The origin of the mid-infrared nuclear polarization of active galactic nuclei

We combine new (NGC 1275, NGC 4151, and NGC 5506) and previously published (Cygnus A, Mrk 231, and NGC 1068) sub-arcsecond resolution mid-infrared (MIR; 8-13 $\mu$m) imaging- and spectro-polarimetric observations of six Seyfert galaxies using CanariCam on the 10.4-m Gran Telescopio CANARIAS. These observations reveal a diverse set of physical processes responsible for the nuclear polarization, and permit characterization of the origin of the MIR nuclear polarimetric signature of active galactic nuclei (AGN). For all radio quiet objects, we found that the nuclear polarization is low (<1 per cent), and the degree of polarization is often a few per cent over extended regions of the host galaxy where we have sensitivity to detect such extended emission (i.e., NGC 1068 and NGC 4151). We suggest that the higher degree of polarization previously found in lower resolution data arises only on the larger-than-nuclear scales. Only the radio-loud Cygnus A exhibits significant nuclear polarization ($\sim$11 per cent), attributable to synchrotron emission from the pc-scale jet close to the core. We present polarization models that suggest that the MIR nuclear polarization for highly obscured objects arises from a self-absorbed MIR polarized clumpy torus and/or dichroism from the host galaxy, while for unabsorbed cores, MIR polarization arises from dust scattering in the torus and/or surrounding nuclear dust.