Physically-based reduced order modelling of a uni-axial polysilicon MEMS accelerometer.

Area of Science:

• Mechanical Engineering
• Materials Science
• Micro-electro-mechanical Systems (MEMS)

Background:

• MEMS accelerometers are crucial for impact detection.
• Accurate simulation of their mechanical response under shock is vital for reliability.
• Existing models may lack efficiency or precision for high-g events.

Purpose of the Study:

• To numerically investigate the mechanical response of a polysilicon MEMS accelerometer during drop impacts.
• To develop and validate a simplified, physically-based reduced-order model for dynamic analysis.
• To enhance the model's accuracy for predicting stress fields in critical regions.

Main Methods:

• A simplified two-degrees-of-freedom model of the MEMS accelerometer's movable parts was developed.
• The reduced-order model's accuracy was validated against 3D finite element simulations.
• Experimental data from instrumented samples were used for further model assessment.

Main Results:

• The reduced-order model accurately captures the system dynamics of the MEMS accelerometer.
• A correction factor was introduced to improve stress field predictions, particularly near re-entrant corners.
• The validated model offers an efficient yet accurate approach for analyzing MEMS behavior under shock.

Conclusions:

• Simplified models can effectively simulate MEMS accelerometer dynamics under impact.
• Accounting for local stress amplification is necessary for accurate failure prediction in MEMS devices.
• This work provides a validated, efficient tool for MEMS accelerometer design and reliability analysis.