Invention is not a straightforward process; ideas often bear fruit in applications quite different from what was first intended. Faced with the problem of measuring the position of an explosive shock front, Jonathan Weiss, a physicist at Sandia National Laboratories (Albuquerque, NM), designed a fluorescent optical fiber that is progressively destroyed as the shock front moves forward; pumped by a laser, the fiber returns fluorescent light in proportion to the length of surviving fiber. Pondering his invention, Weiss saw another possibility—the extinguishing of fluorescence without destroying the medium—and came up with an all-optical liquid-level sensor.
Weiss has replaced the fiber with a dye-containing slab, introducing light into the end of the slab at an angle so that it zigzags down the slab via total internal reflection. The slab is placed dipstick-style into a tank of liquid and fixed in place. Light that strikes the face of the slab below the level of the fluid escapes the slab, eliminating fluorescence below that point. Return light collected at the top of the slab provides a measure of the amount of fluorescence occurring in the slab. By bringing pump light into the slab via a fiber plus collimating lens and collecting fluorescence with a fiber bundle, Weiss can place both the light source and detector far from the sensor. Not only are electromagnetic interference problems thus eliminated, but the sensor can be used with flammable liquids such as jet fuel. (In fact, the idea for the sensor came to Weiss only weeks after the TWA 800 jet crash, he explains.)
Made of methyl styrene doped with an organic dye, the slab in the test setup measured 45.7 cm high, 1.27 cm wide, and 5.08 cm deep (its depth could be reduced as long as the side faces are of high-enough optical quality). Pump sources included an argon-ion laser with a wavelength of 476, 488, or 496 nm, and a light-emitting diode with a center wavelength of 471 nm and a full-width-at-half-maximum spectral bandwidth of 40 nm. The performance of the device varied depending on the pump source and wavelength (488 nm was best for the dye used), pointing to the need for matching the light source to the fluorescent medium.
To eliminate discrete jumps in signal as the liquid level is lowered, the collimated beam can be made wide enough so that light strikes the slab faces at all points along the sensor's vertical range. A mirrored surface at the bottom of the slab sends the pump light zigzagging back up and provides a single, large, intended jump in signal when the fluid level dips below the mirror; the jump can serve as an alarm. Although variations of signal with temperature change were not tested, Weiss says that temperature dependence of the dye can be optically monitored (and then factored out) either by including an auxiliary pump beam or by using the fluorescence generated in the first bounce of the existing pump beam as a check.
Absorption of the pump and fluorescent wavelengths by the dye and host medium creates a nonlinear detector response that can be easily calibrated. Other loss mechanisms include efficiency of the optics and the quantum efficiency of the dye. As one alternative to the methyl styrene host (whose chemical resistance to jet fuel remains untested), Weiss envisions inorganic slabs made of fused silica doped with neodymium, pumped in the near-infrared and fluorescing at 900, 1060, and 1350 nm. Not only is fused silica relatively inert to chemicals, but its refractive index is lower than that of methyl styrene, which opens up the possibility of use with liquid oxygen and hydrogen fuels.
