On the origin of the radio emission in IRAS galaxies with high and ultrahigh luminosity: the starburst-AGN controversy

In the scenario of a violent nuclear/circumnuclear starburst, the radio luminosity detected in high-luminosity (HLIRG) and ultraluminous (ULIRG) IRAS galaxies is explained as non-thermal radio emission generated in radio supernovae and their remnants, created in an ongoing massive starburst process. These supernovae and their remnants must have peak radio luminosities 100 to 1000 times the luminosity of Cas A, and thus be radio hypernovae. The presence of radio hypernovae in a circumnuclear starburst naturally explains the radio excess observed in these regions with respect to disc H II regions, and the FIR–radio luminosity relation. Considering the empirical radio luminosity curve of the well-monitored radio hypernova SN1979c, supernova rates $$\Upsilon_\text {SNII}\sim 0.7-3$$ and $$\sim 1.5-10\enspace\text {yr}^{-1}$$ are predicted for HLIRG and ULIRG, respectively. The observed far-infrared (FIR) luminosity, q parameter and emission-line luminosities are best accounted for by a starburst characterized by a Salpeter initial mass function (IMF), with upper mass limit $$M_\text u = 60\text M_\odot$$. If the lower mass limit $$M_1 = 3 M_\odot$$, the star formation rate corresponds to ∼ 40–150 and $$\sim 80-500\enspace\text M_\odot\enspace \text {yr}^{-1}$$ for HLIRG and ULIRG, respectively. The FIR spectral energy distribution of HLIRG and ULIRG differs from that of starburst galaxies and active galactic nuclei (AGN). It is consistent with blackbody dust reradiation at a lower temperature than in normal starbursts. Although there is evidence, at least in some sample galaxies like Mrk 231 or NGC 7469, for an AGN, the radio and hard X-ray luminosities indicate that this will not, in general, be a luminous QSO. Thus the central AGN will not be the main heating and ionizing source in these galaxies.