THE KECK INTERFEROMETER NULLER

The Keck Interferometer Nuller (KIN), the first operational separated-aperture infrared nulling interferometer, was designed to null the mid-infrared emission from nearby stars so as to ease the measurement of faint circumstellar emission. This paper describes the basis of the KIN's four-beam, two-stage measurement approach and compares it to the simpler case of a two-beam nuller. In the four-beam KIN system, the starlight is first nulled in a pair of nullers operating on parallel 85 m Keck-Keck baselines, after which 'cross-combination' on 4 m baselines across the Keck apertures is used to modulate and detect residual coherent off-axis emission. Comparison to the constructive stellar fringe provides calibration. The response to an extended source is similar in the two cases, except that the four-beam response includes a term due to the visibility of the source on the cross-combiner baseline—a small effect for relatively compact sources. The characteristics of the dominant null depth errors are also compared for the two cases. In the two-beam nuller, instrumental imperfections and asymmetries lead to a series of quadratic, positive-definite null leakage terms. For the four-beam nuller, the leakage is instead a series of correlation cross-terms combining corresponding errors in each of the two nullers, which contribute offsets only to the extent that these errors are correlated on the timescale of the measurement. This four-beam architecture has allowed a significant (∼order of magnitude) improvement in mid-infrared long-baseline fringe-visibility accuracies.