[en] Studying present day properties of galaxies can give information both on their past evolution and on the matter content of the Universe. Morphological and kinematical properties can be studied statistically, as a function of environment, with recent observations. To interpret these observed properties, we have modelled the formation and past evolution of galaxies with numerical simulations. The codes that are used, already existing for a part, and developed during the thesis for the other part, model the gravitational dynamics of stars, gas clouds and dark matter, the dissipative hydrodynamic nature of interstellar gas, and star formation. Outputs from the simulations can then be compared to observations to understand which processes have led to the current state of galaxies. In a first part, we study the morphology of isolated galaxies. We show that most spiral galaxies have a bar - an elongated central concentration of stars. This bar should have been destroyed in a few billions year by angular momentum transfer to the interstellar medium. The ubiquity of bars since 10 billion years then requires them to be reformed or strengthen by an external process. This can be explained only if galaxies accrete large amounts of gas from the intergalactic medium. The study of disk lopsidedness in spiral galaxies also argues towards this conclusion. We estimate the typical accretion rate, of w few solar masses per year for a Milky-Way like galaxy. This give strong constraints for cosmological models: the Universe must contain enough baryonic matter, that must not be converted into stars too rapidly, so that enough gas remains to be accreted and increase the mass of spiral galaxies by a few tens of percent below the epoch of redshift z=1. Then, galaxies have grown by a gas accretion process, for a part, by collisions and mergers also play an important role in their evolution. It is well-known that equal-mass galaxy mergers from elliptical galaxies. We show that even mergers will small companions strongly affect the disk of spiral galaxies. Indeed, many mergers over a large range of mass ratios can convert spiral galaxies into lenticular or elliptical systems; the accretion of gas evoked before can reform a thin spiral disk later-on. We also demonstrate that multiple successive minor mergers with dwarf galaxies can lead to the formation of elliptical galaxies, as a major merger with a same-mass galaxy do. This new process for the formation of elliptical galaxies should lead to a revision of estimates of the dark matter content in elliptical galaxies, which may give new information on the nature of this dark matter itself. Others phenomena occur during galaxy collisions (formation of rings, genesis of dwarf galaxies in tidal tails). They have been studied with the help of our numerical models. They help to constrain the properties and nature of dark matter, by indicating how it behaves during galaxy interactions. The comparison of numerical models to observational data indicates that a part of dark matter if condensed into the collisional and tidal debris. Even if higher resolution observations re still required to reach a definitive conclusion, this tends to indicate that a part of dark matter has a collisional dynamics -- probably the baryonic component of dark matter, which favor the hypothesis of dark molecular gas. The statistical comparison of observations and high-resolution numerical simulations has thus given constraints on the physical process that drive galaxy evolution, and the visible and dark matter content of the Universe. In the future, increase capability of observation and numerical simulation, dedicated in particular to distant galaxies, will enable an even better understanding of galaxy evolution, large-scale star formation, and the history of the baryonic content of the Universe. Cosmological scenarios for the evolution of the whole Universe can be tested and/or refined with the help of such studies. (author)

[en] The diversity of galaxy morphologies in the local Universe is still poorly understood: the 'Hubble Sequence' for massive galaxies ranges from star forming disk-dominated spirals to non-star forming, spheroid-dominated elliptical galaxies. Numerical simulations are a powerful tool to study the formation and evolution of galaxies, but all codes have physical and numerical limitations. I will present a critical review of the strengths and limitations of various codes used to model the gravitational dynamics of galaxies and the hydrodynamics of their interstellar medium, and discuss how these aspects change with the rapid evolution of modern supercomputers. In particular, some hydrodynamic codes can now reach high enough spatial resolution and resolve the gas cooling down to relatively low temperatures, so that supersonic turbulence in the interstellar gas can be correctly modeled. Describing the formation of stars on galactic scales is nevertheless still challenging. I will then review recent results on galaxy formation obtained from numerical simulations and observations. Spiral galaxies were not born with their modern shape, but were much more irregular in the young Universe. Observations suggest galaxy collisions and mergers are not the dominant process in their formation. Instead, smooth gas flows assembled massive gas-rich disks, inside which a strong turbulence developed. These primordial disks then fragmented into giant clumps, formed stars actively, and built the bulges and exponential disks of today's spiral galaxies. Star formation and morphological evolution in Milky Way-like galaxies is thus mostly driven by internal processes. At the opposite, galaxies that did not escape violent interactions and mergers became today's ellipticals. Interactions first trigger star formation and massive star cluster formation, but later form spheroidal galaxies in which efficient star formation is naturally quenched, explaining the so-called 'red and read' ellipticals. The diversity of galaxy properties thus seems to result from different formation and evolution mechanisms, but a number of observed properties remain unexplained: I will review the main open questions in the field, and perspectives for observations and numerical simulations. (author)

