On the toughening of bulk metallic glasses (BMGs), successful results in the phase-separated Zr63.8Ni16.2Cu15Al5 BMG have achieved compressive ductility over 15% through the computational-thermodynamic approach. In this study, the phase-separated Zr63.8Ni16.2Cu15Al5 BMG was compressed to nominal strains of 3%, 7%, and 10% at low strain rates (~10-4 s-1) and the results demonstrated that the BMG exhibited apparent uniform deformation initially, followed by visible local shear bands development. Afterwards, a single shear along the principal shear plane was soon developed and mainly dominated the whole deformation process. The principal shear contributed more than 2/3 of the overall plastic strain until failure. It was also found that the local shear strain varied along the principal shear plane and decreased monotonically from the shear band initiation site. Subsequently, in-situ compression experiments were conducted to monitor the change of sample shape during deformation in order to properly correlate with the stress-strain curve. The observed images showed that there was a one-to-one correspondence between the intermittent sample sliding and flow serration in the plastic region of stress-strain curve. Further investigations on flow serration were conducted on the Pd40Ni40P20 BMG through the compression experiments equipped with high-sensitivity strain gauges directly attached to two opposite sides of the test sample. There was an accompanied displacement burst when a shear band starts to propagate during deformation and this displacement burst would be accurately captured by the high-sensitivity strain gauges. Based on the displacement-time profile for one serration, shear-band propagating speed can be estimated and found to be insensitive to the applied strain rates (or the applied crosshead speeds). The disappearance of flow serration at high strain rates should be a result that the signal of displacement burst was overwhelmed by the applied strain rate. Using the shear strain rate data, the measured viscosity within a propagating shear band was found to be relatively low, which is in similar to the viscosity values reported in the supercooled liquid region during homogeneous deformation. In comparison with shear band propagation in the brittle Mg58Cu31Y6Nd5 and Au49Ag5.5Pd2.3Cu26.9Si16.3, moderately ductile Cu50Zr43Al7 and Pd40Ni40P20, and highly ductile phased-separated Zr63.8Ni16.2Cu15Al5 systems, the ductility of BMGs appears to be closely related to the dynamics during shear band propagation. The more ductile in nature the metallic glass is, the slower the shear band propagating speed would become. We also made attempts to investigate the shear band propagation in the porous Mo particles reinforced Mg58Cu28.5Gd11Ag2.5 bulk metallic glass composites (BMGCs) with up to 10% compressive failure strain. It was found that flow serration was absent in the stress-strain curve. Using high-sensitivity strain gauges, no distinct displacement burst was detected in the displacement-time profile. The diappearance of flow serration for the current porous Mo particles reinforced Mg58Cu28.5Gd11Ag2.5 BMGC is apparently associated with the lack of long-range shear band propagagtion. By employing the approach of separating the homogeneous amorphous matrix into many individual compartments, only short-range shear band propgagation is possible in the current Mg-based BMGC. An effective free spacing considering the spacing between two porous Mo particles and porous Mo particle size was applied to interpret the development of shear band propagation and is a useful indicator for the design of BMGC with high ductility.

In this work, the Shear Transformation Zone (STZ) dynamics model is adapted to capture the transitions between different regimes of flow serration in the deformation map of metallic glass. This was accomplished by scaling the STZ volume with a log-linear fit to the strain rate, and also adjusting the activation energy of an STZ with a log-linear fit to maintain constant yield strength at differing strain rates. Twelve simulations are run at each of six different strain rates ranging from 10-5 to 100 s-1, and statistics are collected on simulation behavior and shear band nucleation and propagation rates. The simulations show shear band nucleation has a positive correlation to strain rate, and shear band propagation has a negative correlation to strain rate. This shows that in STZ dynamics, the regime of reduced flow serration arises due to competing rates of nucleation and propagation, supporting the hypothesis proposed by Schuh. A positive correlation between critical shear band nucleus size and strain rate is proposed as an underlying cause of these rate dependencies.