HI as a Probe of the Large Scale Structure in the Post-Reionization Universe

TL;DR

This paper models the distribution of neutral Hydrogen in the post-reionization universe using N-body simulations, highlighting high bias at small scales and evaluating detection prospects with current and future radio telescopes.

Contribution

It introduces a gravity-only simulation-based model for HI distribution that aligns with observations and explores detection strategies with radio telescopes like MWA and GMRT.

Findings

High bias of HI at small scales identified

Detection of HI fluctuations feasible with current telescopes

Direct detection of rare HI peaks possible within 1000 hours

Abstract

We model the distribution of neutral Hydrogen (HI hereafter) in the post-reionization universe. This model uses gravity only N-Body simulations and an ansatz to assign HI to dark matter haloes that is consistent with observational constraints and theoretical models. We resolve the smallest haloes that are likely to host HI in the simulations, care is also taken to ensure that any errors due to the finite size of the simulation box are small. We then compute the smoothed one point probability distribution function and the power spectrum of fluctuations in HI. This is compared with other predictions that have been made using different techniques. We highlight the significantly high bias for the HI distribution at small scales. This aspect has not been discussed before. We then discuss the prospects for detection with the MWA, GMRT and the hypothetical MWA5000. The MWA5000 can detect visibility correlations at large angular scales at all redshifts in the post-reionization era. The GMRT can detect visibility correlations at lower redshifts, specifically there is a strong case for a survey at z=1.3. We also discuss prospects for direct detection of rare peaks in the HI distribution using the GMRT. We show that direct detection should be possible with an integration time that is comparable to, or even less than, the time required for a statistical detection. Specifically, it is possible to make a statistical detection of the HI distribution by measuring the visibility correlation, and, direct detection of rare peaks in the HI distribution at z = 1.3 with the GMRT in less than 1000 hours of observations.