Utilizing mechanical micro-lever coupling structure to enhance sensitivity in mode-localized MEMS accelerometer

The sensitivity of the resonant sensors has been proven to be remarkably enhanced by using mode localization of the weakly coupled resonators as the sensing mechanism. This paper reports a mode-localized resonant accelerometer (ML-RXL) with a novel designed micro-lever mechanical coupler, which is connected with two identical resonators and interconnects the strain energy flow 'leaked' from the non-ideal anchors of the two resonators to enable weak mechanical coupling. The coupling stiffness between the resonators can be reduced by optimizing the anchor to make it more ideal to weaken the leakage energy or optimizing the critical dimension of the micro-lever to weaken the transmission of the leakage energy. A finite element method (FEM) simulation model is established and the simulation results demonstrate that the two methods can effectively weaken the coupling strength between the resonators and improve the sensitivity. The wafer-level vacuum-packaged ML-RXL prototype device is fabricated using the silicon-on-insulator process and is performed in both open-loop and closed-loop tests respectively. The open-loop experiment results demonstrate that the sensitivity of the prototype can be improved from 8.86AR/g to 46.28AR/g, which is consistent with the FEM simulation results. The closed-loop experimental results demonstrate that the sensitivity of the prototype is 46.10AR/g under the linear range of 0.22 g. Moreover, the noise floor decreases from 4.39 mu g/root Hz to 1.71 mu g/root Hz within 10 Hz bandwidth, and the bias-instability decreases from 4.27 mu g to 1.6 mu g.

Automated summary

This paper presents a mode-localized resonant accelerometer (ML-RXL) that uses a novel-designed micro-lever mechanical coupler to enhance the sensitivity. By optimizing the anchor design or the critical dimension of the micro-lever, the coupling strength between the resonators can be effectively weakened. Simulation and experimental results confirm the superior performance of the ML-RXL.