2024ApJ...966...61D

We present detailed spectral energy distribution (SED) modeling of 14 local ultraluminous infrared galaxies (ULIRGs) with outstanding photometric data from the literature covering the ultraviolet-infrared (FIR) and radio bands (∼50 MHz to ∼30 GHz). We employ the CIGALE SED fitting code to model the ultraviolet-FIR-radio SED. For the radio-only SED modeling, we use the UltraNest package, leveraging its nested sampling algorithm. Combining the results from our previous study on 11 luminous infrared galaxies (LIRGs), we discuss the global astrophysical properties of a sample of 25 starburst galaxies (z < 0.5). Their radio spectra are frequently characterized by bends and turnovers, with no indication of ULIRGs exhibiting more complicated SEDs than LIRGs despite showing more signs of interactions. Including radio measurements in the CIGALE modeling constrained the dust luminosity and star formation rate (SFR) estimates by more than 1 order of magnitude better than previously reported for starburst galaxies. We show that total and nonthermal radio luminosity at 1.4 and 4.8 GHz frequencies can be good estimators of recent SFRs for all LIRGs and those ULIRGS with an insignificant influence of active galactic nuclei. A weaker but still significant correlation is observed between radio SFRs at 1.4 GHz and old (averaged over 100 Myr) SFRs based on SED modeling, indicative of multiple episodes of starburst activity during their lifetime. The thermal radio luminosity at 4.8 GHz is a better tracer of recent star formation than the thermal luminosity at 1.4 GHz. Statistically, our modeled nonthermal radio spectral indices do not significantly correlate with redshift, stellar mass, SFR, specific SFR, and dust mass.