Image reconstruction from optical interferometric data

Abstract

Generating images of astronomical sources is a significant challenge for optical interferometry, although this is routinely done for similar observations using radio telescopes. One needs to measure both the ‘fringe-visibility’ amplitude and phase of the incident light waves. However, while the atmosphere usually changes on long timescales at radio wavelengths, in the optical the fluctuations are on the order of a few milliseconds, thus resulting in corrupted fringe-visibility phases. One particular image-reconstruction approach that has recently benefited from significant improvement1, 2 uses the square of the visibility amplitude and the so-called closure phases. A closure phase is the sum of the visibility phases around a triangle of interferometer baselines, resulting in the cancellation of atmospheric-turbulence effects. This technique has reliably reproduced test images,3 but it also leads to the loss of a fraction of the phase information and can reduce the signal-to-noise ratio of the final images. Therefore, one of themajor challenges is the development of techniques that recover as much phase information as possible. We have been working on the generation of such techniques for use with the Navy Prototype Optical Interferometer (NPOI). An important recent achievement is the implementation of an integration method4, 5 that first aligns the phases of the 2ms interferometric observations and subsequently adds the data frames coherently over much longer timescales. This technique increases the observational signal-to-noise ratio and retains the maximum phase information. We have also improved two methods aimed at correcting the observed phases for atmospheric effects. We first developed a differential phase technique6 to recover the phase information Figure 1. Images of the Hα hydrogen emission of the eclipsing binary star β Lyrae at orbital phases of 0.24 (left) and 0.78 (right). The movement of the Hα emission relative to the continuum photocenter (green dots) is evident. The elongation of the Hα images is caused by the shape of the image-reconstruction beam. The green bar corresponds to 2 milliarcseconds.