Before going into details of the properties of diffuse X-ray emission of dwarf galaxies, it may be fruitful to check a few of the assumptions and theoretical predictions in smaller, better controlled, nearby systems, like the superbubbles in the Magellanic Clouds. Fig. 1 shows an H image of the superbubble N51D in the Large Magellanic Cloud (LMC). The structure is a fairly typical example of its class (Meaburn 1980). It shows a complete shell of ionized gas with a diameter of about 100 pc centered on one large OB association. Such a superbubble distinguishes itself physically from a (~ 1 - 50 pc) bubble around a single star (Weis & Duschl 1998) and a kpc-sized supergiant shell (sometimes also called supershell) (Meaburn 1980; Chu 1995) around a large complex of OB associations.
Figure 1. H image of the LMC superbubble N51D with contours of the ROSAT soft and medium band X-ray emission overplotted.
N51D expands with a velocity of ~ 35 km s-1 (Lasker 1980; de Boer & Nash 1982) into the surrounding medium. Despite of some larger loop-like extensions of the H shell to the north and the south-west (as visible in Fig. 1), this superbubble appears to be still a closed volume. The difference in surface brightness between the eastern and western section is partly due a different density of the surrounding gas into which the superbubble expands. The gas is ionized by the OB association inside, which contains one the UV brightest stars in the LMC, the WC5+O8Iaf binary Sk-67 104 (= HD 36402). A detailed study of the OB association (Oey & Smedley 1998) yields an age of 3 Myr and determines from the observed stellar content the total energy input of the stars to the superbubble. N51D emits diffuse X-ray emission (Chu & Mac Low 1990) which is brighter near the eastern rim of the superbubble. Fig. 1 shows the H image of N51D overlayed with contours of the X-ray emission derived from an ROSAT PSPC archival data (Bomans et al. 2001a, in prep). While some X-ray emission inside N51D seems to stem from point sources (most probably X-ray binaries), diffuse emission is clearly present inside the cavity delineated by the H shell. The ROSAT data confirm the surface brightness enhancement at the bright, probably denser, eastern boundary of N51D first detected using EINSTEIN (Chu & Mac Low 1990), but also show considerable substructure here, not following one-to-one the H emission (Bomans et al. 2001a, in prep.). The temperature of the hot gas is about 5 × 106 K, when fitting spectra with collisional equilibrium models (e.g. Raymond & Smith 1977). The X-ray surface brightness and luminosity of N51D does not fit well to the predictions of the Weaver et al. (1977) model of an expanding superbubble and require an additional source of X-ray emission, which can be provided by supernova explosions of massive stars in the OB association, meaning inside the superbubble. As soon as the shock wave hits the dense shell wall, hot gas with high density and therefore surface brightness is produced, a process we may well observe right now at the eastern inner shell wall (Chu & Mac Low 1990, Bomans et al. 2001a, in prep.). It is interesting to note here, that the loop at the south-western edge of N51D is a separate diffuse X-ray region, which is tempting to be identified with a beginning outflow, similar to the one in the LMC superbubble N44 (Chu et al. 1993). With the availability of an HI synthesis map of the LMC (Kim et al. 1998), even some informations about the surrounding neutral gas density and topology can be derived. Therefore in the case of N51D all critical parameters for the evolutions of a bubble can be observationally determined or at least estimated.
The crucial determination of the metallicity of the diffuse hot gas is still missing. The current information on superbubbles does not give reliable metallicities or how metal-enriched and more metal-poor gas is distributed in the bubble and along the shell walls, much less than giving details on the actual mixing processes. ASCA observations of a sample of LMC supernova remnants prove at least the first assumption: the hot gas is indeed metal-enriched by the supernova explosion, but it shows also that mixing of hot metal-rich and cool metal-poor gas starts right away (Hughes et al. 1998).