April 25, 1995
We are investigating the effects of uncompensated dispersion on fringe visibility in a long baseline fiber optic interferometer. Our goals.are to determine how serious the problem of dispersion is and whether the dispersion effects can be minimized by judicious choice of central wavelength. Our analysis extends the work done by W. J. Tango (Applied Optics, vol. 29, pp. 516-521, 1991), who analyzed the effects of dispersion in an air path interferometer, to a silica fiber system where both material and waveguide dispersion must be considered. We expanded the optical path delay in a Taylor series, and used the first five terms to calculate the central fringe visibility versus fiber length difference. The visibility versus fiber length difference is significantly improved at wavelengths near 1300 nm, where the second order waveguide and material dispersion terms cancel. We also measured the visibility versus fiber length difference in a laboratory Mach-Zehnder interferometer to verify our theoretical predictions.
We are also investigating the effects of finite pupil size in a fiber optic stellar interferometer. We have extended the results of S. Shaklan who calculated the fringe visibility of a fiber optic interferometer with the assumption that the telescope pupils are small enough that the mutual intensity is constant over the width of the pupil autocorrelation (Ph. D. Thesis, University of Arizona, 1989). We found that when the width of the pupil autocorrelation is not negligible compared to the feature size of the mutual intensity function, the measured visibility function can become significantly distorted. However, if the pupil functions are known, the measured visibility can be corrected to get the true mutual intensity. We plan to verify this experimentally.