Predicting Reliability of Silicon MEMS

We have developed a method for predicting the reliability of a brittle microstructure. This method enables a MEMS designer to simulate the reliability of a MEMS device and to make design corrections before investing the time and expense of fabrication and development. Our method combines empirical data gathered from fracture of test specimens, analyzed by Weibull methods, and finite element analysis results, in an algorithm that estimates failure probability for a specified load. MEMS devices typically have multiple and varying surface qualities resulting from fabrication processes. In this work, we demonstrate that the accuracy of the prediction method depends on the inclusion of the Weibull parameters for all surface types.