3.2. X-ray Constraints on the Demography and Physics of High-Redshift (z > 4) AGN
Deep X-ray surveys can find z > 4 AGN that are 10-30 times less luminous than the quasars found in wide-field optical surveys (see Figure 6). Such moderate-luminosity AGN are much more numerous and thus more representative of the AGN population than the rare, highly luminous quasars. Furthermore, unlike the rest-frame ultraviolet light sampled in ground-based surveys of z > 4 AGN, X-ray surveys suffer from progressively less absorption bias as higher redshifts are surveyed. At z > 4, hard 2-50 keV rest-frame X-rays are observed.
Optical spectroscopic follow-up of moderate-luminosity X-ray detected AGN at z > 4 is challenging, since such objects are expected to have I-band magnitudes of 23-27 (provided they are not at z 6.5, where they largely disappear from the I bandpass due to intergalactic absorption). Nevertheless, significant constraints on the sky density of such objects have been set via large-telescope spectroscopy and Lyman-break selection (e.g., Alexander et al. 2001; Barger et al. 2003b; Cristiani et al. 2004; Koekemoer et al. 2004; Wang et al. 2004c). The "bottom line" from these demographic studies is that the sky density of z > 4 AGN is 30-150 deg-2 at a 0.5-2 keV flux limit of 10-16 erg cm-2 s-1; for comparison, the Sloan Digital Sky Survey finds a sky density of z = 4-5.4 quasars of 0.12 deg-2 at an i-magnitude limit of 20.2 (e.g., Schneider et al. 2003). Combined demographic constraints from X-ray and wide-field optical surveys indicate that the AGN contribution to reionization at z 6 is small. Significantly better source statistics are needed to refine constraints on the faint end of the AGN X-ray luminosity function at z > 4, as only six moderate-luminosity, X-ray selected AGN at z > 4 are presently published. The requisite improvements are expected as follow-up studies of the surveys in Table 1 progress. The current X-ray constraints at z > 4 are plausibly consistent with the optical luminosity function from COMBO-17 (Wolf et al. 2003).
Once z > 4 AGN have been identified, broad-band spectral energy distribution analyses and X-ray spectral fitting provide information on their accretion processes and environments. The currently available data suggest that z > 4 AGN spanning a broad luminosity range are accreting in basically the same mode as AGN in the local universe (e.g., Vignali et al. 2002, 2003; Brandt et al. 2004c). The X-ray power-law photon indices of z > 4 and low-redshift AGN appear consistent, and their X-ray-to-optical flux ratios also agree after allowing for luminosity effects. The basic consistency observed provides confidence that X-ray surveys should remain an effective way to find AGN at the highest redshifts.
To constrain even lower luminosity AGN populations at z > 4, source-stacking analyses have been employed. These search for an average X-ray signal from a set of high-redshift sources whose individual members lie below the single-source X-ray detection limit. The most sensitive z > 4 source-stacking analyses to date have employed samples of 250-1700 Lyman break galaxies (e.g., Lehmer et al. 2005; B, V, and i dropouts) and 100 Ly emitters (e.g., Wang et al. 2004b). The resulting average X-ray detections and upper limits are consistent with any X-ray emission from these objects arising from star formation; thus no AGN emission has clearly been detected. These average constraints at luminosities below those that can be probed by single-source analyses further limit the contribution that AGN could have made to reionization at z 6. A complementary average constraint, derived by considering the unresolved component of the 0.5-2 keV CXRB, provides additional evidence that AGN and lower mass black holes did not dominate reionization (Dijkstra, Haiman & Loeb 2004).