Shells, Bubbles and Holes: the Porosity of the Interstellar Medium in Galaxies
We present an analysis of the properties of HI holes detected in 20 galaxies that are part of “The HI Nearby Galaxy Survey” (THINGS). We detected more than 1000 holes in total in the sampled galaxies. Where they can be measured, their sizes range from about 100 pc (our resolution limit) to about 2 kpc, their expansion velocities range from 4 to 36 km/s, and their ages are estimated to range between 3 and 150 Myr. The holes are found throughout the discs of the galaxies, out to the edge of the HI disc; 23% of the holes fall outside R25. We find that shear limits the age of holes in spirals; shear is less important in dwarf galaxies which explains why HI holes in dwarfs are rounder, on average than in spirals. Shear, which is particularly strong in the inner part of spiral galaxies, also explains why we find that holes outside R25 are larger and older. We derive the scale height of the HI disc as a function of galactocentric radius and find that the disc flares at large radii in all galaxies. We proceed to derive the surface and volume porosity (Q2D and Q3D) and find that this correlates with the type of the host galaxy: later Hubble types tend to be more porous. The size distribution of the holes in our sample follows a power law with a slope of a=−2.9. Assuming that the holes are the result of massive star formation, we derive values for the supernova rate (SNR) and star formation rate (SFR) which scales with the SFR derived based on other tracers. If we extrapolate the observed number of holes to include those that fall below our resolution limit, down to holes created by a single supernova, we find that our results are compatible with the hypothesis that HI holes result from star formation. We use HI data from THINGS, 8μm, 24μm, 70μm and HI maps from SINGS, CO(2–1) data from HERACLES and FUV data from NGS to present a visual comparison of these maps with respect to the locations of HI holes. We find that the vast majority of HI holes are also prominent in the 8μm map and to some extent in the 24μm map. There is a lack of molecular gas from the interior of nearly all the holes, which is consistent with the idea that the latter are filled with hot gas. About 60% of young holes have FUV emission detected in their interiors highlighting the presence of the parent OB association. In addition, FUV is detected on the rims of some of the older HI holes, presumably due to the dispersion of the OB association with respect to the gas. We describe the development of a 2–D cross-correlation method to compare multi-wavelength maps in a quantitative way (quantified by Ccoef ) and give some first results from the application of this method to the nearby galaxy NGC2403. We find that the all the dust tracers are well correlated (Ccoef > 0.7) with the 8μm–24μm correlation being the highest (Ccoef > 0.88). Similarly all the star formation tracers are well linked as expected (Ccoef > 0.6). With respect to the relations between star formation and dust tracers we found that most are well matched (Ccoef > 0.7) as dust grains are heated by radiation in star forming regions. At smaller scales (15") FUV correlates poorly (Ccoef ~ 0.3) with the dust tracers, a direct consequence of the absorption of FUV photons by dust. We find that the HI is reasonably well correlated with the 8μm emission (Ccoef ~ 0.6) illustrating the fact that HI is mixed with PAH’s. Interestingly, the HI map shows some correlation with the SF map (Ccoef ~ 0.4) even though FUV and HI emissions were found to be completely uncorrelated (Ccoef ~ 0).