Stabilising a nulling interferometer using optical path difference dithering

Context. Nulling interferometry has been suggested as the underlying principle for the Darwin and TPF-I exoplanet research missions. Aims. There are constraints both on the mean value of the nulling ratio, and on its stability. Instrument instability noise is most detrimental to the stability of the nulling performance. Methods. We applied a modified version of the classical dithering technique to the optical path difference in the scientific beam. Results. Using only this method, we repeatedly stabilised the dark fringe for several hours. This method alone sufficed to remove the 1/f component of the noise in our setup for periods of 10 min, typically. These results indicate that performance stability may be maintained throughout the long-duration data acquisitions typical of exoplanet spectroscopy. Conclusions. We suggest that further study of possible stabilisation strategies should be an integral part of Darwin/TPF-I research and development.