The star formation history of the universe: An infrared perspective

A simple and versatile parameterized approach to star formation history allows a quantitative investigation of constraints from far-infrared and submillimeter counts and background intensity measurements. The models include four spectral components: infrared cirrus (emission from interstellar dust), an M82-like starburst, an Arp 220-like starburst, and an active galactic nucleus (AGN) dust torus. The 60 km luminosity function is determined for each chosen rate of evolution using the Point Source Catalog Redshift Survey (PSCz) redshift data for 15,000 galaxies. The proportions of each spectral type as a function of 60 mum luminosity are chosen for consistency with IRAS and SCUBA color-luminosity relations, and with the fraction of AGNs as a function of luminosity found in 12 mum samples. The luminosity function for each component at any wavelength can then be calculated from the assumed spectral energy distributions (SEDs). With assumptions about the optical SEDs corresponding to each component and, for the AGN component, an assumed dependence of the dust covering factor on luminosity, the optical and near-infrared counts can be accurately modeled. High- and low-mass stars are treated separately, since the former will trace the rate of star formation, while the latter trace the cumulative integral of the star formation rate. A good fit to the observed counts at 0.44, 2.2, 15, 60, 90, 175, and 850 mum can be found with pure luminosity evolution in all three cosmological models investigated: Ohm (0) = 1, Ohm (0) = 0.3 (Lambda = 0), and Ohm (0) = 0.3, Lambda = 0.7. All three models also give an acceptable fit to the integrated background spectrum. Selected predictions of the models, for example redshift distributions for each component at selected wavelengths and fluxes, are shown. The effect of including an element of density evolution is also investigated. The total mass-density of stars generated is consistent with that observed, in all three cosmological models.