2025A&A...697A.238C

The population of high-redshift radio galaxies (HzRGs) is still poorly studied because only a few of these objects are currently known. We here present the results of a pilot project of spectroscopic identification of HzRG candidates. The candidates are selected by combining low-frequency radio and optical surveys that cover a total of ∼2000 deg² using the dropout technique, that is, the presence of a redshifted Lyman break in their photometric data. We focused on 39 g-dropout sources, which is about one -third of the selected sources, that are expected to be at 3.0 < z < 4.5. We considered single and double radio sources separately and searched for g-dropout sources at the location of the midpoint of the radio structure for the latter. The host galaxy is expected to be located there. We confirm only one out of 29 candidate HzRG associated with an extended radio source. For the compact radio sources, we instead reach a success rate of 30% by confirming 3 out of 10 HzRG targets. The four newly discovered HzRGs show a wide range of spectral radio slopes. This supports the idea that not all HzRGs are ultrasteep radio sources (USSs). The criterion for USSs is most commonly used to find HzRGs, but this method only selects a subpopulation. We discuss various contamination sources for the objects that are selected with the Lyman-break method and conclude that they are likely mainly HzRGs, but with a Lyα line that is underluminous with respect to expectations.