Le 6 décembre 2013
De 10h30 à 11h30
Lucio MAYER
University of Zurich (Suisse)
The theory of galaxy formation within the standard LCDM paradigm has faced several problems that in the past have generated serious doubts on the cosmological model itself. The progess of numerical simulations in the last few years, however, has convincingly shown that the formation of galaxies with realistic structural properties is viable in LCDM, but it is strongly dependent on our understanding of star formation and feedback. The progress has been made possible by the combination of increased numerical resolution and more careful modeling of sub-grid astrophysical processes. In particular, some numerical simulations, both SPH and AMR, rely on the ability to resolve at least the major star forming regions corresponding to Giant Molecular Clouds (GMCs), thus capturg the clustered mode of star formation and driving naturally strong supernovae outflows. I will show how the same methodology can deliver galaxies with realistic properties from the faintest gas-rich dwarfs to spiral galaxies similar to our own Milky Way portrayed in the Eris simulations. In low mass galaxies a sizable dark matter cores forms as a result of repeated baryonic mass loss episodes due to supernovae winds. The lowest mass dwarfs, with virial masses below 10^9 Mo, show little or no star formation due to the inability to host gas dense enough to form stars. At larger mass scales, the Eris simulations match nearly all the observational constraints available on the Milky Way, including the kinematics of halo stars. However, the same simulations require the suppression of cooling via metal lines to avoid overproducing stellar masses of galaxies. Increasing supernovae feedback to compensate for cooling leads to thick, turbulent disks that do not resemble those of spiral galaxies. Hence this leaves us with one remaining challenge, namely we need to identify an alternative feedback mechanism that only acts to suppress cooling without having significant dynamical effects.