Sergey Mashchenko

Stellar feedback in dwarf galaxy formation.

Dwarf galaxies pose substantial challenges for cosmological models. In particular, current models predict a dark-matter density that is divergent at the center, which is in sharp contrast with observations that indicate a core of roughly constant density. Energy feedback, from supernova explosions and stellar winds, has been proposed as a major factor shaping the evolution of dwarf galaxies. We present detailed cosmological simulations with sufficient resolution both to model the relevant physical processes and to directly assess the impact of stellar feedback on observable properties of dwarf galaxies. We show that feedback drives large-scale, bulk motions of the interstellar gas, resulting in substantial gravitational potential fluctuations and a consequent reduction in the central matter density, bringing the theoretical predictions in agreement with observations.

The removal of cusps from galaxy centres by stellar feedback in the early Universe

The standard cosmological model, now strongly constrained by direct observations of the Universe at early epochs, is very successful in describing the evolution of structure on large and intermediate scales. Unfortunately, serious contradictions remain on smaller, galactic scales. Among the main small-scale problems is a significant and persistent discrepancy between observations of nearby galaxies, which imply that galactic dark matter haloes have a density profile with a flat core, and the cosmological model, which predicts that the haloes should have divergent density (a cusp) at the centre. Here we report numerical simulations that show that random bulk motions of gas in small primordial galaxies, of the magnitude expected in these systems, will result in a flattening of the central dark matter cusp on relatively short timescales (∼108 years). Gas bulk motions in early galaxies are driven by supernova explosions that result from ongoing star formation. Our mechanism is general, and would have operated in all star-forming galaxies at redshifts z ≥ 10. Once removed, the cusp cannot be reintroduced during the subsequent mergers involved in the build-up of larger galaxies. As a consequence, in the present Universe both small and large galaxies would have flat dark matter core density profiles, in agreement with observations.The standard cosmological model, now strongly constrained by direct observation at early epochs, is very successful in describing the structure of the evolved universe on large and intermediate scales. Unfortunately, serious contradictions remain on smaller, galactic scales. Among the major small-scale problems is a significant and persistent discrepancy between observations of nearby galaxies, which imply that galactic dark matter (DM) haloes have a density profile with a flat core, and the cosmological model, which predicts that the haloes should have divergent density (a cusp) at the centre. Here we use numerical N-body simulations to show that random bulk motions of gas in small primordial galaxies, of the magnitude expected in these systems, result in a flattening of the central DM cusp on short timescales (of order 10^8 years). Gas bulk motions in early galaxies are driven by supernova explosions which result from ongoing star formation. Our mechanism is general and would have operated in all star-forming galaxies at redshifts z>~ 10. Once removed, the cusp cannot be reintroduced during the subsequent mergers involved in the build-up of larger galaxies. As a consequence, in the present universe both small and large galaxies would have flat DM core density profiles, in agreement with observations.

GLOBULAR CLUSTERS WITH DARK MATTER HALOS. II. EVOLUTION IN A TIDAL FIELD

In this second paper in our series, we continue to test primordial scenarios of globular cluster formation, which predict that globular clusters formed in the early universe in the potential of dark matter minihalos. In this paper we use high-resolution N-body simulations to model tidal stripping experienced by primordial dark matter-dominated globular clusters in the static gravitational potential of the host dwarf galaxy. We test both cuspy Navarro-Frenk-White (NFW) and flat-core Burkert models of dark matter halos. Our primordial globular cluster with an NFW dark matter halo survives severe tidal stripping and after 10 orbits is still dominated by dark matter in its outskirts. Our cluster with a Burkert dark matter halo loses almost all its dark matter to tidal stripping and starts losing stars at the end of our simulations. The results of this paper reinforce our conclusion in Paper I that current observations of globular clusters are consistent with the primordial picture of globular cluster formation.

Globular Clusters with Dark Matter Halos. I. Initial Relaxation

In a series of two papers, we test the primordial scenario of globular cluster formation using results of high-resolution N-body simulations. In this first paper we study the initial relaxation of a stellar core inside a live dark matter minihalo in the early universe. Our dark matter-dominated globular clusters show features that are usually attributed to the action of the tidal field of the host galaxy. Among them are the presence of an apparent cutoff (tidal radius) or a break in the outer parts of the radial surface brightness profile and a flat line-of-sight velocity dispersion profile in the outskirts of the cluster. The apparent mass-to-light ratios of our hybrid (stars + dark matter) globular clusters are very close to those of purely stellar clusters. We suggest that additional observational evidence, such as the presence of obvious tidal tails, is required to rule out the presence of significant amounts of dark matter in present-day globular clusters.