[en] We present hydrodynamic simulations of a major merger of disk galaxies, and study the interstellar medium (ISM) dynamics and star formation (SF) properties. High spatial and mass resolutions of 12 pc and 4 x 10⁴ M _sun allow us to resolve cold and turbulent gas clouds embedded in a warmer diffuse phase. We compare lower-resolution models, where the multiphase ISM is not resolved and is modeled as a relatively homogeneous and stable medium. While merger-driven bursts of SF are generally attributed to large-scale gas inflows toward the nuclear regions, we show that once a realistic ISM is resolved, the dominant process is actually gas fragmentation into massive and dense clouds and rapid SF therein. As a consequence, SF is more efficient by a factor of up to ∼10 and is also somewhat more extended, while the gas density probability distribution function rapidly evolves toward very high densities. We thus propose that the actual mechanism of starburst triggering in galaxy collisions can only be captured at high spatial resolution and when the cooling of gas is modeled down to less than 10³ K. Not only does our model reproduce the properties of the Antennae system, but it also explains the 'starburst mode' recently revealed in high-redshift mergers compared to quiescent disks.

[en] Spiral galaxies have most of their stellar mass in a large rotating disk, and only a modest fraction in a central spheroidal bulge. This challenges present models of galaxy formation: galaxies form at the center of dark matter halos through a combination of hierarchical merging and gas accretion along cold streams. Cosmological simulations thus predict that galaxies rapidly grow their bulge through mergers and instabilities and end up with most of their mass in the bulge and an angular momentum much below the observed level, except in dwarf galaxies. We propose that the continuous return of gas by stellar populations over cosmic times could help to solve this issue. A population of stars formed at a given instant typically returns half of its initial mass in the form of gas over 10 billion years, and the process is not dominated by supernovae explosions but by the long-term mass-loss from low- and intermediate-mass stars. Using simulations of galaxy formation, we show that this gas recycling can strongly affect the structural evolution of massive galaxies, potentially solving the bulge fraction issue, as the bulge-to-disk ratio of a massive galaxy can be divided by a factor of 3. The continuous recycling of baryons through star formation and stellar mass loss helps the growth of disks and their survival to interactions and mergers. Instead of forming only early-type, spheroid-dominated galaxies (S0 and ellipticals), the standard cosmological model can successfully account for massive late-type, disk-dominated spiral galaxies (Sb-Sc).

[en] We present an analytical model of the relation between the surface density of gas and star formation rate in galaxies and clouds, as a function of the presence of supersonic turbulence and the associated structure of the interstellar medium (ISM). The model predicts a power-law relation of index 3/2, flattened under the effects of stellar feedback at high densities or in very turbulent media, and a break at low surface densities when ISM turbulence becomes too weak to induce strong compression. This model explains the diversity of star formation laws and thresholds observed in nearby spirals and their resolved regions, the Small Magellanic Cloud, high-redshift disks and starbursting mergers, as well as Galactic molecular clouds. While other models have proposed interstellar dust content and molecule formation to be key ingredients to the observed variations of the star formation efficiency, we demonstrate instead that these variations can be explained by ISM turbulence and structure in various types of galaxies.

[en] The question of what regulates star formation is a longstanding issue. To investigate this issue, we run simulations of a kiloparsec cube section of a galaxy with three kinds of stellar feedback: the formation of H ii regions, the explosion of supernovae, and ultraviolet heating. We show that stellar feedback is sufficient to reduce the averaged star formation rate (SFR) to the level of the Schmidt–Kennicutt law in Milky Way–like galaxies but not in high-redshift gas-rich galaxies, suggesting that another type of support should be added. We investigate whether an external driving of the turbulence such as the one created by the large galactic scales could diminish the SFR at the observed level. Assuming that the Toomre parameter is close to 1 as suggested by the observations, we infer a typical turbulent forcing that we argue should be applied parallel to the plane of the galactic disk. When this forcing is applied in our simulations, the SFR within our simulations closely follows the Schmidt–Kennicutt relation. We found that the velocity dispersion is strongly anisotropic with the velocity dispersion alongside the galactic plane being up to 10 times larger than the perpendicular velocity.

