Constraining astrophysics using line intensity mapping in the post-reionization regime

Abstract

Line Intensity Mapping (LIM) is an emerging technique in cosmology for quantifying clustering and strength of astrophysical sources at large length scales. In this method, instead of resolving each individual source of emission, one generates a coarse map of the total emission being observed in a broad resolution element. This helps us to detect emissions from very faint and distant sources, which are otherwise difficult to detect with the resolution and sensitivity of the presently operating telescopes. This technique thus presents a unique window for studying the large-scale structures by generating tomographic maps of the emission lines, tracking the evolution of the structures through different redshifts. Many surveys, using LIM, are ongoing in their initial phase, and many future surveys are planned to extend upon the instrumental techniques derived from them. Carbon monOxide Mapping Array Project (COMAP) and MeerKAT radio telescope are two such experiments targeting the CO(1-0) and HI 21-cm line transition, respectively, reporting upper limits at 𝑧 ≈ 3 and detections at 𝑧 ≈ 0.32 and 0.44, respectively. In this study, we analyze data from the COMAP and MeerKAT surveys to constrain the model parameters 𝛼, 𝛽, and neutral hydrogen density parameter, ΩHI, which govern an empirical CO(1-0) emission model and a semi-numerical framework for HI 21-cm line emission. We employ a Bayesian framework with Markov Chain Monte Carlo (MCMC) sampling to explore the model parameter space. Our analysis shows that the current data is insufficient to constrain the parameters 𝛼 and 𝛽, whereas we can obtain stringent constraints on ΩHI, given a specific HI injection model for halos. At 𝑧 = 0.32, we find ΩHI = 6.7+0.1 −0.3 × 10−4 when priors drawn from galaxy surveys are applied, and ΩHI = 7.2+0.3 −0.3 × 10−4 when no priors are imposed. Similarly, at 𝑧 = 0.44, the parameter is constrained to ΩHI = 10.2+0.1 −0.3 × 10−4 with priors from radio observations, and ΩHI = 10.9+0.3 −0.3 × 10−4 without priors. We additionally perform a joint analysis combining data from both surveys; however, this does not yield significant constraints on the model parameters, underscoring the need for higher signal-to-noise ratio (SNR) observations and improved data quality for robust joint inference.