Building the Stellar Halo through Feedback in Dwarf Galaxies

We present a new model for the formation of stellar halos in dwarf galaxies. We demonstrate that the stars and star clusters that form naturally in the inner regions of dwarfs are expected to migrate from the gas rich, star forming centre to join the stellar spheroid. For dwarf galaxies, this process could be the dominant source of halo stars. The effect is caused by stellar feedback-driven bulk motions of dense gas which, by causing potential fluctuations in the inner regions of the halo, couple to all collisionless components. This effect has been demonstrated to generate cores in otherwise cuspy cold dark matter profiles and is particularly effective in dwarf galaxy haloes. It can build a stellar spheroid with larger ages and lower metallicities at greater radii without requiring an outside-in formation model. Globular cluster-type star clusters can be created in the galactic ISM and then migrate to the spheroid on 100Myr timescales. Once outside the inner regions they are less susceptible to tidal disruption and are thus long lived; clusters on wider orbits may be easily unbound from the dwarf to join the halo of a larger galaxy during a merger. A simulated dwarf galaxy (Mvir ≃ 10 9 M⊙ at z = 5) is used to examine this gravitational coupling to dark matter and stars.

Impact of ultraviolet radiation from giant spirals on the evolution of dwarf galaxies

We show that ultraviolet (UV) radiation, with wavelengths shorter than 2000 A, escaping from the discs of giant spirals could be one of the principal factors affecting the evolution of low-mass satellite galaxies. We demonstrate, using an analytical approach, that the Lyman continuum part of the radiation field can lead to the ionization of the interstellar medium (ISM) of dwarf galaxies through the process of photoevaporation, making the ISM virtually unobservable. The far-ultraviolet part (912 < λ < 2000 A) is shown to dominate over the internal sources of radiation for most of the Galactic dwarf spheroidals. The proposed environmental factor could be at least partially responsible for the bifurcation of the low-mass protogalaxies into two sequences: dwarf irregulars and dwarf spheroidals. We discuss many peculiarities of the Local Group early-type dwarfs, which can be accounted for by the impact of the UV radiation from the host spiral galaxy (Milky Way or M31).

The Universality of Initial Conditions for Globular Cluster Formation

We investigate a simple model for globular cluster (GC) formation. We simulate the violent relaxation of initially homogeneous isothermal stellar spheres and show that it leads to the formation of clusters with radial density profiles that match the observed profiles of GCs. The best match is achieved for dynamically unevolved clusters. In this model, all the observed correlations between global GC parameters are accurately reproduced if one assumes that all the clusters initially had the same value of the stellar density and the velocity dispersion. This suggests that the gas that formed GCs had the same values of density and temperature throughout the universe.

FORMATION OF MINIGALAXIES IN DEFUNCT COSMOLOGICAL H ii REGIONS

Using a large set of high-resolution numerical simulations incorporating nonequilibrium molecular hydrogen chemistry and a constant source of external radiation, we study gas collapse in previously photoionized minigalaxies with virial temperatures less than 104 K in the early universe (redshifts z = 10-20). We confirm that the mechanism of positive feedback of ionizing radiation on star formation in minigalaxies proposed by Ricotti and coworkers can be efficient despite a significant flux of metagalactic photodissociating radiation. We derive critical fluxes for the Lyman-Werner background radiation sufficient to prevent the collapse of gas in minigalaxies as a function of the virial mass of the halo and redshift. In our model, the formation of minigalaxies in defunct H II regions is most efficient at large redshifts (z 15) and/or for large local gas overdensity δ 10. We show that nonequilibrium chemistry plays an important dynamical role not only during the initial evolutionary phase, leading to the gas becoming gravitationally unstable inside the minihalo, but also at the advanced stages of the core collapse, resulting in efficient gas accretion in the core region. We speculate on a possible connection between our objects and metal-poor globular clusters and dwarf spheroidal galaxies.

Complex lens design: searching for a needle in a haystack

Optical design of complex (multi-element) lenses is traditionally considered to be part science and part art, primarily because of the enormous complexity of the problem. Recent advances in high performance computing (HPC) made it feasible to adopt a purely scientific approach in discovering new lens designs. In this paper, I formulate the task of finding a new lens design that satisfies a given set of constraints as a search for the global minimum of a function of unknown and very large (~ 30 – 100) number of dimensions. I address the significant complication that only a tiny fraction of the volume of the free parameters space is physically accessible. I propose a smart lens drafting algorithm which circumvents this difficulty. I present my numerical code which can be used to discover novel complex lens designs in a fully automatic fashion. I discuss the HPC aspects of the problem of searching for minima of high dimensionality functions.

A case for extended dark matter halos in dwarf spheroidal galaxies

We explore the recently proposed idea that the Galactic dwarf spheroidal galaxies are significantly (by 2 orders of magnitude) more massive than the conventional mass estimates of ∼10 M . In the larger mass case, the observed distribution of stars in these galaxies should have been entirely shaped by internal processes (formation and dynamic relaxation of stars in the potential of the dark matter halo), and not by the Galactic tidal field. We carried out numerical n-body simulations aimed at testing this scenario. Observed properties of three Galactic dwarf spheroidal galaxies were found to be consistent with our model. From our analysis, these dwarfs appear to be massive enough to alleviate the “missing satellites” problem of cold dark matter cosmologies.