Dynamical triggering of starbursts

Galaxy interactions/mergers, gravitational instabilities and density waves, such as bars, are frequently invoked to trigger starbursts. These mechanisms have been explored through numerical simulations, with the help of various star formation recipes. Gravitational instabilities are necessary to initiate star formation, but the main trigger might be the gas flows, to provide sufficient fuel in a short time-scale. Gas accretion is also acting on the dynamics, in favoring bars/spirals, which will drive the gas inwards. Large amounts of external gas accretion are required to explain the bar frequency, and this accretion rate can be provided by the cosmic filaments, as supported by cosmological simulations. Subsequent interactions can then trigger starbursts by driving this accreted gas inwards. OBSERVATIONAL EVIDENCE OF DYNAMICAL TRIGGERING It is now widely accepted that strong galaxy interactions can produce violent starbursts, and ultra-luminous infra-red galaxies (ULIRGs) are all mergers (e.g. Sanders & Mirabel 1996). But interacting galaxies in general do not show intense starbursts (Bergvall et al 2003), or only in their centers. The tidal parameter, depending on mass and distance of companions, is not correlated to any star formation tracer. Starbursts are rare in the local universe, and much rarer than galaxy interactions. The only perceptible effect is that interacting galaxies have more mass gas, dust & young stars, and are more concentrated to the center (Bergvall et al 2003). One of the main conclusions that can be drawn for giant galaxy starbursts is that galaxy interactions is a necessary condition to trigger them, but not sufficient. Certainly, other parameters have to be taken into account, such as the gas content, the geometry/velocity of the encounter, and also the phases of the interaction, the starburst occuring only during about 10% of the interaction time, according to simulations (e.g. Mihos & Hernquist 1996). For dwarf galaxies, interactions are even not a necessary condition for triggering: starbursting dwarfs have no more companions (Brosch et al 2004), although it might be different for Blue Compact Dwarfs (BCD, Hunter & Elmegreen 2004). Dwarf galaxies with starbursts present asymmetries, and a possible trigger could just come from gas accretion: the gas infalling in dark-matter haloes may experience sloshing and oscillations favoring compression and instabilities. This mechanism might be relevant for primordial galaxies (Brosch et al 2004). One of the best evidence of triggering can be obtained by tracing the fossil records of the star formation history in nearby galaxies. In the Small Magellanic Cloud, the number of stars at a given age has been determined by Zaritsky & Harris (2004). They noticed that the derived SMC star formation history reveals some bursts corresponding to pericenters of its orbit around the Milky Way. A model taking the tidal triggering FIGURE 1. Star formation history in the SMC derived from a tidally triggered model (dash-line), or a tidally triggered model with infall (solid line), compared to the observed number of stars at each age, from Zaritsky & Harris (2004). The insert is a zoom of the recent 2.5 Gyr. into account fits better the observations, although the quantitative influence of the tide is uncertain, between 10 and 70% of the star formation could be tidal in origin, cf figure 1. One of the problem in the fit comes from the too large increase in star formation about 3-4 Gyrs ago, which can only be explained by external gas accretion. A significant gas infall, by about 50% of the total gas mass, improves the fit, and accounts also for the observed age-metallicity relation. The metallicity is not monotonously increasing with decreasing age, but remains constant for old stars until 3 Gyrs, indicating that dilution is required from deficient gas infall.