1 Introduction

The study of stellar environment offers a number of exciting topics for which detection and imaging of faint morphological features close to an unresolved source are concerned. This goal relates to both compact features such as faint companions or planets and diluted features such as circumstellar envelopes, protoplanetary disks, jets and other structures. In this paper we consider instead detection of faint companions.

Besides nulling techniques applied to diluted apertures (Bracewell [1978]; Woolf & Angel [1997]; Mennesson & Marriotti [1997]) the use of coronagraphy with compact apertures nowadays concerns the classical Lyot Coronagraphy (Lyot [1930]) adapted to stellar studies (Malbet [1996]; Beuzit et al. [1997]), the Achromatic Interfero Coronagraphy (Gay & Rabbia [1996]), and the Phase Mask Coronagraphy (Roddier & Roddier [1997]).

The classical Lyot coronagraphy uses an opaque mask located at an intermediate image plane, so as to obscure the center of the field. This technique does not allow a very close-sensing of the star environment since the central area of the field is masked up to a number of Airy rings (say 3 to 10) and provides sensing capabilities ranging between 0.3 and 2 arcsec from the star (Beuzit et al. [1997]; Nakajima et al. [1995]). Also, removal of the light from the star is not complete because diffraction features in the image plane cannot be totally eliminated even in perfect optical conditions (in spite of using a Lyot stop in the next pupil plane). On the contrary, in similar conditions our coronagraph (simply by its principle) is free from such diffraction features.

In the phase mask approach the opaque mask is replaced by a transparent and dephasing one and should allow the sensing of the source's environment as close as the first Airy dark ring. However, even in perfect optical conditions (as for the Lyot approach) its extinction capability suffers from spurious diffraction features. Moreover extinction is intrinsically chromatic which limits the working spectral bandwidth.

Our coronagraph is likely to be a promising concept since it is able to look very close to the star (as close as a fraction of the first Airy ring), avoids the use of a Lyot stop (which would reduce the effective aperture diameter) and can work at a large spectral bandwidth, which tends to increase detection capabilities and allows spectroscopic study of faint companions.

In Sect. 2 we recall the principle of the concept and the associated generic set-up as well as specific features and inherent limitations in the ideal case. In Sect. 3 departures from the ideal case are considered through a theoretical algebraic derivation. In Sect. 4, expected detection capabilities for ground-based operation are evaluated.