[en] We point out a natural mechanism for quenching of star formation in early-type galaxies (ETGs). It automatically links the color of a galaxy with its morphology and does not require gas consumption, removal or termination of gas supply. Given that star formation takes place in gravitationally unstable gas disks, it can be quenched when a disk becomes stable against fragmentation to bound clumps. This can result from the growth of a stellar spheroid, for instance by mergers. We present the concept of morphological quenching (MQ) using standard disk instability analysis, and demonstrate its natural occurrence in a cosmological simulation using an efficient zoom-in technique. We show that the transition from a stellar disk to a spheroid can be sufficient to stabilize the gas disk, quench star formation, and turn an ETG red and dead while gas accretion continues. The turbulence necessary for disk stability can be stirred up by sheared perturbations within the disk in the absence of bound star-forming clumps. While other quenching mechanisms, such as gas stripping, active galactic nucleus feedback, virial shock heating, and gravitational heating are limited to massive halos, MQ can explain the appearance of red ETGs also in halos less massive than ∼10¹² M _sun. The dense gas disks observed in some of today's red ellipticals may be the relics of this mechanism, whereas red galaxies with quenched gas disks could be more frequent at high redshift.

[en] The formation of thick stellar disks in spiral galaxies is studied. Simulations of gas-rich young galaxies show formation of internal clumps by gravitational instabilities, clump coalescence into a bulge, and disk thickening by strong stellar scattering. The bulge and thick disks of modern galaxies may form this way. Simulations of minor mergers make thick disks too, but there is an important difference. Thick disks made by internal processes have a constant scale height with galactocentric radius, but thick disks made by mergers flare. The difference arises because in the first case, perpendicular forcing and disk-gravity resistance are both proportional to the disk column density, so the resulting scale height is independent of this density. In the case of mergers, perpendicular forcing is independent of the column density and the low-density regions get thicker; the resulting flaring is inconsistent with observations. Late-stage gas accretion and thin-disk growth are shown to preserve the constant scale heights of thick disks formed by internal evolution. These results reinforce the idea that disk galaxies accrete most of their mass smoothly and acquire their structure by internal processes, in particular through turbulent and clumpy phases at high redshift.

[en] We study the evolution of galactic bars and the link with disk and spheroid formation in a sample of zoom-in cosmological simulations. Our simulation sample focuses on galaxies with present-day stellar masses in the 10¹⁰-10¹¹ M_☉ range, in field and loose group environments, with a broad variety of mass growth histories. In our models, bars are almost absent from the progenitors of present-day spirals at z > 1.5, and they remain rare and generally too weak to be observable down to z ≈ 1. After this characteristic epoch, the fractions of observable and strong bars rise rapidly, bars being present in 80% of spiral galaxies and easily observable in two thirds of these at z ≤ 0.5. This is quantitatively consistent with the redshift evolution of the observed bar fraction, although the latter is presently known up to z ≈ 0.8 because of band-shifting and resolution effects. Our models hence predict that the decrease in the bar fraction with increasing redshift should continue with a fraction of observable bars not larger than 10%-15% in disk galaxies at z > 1. Our models also predict later bar formation in lower-mass galaxies, in agreement with existing data. We find that the characteristic epoch of bar formation, namely redshift z ≈ 0.8-1 in the studied mass range, corresponds to the epoch at which today's spirals acquire their disk-dominated morphology. At higher redshift, disks tend to be rapidly destroyed by mergers and gravitational instabilities and rarely develop significant bars. We hence suggest that the bar formation epoch corresponds to the transition between an early 'violent' phase of spiral galaxy formation at z ≥ 1 and a late 'secular' phase at z ≤ 0.8. In the secular phase, the presence of bars substantially contributes to the growth of the (pseudo-)bulge, but the bulge mass budget remains statistically dominated by the contribution of mergers, interactions, and disk instabilities at high redshift. Early bars at z > 1 are often short-lived, while most of the bars formed at z ≤ 1 persist down to z = 0, late cosmological gas infall being necessary to maintain some of them.

[en] Galaxies above redshift 1 can be very clumpy, with irregular morphologies dominated by star complexes as large as 2 kpc and as massive as a few x10⁸ or 10⁹ M _sun. Their co-moving densities and rapid evolution suggest that most present-day spirals could have formed through a clumpy phase. The clumps may form by gravitational instabilities in gas-rich turbulent disks; they do not appear to be separate galaxies merging together. We show here that the formation of the observed clumps requires initial disks of gas and stars with almost no stabilizing bulge or stellar halo. This cannot be achieved in models where disk galaxies grow by mergers. Mergers tend to make stellar spheroids even when the gas fraction is high, and then the disk is too stable to make giant clumps. The morphology of high-redshift galaxies thus suggests that inner disks assemble mostly by smooth gas accretion, either from cosmological flows or from the outer disk during a grazing interaction.