Non-universal behavior of intermittent flow in microplasticity

The classical view of plasticity in metals assumed that any fluctuations in the underlying dislocation motion were essentially Gaussian and could therefore be accurately represented by large-scale averages in a homogenous continuum. However, more modern experiments have shown that plasticity of metals demonstrates stochastically intermittent and scale-free “dislocation avalanche” behavior during deformation. The scale-free nature of these deformation events means that they have no well-defined mean, and therefore are somehow fundamentally incompatible with a model based on averaged quantities. \nIn order to describe collective dislocation motion despite this difficulty, various theoretical models or simulation-based predictions for dislocation avalanche behavior (generally focused on the statistics of the avalanches) have been developed, but no consensus has been established on which model is correct. In addition, some models predict “universal” behavior that should not be affected by microstructural details. This is a very unexpected prediction from the viewpoint of materials science, so (as can be inferred from the title of this dissertation) we set out to experimentally investigate whether this behavior was truly universal, or instead had aspects that were affected by microstructural details.\nFirst, we verified that the experimental setup produces results that accurately represent the underlying behavior of the material, rather than being limited by the physical qualities of the transducer, particularly for our high-time-resolution dynamic measurements. We find that the used nanoindenter can indeed accurately trace very rapid dynamic slip in microcrystals due to its low mobile mass and capacity for high data acquisition rates.\nNext, we examine the avalanche size statistics across a wide range of microstructural and experimental variables, and find that while the observed size statistics follow the general predictions (and some previous experimental results) of a truncated power-law distribution, the power-law scaling exponent is not found to be “universal” as some theories predicted, instead being significantly affected by material (especially crystal structure), loading orientation, and drive rate, with the exponent values observed spanning a range that includes the predictions of both the pinning/depinning transition and jamming/unjamming models.\nSimilarly, the observed avalanche dynamics (avalanche velocities and their scaling with avalanche size, as well as averaged velocity-time profiles at fixed avalanche size or duration) are found to at least partially match the theoretical predictions, but still show microstructure-specific behavior and are thus non-universal. In some cases of said non-universal behavior, there are pre-existing extensions to the basic models that can account for these changes, which is encouraging in terms of the ability to produce theoretical models that represent realistic material behavior.\nFinally, we propose that the transition from intermittent flow at the microscale to apparently-smooth flow in bulk plasticity is due to the natural velocity scales of the dislocation avalanches and the fact that equivalent strain rates at larger sample sizes will produce higher absolute drive rates, eventually overtaking and eliminating any detectable avalanche behavior.\nThis dissertation focuses on single-crystal metals in order to cleanly examine the most fundamental effects on the avalanche statistics. Future work might expand this by investigating how the confined and randomized-orientation nature of a polycrystalline aggregate affects the overall deformation behavior relative to the single-crystal results, which could lead to improved models of deformation for commonly used metallic materials in engineering applications.