Annu. Rev. Astron. Astrophys. 2001. 39: 249-307 Copyright © 2001 by Annual Reviews. All rights reserved

4.3. The Radio-Infrared Background Connection

IRAS data revealed a remarkable correlation between the radio and far-infrared fluxes of galaxies, stretching over four orders of magnitude in infrared flux density (Helou 1991). The correlation holds separately for the thermal and synchrotron components of the radio emission as well as with their sum (Price & Duric 1992). In hindsight, this correlation should not be surprising because the two fluxes arise from the formation of massive stars (Lisenfeld et al. 1996). These stars do most of the heating of the dust that emits the far-infrared radiation. They also form H II regions, which dominate the radio thermal emission, and, after their explosive death, accelerate the high-energy particles that produce the synchrotron emission. The correlation does not hold for active galactic nuclei (AGN)-dominated galaxies, in which the radio emission is not associated with massive star formation.

Haarsma & Partridge (1998) used the radio-infrared correlation to estimate the contribution of star-forming galaxies to the cosmic radio background (CRB). Assuming that the sources of the CIB are predominantly star-forming galaxies (Section 5.1), and that the local radio-infrared relation can be extended to all star-forming galaxies, they found that star-forming galaxies account for about 50% of the CRB at 170 cm, exclusive of the CMB contribution. They also showed that the integrated flux from discrete radio sources associated with AGN is about half the CRB at 75 cm, providing a consistent model for the origin of the CRB.

An important implication is that the cosmic star formation history can be constructed from surveys of faint radio sources, avoiding the need to correct for the effects of dust extinction. Haarsma et al. (2000) used this technique to study the star formation rate in the range z = 0-1.5, confirming the rapid rise from z = 0 to 1, but finding rates considerably higher than extinction-corrected rates based on optical data.