2003A&A...412..615G

We revisit in this paper the location of the various components observed in the AGN of NGC1068. Discrepancies between previously published studies are explained, and a new measurement for the absolute location of the K-band emission peak is provided. It is found to be consistent with the position of the central engine as derived by Gallimore et al. (1997Natur.388..852G), Capetti et al. (1997ApJ...476L..67C) and Kishimoto (1999ApJ...518..676K). A series of map overlays is then presented and discussed. Model predictions of dusty tori show that the nuclear unresolved NIR-MIR emission is compatible with a broad range of models: the nuclear SED alone does not strongly constrain the torus geometry, while placing reasonable constraints on its size and thickness. The extended MIR emission observed within the ionizing cone is shown to be well explained by the presence of optically thick dust clouds exposed to the central engine radiation and having a small covering factor. Conversely, a distribution of diffuse dust particles within the ionizing cone is discarded. A simple model for the H₂ and CO emission observed perpendicularly to the axis of the ionizing cone is proposed. We show that a slight tilt between the molecular disc and the Compton thick central absorber naturally reproduces the observed distribution of H₂ of CO emission.