field of view. In the NGC 6329 pointing, we were not able to reliably identify any optical counterpart of the 7 strong X-ray point sources in the field of view. Therefore, we had to adopt that the X-ray and optical centres of these three galaxies are aligned. This appears to be a sensible assumption, because the X-ray emission of all three is strongly peaked and - in NGC 2885 - time variable. The distribution of the surrounding diffuse emission in NGC 2832 and NGC 6329 is radial symmetric with respect to the location of the strong central source. The pointing offset corrections for all six galaxies lie in the range from to .

We used the PROS task imcnts to sum up the total number of photons received from all objects after background subtraction. In all cases, the integration was performed in a circular aperture with a radius of around the optical centre of the galaxy (which, for all practical purposes, spatially coincides with the centre of the X-ray emission for all six targets). The resulting count rates were used to calculate the source fluxes.

© European Southern Observatory (ESO) 1998

Online publication: March 23, 1998
helpdesk@link.springer.de