The FIR SEDs of z~2 galaxies: evidence for scaled up cool galaxies

Last week we talked about a paper that calibrated the relationship between the observed 24um (= 8um in the restframe) flux and SFR for z~2 galaxies. This paper by Muzzin et al. looks at a similar issue, but takes a slightly different approach.

The Chary & Elbaz templates describe the infrared SEDs of local galaxies. These templates have a luminosity dependence, such that IR-luminous galaxies have templates that peak at shorter wavelengths, due to increased dust temperatures.

So one method of estimating the SFR of a galaxy is to see which template corresponds to the observed restframe 8um luminosity, assume that the template accurately describes the SED throughout the rest of the IR regime, and to convert the total IR luminosity of the template into an SFR.

The red dashed curves in the SEDs shown above illustrate the procedure. But Muzzin et al. have compared the templates to the observed fluxes at different IR wavelengths for these two galaxies, and obviously the templates don't work very well. But the solid curves, which are templates for low-luminosity galaxies that have been scaled up to match the data points, do provide good fits. This shows that the correlation between IR luminosity and dust temperature that we see in the local universe doesn't work at higher redshifts, and that even IR-luminous galaxies at z~2 can have relatively cool dust. Previous studies that relied on this correlation may have overestimated the SFR by a factor of several.

Muzzin et al. go on to show general agreement between the SFRs as derived from the IR luminosity and from the dust-corrected Halpha luminosity (where the extinction in Halpha is taken to be roughly twice the extinction to the stellar continuum, which in turn is estimated using stellar population synthesis modeling).