The under-abundance of very massive galaxies in the universe is frequently attributed to the effect of galactic winds. Although ionized galactic winds are readily observable most of the expelled mass is likely in cooler atomic and molecular phases. E
xpanding molecular shells observed in starburst systems such as NGC 253 and M 82 may facilitate the entrainment of molecular gas in the wind. While shell properties are well constrained, determining the amount of outflowing gas emerging from such shells and the connection between this gas and the ionized wind requires spatial resolution <100 pc coupled with sensitivity to a wide range of spatial scales, hitherto not available. Here we report observations of NGC 253, a nearby starburst galaxy (D~3.4 Mpc) known to possess a wind, which trace the cool molecular wind at 50 pc resolution. At this resolution the extraplanar molecular gas closely tracks the H{alpha} filaments, and it appears connected to molecular expanding shells located in the starburst region. These observations allow us to directly measure the molecular outflow rate to be > 3 Msun/yr and likely ~9 Msun/yr. This implies a ratio of mass-outflow rate to star formation rate of at least {eta}~1-3, establishing the importance of the starburst-driven wind in limiting the star formation activity and the final stellar content.
We compare atomic gas, molecular gas, and the recent star formation rate (SFR) inferred from H-alpha in the Small Magellanic Cloud (SMC). By using infrared dust emission and local dust-to-gas ratios, we construct a map of molecular gas that is indepe
ndent of CO emission. This allows us to disentangle conversion factor effects from the impact of metallicity on the formation and star formation efficiency of molecular gas. On scales of 200 pc to 1 kpc we find a characteristic molecular gas depletion time of ~1.6 Gyr, similar to that observed in the molecule-rich parts of large spiral galaxies on similar spatial scales. This depletion time shortens on much larger scales to ~0.6 Gyr because of the presence of a diffuse H-alpha component, and lengthens on much smaller scales to ~7.5 Gyr because the H-alpha and H2 distributions differ in detail. We estimate the systematic uncertainties in our measurement to be a factor of 2-3. We suggest that the impact of metallicity on the physics of star formation in molecular gas has at most this magnitude. The relation between SFR and neutral (H2+HI) gas surface density is steep, with a power-law index ~2.2+/-0.1, similar to that observed in the outer disks of large spiral galaxies. At a fixed total gas surface density the SMC has a 5-10 times lower molecular gas fraction (and star formation rate) than large spiral galaxies. We explore the ability of the recent models by Krumholz et al. (2009) and Ostriker et al. (2010) to reproduce our observations. We find that to explain our data at all spatial scales requires a low fraction of cold, gravitationally-bound gas in the SMC. We explore a combined model that incorporates both large scale thermal and dynamical equilibrium and cloud-scale photodissociation region structure and find that it reproduces our data well, as well as predicting a fraction of cold atomic gas very similar to that observed in the SMC.
We use high spatial resolution observations of CO to systematically measure the resolved size-line width, luminosity-line width, luminosity-size, and the mass-luminosity relations of Giant Molecular Clouds (GMCs) in a variety of extragalactic systems
. Although the data are heterogeneous we analyze them in a consistent manner to remove the biases introduced by limited sensitivity and resolution, thus obtaining reliable sizes, velocity dispersions, and luminosities. We compare the results obtained in dwarf galaxies with those from the Local Group spiral galaxies. We find that extragalactic GMC properties measured across a wide range of environments are very much compatible with those in the Galaxy. We use these results to investigate metallicity trends in the cloud average column density and virial CO-to-H2 factor. We find that these measurements do not accord with simple predictions from photoionization-regulated star formation theory, although this could be due to the fact that we do not sample small enough spatial scales or the full gravitational potential of the molecular cloud. We also find that the virial CO-to-H2 conversion factor in CO-bright GMCs is very similar to Galactic, and that the excursions do not show a measurable metallicity trend. We contrast these results with estimates of molecular mass based on far-infrared measurements obtained for the Small Magellanic Cloud, which systematically yield larger masses, and interpret this discrepancy as arising from large H2 envelopes that surround the CO-bright cores. We conclude that GMCs identified on the basis of their CO emission are a unique class of object that exhibit a remarkably uniform set of properties from galaxy to galaxy (abridged).
