Final Report Summary - PARSECDUST (Supermassive black holes under the microscope: Revealing the dusty environment of AGN on parsec scales with IR and sub-mm interferometry)
1 Introduction and motivation
The project aimed at exploring the origin, structure, and physical properties of the dusty environment of actively accreting supermassive black holes – or active galactic nuclei (AGN).
It has long been held that the bulk of the dust is arranged in a toroidal geometry, also known as the “dusty torus”, which contains a significant amount of mass to grow the black hole and obscure the line-of-sight to the AGN when seen edge-on. However, with the help of infrared interferometry, I was able to demonstrate recently that the majority of the infrared light emitted by the dust originates from a region where the standard torus model would have predicted no dust to be present. I formulated a new hypothesis where the dust consists of two major elements: (1) a thin disk of dusty gas transporting mass from galactic scales towards the black hole, and (2) a dusty wind that blows away a significant part of the accreted dusty gas, driven by the pressure of the strong radiation of the hot gas around the black hole. If correct, this would have major implications on how AGN affect the evolution of their host galaxies.
2 Project outline
In the course of the Marie Curie fellowship, I wanted to test this new hypothesis. In particular, I tried to understand if these dusty wind features are ubiquitous in AGN and if they are, indeed, caused by radiation pressure. Moreover, I wanted to understand how the radiation changes the composition of the dust as it makes its way from galactic scales inward to the black hole. This affects the way we see the dusty region in different wavebands. To achieve these goals, new observations with the largest infrared telescopes were performed and new models have been developed.
3 Results
One of the key results from this project that the infrared emission is, indeed, predominantly originating from the polar region in the large majority of typical AGN (Hoenig et al., in prep; Asmus, Hoenig & Gandhi 2016). Most strikingly, we demonstrated that all AGN that displayed a hint of extension in single 8m-telescope observations were extended into the polar region. Since the observed scales in this study are larger by a factor of 10 than what we saw in previous and the new interferometry data taken in the course of the project, the “dusty wind” hypothesis received significant support: It becomes increasingly difficult to explain the new observations in the framework of the old torus paradigm.
The second aspect of the project concerned the physical interpretation of the dusty wind features: what physical process drives the wind? I developed a dynamical model based on an ensemble of stable dusty gas clouds orbiting the supermassive black hole. The model assumes that the motion of clouds is influenced by gravity from the black hole and all other clouds in the system, as well as the ultraviolet and optical radiation pressure from the accretion around the black hole and the reemission of infrared radiation from each cloud. In particular the latter part is computationally extremely expensive, which led me to develop very effective methods to solve the radiative transfer problem within a highly-parallel code (typically involving 16-64 CPU cores). The simulations showed that radiation pressure from the clouds, from the AGN, and self-gravity of the clouds dominate in different regimes of the parameter space of black hole mass and AGN luminosity. Interestingly, the strongest push into the polar region occurs when the accretion disk radiates at about 5-8% of its maximum possible luminosity for a given black hole mass. This corresponds to the typical AGN we see in our cosmic neighbourhood and those probed by the interferometric and single-telescope observations. These results lend further strong support to the dusty wind hypothesis and will be published in an upcoming paper (Hoenig 2016, in prep.).
Finally, it was investigated how the dust composition changes when the dust is heated to the highest temperatures, being close to sublimation. This hot dust can be observed in the near-infrared. I requested new observations to test this regime but, due to bad weather, mostly reverted to simulations and archival data. I found that coming up with an assessment of how different chemical composition influences the observed emission is difficult since the expected flux in models is very sensitive to the physical assumptions of the model. On the other hand, we found that the hot-dust region forms a boundary to the strong iron emission seen in X-rays, with the majority of the iron emission coming from very close to the sublimation radius (Gandhi, Hoenig & Kishimoto 2015). The presence of highly ionised iron in this region suggests that we, indeed, see sublimation of silicate dust grains – one of the major constituents of astronomical dust – while graphite grains still survive. I built a new radiative transfer model based on this phenomenology, which has been applied to interpret the infrared emission of AGN by my co-supervised student Judit Garcia (Hoenig et al. 2014; Garcia Lopez, Alonso Herrero, Hoenig, et al., in prep.)
4 Unexpected opportunities
While investigating the near-infrared emission of AGN, I discovered that the combination of time-lag measurements and infrared interferometry of the hot dust region can be used to measure direct geometric distances to AGN (Hoenig et al. 2014, Nature). This is quite relevant since it is a new and independent way of determining the cosmologically important Hubble constant, which measures the current expansion rate of our universe. I started a new observational campaign to extend these measurements to a sample of about a dozen AGN with the goal of measuring the Hubble constant with 3-4% accuracy.
