The Chandra deep field north survey. VI. The nature of the optically faint X-ray source population

We provide constraints on the nature of the optically faint (I > 24) X-ray source population from a 1 Ms Chandra exposure of a 8′.4 × 8′.4 region within the Hawaii flanking-field area containing the Hubble Deep Field North region. We detect 47 (2400 ⁺⁴⁰⁰_-350 deg ^-2) optically faint sources down to 0.5-2.0 keV and 2.0-8.0 keV fluxes of ≈3 × 10^-17 ergs cm^-2 s^-1 and ≈2 × 10^-16 ergs cm^-2 s^-1, respectively; these sources contribute ≈14% and ≈21% of the 0.5-2.0 keV and 2.0-8.0 keV X-ray background radiation, respectively. The fraction of optically faint X-ray sources is approximately constant (at ≈35%) for 0.5-8.0 keV fluxes from 3 × 10^-14 ergs cm^-2 s^-1 down to the X-ray flux limit. A considerable fraction (30⁺¹⁴_-10%) of the optically faint X-ray sources are Very Red Objects (I-K ≥ 4). Analysis of the optical and X-ray properties suggests a large number of optically faint X-ray sources are likely to host obscured active galactic nucleus (AGN) activity at z = 1-3. From these results we calculate that a significant fraction (≈5%-45%) of the optically faint X-ray source population could be obscured QSOs (rest-frame unabsorbed 0.5-8.0 keV luminosity > 3 x 10⁴⁴) at z ≤ 3. Given the number of X-ray sources without I-band counterparts, there are unlikely to be more than ≈15 sources at z > 6. We provide evidence that the true number of z > 6 sources is considerably lower. We investigate the multiwavelength properties of optically faint X-ray sources. Nine optically faint X-ray sources have μJy radio counterparts; ≈53⁺²⁴_-17% of the optically faint μJy radio sources in this region. The most likely origin of the X-ray emission in these X-ray detected, optically faint μJy radio sources is obscured AGN activity. However, two of these sources have been previously detected at submillimeter wavelengths, and the X-ray emission from these sources could be due to luminous star formation activity. Assuming the spectral energy distribution of NGC 6240, we estimate the 175 μm flux of a typical optically faint X-ray source to be less than 10 mJy; however, those sources with detectable submillimeter counterparts (i.e., f_850μm > 3 mJy) could be substantially brighter. Hence, most optically faint X-ray sources are unlikely to contribute significantly to the far-IR (140-240 μm) background radiation. However, as expected for sources with AGN activity, the two optically faint X-ray sources within the most sensitive area of the ISOCAM HDF-N region have faint (≲50 μJy) 15 μm counterparts. We also provide constraints on the average X-ray properties of classes of optically faint sources not individually detected at X-ray energies. Stacking analyses of optically faint μJy radio sources not individually detected with X-ray emission yields a possible detection (at 98.3% confidence) in the 0.5-2.0 keV band; this X-ray emission could be produced by star formation activity at z = 1-3. None of the optically faint AGN-candidate sources in the HDF-N itself are detected at X-ray energies either individually or with stacking analyses, showing that these sources have low X-ray luminosities if they are indeed AGNs.