Zoom-in on the dust-obscured phase of galaxy formation with gravitational lenses

Over the last 20 years gravitational lensing has become an essential instrument to investigate the structures within the Universe and the Universe itself. It directly traces the gravity of matter, whether baryonic or not, and so it’s essential for a systematic study of dark matter and its distribution on both small and large scales. Moreover, the magnification generated by a foreground lensing system, like a massive elliptical galaxy, on a background source allows us to study high-redshift galaxy structures down to scales difficult to probe with the largest telescope at present, and to detect intrinsically faint objects. In this PhD thesis I describe the advantages that gravitational lensing offers in the study of high-redshift (z > 1.5) dusty star forming galaxies (DSFGs), progenitors of the early-type (ETGs) observed in the local Universe. DSFGs are the major contributors to the cosmic star formation activity in the Universe and, as such, they represent the key to understand the build-up of galaxies. Dust absorption of UV/optical radiation from newborn stars is re-emitted in the far-infrared/sub-mm bands, making DSFGs particularly bright at those wavelengths. In order to extract information from the galaxy-galaxy strong lensing events involving DSFGs, I have written a Python code performing lens modelling and source reconstruction, based on the Regularized Semilinear Inversion method by Warren & Dye (2003), as outlined in Enia et al. (2018). This method reconstructs the intrinsic (i.e. un-lensed) surface brightness of the background galaxy without any analytic pre-assumption on its distribution, while searching in the parameters space for the mass distribution of the lens. Since DSFGs are the main focus of this project, and since they are very bright at FIR/sub-mm wavelengths, I have extended the formalism to the uv plane, in order to deal with interferometric data. In fact, interferometry is the best observational technique to achieve high resolution imaging in the sub-mm/mm bands, thanks to facilities like the Atacama Large Millimeter/sub-millimeter Array (ALMA). Furthermore, since the lens galaxy is usually a massive elliptical, the main emission in the sub-mm/mm is due almost exclusively from the background galaxy. DSFGs show extremely steep number counts, so that any DSFGs with a very high flux density (e.g. above ∼ 100mJy at 500μm) is expected to be lensed. The selection of gravitationally lensed galaxies based on a simple cut in flux density has proven to be extremely efficient in the search for these sources in wide area extragalactic surveys such as the Herschel Astrophysical Terahertz Large Area Survey (H-ATLAS). This survey found 80 candidate lensed galaxies, 20 of which have already been confirmed to be lensing systems by a number of follow-up observations mainly with the Sub-Millimeter Array (SMA), the Hubble Space Telescope and the Keck telescope. In Enia et al. (2018) I have applied my code to the SMA observations of 12 strongly lensed galaxies from the H-ATLAS in order to derive their morphologies, sizes and magni cations. The derived lens model parameters are in general consistent with previous findings (i.e. Bussmann et al., 2013), however the estimated magnification factors, ranging from 3 to 10, are lower. These discrepancies are observed in particular where the reconstructed source hints at the presence of multiple knots of emission. An appropriate estimate of the magnification factor is essential to properly retrieve the physical properties of the sources, i.e. CO line luminosities, star formation rates, or SFR surface densities. In Massardi, Enia et al., 2018 multiwavelength observations of two strongly lensed sources are presented. H-ATLAS J090740.0-004200, also known as SDP.9, and H-ATLAS J091043.1-000322, also known as SDP.11, both come from the H-ATLAS sample. The observations were carried out with Chandra, HST and ALMA, covering a large portion of the electromagnetic spectrum. These multiwavelength observations probed the presence of highly obscured nuclear activity in the galaxy, with X-ray emissions generated in the nuclear area, allowing an insight on the co-evolution between the central SMBH and the galaxy, as predicted by various evolutionary theories for galaxy formation and evolution. I applied the code to SDP.9, reconstructing the background source in the different bands, obtaining a clear cospatiality in the source plane between the sub-mm emission, tracing the star formation, and the X-ray signal, tracing the nuclear activity, within a circle of ∼ 400 pc diameter. This analysis will be further exploited in the future thanks to the large number of follow-up campaigns in different wavelength ranges currently ongoing. In Rodighiero, Enia et al., ApJL submitted, a study of the statistical properties of a sample of dusty sources with very efficient star formation rates (SFR) is performed, in order to understand the role of enhanced SFR in triggering the Black-Hole Accretion Rate. These sources are Herschel-selected in the COSMOS field, with SFRs elevated 4× above the star-forming ”main sequence”, classifying them as starbursts (SB). Here, by means of a multicomponent spectral energy distribution fitting analysis, the contribution of stars, AGN torus, and star formation to the total emission at different wavelengths is separated, spanning the range from the UV to the far-IR. The sample is divided into active SBs (dominated by an AGN emission, SBs-AGN) and purely star-forming SBs (SBs-SFR). From visual inspection of the HST-UV morphology, the two classes have statistically different morphologies: SBs-SFR are generally irregular systems, while a large majority (∼ 65%) of SBs-AGN are instead dominated by regular compact and symmetric morphologies. Searching in the ALMA public archive, I found continuum counterparts with a secure detection above 3σ for 24 galaxies (10 SBs-AGN and 14 SBs-SFR). Then, dust and total molecular gas masses are computed, finding that SBs turn to be gas rich systems (fgas =45%−85%),with similar gas fractions in the two classes, and therefore no direct evidence of AGN feedback depleting the parent hosts. This results are discussed in the context of the co-evolution scenario. The SB population is consistent with a mixture of: low-mass primordial galaxies, rapidly accreting their M∗ together with their MBH (mainly the more compact SBs-AGN), and a class of highly star-forming merging systems (dominating the SBs-SFR). Anyway, feedback effects have not reduced yet the fgas of the objects. Alternatively, feedback processes (in form of galactic outflows from the SMBH) are not efficient enough to significantly deplete the gas masses of the host galaxies.