The stellar masses and specific star-formation rates of submillimetre galaxies

Establishing the stellar masses, and hence specific star-formation rates of submillimetre galaxies is crucial for determining the role of such objects in the cosmic history of galaxy/star formation. However, there is as yet no consensus over the typical stellar masses of submillimetre galaxies, as illustrated by the widely differing results reported from recent optical-infrared studies of submillimetre galaxies with spectroscopic redshifts z ≃ 2-3. Specifically, even for the same set of submillimetre galaxies, the reported average stellar masses have ranged over an order of magnitude, from ≃5 × 1010 M&sun; to ≃5 × 1011 M&sun;. Here we study how different methods of analysis can lead to such widely varying results. We find that, contrary to recent claims in the literature, potential contamination of IRAC 3-8 μm photometry from hot dust associated with an active nucleus is not the origin of the published discrepancies in derived stellar masses. Instead, we expose in detail how inferred stellar mass depends on assumptions made in the photometric fitting, and quantify the individual and cumulative effects of different choices of initial mass function, different "brands" of evolutionary synthesis models, and different forms of assumed star-formation history. We review current observational evidence for and against these alternatives as well as clues from the hydrodynamical simulations, and conclude that, for the most justifiable choices of these model inputs, the average stellar mass of luminous (S850 ≳ 5 mJy) submillimetre galaxies is ≃2 × 1011 M&sun; to within a factor ≃2. We also check and confirm that this number is perfectly reasonable in the light of the latest measurements of the dynamical masses of these objects (≃2-6 × 1011 M&sun; from CO (1-0) observations), and the evolving stellar mass function of the overall galaxy population. Galaxy stellar masses of this order imply that the average specific star-formation rate of submillimetre galaxies is comparable to that of other star-forming galaxies at z > 2, at 2-3 Gyr-1. This supports the view that, while rare outliers may be found at any stellar mass, most submillimetre galaxies simply form the top end of the "main-sequence" of star-forming galaxies at these redshifts. Conversely, this argues strongly against the viewpoint, frequently simply asserted in the literature, that submillimetre galaxies are extreme pathological objects, of little relevance in the cosmic history of star-formation. Appendix A is available in electronic form at http://www.aanda.org