Influence of material stiffness and geometrical variations on the electro-thermally driven microactuator performance

In order to make robust design of microdevices, it is important to implement a model considering uncertainty analysis. The high cost of experimentation and product development has led to an emphasis on simulated based design to achieve the success in the first-pass design and reliability. In this paper, finite element model is developed to investigate the effects of geometrical and material stiffness variations on an electro-thermally driven microactuator performance as an example. These microactuators are generally composed of two suspended beam (arm) joined at the free end. This device generates deflection through asymmetric heating of the hot and cold polysilicon arms with variable length or cross-section. These microactuators based on the force and deflection characteristic of elastic members are particularly sensitive to uncertainty in material properties. Polysilicon as a microdevice material shows a wide variation rang for Young’s modulus from batch to batch fabrication process. Also, the microfabrication process that are utilized for microdevice production can yield sometimes shapes that are not geometrically perfect, either due to the microfabrication process limitations itself or because of phenomena that take place during or after microfabrication. These geometrical errors can decrease the net cross section and will affect both modeling and experimental results. The effect of fillet radius at the juncture of this flexible microactuator and contact pads is investigated in this model. The residual stresses due to the fabrication process are taken into account too. The analysis is performed at multiple input voltages to estimate uncertainty bands around the deflection curve of the microactuator. Simulation results are compared with experimental results in literature. The results demonstrate how each of these factors affects the microactuator performance and justifies the deviation of previous nominal results from experimental results.