To Boldly Go Where No Sensor Has Gone Before

Breck Hitz

Unlike electrical sensors, fiber optic sensors are impervious to electromagnetic interference, and they can be daisy-chained — connected in series — so that many sensors can be monitored simultaneously with a single detection instrument. But fiber optic sensors and electrical sensors share an inability to function at high temperatures.

Recently, Y. Jun-Jiang Rao and his colleagues at the Research Center for Optical Fiber Technology at the University of Electronic Science and Technology of China in Chengdu have demonstrated a fiber optic strain sensor that can operate at a temperature of 800 °C, hundreds of degrees higher than other sensors can withstand. Moreover, their simple fabrication technique makes the sensors easy to mass-produce inexpensively. They believe that these low-cost, high-quality sensors may lead to a revolution in the field of fiber optic sensors.

Figure 1. The scientists fabricated the Fabry-Perot sensor by micromachining a gap in the core of a photonic crystal fiber, whose cross-sectional structure is shown on the left. The photo on the right shows the ~46-μm gap that they machined with 1600 pulses from an F₂ excimer laser. Images reprinted with permission of Optics Letters.

Today’s commercial fiber optic sensors are based on optical interference in fiber Bragg gratings or in Fabry-Perot interferometers. In either case, when the device is strained, its interference spectrum shifts in proportion to the amount of strain. A problem with Bragg grating sensors is that it’s very difficult to separate the effects of temperature from those of strain. Fabry-Perot sensors don’t have this problem, but because they are dependent on nonfiber optical elements, they cannot function at elevated temperatures.

Ideally, one would like an in-line Fabry-Perot; that is, one completely built into the fiber. Previous attempts at this goal have centered on splicing a short-hollow-core fiber between two lengths of single-mode fiber, so that the Fabry-Perot cavity is defined by the two end facets of the single-mode fibers. Although this approach is successful up to a point, it entails time-consuming manual assembly that not only is expensive but also invites the possibility of contamination.

Figure 2. The fringes generated by the Fabry-Perot cavity in Figure 1 had a visibility of ∼26 dB.

The scientists from China have taken a different approach to fabricating an in-line Fabry-Perot strain sensor. Using the 157-nm radiation from a fluorine-excimer laser made by Coherent Inc. of Santa Clara, Calif., they micromachined a tiny Fabry-Perot resonator in an endlessly single-mode photonic crystal fiber from Crystal Fibre A/S of Birkerød, Denmark. Sixteen hundred laser pulses at 20 Hz produced a clean break in the fiber core, with mirrorlike surfaces on the ablated facets (Figure 1). The resulting interference fringes had a visibility of ~26 dB, which the investigators believe to be the best that has ever been observed from an uncoated Fabry-Perot strain sensor (Figure 2).

To test the sensor’s sensitivity to strain and insensitivity to temperature, they placed it in an oven and applied a carefully calibrated stress. The resulting strain was linear with the variation in cavity length that the scientists calculated from the shift of the interference peaks in Figure 2, and the measurement was almost entirely independent of temperature (Figure 3a).

Figure 3. Measurements at 20, 500 and 800 °C were virtually identical (a). The small but seemingly random temperature dependence was the result of thermal compensation in the interferometer’s structure (b).

The temperature independence arose from the fact that the Fabry-Perot is, in effect, a thermally compensated interferometer. A temperature increase expands the core and so pushes the two interferometer facets closer together. At the same time, the cladding also expands, pushing the facets apart. These two effects compete with each other, resulting in a small but haphaz-ard temperature dependence (Figure 3b).

Optics Letters, Nov. 1, 2007, pp. 3071-3073.