The impact of thermal-noise on bifurcation MEMS sensors

In this work, we investigate and quantify intrinsic noise sources in electrostatic MEMS and study their impact on the mass sensitivity of bifurcation-based inertial sensors. Experiments are carried out to measure the power spectral densities (PSD) of the two leading intrinsic noise sources, mechanical-thermal noise and electric-thermal noise. We also present a stochastic model of the sensor that encompasses those noise sources. The model is deployed to obtain the probability distribution of the cyclic-fold bifurcation point in the presence of noise. It shows that thermal noise leads to an uncertainty of 2Hz in the bifurcation frequency. Within that range, stochastic switching occurs between the two co-existing stable orbits. Therefore, the closest operating frequency of the sensor should exclude this region to protect against false positives. This represents the lower bound on the sensor's minimum detectable mass. The predictions of this model were found to be in close agreement with previous experimental results.