4. Deep Field VLBI observations of the HDF

On 12-14 November 1999 the EVN conducted the first "pilot" VLBI "blank field" observations of the radio sky. The field chosen was the HDF-N - an area that is just about as "blank" and undistinguished as the radio sky gets. The brightest source in the ~ 2 arcminute radial field of view was an FR-I radio galaxy with a total WSRT 1.4 GHz flux density of ~ 1.6 mJy.

The data were recorded at a rate of 256 Mbits/sec for 32 hours - a sustained capability that is unique to the EVN (and has recently been extended to 512 Mbits/sec). Observing in phase-reference mode, a total of ~ 14 hours of "on-source" data were collected. With a resolving beam area 1 million times smaller than the WSRT HDF-N observations (see section 3), the EVN imaged an area of about 12 arcmin². Six HDF-N radio sources were thus targeted simultaneously (using wide-field imaging techniques - see Garrett et al. 1999). The final naturally weighted images have an r.m.s. noise level of ~ 33 µJy/beam - much larger than that expected from thermal noise considerations (~ 11 µJy/beam). The images are probably limited by the inclusion of poorly calibrated or completely corrupt data - difficult to identify in this case, simply by inspection.

Nevertheless, the EVN simultaneously detected three radio sources above the 165 µJy (~ 5) limit, in the inner part of HDF-Nm region (see Fig. 5). These include: VLA J123644+621133 (a z = 1.013, low-luminosity and extremely distant FR-I radio source which is resolved by the EVN into a core plus hot-spots associated with the larger scale jet), VLA J123642+621331 (a dust enshrouded, optically faint, z = 4.424 starburst system - Waddington et al. 1999) and the faintest detection, VLA J123646+621404 (a face-on spiral galaxy at z = 0.96 with a total EVN flux density of 180 µJy/beam). The diversity of optical type is interesting but the real surprise is the detection of a radio-loud AGN in the dust obscured, z = 4.4, starburst system. This argues that at least some fraction of the optically faint radio source population harbour hidden AGN (this may be similar to the same obscured population detected by Chandra). These AGN powered systems might be quite difficult to detect with SCUBA, assuming the dust temperatures are higher than that associated with pure star forming systems. In any case, the detection of this system highlights the use of VLBI as a powerful diagnostic - able to distinguish in principle (via brightness temperature arguments) between radio emission generated by nuclear starbursts and AGN.

Figure 5. EVN detections in the HDF: the distant z = 1.01 FRI (left), the z = 4.4 dusty obscured starburst hosting a hidden AGN (middle) and the z = 0.96 AGN spiral (right).