SIMBAD references

We have measured the near-infrared colors and the fluxes of individual pixels in 68 galaxies common to the Spitzer Infrared Nearby Galaxies Survey and the Large Galaxy Atlas Survey. Pixels from each galaxy are grouped into regions of increasingly red near-infrared colors. As expected, the majority of pixels are shown to have relatively constant NIR flux ratios (log₁₀ I_3.6/I_1.25= -0.30±0.07 and log₁₀I_4.5/I_3.6= -0.19±0.02), representing the blackbody continuum emission of main sequence stars. However, pixels with red NIR colors correspond to pixels with higher H_αemission and dust extinction. We show that the NIR colors are correlated to both quantities, with the strongest correlation to the intrinsic H_αemission. In addition, in regions of high star formation, the average intensity of pixels in red-excess regions (at 1.25 µm, 3.6 µm, 4.5 µm, 5.6 µm, 8.0 µm and 24 µm) scales linearly with the intrinsic intensity of Hα emission, and thus with the star formation rate (SFR) within the pixel. This suggests that most NIR-excess regions are not red because their light is being depleted by absorption. Instead, they are red because additional infrared light is being contributed by a process linked to star formation. This is surprising because the shorter wavelength bands in our study (1.25 µm-5.6 µm) do not probe emission from cold (10-20 K) and warm (50-100 K) dust associated with star formation in molecular clouds. However, emission from hot dust (700-1000 K) and/or polycyclic aromatic hydrocarbon (PAH) molecules can explain the additional emission seen at the shorter wavelengths in our study. The contribution from hot dust and/or PAH emission at 2 µm-5 µm and PAH emission at 5.6 µm and 8.0 µm scales linearly with warm dust emission at 24 µm and the intrinsic Hα emission. Since both are tied to the SFR, our analysis shows that the NIR excess continuum emission and PAH emission at ∼1-8 µm can be added to spectral energy distribution models in a very straightforward way, by simply adding an additional component to the models that scales linearly with SFR.