The Einstein Observatory Extended Medium-Sensitivity Survey. II. The Optical Identifications

The optical identifications of the Einstein Extended Medium Sensitivity Survey (EMSS) are presented. At present over 96% of the 835 X-ray sources in the survey have been successfully identified in the following proportions: active galactic nuclei (QSOs, quasars, and Seyferts), 51.1%; BL Lacertae objects, 4.3%; clusters of galaxies, 12.2%; normal galaxies, 2.1%; cooling flow galaxies, 0.6%; Galactic stars, 25.8%; and unidentified, 3.9%. All of the EMSS sources are identified with previously known classes of X-ray emitters except for five sources now identified with cooling flow galaxies whose novel attributes are described. The Galactic stars are primarily normal dwarf stars but also include RS CVn and W UMa binary systems, two FK Comae stars, five cataclysmic variables including new examples of AM Herculis stars, one white dwarf, and one X-ray binary candidate. The extragalactic sources are predominantly AGNs, which range from low-z, low-luminosity (z < 0.1; L_x_ < 10^43^ ergs s^-1^) Seyfert galaxies to high-z QSOs (z > 2) but also include distant clusters of galaxies (0.05 < z < 0.6), BL Lac objects (0.15 < z < 0.65) and nearby, normal galaxies (z <= 0.1). Most of the individual optical counterparts are previously unknown objects and so constitute large statistical samples independent of previous selection methods. Confusion levels due to chance projection onto the X-ray error circles and to spatially unresolved close pairs of X-ray sources are estimated to be quite small (~20 sources), consistent with the numbers of such sources identified within the EMSS already. The small confusion levels within the EMSS validate the identification procedure in which spectra are obtained within and adjacent to each error circle until a plausible counterpart is discovered. The criteria used to determine whether the proposed optical counterpart to the X-ray source is a plausible identification are described in detail. Plausibility is based upon the optical classification of the counterpart (e.g., AGN, cluster, G star) and the X- ray-to-optical flux ratios previously observed for these classes of X-ray emitters. Two independent schemes of optical classification of the counterparts are used to check the plausibility of these identifications; one scheme based upon moderate-resolution optical spectroscopy (7A resolution in the range 3400-6400 A) and R-band CCD direct imaging, and the other based upon the inferred X-ray luminosity and the overall (radio-optical-X-ray) energy distribution. Counterparts whose optical classification is somewhat ambiguous are a small fraction of the total and include sources which are either (a) low- luminosity AGNs or high-luminosity normal spiral galaxies; (b) low- luminosity clusters of galaxies or high-luminosity elliptical galaxies; (c) AGNs with unusual spectra lacking convincing evidence of broad emission lines; (d) AGNs with extremely weak emission lines (W_lambda_ <= 30 A); and (e) BL Lac objects with modest starlight contributions. Overall, despite X-ray error circles whose 90% confidence radii are ~50" and despite the extremely diverse nature of the optical counterparts, we conclude that the optical identification of faint X-ray sources within the parameter range of the EMSS (f_x_ >= 8 x 10^-14^ ergs s^-1^ cm^-2^ in the [0.3-3.5] keV band); |b| >= 20^deg^) is possible at an extremely high confidence level. Thus this sample can be used for detailed statistical investigations (e.g., luminosity functions, determination of cosmological evolution and X-ray background contribution). Our results also suggest that, although time-consuming and difficult, the complete identification of high Galactic latitude ROSAT all-sky survey sources is possible with present ground-based telescopes and instrumentation.