Date of Award
Spring 2023
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Mechanical Engineering & Materials Science (ENAS)
First Advisor
O'Hern, Corey
Abstract
Metallic glasses are alloys with amorphous atomic structures. They possess promising mechanical properties such as larger values for the strength and elastic limit compared to those for conventional crystalline alloys; however, they are typically brittle under tensile loading at room temperature. To use metallic glasses as structural materials, it is important to develop methods to predict their mechanical response as a function of the microstructure prior to loading. We developed a novel coarse-grained spring network model, which describes the mechanical response of metallic glasses using an equivalent series network of springs that can break and re-form to mimic atomic rearrangements during deformation. To validate the spring network model, we performed numerical simulations of quasistatic, uniaxial tensile deformation of Lennard-Jones and embedded atom method (EAM) potentials for Cu50Zr50 metallic glasses in the absence of large-scale shear band formation. We considered samples prepared using a wide range of cooling rates and with different amounts of crystalline order. By specifying five parameters in the spring network model that can be directly related to five key features of the shape of the stress versus strain curve, we can accurately describe the form of the stress-strain curves during uniaxial tension for the computational studies of Cu50Zr50, as well as experimental studies of Zr65Al10Ni10Cu15 metallic glasses deformed at high temperatures. For the computational studies of Cu50Zr50, we find that the elements in the spring network model can withstand longer elongations for slowly cooled glasses compared to rapidly cooled glasses. In addition, the average number of new springs and their rate of formation decreases with decreasing cooling rate. These effects offset each other at large strains, causing the stress-strain curve to become independent of the sample preparation protocol in this regime. Finally, we show the preliminary work on using the spring network model on more brittle samples prepared with hybrid molecular dynamics and Monte Carlo simulations. We found that the spring network model gives a good quality fit to the mechanical responses of these samples. However, the error of the fits increases as the effective cooling rate is lowered. Future studies would include extracting the parameters that define the spring network model directly from atomic rearrangements that occur during uniaxial deformation, applying the spring network model to brittle samples that develop shear bands, and improving the fits of the spring network model to stress versus strain curves obtained from brittle samples.
Recommended Citation
Nawano, Aya, "Computational Studies of the Mechanical Properties of Metallic Glasses" (2023). Yale Graduate School of Arts and Sciences Dissertations. 1065.
https://elischolar.library.yale.edu/gsas_dissertations/1065
