Dissertation Talk: High-Q Strong Coupling Capacitive-Gap Transduced RF Micromechanical Resonators
Presentation: Berkeley Sensor & Actuator Center (BSAC): EE | December 3 | 2-3 p.m. | Cory Hall, Swarm Immersion, 490 Cory Hall
Alper Ozgurluk, Prof. Clark T.-C. Nguyen Research Lab, BSAC
Single-digit-nanometer electrode-to-resonator gaps have enabled 200-MHz radial-contour mode polysilicon disk resonators with motional resistance Rx as low as 144Ohm while still posting Qs exceeding 10,000, all with only 2.5V dc-bias. The demonstrated gap spacings down to 7.98nm are the smallest to date for upper-VHF micromechanical resonators and fully capitalize on the fourth power dependence of motional resistance on gap spacing. High device yield and ease of measurement debunk popular prognosticated pitfalls often associated with tiny gaps, e.g., tunneling, Casimir forces, low yield, none of which appear. The tiny motional resistance, together with kt2s up to 1% at 4.7V dc-bias and kt2-Q products exceeding 100, propel polysilicon capacitive-gap transduced resonator technology to the forefront of MEMS resonator applications that put a premium on noise performance, such as radar oscillators.
Pursuant to integrating such a sub-10nm MEMS resonator atop CMOS, this dissertation also reconsiders metal as a viable resonator structural material. To this end, introduction of tensile stress via localized Joule heating has yielded some of the highest metal MEMS resonator Qs measured to date, as high as 48,919 for a 12-MHz ruthenium micromechanical clamped-clamped beam (CC-beam). The high Qs continue into the VHF range, with Qs of 7,202 and 4,904 at 61 and 70 MHz, respectively. These marks are substantially higher than the 6,000 at 10 MHz and 300 at 70 MHz previously measured for polysilicon CC-beams, defying the common belief that metal Q cannot compete with conventional micromachinable materials. The low-temperature ruthenium metal process, with highest temperature of 450°C and paths to an even lower ceiling of 200°C, further allows for MEMS post-processing directly over finished foundry CMOS wafers, thereby offering a promising route towards fully monolithic realization of CMOS-MEMS circuits, such as needed in communication transceivers. This, together with its higher Q, may eventually make ruthenium metal preferable over polysilicon in some applications.