Unveiling the evolution of the CO excitation ladder through cross-correlation of CONCERTO-like experiments and galaxy redshift surveys

Abstract

Context: Rotational CO transitions, while acting as a foreground for [C II] line-intensity mapping (LIM) experiments, trace the physical conditions of cold gas in galaxies at lower redshifts. Studying these transitions is also crucial for improving component-separation methods as LIM sensitivity increases. Aims: Galaxy-evolution models have so far predicted only the total CO LIM signal. We explore the potential of cross-correlating millimeter-wave LIM data with spectroscopic galaxy surveys to constrain individual CO-line contributions, measure the CO-background spectral line energy distribution (SLED), and derive the cosmic molecular gas density, $\rho_{\mathrm{H2}}(z)$, up to $z = 3$. Methods: We built 12 light cones of $9~\mathrm{deg}^2$ from the Simulated Infrared Extragalactic Sky (SIDES) simulation. By analyzing cross-power spectra between different CO transitions and the galaxy density field, we recovered the CO background SLED. Combining it with bias-weighted line intensities yielded $\rho_{\mathrm{H2}}(z)$. We also assessed the detectability of the CO(4--3) cross-power spectrum with a CONCERTO-like experiment. Results: For a realistic spectroscopic depth, the CO background SLED is accurately recovered up to $J_{\mathrm{up}} = 6$ with $\leq 20%$ uncertainties. Reconstructing $\rho_{\mathrm{H2}}$ from millimeter LIM data requires an excitation correction relative to CO(1--0). Interloper-induced variance does not prevent precise $\rho_{\mathrm{H2}}$ estimation. In the two-star-formation-mode SIDES model, starbursts dominate the SLED at $J_{\mathrm{up}} \geq 6$ but do not bias $\rho_{\mathrm{H2}}$ estimates from $2 \leq J_{\mathrm{up}} \leq 6$. However, CONCERTO lacks the sensitivity to detect the CO$\times$galaxy cross-power on relevant scales, even under ideal conditions.