|Annu. Rev. Astron. Astrophys. 1992. 30:
Copyright © 1984 by Annual Reviews. All rights reserved
The energy density of the X-ray sky is dominated by a diffuse radiation which is mostly of cosmic origin: the X-ray background (XRB). It was the first cosmic background radiation discovered and was first clearly detected during a rocket flight intended to study X-rays from the Moon. The observers instead found the first extra-solar X-ray source (Sco X-1) and the XRB (Giacconi et al 1962). The Moon has only recently been detected as an X-ray source by the ROSAT X-ray telescope (Figure 1). The bright side of the Moon is observed to reflect Solar X-rays whereas the dark side is clearly darker than the surrounding sky emission, which is from the XRB.
Despite the enormous observational progress that modern X-ray detectors and telescopes have brought to the study of the XRB in the intervening 30 years, its origin is still unknown. The detection of many individual X-ray sources shows that some fraction of the XRB must be due to unresolved sources, but so far no class of known sources has clearly been shown to contribute more than 50% of the total intensity. In other words, even with the most recent X-ray images, there remains an unresolved and unexplained component of the XRB, amounting to at least half its intensity. This component does not necessarily have to be a truly diffuse component in the XRB, only that not enough angular resolution and sensitivity are yet available to identify all the sources of the XRB.
Figure 1. ROSAT X-ray image of the moon in the 0.1-2 keV band (from Schmitt et al 1991). Note the decrease in the XRB in the dark side of the Moon.
The isotropy of the 3 keV XRB strongly suggests that it is extragalactic. A weak Galactic anisotropy is observed in the 2-10 keV band (Warwick et al 1980, Iwan et al 1982), but otherwise the X-ray sky is isotropic to within a few per cent on large scales. Figure 2 shows all-sky maps of the XRB in the energy band 2-10 keV obtained by the HEAO-1 A2 experiment where both the Galactic anisotropy and the high-Galactic latitude isotropy are evident. Much weaker than the Galactic contamination, a dipole radiation field is observed on the largest scale. This is consistent with the Compton-Getting effect due to the motion of our Galaxy in the direction measured from the dipole in the microwave background (MWB) (Shafer 1983, Shafer & Fabian 1983, Boldt 1987). This provides further evidence for the distant extragalactic origin of the 3 keV XRB. There is a strong distinction between the XRB which is reviewed here and the soft XRB (energies key), the extragalactic contribution of which is uncertain but which is less than 10% of the observed intensity below about 1/4 keV (see e.g. McCammon & Sanders 1990). The uncertainty is mainly due to a strong Galactic component dominating the soft XRB below about 0.5 keV and to the attenuation of extragalactic X-rays by photoelectric absorption in our Galaxy's interstellar medium at lower energies.
Figure 2. (Top) The X-ray sky from HEAO-1 A2 in Galactic coordinates at 3 resolution. Bright sources and diffuse emission from the Galaxy dominate at low Galactic latitudes. (Bottom) The same map with the contrast enhanced in order to emphasize fluctuations in the XRB at high Galactic latitude. In particular a cold spot is seen near the south Galactic pole. Both maps have been kindly provided by K. Jahoda.
A point that must be emphasized is that the XRB is detected over a wide energy band from soft X-ray energies of about 0.1 keV to gamma rays above an MeV. This corresponds to over four decades of frequency or wavelength and so is at least an order of magnitude wider than the infrared/optical/UV band stretching from 100 µ to 1000 Å. The large bandwidth should be borne in mind whenever considering models for the XRB based on data from only a small part of the whole XRB band. The spectrum of an object detected at say 100 µ needs to be known with great accuracy before confident predictions can be made about its contribution in the UV. Similarly, we cannot yet be sure of the contribution of quasars detected below 1 keV to the XRB at 40 keV. This example is especially relevant here, since current imaging X-ray telescopes operate at energies 3 keV where quasars are a prominent component of source counts, but most of the energy density in the XRB is concentrated at ~ 20-40 keV. This situation will change soon with the launch of telescopes capable of imaging 10 keV X rays. In Figure 3 we plot I (where a horizontal line represents equal energy per decade) from radio to rays. Apart from the very bright MWB and the highly uncertain optical and UV backgrounds, the extragalactic sky is dominated by the XRB at ~ 40 keV.
Figure 3. Spectrum of the extragalactic sky. I from the radio band to -rays. Arrows denote upper limits. The microwave and submillimeter background spectrum is from Mather et al (1990), the optical from the review by Mattila (1990), the ultraviolet from Paresce (1990), Bowyer 1991), and Martin et at (1991), and the X-ray spectra are from (1/4 keV limit) McCammon & Sanders (1990), (dots) Wu et at (1991), (smooth curve from Equations 2 and 3) Marshall et al (1980), Gruber et al (1984) and (final dot) Fichtel et al (1978). The limit at 912 Å (3.3 x 1015 Hz) is inferred from sharp H I cutoffs to nearby spiral galaxies (see discussion by Kenney 1990).
No features have yet been detected in the spectrum or appearance of the unresolved XRB with which the emission redshift can be unambiguously determined. Several models for its overall spectral shape (as discussed in Section 5) suggest that the dominant redshift z is about 2. Unless the photons now observed as X rays originated as rays, with consequent energy problems, it is most unlikely that the XRB originates from beyond the redshift range 0 < z < 10. It therefore represents the integrated X-ray emission from all objects and processes since the Universe was about a billion years old (including unresolved nearby objects). Its intensity and spectrum measure the total X-ray power of the Universe since then, and its anisotropies reflect the spatial distribution of the sources (Rees 1979, Bagoly et al 1988, Barcons & Fabian 1988, Meszaroz & Meszaros 1988). This is of importance to studies of large-scale structure of the Universe and galaxy formation, since, unlike the MWB, the XRB can probe epochs in which the large-scale structural inhomogeneities are being formed.
After a brief review of the main observational features (spectrum and isotropy) of the XRB, we discuss the contribution of known classes of discrete sources. Models are then presented for the residual background intensity, with the problems arising from a diffuse origin emphasized. We then make the case for some new classes of sources, or rather old classes with new faces, dominating the XRB. Finally, the importance of the isotropy of the XRB for the study of the large-scale structure of the Universe is briefly discussed.