On Balanced Tradeoffs between Stiffness and Design Area in CMOS-MEMS Accelerometer Springs

This work shows the application of an evolutionary algorithm to find a balanced tradeoff between the design area occupied by the layout of a metal folded beam spring and its stiffness, in CMOS-MEMS inertial sensor developments. CMOS-MEMS is the set of techniques aiming to integrate micro-electromechanical structures with a variety of purposes ranging from purely capacitive to thermal and radiation-sensitive applications within the metal and semiconductor layers found in most CMOS standard fabrication processes. This integration allows to create monolithic systems where transducers and signal processing stages coexist. The algorithm used to analyze the mentioned relations can be characterized as a genetic algorithm with three variables and two objectives functions, since these two objective (goals) are in conflict, the acquisition and discussion of results will be driven by the Pareto Optimality criteria. The solutions given by the execution of the algorithm may be taken as a start point to fine-tune the final design according to the designer preferences.