Annu. Rev. Astron. Astrophys. 2000. 38: 667-715 Copyright © 2000 by Annual Reviews. All rights reserved

4.8. Radio sources

The HDF-N has become one of the best-studied regions of the sky at radio wavelengths, and the availability of extremely deep optical/IR imaging and extensive spectroscopy has made it arguably the most important survey for understanding the properties of "average" radio sources at the millijansky and microjansky level. The VLA and MERLIN arrays have been used to map regions including the HDF-N at 3.5 cm and 20cm wavelengths [Fomalont et al. 1997, Richards et al. 1998, Richards et al. 1999, Muxlow et al. 1999, Richards 2000]. The combined MERLIN + VLA data have been used to make high-resolution maps of 91 HDF radio sources, permitting a detailed study of radio source extent, as well as pinpointing locations for a few ambiguous identifications and confirming the blank-field character of others. Altogether, 16 radio sources are located in the central HDF-N WFPC2 field, and many more in the flanking fields.

The optical counterparts for most of the radio sources (72 of 92 in the region with suitably deep imaging) are relatively bright galaxies (<I> 22) at 0.4 < z < 1. Apart from a few manifestly AGN powered objects, usually giant elliptical radio galaxies (including the z = 1.02 galaxy HDF 4-752.1), most are disk galaxies, sometimes disturbed or interacting. For nearly all, the radio emission is resolved with a typical size ~ 1", reinforcing the evidence that it primarily arises from star formation. However, about 20% of the sources have very faint or optically invisible counterparts with I 25. Many of these have K-band identifications with very red optical-IR colors [Richards et al. 1999]. [Waddington et al. 1999] have studied one red HDF radio source which is apparently at z = 4.4 (from a single-line spectroscopic redshift). Several have been detected in 850 µm observations with SCUBA [Barger et al. 1999]. These "invisible" radio sources appear to be a population discontinuous from the majority of the optically brighter sources, and may be high-redshift, dust-obscured, vigorously star-forming systems. Finally, two VLA sources in the HDF proper have no optical counterparts to the limits of the WFPC2 data, and one (123646+621226) has no counterpart to the extremely faint limit of the NICMOS GTO HDF field.

The HDF-S was observed by the Australia Telescope National Facility (ATNF) in 1998. The first year's observations covered a 60 diameter region at 20 cm detecting 240 sources down to 100 µJy (13 in the primary WFPC2 field) and correspondingly smaller fields at 13, 6 and 3 cm. These have now been augmented by substantially deeper observations which are currently being analyzed (Norris et al. 2000), but which should reach 5 sensitivities of 40 µJy at 13 and 6 cm.