Linear spectropolarimetry of polarimetric standard stars with VLT/FORS2

TL;DR

This study analyzed archival spectropolarimetric data of standard stars from VLT/FORS2, providing precise polarization measurements, testing instrument stability, and refining polarization standards for scientific and technical use.

Contribution

It offers the first detailed calibration and stability assessment of FORS2 spectropolarimetry, improving the accuracy of polarization standards and revealing wavelength-dependent instrumental effects.

Findings

Instrumental polarization stability at ≤0.1%

Polarization degree and angle stable within 0.1% and 0.2°

Systematic deviation in Serkowski parameters from previous studies

Abstract

We reduced ESO's archival linear spectropolarimetry data (4000-9000\AA) of 6 highly polarized and 8 unpolarized standard stars observed between 2010 and 2016, for a total of 70 epochs, with the FOcal Reducer and low dispersion Spectrograph (FORS2) mounted at the Very Large Telescope. We provide very accurate standard stars polarization measurements as a function of wavelength, and test the performance of the spectropolarimetric mode (PMOS) of FORS2. We used the unpolarized stars to test the time stability of the PMOS mode, and found a small ($\leq$0.1%), but statistically significant, on-axis instrumental polarization wavelength dependency, possibly caused by the tilted surfaces of the dispersive element. The polarization degree and angle are found to be stable at the level of $\leq$0.1% and $\leq$0.2 degrees, respectively. We derived the polarization wavelength dependence of the polarized standard stars and found that, in general, the results are consistent with those reported in the literature, e.g. Fossati et al. (2007) who performed a similar analysis using FORS1 data. The re-calibrated data provide a very accurate set of standards that can be very reliably used for technical and scientific purposes. The analysis of the Serkowski parameters revealed a systematic deviation from the width parameter $K$ reported by Whittet et al. (1992). This is most likely explained by incorrect effective wavelengths adopted in that study for the R and I bands.