Astron. Astrophys. 353, 25-40 (2000)

1. Introduction

The AGN/QSO luminosity function and its evolution with cosmic time are key observational quantities on understanding the origin of and accretion history onto supermassive balckholes, which are now believed to occupy the centers of most galaxies. Since X-ray emission is one of the prominent characters of the AGN activity, X-ray surveys are effective means of sampling AGNs for the luminosity function and evolution studies. The Röntgen satellite (ROSAT ), with its unprecedented imaging capabilities, provided us with soft X-ray surveys with various depths, ranging from the ROSAT All-Sky Survey (RASS) to the ROSAT Deep Survey (RDS) on the Lockman Hole (Hasinger et al. 1998). Various optical identification programs of the survey fields have been conducted and the combination of these now enabled us to construct the soft X-ray luminosity function (SXLF) as a function of redshift.

The evolution of SXLF has already been seen in the Extended Medium Sensitivity Survey (EMSS) AGNs (Maccacaro et al. 1991; Della Ceca et al. 1992) for high-luminosity AGNs. Combining results from deep ROSAT PSPC surveys and the EMSS has extended the sample into the higher-redshift lower-luminosity regime, providing much wider baseline to explore the evolution properties (e.g. Boyle et al. 1994; Jones et al. 1996; Page et al. 1996). All of these were characterized by a pure luminosity evolution model (PLE) with approximately up to , and consistent with no evolution beyond that point. Using a larger ROSAT sample, Page et al. (1997) found that PLE underpredicts the number of high-redhsift low luminosity AGNs for . Simple extrapolations of any of the PLE expressions only explain of the soft X-ray Background (0.5-2 keV) by AGN.

Because of the relatively large PSF of the ROSAT PSPC, the identifications of the deepest ROSAT PSPC surveys are sometimes ambiguous and misidentifications can occur. Based on results of the optical followup studies of ROSAT PSPC surveys, a number of groups, including the Deep ROSAT Survey (DRS; Griffiths et al. 1996) and UK Deep Survey (UKD; McHardy et al. 1998) report a population of X-ray sources called "Narrow Emission-line Galaxies" (NELG) at faint fluxes. On the other hand, faint X-ray sources found in the ROSAT Deep Survey on the Lockman Hole (RDS-LH), which have accurate source positions from 1 million seconds of ROSAT HRI data, are still predominantly AGNs down to the faintest fluxe in the survey (Schmidt et al. 1998; Hasinger et al. 1999). Some of these have optical spectra which apparently show only narrow-lines but have other signs of an AGN activity and might have been classified as "NELGs" at the criteria of other groups. On the other hand, Lehmann et al. (1999b) have compared redshift distributions of the RDS X-ray AGNs, UKD X-ray sources, non X-ray emitting (at the RDS-LH limit) field galaxies showing narrow emission-lines. They found that the redshift distribution of UKD X-ray sources has a significant excess over that of the RDS-LH sources at . This excess was dominated by "NELGs", whose redshift distribution was similar to that of non X-ray source narrow emission-line field galaxies. This shows that a significant fraction of "NELGs" are likely to be misidentifications by chance coincidences. This observation seems to contradict with estimations of the relatively low probabilities of such chance coincidences by the DRS and UKD groups. A more detailed comparison is urgently needed. Misidentifications affect SXLF estimates in two ways, i.e., by putting a wrong object into the sample and by missing the true identifications. Thus it is important to have a high spatial resolution image to obtain unambiguous identifications, especially in the faintest regime.

In this study, we investigate the global behavior of the soft X-ray luminosity function (SXLF) of AGNs from a combined sample of various ROSAT surveys. We use the term "AGN" for both Seyfert galaxies, including type 1's and type 2's, and QSOs. Preliminary work, using earlier versions of the combined sample, have been reported in Hasinger (1998) and Miyaji et al. (1999a) (hereafter M99a), while in this work, we have made a more extensive analysis with updated ROSAT Bright Survey (RBS) and ROSAT Deep Survey (RDS) catalogs including new identifcations from observations made in the winter-spring season of 1999. In this paper, we put emphasis on the expressions representing the global behavior of the SXLF. Presenting separate expressions in several redshift intervals, giving more accurate representation of the data in the redshift ranges of interest will be a topic of a future paper (Miyaji et al. in preparation, paper II). In paper II, we will also present tables of full numerical values of the binned SXLF.

We use a Hubble constant . The dependences are explicitly stated. We calculate the results with common sets of cosmological parameters: (1.0,0.0) and (0.3,0.0). For some important parameterized expressions, we also show the results for (0.3,0.7).

© European Southern Observatory (ESO) 2000

Online publication: December 8, 1999
helpdesk.link@springer.de