Periodic Reporting for period 2 - BiD4BEST (Big Data applications for Black hole Evolution STudies)
Periodo di rendicontazione: 2022-03-01 al 2024-02-29
The overall objective was to train doctoral researchers in the formation of supermassive black holes and their impact on the evolution of galaxies. The study focused on four scientific Work Packages with ambitions that these researchers would acquire outstanding academic expertise, a solid theoretical framework and master state-of-the-art data-science tools in machine learning (ML) and statistical analysis.
WP1-INFANCY The early black hole growth in highly starforming, dust-enshrouded galaxies
WP2-ADOLESCENCE Feedback and Outflows, The effects of AGN on their host galaxies
WP3-ADULTHOOD Galaxy properties and AGN
WP4-Bridging theoretical models to key observables
The project has ensured the creation in Europe of a critical mass of experts in supermassive black hole physics with comprehensive scientific, computational, mathematical, statistical and soft skills. Our ESRs have published 18 papers, contributed to 189 conferences and highlighted to the public the field of supermassive black hole evolution.
WP1-INFANCY The early black hole growth in highly starforming, dust-enshrouded galaxies
WP2-ADOLESCENCE Feedback and Outflows, The effects of AGN on their host galaxies
WP3-ADULTHOOD Galaxy properties and AGN
WP4-Bridging theoretical models to key observables
The project has ensured the creation in Europe of a critical mass of experts in supermassive black hole physics with comprehensive scientific, computational, mathematical, statistical and soft skills. Our ESRs have published 18 papers, contributed to 189 conferences and highlighted to the public the field of supermassive black hole evolution.
WP1 probed the first phases of growth of SMBHs believed to occur in the early, dust-enshrouded growth phases of the host galaxies. We catalogued samples of AGN from IR/radio/X-ray at different redshifts, with some of them specifically selected to be dust-enshrouded quasars, in their early growth phase, the so-called “red quasars”. We used high-resolution ALMA observations to show that the obscured AGN are often hosted in extreme compact starburst galaxies, with large obscuration levels around the central growing black hole, in tension with canonical pure orientation models for AGN obscuration. We adopted a Bayesian-based multi-wavelength methodology to characterize the level of obscuration in deep Chandra COSMOS-legacy field data. We found that the obscured quasars have both higher star-formation IR luminosity and excess radio emission when compared to unobscured quasars. We also found an overall fraction of obscured AGN which is steadily decreasing with cosmic time, from ~60% at z~3, to ~20% in the local Universe.
WP2 probed the phase in which the SMBH is mature enough to eject energetic winds and jets with a potential to impact the host galaxy. This process is dubbed “AGN feedback”. We analysed the incidence of AGN feedback features in multiwavelength AGN samples and assessed the energetics associated with AGN outflows. We integrated ML techniques to enhance the completeness and efficiency of identifying AGN with outflows, resulting in a more than a five-fold increase in sample size compared with standard techniques. We selected ~1200 candidates AGN in the feedback phase out of a parent sample of ~14500 AGN at z>0.5. We performed a parametric analysis of the [OIII] emission lines with the aim of searching for outflows and studied the outflow properties. We analysed Type-2 quasars in the nearby Universe along with the MEGARA datacubes to characterise the emission line kinematics. We found that the kinetic power represents 0.1% of the quasar bolometric luminosity.
WP3 focused on AGN host galaxies and their environment. Via the use of Convolutional Neural Networks trained on simulated mergers and applied real time data from the Sloan Digital Sky Survey (SDSS), we found that while galaxy mergers do not cause black hole accretion overall, there is a correlation between mergers and AGN activity in galaxies with ongoing star formation. Via semi-empirical models, we found that the enhancement of AGN activity at high redshift in dense structures appears associated with galaxies that are located on the outskirts of galaxy clusters. We enhanced the SED-fitting code CIGALE for modeling low accretion rate AGN and found that black hole nuclear activity impacts both the global and radial star formation rate profiles in X-ray-selected AGN within the miniJPAS field.
WP4 aimed to provide a robust theoretical background to all observational projects developed in WP1-3, using phenomenological/data-driven models, semi-analytic models, and hydrodynamic simulations. Starting from the growth of massive “seed” BHs in high-redshift, dust-enshrouded galaxies, we built a full-scale BH mass function, from stellar mass BHs to supermassive BHs at different cosmic epochs. We included state-of-the-art subgrid models of BH growth and feedback from supernovae and AGN in the COLBRE hydrodynamic simulations, which allowed us to investigate the effects of the AGN feedback in galaxies in detail and for a large sample of galaxies. We used ML techniques to calibrate the strength of AGN and supernova feedback in the COLIBRE model. We developed a cutting-edge SMBH spin evolution model which will be coupled to a prescription for energy injection aimed at reproducing the interaction of jets from AGN with the surrounding medium. We developed DECODE, the Discrete statistical semiempirical model, specifically designed to predict rapidly and efficiently, in a full ΛCDM cosmological context, galaxy and SMBH assembly and merger histories for any given input stellar mass-halo mass and SMBH-galaxy mass relations. Via DECODE we have built a holistic and self-consistent view among several galactic-related properties, such as the galaxystar formation histories, merger histories, satellite abundances, intracluster light and probed galaxystar formation shutdown via halo and SMBH quenching. We predicted in semi-analytic and semi-empirical models the roles of both evolutionary and orientation models in shaping obscuration in AGN. We found that a central torus component seems to be a key component in many highly Compton-thick sources at any stage of their evolution.
