Design, fabrication, modeling and experimental testing of a piezoelectric flextensional microactuator

by 1971- Cheong, Jongpil

iii The necessity for reliable and low-cost microactuators capable of providing high resolution displacements and controlled force has been consistently being brought up because of their applications in RF and optical switches, microfluidic pumps and valves, micromanipulators for nanoscale handling and atomic force microscope drives. In this thesis, a piezoelectric flextensional microactuator is designed, fabricated, modeled, and tested for the amplification of the small piezoelectric strain to achieve large displacements. Bulk PZT material available in the form of 500 µm thick polished substrate is integrated with a precision micromachined silicon beam structure to fabricate the clamped-clamped flextensional microactuator. A high strength, high precision, and low temperature (~200 °C) In/Sn solder bonding process is developed and used to bond the PZT and the beam. This process can also be used for the heterogeneous integration of various materials for MEMS fabrication as well as for their packaging. The experimentally measured static deflection characteristics of the silicon micromachined beam show a flextensional gain factor of 20 with a large amplitude stroke of ~8 µm when actuated using -100 V to 100 V. The fabrication process produces devices with variable initial imperfection so a nonlinear model is developed to predict static and dynamic performance based on a polynomial curve fit of the initial shape. A static analysis of this model shows that for maximal actuator displacement, gain factor, and blocked force, the microactuator should have a thin beam structure and initial imperfection tuned to the maximum contraction provided by the PZT. For large PZT displacement, a perfect beam provides maximum gain factor but some imperfection is required to guarantee actuator iv movement in the desired (up or down) direction. For small PZT displacement, more initial imperfection improves performance. The theoretical model, based on the measured initial beam shape, predicts the experimentally measured direction and magnitude of beam displacement for three devices. The dynamic response of the microactuator results from buckling of the clamped beam in response to contraction of the bonded PZT support. Unlike previous research where sinusoidal initial beam shapes are analyzed, in this work the polynomial initial beam shape enables more accurate prediction of beam natural frequencies, frequency response, and time response when compared with experimental results. The inclusion of squeeze film damping between the beam and PZT support enables the model to predict response times. Experiments show that mounting the PZT with soft carbon tape limits PZT vibration.