3 Identification of radio-loud RASS NEP sources

3.1 RASS-VLA correlation

A list of radio-loud RASS sources was constructed by finding all unresolved RASS and VLA sources with offsets $\leq$25$^{\prime\prime}$($\leq$40$^{\prime\prime}$) in the central (outer) 3$^\circ \times$ 3$^\circ$ fields. For sources resolved in either band, a larger offset of 60$^{\prime\prime}$(100$^{\prime\prime}$) in the central (outer) 3$^\circ \times$ 3$^\circ$ fields was permitted to capture extended X-ray clusters containing a radio galaxy or extended radio sources associated with an X-ray AGN. A total of 74 reliable matches were found. Table 1 gives the resulting RASS-VLA unresolved sources (28 matches) and Table 2 contains RASS-VLA sources where either the radio or X-ray source is extended or multiple (46 matches representing at least 37 separate radio sources).

The reliability of the associations was evaluated by repeatedly recorrelating the VLA catalog with X-ray source positions randomly shifted from the RASS positions by a range of angular distances, typically a few arc minutes. The simulations indicate that Tables 1 and 2 have high reliability with % spurious RASS-VLA associations. Table 3 contains possible additional sources (34 in number) with greater angular separations. Here the chance of spurious associations is around 50% in the central field and 20% in the outer fields. Most of the spurious listings will have faint (1-2 mJy) radio flux densities.

The radio and X-ray data are summarized in the first 11 columns of the tables. Columns 1-5 are extracted from Kollgaard et al.(1994) and give the VLA source name, peak position in J2000, integrated flux density at 1.5 GHz in mJy with its error. Secondary radio peaks which were individually cataloged in Kollgaard et al.are also listed. Column 6 gives the name of the corresponding ROSAT Survey source and in Col. 7 we list the positional offsets between the VLA and the X-ray sources in arcsec. Column 8 indicates whether the source lies in the inner 3$^\circ$ (i) or outer region (o) of the RASS NEP field. Column 9 gives the likelihood value from the EXSAS processing. Columns 10-11 present the 0.1-2.4 keV X-ray flux (with error) in units of 10^-14 erg cm^-2 s^-1, assuming a power law spectrum with photon index $\Gamma = 2.1$ and Galactic absorption.

3.2 Optical counterparts

The RASS-VLA matches were correlated with optical catalogs produced from densitometric scans of blue plates from the Palomar Observatory Sky Surveys (POSS). As indicated in Sect. 2.2, two systems are used: APM scans of the POSS-I O plates and COSMOS scans of the POSS-II J plates. Optical positions within 4$^{\prime\prime}$ of VLA radio positions were considered likely matches. The resulting objects are listed in Cols. 12-15 of Tables 1-3 with optical morphological class, optical minus radio positional offsets in arcsec, and estimated B magnitudes. A few objects with offsets up to 8$^{\prime\prime}$ are included when the identification is indicated in the literature or the radio source is extended. Additional nearby optical objects are given in the table notes with offset $(\Delta\alpha,\Delta\delta)$ in arcsec. As the radio sources in Table 2 are often extended, the absence of a listed optical counterpart is not a reliable indication that a counterpart is not present.

We found significant discrepancies, up to 3 magnitudes, between APM and COSMOS blue magnitudes and, where available, between both systems and published magnitudes. These differences are much greater than expected from the changes in emulsion-filter combinations used for POSS-I and II. We constructed a rough magnitude system by calibrating COSMOS measurements to APM magnitudes for stellar sources within 1$^\prime$ of RASS-VLA sources. A linear regression between COSMOS MCOR values and $B\rm _{APM}$ gives MCOR = $(0.56\ \pm\ 0.02) \times B_{\rm APM} + (-13.3 \ \pm\ 0.3)$. This formula was used to convert COSMOS measurements into magnitudes when an object was found on both plates. Values are rounded to the nearest 0.5 magnitude. However, many objects are found only in one of the APM or COSMOS catalogs, and a cross-calibration could not be performed in these cases. We caution that magnitudes reported in the tables are based on heterogeneous magnitude systems, and may suffer systematic uncertainties around $\Delta B \sim \pm$1. When a source was not detected by either APM or COSMOS blue plates, a limiting magnitude of B > 22.5 from the deeper COSMOS plates is adopted.

In Tables 1-3, Col. 12 presents an optical morphology classification derived in most cases from the COSMOS object catalog: E = empty field; F = too faint to classify; G = galaxy; Gp = galaxy group; S = stellar. The Gp classification was made by us as described in Sect. 4.2. Columns 13-14 give the optical minus VLA positional offsets in arcsec. Column 15 gives approximate B magnitudes derived as follows: a = APM object catalog; b = COSMOS object catalog calibrated to APM magnitudes; c = COSMOS object catalog. The three objects with an "a" code occurred in the region where no blue POSS II plate was available (one case) or were not detected on the blue POSS II plate but were detected on the red plate and/or POSS I. Magnitudes without codes were obtained from the literature cited in the table notes. Table notes give cross-identifications and associated data obtained from the NASA Extragalactic Database (NED) or from Véron-Cetty & Véron (1996), and indicate whether radio or X-ray extended structure is present. Table notes also list nearby optical objects (from COSMOS or APM catalogs) with magnitudes and $(\Delta\alpha,\Delta\delta)$ in arcsec.