Astron. Astrophys. 319, 413-429 (1997)

2. The data

We have compiled a sample of all quasars from the Véron-Cetty - Véron catalogue (1993, from now on VV93) detected by ROSAT in the ROSAT All-Sky Survey (Voges 1992), as targets of pointed observations, or as serendipitous sources from pointed observations as available publicly from the ROSAT point source catalogue (ROSAT-SRC, Voges et al. 1995).

Many of the radio-loud quasars presented here from the ROSAT All-Sky Survey  (RASS) have been published recently from correlations of the RASS with the Molonglo 408 MHz Survey (Brinkmann et al. 1994, Paper I) and the Greenbank 5 GHz Survey (Brinkmann et al. 1995, Paper II). In total, these papers contain 360 objects, of which 346 are listed in VV93. This list is supplemented with 164 objects from the VV93 catalogue found additionally in more recent data from the ROSAT All-Sky Survey re-processing (SASS) or outside the spatial and flux limits of the two radio lists above, and a few new objects from recent publications.

The cross correlation of the VV93 radio quasars with the ROSAT-SRC, using a distance criterion between the optical and the X-ray source of , yields a list of 294 objects detected in pointed observations, many of them observed more than once. Amongst them, 100 objects are already included in Paper I and Paper II.

Taking into account the overlaps, the total number of ROSAT detected radio quasars from the above three sources is 654 objects, of which 360 were seen in the RASS only, 126 are only seen in pointed observations, and 168 were seen in both the RASS and pointed observations. It should be noted that the ROSAT All-Sky Survey detected sources form a well defined sample as the survey's limiting sensitivity is rather uniform (a few times erg cm ) while the sources from pointed observations have vastly different exposures.

For objects which have been seen in both the ROSAT All-Sky Survey and pointed observations, we use the data from pointed observations for the determination of the spectral parameters, as Survey data suffer from much larger statistical errors due to the fact that the exposure in pointed observations is generally longer by more than an order of magnitude.

For 92 objects with sufficient numbers of source counts, spectral fits were made using a simple power law model with neutral absorption. For the rest, the power law indices (assuming Galactic or free absorption) were estimated using the two hardness ratios given by the SASS (Voges et al. 1992) by the method described in Paper I. We estimate the errors of the power law indices and the values from the errors of the hardness ratios as described by Schartel (1995).

For 123 objects, no reasonable power law index and absorption column density could be obtained from the given hardness ratios, either due to the low number of source counts (with correspondingly large errors of the hardness ratios) or intrinsic deviations of the source spectra from simple power laws. For some objects published data exist in the literature; we use these results if they are of superior quality and mark them with their references.

82 objects have been observed in pointed observations more than once. For them, the mean values of and , as well as their corresponding 1 errors are calculated, assuming that the source spectra do not vary between the different measurements. In fact, we found that in most cases the changes of the spectral properties are not significant and within the errors of the fitted parameters. The intensity variations of the sources will be discussed in Sect. 3.

The X-ray fluxes in the ROSAT (0.1-2.4 keV) band are calculated from the count rates using the energy to counts conversion factor (see Paper I) for power law spectra and Galactic absorption. For the power law index we use the value obtained for the individual source if the estimated error is , and we take an average value for the objects with low signal-to-noise ratios. It must be noted that slight differences in the power law slope do not affect the flux determination notably; the dominant parameter is the amount of absorbing matter in the line of sight.

In Table I we list the relevant information of all 654 quasars, starting with the IAU designation and the common name. Objects for which radio emission has been detected but which don't qualify as radio-loud according to the above mentioned flux criterion, are marked with a star. Then, following the J2000 positions, we list the redshift and optical magnitudes as found in VV93. In columns 5 and 6 we give the radio flux at 5 GHz as well as the radio spectral index . The 5 GHz flux densities are taken from the 87GB radio survey (Paper II), from the NED data base, and from VV93. For most objects multi frequency radio data are available from NED which allows a good determination of the radio spectra. For objects with obvious power law spectra or objects with large scatter in the measured fluxes the spectral indices are determined by fitting a power law slope to the data. For objects showing a distinct non-power law spectral shape or when only few flux values are measured we calculated two point indices between 5 GHz and 1.4 GHz, if available.

In column 8 we list the (0.1 - 2.4 keV) unabsorbed X-ray flux obtained, as described above, with the assumption of a power law spectrum (i.e. the flux which would be measured without interstellar absoption). The given errors are the statistical errors from the count rates only. However, for sources with a small number of total counts (mostly from the Survey) the systematic errors can be of the order of % (see Paper I). For strong sources the assumed simple power law slope is often an inappropriate representation of the spectrum. In both cases the systematic uncertainties can be considerably larger than the purely statistical errors and the errors given in Table I should, therefore, be taken as lower limits. We then give the X-ray power law photon index with errors obtained under the assumption of Galactic absorption, an indication whether the object was seen in the Survey only (S), in a pointed observation (P), or in both (SP). The last entry gives a reference to other published results.

From the point of view of a radio classification the list contains a total of 297 flat spectrum and 201 steep spectrum quasars. We considered sources with as flat spectrum objects and as steep spectrum sources those with . 24 objects were classified as 'Gigahertz Peaked Spectrum' (GPS) and 12 as 'Compact Steep Spectrum' (CSS) sources. For the rest of the objects the radio information is insufficient for a classification, unclear, or the objects are not radio-loud.

Many of the radio-loud objects in the VV93 compilation are not primarily radio detected quasars but have been discovered in other wavelength bands (X-rays, optical) and subsequently found to be radio emitters as well. To avoid any possible detection biases which might influence the correlations of the broad band properties of the sources we will restrict the statistical analyses to objects only, which were first detected in well defined radio surveys. These are, ordered according to the number of objects entering our list, the Kühr 1Jy sample (Kühr et al. 1981a), the Parkes Survey (Bolton et al. 1979), the B2 catalogue (Grueff & Vigotti 1975), the 3CR sample (Laing, Riley & Longair 1983), the Molonglo Reference Catalogue (Large et al. 1981), the 4C list (Gower et al. 1967), and the S4 (Pauliny-Toth et al. 1978) and S5 catalogues (Kühr et al. 1981b). This gives a total of 465 sources radio detected in well defined samples at different frequencies from large regions of the sky. The reduced sample is still large enough to have a sufficient number of objects in any subclass of radio quasars for a high quality statistical analysis.

© European Southern Observatory (ESO) 1997

Online publication: July 3, 1998
helpdesk@link.springer.de