Clumpy Dust Tori in Active Galactic Nuclei

Abstract

Active Galactic Nuclei (AGN) are amongst the most luminous objects in the universe. The source of their activity is accretion onto a supermassive black hole in the center of the galactic nucleus. The various phenomena observed in AGN are explained in a common unification scheme. The cornerstone of this unification scheme of AGN is the presence of an optically and geometrically thick dust torus which surrounds the central accretion disk and broad-line region (BLR). This parsec-scaled torus is responsible for the apparent difference between type 1 and type 2 AGN. If the line-of-sight intersects with the torus, the accretion disk and BLR are not visible and the AGN is classified as a type 2 object. On the other hand, if the torus is seen nearly face-on, the accretion disk and BLR are directly exposed to the observer, so that the galaxy appears as a type 1 AGN.<br /> Near- (NIR) and mid-infrared (MIR) interferometry has resolved, for the first time, the dust torus around the nearby prototypical Seyfert 2 AGN NGC 1068. These observations provided an insight into the structure of the torus: Apparently, the dust is not smoothly distributed in the torus but arranged in clumps -- contrary to what has been commonly used in models.<br /> We developed a new radiative transfer model of clumpy dust tori which is a key tool to interpret NIR and MIR observations of AGN. The model accounts for the 3-dimensional arrangement of dust clouds. Model SEDs and images can be obtained for a number of different physical parameters (e.g., radial and vertical dust density distribution, cloud radii, optical depths, etc.). It was shown that the model SEDs are in agreement with observed spectral properties. Moreover, we applied our new model to the data of NGC 1068. It was possible, for the first time, to simultaneously reproduce NIR and MIR interferometry and photometry of the nucleus of NGC 1068. In particular, the model follows the trend of the deeper 9.7 micron silicate absorption features in the correlated fluxes than in the total fluxes, as observed with VLTI/MIDI in the 8-13 micron band. Comparison with the NGC 1068 multi-wavelength SED from Radio to the infrared shows that most of the unresolved MIR flux comes from thermal dust emission inside the torus, while in the NIR a possible synchrotron source or the accretion disk might be seen through "holes" in the clumpy torus.<br /> To get a better idea how much the accretion disk contributes to the NIR emission of AGN, we studied NIR colors of a sample of type 1 AGN which were observed in J-, H-, and K-band with HST/NICMOS. By comparing the observed colors with those expected from torus models, we found out that the accretion disk contributes typically < 25% to the K-band flux. The observed colors also indicate that the sublimation temperature is probably close to ~1500 K, but not significantly higher. In addition, reverberation radii of type 1 AGN were compared to theoretical predictions for the dust sublimation radius. Apparently, the reverberation radii are about a factor of 3 smaller than the expected sublimation radius for standard ISM dust grains. This discrepancy can be solved if the inner torus region is dominated by large carbon grains.<br /> We studied the feedback of AGN radiation on the dust torus. It was found out that dust which is smoothly distributed cannot withstand the radiation pressure from the AGN. On the other hand, self-gravitating clouds in clumpy tori can efficiently compensate the AGN radiation pressure. A physically-motivated clumpy torus model was used to study the impact of the AGN radiation on obscuration properties of the torus. We showed that below an AGN luminosity of ~10^42 erg/s, the associated low accretion rates can no longer support an obscuring torus. In the high-luminosity regime, large clouds become unbound so that the torus is dominated by smaller clouds. As a result, the covering factor and apparent scale height decrease with luminosity, so that the fraction of type 1 AGN should become larger at higher luminosities (and high radiative efficiencies). This picture offers a physical explanation for the long-standing "receding torus" phenomenon.<br /> One of the major astronomical discoveries within the last year was the identification of type 2 counterparts of QSOs. These objects were the "missing link" in the unification scheme. We studied restframe optical-to-MIR SEDs of a sample of 21 obscured QSOs with our clumpy torus model. It was found out that the observed SEDs favor models with compact geometries and, apparently, no flaring. In some objects, the combination of blue NIR color and very deep silicate absorption is in contradiction to expectations from torus models. We propose that in such cases, the torus is actually seen face-on, and a detached cold absorber in the host galaxy (e.g., a dust lane or cloud) is responsible for the deep silicate absorption feature. According to this picture, some of the obscured QSOs are mimicking type 2 AGN although their torus orientation might be similar to a type 1 AGN.