In this two-part paper, the optimization of the electromechanical coupling coefficient for thin-film piezoelectric devices is investigated both analytically and experimentally. The electromechanical coupling coefficient is crucial to the performance of piezoelectric energy conversion devices. A membrane-type geometry is chosen for the study. In part I a one-dimensional model is developed for a membrane composed of two layers, a passive elastic material and a piezoelectric material. The lumped-parameter model is then used to explore the effect of design and process parameters, such as residual stress, substrate thickness, piezoelectric thickness and electrode coverage, on the electromechanical coupling coefficient. The model shows that the residual stress has the most substantial effect on the electromechanical coupling coefficient. For a given substrate material and thickness an optimum piezoelectric thickness can be found to achieve the maximum coupling coefficient. The substrate stiffness affects the magnitude of the maximum coupling coefficient that can be obtained. Electrode coverage was found to be important to electromechanical coupling. The model predicts an optimum electrode coverage of 42% of the membrane area. The model developed in part I formed the basis for the parameters studied experimentally in part II.

In this work a thermal interface material fabricated by thermocompression bonding of vertically aligned carbon nanotube turf (VACNT) to metallized substrates was characterized. The VACNT structure was fabricated onto silicon substrates using chemical vapor deposition. The structures were then transferred to metallized substrates using thermocompression bonding. The resulting structure consisted of VACNT turf sandwiched between two layers of Au. Two configurations of VACNT, full coverage and patterned, were fabricated and tested. In addition, the thermal interface resistance of structures at intermediate steps in the thermocompression bonding process were measured. For the full coverage turf a thermal interface resistance of 1.082 cm(2) degrees C W(-1) at an applied load of 1 N was measured, while a thermal interface resistance of 0.044 cm(2) degrees C W(-1) at a load of 1 N was measured for the patterned turf configuration.