X iv : a st ro - p h / 01 02 41 1 v 1 2 3 Fe b 20 01 The Chandra Deep Survey of the Hubble Deep Field North Area . IV . An Ultradeep Image

We present results from a 479.7 ks Chandra exposure of the Hubble Deep Field North (HDF-N) and its immediate vicinity. In our X-ray image, the deepest ever reported with a 0.5–2.0 keV flux limit of ≈ 4.9 × 10 −17 erg cm −2 s −1 , we detect four new HDF-N X-ray sources bringing the total number of such sources to 12. The new sources include two optically bright (R = 18.3–18.8), low-redshift (z < 0.15) galaxies, a Fanaroff-Riley I radio galaxy, and an edge-on spiral hosting either a powerful starburst or a low-luminosity active galactic nucleus (AGN). Notably, we have now detected X-ray emission from all luminous galaxies (M V < −18) with z < 0.15 known in the HDF-N. We have also detected the remarkable µJy radio source VLA J123642.09+621331.4, which is located just outside the HDF-N and has a likely redshift of z = 4.424. The observed X-ray emission supports the presence of an AGN in this object, and its X-ray-to-optical flux ratio (i.e., α ox) is consistent with what is seen for low-redshift AGN. We have detected X-ray variability from two of the previously known HDF-N X-ray sources, and spectral fitting shows clear evidence for X-ray absorption in the brightest X-ray source in the HDF-N, CXOHDFN J123646.3+621404, a z = 0.960 broad-line AGN with associated Mg ii absorption. Stacking analyses of optically bright HDF-N galaxies not individually detected in X-rays have provided estimates of their average X-ray fluxes, and we find that the X-ray luminosities of " normal " spirals at z ≈ 0.5 are not more than a factor of ≈ 3 larger than those of spirals in the local Universe (z < 0.01). This constrains models for the evolution of low-mass X-ray binary populations in galaxies in response to the declining cosmic star-formation rate. Monte-Carlo simulations support the validity of the stacking analyses and show that the Chandra Advanced CCD Imaging Spectrometer (ACIS) performs source detection well even with effective exposure times of ≈ 8 Ms.