High Angular Resolution Radio Observations of Luminous Infrared Galaxies

Abstract

The overall goal of this thesis has been to improve our understanding of the radio emission and absorption processes taking place in luminous and ultra-luminous infrared galaxies (U/LIRGs) in the local Universe, using state-of-the-art radio interferometers that offer high angular resolution (better than 1 arcsecond) and sensitivity at multiple frequencies. These radio interferometric observations allow us to characterise the spectral energy distribution in the GHz frequency range, where synchrotron emission, as well as free-free thermal emission, is known to contribute significantly (Condon, 1992). I have presented results from the e-MERLIN LIRGI sample, whose overall objective is to characterise the phenomenological evolution of the core of a starburst. One of the immediate goals is to reveal the source responsible for the heating of the dust and gas in the nuclear regions of these galaxies (AGN or starburst). Although we present here preliminary results from LIRGI in the 5 GHz band, the results show the enormous potential of using radio interferometry at resolutions better than arcsecond to study outbursts and nuclear regions in the local Universe. In the first part of this paper we made a comprehensive study of the LIRG Arp 299, where we present the first observations of this LIRG from the Jansky Very Large Array (JVLA) at frequencies between 1.4 and 8.4 GHz combined with the first observations of this LIRG obtained with the LOw Frequency ARray (LOFAR), including the international stations. This work has included a study of its magnetic field, of its emission measure and therefore its corresponding electron density, of its spectral index, and of its structural characteristics at these frequencies. One of the most important results obtained from LOFAR observations, in combination with JVLA observations at various frequencies, is the characterisation of the interstellar medium in the cores of LIRGs. For this purpose, we fit the spectral energy distribution of the cores, between 150 MHz and 8.4 GHz, using two different models of the absorbing/emitting thermal gas: in one model, emitting/absorbing particles are uniformly distributed (continuous model) (Condon, 1992), while in the second one we assume a clumpy medium (Conway, Elitzur, and Parra, 2018), where there is a nonuniform distribution. Both models fit the existing data well. The continuum model can account for the SED of nuclei with a standard population of relativistic electrons subject to synchrotron, Bremsstrahlung and ionisation losses, which are expected to be significant due to the large densities found in the central regions of the U/LIRGs (Lacki, Thompson, and Quataert, 2010). The clumpy model can explain the data by a relativistic electron population with negligible energy losses, and predicts thermal fractions that are more typical of star-forming galaxies, compared to the continuum model. We propose LOFAR observations at frequencies below 100 MHz, or uGMRT observations at 600 MHz to discern between the two models. In any case, these results highlight the relevance of low-frequency, high angular resolution observations for tracing the diffuse interstellar medium in galaxies. In the second part of the thesis I have addressed the study of the physical properties of the LIRGI sample, as well as the characterisation of the emission and absorption processes in local galaxies, where we can do so in extraordinary detail. This will allow us to better understand the properties of galaxies with outbursts of star formation at cosmological distances (Magnelli et al., 2009), where we know that they were much more abundant, but the angular resolution does not allow us to resolve the structures in adequate detail. Finally, I studied the supernova remnant luminosity function (SNR) in normal (Chomiuk and Wilcots, 2009) galaxies, with the aim of constructing a universal luminosity function. To do so, I wrote a code that uses a non-uniform bin size, which avoids introducing biases in the study due to the small size of some SNR samples. The main result is to obtain a new (true) completeness limit on the global sample due to the presence of starburst galaxies.