Cosmic evolution of the Civ in high-resolution hydrodynamic simulations

We investigate the properties of triply ionized carbon (Civ) in the intergalactic medium (IGM) using a set of high resolution and large box size cosmological hydrodynamic simulations of a Lambda cold dark matter (ΛCDM) model. We rely on a modification of the publicly available TreeSPH code gadget-2 that self-consistently follows the metal enrichment mechanism by means of a detailed chemical evolution model. We focus on several numerical implementations of galactic feedback: galactic winds in the energy-driven and momentum-driven prescriptions, galactic winds hydrodynamically coupled to the surrounding gas and active galactic nuclei (AGNs) powered by gas accretion on to massive black holes. Furthermore, our results are compared to a run in which galactic feedback is not present and we also explore different initial stellar mass function. After having addressed some global properties of the simulated volume like the impact on the star formation rate and the content in carbon and Civ, we extract mock IGM transmission spectra in neutral hydrogen (Hi) and Civ and perform Voigt profile fitting. The results are then compared with high-resolution quasar (QSO) spectra obtained with the Ultraviolet Echelle Spectrograph (UVES) at the Very Large Telescope (VLT) and the High Resolution Echelle Spectrograph (HIRES) at Keck. We find that feedback has little impact on statistics related to the neutral hydrogen, while Civ is more affected by galactic winds and/or AGN feedback. The feedback schemes investigated have a different strength and redshift evolution with a general tendency for AGN feedback to be more effective at lower redshift than galactic winds. When the same analysis is performed over observed and simulated Civ lines, we find a reasonably good agreement between data and simulations over the column density range NCIV= 1012.5-15cm-2. Also the Civ linewidth distribution appears to be in agreement with the observed values, while the Hi Doppler parameters, bHI, are in general too large, showing that the diffuse cosmic web is heated more than what is inferred by observations. The simulation without feedback fails in reproducing the Civ systems at high column densities at all redshift, while the AGN feedback case agrees with observations only at z < 3, when this form of feedback is particularly effective. We also present scatter plots in the b-N and in the NCIV-NHI planes, showing that there is rough agreement between observations and simulations only when feedback i