2017ApJ...851...55S

The mechanism for producing polarized emission from protostellar disks at (sub)millimeter wavelengths is currently uncertain. Classically, polarization is expected from non-spherical grains aligned with the magnetic field. Recently, two alternatives have been suggested. One polarization mechanism is caused by self-scattering from dust grains of sizes comparable with the wavelength, while the other mechanism is due to grains aligned with their short axes along the direction of radiation anisotropy. The latter has recently been shown as a likely mechanism for causing the dust polarization detected in HL Tau at 3.1 mm. In this paper, we present ALMA polarization observations of HL Tau for two more wavelengths: 870 µm and 1.3 mm. The morphology at 870 µm matches the expectation for self-scattering, while that at 1.3 mm shows a mix between self-scattering and grains aligned with the radiation anisotropy. The observations cast doubt on the ability of (sub)millimeter continuum polarization to probe disk magnetic fields for at least HL Tau. By showing two distinct polarization morphologies at 870 µm and 3.1 mm and a transition between the two at 1.3 mm, this paper provides definitive evidence that the dominant (sub)millimeter polarization mechanism transitions with wavelength. In addition, if the polarization at 870 µm is due to scattering, the lack of polarization asymmetry along the minor axis of the inclined disk implies that the large grains responsible for the scattering have already settled into a geometrically thin layer, and the presence of asymmetry along the major axis indicates that the HL Tau disk is not completely axisymmetric.