The informativeness of [C II] line-intensity mapping as a probe of the H I content and metallicity of galaxies at the end of reionization

TL;DR

This paper develops a new model linking halo properties to [C II] emission, predicts the detectability of the [C II] signal at high redshift with upcoming surveys, and explores how model parameters influence the intensity distribution.

Contribution

It introduces a novel halo-[C II] connection model based on FIRE simulations and forecasts the detectability of the [C II] LIM signal at z>6 with future experiments.

Findings

A 4000-hour CCAT survey can detect the [C II] power spectrum at 5σ significance at z~6.

Extended surveys could improve detection significance to 48σ at z~6.

Model parameters influence the one-point statistics and can be constrained through intensity distribution analysis.

Abstract

Line-intensity mapping (LIM) experiments coming online now will survey fluctuations in aggregate emission in the [C II] ionized carbon line from galaxies at the end of reionization. Experimental progress must be matched by theoretical reassessments of approaches to modelling and the information content of the signal. We present a new model for the halo-[C II] connection, building upon results from the FIRE simulations suggesting that gas mass and metallicity most directly determine [C II] luminosity. Applying our new model to an ensemble of peak-patch halo lightcones, we generate new predictions for the [C II] LIM signal at $z\gtrsim6$. We expect a baseline 4000-hour LIM survey from the CCAT facility to have the fundamental sensitivity to detect the [C II] power spectrum at a significance of $5\sigma$ at $z\sim6$, with an extended or successor Stage 2 experiment improving significance to $48\sigma$ at $z\sim6$ and achieving $11\sigma$ at $z\sim7.5$. Cross-correlation through stacking, simulated against a mock narrow-band Lyman-break galaxy survey, would yield a strong detection of the radial profile of cosmological [C II] emission surrounding star-forming galaxies. We also analyse the role of a few of our model's parameters through the pointwise relative entropy (PRE) of the distribution of [C II] intensities. While the PRE signature of different model parameters can become degenerate or diminished after factoring in observational distortions, various parameters do imprint themselves differently on the one-point statistics of the intrinsic signal. Further work can pave the way to access this information and distinguish different sources of non-Gaussianity in the [C II] LIM observation.