During the most active period of star formation in galaxies, which occurs in the redshift range 1 < z < 3, strong bursts of star formation result in significant quantities of dust, which obscures new stars being formed as their UV/optical light is absorbed and then re-emitted in the infrared, which redshifts into the mm/sub-mm bands for these early times. To get a complete picture of the high- z galaxy population, we need to survey a large patch of the sky in the sub-mm with sufficient angular resolution to resolve all galaxies, but we also need the depth to fully sample their cosmic evolution, and therefore obtain their redshifts using direct mm spectroscopy with a very wide frequency coverage. This requires a large single-dish sub-mm telescope with fast mapping speeds at high sensitivity and angular resolution, a large bandwidth with good spectral resolution and multiplex spectroscopic capabilities. The proposed 50-m Atacama Large Aperture Submillimeter Telescope (AtLAST) will deliver these specifications. We discuss how AtLAST allows us to study the whole population of high-z galaxies, including the dusty star-forming ones which can only be detected and studied in the sub-mm, and obtain a wealth of information for each of these up to z ∼ 7: gas content, cooling budget, star formation rate, dust mass, and dust temperature. We present worked examples of surveys that AtLAST can perform, both deep and wide, and also focused on galaxies in proto-clusters. In addition we show how such surveys with AtLAST can measure the growth rate f σ ₈ and the Hubble constant with high accuracy, and demonstrate the power of the line-intensity mapping method in the mm/sub-mm wavebands to constrain the cosmic expansion history at high redshifts, as good examples of what can uniquely be done by AtLAST in this research field.

We present the initial results of an ongoing survey with the Karl G. Jansky Very Large Array targeting the CO(J = 1-0) transition in a sample of 30 submillimeter-selected, dusty star-forming galaxies (SFGs) at z = 2-5 with existing mid-J CO detections from the Atacama Large Millimeter/submillimeter Array and NOrthern Extended Millimeter Array, of which 17 have been fully observed. We detect CO(1-0) emission in 11 targets, along with three tentative (∼1.5σ-2σ) detections; three galaxies are undetected. Our results yield total molecular gas masses of 6-23 × 10¹⁰ (α _CO/1) M _⊙, with gas mass fractions, f _gas = M _mol/(M _*+M _mol), of 0.1-0.8 and a median depletion time of (140 ± 70) Myr. We find median CO excitation ratios of r ₃₁ = 0.75 ± 0.39 and r ₄₁ = 0.63 ± 0.44, with significant scatter. We find no significant correlation between the excitation ratio and a number of key parameters such as redshift, CO(1-0) line width, or Σ_SFR. We only find a tentative positive correlation between r ₄₁ and the star-forming efficiency, but we are limited by our small sample size. Finally, we compare our results to predictions from the SHARK semi-analytical model, finding a good agreement between the molecular gas masses, depletion times, and gas fractions of our sources and their SHARK counterparts. Our results highlight the heterogeneous nature of the most massive SFGs at high redshift, and the importance of CO(1-0) observations to robustly constrain their total molecular gas content and interstellar medium properties.