Hierarchical models of structure formation suggest that dwarf galaxies are the building blocks of larger galaxies, merging at high redshifts to form the today's distribution of galaxies (e.g. Padmanabhan 1993 and references therein; Tegmark et al. 1997). It is therefore crucial to understand both their formation and evolution. Besides the importance for the formation of spiral and elliptical galaxies, dwarf galaxies are ideal targets for investigation of the interrelations between galaxies and the contents and dynamics of the intergalactic medium, since dwarfs are known to support strong outflows via efficient and fast starburst events with accompanying supernova activity (cf. Heckman 1998 and references therein). The fact that the bolometric infrared luminosity, an indicator of starforming intensity, shows little correlation with galaxy rotation speed (an indicator of galaxy mass) over - of M(), suggests that early generations of premerged starbursting dwarfs had vigorous outflow phases, aided by their low escape velocities (cf. MacLow & Ferrara 1999).
The intergalactic medium is enriched with metals (e.g. Renzini 1997); it also contains large-scale magnetic fields, at least in and around clusters of galaxies (cf. Kronberg 1994). The metallicity enhancement of the intergalactic medium can be explained by galactic winds from starbursting dwarfs (Trentham 1994; Nath & Chiba 1995). Kronberg et al. (1999) have shown that the observed magnetization of the intergalactic medium (IGM), in principle, can also be caused by primeval starbursting dwarf galaxies. They used model calculations and present-day starburst galaxies to show that a significant fraction of the intergalactic medium can be filled with magnetic fields that originally emanated from the transport of magnetic flux into the IGM from primeval dwarf galaxies. Evidence for strong, magnetized outflows come from observations of M82 (e.g. Reuter et al. 1992; Shopbell & Bland-Hawthorn 1998), NGC 3079 (e.g. Duric et al. 1983), NGC 4631 (e.g. Hummel et al. 1988) and NGC 4666 (e.g. Dahlem et al. 1997), all of which have superwinds and bubbles as magnetohydrodynamical consequences of the mechanical energy deposited inside starbursts (cf. also Heckman 1998 and references therein).
The calculations of Kronberg et al. (1999) focused on the volume distribution in the IGM of magnetic flux, and did not attempt a quantitative prediction of its strength. Recently, Birk et al. (1999) have shown that in the rather complex filamentary outflow of M82 (cf. Reuter et al. 1992; McKeith et al. 1995) unstable Kelvin-Helmholtz (KH) modes can be excited and that these result in magnetic field regeneration. They concluded that these KH modes can be responsible for the relatively strong and ordered halo magnetic fields. The resulting fields show significant deviation from estimates based on flux conservation only. The amplification of magnetic fields in magnetized plasma flows due to a strong curling and twisting of field lines in magnetized boundary layers is a generic feature of KH modes (e.g. Clarke 1993; Frank et al. 1996).
In this contribution we focus on the role of KH instabilities in superwinds of primeval galaxies. We consider a situation which is characterized by the interaction of the galactic winds and the IGM. Whereas one should expect a huge number of KH resonant layers that are convected by the outflowing material and may interact with each other, as a first step, we will concentrate on one KH unstable boundary layer.
The winds and the IGM may be partially ionized , in particular, at very early epochs of the newly formed dwarf galaxies at which neutral gas is a large component of the galaxy's mass. Moreover, some of the expelled hot wind plasma cools and may fall back (probably into a neighbouring galaxy) as a mixture of ionized and neutral gas, thereby interacting with outflowing plasma in the process. This type of scenario is supported by observations of present day superwinds, that clearly indicate that neutral gas is a significant component of the material expelled from the galactic plane of starburst galaxies (e.g. Heckman et al. 1993; Weiß et al. 1999).
The consequence of this is that a two-fluid treatment of KH dynamics rather than a pure MHD description seems to be the appropriate approach to understanding the evolution of the field, and the energy equipartition. For recent detailed investigations of KH modes in the pure MHD regime see Frank et al. (1996) and Miniati et al. (1999) and references therein. In the next section we develop analytic estimates of the relevant time scales of the KH modes. These estimates are based on the analysis of the general dispersion relation of unstable KH modes in partially ionized plasmas. They take account of the different bulk velocities and mass densities of both the neutral and the ionized gas components on both sides of the KH interface. In Sect. 3 we present numerical simulations that model the KH dynamics in a self-consistent way.
© European Southern Observatory (ESO) 2000
Online publication: December 8, 1999