Hybridizing Phononic Crystals and Laterally Vibrating MEMS Contour Mode Resonators in Piezoelectric Aluminum Nitride on Silicon-on-Insulator Technology

Piezoelectric films respond to mechanical strain by generating electric charge and likewise inthe inverse. They are widely applied as electromechanical transducers. Aluminum Nitride(AlN) thin-films are piezoelectric and the process of their deposition is compatible with thatof fabricating silicon (Si) integrated circuits (ICs). By applying AlN on Si, piezoelectrically-activatedmechanically vibrating elements (called resonators) can be implemented. Thin-filmpiezoelectric AlN-on-Si (TPoS) resonators can leverage on existing IC fabrication technology.In converting between electrical and mechanical domains, their coupling factors are severalorders of magnitude that of Si resonators capactively-activated via air gaps. However, TPoSresonators suffer a much lower quality factor (Q) compared to Si resonators; Q defines thefrequency selectivity of the resonator and a high value is generally desired.It is generally recognized that one of the dominant causes of limiting Q in TPoS resonators isleakage of acoustic energy through structures that help to suspend the free-standing resonatorcalled tethers. Existing methods to reduce this source of loss are difficult to scale to higherfrequencies (up to GHz – the range at which most AlN resonators have been reported) by thenature of the structures involved.This project aims to create new device concepts by merging TPoS resonators with phononiccrystals (PnCs) – two different classes of devices not normally associated with each other.More specifically here, we focus on two-dimensional (2D) PnCs that should prove easier tofabricate and scale to GHz. PnCs can be likened to acoustic manipulators, which we engineerto reflect acoustic energy leaking out from the tethers back into the TPoS resonator. Thisapproach thus allows us to boost Q but without interfering with the TPoS resonator. Ourpreliminary experimental results have confirmed doubling of Q by the proposed 2D-PnCapproach. Here in this proposal, the challenge of scaling up the frequency of the 2D-PnCresonatorhybrids is split between the first two research objectives. In the first objective, wewill test out different PnC-structures within the conservative range until 0.45GHz. In thesecond objective, based on the findings from first objective, we push the target frequencyfurther to 0.9GHz. Finally, we compare our approach to a recently reported approach thatrelies on tethers created out of 1D-PnCs.Based on our excellent preliminary results and our expertise in modeling and characterization,we are confident that these research objectives are achievable. The impact of the proposal’soutcomes will be significant scientifically and technologically.