Resolution and calibration effects in high contrast polarimetric imaging of circumstellar scattering regions

Context . Many circumstellar dust scattering regions have been detected and investigated with polarimetric imaging. However, the quantitative determination of the intrinsic polarization and of dust properties is difficult because of complex observational effects. Aims . This work investigates the instrumental convolution and polarimetric calibration effects for high contrast imaging polarimetry with the aim of defining the measuring parameters and calibration procedures for accurate measurements of the circumstellar polarization. Methods . We simulated the instrumental convolution and polarimetric cancellation effects for two axisymmetric point spread functions (PSFs), a Gaussian PSF G and an extended PSF AO , typical for a modern adaptive optics system. The PSFs have the same diameter D PSF for the PSF peak. Further, polarimetric zero-point corrections (zp-corrections) were simulated for different cases, including coronagraphic observations and systems with barely resolved circumstellar scattering regions. Results . The PSF convolution reduces the integrated azimuthal polarization, Σ Q ϕ , for the scattering region, while the net Stokes signals Σ Q and Σ U are not changed. For non-axisymmetric systems, a spurious U ϕ signal is introduced by the convolution. These effects are strong for compact systems and for the convolution with an extended PSF AO . Compact scattering regions can be detected down to an inner working angle of r ≈ D PSF based on the presence of a net Σ Q ϕ signal. Unresolved central scattering regions can introduce a central Stokes Q, U signal that can be used to constrain the scattering geometry even at separations r < D PSF . The smearing by the halo of the PSF AO produces an extended, low surface brightness polarization signal. These effects change the angular distribution of the azimuthal polarization, Q ϕ ( ϕ ), but the initial Q ϕ ′ ( ϕ ) signal can be partly recovered with the analysis of measured Stokes Q and U quadrant pattern. We find that applying a polarimetric zp-correction for the removal of offsets from instrumental or interstellar polarization depends on the selected reference region and can introduce strong bias effects for Σ Q and Σ U and the azimuthal distribution of Q ϕ ( ϕ ). Strategies for the zp-correction are described for different data types, such as coronagraphic data or observations of partly unresolved systems. These procedures provide polarization parameters that can be easily reproduced with model simulations. Conclusions . The simulations describe the impact of the PSF convolution and of calibration offsets for imaging polarimetry in a systematic way, and they show when these effects are strong and how they can be considered in the analysis. This defines also suitable measuring parameters and procedures for the quantitative characterization of the intrinsic scattering polarization Q ϕ ′ for an accurate determination of the properties of the circumstellar dust.