It has been known all along that definite detection of NT X-ray emission in clusters is a challenging task. This is due to several factors, major among which are its intrinsically weak level - swamped as it is by the intense primary thermal emission - and the difficult to achieve requisite high sensitivity and low detector background in the 10-100 keV band. As has been emphasised, results of the search are not unequivocal, even for Coma and A 2256 for which detection (by RXTE and BeppoSAX) was claimed to be at moderately high level of statistical significance. This is mainly due to source confusion and a complete lack of spatial information. The search should continue (with Suzaku and future satellites), spurred by its sound physical basis - the ubiquity of the CMB, and the observed radio synchrotron emission from relativistic electrons.
The expected level of NT emission of Compton origin depends steeply on the mean magnetic field strength in the central cluster region. Clearly, if this is typically a fraction of a µG, as has been deduced in the analyses described in the previous section, then prospects for more definite detection of NT emission - perhaps already with deep Suzaku measurements - are indeed good. However, significantly higher Bfr values - a few µG - are deduced from FR measurements; had these been meaningful estimates of the volume-averaged field strength, then definite detection of NT emission would have been seriously questioned.
As we have noted already (in Sect. 2.2), the apparent discrepancy between deduced values of Brx and Bfr has been investigated at some length. These two field measures are quite different; the former is essentially a volume average of the relativistic electron density and (roughly) the square of the field, the latter is an average along the line of sight of the product of the field and gas density. All these quantities vary considerably across the cluster; in addition, the field is very likely tangled, with a wide range of coherence scales which can only be roughly estimated. These make the determination of the field by both methods considerably uncertain. Thus, the unsatisfactory observational status (stemming mainly from lack of spatial information) and the intrinsic difference between Brx and Bfr, do not allow a simple comparison of these quantities. Even if the large observational and systematic uncertainties, the different spatial dependences of the fields, relativistic electron density, and thermal electron density, already imply that Brx and Bfr will in general be quite different. This was specifically shown by Goldshmidt & Rephaeli (1993) in the context of reasonable assumptions for the field morphology, and the known range of IC gas density profiles. They concluded that Brx is indeed generally smaller than Bfr. The implication is that prospects for detection of NT emission should not be based on the relatively high deduced values of Bfr.
In conclusion, RXTE and BeppoSAX measurements have yielded appreciable evidence for power-law X-ray emission in four clusters. These results motivate further measurements and theoretical studies of NT phenomena on cluster and cosmological scales.
The authors thank ISSI (Bern) for support of the team "Non-virialized X-ray components in clusters of galaxies".