We measure the star formation efficiency (SFE), the star formation rate per unit gas, in 23 nearby galaxies and compare it to expectations from proposed star formation laws and thresholds. We use H I maps from THINGS and derive H 2 maps from CO measured by HERACLES and BIMA SONG. We estimate the star formation rate by combining GALEX FUV maps and SINGS 24µm maps, infer stellar surface density profiles from SINGS 3.6µm data, and use kinematics from THINGS. We measure the SFE as a function of: the free-fall and orbital timescales; midplane gas pressure; stability of the gas disk to collapse (including the effects of stars); the ability of perturbations to grow despite shear; and the ability of a cold phase to form. In spirals, the SFE of H 2 alone is nearly constant at 5.25 ± 2.5 × 10 −10 yr −1 (equivalent to an H 2 depletion time of 1.9 × 10 9 yr) as a function of all of these variables at our 800 pc resolution. Where the ISM is mostly H I, on the other hand, the SFE decreases with increasing radius in both spiral and dwarf galaxies, a decline reasonably described by an exponential with scale length 0.2-0.25 r 25 . We interpret this decline as a strong dependence of GMC formation on environment. The ratio of molecular to atomic gas appears to be a smooth function of radius, stellar surface density, and pressure spanning from the H 2 -dominated to H I-dominated ISM. The radial decline in SFE is too steep to be reproduced only by increases in the free-fall time or orbital time. Thresholds for large-scale instability suggest that our disks are stable or marginally stable and do not show a clear link to the declining SFE. We suggest that ISM physics below the scales that we observe -phase balance in the H I, H 2 formation and destruction, and stellar feedback -governs the formation of GMCs from H I.

We present a comprehensive analysis of the relationship between star formation rate surface density, Σ SFR , and gas surface density, Σ gas , at sub-kpc resolution in a sample of 18 nearby galaxies. We use high resolution H i data from THINGS, CO data from HERACLES and BIMA SONG, 24 µm data from the Spitzer Space Telescope, and UV data from GALEX. We target 7 spiral galaxies and 11 late-type/dwarf galaxies and investigate how the star formation law differs between the H 2 -dominated centers of spiral galaxies, their H i-dominated outskirts and the H i-rich late-type/dwarf galaxies. We find that a Schmidt-type power law with index N = 1.0 ± 0.2 relates Σ SFR and Σ H2 across our sample of spiral galaxies, i.e., that H 2 forms stars at a constant efficiency in spirals. The average molecular gas depletion time is ∼ 2 · 10 9 years. The range of Σ H2 over which we measure this relation is ∼ 3 − 50 M ⊙ pc −2 , significantly lower than in starburst environments. We find the same results when performing a pixel-by-pixel analysis, averaging in radial bins, or when varying the star formation tracer used. We interpret the linear relation and constant depletion time as evidence that stars are forming in GMCs with approximately uniform properties and that Σ H2 may be more a measure of the filling fraction of giant molecular clouds than changing conditions in the molecular gas. The relationship between total gas surface density (Σ gas ) and Σ SFR varies dramatically among and within spiral galaxies. Most galaxies show little or no correlation between Σ HI and Σ SFR . As a result, the star formation efficiency (SFE), Σ SFR /Σ gas , varies strongly across our sample and within individual galaxies. We show that this variation is systematic and consistent with the SFE being set by local environmental factors: in spirals the SFE is a clear function of radius, while the dwarf galaxies in our sample display SFEs similar to those found in the outer optical disks of the spirals. We attribute the similarity to common environments (low-density, low-metallicity, H i-dominated) and argue that shear (which is typically absent in dwarfs) cannot drive the SFE. In addition to a molecular Schmidt law, the other general feature of our sample is a sharp saturation of H i surface densities at Σ HI ≈ 9 M ⊙ pc −2 in both the spiral and dwarf galaxies. In the case of the spirals, we observe gas in excess of this limit to be molecular.

No abstract

We present rotation curves of 19 galaxies from THINGS, The H I Nearby Galaxy Survey. The high spatial and velocity resolution of THINGS make these the highest quality H I rotation curves available to date for a large sample of nearby galaxies, spanning a wide range of H I masses and luminosities. The high quality of the data allows us to derive the geometrical and dynamical parameters using H I data alone. We do not find any declining rotation curves unambiguously associated with a cut-off in the mass distribution out to the last measured point. The rotation curves are combined with 3.6 µm data from SINGS (Spitzer Infrared Nearby Galaxies Survey) to construct mass models. Our best-fit, dynamical disk masses, derived from the rotation curves, are in good agreement with photometric disk masses derived from the 3.6 µm images in combination with stellar population synthesis arguments and two different assumptions for the stellar Initial Mass Function (IMF). We test the Cold Dark Matter-motivated cusp model, and the observationally motivated central density core model and find that (independent of IMF) for massive, disk-dominated galaxies, all halo models fit apparently equally well; for low-mass galaxies, however, a core-dominated halo is clearly preferred over a cuspy halo. The empirically derived densities of the dark matter halos of the late-type galaxies in our sample are half of what is predicted by CDM simulations, again independent of the assumed IMF. Stellar mass-to-light ratio trendsAn overview of the derived Υ 3.6 ⋆ values of the main disk components of our sample galaxies is given in Fig. 59. Here we plot the photometric (fixed) and dynamical (free) values for Υ 3.6 ⋆ against the colors and luminosities of the galaxies. The left-hand panels show the fixed and free Υ 3.6 ⋆ values of the disks for both the ISO and NFW model plotted against (J − K) color. We also distinghuish between our two choices for the IMF. Also shown is the predicted relation between Υ 3.6 ⋆ and color (Eqs. 4 and 5), again for both the diet-Salpeter and the Kroupa IMFs. By definition, the photometric Υ 3.6 ⋆ values follow their respective relations. The small scatter is caused by the color gradients present in these galaxies which make the photometric Υ 3.6 ⋆ differ slightly from those with a constant Υ 3.6 ⋆ as a function of radius.In general, we see that the diet-Salpeter curve defines an approximate upper limit to the distribution of the majority of the best fit values (apart from a few obviously discrepant cases). Accepting the best fit (free) Υ 3.6 ⋆ values at face value, this would suggest that a diet-Salpeter Υ 3.6 ⋆ analysis overestimates the disk masses slightly. Indeed, the Kroupa fixed Υ 3.6 ⋆ values seem to be a better match to the free Υ 3.6 ⋆ values. Alternative explanations could be effects of star formation history, or unexpectedly large contamination by PAHs, AGBs or hot dust in the 3.6 µm maps, though it is likely that the IMF plays the dominant role. Future rigorous population synthesis modeling should shed light on s...