1999A&A...344..668G

Polarization at 800 and 1100µm has been observed in the star-forming regions Mon R2, DR21 and W3-IRS4. These data have been used to test proposed mechanisms for producing polarized dust emission in the submillimetre. It was found that illumination effects are not sufficient to explain the observed polarization, and that some degree of grain alignment is required. In general, mechanical alignment (via collisions with molecules) and radiative alignment of asymmetrical grains are the most probable mechanisms. The degrees and directions of polarization observed in DR21 are found to be wavelength dependent at 800 to 1100µm. A search of previous JCMT polarimetry results shows that for about half the sources observed, p(1100) is greater than p(800), while in most of the other sources the two values of p are in agreement. A possible explanation for this result is that there are two (or more) non-cospatial grain populations along the line of sight, with different degrees of polarizability and different opacity indices. This can produce bias effects in the polarization percentage and position angle. Thus the choice of wavelength is not immaterial in dust emission polarimetry, as is commonly believed.