Abstract
In order to expand undergraduate education in microelectromechanical systems (MEMS) and nanotechnology, a series of sensors were designed with the intent of integrating the design process into the project portion of a micro/nano systems course. The majority of the design work focused on piezoresistive, diaphragm-based pressure sensors, utilizing multiple diaphragm sizes and geometries. These sensors were chosen for their geometrical simplicity and their ability to be manufactured using available photolithographic techniques. Finite Element Analysis (FEA) was used to gain a deeper understanding of stress distribution, leading to better design decisions. Results were validated using analytical models, and multiple iterations were performed to understand stress variation relative to diaphragm specifications. This analysis procedure aided in optimizing piezoresistor placement and assessing the effect of process variation on device performance. The documented procedures were followed by students to design and analyze the sensor type of their choice.
The Challenge of MEMS Education
Teaching MEMS and nanotechnology at the undergraduate level is inherently difficult due to its interdisciplinary nature. Students from Electrical, Mechanical, and Materials Engineering backgrounds often lack the specific CAD or mechanics knowledge required to design micro-scale devices effectively.
There is frequently a gap between "recipe-based" lab work—where students follow fixed instructions—and true engineering design. This curriculum was developed to allow students to rationally design a sensor, understand the physics of its stress distribution, and predict performance using computational tools before entering the cleanroom.
Methodology: A Three-Pillar Approach
The design and analysis workflow focuses on diaphragm-based piezoresistive pressure sensors through three main pillars:
- Analytical Modeling: Using classical mechanics to predict maximum radial and tangential stress in circular and square diaphragms.
- Computational Analysis (FEA): Utilizing FEA to visualize von Mises stress across the entire diaphragm surface, providing a more nuanced understanding than simple edge-case equations.
- Validation & Iteration: FEA results are validated against analytical models to ensure simulation accuracy, particularly regarding the relationship between stress and diaphragm thickness.
Key Insights and Design Optimization
- The "20% Rule": FEA revealed that minimum stress occurs at a distance from the edge equivalent to approximately 20% of the diaphragm's length. Placing a piezoresistor at this location would result in near-zero sensitivity.
- Sensitivity vs. Robustness: Circular diaphragms are more sensitive, experiencing higher maximum stress for the same applied pressure, but they have a lower fracture threshold compared to square geometries.
- Thickness Sensitivity: The relationship between stress and thickness follows an inverse square law ($1/t^2$). For thin diaphragms (~5 μm), minor manufacturing variations lead to massive changes in sensor output, emphasizing the need for strict process control.
Validation Results
The FEA results for normalized stress as a function of thickness showed an almost perfect correlation ($R^2 = 0.9998$) with the expected analytical exponent of -2, as shown in the power-law fit of the simulated data.
Educational Efficacy and Future Work
The pilot course demonstrated that providing a "reference design" allowed students from different majors to bridge the gap between mechanical stress analysis and electrical signal transduction. All students successfully produced valid mask designs.
Future work aims to close the fabrication gap by sending student-designed masks to external foundries, allowing students to test their own hardware. Additionally, there are plans to expand the curriculum to include a wider variety of MEMS devices, such as accelerometers and micro-actuators, to satisfy student interest in diverse applications.
Summary updated