Comparison of capacitive and frequential readout when scaling accelerometers down from Micro- to Nano- Electro Mechanical Systems

TL;DR

This paper compares capacitive and resonant readout methods for scaled silicon accelerometers from MEMS to NEMS, analyzing noise, sensitivity, and scaling effects to identify optimal sensing strategies at nanometer dimensions.

Contribution

It models and compares noise and detection limits of capacitive and resonant accelerometers during downscaling from MEMS to NEMS, highlighting advantages of resonant sensing at nanoscales.

Findings

Resonant sensing outperforms capacitive sensing below a few hundred nanometers.

Quality factor and beam width have limited influence on detection limits in resonant devices.

Scaling rules suggest resonant sensors are better suited for ultra-small accelerometers.

Abstract

This paper shows the effect of scaling silicon accelerometers down from MEMS to NEMS. It models both electronics and Brownian noise sources for both capacitive and resonant devices, and computes the minimum detectable signal attainable. Computed results are remarkably close to published experimental results. It shows the relatively low influence of the quality factor and of the beam width in the resonant case. Different scaling rules are investigated, and it appears that resonant sensing may satisfy some new application requirements, in particular for critical dimensions below a few hundreds of nm, when it is better resolved than capacitive sensing.