A particular population of galaxies have drawn much interest recently, which are as faint as typical dwarf galaxies but have the sizes as large as L * galaxies, the so called ultra-diffuse galaxies (UDGs). The lack of tidal features of UDGs in dense environments suggests that their host halos are perhaps as massive as that of the Milky Way. On the other hand, galaxy formation efficiency should be much higher in the halos of such masses. Here we use the model galaxy catalog generated by populating two large simulations: the Millennium-II cosmological simulation and Phoenix simulations of 9 big clusters with the semi-analytic galaxy formation model. This model reproduces remarkably well the observed properties of UDGs in the nearby clusters, including the abundance, profile, color, and morphology, etc. We search for UDG candidates using the public data and find 2 UDG candidates in our Local Group and 23 in our Local Volume, in excellent agreement with the model predictions. We demonstrate that UDGs are genuine dwarf galaxies, formed in the halos of ∼ 10 10 M . It is the combination of the late formation time and high-spins of the host halos that results in the spatially extended feature of this particular population. The lack of tidal disruption features of UDGs in clusters can also be explained by their late infall-time.

We update our recently published model for GAlaxy Evolution and Assembly (GAEA), to include a self-consistent treatment of the partition of cold gas in atomic and molecular hydrogen. Our model provides significant improvements with respect to previous ones used for similar studies. In particular, GAEA (i) includes a sophisticated chemical enrichment scheme accounting for non-instantaneous recycling of gas, metals, and energy; (ii) reproduces the measured evolution of the galaxy stellar mass function; (iii) reasonably reproduces the observed correlation between galaxy stellar mass and gas metallicity at different redshifts. These are important prerequisites for models considering a metallicity dependent efficiency of molecular gas formation. We also update our model for disk sizes and show that model predictions are in nice agreement with observational estimates for the gas, stellar and star forming disks at different cosmic epochs. We analyse the influence of different star formation laws including empirical relations based on the hydrostatic pressure of the disk, analytic models, and prescriptions derived from detailed hydrodynamical simulations. We find that modifying the star formation law does not affect significantly the global properties of model galaxies, neither their distributions. The only quantity showing significant deviations in different models is the cosmic molecular-to-atomic hydrogen ratio, particularly at high redshift. Unfortunately, however, this quantity also depends strongly on the modelling adopted for additional physical processes. Useful constraints on the physical processes regulating star formation can be obtained focusing on low mass galaxies and/or at higher redshift. In this case, self-regulation has not yet washed out differences imprinted at early time.

We use the shear catalog from the CFHT Stripe-82 Survey to measure the subhalo masses of satellite galaxies in redMaPPer clusters. Assuming a Chabrier Initial Mass Function (IMF) and a truncated NFW model for the subhalo mass distribution, we find that the sub-halo mass to galaxy stellar mass ratio increases as a function of projected halo-centric radius r p , from M sub /M star = 4.43 +6.63 −2.23 at r p ∈ [0.1, 0.3] h −1 Mpc to M sub /M star = 75.40 +19.73 −19.09 at r p ∈ [0.6, 0.9] h −1 Mpc. We also investigate the dependence of subhalo masses on stellar mass by splitting satellite galaxies into two stellar mass bins: 10 < log(M star /h −1 M ) < 10.5 and 11 < log(M star /h −1 M ) < 12. The best-fit subhalo mass of the more massive satellite galaxy bin is larger than that of the less massive satellites: log(M sub /h −1 M ) = 11.14 +0.66 −0.73 (M sub /M star = 19.5 +19.8 −17.9 ) versus log(M sub /h −1 M ) = 12.38 +0.16 −0.16 (M sub /M star = 21.1 +7.4 −7.7 ).

The estimate of stellar metallicities (Z *) of high-z galaxies are of paramount importance in order to understand the complexity of dust effects and the reciprocal interrelations among stellar mass, dust attenuation, stellar age and metallicity. Benefiting from uniquely deep FUV spectra of > 500 star-forming galaxies at redshifts 2 < z < 5 extracted from the VANDELS survey and stacked in bins of stellar mass and UV continuum slope (β), we estimate their stellar metallicities Z * from stellar photospheric absorption features at 1501 and 1719 Å, which are calibrated with Starburst99 models and are largely unaffected by stellar age, dust, IMF, nebular continuum or interstellar absorption. Comparing them to photometric based spectral slopes in the range 1250-1750 Å, we find that the stellar metallicity increases by ∼ 0.5 dex from β ∼ −2 to β ∼ −1 (1 A 1600 3.2), and a dependence with β holds at fixed UV absolute luminosity M UV and stellar mass up to ∼ 10 9.65 M. As a result, the metallicity is a fundamental ingredient for properly rescaling dust corrections based on M UV and M *. Using the same absorption features, we analyze the mass-metallicity relation (MZR), and find it is consistent with the previous VANDELS estimation based on a global fit of the FUV spectra. Similarly, we do not find a significant evolution between z ∼ 2 and z ∼ 3.5. Finally, the slopes of our MZR and Z *-β relation are in agreement with the predictions of wellstudied semi-analytic models of galaxy formation (SAM), while some tensions with observations remain as to the absolute metallicity normalization. The relation between UV slope and stellar metallicity is fundamental for the exploitation of large volume surveys with next generation telescopes and for the physical characterization of galaxies in the first billion years of our Universe.

We present a comprehensive analysis of atomic hydrogen (HI) properties using a semianalytical model of galaxy formation and N-body simulations covering a large cosmological volume at high resolution. We examine the HI mass function and the HI density, characterizing both their redshift evolution and their dependence on hosting halo mass. We analyze the HI content of dark matter haloes in the local Universe and up to redshift z = 5, discussing the contribution of different galaxy properties. We find that different assembly history plays a crucial role in the scatter of this relation. We propose new fitting functions useful for constructing mock HI maps with HOD techniques. We investigate the HI clustering properties relevant for future 21 cm Intensity Mapping (IM) experiments, including the HI bias and the shot noise level. The HI bias increases with redshift and it is roughly flat on the largest scales probed. The scale dependency is found at progressively larger scales with increasing redshift, apart from a dip feature at z = 0. The shot-noise values are consistent with the ones inferred by independent studies, confirming that shot-noise will not be a limiting factor for IM experiments. We detail the contribution from various galaxy properties on the HI power spectrum and their relation to the halo bias. We find that HI poor satellite galaxies play an important role at the scales of the 1-halo term. Finally, we present the 21 cm signal in redshift space, a fundamental prediction to be tested against data from future radio telescopes such as SKA.