4.4. Deep Chandra and XMM-Newton Hard X-ray Surveys
Because the ROSAT Deep Surveys resolved most of the soft X-ray background, much of the emphasis with Chandra and XMM-Newton has been on hard X-ray surveys 7. It also seems that at the faintest fluxes reached by Chandra (~ 10-16 ergs cm-2 s-1 in the 2 - 10 keV band, corresponding to 10 counts in the deepest 2 × 106 s exposures), the relative fraction of objects which are AGN declines rapidly (Hornscheimer et al. 2003), indicating that one has reached the "end of AGN-ness", similar to what is seen in the deep radio surveys.
The major advantage of the Chandra data is that, to very faint optical magnitudes (I ~ 28), there is almost always an "unique" optical identification (or lack thereof, Barger et al. 2003a; Koekemoer et al. 2004). Thus, as opposed to almost all other surveys (with the exception of radio data), the identifications are certain, and one does not need to rely on optical spectroscopy to confirm the identification.
The nature of the hard Chandra sources has been rather surprising (Mushotzky et al. 2000; Barger et al. 2001; Alexander et al. 2001). As summarized in Barger et al. (2003b), less than 30% of the optical counterparts have strong broad lines or a non-thermal continuum, and many of the other 70% are pure absorption-line objects. Most of the light from these sources is due to stars in both the optical (Barger et al. 2002) and IR (Crawford et al. 2001) bands; even with HST images, the nuclei are almost invisible (Grogin et al. 2003; Cowie & Barger, this volume). Most of these objects have stellar luminosities near L* or brighter in the K-band. Given the very high areal density of the Chandra sources (more than 3000 deg-2 at FX ~ 10-16 ergs cm-2 s-1 in the 2 - 8 keV band), this makes these "optically-dull" (a nomenclature first used by Elvis et al. 1981) objects the most numerous class of AGN. Similar results are seen in the XMM-Newton surveys (Barcons et al. 2002). Comparison of the optical colors of the XMM-Newton sources shows that only about half are consistent with the region used by the SDSS to find AGN (Richards et al. 2002).
The exquisite Chandra positions have allowed the detection of the X-ray-selected AGN population that (a) dominates the AGN numbers, (b) are not found by standard optical selection criteria, and (c) would not have been found in previous X-ray surveys. This is a stunning example of the "lamppost effect" alluded to in the introduction and should open our eyes to the possibility of even more such surprises when IR selection techniques are well developed.
7 For the most recent update on the Chandra and XMM-Newton surveys, see Barcons (2003). Back.