WP2 probed the phase in which the SMBH is mature enough to eject energetic winds and jets with a potential to impact the host galaxy. This process is dubbed “AGN feedback”. We analysed the incidence of AGN feedback features in multiwavelength AGN samples and assessed the energetics associated with AGN outflows. We integrated ML techniques to enhance the completeness and efficiency of identifying AGN with outflows, resulting in a more than a five-fold increase in sample size compared with standard techniques. We selected ~1200 candidates AGN in the feedback phase out of a parent sample of ~14500 AGN at z>0.5. We performed a parametric analysis of the [OIII] emission lines with the aim of searching for outflows and studied the outflow properties. We analysed Type-2 quasars in the nearby Universe along with the MEGARA datacubes to characterise the emission line kinematics. We found that the kinetic power represents 0.1% of the quasar bolometric luminosity.
WP3 focused on AGN host galaxies and their environment. Via the use of Convolutional Neural Networks trained on simulated mergers and applied real time data from the Sloan Digital Sky Survey (SDSS), we found that while galaxy mergers do not cause black hole accretion overall, there is a correlation between mergers and AGN activity in galaxies with ongoing star formation. Via semi-empirical models, we found that the enhancement of AGN activity at high redshift in dense structures appears associated with galaxies that are located on the outskirts of galaxy clusters. We enhanced the SED-fitting code CIGALE for modeling low accretion rate AGN and found that black hole nuclear activity impacts both the global and radial star formation rate profiles in X-ray-selected AGN within the miniJPAS field.
WP4 aimed to provide a robust theoretical background to all observational projects developed in WP1-3, using phenomenological/data-driven models, semi-analytic models, and hydrodynamic simulations. Starting from the growth of massive “seed” BHs in high-redshift, dust-enshrouded galaxies, we built a full-scale BH mass function, from stellar mass BHs to supermassive BHs at different cosmic epochs. We included state-of-the-art subgrid models of BH growth and feedback from supernovae and AGN in the COLBRE hydrodynamic simulations, which allowed us to investigate the effects of the AGN feedback in galaxies in detail and for a large sample of galaxies. We used ML techniques to calibrate the strength of AGN and supernova feedback in the COLIBRE model. We developed a cutting-edge SMBH spin evolution model which will be coupled to a prescription for energy injection aimed at reproducing the interaction of jets from AGN with the surrounding medium. We developed DECODE, the Discrete statistical semiempirical model, specifically designed to predict rapidly and efficiently, in a full ΛCDM cosmological context, galaxy and SMBH assembly and merger histories for any given input stellar mass-halo mass and SMBH-galaxy mass relations. Via DECODE we have built a holistic and self-consistent view among several galactic-related properties, such as the galaxystar formation histories, merger histories, satellite abundances, intracluster light and probed galaxystar formation shutdown via halo and SMBH quenching. We predicted in semi-analytic and semi-empirical models the roles of both evolutionary and orientation models in shaping obscuration in AGN. We found that a central torus component seems to be a key component in many highly Compton-thick sources at any stage of their evolution.
The project has contributed to the European knowledge base, made major progress in the field of supermassive black hole evolution, and created robust solutions to deal with AGN science, huge volumes and selection effects. We established a dialogue with society through 74 public engagement and outreach activities and 55 publications. This transfer of our scientific knowledge will inform society and decisions that will shape our future and impact everyday lives, such as through the creation of jobs and the introduction of novel technologies.
WP1: Our algorithms and determination of redshift dependent AGN obscured fractions are of extreme relevance for observational determination of AGN in large surveys, and to constrain cosmological models of the co-evolution of black holes and galaxies.
WP2: We have developed sound strategies to measure mass outflow rates and energetics from observational data to be compared against model predictions.
WP3: We aim to release/publish ad-hoc ML-based algorithms to determine the merger/morphology status on very large samples of active and inactive galaxies (e.g. LSST, Euclid). We will have detailed measures of SMBH Eddington ratio distributions and host galaxy properties, that will be invaluable to constrain SMBH scaling relations and cosmological models.
WP4: Our phenomenological method provides sound constraints for cosmological models on the existence of massive BHs at high redshift, and to the future LISA Gravitational Wave background signal. We have set out a semi-empirical model DECODE, to predict the mean stellar mass assembly and merging histories of galaxies of any given final mass for any input stellar mass-halo mass relation. We will make available our hydrodynamic simulations and SMBH feeding and feedback recipes for others to compare with and/or implement in their models.
WP1: Our algorithms and determination of redshift dependent AGN obscured fractions are of extreme relevance for observational determination of AGN in large surveys, and to constrain cosmological models of the co-evolution of black holes and galaxies.
WP2: We have developed sound strategies to measure mass outflow rates and energetics from observational data to be compared against model predictions.
WP3: We aim to release/publish ad-hoc ML-based algorithms to determine the merger/morphology status on very large samples of active and inactive galaxies (e.g. LSST, Euclid). We will have detailed measures of SMBH Eddington ratio distributions and host galaxy properties, that will be invaluable to constrain SMBH scaling relations and cosmological models.
WP4: Our phenomenological method provides sound constraints for cosmological models on the existence of massive BHs at high redshift, and to the future LISA Gravitational Wave background signal. We have set out a semi-empirical model DECODE, to predict the mean stellar mass assembly and merging histories of galaxies of any given final mass for any input stellar mass-halo mass relation. We will make available our hydrodynamic simulations and SMBH feeding and feedback recipes for others to compare with and/or implement in their models.