The Cosmic Evolution Survey (COSMOS) has become a cornerstone of extragalactic astronomy. Since the last public catalog in 2015, a wealth of new imaging and spectroscopic data have been collected in the COSMOS field. This paper describes the collection, processing, and analysis of these new imaging data to produce a new reference photometric redshift catalog. Source detection and multiwavelength photometry are performed for 1.7 million sources across the 2 deg2 of the COSMOS field, ∼966,000 of which are measured with all available broadband data using both traditional aperture photometric methods and a new profile-fitting photometric extraction tool, The Farmer, which we have developed. A detailed comparison of the two resulting photometric catalogs is presented. Photometric redshifts are computed for all sources in each catalog utilizing two independent photometric redshift codes. Finally, a comparison is made between the performance of the photometric methodologies and of the redshift codes to demonstrate an exceptional degree of self-consistency in the resulting photometric redshifts. The i < 21 sources have subpercent photometric redshift accuracy and even the faintest sources at 25 < i < 27 reach a precision of 5%. Finally, these results are discussed in the context of previous, current, and future surveys in the COSMOS field. Compared to COSMOS2015, it reaches the same photometric redshift precision at almost one magnitude deeper. Both photometric catalogs and their photometric redshift solutions and physical parameters will be made available through the usual astronomical archive systems (ESO Phase 3, IPAC-IRSA, and CDS).

We report Atacama Large Millimeter Array observations of the neutral atomic carbon transitions [C i] and multiple CO lines in a sample of ∼30 main-sequence galaxies at , including novel information on [C i] and CO for 7 of such normal objects. We complement our observations with a collection of >200 galaxies with coverage of similar transitions, spanning the z = 0–4 redshift interval and a variety of ambient conditions from local to high-redshift starbursts. We find systematic variations in the [C i]/IR and [C i]/high-J upper (J upper = 7) CO luminosity ratios among the various samples. We interpret these differences as increased dense molecular gas fractions and star formation efficiencies in the strongest high-redshift starbursts with respect to normal main-sequence galaxies. We further report constant  / ratios across the galaxy populations and redshifts, suggesting that gas temperatures T exc traced by [C i] do not strongly vary. We find only a mild correlation with T dust and that, generally, T exc ≲ T dust. We fit the line ratios with classical photodissociation region models, retrieving consistently larger densities and intensities of the UV radiation fields in submillimeter galaxies than in main-sequence and local objects. However, these simple models fall short in representing the complexity of a multiphase interstellar medium and should be treated with caution. Finally, we compare our observations with the Santa Cruz semi-analytical model of galaxy evolution, recently extended to simulate submillimeter emission. While we confirm the success in reproducing the CO lines, we find systematically larger [C i] luminosities at fixed IR luminosity than predicted theoretically. This highlights the necessity of improving our understanding of the mechanisms regulating the [C i] emission on galactic scales. We release our data compilation to the community.

We present the detection of multiple carbon monoxide CO line transitions with ALMA in a few tens of infrared-selected galaxies on and above the main sequence at z = 1.1−1.7. We reliably detected the emission of CO (5 − 4), CO (2 − 1), and CO (7 − 6)+[C I](3P2  −  3P1) in 50, 33, and 13 galaxies, respectively, and we complemented this information with available CO (4 − 3) and [C I](3P1  −  3P0) fluxes for part of the sample, and by modeling of the optical-to-millimeter spectral energy distribution. We retrieve a quasi-linear relation between LIR and CO (5 − 4) or CO (7 − 6) for main-sequence galaxies and starbursts, corroborating the hypothesis that these transitions can be used as star formation rate (SFR) tracers. We find the CO excitation to steadily increase as a function of the star formation efficiency, the mean intensity of the radiation field warming the dust (⟨U⟩), the surface density of SFR (ΣSFR), and, less distinctly, with the distance from the main sequence (ΔMS). This adds to the tentative evidence for higher excitation of the CO+[C I] spectral line energy distribution (SLED) of starburst galaxies relative to that for main-sequence objects, where the dust opacities play a minor role in shaping the high-J CO transitions in our sample. However, the distinction between the average SLED of upper main-sequence and starburst galaxies is blurred, driven by a wide variety of intrinsic shapes. Large velocity gradient radiative transfer modeling demonstrates the existence of a highly excited component that elevates the CO SLED of high-redshift main-sequence and starbursting galaxies above the typical values observed in the disk of the Milky Way. This excited component is dense and it encloses ∼50% of the total molecular gas mass in main-sequence objects. We interpret the observed trends involving the CO excitation as to be mainly determined by a combination of large SFRs and compact sizes, as a large ΣSFR is naturally connected with enhanced dense molecular gas fractions and higher dust and gas temperatures, due to increasing ultraviolet radiation fields, cosmic ray rates, as well as dust and gas coupling. We release the full data compilation and the ancillary information to the community.

We present new observations, carried out with IRAM NOEMA, of the atomic neutral carbon transitions [C i]( 3 P 1 -3 P 0 ) at 492 GHz and [C i]( 3 P 2 -3 P 1 ) at 809 GHz of GN20, a well-studied star-bursting galaxy at z = 4.05. The high luminosity line ratio [C i]( 3 P 2 -3 P 1 ) /[C i]( 3 P 1 -3 P 0 ) implies an excitation temperature of 48 +14 −9 K, which is significantly higher than the apparent dust temperature of T d =33 ± 2 K (β = 1.9) derived under the common assumption of an optically thin far-infrared dust emission, but fully consistent with T d = 52 ± 5 K of a general opacity model where the optical depth (τ) reaches unity at a wavelength of λ 0 = 170 ± 23 µm. Moreover, the general opacity solution returns a factor of ∼ 2× lower dust mass and, hence, a lower molecular gas mass for a fixed gas-to-dust ratio, than with the optically thin dust model. The derived properties of GN20 thus provide an appealing solution to the puzzling discovery of starbursts appearing colder than main-sequence galaxies above z > 2.5, in addition to a lower dust-to-stellar mass ratio that approaches the physical value predicted for starburst galaxies.

We investigate the fraction of close pairs and morphologically identified mergers on and above the star-forming main sequence (MS) at 0.2 ≤ z ≤2.0. The novelty of our work lies in the use of a non-parametric morphological classification performed on resolved stellar mass maps, reducing the contamination by non-interacting, high-redshift clumpy galaxies. We find that the merger fraction rapidly rises to ≥70% above the MS, implying that -already at z 1 -starburst (SB) events (∆ MS ≥ 0.6) are almost always associated with a major merger (1:1 to 1:6 mass ratio). The majority of interacting galaxies in the SB region are morphologically disturbed, late-stage mergers. Pair fractions show little dependence on MS offset and pairs are more prevalent than late-stage mergers only in the lower half of the MS. In our sample, major mergers on the MS occur with a roughly equal frequency of ∼5-10% at all masses 10 10 M . The MS major merger fraction roughly doubles between z = 0.2 and z = 2, with morphological mergers driving the overall increase at z 1. The differential redshift evolution of interacting pairs and morphologically classified mergers on the MS can be reconciled by evolving observability time-scales for both pairs and morphological disturbances. The observed variation of the late-stage merger fraction with ∆ MS follows the perturbative 2-Star Formation Mode model, where any MS galaxy can experience a continuum of different SFR enhancements. This points to an SB-merger connection not only for extreme events, but also more moderate bursts which merely scatter galaxies upward within the MS, rather than fully elevating them above it.