Resolving the Radio Source Background: Deeper Understanding Through Confusion

We used the Karl G. Jansky Very Large Array (VLA) to image one primary beam area at 3 GHz with 8 arcsec FWHM resolution and 1.0 microJy/beam rms noise near the pointing center. The P(D) distribution from the central 10 arcmin of this confusion-limited image constrains the count of discrete sources in the 1 < S(microJy/beam) < 10 range. At this level the brightness-weighted differential count S^2 n(S) is converging rapidly, as predicted by evolutionary models in which the faintest radio sources are star-forming galaxies; and ~96$% of the background originating in galaxies has been resolved into discrete sources. About 63% of the radio background is produced by AGNs, and the remaining 37% comes from star-forming galaxies that obey the far-infrared (FIR) / radio correlation and account for most of the FIR background at lambda = 160 microns. Our new data confirm that radio sources powered by AGNs and star formation evolve at about the same rate, a result consistent with AGN feedback and the rough correlation of black hole and bulge stellar masses. The confusion at centimeter wavelengths is low enough that neither the planned SKA nor its pathfinder ASKAP EMU survey should be confusion limited, and the ultimate source detection limit imposed by "natural" confusion is < 0.01 microJy at 1.4 GHz. If discrete sources dominate the bright extragalactic background reported by ARCADE2 at 3.3 GHz, they cannot be located in or near galaxies and most are < 0.03 microJy at 1.4 GHz.