4 Results and discussion

4.1 Identification results

Out of 34 X-ray sources, thirty-one have been fully identified. The other three proposed optical counterparts to the X-ray sources are too faint to recognize any features from their spectra or even to do spectrometry with a 2 m telescope. The optical counterparts are all the proposed objects except 1RXS J085930.1+745510 whose counterpart is a bright $(E=15.8^{\rm m})$ AGN located southwest the proposed one. Final identification results are provided in Table 1. The identification classification of the counterpart based upon the criteria described in Sect. 3 is listed in Col. 10. For most of the extragalactic objects, redshifts are determined and given in Col. 10. Column 11 comments on main features present in the spectrum that give clues to the spectral classification. The distribution of two hardness ratios for objects in the sample is shown in Fig. 4.

 

Table 1: X-ray identification information

 

Figure 4: Hardness ratios for the objects in the sample

4.2 Selection efficiency for BL Lacertae objects

BL lacertae objects are rare AGN which are characterized by "featureless'' spectra, highly variable, and strong polarization in the radio to optical emission. Its peculiar spectral characteristic prevents it from being discovered in optical domain (Green et al. 1986; Flemming et al. 1991). At first, BL Lacertae objects have been discovered almost by their radio emission (Kühr et al. 1990; Stickel et al. 1991; Stickel et al. 1993; Burbidge & Hewitt 1987). Since last decade the identification of X-ray sources has been proved to be a more efficient way to search for BL Lac objects (EMSS, Stocke et al. 1991; Giommi et al. 1991; Elvis et al. 1992). Because the identification of large X-ray sky survey sample is very time consuming, some kind of pre-selection ought to be made to get a higher BL Lac object fraction. Several teams made use of the fact that BL Lac objects have typical broadband spectral indices in their finding BL Lac objects program (Schachter et al. 1993; Nass et al. 1996). Considerations both on broadband spectral distribution and optical polarimetry lead to fruitful results (Kock et al. 1996).

Seven BL Lacertae objects found in the sample are all newly discovered. The efficiency of the high X-ray-to-optical selection $(\gt 20\%)$ is higher than for most other techniques (Kock et al. 1996) and much higher than the prediction for the potential of RASS in finding BL Lac objects. We have noted that the X-ray information quoted in RASS-BSC can give more constraint on pre-selection. Most BL Lacertae objects are point sources in X-ray and possess HR1 larger than -0.5. If these two restrictions had been supplied at the beginning of sample selection, we would have found 7 BL Lac objects out of 18 X-ray sources. The fraction of BL Lacertae objects would have been doubled.

4.3 Soft X-ray excess of emission line AGNs

ROSAT PSPC worked on a rather soft X-ray band (0.1-2.4 keV) compared with previous X-ray missions. This situation provides a probe into the soft X-ray property of X-ray emitters. Among the ten AGN detected in the sample, nine have rather small HR1 (HR1 < 0), one exception is 1RXS J085930.1+745510 whose counterpart is not the proposed one. By definition small HR1, HR1< 0, is to say that more than $50\%$ of the counts are in soft band (0.1-0.4 keV). Generally, a negative HR1 presents a steeper X-ray spectrum than a positive one (Grupe et al. 1997). The steep X-ray spectra can indicate either the presence of a soft X-ray excess over a flat hard X-ray continuum as a part of a Big Blue Bump (BBB) which extends from the UV spectral range to soft X-ray energies, or a systematic lack of emission around 1 keV and above which may be caused by warm absorbers (Boller et al. 1996). If we take the sample selection criterion into consideration, the former case should be followed.

Supposing intrinsic optical to X-ray indices for AGN are the same, lack of emission at hard X-ray caused by warm absorber or other possible mechanic would reduce total emission in X-ray band. As a result, an AGN with steep X-ray spectrum must be less luminous than those with flat spectra in X-ray band. It is obviously opposite to the identification results of our sample because it is a high X-ray-to-optical flux ratio selected one. We conclude that the soft HR1 of AGN in the sample is an indicator of soft X-ray excess.

4.4 BL Lac object candidates as the counterparts

Three proposed counterparts of X-ray sources are classified as BL Lac object candidates. Their spectra show marked normal galaxy features such as CaII break, "G'' band etc. We did not take them as normal galaxies because their X-ray luminosity largely exceeds the typical value for elliptical galaxies. We use lower conversion factor between the count rate and flux, 10^-11 erg cm^-2 s^-1, which corresponds to a relatively low galactic foreground absorption, to estimate the lower limit X-ray flux of X-ray sources. For an Einstein-de Sitter universe model ($q_0=\frac{1}{2}$), the luminosity distance takes the form: The redshift, $D^{\rm L}$ and estimated X-ray luminosity for the three proposed galaxies are listed in Table 2. The X-ray luminosity of the three objects is largely in excess of the constraint for elliptical galaxies drawn from EMSS (Stocke et al. 1991) even if the difference in the bandpass between Einstein IPC and ROSAT PSPC is concerned. So it is impossible to take normal galaxies as the counterparts of those three X-ray sources.

BL Lacertae objects are very luminous X-ray emitters, a galaxy with an active galactic nucleus can count for such high X-ray luminosity. On the other hand, BL Lacertae object is featureless in optical regime, it can be easily submerged with starlight from the host galaxy. Browne and Marchã have noticed that the possibility of misclassifying X-ray sources as elliptical galaxies or clusters of galaxies is marked (Browne & Marchã 1993). All of the three sources can also find their entries in the NRAO VLA Sky Survey (Condon et al. 1998). Table 2 gives the radio flux from NVSS for the three sources. So long as these facts are concerned, the probability that the three sources are in fact BL Lac objects is very high.

  

Table 2: Evaluated X-ray parameters for three normal galaxies$^ \ddag$

$H_0=50\,$km s^-1 Mpc^-1 is assumed.

Approach to find other plausible counterparts is failed since there is no other objects inside or just outside the error circle in the DSS image. However, the probable counterparts to the three X-ray sources may also be,

(a)

clusters of galaxies. The proposed counterparts are the major cD galaxies of the clusters, other cluster members are too faint to present in DSS images. However, redshift around 0.2 is a quite medium value for clusters of galaxies, clusters with such a redshift should have presented cluster morphology in DSS (Stocke et al. 1991) and extent X-ray emission.

(b)

The proposed optical counterparts are projections by chance, and the real counterparts are too faint to be found in DSS images. Since optical blank field cases are common in X-ray sources identification, projection by chance is an acceptable explanation. The real counterparts to them may be distant clusters of galaxies or BL Lacertae objects.

4.5 Efficiency in finding WD and CVs

The proportions of WDs and CVs are high in the sample. That is another beneficial result of the softness of ROSAT PSPC bandpass: efficient in finding white dwarfs and cataclysmic variable stars. Most white dwarfs release energy mainly at bandpass below 0.4 keV (see Fig. 3), and half numbers of CVs possess soft X-ray excess according to Richman's (1996) investigation. The peculiar X-ray spectrum of these sources makes them difficult to detect in energy bands harder than that of ROSAT.