A spherical wave AVO response is investigated by measuring ultrasonic reflection amplitudes from a water/Plexiglas interface. The experimental results show substantial deviation from the plane-wave reflection coefficients at large angles. However there is an excellent agreement between experimental data and full-wave numerical simulations performed with the reflectivity algorithm. By comparing the spherical-wave AVO response, modeled with different frequencies, to the plane-wave response, I show that the differences between the two are of such magnitude that three-term AVO inversion based on AVA curvature can be erroneous. I then propose an alternative approach to use critical angle information extracted from AVA curves, and show that this leads to a significant improvement of the estimation of elastic parameters. Azimuthal variation of the AVO response of a vertically fractured model also shows good agreement with anisotropic reflectivity simulations, especially in terms of extracted critical angles which indicated that (1) reflection measurements are consistent with the transmission measurements; (2) the anisotropic numerical simulation algorithm is capable of simulating subtle azimuthal variations with excellent accuracy; (3) the methodology of picking critical angles on seismograms using the inflection point is robust, even in the presence of random and/or systematic noise.

We investigated dynamic response of Cu46Zr54 metallic glass under adiabatic planar shock wave loading (one-dimensional strain) with molecular dynamics simulations, including Hugoniot (shock) states, shock-induced plasticity, and spallation. The Hugoniot states are obtained up to 60 GPa along with the von Mises shear flow strengths, and the dynamic spall strengths, at different strain rates and temperatures. For the steady shock states, a clear elastic-plastic transition is identified. The local von Mises shear strain analysis is used to characterize local deformation, and the Voronoi tessellation analysis, the corresponding local structures at various stages of shock, release, tension and spallation. The plasticity in this glass, manifested as localized shear transformation zones, is of local structure rather than thermal origin, and void nucleation occurs preferentially at the highly shear-deformed regions. The Voronoi and shear strain analyses show that the atoms with different local structures are of different shear resistances that lead to shear localization. Additionally, we performed large-scale molecular dynamics simulations to investigate plasticity in Cu/Cu46Zr54 glass nanolaminates under uniaxial compression. Partial and full dislocations are observed in the Cu layers, and screw dislocations, near the amorphous−crystalline interfaces (ACIs). Shear bands are directly induced by the dislocations in the crystalline Cu layer through ACIs, and grow from the ACIs into the glass layers and absorb ambient shear transformation zones. Plasticity in the glass layers is realized via pronounced, stable shear banding. As the last part of the dissertation, we investigated with nonreactive molecular dynamics simulations, the dynamic response of phenolic resin and its carbon-nanotube (CNT) composites to shock wave compression. For phenolic resin, our simulations yielded shock states in agreement with experiments on similar polymers, except the "phase change" observed in experiments, indicating that such phase change is chemical in nature. The elastic–plastic transition is characterized by shear stress relaxation and atomic-level slip, and phenolic resin shows strong strain hardening. Shock loading of the CNT-resin composites was applied parallel or perpendicular to the CNT axis, and the composites demonstrated anisotropy in wave propagation, yield and CNT deformation. Our simulations suggested that the bulk shock response of the composites depends on the volume fraction, length ratio, impact cross-section, and geometry of the CNT components; the short CNTs in current simulations had insignificant effect on the bulk response of resin polymer.