Novel Architecture for Inertial Grade SOI MEMS Inertial Sensors Ahmed Abdel Aziz 1 , Abdel-Hameed Sharaf 1, 2 , Mohamed Serry 3 and Sherif Sedky 1, 4 1 Science and Technology Research Center (STRC), American University in Cairo (AUC), Cairo, Egypt 2 National Center for Radiation Research and Technology (NCRRT), Egyptian Atomic Energy Authority (EAEA), Cairo, Egypt 3 King Abdullah University of Science and Technology (KAUST), Thuwal, Saudi Arabia 4 Physics Department, AUC, Cairo, Egypt E-mail: sedky@aucegypt.edu and aksa@aucegypt.edu This work reports, for the first time, on a novel architecture for realizing high performance miniaturized micromachined inertial sensors based on Silicon on Insulator (SOI) technology. The new sensors are essential components for high precision Inertial Navigation Systems (INS) serving a wide range of applications varying from automotive to space. Similar to the state-of-the-art designs and architectures, our novel architecture utilizes SOI technology, which is crucial for realizing such electro-mechanical sensors. This is mainly due to the fact that it allows for having stress free suspensions and relatively very large proof masses (for bulk rather than surface micromachined MEMS) coupled with the flexibility of choosing between the presence or absence of electrical isolation between the upper and lower parts of the SOI wafer depending on the sensor’s layout. This accordingly enables having the electrical contacts isolated from the rest of the sensor’s structure. Typically, vibratory micromachined gyroscopes (VMG) and accelerometers are designed in-plane, i.e. the suspension and the proof mass vibrate in the same plane parallel to the substrate surface. Such in-plane architecture reduces dramatically the sensor's fill factor (typically around 17%); hence the Brownian noise level and overall sensitivity are degraded due to the reduced proof mass. Thus, a dramatic enhancement in performance can be achieved by improving the fill factor as this has a direct impact on increasing the proof mass significantly. Accordingly, we introduced, for the first time, the vertically suspended micromachined inertial sensor displayed in Figs. 1 and 2. Such design allows increasing the fill factor by more than a factor of four (up to 80-90%). Moreover, this design offers a decoupled geometric design of the proof mass and the support stiffness dimensions. The performance of the proposed sensors was analyzed using Finite Element Method (FEM) to determine the natural mode shapes and frequencies in addition to the mechanical stability and the drive-sense mode decoupling (1-2% similar to common designs) for the MEMS gyroscopes. Tables I and II give a comparison between the proposed design and two of the state-of-the-art designs [1, 2]. By inspecting Table I, it is clear that for the same device area, the proof mass is increased by more than a factor of 20 due to the increased fill factor (80%). In addition, the proof mass thickness is increased to 200 µm. It should be noted that the proposed design offers the flexibility of increasing the thickness of the proof mass up to the whole wafer thickness without introducing any additional lithography masks as in the conventional SOI architectures where extra-mass is employed [1]. Moreover, the mode resonance frequency is significantly reduced due to the reduced mode stiffness. This serves to suppress the thermo-elastic damping and support losses that dominate in high performance vacuum operation. The new architecture also suppresses the noise floor by more than one order of magnitude, enhances signal sensitivity by more than two orders of magnitude and Signal to Noise Ratio (SNR) up to three orders of magnitude. Table II reflects similar improvements for the proposed inertial grade MEMS accelerometer. Furthermore, the proposed process flow for the novel sensors consists only of four masks where all the steps rely only on Deep Reactive Ion Etching (DRIE). This reveals that the main limitation to fabrication is the DRIE aspect ratio. The etching aspect ratio has been found experimentally to be as high as 1:50 as illustrated by the Scanning Electron Microscopy (SEM) image in Fig. 3 for deep 80 µm Silicon trenches. It is worth noting that such a high ratio does not impose a limitation to the fabrication of the various sensors’ geometries investigated. Finally, the new architecture allows additionally, and for the first time, to realize a quad mass sensing scheme, i.e. dual differential sensing, for common mode 978-1-4244-5232-3/09/$25.00 ©2009 IEEE