Modeling the neutral hydrogen distribution in the post-reionization Universe: intensity mapping

Abstract

We model the distribution of neutral hydrogen (HI) in thepost-reionization era and investigate its detectability in 21 cmintensity mapping with future radio telescopes like the SquareKilometer array (SKA). We rely on high resolution hydrodynamicalN-body simulations that have a state-of-the-art treatment of the lowdensity photoionized gas in the inter-galactic medium (IGM). The HI is assigned a-posteriori to the gas particlesfollowing two different approaches: a halo-based method in which HI isassigned only to gas particles residing within dark matter halos;a particle-based method that assigns HI to all gas particlesusing a prescription based on the physical properties of theparticles.The HI statistical properties are then compared to the observationalproperties of Damped Lyman-α Absorbers (DLAs) and of lowercolumn density systems and reasonable good agreement is found for allthe cases. Among the halo-based method, we further considertwo different schemes that aim at reproducing the observed propertiesof DLAs by distributing HI inside halos: one of this results in a muchhigher bias for DLAs, in agreement with recent observations, whichboosts the 21 cm power spectrum by a factor ∼ 4 with respect tothe other recipe. Furthermore, we quantify the contribution of HI inthe diffuse IGM to both ΩHI and the HI power spectrumfinding to be subdominant in both cases.We compute the 21 cm power spectrum from the simulated HI distributionand calculate the expected signal for both SKA1-mid and SKA1-lowconfigurations at 2.4 ⩽ z ⩽ 4. We find that SKA will be ableto detect the 21 cm power spectrum, in the non-linear regime, up tok ∼ 1 h/Mpc for SKA1-mid and k ∼ 5 h/Mpc for SKA1-low with100 hours of observations.We also investigate the perspective of imaging the HI distribution.Our findings indicate that SKA1-low could detect the most massive HIpeaks with a signal to noise ratio (SNR) higher than 5 for anobservation time of about 1000 hours at z = 4, for a synthesized beamwidth of 2′. Detection at redshifts z⩾2.4 with SKA1-midwould instead require a much longer observation time to achieve acomparable SNR level.