Previous: Observing with FOS
Up: Observing Procedures
Next: Observing with Fibres
Previous Page: Data reduction
Next Page: How to derive the Stokes parameters
Different modes of using the polarisation optics are conceivable; here we will describe what may be considered a common-user mode for measuring linear polarisation, which is adequate for most purposes. This so-called ``staring'' mode uses the calcite plate analyzer to separate the light into an ordinary and an extra-ordinary beam which results in two spectra on the detector. One component of the polarisation vector (e.g. the Stokes Q parameter) of the incoming beam is converted to an intensity difference between the two spectra. Since both spectra are taken under exactly the same conditions, the intensity ratio is independent of sky transparency. In order to account for differences in the response of the spectrograph and the detector to the highly polarised o and e rays a second exposure with the halfwave plate rotated by 45 degrees is required. This offset of 45 degrees in the halfwave plate position angle results in a rotation of the incoming polarisation vector by 90 degrees. Due to this rotation, on passing the calcite plate, the intensity difference will be inverted relative to the first exposure while leaving the instrumental response identical; the instrumental response can then be taken out by comparing the two exposures.
To measure the full linear polarisation vector (i.e. both Stokes Q and U) a second set of two exposures is required with the halfwave plate set at 22.5 and 67.5 degrees.
A standard sequence of exposures is:
Exposures 1 and 2 yield the Stokes Q spectrum, and exposures 3 and 4 the Stokes U spectrum. The angle of the (linear) polarisation vector, as defined by its Q and U components, is given relative to some instrumental coordinate system. The halfwave plate positions angle may be given any (constant) offset which will only affect the instrumental reference system. The orientation of the instrumental reference system relative to the N-S meridian has to be calibrated by observing a polarisation standard star. It is important to keep the orientation of the instrument fixed relative to the sky ( e.i. set Cassegrain rotator tracking). Be sure to write down the position angle of the halfwave plate and of the Cassegrain rotator and to understand the exact definitions.
Since the calcite plate produces two slit images slightly offset in the spatial direction (the o and e rays) a dekker mask has to be used to prevent confusion between different parts of the slit. This implies that continuous long-slit observations are not feasible in this observing mode. To facilitate semi long-slit observations a comb Dekker may be used to observe a series of regularly spaced apertures along the slit simultaneously. For a more detailed description of observational considerations see the ISIS Spectropolarimetry Users' Manual.
The ISIS polarisation module has been proven to be capable of measuring polarisation reliably to an accuracy better than 0.1 %. High-accuracy polarimetry requires many photons: as a rule of thumb, for planning observations, the uncertainty in one Stokes parameter , where N is the total number of photons --per resolution element-- obtained in two exposures. The generally high count rates necessary for spectropolarimetric observations (and certainly for the usually bright polarimetric standard stars) a CCD is strongly preferred as a detector over the IPCS.
Currently the polarisation optics have been used with the red and blue cameras of ISIS. In principle the polarisation module also functions with FOS but this option has not been tested extensively.
Note that experiments with the dichroic beamsplitter in position showed reflected light from the rear of this component. Such light is displaced along the slit, partly into the spectrum of the other polarisation; this spoils the polarimetry, so for the time being we must, reluctantly, advise against use of the dichroic.