(Abridged) Spitzer data at 24, 70, and 160 micron and ground-based H-alpha images are analyzed for a sample of 189 nearby star-forming and starburst galaxies to investigate whether reliable star formation rate (SFR) indicators can be defined using the monochromatic infrared dust emission centered at 70 and 160 micron. We compare recently published recipes for SFR measures using combinations of the 24 micron and observed H-alpha luminosities with those using 24 micron luminosity alone. From these comparisons, we derive a reference SFR indicator for use in our analysis. Linear correlations between SFR and the 70 and 160 micron luminosity are found for L(70)>=1.4x10^{42} erg/s and L(160)>=2x10^{42} erg/s, corresponding to SFR>=0.1-0.3 M_sun/yr. Below those two luminosity limits, the relation between SFR and 70 micron (160 micron) luminosity is non-linear and SFR calibrations become problematic. The dispersion of the data around the mean trend increases for increasing wavelength, becoming about 25% (factor ~2) larger at 70 (160) micron than at 24 micron. The increasing dispersion is likely an effect of the increasing contribution to the infrared emission of dust heated by stellar populations not associated with the current star formation. The non-linear relation between SFR and the 70 and 160 micron emission at faint galaxy luminosities suggests that the increasing transparency of the interstellar medium, decreasing effective dust temperature, and decreasing filling factor of star forming regions across the galaxy become important factors for decreasing luminosity. The SFR calibrations are provided for galaxies with oxygen abundance 12+Log(O/H)>8.1. At lower metallicity the infrared luminosity no longer reliably traces the SFR because galaxies are less dusty and more transparent.
Download