2015ApJ...799..187K

Understanding the late-stage evolution of the most massive stars such as η Carinae is challenging because no true analogs of η Car have been clearly identified in the Milky Way or other galaxies. In Khan et al., we utilized Spitzer IRAC images of 7 nearby ( ≲ 4 Mpc) galaxies to search for such analogs, and found 34 candidates with flat or red mid-IR spectral energy distributions. Here, in Paper II, we present our characterization of these candidates using multi-wavelength data from the optical through the far-IR. Our search detected no true analogs of η Car, which implies an eruption rate that is a fraction 0.01 ≲ F ≲ 0.19 of the core-collapse supernova (ccSN) rate. This is roughly consistent with each M_ZAMS ≳ 70 M_☉ star undergoing one or two outbursts in its lifetime. However, we do identify a significant population of 18 lower luminosity (log (L/L_☉) ≃ 5.5-6.0) dusty stars. Stars enter this phase at a rate that is a fraction 0.09 ≲ F ≲ 0.55 of the ccSN rate, and this is consistent with all 25 < M_ZAMS< 60 M_☉ stars undergoing an obscured phase at most lasting a few thousand years once or twice. These phases constitute a negligible fraction of post-main-sequence lifetimes of massive stars, which implies that these events are likely to be associated with special periods in the evolution of the stars. The mass of the obscuring material is of order ∼M_☉, and we simply do not find enough heavily obscured stars for theses phases to represent more than a modest fraction (∼10% not ∼50%) of the total mass lost by these stars. In the long term, the sources that we identified will be prime candidates for detailed physical analysis with the James Webb Space Telescope.