Infrared polarimetry and the galactic magnetic field. II. Improved models

The observed trend in interstellar polarization with extinction at 2.2 μm can be well modeled by assuming that interstellar polarization is independent of magnetic field strength and depends only on the geometry of the field. Two models, one invoking Alfvén waves with random phase and one combining ad hoc random and constant components, are developed. The wave model best fits the data when there is equipartition between the magnetic and turbulent energy densities in the ISM. The two-component model best fits the data when there is equal energy in the random and constant components. The observed dispersion in polarization magnitude and position angle requires that the random component of both models decorrelate over an optical depth interval (not physical path length) close to a value of Δτ ∼ 0.1 at 2.2 μm (A_v ∼ 1). This constraint on the decorrelation optical depth is found to hold for lines of sight through both the diffuse ISM and dense dark clouds, even though the corresponding physical path lengths are widely different. This suggests that the geometry of the magnetic field is at least partially preserved when molecular clouds contract out of the diffuse ISM. Two simple scenarios for cloud formation, one involving reducing the space between very low filling factor cloudlets and the other involving one-dimensional compression across field lines are discussed.