2002AJ....124.1839B

We present the optical, near-infrared, submillimeter, and radio follow-up catalog of the X-ray-selected sources from an ~1 Ms Chandra observation of the Hubble Deep Field North region. We have B, V, R, I, and z' magnitudes for the 370 X-ray point sources, HK' magnitudes for 276, and spectroscopic redshifts for 182. We present high-quality spectra for 175 of these. The redshift distribution shows indications of structures at z=0.843 and 1.0175 (also detected in optical surveys), which could account for a part of the field-to-field variation seen in the X-ray number counts; however, these structures do not dominate the number of X-ray sources in the sample and hence should not strongly affect the redshift distribution. All of the X-ray sources with z>1.6 are either broad-line active galactic nuclei (AGNs) or have narrow Lyα and/or C III] 1909 Å emission; none of the known z>1.6 absorption-line galaxies in the field are detected individually in X-rays. We estimate photometric redshifts for the sources with (B-I)>1.5 (bluer than this and it is hard to distinguish between low-redshift irregulars and luminous high-redshift AGNs) and find agreement (most are within 25%) with the available spectroscopic redshifts. The majority of the galaxies in both the 2-8 keV (hard) and 0.5-2 keV (soft) samples have absolute magnitudes comparable to or more luminous than M^*_I=-22. The flux contributions separated into unit bins of redshift show that the z<1 spectroscopically identified sources already contribute about one-third of the total flux in both the hard and soft bands. Thus, major accretion onto supermassive black holes has occurred since the universe was half its present age. We find from ratios of the X-ray counts that the X-ray spectra are well described by absorption of an intrinsic Γ=1.8 power law, with N_Hvalues ranging from about 10²¹ cm^–2 to 5x10²³ cm^–2. We find very little evolution in the maximum rest-frame opacity-corrected and K-corrected 2-8 keV X-ray luminosities with decreasing redshift until z≲0.5, where the volume becomes too small to probe effectively very high luminosity sources. We estimate that the Chandra sources that produce 87% of the HEAO A X-ray background (XRB) at 3 keV produce 57% at 20 keV, provided that at high energies the spectral shape of the sources continues to be well described by a Γ=1.8 power law. However, when the Chandra contributions are renormalized to the BeppoSAX XRB at 3 keV, the shape matches fairly well the observed XRB at both energies. Thus, whether a substantial population of as yet undetected Compton-thick sources or some change in the spectral shape of the current sources from the simple power-law dependence is required to completely resolve the XRB above 10 keV depends critically on how the currently discrepant XRB measurements in the 1-10 keV energy range tie together with the higher energy XRB.