Annu. Rev. Astron. Astrophys. 1992. 30: 429-456 Copyright © 1992 by Annual Reviews. All rights reserved

4.2 Surveys at Energies Less Than 3 keV

The extension of the Einstein Observatory Medium Sensitivity Survey (EMSS) (Gioia et al 1990b) is currently the major sample of identified X-ray sources at high Galactic latitude. It consists of 835 sources, of which at least 804 have been optically identified (Stocke et al 1991); 427 of this last set are AGN, representing 73% of the extragalactic fraction. The EMSS has limiting sensitivities of 5 x 10^-14 - 3 x 10^-12 erg cm^-2 s^-1 in the 0.3-3.5 keV band, over which ~ 0.1 and 776 deg² of sky have been surveyed, respectively. The source counts from the EMSS extend down to the Einstein Observatory Deep Survey counts (Giacconi et al 1979, Griffiths et al 1987, Primini et al 1991). Analysis of the ROSAT survey carried out in the band 0.1-2.4 keV from July 1990 to January 1991 is expected to produce a sample of at least 50,000 sources from the whole sky. Complete identifications of only 10% or so of this large sample are unlikely to be available for some years.

The mean redshift of the AGN in the EMSS is 0.42. This low value reflects the slower evolution observed in the X-ray luminosity of AGN than in their optical luminosity. Maccacaro et al (1991) have carried out a detailed ``V/V_max'' analysis of this AGN sample and have obtained luminosity functions as a function of redshift. They find that a good fit to the data is obtained with pure luminosity evolution of the form either L(z) = L₀ exp(C) with C = 4.18±0.3 and the look-back time in units of the age of the Universe, or L(z) = L₀(1 + z)^C with C = 2.56±0.17. The luminosity function at z = 0 is modeled as a broken power law with a slope of 1.35 at L₀ < 2.5 x 10⁴³ erg s^-1 and 3.05 at higher luminosities (H₀ = 50 km s^-1 Mpc^-1). Maccacaro et al (1991) then integrate this evolving function over the luminosity and redshift ranges of 10⁴¹-10⁴⁷ erg s^-1 and 0-3 respectively to obtain the expected background intensity. They assume an energy index of 1 for the sources [as obtained from spectral analysis (Maccacaro et al 1988)] and obtain an intensity at 2 keV of 2.23 keV cm^-2 s^-1 sr^-1 keV^-1, which is 37% of the intensity obtained by extrapolating Boldts formula (Equation 1; see Figure 4). They estimate that the uncertainty on this figure is about ± 20% at the 95% confidence level. Note that the shape of the luminosity function means that the result is not very sensitive to the adopted extrema of the luminosity function and that most (85%) of the intensity is due to AGN at luminosities between 10⁴² and 10⁴⁴ erg s^-1.

The slow X-ray evolution of AGN means that a direct determination of the distant quasar luminosity function and background contribution requires deeper observations than the EMSS or the Deep Survey. These have now been obtained by Hasinger et al (1991) and by Shanks et al (1991) using ROSAT. The limiting sensitivity of these images is ~ 9 x 10^-15 erg cm^-2 s^-1 in the 0.1-2.4 keV band. The last group has identified 24 quasars as X-ray sources within a 40-arcmin diameter field, corresponding to about 70 quasars per square degree. The quasars have a similar redshift distribution to that of faint optically-identified quasars. Direct integration of the observed quasar emission (0.5-2 keV) accounts for about 35% of the soft XRB observed in the same image. These data and further, deeper observations in the ``Lockman hole'' (a region of very low Galactic hydrogen density) reviewed by Hasinger (1992) directly show that the integral X-ray source counts flatten below a 0.1-2.4 keV flux level of about 10^-14 erg cm^-2 s^-1 so that the slope of the integral source counts change from a value of about 1.5 to 1.2 or less. This confirms the results from P(D) analyses of both Einstein and ROSAT fields, which required a flattening of the source counts and showed no clear evidence for any further changes down to a level at which there is about 200 sources per square degree.

The good agreement between the above results [and to some extent with the predictions of Schmidt & Green (1986)] shows that the luminosity functions produced by Maccacaro et al (1991) are likely to apply also to the ROSAT data (when modified for the small difference in energy bands). Note that the two fractions of the background intensity obtained thereby should not be added together, since they represent the same objects. How important these fractions are for the 3-300 keV XRB is not yet clear, since what the observations may have done is to clearly resolve the soft excess in the XRR spectrum (Figure 4). The resolved sources do not produce all the XRB, even at soft X-ray energies. We know from these and previous observations that the soft X-ray spectra of quasars and AGN are steep (they must rise in the soft X-ray/EUV band to match their optical spectra). Only if quasar spectra flatten at harder energies (the acceptable power-law spectra now fitted to the observed 0.1-2 keV spectra of quasars over the wide redshift range of 0.1-2.5 argues against this) can they make a significant contribution ( 10%) above 10 keV.

The contribution of observed clusters of galaxies to the soft XRB is less than about 10%. Since they are spatially extended and the EMSS was obtained using an algorithm searching for point sources, there is an uncertain flux correction for this (Gioia et al 1990b, Pesce et al 1990). The negative evolution of clusters discussed earlier argues against a strong contribution to the 1 keV XRB from clusters with z < 0.5.