MEMS Navigation-Grade Electro-Optical Accelerometer

Accurately measuring displacement of mechanical components on micro-scale devices is often the limiting factor affecting resolution for many MEMS-based sensors including accelerometers, Coriolis-based gyroscopes, pressure sensors, temperature sensors, and magnetometers. This paper introduces a novel electro-optical displacement detection scheme which has been applied to navigation-grade MEMS accelerometers under funding from the Office of Naval Research (ONR). The optical detection method is extremely sensitive to displacements, and generates large steady-state currents thereby eliminating the necessity to co-integrate sensitive amplification electronics next to the sensing element. Experimental measurements have resulted in a 10 femtometer noise floor without the need to co-integrate amplifying control electronics within the same package as the accelerometer sensor. This technique is based on the monolithic integration of a Fabry-Perot optical interferometer and a photodiode. 1.0 INTRODUCTION High sensitivity accelerometers are critical for the next generation navigation and guidance systems including tight coupling to existing GPS engines, pressure sensors, and platform stabilization for space applications. The impetus for a MEMS-based inertial accelerometer is based upon the hopes of realizing a low cost, small, lightweight and highly sensitive alternative to existing macro-scale approaches. The successful fabrication of a low cost, high-sensitivity MEMS accelerometer will result in new applications for both consumer and military users that aren’t feasible with current technologies. Examples include personal hand held navigators for military and consumer applications, as well as GPS-denied navigation applications such as in valleys, urban areas, and within buildings and caves. Numerous MEMS devices used for the detection of motion, position, pressure, and temperature rely upon the precise measurement of the displacement of a proof mass attached to a spring. To detect these displacements several techniques have been employed including measurement of charge on a variable capacitor [1-4], change in optical transmission through a Bragg grating [5], change in resistance of a piezoresistor [6-9], and more recently measurement of tunneling current through a well-controlled airgap [10-13]. Of these techniques, tunneling displacement sensors hold the best promise for realizing small, highly sensitive transducers required for navigation and acoustic applications. Tunneling transducers take advantage of the exponential sensitivity in tunneling current to the tunneling gap distance in order to realize appreciable changes in current with input acceleration. Typical steady state tunneling current is Waters, R.L.; Jones, T.E. (2007) MEMS Navigation-Grade Electro-Optical Accelerometer. In Military Capabilities Enabled by Advances in Navigation Sensors (pp. 12-1 – 12-16). Meeting Proceedings RTO-MP-SET-104, Paper 12. Neuilly-sur-Seine, France: RTO. Available from: http://www.rto.nato.int. MEMS Navigation-Grade Electro-Optical Accelerometer 12 2 RTO-MP-SET-104 UNCLASSIFIED/UNLIMITED UNCLASSIFIED/UNLIMITED approximately 1-2 nA [14]. In order to achieve this steady-state current, an airgap on the order of approximately 10 Å (1.0 nm) must be maintained via force rebalancing. Due to the limited tunneling area (one metal atom on the surface of each side of the airgap), larger tunneling currents are difficult to obtain. In addition, tunneling transducers may have high temperature-sensitivity since thermal expansion coefficient mismatches can alter the tunneling gap distance [15]. Finally, tunneling accelerometers are reported to have a significant 1/f (flicker) noise contribution and a high variability in the tunneling barrier height [16]. Capacitive-based accelerometers have been the most widely developed and commercialized due to the operational simplicity of the device and available fabrication methods. Obtaining high resolution, high stability capacitive accelerometers, however, has been difficult due to the low steady-state capacitances and the tight processing tolerances. To increase the sensitivity of these capacitive accelerometers, the size and surface area of the devices are generally increased, sense gaps decreased to a micron or less, and the quality factor, Q, increased. All of these factors make high resolution, high stability capacitive sensors difficult to realize. We present a novel optical transducer concept for the precise measurement of a proof mass attached to a spring that has raw sensitivity greater than that of a tunneling transducer and stability exceeding those of current capacitive accelerometers. The concept involves the integration of a Fabry-Perot interferometer and a photodiode on a (100) Si substrate using surface and / or bulk micro-machining techniques resulting in a compact device with minimal parasitic elements. MEMS-based Fabry-Perot Interferometers have recently been investigated for use in numerous applications including chemical sensing [17], Wavelength Division Multiplexed (WDM) optical communications [18], and pressure sensors [19]. The monolithic integration of a Fabry-Perot interferometer and a p-n photodiode on a (111) silicon substrate has been reported elsewhere for use as a versatile switch and amplifier [20]. 2.0 CONCEPTUAL THEORY A Fabry-Perot cavity is extremely sensitive to changes in the effective cavity length. A Fabry-Perot cavity first devised by C. Fabry and H. Perot in 1899 utilizes multiple beam interference. In the field of optics, it is usually used to measure wavelengths of light with high precision and study the fine structure of spectral lines. In its simplest form, a Fabry-Perot cavity consists of two optically flat, partially transmissive and parallel mirrors separated by a distance, y. In the accelerometer application, that distance is designed to be a function of the acceleration, a. In such a case where one of the mirrors is allowed to move, the structure is referred to as an interferometer. The two mirrors form an optically resonant cavity whereby the transmission of monochromatic light through the cavity can be made to be highly sensitive to displacement of one mirror with reference to the second fixed mirror. For an ideal Fabry-Perot cavity, the two reflecting surfaces are separated by a medium (generally air) with thickness, y(a), refractive index, n, and absorption coefficient, α. The ideal Fabry-Perot cavity is illustrated in Figure 1 below. Assuming a symmetric structure (r01 = -r12 and t01 = t12) with no absorption within the mirrors or the cavity at the wavelength of interest, an analytical expression for the photo-generated current as a function of the effective airgap displacement, y(a), can be obtained as 2 1 1 U F P I in ph + R = (1) ( ) ( )      = a y n U 1 2 sin λ π (2) MEMS Navigation-Grade Electro-Optical Accelerometer UNCLASSIFIED/UNLIMITED In equations (1-2), F is the Finesse of the cavity, Pin is the input optical power incident normal to the surface of the upper mirror, is the responsivity of the photodiode in amperes/watt (A/W), and λ is the wavelength of the incident light. In equations (1-2) it is assumed that the mirrors are lossless and absorption occurs only in the medium between the mirrors. R