Constraining structure formation using EDGES

The experiment to detect the global epoch of reionization signature (EDGES) collaboration reported the detection of a line at 78 MHz in the sky-averaged spectrum due to neutral hydrogen (/ion{H}{i}) 21-cm hyperfine absorption of cosmic microwave background (/cmb) photons at redshift $z\sim 17$. This requires that the spin temperature of /ion{H}{i} be coupled to the kinetic temperature of the gas at this redshift through the scattering of /lya photons emitted by massive stars. To explain the experimental result, star formation needs to be sufficiently efficient at $z\sim 17$ and this can be used to constrain models in which small-scale structure formation is suppressed (/dmf models), either due to dark matter free-streaming or non-standard inflationary dynamics. We combine simulations of structure formation with a simple recipe for star formation to investigate whether these models emit enough Lyman-$\alpha$ photons to reproduce the experimental signal for reasonable values of the star formation efficiency, $f_{\star }$. We find that a thermal warm dark matter (/wdm) model with mass $m_{WDM}\sim 4.3keV$ is consistent with the timing of the signal for $f_{\star }\lesssim 2\%$. The exponential growth of structure around $z\sim 17$ in such a model naturally generates a sharp onset of the absorption. A warmer model with $m_{WDM}\sim 3keV$ requires a higher star formation efficiency, $f_{\star }\sim 6\%$, which is a factor of few above predictions of current star formation models and observations of satellites in the Milky Way. However, uncertainties in the process of star formation at these redshifts do not allow to derive strong constrains on such models using 21-cm absorption line. The onset of the 21-cm absorption is generally slower in /dmf than observed in cold dark matter (/cdm) models, unless some process significantly suppresses star formation in halos with masses below ~10⁸h^-1M_⊙.