21-cm cosmology

TL;DR

The review surveys how the redshifted 21 cm line from neutral hydrogen can illuminate the cosmic dawn, the epoch of reionization, and beyond. It combines atomic physics (spin temperature coupling), global-signal evolution, 3D tomography, and intensity mapping to provide a comprehensive framework for interpreting current and upcoming radio observations. Key contributions include analytic formalisms for coupling and heating fluctuations, excursion-set-based models of ionization topology, and pragmatic discussions of simulation approaches and detectability. The work underscores the potential of 21 cm observations to constrain cosmology, structure formation, and fundamental physics, while also outlining substantial observational and theoretical challenges ahead.

Abstract

Imaging the Universe during the first hundreds of millions of years remains one of the exciting challenges facing modern cosmology. Observations of the redshifted 21 cm line of atomic hydrogen offer the potential of opening a new window into this epoch. This would transform our understanding of the formation of the first stars and galaxies and of the thermal history of the Universe. A new generation of radio telescopes is being constructed for this purpose with the first results starting to trickle in. In this review, we detail the physics that governs the 21 cm signal and describe what might be learnt from upcoming observations. We also generalize our discussion to intensity mapping of other atomic and molecular lines.

Figures (20)

• Figure 1: The 21-centimeter cosmic hydrogen signal. (a) Time evolution of fluctuations in the 21-cm brightness from just before the first stars formed through to the end of the reionization epoch. This evolution is pieced together from redshift slices through a simulated cosmic volume santos2008. Coloration indicates the strength of the 21-cm brightness as it evolves through two absorption phases (purple and blue), separated by a period (black) where the excitation temperature of the 21-cm hydrogen transition decouples from the temperature of the hydrogen gas, before it transitions to emission (red) and finally disappears (black) owing to the ionization of the hydrogen gas. (b) Expected evolution of the sky-averaged 21-cm brightness from the "dark ages" at redshift 200 to the end of reionization, sometime before redshift 6 (solid curve indicates the signal; dashed curve indicates $T_b=0$). The frequency structure within this redshift range is driven by several physical processes, including the formation of the first galaxies and the heating and ionization of the hydrogen gas. There is considerable uncertainty in the exact form of this signal, arising from the unknown properties of the first galaxies.
• Figure 2: Left panel: Hyperfine structure of the hydrogen atom and the transitions relevant for the Wouthuysen-Field effect pritchard2006. Solid line transitions allow spin flips, while dashed transitions are allowed but do not contribute to spin flips. Right panel: Illustration of how atomic cascades convert Ly$n$ photons into Ly$\alpha$ photons.
• Figure 3: Cartoon of the different phases of the 21 cm signal. The signal transitions from an early phase of collisional coupling to a later phase of Ly$\alpha$ coupling through a short period where there is little signal. Fluctuations after this phase are dominated successively by spatial variation in the Ly$\alpha$ , X-ray, and ionizing UV radiation backgrounds. After reionization is complete there is a residual signal from neutral hydrogen in galaxies.
• Figure 4: Top panel: Evolution of the CMB temperature $T_{\rm CMB}$ (dotted curve),the gas kinetic temperature $T_K$ (dashed curve), and the spin temperature $T_S$ (solid curve). Middle panel: Evolution of the gas fraction in ionized regions $x_i$ (solid curve) and the ionized fraction outside these regions (due to diffuse X-rays) $x_e$ (dotted curve). Bottom panel: Evolution of mean 21 cm brightness temperature $T_b$. In each panel we plot curves for model A (thin curves), model B (medium curves), and model C (thick curves). pritchard2008
• Figure 5: Optical depths per time for various photon-IGM processes, in units of the Hubble time, at z = 300, assuming a neutral IGM. These include processes which deposit energy directly into the IGM (pair production and photoionization), processes which redistribute photons (2$\gamma\rightarrow2\gamma$) and ones that do both (Compton). At very low energies, photoionization is the dominant process; at very high energies, e$\pm$ pair production dominates mack2008.