Sensor: The essence of the Faraday effect is that the polarization direction of light in a crystal rotates in the presence of a magnetic field along the propagation direction. The rotation direction is the same even when the light propagation direction is reversed, so that the rotation angle is doubled by sending the light back through the crystal with a mirror. This configuration has long been used in the fabrication of optical isolators. The crystal is placed in a magnetic field, and polarizers oriented at 45 degrees are positioned at either end to allow light to travel in one direction but not the reverse.
A Faraday sensor can be made by removing the magnet and detecting changes in the transmitted light intensity in the presence of a field. Often the rear surface of the crystal is made reflective to serve as the mirror, doubling the sensitivity. Unpolarized light from a light emitting diode can be transported to the crystal in an optical fiber, and a single piece of Polaroid film serves as both polarizer and analyzer. The change in return light intensity from the crystal is monitored and used to determine the presence or absence of a magnetic field.
The proportionality constant for Faraday rotation, the Verdet constant, is a property of the crystal material, but it is particularly large for a class of materials known as iron-garnets. Furthermore, the Faraday effect in iron garnets saturates at large fields, so that the crystal can be made to produce not more than 90 degrees of total rotation at saturation to ensure that the transmitted signal is a monotonic function of the field. Although wheel sensors have been made from yttrium iron garnet crystals,1 we are especially interested in bismuth iron garnet (BIG) because its Verdet constant is so large that the crystal can be a film thin enough to eliminate the need for a collimating lens, a significant cost savings in a sensor. However domain boundaries in iron garnets can diffract a significant fraction of the light out of the beam when the beam diameter is small, as is the case when the lens is eliminated. The diffraction decreases, reducing the signal, in the presence of a magnetic field because the domains are aligned in a field. Thus if there is significant diffraction, the signal increases with magnetic field just when it is decreased by the Faraday rotation, and the two effects oppose each other. A further complication arises because the domain structure is unstable in the presence of the changing fields; otherwise a satisfactory sensor could be made using the domain effect itself simply by using unpolarized light.
Since the Faraday rotation depends on the component of the optical path along the magnetic field direction, all the light is rotated by the same angle, whether or not it has been diffracted, so long as the field is uniform over the crystal dimensions. We have found that by carefully adjusting the BIG film thickness, the fiber optic geometry, and the mirror position, we can collect similar large fractions of the diffracted and undiffracted light. In this way we have succeeded in producing a configuration in which the diffractive effects are minimized, leaving a stable signal that is determined primarily by the Faraday effect.
Conclusion: Although it is possible to fabricate a fiber optic wheel rotation sensor using diffractive effects, eliminating the need for a polarizer,2 a more economical configuration is realized by utilizing a polarizer and the Faraday effect. Crystals of BIG can be made thin enough to allow recollection of easily-measurable amounts of modulated light into an optical fiber without a lens. The resulting sensor is reliable, and like other fiber optic sensors, it can be sealed to keep the light in and dirt and grease out. In this way we have met our objectives in developing a useful fiber optic wheel speed sensor that we believe will also be rugged and economical to fabricate.