Identification of Active Galactic Nuclei through different selection techniques

The most recent studies suggest that almost all galaxies in the Local Universe host Super-Massive Black Holes (SMBHs) in their center. SMBHs are responsible for the most powerful phenomena in the Universe, the Active Galactic Nuclei (AGN). When SMBHs accrete matter, strong radiation is emitted at all wavelengths of the electromagnetic field. The complete census of the various populations of AGN and the construction of samples with high statistical significance are essential to understand the evolution of the SMBHs and the host galaxies in the history of the Universe. The present doctoral thesis focused on comparing different AGN detection techniques. In particular, AGN selection techniques based on photometric and spectroscopic data were used from different wavelength regimes of the electromagnetic spectrum (optical, infrared, X-rays) as well as techniques that are based on flux variability. Comparing the methods with each other, the observational and physical AGN properties were studied. The main conclusion is that the different methods used to identify AGN are complementary to each other and equally important in building the whole picture of the AGN population. This doctoral thesis is divided into three distinct parts according to the fields studied and the techniques applied: A. Robust identification of Active Galactic Nuclei through HST optical variability in GOODS-S: Comparison with the X-ray and mid-IR selected samplesIdentifying AGN through their X-ray emission is efficient, but necessarily biased against X-ray-faint objects. In order to characterize this bias, we compare the X-ray-selected AGN to the ones identified through optical variability and mid-IR colours. We present a catalogue of AGN selected through optical variability using all publicly available z-band Hubble Space Telescope images in the GOODS-South field. For all objects in the catalogue, we compute X-ray upper limits or discuss detections in the deepest available ~7 Ms Chandra Deep Field South images and present the Spitzer/IRAC mid-IR colours. For the variability study, we consider only sources observed over at least five epochs and over a time baseline of up to ten years. We adopt the elevated median absolute deviation as a variability indicator robust against individual outlier measurements and identify 113 variability-selected AGN candidates. Among these, 26 have an X-ray counterpart and lie within the conventional AGN area in the Fx/Fopt diagram. The candidates with X-ray upper limits are on average optically fainter, have higher redshifts compared to the X-ray detected ones and are consistent with low luminosity AGN. Out of 41 variable optical sources with IR detections, 13 fulfill the IR AGN colour selection criteria. Our work emphasizes the importance of optical variability surveys for constructing complete samples of AGN including the ones that remain undetected even by the deepest X-ray and IR surveys. The results in this field were published in the journal Monthly Notices of the Royal Astronomical Society (MNRAS) as Pouliasis et al. (2019).B. The "Hubble Catalog of Variables": AGN identification in CANDELS and Frontier FieldsWithin the framework of the "Hubble Catalog of Variables" (HCV) project, AGN candidates were selected through variability. The purpose of the HCV program is to detect variable point-like sources as well as extended sources using data from the Hubble Source Catalog. In particular for the AGN detection, fields that have deep observations with long exposure times and have also been observed many times are necessary using all the available filters. Such are the CANDELS and Frontiers fields. The results in the GOODS-South field were used as a keystone in the selection process and confirmation and verification of the results in the latter fields. Depending on the available data from the entire electromagnetic spectrum in each field, different AGN selection methods were used to compare the results. Part of the results in the CANDELS fields were included in the publication of Bonanos et al. (2019) in the journal Astronomy & Astrophysics.C. An obscured AGN population hidden in the VIPERS galaxies: identification through spectral energy distribution decomposition The X-ray emission constitutes a reliable and efficient tool for the selection of Active Galactic Nuclei (AGN) although, it may be biased against the most heavily obscured AGN. Simple mid-IR broad-band selection criteria could identify a large number of luminous and absorbed AGN yet again host contamination could lead to non-uniform and incomplete samples. The Spectral Energy Distribution (SED) decomposition could decouple the emission from the AGN and that from star-forming regions revealing weaker AGN components. We aim to identify an obscured AGN population through SED modelling among galaxies in the VIPERS survey in the CFHTLS W1 field. We construct SEDs for 6,860 sources and identify 160 AGN at a high confidence level using a Bayesian approach. Using optical spectroscopy, we confirm the nature of ~85% of the AGN. Our AGN sample is highly complete (~92%) compared to mid-IR colour AGN, but also includes a significant number of galaxy-dominated systems with lower luminosities. In addition to lack of X-ray emission (80%), the SED fitting results suggest that the majority of the sources are obscured. We use a number of diagnostic criteria in the optical, infrared and X-ray regime to verify these results. Interestingly, only 35% of the most luminous mid-IR selected AGN have X-ray counterparts suggesting strong obscuration. Our work emphasizes the importance of using SED decomposition techniques to select a population of type II AGN, which may remain undetected by either X-ray or IR colour surveys. The results in this field have been submitted in the journal MNRAS. The conclusions derived from this thesis are that the different methods used to detect AGN are complementary to each other and equally important in determining the full picture of the AGN demographics. Each method selects AGN samples with different physical and observational properties. Mid-IR identification selects AGN with large amounts of dust around the black hole, which may obscure some or all of the X-ray emission. On the other hand, the X-ray selected AGN could not be selected in the infrared regime, as they may contain less amount of dust and often be contaminated by the host galaxy. Furthermore, the optical variability is able to identify a large number of low-luminosity AGN, especially at high redshifts, including those that remain undetectable even at the deeper observations in the X-rays or the infrared. This population is critical for studying the faint end of the AGN luminosity function and may be the key between normal galaxies and AGN. In addition, the importance of using SED decomposition techniques to detect less luminous AGN of high absorption of radiation at optical wavelengths and X-rays is emphasized, which other methods may not be as effective.