| The need for miniaturized frequency-selective components in electronic systems is clear. The questions are whether and how micro-electro-mechanical systems (MEMS) can satisfy the need. This dissertation aims to address these questions from a scientific perspective. Silicon is the focus of this work, as it can benefit from scaling of the semiconductor industry. Silicon also offers many technical advantages. The characteristics of silicon resonators from 32 kHz to 1 GHz are described. The temperature stability and phase noise of a 6-MHz temperature-compensated oscillator and a 100-MHz temperature-controlled oscillator are reported.;Silicon resonators are discussed in the context of frequency references. Resonator design and characterization, with a focus on quality factor, linearity, and the electrical equivalent circuit, are included. Electrical tuning, electromechanical coupling, finite element modeling, and unexpected findings of these resonators are also described. A manufacturability technique employing batch process compensation is demonstrated.;The significance of this work will be unknown for some time. The frequency control and timing industry is not transparent to many researchers, suppliers, and end-users. The applications and their requirements vary across the gamut. The answer to whether MEMS is beneficial to the industry is a resounding 'Yes!' The aim of this research is to explore the fundamental limitations, provide a foundation for future work, and also paint a clearer picture on how micromechanical resonators can complement alternative technologies. |
