Astron. Astrophys. 356, 445-462 (2000)

1. Introduction

Active Galactic Nuclei (AGN) exhibit a wide variety of observable properties resulting in a complex classification scheme. Current unification schemes attribute many of the observational differences to different viewing angles of the sources' anisotropic emission (e.g., Antonucci 1993, Orr & Browne 1982, Barthel 1989, Urry & Padovani 1995). Large, unbiased samples of AGN are powerful tools for understanding the different physical conditions in the cores of these objects and to test unification schemes. All classes of AGN appear in X-ray surveys. Samples based on the ROSAT All-Sky Survey (RASS), the first soft X-ray survey of the whole sky using an imaging X-ray telescope (Trümper 1983), are well suited for investigating the unification of various AGN classes. There are more than 80.000 RASS sources, of which some 60% are expected to be AGN (Voges et al. 1999), ensuring a statistically useful number of objects belonging to the various AGN subclasses. Unfortunately, the large majority of the ROSAT sources remain to be optically identified.

One of the largest obstacles to identifying RASS sources is that the positional accuracy is limited to , making identification of optical counterparts problematic. Correlation with large-scale radio surveys can alleviate this problem, but at the expense of preferentially selecting radio-loud AGN; however, deep radio surveys are capable of detecting radio-quiet AGN, since `radio-quiet' does not necessarily mean radio-silent. Formally, the radio-loudness is given by the parameter, R, where (see Stocke et al. 1992, for a more precise definition) and defines the radio-loud/-quiet boundary. Radio-loud AGN are generally more active; i.e., they exhibit larger scale and more frequent variability and are also more luminous at X-ray and ray energies for a given optical luminosity. This last characteristic implies the selection effect introduced by correlation of a flux-limited X-ray survey with a radio catalogue is less dire than is initially apparent.

Estimates place the fraction of radio-loud AGN at 10%, but sources do not have to be optically faint to reach the radio-loud/-quiet boundary, e.g., a magnitude object will be classified as radio-loud only if its radio flux is greater than 1 mJy - a value well below previous radio survey flux limits. This casts doubt on our estimate of the fraction of radio-loud vs. -quiet objects in addition to our understanding of the sharpness of the transition between radio-loud and -quiet source populations since only a handful of large unbiased quasar samples have had deep follow-up radio observations.

Previous cross-correlations of the RASS source list (Voges 1992) with radio catalogues yielded large samples of radio-loud extragalactic X-ray sources (Brinkmann et al. 1994, Brinkmann et al. 1995, B95). The study of the correlation with the 5 GHz Green Bank survey (Condon et al. 1989) in particular demonstrated that the majority of the new X-ray - radio sources have broad-band properties between those of traditional radio- and X-ray-selected AGN. Contrary to previous results (e.g., Kellermann et al. 1989), this ROSAT - Green Bank (RGB, Brinkmann et al. 1997b, Laurent-Muehleisen et al. 1997) study also found no evidence for a bimodal radio-loudness distribution, indicating that the traditional division between radio-quiet and radio-loud AGN may not be warranted (Brinkmann et al. 1997b).

The RGB sources' X-ray spectra, optical colors and morphologies, and spectral energy distributions indicate that most of the spectroscopically unclassified sources are quasars, although some BL Lacs, radio galaxies, Seyferts and clusters are undoubtedly present. However, the broad-band properties of these unclassified sources differ noticeably from those of brighter radio-loud AGN such as those from the 1 Jy catalogue (Kühr et al. 1979). Specifically, the broad-band properties are intermediate between those of traditional radio- and X-ray-selected AGN (Laurent-Muehleisen et al. 1998). Deeper X-ray and radio surveys thus offer the possibility of revealing a substantially different AGN population with radio properties intermediate between the traditional radio-loud AGN (e.g., PKS and 1 Jy samples) and radio-quiet AGN.

We present the broad-band properties of objects from a correlation of the RASS source list with the FIRST (Faint Images of the Sky at Twenty centimeters) VLA ² radio survey (Becker et al. 1995, White et al. 1997). The FIRST survey's combination of high sensitivity (1 mJy at 1.4 GHz) and positional accuracy ( 1") makes it possible to unambiguously identify optical counterparts for sources several times fainter than was previously feasible with any other large area radio survey. In Fig. 1 we plot the 5 GHz luminosities of all spectroscopically classified objects in the RGB sample as a function of their redshifts. The solid line represents the K-corrected luminosity limit of sources detectable in the FIRST 1.4 GHz survey. This figure clearly demonstrates the potential to study objects over a much wider range of radio luminosities than was previously possible.

Fig. 1. 5 GHz radio luminosities of spectroscopically classified sources from the RGB sample (diamonds, flux limit mJy) and the luminosity limits of objects at different redshifts for the FIRST survey (straight line, flux limit mJy).

The good sensitivity of the FIRST survey also enlarges the available broad-band parameter space, providing the opportunity to adequately sample the spectral energy distributions of a wide range of radio-quiet, -loud, and -intermediate objects. Previous X-ray-radio catalogues are biased towards sources with high fluxes in at least one of the observing bands, (cf. B95). The current sample reaches fainter fluxes and does so with a nearly uniform coverage over a large part of the sky, shifting phase-space boundaries into unexplored regions and possibly discovering new source populations. Finally, the current sample provides a stringent test of the accuracy of previous claims of correlations between the luminosities in different energy bands (Avni & Tananbaum 1986, Kembhavi et al. 1986, Worrall et al. 1987, Wilkes et al. 1994, Brinkmann et al. 1994, 1995 (B95), B97).

In the next section we overview the content and the properties of the catalogues, the selection criteria, and also present the source catalogue with the relevant X-ray, optical, and radio information. We also discuss and present the results of new spectroscopic data. In Sect. 3 we discuss the general radio through X-ray properties of the objects and compare the results to those obtained for the previously known objects in this and other samples. Sect. 4 is devoted to the broad-band study of the most prominent object classes. Throughout the paper we use a Friedman cosmology with km s^-1Mpc^-1 and q₀ = 0.5 for the computation of the K-corrected luminosities.

© European Southern Observatory (ESO) 2000

Online publication: April 10, 2000
helpdesk.link@springer.de