Dislocation Interactions In Thin Films PDF Download

Nanotechnology has played a significant role in the development of useful engineering devices and in the synthesis of new classes of materials. For the reliable design of devices and for structural applications of materials with micro- or nano-sized features, nanotechnology has always called for an understanding of the mechanical properties of materials at small length scales. Thus, it becomes important to develop new experimental techniques to allow reliable mechanical testing at small scales. At the same time, the development of computational techniques is necessary to interpret the experimentally observed phenomena. Currently, microcompression testing of micropillars, which are fabricated by focused-ion beam (FIB) milling, is one of the most popular experimental methods for measuring the mechanical properties at the micrometer scale. Also, dislocation dynamics codes have been extensively developed to study the local evolution of dislocation structures. Therefore, we conducted both experimental and theoretical studies that shed new light on the factors that control the strength and plasticity of crystalline materials at the sub-micrometer scale. In the experimental work, we produced gold nanopillars by focused-ion beam milling, and conducted microcompression tests to obtain the stress-strain curves. Firstly, the size effects on the strength of gold nanopillars were studied, and "Smaller is Stronger" was observed. Secondly, we tried to change the dislocation densities to control the strength of gold nanopillars by prestraining and annealing. The results showed that prestraining dramatically reduces the flow strength of nanopillars while annealing restores the strength to the pristine levels. Transmission electron microscopy (TEM) revealed that the high dislocation density (~1015 m-2) of prestrained nanopillars significantly decreased after heavy plastic deformation. In order to interpret this TEM observation, potential dislocation source structures were geometrically analyzed. We found that the insertion of jogged dislocations before relaxation or enabling cross-slip during plastic flow are prerequisites for the formation of potentially strong natural pinning points and single arm dislocation sources. At the sub-micron scale, these conditions are most likely absent, and we argue that mobile dislocation starvation would occur naturally in the course of plastic flow. Two more outstanding issues have also been studied in this dissertation. The first involves the effects of FIB milling on the mechanical properties. Since micropillars are made by FIB milling, the damage layer at the free surface is always formed and would be expected to affect the mechanical properties at a sub-micron scale. Thus, pristine gold microparticles were produced by a solid-state dewetting technique, and the effects of FIB milling on both pristine and prestrained microparticles were examined via microcompression testing. These experiments revealed that FIB milling significantly reduces the strength of pristine microparticles, but does not alter that of prestrained microparticles. Thus, we confirmed that if there are pre-existing mobile-dislocations present in the crystal, FIB milling does not affect the mechanical properties. The second issue is the scaling law commonly used to describe the strength of micropillars as a function of sample size. For the scaling law, the power-law approximation has been widely used without understanding fundamental physics in it. Thus, we tried to analyze the power-law approximation in a quantitative manner with the well-known single arm source model. Material parameters, such as the friction stress, the anisotropic shear modulus, the magnitude of Burgers vector and the dislocation density, were explored to understand their effects on the scaling behavior. Considering these effects allows one to rationalize the observed material-dependent power-law exponents quantitatively. In another part of the dissertation, a computational study of dislocation dynamics in a free-standing thin film is described. We improved the ParaDiS (Parallel Dislocation Simulator) code, which was originally developed at the Lawrence Livermore National Laboratory, to deal with the free surface of a free-standing thin film. The spectral method was implemented to calculate the image stress field in a thin film. The faster convergence in the image stress calculation were obtained by employing Yoffe's image stress, which removes the singularity of the traction at the intersecting point between a threading dislocation and free surface. Using this newly developed code, we studied the stability of dislocation junctions and jogs, which are the potential dislocation sources, in a free standing thin film of a face-centered-cubic metal and discussed the creation of a dislocation source in a thin film. In summary, we have performed both microcompression tests and dislocation dynamics simulations to understand the dislocation mechanisms at the sub-micron scale and the related mechanical properties of metals. We believe that these experimental and computational studies have contributed to the enhancement of our fundamental knowledge of the plasticity of metals at the sub-micron scale.