Sputter deposition and characterization of thin film nickel titanium shape memory alloy for microelectromechanical systems application

Shape memory alloys (SMA) have the unique ability to recover large deformations when heated above a phase transformation temperature. The NiTi system gives the largest strain recovery and is the focus of this thesis. Large displacement and force make NiTi SMA an ideal candidate for MEMS actuators.; This dissertation provides a background review of the mechanism behind the Shape memory effect and examines the difficulties of sputter depositing NiTi SMA. Successful methods of sputtering NiTi SMA employed by other researchers are reviewed. A novel method for depositing NiTi thin film by DC sputtering is presented. This new method produces films with transformation temperatures very close to that of the target. The new process involves heating the target and does not require compositional modification of a near equiatomic NiTi target. Results from XRD, DSC, 4point probe, RBS, and TEM are presented. These results indicate that compositional modification can be produced by varying the target temperature. Films that were produced by gradual heating of the target produced a gradation of composition through the film thickness. This gradation produced films exhibiting the two-way SME. The simplicity of this new process should facilitate the incorporation of NiTi films into various microactuator devices.; The cause for the difference in composition between a heated target and a cooled target during the sputter deposition of NiTi is examined. Silicon substrates were placed radially at 18 degree intervals to capture the semicircular sputtering profile. Composition was measured using RBS. Plot of composition versus polar angle showed that Ti/Ni flux was greater at higher angles for the cooled target conditions. In the heated target the Ti/Ni flux was more uniform or constant from 0 to 90 degrees from the target normal.; The change in composition between a heated target and a cooled target was attributed to a change in composition of the top altered layer of the target. Diffusion calculation was done to determine if diffusion rates were large enough to account for this change in the top alter layer of the target. Diffusion was significant but no drift in bulk target composition was observed.