Development of PZT thin-film membrane actuators

Lead Zirconate Titanate Oxide (PZT) thin-film actuators have received large attention in Microelectromechanical systems (MEMS) because they have large frequency bandwidth, large energy density, and can generate large force. To develop a sub-millimeter PZT thin-film membrane actuator and understand its behaviors in complex loading conditions is of both academic and practical interests. In this thesis, I have illustrated the fabrication, testing, design and simulation of such an actuator.;I have developed procedures to fabricate a PZT thin-film membrane actuator. These procedures consist of fabrication of a sol-gel PZT thin film and a membrane structure. I have also developed procedures for post-processing which consist of dicing, wiring, poling, and packaging to prepare a PZT thin-film membrane actuator for testing with external fluid loading.;I found that the actuation of the PZT membrane actuator is strong enough and is not significantly affected by external fluid loading. Packaging, however, can have significant influence on the membrane actuation. A preliminary Finite Element Analysis (FEA) of the PZT membrane actuator was developed to optimize the actuator performance. The FEA shows that the relative sizes of electrode and silicon residue are two critical parameters. The actuator achieves maximal actuation when the size of the top electrode is about 50% ∼ 60% of that of the inner diameter of the silicon residue. An experiment has verified this trend suggested by the FEA. An analytical model has also been developed to predict the optimal size of the top electrode which contributes the maximal actuation. To improve the correlation of the simulation and the experiment quantitatively, I have improved the measurement accuracy with refined setup and parameter settings. I have also validated the geometric and material parameters used in the FEA.;I have tested the validity of the preliminary FEA model with SEM measurements of the geometry and parametric studies. The results showed the increment of both the silicon's and PZT's Young's modulus improve the correlation between the FEA and the experiment. Reduction of the PZT density also improves the correlation. Uncertainty of the SEM measurement can contribute to variation of correlation. However, these parameters do not contribute significant enough to explain the still-wide difference between the experimental and FEA results. I found that the addition of residual stress in the FEA model can improve the correlation. I also found the nonlinear effect is a possible effect to include in the FEA model, and a static analysis is a valid approach for the actuation simulation other than a harmonic analysis.