Section 5 (click here) gives the effective performance of the instrumental configuration used. We discuss them in comparison with theoretical calculations and others systems.
First, the rejection capability of the coronograph appears to be in
good agreement with theoretical expectations, even though it is
slightly reduced. The following effects may explain the remaining
discrepancy. First, the optics are not perfect. In particular, the
mirror scattering decreases the efficiency of a coronograph. For
instance, the latest observations occurred after the primary mirror had
been re-aluminized and the improvement in the image quality was
spectacular. Second, the wavefront
after the adaptive optics system
is not perfectly flat and the wings of the PSF are correspondingly
modified. Also, since the efficiency of the coronograph critically
depends on the accurate centering of the star, a slight variation in
the position during observation, possibly due to a small mechanical
deformation, affects the whole shape of the residual light profile.
The importance of the star position stability underlined below explains why observing with little, or no image correction does not allow the use of very small masks. Consequently, the close vicinity of the star may not be investigated without AO. Quantitatively, no precise measurements have been achieved closer than 3 from a star without AO correction (Kalas & Jewitt 1995) or closer than 2.5 with tip-tilt correction (Golimovski et al. 1993). This is to be compared to our present limit of 1.5, where the uncertainty of the whole observational and reduction procedure discussed in Sect. 5.2 is lower than the disk brightness (Mouillet et al. 1995, here after Paper II). We have to stress that our observational parameters were optimized for observation close to the star. Different observational parameters would be chosen to observe at greater distances. A larger mask would provide a better S/N with longer exposure times. The compromise between the pixel sampling and the field of view depends on the particular application concerned.
Finally, the use of a coronograph may also be discussed. In terms of the requested dynamic range, our detection limit at 2 and beyond is around times fainter than the peak intensity of the non occulted star observed with the adaptive optics system under the same conditions. This limit could even be improved by a total exposure time larger than 10 minutes since the limiting factor is in this case the low flux noise. Yet to achieve the same result without a coronograph, the detector intrinsic dynamic range (in the linear regime) would need to be much larger. Indeed, the subtraction of a reference profile requires a very good signal to noise ratio in the wings and this implies a significant number of electrons in the low flux regions.