Figure 1 (click here) is a sketch of the device which consists of two wedges which
split the pupil image and deviate all rays above/below the optical
axis thus preventing vignetting at the interface between the two
systems of prisms. The rays than
enter into the two Wollastons which have crystal axis at 
, and
emerge at 4 different angles with relative intensities depending on the
polarization angle and degree of the input light. The four images on the
array correspond therefore to measurements performed with
polarizers at 0, 90, 45 and 135 degrees, i.e. the parameters needed
to derive the first 3 elements of the Stokes vector (e.g. Shurcliff 1962).
The wedges and the Wollaston prisms are not necessarily separated elements, i.e. the wedge could be obtained by simply cutting the entrance face of the first Wollaston prism at a suitable angle. This simpler solution could be useful in cases where lateral chromatism is not a crucial issue (Sect. 3.1).
From the practical point of view, integrating a WeDoWo inside an instrument is as simple as using Wollastons, i.e. the device can be mounted in the filter or grism wheels of any focal reduced with a corrected, parallel pupil image. The only difference is that the input field mask should be designed taking into account that the WeDoWo produces four images.
Figure 2: Left: schematic representation of a thin
WeDoWo device useful for spectro-polarimetry at visual
(CaCO
prisms) and IR (LiNbO
prisms) wavelengths.
Center, right: values of the 
and 
angles necessary to
create four non-overlapping images of a slit. The parameter 
is the
angle projected by the slit length onto the pupil image (Eq. 1).
All computations are based on room temperature refractive indices
of CaCO
(Bennet 1995) while those of LiNbO
are at 77 K
using refraction indices and thermo-optic coefficients from
Smith et al. (1976)