Deep, stable nulling of starlight requires careful control of the amplitudes and phases of the beams that are being combined. The detection of earth-like planets using the interferometer architectures currently being considered for the Terrestrial Planet Finder mission require that the E-field amplitudes are balanced at the level of similar to 0. 1%, and the phases are controlled at the level of 1 mrad (corresponding to similar to 1.5 nm for a wavelength of 10 mum). These conditions must be met simultaneously at all wavelengths across the science band, and for both polarization states, imposing unrealistic tolerances on the symmetry between the optical beamtrains. We introduce the concept of a compensator that is inserted into the beamtrain, which can adaptively correct for the mismatches across the spectrum, enabling deep nulls with realistic, imperfect optics. The design presented uses a deformable mirror to adjust the amplitude and phase of each beam as an arbitrary function of wavelength and polarization. A proof-of-concept experiment will be conducted at visible/near-IR wavelengths, followed by a system operating in the Mid-IR band.
