Utilizing Optomechanics To Enhance Mems Oscillators And Inertial Sensors

Abstract

Virtually every communications system today requires some form of reference oscillators. The need for miniaturized, batch manufacturable oscillators as chip scale timing references stems from the need to replace the well-established, high performing, albeit expensive quartz oscillators without compromising on performance. MEMS oscillators have recently found applications in various consumer electronic applications. With numerous advances in fabrication technology and materials processing, these oscillators are being pushed to create a presence in the high performance base-band market and high frequency applications. Scaling MEMS oscillators to high frequencies presents challenges in terms of reduced transduction efficiencies and material limits on quality factors. Opto-mechanical transduction offers higher sensitivity and opens up possibilities to interrogate high frequency mechanical resonances hitherto inaccessible. MEMS has also found a strong foothold in the several markets for inertial sensors. Typical MEMS accelerometers and gyroscopes operate at low frequencies (kHz) through electrostatic drive and sense. A limiting factor in the sensitivity of these sensors is the electronic noise in the sense circuitry. This thesis investigates how opto-mechanics can benefit the field of MEMS in low-phase noise oscillators and inertial sensors (gyroscopes). Using a previously demonstrated opto-acoustic oscillator (OAO), a driving scheme is presented for an OAO by simultaneously exploiting radiation-pressure (RP) and RF feedback oscillation mechanisms to achieve significantly lower phase noise than could be realized by either phenomenon solely. A theoretical model and experimental results are presented corroborating this scheme, demonstrating a silicon OAO operating at 175 MHz with a phase noise of -128.6 dBc/Hz at 1 MHz offset with 2.77 dBm RF output power, resulting in a 10dB far-from-carrier phase noise improvement. An opto-mechanical transduction scheme is then presented which is selective to sensing the wine glass modes in an on-chip integrated silicon mechanical resonator. By utilizing two orthogonal waveguides for opto-mechanical sensing at the anti-nodes of the desired wine glass mode, the phase difference between the wine glass modes can be used to selectively sense the desired wine glass modes as compared to other modes. The theoretical model for this transduction scheme is presented followed by the mechanical and optical design of several potential resonators. The design of the resonator, drive electrodes, and tunable waveguides are all explored. These designs are experimentally characterized for the 22.3 MHz and 25 MHz wineglass modes. Optical resonances were shown with loaded optical quality factors as high as 95k. Tunable optical coupling was shown with a DC bias gap-closing waveguide design that reduced the extinction of the optical resonance by 7dB with an applied voltage of -43.9V. Differential electrostatic drive was implemented and showed a 15dB suppression of the unwanted wineglass mode. The RF transmission of the wineglass modes for both waveguides showed out-of-phase signals for the wineglass mode of interest and in-phase signals for the other wineglass mode, consistent with what is needed for balanced differential detection. An application considered for these resonator designs and differential sensing scheme is as a vibratory gyroscope. The resonator designs are modified to reduce the wineglass mode splitting to 50-76kHz. Initial COMSOL simulations were performed by applying external rotations and measuring the displacements at the drive and sense from the resulting coriolis accelerations.