In many astrophysical processes, the complete characterization of the polarization properties of radiation gives important information about the source or the interaction between radiation and matter. Stokes imaging polarimeters based on liquid crystal variable retarders (LCVRs) are widely used to determine the polarization state of radiation. But these devices have errors due to the nature of the liquid crystals that induce an inhomogeneous retardance over the complete aperture, and their fast-axis position depends on the applied voltage. We present the analysis of the impact of variations in the retardance and orientation of the LCVRs, in the optimization of a Stokes imaging polarimeter. An optimized Stokes polarimeter based on two LCVRs, was simulated. First considering individual errors in the retardance and orientation, then one million cases were simulated adding random errors in both LCVRs. The condition number was calculated to analyze the effect in the optimization of the polarimeter. Also, a Stokes imaging polarimeter was calibrated, to verify experimentally the effect of these errors in a real polarimeter. For the calibration the instrument matrix and condition number were calculated individually in every pixel of the image. The results show that these errors lead to polarimetric measurements different from the ideal and to a poorly optimized polarimeter, demonstrating that a condition number optimization of the instrument matrix and a calibration are not sufficient to guarantee accurate polarimetric measurements.
|