Modelling continuum optical and ultraviolet polarization of active galactic nuclei

We present results from a new Monte Carlo radiative transfer computer code, STOKES, developed to model polarization induced by scattering off free electrons and dust grains. STOKES is freely available on the web and can be used to model scattering in a wide variety of astrophysical situations. For edge-on (type-2) viewing positions, the polarization produced by a dusty torus alone is largely wavelength-independent. This is because the torus is optically thick and the change in albedo with wavelength is slight. Wavelength-independent polarization therefore does not necessarily imply electron scattering. We are able to fully explain wavelength independent type-2 polarization without the need to invoke electron scattering cones and consider the dusty torus to be main origin of the polarization in type-2 Seyferts. Polarization perpendicular to the axis of symmetry in type-2 AGNs requires torus half-opening angles of <∼ 60. The degree of polarization is a strong function of the half-opening angle. The high polarizations observed in type-2 AGNs imply half-opening angles ≤ 30. The polarization parallel to the axis of symmetry seen in many type-1 (face-on) AGNs cannot be explained by scattering from a torus alone, or by a torus plus scattering cones. For type-1 objects, the torus gives negligible polarization and electron cones give the wrong angle of polarization. The parallel polarization must arise from a flattened electron-scattering distribution inside the torus. We demonstrate how both the accretion disc itself, and a flattened nearby electron-scattering region can produce the required degree of polarization. We show how the expected polarization varies with the flattening of the regions and source coverage. The dominant parameter influencing the parallel polarization is the electron-scattering optical depth. For a highly ionized emitting and scattering accretion disk an optical depth of only a few tenths is sufficient to produce the observed polarization.