Searching for inflationary B modes: can dust emission properties be extrapolated from 350 GHz to 150 GHz?

Recent Planck results have shown that radiation from the cosmic microwave background passes through foregrounds in which aligned dust grains produce polarized dust emission, even in regions of the sky with the lowest level of dust emission. One of the most commonly used ways to remove the dust foreground is to extrapolate the polarized dust emission signal from frequencies where it dominates (e.g. similar to 350 GHz) to frequencies commonly targeted by cosmic microwave background experiments (e.g. similar to 150 GHz). In this Letter, we describe an interstellar medium effect that can lead to decorrelation of the dust emission polarization pattern between different frequencies due to multiple contributions along the line of sight. Using a simple 2-cloud model we show that there are two conditions under which this decorrelation can be large: (a) the ratio of polarized intensities between the two clouds changes between the two frequencies; (b) the magnetic fields between the two clouds contributing along a line of sight are significantly misaligned. In such cases, the 350 GHz polarized sky map is not predictive of that at 150 GHz. We propose a possible correction for this effect, using information from optopolarimetric surveys of dichroicly absorbed starlight.