The project aimed at exploring the origin, structure, and physical properties of the dusty environment of actively accreting supermassive black holes – or active galactic nuclei (AGN).
It has long been held that the bulk of the dust is arranged in a toroidal geometry, also known as the “dusty torus”, which contains a significant amount of mass to grow the black hole and obscure the line-of-sight to the AGN when seen edge-on. However, with the help of infrared interferometry, I was able to demonstrate recently that the majority of the infrared light emitted by the dust originates from a region where the standard torus model would have predicted no dust to be present. I formulated a new hypothesis where the dust consists of two major elements: (1) a thin disk of dusty gas transporting mass from galactic scales towards the black hole, and (2) a dusty wind that blows away a significant part of the accreted dusty gas, driven by the pressure of the strong radiation of the hot gas around the black hole. If correct, this would have major implications on how AGN affect the evolution of their host galaxies.
2 Project outline
In the course of the Marie Curie fellowship, I wanted to test this new hypothesis. In particular, I tried to understand if these dusty wind features are ubiquitous in AGN and if they are, indeed, caused by radiation pressure. Moreover, I wanted to understand how the radiation changes the composition of the dust as it makes its way from galactic scales inward to the black hole. This affects the way we see the dusty region in different wavebands. To achieve these goals, new observations with the largest infrared telescopes were performed and new models have been developed.
3 Results
One of the key results from this project that the infrared emission is, indeed, predominantly originating from the polar region in the large majority of typical AGN (Hoenig et al., in prep; Asmus, Hoenig & Gandhi 2016). Most strikingly, we demonstrated that all AGN that displayed a hint of extension in single 8m-telescope observations were extended into the polar region. Since the observed scales in this study are larger by a factor of 10 than what we saw in previous and the new interferometry data taken in the course of the project, the “dusty wind” hypothesis received significant support: It becomes increasingly difficult to explain the new observations in the framework of the old torus paradigm.
The second aspect of the project concerned the physical interpretation of the dusty wind features: what physical process drives the wind? I developed a dynamical model based on an ensemble of stable dusty gas clouds orbiting the supermassive black hole. The model assumes that the motion of clouds is influenced by gravity from the black hole and all other clouds in the system, as well as the ultraviolet and optical radiation pressure from the accretion around the black hole and the reemission of infrared radiation from each cloud. In particular the latter part is computationally extremely expensive, which led me to develop very effective methods to solve the radiative transfer problem within a highly-parallel code (typically involving 16-64 CPU cores). The simulations showed that radiation pressure from the clouds, from the AGN, and self-gravity of the clouds dominate in different regimes of the parameter space of black hole mass and AGN luminosity. Interestingly, the strongest push into the polar region occurs when the accretion disk radiates at about 5-8% of its maximum possible luminosity for a given black hole mass. This corresponds to the typical AGN we see in our cosmic neighbourhood and those probed by the interferometric and single-telescope observations. These results lend further strong support to the dusty wind hypothesis and will be published in an upcoming paper (Hoenig 2016, in prep.).
Finally, it was investigated how the dust composition changes when the dust is heated to the highest temperatures, being close to sublimation. This hot dust can be observed in the near-infrared. I requested new observations to test this regime but, due to bad weather, mostly reverted to simulations and archival data. I found that coming up with an assessment of how different chemical composition influences the observed emission is difficult since the expected flux in models is very sensitive to the physical assumptions of the model. On the other hand, we found that the hot-dust region forms a boundary to the strong iron emission seen in X-rays, with the majority of the iron emission coming from very close to the sublimation radius (Gandhi, Hoenig & Kishimoto 2015). The presence of highly ionised iron in this region suggests that we, indeed, see sublimation of silicate dust grains – one of the major constituents of astronomical dust – while graphite grains still survive. I built a new radiative transfer model based on this phenomenology, which has been applied to interpret the infrared emission of AGN by my co-supervised student Judit Garcia (Hoenig et al. 2014; Garcia Lopez, Alonso Herrero, Hoenig, et al., in prep.)
4 Unexpected opportunities
While investigating the near-infrared emission of AGN, I discovered that the combination of time-lag measurements and infrared interferometry of the hot dust region can be used to measure direct geometric distances to AGN (Hoenig et al. 2014, Nature). This is quite relevant since it is a new and independent way of determining the cosmologically important Hubble constant, which measures the current expansion rate of our universe. I started a new observational campaign to extend these measurements to a sample of about a dozen AGN with the goal of measuring the Hubble constant with 3-4% accuracy.