Introduction

At several places in the VLTI some form of spatial filtering is performed, whereby the light from certain parts of the image plane is selected to continue along one path through the VLTI, and other light is rejected or sent along a different path. Spatial filtering processes such as this couple the Zernike piston mode to high-order Zernike modes, meaning that the piston mode in the interferometric beam is modulated according to the wavefront corrugations across the aperture plane, as well as modifying the spectral sensitivity distribution (see Section 3.3 for further details of both these effects). As a result, it is necessary to study in detail the expected wavefront corrugations immediately before each VLTI component which performs spatial filtering, which requires numerical simulations of the optical wavefronts at these locations. The first of these components is the StS, which separates the light from one star into two beams when operating in calibration mode. The separation of the beams is done using a roof mirror in the image plane which produces the same effect as a knife-edge test or Schlieren type wavefront detector (see [13,14]) in the individual output beams (with each output beam having the knife-edge on the opposite side of the image plane). The amplitude and phase of the output beams are thus strongly (and non-linearly) coupled to the wavefront corrugations in the input beam.

The StS calibration mode relies on both of the output beams from a single star having the same phase. In order investigate the expected performance of this calibration procedure it will be necessary to assess the amount of difference in the wavefront amplitude and phase expected in the two outputs from the StS.