Advanced HARPSS Processes for High Q and High Frequency Inertial Sensors

Abstract

The surging growth of the Internet of Things (IoT) is driving demand for microelectromechanical systems (MEMS) devices in areas such as navigation and tracking systems, smart grids, 5G and building automation. The objective of this thesis is to introduce several advances into the HARPSS fabrication platform to improve the performance metrics of such resonant MEMS devices and enable new functionalities. Many resonators including bulk acoustic wave (BAW) gyroscopes require ultra-high quality factor (Q) in the order of 1M to improve signal to noise ratio. The use of isotropic device layers such as thick epitaxially-grown polycrystalline silicon (epipoly) can enable close to Akhiezer limited Qs were used to fabricate high-Q gyroscopes was implemented. These gyroscopes were mode matched to yield an ARW of 0.01deg/rthr, with a scale factor as large as 7nA/dps, which can be used in IoT applications such as smart cars and wearable health monitoring, where detection of small movements is of prime importance. The second part of this thesis involves research on a novel design of wafer-level-packaged (WLP) high-Q capacitive Distributed Lame Mode resonators (DLR), which enable significant improvement in their figure of merit, extending their frequency into the VHF range while keeping their motional resistance low (<1kΩ), with high Qs (250k). Lastly, the multi-transduction capability of capacitive resonators can be combined using piezoelectric transduction, and by scaling the capacitive gaps to sub-100nm, high performance BAW sensors can be enabled. In the third part of this thesis, a 360kHz frequency-output piezoelectrically-transduced resonant BAW accelerometer is introduced, combining a sub-100nm gap HARPSS process with a metal-less high-frequency AlN-on-silicon resonator, with a quality factor of 3800 in air. Such high BW accelerometers can be used in IoT health monitoring applications.