Experimental study of inhomogeneous deformation in bulk metallic glass and their composite during wedge-like cylindrical indentation

Bulk metallic glasses (BMGs) are amorphous metals with impressive mechanical properties, such as high elastic strain up to 2%, high strength (up to 2% of Young\u27s modulus) and high hardness. Their weight normalized properties exceed the high strength to weight ratio of titanium alloys. Because of the lack of crystalline defects such as grain boundaries and dislocations, they have good corrosion resistance and good formability. The unique die molding properties of BMGs render them as excellent candidates for micro-scale machine parts, pressure sensor, golf clubs and casings. BMG\u27s also exhibit enhanced plastic creep resistance, since homogeneous plastic deformation is inhibited at room temperature. Below the glass transition temperature, BMGs exhibit inhomogeneous plastic flow through the formation of localized shear bands. Under unconfined loading geometry, BMGs fails in a brittle material manner with unstable propagation of a single shear band. However, under confined geometry, BMG\u27s show increased ductility due to the ability to nucleate and propagate multiple shear bands. This dissertation focuses on experimentally analyzing evolution and propagation of the shear bands in BMGs and their composites, by monitoring the deformation mechanisms at the scale of the shear band under confined geometry. Wedge-like cylindrical indentation has been used to provide a stable loading configuration for in-situ observation of the inhomogeneous deformation zone underneath the indenter. High resolution digital camera has been employed to capture surface images of the evolution of the process-zone. An in-house digital image correlation (DIC) program has been developed, utilizing MATLAB commercial software, to calculate the in-plane finite strain distribution at the scale of the shear band. First, the plastic deformation and flow field under the indenter are studied in both aluminum and copper alloys with different grain sizes to verify and validate the analysis protocol. The measured plastic zone size is comparable with the one predicted by the simplified cavity model and there is a unique correlation of the strain distribution along the radial line with different angular positions originating from the indentation center. The deformation zones developed under indenters with different radii are found to be self-similar. In the elastic domain, the measured strain distribution agree with FEM predictions; in the elastic-plastic domain, extra hardening is observed, which could be the result of constrained deformation. Second, the inhomogeneous deformation behavior of Vitreloy-1 bulk metallic glass is examined at room temperature. To overcome the resolution limit of the DIC technique to resolve the strain within a single shear band having 10-20nm width, an alternative method is implemented, addressing the strain jump within the band and the surrounding matrix. The results show that the BMG can deform homogenously to a large elastic strain level of about 4-6% before the onset of inhomogeneous deformation via localized shear bands. Such observation indicates the ability of BMG to withstand such high levels of stresses and strains if unstable shear band can be suppressed from the nucleation from the surface, such as the case of tension or bending. Following the perturbation analysis of Hwang et al (2004) and utilizing the same material parameters, it is found that homogenous nucleation strain is of the same order. The experimental measurements show more subtle details about the kinematics of shear band propagations. The shear band propagates intermittently at the expense of the surrounding matrix stored elastic strain energy. The surrounding matrix ceases to deform, during the activity of the shear band, however, no unloading is observed. The accumulated strain level inside of the shear band is about 3 orders higher than the one in the surrounding matrix. By tracking the strain increments of a single shear band and its surrounding matrix, the deformation filed has been shown to be self-similar, within the surrounding matrix. While the stress state at the observation point is defined by the global indentation filed, the local stress state within the shear band is a simple shear state, with respect to the band propagation direction. Relative to the band-propagation direction and the corresponding normal, the surrounding matrix deforms in a pure shear-state to accommodate shear band deformation. The experimental protocol is also utilized to study the kinematics of shear band initiation, propagation and arrest or hindrance by a secondary ductile phase. The deformation mechanisms in BMG composite with brass particles are examined. The composite is manufactured by warm extrusion of a mix of gas atomized powders of Ni-based BMG and brass. The resulting composite has an elongated particulate structure in the extrusion direction. The fracture toughness and toughening mechanism of the BMG composites are examined in the parallel and normal directions to the extrusion axis. This composite shows highly anisotropic properties along different loading directions. For the normal direction loading, brass reinforcements not only trigger the initial localized shear band, but also modify the crack propagation by crack bridging mechanisms. Also, microcracking is another important toughening mechanism. For the parallel direction loading, interface debonding is the main failure mechanism. Using FEM simulations, it is shown that local fracture is strain-controlled along the normal loading direction and stress controlled along the parallel loading direction. The proposed experimental framework is further extended for fracture match applications in forensic science. The likelihood of matching broken pieces, wherein a macroscopic crack trajectory cannot be established is analyzed via spectral analysis of the 3D fracture surfaces. The surface topographies are acquired using a non-contact 3D optical surface profilometer. A quantitative signature of the fracture surface, employing the different length scales of the fracture process zone is derived and used to establish class and sub-class matching. The details of the algorithm and its applications are detailed in the Appendix