Variation in the dust emissivity index across M33 with Herschel and Spitzer (HerM 33es)

We study the wavelength dependence of the dust emission as a function of position and environment across the disk of M33 using Spitzer and Herschel photometric data. M33 is a Local Group spiral with slightly subsolar metallicity, which makes it an ideal stepping-stone to less regular and lower-metallicity objects such as dwarf galaxies and, probably, young-universe objects. Expressing the emissivity of the dust as a power law, the power-law exponent (beta) was estimated from two independent approaches designed to properly treat the degeneracy between beta and the dust temperature (T). Both beta and T are higher in the inner than in the outer disk, contrary to reported beta - T anti-correlations found in other sources. In the cold + warm dust model, the warm component and the ionized gas (H alpha) have a very similar distribution across the galaxy, demonstrating that the model separates the components in an appropriate way. Both cold-and warm-dust column densities are high in star-forming regions and reach their maxima toward the giant star-forming complexes NGC 604 and NGC595. beta declines from close to 2 in the center to about 1.3 in the outer disk. beta is positively correlated with star formation and with the molecular gas column, as traced by the H alpha and CO emission. The lower dust-emissivity index in the outer parts of M33 is most likely related to the reduced metallicity (different grain composition) and possibly to a different size distribution. It is not due to the decrease in stellar radiation field or temperature in a simple way because the far-infrared-bright regions in the outer disk also have a low beta. Like most spirals, M33 has a (decreasing) radial gradient in star formation and molecular-to-atomic gas ratio such that the regions bright in H alpha or CO tend to trace the inner disk, which makes it difficult to distinguish between their effects on the dust. The assumption of a constant emissivity index beta is obviously not appropriate.