We have shown the first log(N)-log(S) relations based on the XMM-Newton observatory. Given the still existing systematic uncertainties, the data is fully consistent with the ROSAT and Chandra source counts in the 0.5-2 keV band. This demonstrates on one hand that the combination of EPIC detectors is not yet confusion limited in a 100 ksec observation, on the other hand that cosmic variance between different fields does not affect the source counts significantly at the currently achieved flux levels, at least not in the soft band. In the 2-10 keV band there is an inconsistency of about 40% between the two recent Chandra datasets by Mushotzky et al. (2000) and Giacconi et al. (2000), the latter one having a lower normalisation. The new XMM data are consistent with the Giacconi et al. log(N)-log(S), maybe even somewhat flatter, and clearly confirm a break in the slope compared to the quasi-Euclidean behaviour at brighter fluxes. In the 5-10 keV band the XMM data go more than an order of magnitude deeper than the previous BeppoSAX counts (Fiore et al. 1999). There is so far relatively little deviation from a Euclidean slope and the data is fully consistent with the predictions from recent population synthesis models for the X-ray background (Gilli et al. 1999). Adding up the source counts, we resolve about 60% of the 5-10 keV X-ray background.
The diagnostic power of XMM lies in its wide energy band and its unprecedented sensitivity in the hard band. With the help of X-ray colour-colour diagrams and the ``training set'' of about 60 previously identified ROSAT sources in the same field it is possible to characterise the new XMM sources as typically harder, probably intrinsically absorbed sources. A small number of objects with similar X-ray colours has already been identified in the deepest ROSAT survey (Lehmann et al. 2000), they are type-2 Seyferts or unidentified objects with extremely red optical/NIR colours (R-K > 5). The new XMM source population is therefore very likely dominated by obscured AGN, as predicted by the AGN population synthesis models for the X-ray background.
Acknowledgements
We thank the whole XMM-Newton team for providing such a wonderful observatory and producing well-calibrated data at an early time in the mission. We thank R. Gilli for providing the model prediction in Fig. 3. We acknowledge very useful coments by the referee, G. Zamorani. Part of this work was supported by the German Deutsches Zentrum für Luft- und Raumfahrt, DLR project numbers 50 OX 9801 and 50 OR 9908.
© ESO 2001