In the WW spectropolarimeter, we have already eliminated the problem of rotation-angle dependence of the retardance in the achromatic QWPs, as noted by Eversberg et al. (1998), by providing better mountings for the two QWPs. The next step is to eliminate the illumination problem of the fibers (the same procedure could be applied to MuSiCoS). One possibility, in principle, would be to add a lenslet in front of each fiber, to image the entrance pupil instead of the star onto the fibers. This is however a very awkward solution, given that the instrument is built already; such lenslets would also have to be extremely small and precisely mounted in the two beams.
We believe that a much simpler solution to try is to mount a polished, highly uniform interface plate onto the fiber input ends, using a cement of similar index of refraction. In this way, the coupling of the beams to the fibers will be much less susceptible to sensitivity variations across the fiber faces. This is being explored on the WW spectropolarimeter and could be done on MuSiCoS, as well as the new ultra-efficient spectropolarimeter ESPaDOnS planned for use at the Canada-France-Hawaii telescope in the near future. If 1% fluctuations are important, this should also be considered for any other type of fiber-fed device.
© European Southern Observatory (ESO) 2000
Online publication: June 8, 2000