Measuring the CXB is extremely challenging with currently flying instruments as none of them has been directly designed to make such measurement. The main difficulty here is the capability, by design or analysis, to discriminate the CXB among the other, sometimes dominating, instrumental background components. In the case of INTEGRAL and Swift, it is even worst as their high-energy detectors (IBIS and BAT [20, 21]) are coded-mask detectors expressly designed for the study of point-like sources. Measurements of the CXB have thus to rely on indirect techniques as the one based on the modulation of the CXB signal by the passage of the earth in the field of view (FOV). All the recent measurements performed by INTEGRAL, BeppoSAX and Swift/BAT [22, 18, 23] use the earth occultation technique. These measurements, see Fig. 1, are more consistent with the original HEAO1 CXB spectrum rather than with its renormalization by ~ 1.3 (see discussion in the previous section). The INTEGRAL and Swift/BAT CXB spectra lie ~ 10% above the HEAO1 spectrum. Unfortunately, the albedo emission from the earth, which must be taken into account during the occultation analysis, does not allow to make a strong test of the shape of the CXB spectrum. This test becomes only possible when different measurements and instruments are used. Taking into account all the newest measurements and neglecting the broadband measurement of HEAO1 it is possible to show that a simple description of the CXB spectrum is achieved using a smoothly joined double power-law function . From this analysis, it appears that the CXB spectrum, as measured most recently is ~ 30% larger than the HEAO1 spectrum below 10 keV while only 10% larger than it above this energy. A new measurement performed in the 1-7 keV band with the X-ray telescope (XRT), on board Swift, confirms this picture .
Figure 1. Comparison of measurements of the peak of the Cosmic X-ray background (adapted from ).