2 Experimental set-up

 

  

Figure 1: Left: laboratory interferometric workbench for testing integrated optics beam combiners based on a Mach-Zehnder interferometer concept. The collimated beam provided by the sources (laser, laser diode or white-light source) is splitted in two beams which are focused onto low-birefringence single-mode fibers by off-axis parabolae. Fibers are directly connected to the integrated optics beam combiner. The optical path difference between the beams is modulated by a translating stage to produce a delay. The three outputs of the beam combiner are imaged on a cooled infrared HgCdTe array. The angle between the incident and reflected beams on the beamsplitter is close to $20^\circ$. Right: integrated optics two-way beam combiner. The light from the telescopes is coupled into the integrated optics beam combiner thanks to optical fibers. Two direct Y-junctions provide photometric calibration signals for each beam. A reverse Y-junction combines the two input beams. Each output is imaged by a lens onto an infrared array

We carried out laboratory tests with off-the-shelves integrated optics components designed for micro-sensor applications. The waveguides are made by ion exchange (here potassium or silver) on a standard glass substrate thanks to photolithography techniques [Schanen-Duport et al. (1996)]. The exchanged area is analogous to the core of an optical fiber and the glass substrate to the fiber cladding. Our $5 \mbox{ mm} \times 40 \mbox{ mm}$ component is schematically depicted in the right part of Fig. 1. We use it as a two-way beam combiner with two photometric calibration signals. The component operates in the H atmospheric band (1.43 - 1.77 ) and its waveguides are single-mode in that domain. From an optical point of view, the reverse Y-junction acts as one of the two outputs of a classical beam splitter. The second part of the interferometric signal with a $\pi$ phase shift is radiated at large scale in the substrate. Light is carried to the component thanks to standard non-birefringent silica fibers.

We have set up a laboratory Mach-Zehnder interferometer to test the interferometric capabilities of our components (see the left part of Fig. 1). The available sources are: a 1.54-He-Ne laser, a 1.55-laser diode and an halogen white-light source. The latter is used with an astronomical H filter.

We scan the interferograms by modulating the optical path difference (OPD) with four points per fringe. The delay line speed is restricted by the integration time ($\sim$1 ms for laser sources and $\sim$10 ms for the white-light source to get a sufficient signal-to-noise ratio) and the frame rate (50 ms of read-out time for the full frame). The OPD scan and the data acquisition are not synchronized, but for each image the translating stage provides a position with an accuracy of 0.3 $\!$. The simultaneous recording of the photometric and interferometric signals allows to unbias the fringe contrast from the photometric fluctuations as suggested by [Connes et al. (1984)] and validated by [Coudé du Foresto (1994)].

A typical white light interferogram I₀ is plotted in Fig. 2a together with the simultaneous photometric signals P₁ and P₂. To correct the raw interferogram from the photometric fluctuations, we substract a linear combination of P₁ and P₂ from I₀. The expression of the corrected interferogram is then

$$I_{\rm c}=\frac{I_{0}-\alpha P_{1}-\beta P_{2}}{2\sqrt{\alpha P_{1}\,\beta P_{2}}}$$ (1)

with $\alpha$ and $\beta$ coefficients determined by occulting alternatively each input beam. The resulting corrected interferogram is displayed in Fig. 2b.

  

Figure 2: White-light interferograms obtained in laboratory with integrated optics components: potassium ion exchange component connected with low-birefringent fibers a, b) and silver ion exchange component connected with high-birefringent fibers c). a) Raw interferogram in the H atmospheric band with photometric calibration signals (upper and lower curves with vertical shifts of +0.1 and -0.1) in normalized units. b) Interferogram corrected from the photometry b) in same units. c) Photometry-corrected interferogram obtained with input beams polarized along the neutral axis of high-birefringent fibers