Microelectromechanical resonators are advantageous over traditional LC tanks, off-chip quartz crystals, and SAW devices due to their high quality factors, small size and low power consumption. FET-sensing has been demonstrated in Resonant Body Transistors (RBTs) to reach an order of magnitude higher frequencies than possible with passive resonators due to decoupling of the drive and sense transduction mechanism and the greater sensing efficiency of FET sensing over traditional mechanisms such as capacitive sensing. To-date, two of the most common transduction mechanisms for MEMS resonators are electrostatic and piezoelectric. This thesis aims to develop FET-sensed Si-based resonators with dielectric and piezoelectric materials for high frequency applications.
Monolithic integration of dielectric resonators into CMOS is critical for commercial applications due to reduced size, power, weight and parasitics. A vast majority of CMOS-integrated resonators require a release step to freely suspend their vibrating structures, necessitating costly, complex encapsulation methods. This thesis proposes the development of fully unreleased resonators in CMOS using acoustic isolation structures such as acoustic Bragg reflectors and phononic crystals, this being the first implementation of the latter in CMOS. These dielectrically driven, FET-sensed resonators may be fabricated at the transistor-level of a standard CMOS process, which allows for seamless integration without the need for post-processing or packaging.
While electrostatic drive-based resonators have been primarily explored in this work, piezoelectric resonators have been commercially popular for their high electromechanical coupling factors and resulting low insertion losses for applications ranging from communications to microprocessor clocking. This work explores the integration of an active sensing element such as a FET into a sidewall piezoelectric resonator with CMOS-friendly materials such as AlN, for improved transduction efficiency and to allow scaling of these devices to multi-GHz frequencies.
Prof. Dana Weinstein (thesis supervisor)
Prof. Akintunde Ibitayo Akinwande
Prof. Jeffrey Lang