The Zr-Pt system has proven to be ideal for illustrating how changes in solute levels modify liquid and amorphous structures, yielding contrasting phase selection during solidification or crystallization. For instance, we previously demonstrated that a sequential formation of metastable, structurally-related phases develop during quenching of eutectic Zr 80 Pt 20 liquids having oxygen contents ranging from < 200 to approximately 5000 ppm mass.

The accompanying slide explains how solute levels of oxygen modify both the as-quenched amorphous structure and crystallization pathway of hyper-eutectic Zr 75 Pt 25 melt spun ribbons. The combined DSC and TEM (1) illustrate that oxygen modifies the crystallization kinetics and dramatically alters phase selection during heating; metastable quasicrystalline and 'big cube' phases form, respectively, in glasses with lower and higher oxygen levels. A combined simulation and experimental strategy was followed to examine the structures of the glasses with different oxygen levels. From the partial pair correlations obtained using ab initio simulations of amorphous Zr 75 Pt 25 ( 2 ), we deduced that, on average, the bond length of Zr-Pt are shorter than Zr-Zr. Using high-energy synchrotron X-ray diffraction to access large reciprocal-space distances, we were able to obtain total pair correlation functions, g(r), having sufficient statistical accuracy to reveal how the glass structure changes with small additions of oxygen ( 3 ). Specifically, the strong bonding between Zr and O (Zr-O ~776 kJ mol -1 ) leads to a decrease in the frequency of Zr-Zr bonds (~298 kJ mol -1 ), which is illustrated as a decreases in the Zr-Zr contribution to the first peak of the total pair correlation function. Similarly, as Zr-Zr bonding decreases, the relative amount of stronger Zr-Pt bonds (~376 kJ mol -1 ) increases.

These results illustrate that even small levels of solute additions such as oxygen can alter the average structure and crystallization behavior of a metallic glass. Further work is in progress to perform large-scale simulations (e.g., reverse Monte Carlo 4 ) to yield 3-D models for detailed topological analysis of atomic configurations that develop as oxygen is added to Zr-Pt glasses. Ultimately, we hope to combine time-resolved experimental data with simulations to better understand what dictates phase selection of a specific Zr-Pt glass composition and solute content.