One of the most challenging and exciting subjects in modern astrophysics is that of galaxy formation at the epoch of reionisation. The SKA, with its revolutionary capabilities in terms of frequency range, resolution and sensitivity, will allow to exp
lore the first Gyr of structure formation in the Universe, in particular, with the detection and study of the earliest manifestations of the AGN phenomenon. The tens of QSOs that are currently known out to the highest redshifts (z~7), many of them exhibiting powerful radio emission, imply that super-massive black holes can be grown on a very short timescale and support the existence of very high redshift (z > 7) radio loud sources - sources that have so far escaped detection. Not only would such detections be paramount to the understanding of the earliest stages of galaxy evolution, they are necessary for the direct study of neutral hydrogen in the Epoch of Reionisation, through observations of the HI 21cm forest against such background sources. In order to understand how SKA and SKA1 observations can be optimised to reveal these earliest AGN, we have examined the effect of a hot CMB on the emission of powerful and young radio galaxies. By looking at the SKA1 capabilities, in particular in terms of wavelength coverage and resolution, we determine how the effects of CMB-muting of a radio loud source can be observationally minimised and how to identify the best highest-redshift radio candidates. Considering different predictions for the space density of radio loud AGN at such redshifts, we identify the survey characteristics necessary to optimize the detection and identification of the very first generation of radio loud AGN in the Universe.
Ultra Steep Spectrum (USS) radio sources have been successfully used to select powerful radio sources at high redshifts (z>~2). Typically restricted to large-sky surveys and relatively bright radio flux densities, it has gradually become possible to
extend the USS search to sub-mJy levels, thanks to the recent appearance of sensitive low-frequency radio facilities. Here a first detailed analysis of the nature of the faintest USS sources is presented. By using Giant Metrewave Radio Telescope and Very Large Array radio observations of the Lockman Hole at 610 MHz and 1.4 GHz, a sample of 58 USS sources, with 610 MHz integrated fluxes above 100 microJy, is assembled. Deep infrared data at 3.6 and 4.5 micron from the Spitzer Extragalactic Representative Volume Survey (SERVS) is used to reliably identify counterparts for 48 (83%) of these sources, showing an average total magnitude of [3.6](AB)=19.8 mag. Spectroscopic redshifts for 14 USS sources, together with photometric redshift estimates, improved by the use of the deep SERVS data, for a further 19 objects, show redshifts ranging from z=0.1 to z=2.8, peaking at z~0.6 and tailing off at high redshifts. The remaining 25 USS sources, with no redshift estimate, include the faintest [3.6] magnitudes, with 10 sources undetected at 3.6 and 4.5 micron (typically [3.6]>22-23 mag, from local measurements), which suggests the likely existence of higher redshifts among the sub-mJy USS population. The comparison with the Square Kilometre Array Design Studies Simulated Skies models indicate that Fanaroff-Riley type I radio sources and radio-quiet Active Galactic Nuclei may constitute the bulk of the faintest USS population, and raises the possibility that the high efficiency of the USS technique for the selection of high redshift sources remains even at the sub-mJy level.
