Atomistic simulation of defect structure evolution and mechanical properties at long time scales

Abstract

This thesis is a computational and theoretical investigation of the response of materials' mechanical properties to a wide range of environmental conditions, with a particular focus on the coupled effects of strain rate and temperature. The thesis provides original contributions to the fundamental understanding of how the materials mechanical properties change, as manifested by defect structure evolution, with temperature and strain rate conditions, as well as to the development of methodology used for enabling the investigation of dislocation-defect interactions over a much wider range of time scales than of reach to traditional techniques. This thesis advanced the capabilities of a recently developed activation-relaxation based atomistic method to enhance the accuracy of kinetic predictions, and to enable the investigation of dislocation-defect interactions dynamically at long time scales. We took the Autonomous Basin Climbing (ABC) method as a starting point, and incorporated the ability to sample multiple transition pathways associated with a given state. This new feature addresses the problem of overestimating the system evolution time due to the one-dimensional nature of the original ABC algorithm. The ABC method was further implemented in a dynamic framework, which makes it possible for the first time to directly simulate the dislocation-obstacle interactions at very low strain rates. This approach allows for a new way to connect the atomistic results to models at the meso-scale for simulating the plasticity of metals. We analytically derived how the applied strain rate couples with the thermal activation process, based on the framework of transition state theory informed by the atomistic approach described above. We demonstrated the coupling effect is a common mechanism behind many important phenomena, and provide three examples from the atomic level on the dislocation mobility and dislocation interactions with radiation induced defects. (i) A well-known universal flow stress upturn behavior in metals has been examined. We provide a simple physically based model to predict the flow stress at various strain rates, without invoking any assumed mechanisms or fitting parameters as in the traditional constitutional models. (ii) We implemented this new model in (i) to investigate the dislocation-obstacle interactions. The approach enabled us to map the interaction between an edge dislocation and a self interstitial atom (SIA) cluster in Zr in a two-parameter space consisting of temperature and strain rate. This approach allows the direct atomistic simulation of dislocation-obstacle interactions at experimental time scale, namely at low strain rates, which cannot be reached by traditional atomistic techniques. The dislocation is found to absorb the SIA cluster and climb at low strain rates and high temperatures, while it passes through the SIA cluster at high strain rates and low temperatures. The predicted mechanism map is able to reconcile the seeming controversy between previous experimental and computational findings. (iii) A dislocation-void interaction in bcc Fe at prescribed strain rate is also investigated. We demonstrated that different applied strain rates can affect the interaction mechanism and the defect microstructure, and eventually lead to a negative strain rate sensitivity (nSRS) of yield strength below a critical strain rate. This finding at the unit process level supplements the previous explanations of the nSRS with higher level constitutive relations. Beyond the specific cases analyzed in metals in this thesis, the insights gained on the coupling between strain rate and thermal activation can be used to explain the dependence on strain rate and temperature in other important classes of materials (e.g. colloids, cement) and phenomena (e.g. corrosion, creep).