Overcoming Design and Process Challenges in Next-Generation SRAM Cell Architectures
March 8, 2021
Connecting SEMulator3D to Third-Party Design and Analysis Software Using Meshing
April 15, 2021
Micro-Electrical-Mechanical Systems (MEMS) based inertial sensors are used measure acceleration and rotation rate. These sensors are integrated into units to measure motion, direction, acceleration or position, and can be found in a wide range of applications including smart phones, consumer electronics, medical devices, transportation systems, oil/gas exploration, military, aeronautical and space sensor systems. These sensors are almost exclusively microfabricated using deep silicon trench etching. Performance is strongly dependent on the trench profile used to define the shape of all of the components that make up the sensor and can normally only be ascertained after fabrication via electrical testing. Here at Coventor, we’ve been working to further improve our solutions to predict inertial MEMS performance from simulation, with a particular focus on process dependency.
The solution employs our MEMS+ simulation platform and a process sensitive device model, which enables trench sidewall-angle and CD loss to be varied. Simulations are script-driven in a Design of Experiments (DOE) loop via our MATLAB interface and the output post-processed to directly compute quadrature error and sense frequency shift. You can see a reverse engineered gyroscope used as a test vehicle in Figure 1 below [1,2]. The colored structures, one of which is magnified, are comprised of 2um wide beams and suspend the central mass (colored green) to allow it to move and sense rotation. In this example, trench sidewall angle and CD are both defined as a function of die location on the wafer x-axis, which is varied in the DOE. Figure 2 shows predicted quadrature error and sense frequency shift due to representative shifts in trench angle of from 0 to 0.2° and CD loss of from 0 to 150nm.
Figure 1: Reverse engineered gyroscope, with suspension spring displayed in the call-out circle
Figure 2: Predicted gyroscope performance, as a function of die location.
Predicting device performance from simulation offers huge value for a product lifecycle. Here, the impact of process changes (trench profile and CD loss) on sensor performance and yield can be investigated without expensive and time-consuming build and test iterations. This enables engineers to digitally explore design ideas, and together with process experts, gain insights into design and process sensitivities, yield and failures modes – all to meet highly defined customer needs. Work continues in this exciting area and we look forward to providing additional updates a in future blog!
References
[1] Yan Loke (STMicroelectronics), “The THELMA MEMS Process from ST and Availability from the CMP Platform”, SEMICON West 2013
[2] Chipworks blog, “Apple uses Nine Degrees-of-Freedom Sensing in iPhone 4”
![Fig. 2. Silicon spin qubit centered in the dotted circle, control and readout signals (M, P, R, T, and Q) are shown in the inset. Simplified schematics of the quantum point contact and corollary circuits are shown. The voltage source is implemented as a digital-to-analog converter at room temperature. (Taken from [6])](https://www.coventor.com/wp-content/uploads/2022/05/Fig.-2.-Silicon-spin-qubit-centered-in-the-dotted-circle-control-and-readout.png)
![Figure 2: Schematic view of standard CMOS including an electro-mechanical switch monolithically integrated into BEOL [2]. Courtesy: University of California, Berkeley](https://www.coventor.com/wp-content/uploads/2022/04/Figure-2-April-2022-MEMS-Blog-for-Coventor-website.png)
