Development of nonlinear and coupled microelectromechanical oscillators for sensing applications.

摘要

Microelectromechanical systems (MEMS) have gained a great deal of interest over the years due to their small size, low power consumption, ability to be batch fabricated, and ability to be integrated with on-chip electronics. These benefits, coupled with the fact that their minute masses and high frequencies lead to unprecedented sensitivities, make MEMS extremely attractive for sensing applications. To date, the conventional approach to these applications has been to utilize linear, uncoupled microresonators. While this method has proved to be successful, nonlinear and/or coupled MEMS resonators possess rich dynamic behavior that is not obtainable with linear, uncoupled MEMS resonators. This dissertation is focused on exploiting nonlinearity and coupling to improve the performance of MEMS resonators in sensing applications.;The first part of this dissertation investigates a highly tunable nonlinear parametrically excited MEMS for filtering and mass sensing applications. The device utilizes two sets of noninterdigitated comb drives, one for tuning and one for actuation, that allow the effective linear and nonlinear stiffness to be electrostatically tuned. This work focuses on the design, fabrication, and testing of tunable parametrically excited MEMS oscillators to demonstrate a novel linear and nonlinear tuning scheme developed for creating bandpass filters with nearly ideal stopband rejection and sharp response roll-off. The tunability of these nonlinear oscillators is also shown to allow chaos under certain conditions. Using Melnikov's method, a criterion for the existence of chaos is derived. Numerical and experimental investigations verify the existence of chaos for certain parameter sets.;The second part of this dissertation details the theoretical and experimental study of a novel mass sensor array capable of detecting multiple analytes with a single input and single output (SISO) signal. This sensor differs from traditional mass sensor arrays in that it exploits vibration localization in an array of coupled, frequency mistuned microbeams. The modeling, analysis, and design of these mass sensors is presented. In the experimental portion of this work a variety of devices are created to help prove the concept and evaluate the viability of the sensor. In the process, SISO detection and identification of multiple chemicals is demonstrated for the first time and an innovative concept for a high-Q SISO multi-analyte biosensor is presented. The current SISO multi-analyte chemical sensor achieves femtogram mass sensitivities when operated in air, which shows promise for this technology in future sensing applications.