Astrophysical Constraints from the SARAS3 non-detection of the Cosmic Dawn Sky-Averaged 21-cm Signal

Observations of the redshifted 21-cm line of atomic hydrogen have resulted in several upper limits on the 21-cm power spectrum and a tentative detection of the sky-averaged signal at $z\sim 17$. Made with the EDGES Low-Band antenna, this claim was recently disputed by the SARAS3 experiment, which reported a non-detection and is the only available upper limit strong enough to constrain cosmic dawn astrophysics. We use these data to constrain a population of radio-luminous galaxies $\sim 200$ million years after the Big Bang ($z\approx 20$). We find, using Bayesian data analysis, that the data disfavours (at 68% confidence) radio-luminous galaxies in dark matter halos with masses of $4.4\times {10}^5$ M${}_{\odot }\lesssim M\lesssim 1.1\times {10}^7$M${}_{\odot }$ (where $M_{\odot }$ is the mass of the Sun) at $z=20$ and galaxies in which $>5$% of the gas is converted into stars. The data disfavour galaxies with radio luminosity per star formation rate of $L_{\mathrm{r}}\mathrm{/}\mathrm{S}\mathrm{F}\mathrm{R}\gtrsim 1.549\times {10}^{25}$ W Hz${}^{-1}$M${}_{\odot }^{-1}$ yr at 150 MHz, a thousand times brighter than today, and, separately, a synchrotron radio background in excess of the CMB by $\gtrsim 6$ at 1.42 GHz.