A comprehensive and critical re-analysis of FORS spectropolarimetric observations: the case of HD92207

Stellar spectropolarimetry has become extremely popular during the last decade, and has led to major advances in our understanding of stellar magnetic fields and of their impact on stellar structure and evolution. Many important discoveries have been obtained thanks to observations performed with the FORS low-resolution spectropolarimeters of the ESO Very Large Telescope. We first review and summarise the major results of a homogeneous re-reduction and analysis of all single-slit FORS1 spectropolarimetric observations. This work revealed a non-negligible dependence of the results upon the adopted reduction and analysis procedure, as well as the presence of instabilities, revealing that photon noise is not the only source of uncertainty. As a consequence of our new analysis and assessment of the uncertainties, we are not able to confirm a large number of magnetic field detections presented in the past for a variety of stars. We further summarise the results of FORS2 spectropolarimetric observations of the A0 supergiant HD92207 which allowed us to explore further the nature of the instabilities, roughly constraining their maximum impact on the derived Stokes profiles and magnetic field values. We finally present new results obtained with a further independent pipeline on the FORS2 data of HD92207, confirming our previous analysis, and discuss simple quality-check controls which can be performed on the data in order to distinguish between genuine and spurious signals. All together, our results reveal that the FORS spectropolarimeters are indeed reliable instruments, when their capabilities are not pushed beyond the limits of a Cassegrain mounted low-resolution spectrograph.