Tiny Fiber Fabry-Perot Created with Femtosecond Pulses

Breck Hitz

Most Fabry-Perot interferometers in optical fiber are intrinsic; that is, they are formed in the fiber core between a pair of reflectors. The alternative is an extrinsic Fabry-Perot, which is formed in the air between two sections of the fiber core. Sensors based on either type of Fabry-Perot can detect a change in strain or temperature, which causes the length of the interferometer to change and thereby changes its resonant frequency.

However, fabricating an extrinsic Fabry-Perot entails complex microassembly of various components, which is time-consuming and expensive and results in a device that is less robust than a monolithic intrinsic fiber Fabry-Perot. Nonetheless, extrinsic Fabry-Perot systems have one distinct advantage over their intrinsic brethren: They can serve as chemical sensors. If the air in an extrinsic Fabry-Perot is replaced by another chemical, the interferometer’s resonant frequency will shift as a result of the chemical’s refractive index.

Recently, Hai Xiao and his colleagues at Missouri University of Science and Technology in Rolla fabricated an all-glass extrinsic Fabry-Perot that requires no microassembly. Using a Coherent Ti:sapphire femtosecond laser, they simply cut a micronotch in a piece of single-mode fiber. The notch was deep enough to cut the fiber’s core, forming an air gap between two reflecting surfaces.

Of course, they had to cut the notch very carefully. They mounted the fiber on a precision translation stage under the focused femtosecond laser and monitored the reflection from the Fabry-Perot as it was created with a tunable laser and a power meter. A scanning electron microphotograph of the notch showed that it was about 30 μm long and 72 μm deep — just deep enough to cut the fiber core (Figure 1).

Figure 1. The extrinsic Fabry-Perot interferometer was created by cutting a micronotch (L ∼30 μm) in a single-mode fiber (a). Scanning electron micro-photographs show the notch from the top (b) and in a cutaway cross section (c). Images reprinted with permission of Optics Letters.

The reflection spectrum of the Fabry-Perot indicated that it was fairly lossy, with a maximum reflectivity of only ∼5 percent (Figure 2). Still, the fringe visibility — better than 14 dB — was sufficient for most sensing applications.

The scientists believe that the loss arises from three distinct causes: 1) the roughness of the ablated surfaces, 2) the surfaces’ not being perpendicular to the core, and 3) coupling loss between the free-space interferometer and the fiber core.

They believe that the first problem could be alleviated by ablating smaller amounts of glass with each laser pulse, albeit at the expense of increasing the time required to create the notch. The second problem -- a residual tilt of the notch edges — might be reduced by positioning the fiber more carefully during the cut. The coupling loss, however, increases with the Fabry-Perot cavity length and ultimately may limit the width of the notch.

Figure 2. The low finesse of the Fabry-Perot interferometer results chiefly from imperfections in the laser-ablated sides of the notch shown in Figure 1. The inset shows the reflectance calibrated in decibels.

From the fringe spacing in Figure 2, the scientists calculated that the interferometer length was 30.797 μm, which was in good agreement with the ∼30 μm measured physically from Figure 1.

To demonstrate the robustness of their interferometer, the scientists placed it in an oven and monitored the device’s cavity length at temperatures up to 1100 °C (the oven’s limit). They observed a linear dependence (Figure 3), but when they calculated the coefficient of linear expansion for the glass from these data, they came up with a value that was four times higher than expected.

Figure 3. In a harsh, high-temperature environment, the Fabry-Perot cavity length showed a linear dependence on temperature.

These perplexing results were completely reproducible during several experimental runs, and the scientists finally concluded that they resulted from a bending of the mechanically asymmetric interferometer structure as it was heated. Their observation that fringe visibility decreased somewhat at high temperatures also indicated that the interferometer was bending.

They intend to study the interferometer’s bending behavior in future experiments. Nonetheless, they believe that the device’s straightforward fabrication, its robustness even in harsh environments and its readily accessible cavity make it uniquely qualified as a compact and rugged chemical sensor.

Optics Letters, March 15, 2008, pp. 536-538.