The REX survey: A search for radio-emitting X-ray sources

We present the scientific goals, strategy, and first results of the REX project, an effort aimed at creating a sizable and statistically complete sample of radio-emitting X-ray sources (REXs) using the available data from a VLA survey (NVSS) and the ROSAT PSPC archive. Through a positional cross-correlation of the two data sets, we have derived a sample of about 1600 REXs. Among the 393 REXs identified so far (either from literature or from our own spectroscopic observations) a high fraction are active galactic nuclei (AGNs; about 60%-80%), typically radio-loud QSOs, and BL Lac objects. The remaining sources are galaxies, typically radio galaxies isolated or in clusters. Thanks to the low flux limits in the radio (5 mJy at 1.4 GHz) and in the X-ray band (similar to 5 x 10(-14) ergs s(-1) cm(-2), 0.5-2.0 keV) and the large area of sky covered by the survey (2183 deg(2)), we intend to derive a new complete and unbiased sample of BL Lac objects that will contain both radio-selected (RBL) and X-ray-selected (XBL) objects. In this way, the apparent dichotomy resulting from the current samples of BL Lac objects can be directly analyzed in a unique sample. Moreover, the high number of BL Lac objects expected in the REX sample (similar to 200) will allow an accurate estimate of their statistical properties, such as the X-ray, radio, and optical luminosity functions and the cosmological evolution. For these reasons, the REX sample will be a powerful tool to accurately test the current theoretical models proposed for BL Lac objects. To date, we have discovered 15 new BL Lac objects and 11 BL Lac candidates with optical properties intermediate between those of a typical elliptical galaxy and those of a typical BL Lac object. These objects could harbor weak sources of nonthermal continuum in their nuclei, and if confirmed they could represent the faint tail of the BL Lac population. The existence of such "weak" BL Lac objects has been matter of discussion in recent literature and could lead to a reassessment of the defining criteria of a BL Lac object, and consequently to a revision of their cosmological and statistical properties. Finally, the sample of similar to 800 emission-line AGNs generated by the REX survey will be useful in addressing many of the open questions regarding AGN phenomenology, such as the relationship between radio-loud and radio-quiet AGNs.