2 Sample and observations

2.1 The sample

The RASS was conducted from August 1990 until January 1991, with supplemental observations taken in February and August 1991. The 2 PSPCs on board ROSAT were the instruments in use for the entire RASS. Descriptions of the observing methods employed during the survey, as well as the data reduction techniques, can be found in Cruddace et al. (1991).

The raw PSPC survey data were processed automatically using the ROSAT Standard Analysis Software System (SASS). The SASS time-tagged, gain-corrected, and applied an attitude solution to each detected X-ray photon. During the first processing of the RASS data, the SASS was also used to compile a first-cut catalogue of RASS sources (Voges 1992). A maximum likelihood (ML) algorithm was run automatically on the RASS data to find all X-ray sources above a likelihood threshold of 11. The source detection algorithms were run on images of different energy bands so that the hardness ratios could be computed for all sources.

In order to calibrate various methods for identifying the contents of the RASS, it was decided to identify all RASS sources in four fields, covering a 575 square degree area of the sky, using optical spectroscopy in a manner similar to the EMSS (Stocke et al. 1991). Telescope time for this project was granted under a Key program at ESO and four study areas were chosen at high Galactic latitude on the southern sky (Danziger et al. 1990).

In order to identify the optical counterparts of the X-ray sources, low-resolution (7 Å and 16 Å) spectra of the objects inside the error circle of the X-ray position were taken, using the Boller&Chivens spectrograph at the ESO 1.5 m telescope and the EFOSC1 in spectrographic mode at the 3.6 m telescope at La Silla. Spectra of bright stars were acquired even if the objects were situated outside of the error circle. The spectra were used to perform a first spectral classification.

For some of the F, G and K stars, intermediate (2 Å) resolution spectra of the H$\alpha$ line were obtained in order to confirm the likely coronal source, through the presence of enhanced chromospheric activity.

Out of the complete sample of RASS sources, the "solar'' type (A to K) stars were selected and are presented in this work. A total of 107 A-K stars were found, out of $\sim$600 sources.

2.2 The photometric observations

The photometric observations were carried out at the European Southern Observatory at La Silla, using the 0.5 m ESO telescope, equipped with a single-channel photon-counting photomultiplier, and the standard ESO $BV(RI)_{\rm c}$ filters. The available U filter was not used, due to the fact that our sample stars are red and for the most rather faint, so that the low counts in the U filter would have led to inaccurate data. Two photomultipliers were used, a red-sensitive dry-ice cooled EMI 9658 and a Peltier-cooled Hamamatsu GaAs (see Table 1).

  

Table 1: Observing periods

Each measurement consisted normally of three 10 s integrations in each filter. For the faintest stars, longer integration times of 20-30 s were used. The transformation coefficients were obtained by observing Cousin's E-region standard stars. Atmospheric extinction coefficients were obtained by observing two E-region standards of very different spectral types at different airmasses during the night.

The errors are of the order 0.01 mag in all filters for stars brighter than V=12. For stars fainter than V=12, the errors are higher, of the order 0.03 in the $V,\ R$, and I filters and 0.04 in the B filter. The results of the observations are listed in Table 3.

Stars brighter than V=5.5 (four of our sources) were not observed, as they were too bright for either photomultiplier tube used. For these stars, we have taken the spectral classification, as well as the photometric data, from the literature. These data are also given in Table 3.

2.3 Spectral classification

The photometric data were used to determine the spectral type and luminosity class of our sample objects, as a complement to the classification already done using the low-resolution spectra available from the identification program.

The photometric spectral classification was done using the colour tables computed by Cutispoto et al. (1996). These tables contain mean values of the colour indices for stars in the spectral range A to K. For each of our sources, a combination of spectral type and luminosity class were selected, by comparing the data with the synthetic colours. The results are summarized in Table 3.

In case that no single star would fit the observations, the object was considered a binary. The observed colours were then compared with the results of the combinations of two single stars, and the combination that best reproduced the data was selected. The results were checked with the help of high-resolution spectra, also obtained in the course of our program (see Metanomski et al. 1997a,b). In most cases, the spectra confirmed the binarity of the source.

We would like to note that this procedure assumes unreddened stars with normal colours, a reasonable assumption, considering that all our stars are found to be close to the Sun. Distant stars, PMS objects and stars with a high activity level could have significantly altered colours.

In Table 3 are also listed the spectral classifications we had either from the literature and SIMBAD database, or from the low-resolution spectra taken in the course of the identification program. In most cases the spectral classification obtained from the colour indices agrees very well with the available spectral classification. The colour indices also gave binary classification in all cases where an object from the sample was known either from the literature or from spectroscopic observations to be a binary. This method cannot, however, give a correct result if the components of a binary system are very similar. We have in some cases used the information taken from our high resolution spectroscopic observations, if an SB2 system was clearly present. In Table 3 visual, photometric and spectroscopic binaries are separately labeled.

2.4 X-ray fluxes, distances and X-ray luminosities.

To determine the X-ray fluxes from the ROSAT countrates, a conversion factor was adopted from Fleming et al. (1995), appropriate for coronal sources:

$$CF = (8.31 + (5.30 \cdot {\rm HR}1)) \ 10^{-12}\ {\rm erg} \ {\rm cm}^{-2} \ {\rm count}^{-1}.$$ (1)

The hardness ratio HR1 is defined as:

$${\rm HR}1 = \frac{CR \left[ch 52-201 \right] - CR \left[ch 11-41 \right]}{CR \left[ch 11-240 \right]}$$

$CR \left[ch a-b \right]$ being the countrate in the channels "a'' to "b''.

Since most of the stars in our sample are closer than 200 pc, extinction can be considered negligible, allowing the use of this relation.

The error for the X-ray flux is determined by both the errors on the countrate and the hardness ratio. Depending on the source's intensity, the uncertainty in the X-ray flux can be as high as a factor 2, due mainly to the error in the HR1.

For five of the sources, the X-ray counts obtained were so low, that the objects could only be seen when summing up all channels. For these objects there is no value for the HR1. For these stars we used the mean value of the hardness ratio to determine the X-ray flux, as was done by Fleming et al. (1995).

The distances were calculated by determining the absolute visual magnitude of the objects, using the observed (B-V) colour indices for single stars, and the spectral types for binaries. The values of M_V as a function of (B-V) were taken from Gliese (1982) or from the lists of Cutispoto et al. (1996), for the early stars and giants/subgiants.

The error in the distance is mainly determined by the error in the (B-V) index for the single stars. This error varies from 0.02 to 0.07 for the faintest objects, leading to an error of 5 to 23% in the distance. For the binaries, an error of one spectral class per component was assumed, leading to an error of up to 30%.

Also to be considered is that for the distance estimation we have used only the photometric spectral classification. This method, as already pointed out in the previous section, assumes "normal'' stars with unreddened colours, and does not allow the identification of PMS objects, nor the correct classification of highly active stars. As a PMS nature or the activity level will only be determined with the help of high-resolution spectroscopic observations, the distances for some of our objects may need to be revised at a later time. We would like to note, however, that this method is consistent with the one used for the EMSS and EXOSAT surveys.