The implications of large dust masses at high redshifts: a first look at galactic evolution in the submillimetre waveband

Abstract

We show that despite the large uncertainty in the absolute dust masses of galaxies, it is possible to calculate the ratio of the dust masses of galaxies at different redshifts with acceptable accuracy as long as the dust masses are estimated from submillimetre fluxes measured at the same rest-frame wavelength. If one assumes that the same fraction of the metals in a galaxy is always incorporated in dust, for which there is some observational evidence, then it is possible to use standard chemical evolution models to estimate how the mass of dust in a galaxy should change with time. Our models show that dust mass is a poor 'galactic clock', because the variation with time depends critically on whether galactic evolution is best represented by a closed-box, inflow or outflow model. Nevertheless, the high dust masses of the high-redshift objects from which submillimetre dust emission has recently been observed do place significant limits on galactic evolution/cosmology. In an no = 1 universe it is possible to reproduce the high dust masses only with the outflow models and with the high-redshift objects being the ancestors of the most massive galaxies in the local Universe. If we live in a low-density universe, it is not possible to reproduce the dust masses with our dust-evolution models, and either one or more of our assumptions must be radically wrong, or gravitational lensing must be amplifying the submillimetre fluxes, and hence the dust masses, or there must be a component of dust in galaxies which is hot at high redshift but which in nearby galaxies is so cold that its presence cannot be detected even at submillimetre wavelengths. We also show that submillimetre observations of galaxies at both high and low redshifts are beginning to put interesting limits on the processes that form dust.