Starburst galaxies are the subject of separate talks in this conference (cf. talks by Dahlem and by Dettmar), so I will briefly review the X-ray evidence, mostly at large scale. The data are fast accumulating but they have not been properly digested yet, and most of the results are still rather qualitative and based on a few examples only.
A large fraction of the total emission of starburst galaxies in the soft X-ray band is due to the hot ISM. The fraction varies in different galaxies, with > 20% in NGC 3256 (Moran et al. 1999; Lira et al 2002), to ~ 50% in the Antennae [Fabbiano et al. 2001], to ~ 80% in M82 [Zezas et al. 2001]. Hot gas is found mainly in three regions: (A) the large scale distribution is associated with extraplanar emission, in the galaxy's halo, usually associated with evidence of halo emission at other wavelengths (HI, radio continuum, H outflows). The emission can be traced to quite large distances: the rather spectacular halo of NGC 253 observed with XMM-Newton in the very soft energy band (Fig. 2) fills a very large region above and below the plane. The halo is usually fueled/fed by a nuclear/galactic outflow/wind. Often "horns" or X-shaped morphologies are observed in galaxies viewed edge-on. A much finer filamentary structure is also observed in halos, e.g. in NGC 3079's halo [Breitschwerdt et al. 2002] with XMM-Newton data and in NGC 253 itself with Chandra data, with spatial fluctuations of ~ 130 pc in size [Strickland et al. 2002a]. In both cases, these point to a close relation with similar structures in the line emission (cf. above and Fig. 1). (B) The disk appears to contain a multi-phase hot ISM. To use the case of NGC 253 again, the data require two temperatures (0.15; 0.5 kT) with no additional absorption and show unquestionable signs of plasma lines, e.g. OVII, FeXVII, that point to emission from a hot ionized medium [Pietsch et al. 2001]. It is possible that the multi-temperature of the gas in the disk is to be attributed in part to contamination from emission at higher galactic latitudes projected onto the disk. (C) Possibly the stronger emission is associated with the nuclear/circumnuclear regions, from which plumes and outflows are likely to be fueled. These regions appear to be hotter than the rest of the disk (e.g. kT ~ 6 keV is found in the inner regions of NGC 253 [again!], pietschpietsch), and might contain a higher metal fraction than the outer regions.
Figure 2. In this figure, the X-ray halo of NGC 253 in the soft (0.2-0.5 keV) energy band observed with XMM-Newton (courtesy of W. Pietsch) is shown, with the hard emission overlayed as contours. The oval shape indicates the approximate size and orientation of the optical disk. A few point sources are evident in the halo region, otherwise the area above and below the plane appears uniformly filled at these energies.
A much larger and better studied sample of objects needs to be examined for more quantitative analysis and a broader range of parameters need to be explored before a classification of the X-ray properties of the hot ISM in these galaxies can be finalized.
Figure 3. XMM-Newton data of NGC 3079 in the soft band indicate a large scale emission above/below the plane of the galaxy, as defined by the optical image. The filamentary structure of the emission is evident. Filaments are traced out to 13-17 kpc. Three main spurs plus a smaller one are responsible for the X-shaped morphology of the halo. From Breitschwerdt et al. 2002.