Radio properties of young SNe/SNRs in Arp220

Paper in proceeding, 2007

Radio observations of young radio supernovae (SNe) and supernova remnants (SNRs) can provide a unique window on the stellar/ISM properties of starburst galaxies. Such observations can potentially constrain issues of cosmological importance such as whether stellar IMFs are radically different in extreme star-forming environments. Recently published observations of the nearest ultra-luminous infra-red galaxy Arp 220 have revealed the radio spectra of a group of SNe/SNR. About half of the sources detected at high frequency have spectral and variability properties consistent with young Type IIn supernovae interacting with their pre-explosion stellar winds. However the high rate of appearance of these sources implies that an unusually large fraction of core-collapse supernovae are highly luminous, which might be at least partly explained by a top heavy IMF. The other half of the compact sources found in Arp 220 were interpreted as SNRs interacting with a dense (104 to 105 cm−3) ISM. In this paper we report on new more sensitive VLBI observations at wavelengths of 2 cm and 3.6 cm. We find that the spectral evolution of most of the suspected SNe sources is consistent with them being of Type IIn. However two rapidly dimming objects may instead be of Type Ib/c or IIb. Most of the long-lived candidate SNR sources show small or undetectable flux density variations, however one almost doubled its 3.6 cm intensity in 11 months. Another source also shows some variability and a complex spectrum. These two objects plus another with a flat spectrum up to 2 cm are the best candidates for an AGN core, though the data does not yet require this interpretation. At least three sources show signs of resolution with diameters in the range 0.1 to 0.2 pc. These sizes put them sightly above, but still consistent with, the size-luminosity correlation for SNRs. The SKA will have sufficient sensitivity to detect the emission from Arp220-like compact sources out to cosmological distances (i.e. up to z≈ 0.5). However the SKA needs global baselines both to separate out the discrete sources from more extended radio emission and to resolve them apart.