2023MNRAS.521..332T

We present a revised measurement of the extragalactic background light (EBL) at radio frequencies based on a near complete compendium of radio source counts. We present the radio-EBL at 150 MHz, 325 MHz, 610 MHz, 1.4 GHz, 3 GHz, 5 GHz, and 8.4 GHz. In all cases the contribution to the radio-EBL, per decade of flux, exhibits a two-humped distribution well matched to the active galactic nucleus (AGN) and star-forming galaxy (SFG) populations, and with each population contributing roughly equal energy. Only at 3 GHz are the source count contributions to the EBL fully convergent, and hence we report empirical lower limits to the radio-EBL in the remaining bands. Adopting predictions from the SHARK semi-analytic model for the form of the SFG population, we can fit the fainter source counts providing measurements of the total contribution to the radio-EBL for the SFG and the AGN populations separately. This constitutes an empirically constrained model-dependent measurement for the SFG contribution, but a fully empirical measurement of the AGN contribution. Using the PROSPECT spectral energy distribution code we can model the ultraviolet-optical-infrared-mm-radio SFG EBL at all frequencies from the cosmic star-formation history and the adoption of a Chabrier initial mass function. However, significant discrepancy remains (5 ×) between our source-count estimates of the radio-EBL and the direct measurements reported from the Absolute Radiometer for Cosmology, Astrophysics, and Diffuse Emission-2 (ARCADE-2) experiment. We can rule out a significant missing discrete source radio population and suggest that the cause of the high ARCADE-2 radio-EBL values may need to be sought either in the foreground subtraction or as a yet unknown diffuse component in the radio sky.