Modelling cosmological reionization and its observational signatures

Bianco, Michele (2021) Modelling cosmological reionization and its observational signatures. Doctoral thesis (PhD), University of Sussex.

[img] PDF - Published Version
Download (17MB)


The Epoch of Reionization is an important period in studying structure formation and evolution of our Universe. The first luminous objects, which may have been star-forming galaxies and quasi-stellar objects, influenced later-day structures formation and evolution. These bright objects produced enough ultra-violet radiation to alter the nature of the host and propagated out into the intergalactic medium. These energetic photons transitioned our Universe from a cold and neutral state to ultimately a hot and ionised state. This interesting period is one of the least understood epochs in the Universe evolution due to the lack of direct observations. The redshifted 21-cm signal of neutral hydrogen can be used as an observable sign of reionisation. The upcoming Square Kilometre Array telescope will be sensitive enough to detect the 21-cm signal and produce images of its spatial distribution throughout reionisation.

This research focuses on improving numerical methods and develop new techniques for understanding and interpreting future observational evidence. Our simulations will play a crucial role and provide numerical support for the upcoming experiments. We proposed a new approach that correctly quantifies the effect of local recombinations on the scale below the large numerical simulation resolution. We present a more general model for the sub-grid gas clumping, depending on the local density. I improved the latter method with an empirical stochastic model based on high-resolution N-body simulation results, and the relevant fluctuations are fully resolved. Moreover, we developed a stable and reliable convolutional neural network, which can identify neutral and ionised regions from noisy 21-cm image observations. The network can identify the regions of interest with greater precision and is less sensitive to the limitation of previous methods. We successfully recover the signal for different instrumental noise levels based on the intensity contrast in the 21-cm signal and from ionised regions simulation independent pattern.

Item Type: Thesis (Doctoral)
Schools and Departments: School of Mathematical and Physical Sciences > Physics and Astronomy
Subjects: Q Science > QB Astronomy > QB0980 Cosmogony. Cosmology
Depositing User: Library Cataloguing
Date Deposited: 13 Sep 2021 11:27
Last Modified: 13 Sep 2021 11:27

View download statistics for this item

📧 Request an update