Improved achromatization of phase mask coronagraphs using colored apodization

Context. For direct imaging of exoplanets, a stellar coronagraph helps to remove the image of an observed bright star by attenuating the diffraction effects caused by the telescope aperture of diameter D. The dual zone phase mask (DZPM) coronagraph constitutes a promising concept since it theoretically offers a small inner working angle (IWA similar to lambda(0)/D where lambda(0) denotes the central wavelength of the spectral range Delta lambda), good achromaticity, and high starlight rejection, typically reaching a 10(6) contrast at 5 lambda(0)/D from the star over a spectral bandwidth Delta lambda/lambda(0) of 25% (similar to H-band). This last value proves to be encouraging for broadband imaging of young and warm Jupiter-like planets. Aims. Contrast levels higher than 10(6) are, however, required for observing older and/or less massive companions over a finite spectral bandwidth. An achromatization improvement of the DZPM coronagraph is therefore mandatory to reach such good performance. Methods. In its design, the DZPM coronagraph uses a gray (or achromatic) apodization. We replaced it by a colored apodization to increase the performance of this coronagraphic system over a wide spectral range. This innovative concept, called colored apodizer phase mask (CAPM) coronagraph, is defined to reach the highest contrast in the exoplanet search area. Once this has been done, we study the performance of the CAPM coronagraph in the presence of different errors to evaluate the sensitivity of our concept. Results. A 2.5 mag contrast gain is estimated from the performance provided by the CAPM coronagraph with respect to that of the DZPM coronagraph. A 2.2 x 10(-8) intensity level at 5 lambda(0)/D separation is then theoretically achieved with the CAPM coronagraph in the presence of a clear circular aperture and a 25% bandwidth. In addition, our studies show that our concept is less sensitive to low than to high-order aberrations for a given value of rms wavefront errors.