Constrains to mass-loss models can be obtained by measuring intrinsic polarization commonly attributed to Thomson scattering by the near totally ionized envelope (Serkowski 1968, 1970). However by classical polarimetric techniques which measure the flux averaged on the object surface, this leads to very constraining accuracies since the polarization degree is small. Besides these usual techniques detect polarization effects inside morphologically dissymmetric envelopes only. On the contrary, an interferometer equipped with a polarimetric mode (called an interfero-polarimeter) allows to study polarization inside any envelope, whatever its symmetry degrees. In practice, the measurement of the fringe visibility through a linear polarizer parallel then perpendicular to the interferometric baseline enables to measure the distortion of the scattering atmosphere seen through the linear polarizer.
Such interfero-polarimetric observations have already been done (Hanbury Brown et al. 1974; Vakili 1981; Rousselet-Perraut et al. 1997a) and have shown the great interest in developing adapted astrophysical models. Such models are mandatory to interpret the recorded data and, above all, to efficiently carry out future observations (choice of objects and adapted interferometer configurations for instance). Now, only one model has been developed for Wolf-Rayet stars (Cassinelli & Hoffmann 1975) and no general formalism exists to describe the interfero-polarimetric observables. Within the context of the definition of the VLTI focal instrumentation (Paresce et al. 1996), we try to design the instrumental requirements as well as the observational constrains of interfero-polarimetry. Thus we develop a general analytical method providing the intensity maps and the corresponding visibilities recorded in polarized light for any scattering environment. In this model, the geometry and the flux contribution of the envelope, the polarizer orientation and the source function can be parametrized. Since we first aim at defining the interferometric observables, we only consider pure single Thomson scattering (Sect. 2). In Sect. 3, we apply our model to a spherical star surrounded by a spherical envelope and we discuss the effect on the visibility of the envelope diameter, its flux contribution and its electron distribution. In Sect. 4, we consider a spherical star surrounded by an ellipsoidal envelope and we study the visibility dependence on flattening and electron distribution. Finally, we discuss the hypotheses of our modeling and the instrumental and observational requirements of interfero-polarimetry, which leads us to propose a specific instrumentation. To conclude we give some astrophysical applications which would be very attractive with future optical aperture synthesis arrays.
Copyright The European Southern Observatory (ESO)