Imaging polarimetry is an important tool for the study of cosmic magnetic fields. In our Galaxy, polarization levels of a few up to $sim$10% are measured in the submillimeter dust emission from molecular clouds and in the synchrotron emission from su
pernova remnants. Only few techniques exist to image the distribution of polarization angles, as a means of tracing the plane-of-sky projection of the magnetic field orientation. At submillimeter wavelengths, polarization is either measured as the differential total power of polarization-sensitive bolometer elements, or by modulating the polarization of the signal. Bolometer arrays such as LABOCA at the APEX telescope are used to observe the continuum emission from fields as large as $sim0fdg2$ in diameter. %Here we present the results from the commissioning of PolKa, a polarimeter for Here we present PolKa, a polarimeter for LABOCA with a reflection-type waveplate of at least 90% efficiency. The modulation efficiency depends mainly on the sampling and on the angular velocity of the waveplate. For the data analysis the concept of generalized synchronous demodulation is introduced. The instrumental polarization towards a point source is at the level of $sim0.1$%, increasing to a few percent at the $-10$db contour of the main beam. A method to correct for its effect in observations of extended sources is presented. Our map of the polarized synchrotron emission from the Crab nebula is in agreement with structures observed at radio and optical wavelengths. The linear polarization measured in OMC1 agrees with results from previous studies, while the high sensitivity of LABOCA enables us to also map the polarized emission of the Orion Bar, a prototypical photon-dominated region.
Context: We investigate non-Zeeman circular polarization and linear polarization levels of up to 1% of $^{12}$CO spectral line emission detected in a shocked molecular clump around the supernova remnant (SNR) IC 443, with the goal of understanding th
e magnetic field structure in this source. Aims: We examine our polarization results to confirm that the circular polarization signal in CO lines is caused by a conversion of linear to circular polarization, consistent with anisotropic resonant scattering. In this process background linearly polarized CO emission interacts with similar foreground molecules aligned with the ambient magnetic field and scatters at a transition frequency. The difference in phase shift between the orthogonally polarized components of this scattered emission can cause a transformation of linear to circular polarization. Methods: We compared linear polarization maps from dust continuum, obtained with PolKa at APEX, and $^{12}$CO ($J=2rightarrow1$) and ($J=1rightarrow0$) from the IRAM 30-m telescope and found no consistency between the two sets of polarization maps. We then reinserted the measured circular polarization signal in the CO lines across the source to the corresponding linear polarization signal to test whether before this linear to circular polarization conversion the linear polarization vectors of the CO maps were aligned with those of the dust. Results: After the flux correction for the two transitions of the CO spectral lines, the new polarization vectors for both CO transitions aligned with the dust polarization vectors, establishing that the non-Zeeman CO circular polarization is due to a linear to circular polarization conversion.
