Variable-energy positron studies of metallic glasses

The temperature dependence of the effective positron diffusion length $L_{+}$ has been studied in various metallic glasses with variable-energy positrons. All the initial measurements in the glasses show a very short diffusion length ($L_{+}\approx 10$ Å) that is alloy dependent. This is more than 2 orders of magnitude smaller than $L_{+}$ in annealed crystalline metals, and is ascribed to positron trapping at a high concentration of intrinsic defects. Larger values of $L_{+}$ are found in the metal-metalloid than in the metal-metal glasses. This is suggested to be caused by boron filling some of the open-volume areas, thereby decreasing the positron trapping rate. Our results are compatible with a density of a few percent of point defects or, alternatively, about ${10}^{13}$ ${\mathrm{cm}}^{-2\mathrm{}}$ of dislocation-type defects. Evidence was found on the existence of a nonhomogeneous defect profile in the near-surface region. Both reversible and irreversible changes in $L_{+}$ are observed during heating and cooling cycles, attributed to positron thermal detrapping and structural relaxation, respectively. Crystallization causes partial removal of the defect structure, but temperatures close to the melting point are required before positron trapping is significantly reduced. In ${\mathrm{Gd}}_{67}$${\mathrm{Co}}_{33}$ positron localization remains even after annealing 50 °C below the melting point. The trapping behavior in this alloy was found to depend on specimen heating and cooling rates indicating an ongoing phase-segregation process. The data for $L_{+}$ are compared with previously obtained bulk angular correlation data as well as Doppler-broadening line-shape parameters to provide a consistent picture of the defects in these systems.