Design for Manufacturability a Holistic Approach for the Development of MEMS Inertial Sensors

Micro-electromechanical system (MEMS) inertial sensors aiming for a high accuracy must anticipate and accommodate for manufacturing variations during the design stage, thereby minimizing reliance on post-fabrication active or passive mitigation measures. In this article, we present a prefabrication design approach that utilizes historical manufacturing data to fit models for fabrication process simulation to predict fabricated device geometry, incorporating account for potential variations in fabrication. Leveraging analytical and finite element modeling, we recreate the manufactured device and its operating conditions to predict its performance. We systematically gathered and analyzed data from hundreds of historical SEM images depicting various trench widths ranging from 1 to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$50~\mu $ </tex-math></inline-formula>m and depths from 10 to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$70~\mu $ </tex-math></inline-formula>m, and fit models to predict key fabrication variations, such as undercut and taper angle. Subsequently, the predicted manufactured device was imported into finite element analysis (FEA), enabling us to anticipate the resonance frequency range of our model while accounting for manufacturing inaccuracies and various damping. Our approach predicted a resonance frequency of 469 Hz and a Q-factor of air 42.6 for the manufacture variation included model (MVIM), aligning closely with experimental findings at 460 Hz and a Q-factor of air 41.8. In contrast, the simulated resonance frequency without considering any fabrication variations was 538 Hz for the designed model (DM). This innovative design for manufacturing shows the significance of incorporating fabrication variations and operating conditions into the simulation stage, aiding the further development of MEMS sensors.