Chemical and physical evolution of the intergalactic medium

The aim of this thesis work is to explore coherently different aspects related to the properties of the Intergalactic Medium (IGM) and to its chemical and physical evolution from high to low redshift, especially regarding metal enrichment and the impact of different feedback processes. In the last years, high resolution spectroscopic observations have shown that the neutral hydrogen (HI) in the IGM at redshift around three, traced by the Lyman-alpha forest lines in absorption, is subjected to metal pollution down to very low column density (10^13 atoms cm^-2) with a metallicity of about 10^-3 in solar units. Since metals are produced only by stars inside galaxies, the diffuse metals in the IGM retain an important trace of the star formation and of the feedback processes from the galaxies to the IGM. The purpose of the thesis is to identify the mechanisms responsible of the enrichment, propose some physically motivated theoretical models and compare the prediction of the models with the latest observational data. In particular we focussed on two different type of enrichment: galactic (energy and momentum driven) winds produced by "starburst" galaxies at redshift z=1.5-4 and AGN feedback associated to the energy released by gas accretion onto super-massive black holes. We analysed in detail the following aspects: dynamic and energetic of wind and AGN feedbacks, IGM contamination efficiency as a function of different astrophysical and cosmological parameters, temperature and chemical composition of the metal systems, nature of the ultraviolet ionizing background (UVB). In the first part of the thesis we investigated the properties of a particular class of cosmological objects, the so called Damped Lyman-alpha systems (DLAs). These are defined as quasar (QSO) absorption systems with neutral hydrogen column density N_HI > 2 x 10^20 atoms cm^-2 (Wolfe et al. 1986). DLAs are considered as an important reservoir and/or sink of gas for the galaxy formation process in the high redshift Universe and their HI content dominate the total neutral hydrogen budget over a large fraction of the cosmic history. The interplay between DLAs and galaxies is thereby fundamental and should be addressed by any galaxy formation model. Using parallel national and internationals supercomputers, we ran high-resolution and large box-size cosmological hydrodynamical simulations of a LambdaCDM model. The numerical code used is a modified version of the Tree Particle-Mesh Smoothed Particle Hydrodynamics code GADGET-2 (Springel 2005). The main modification consists in an accurate modelling of the chemical evolution which allow us to follow the metal release from Type II and Type Ia supernovae (SNII, SNIa), along with low and intermediate mass stars (LIMS) (see Tornatore et al. 2007). We explored the numerical convergence of some relevant physical quantities and we vary the parameters describing: the properties of galactic winds, the initial stellar mass function, the linear dark matter power spectrum and the metal enrichment pattern of the IGM around DLAs. We focussed on the properties of dark matter haloes that are likely to be the hosts of DLAs systems: we predict relatively low star formation rates (0.01-0.1 M_sun/year) and metallicities around 0.1 Z_sun, at least for the bulk of our haloes of masses between 10^9 and 10^10 h^-1 M_sun hosting DLAs. For more massive haloes metallicities and star formation rates depend on the specific wind model. We found that strong galactic winds with speed of about 600 km/s, in an energy-driven wind scenario, are needed in order to match the observed column density distribution function for DLAs and the evolution of the neutral hydrogen content with redshift. The momentum-driven implementation of the galactic wind model, that relates the speed and mass load in the wind to the properties of the dark matter haloes, shows a behaviour which is intermediate between the energy-driven galactic winds of small (100 km/s) and large (600 km/s) velocities. At z=3 the contribution of haloes of masses between 10^9 and 10^10 h^-1 M_sun, for DLAs below 10^20.8 atoms cm^-2, to the column density distribution function, is significant. By interpolating physical quantities along line-of-sights through massive haloes we qualitatively showed how different galactic wind models impact on the IGM around DLAs. Furthermore, we analysed statistics related to the velocity widths of SiII associated to DLAs: while the expanding shells of gaseous matter associated to the wind can account for the observed velocities, the metallicity in the wind seems to be rather clumpy and this produces an underestimation of the observed velocity widths. We outlined possible solutions to this problem. These results are published in the paper Tescari et al. (2009, MNRAS, 397, 411). In the second part of the thesis we turned our attention to the cosmic evolution of the CIV, i.e. triply ionized carbon. Most studies of the high redshift IGM have focussed on CIV absorption, because it is strong and lies redward of the Lyman-alpha forest. Moreover the absorption line is actually a doublet with rest frame wavelengths 1548.204 and 1550.781 Angstrom, so its identification in the observational spectra is easier because of the fixed ratio between the wavelengths of the two components. The cosmological mass density of CIV, Omega_CIV, observed as a function of redshift is a fundamental quantity closely related to the metal enrichment of the IGM. Its apparent lack of evolution in the redshift interval z=[1.5,5] (Songaila 2001; Pettini et al. 2003; Boksenberg et al. 2003) is puzzling since both the physical conditions of the IGM and the properties of the ionizing background are thought to evolve between these epochs. The most recent measurements of CIV absorptions in spectra of QSOs at z=6 seem to indicate a downturn in the CIV mass density at z>5 (Becker et al. 2009; Ryan-Weber et al. 2009). At z<1, recent results based on HST UV data (Cooksey et al. 2009) give Omega_CIV=(6 +- 1) x 10^-8 corresponding to a 2.8 +- 0.5 increase over the 1.5<z<5 values. Our work was the theoretical counterpart of the D'Odorico et al. (2009) paper in which the authors present a new measurement of Omega_CIV in the redshift range [1.5,4] based on a sample of 25 high resolution, high signal-to-noise QSO spectra plus an additional sample of 8 QSO spectra from the literature. The main result is that Omega_CIV is no longer approximately constant in the considered redshift range, but shows a steady increase from z=3-5 to z=1.5-2. Using the same high-resolution and large box-size cosmological simulations of the first part, and adding some new runs with improved feedback parameters prescriptions (for example we run a simulation with the combined effect of wind and AGN feedbacks) we reproduced the Omega_CIV evolution, at least in the range z=1.5-3, by extracting mock but realistic QSO spectra inside the cosmological box and subsequently fit the CIV lines with the public available software package VPFIT. Statistically, with the momentum-driven wind run we fitted, even if with some discrepancies, the observed CIV column density distribution function, the CIV Doppler parameter (b_CIV) probability distribution function and the relation b_CIV-N_CIV, at redshifts z=1.5, 2.25 and 3. A paper containing the results of this work will be soon submitted to MNRAS (Tescari et al. 2010, in preparation).