Diffusion and thermal defects in amorphous metals

Although diffusion in metallic glasses has been investigated for more than two decades, the basic diffusion mechanism is still under debate. Vital information has been obtained from investigations of the pressure and mass dependence of diffusion in fully relaxed metallic glasses. The vanishing pressure dependence of Co diffusion in metal-metalloid glasses and in Co-rich Co-Zr glasses together with the absence of a significant isotope effect are interpreted in terms of a direct, highly cooperative diffusion mechanism involving some ten atoms. On the other hand, the significant pressure dependence of the Zr diffusion in the same Co-Zr glasses is indicative of a defect mechanism. The operation of two different mechanisms for the two constituents provides evidence of the existence of an opposite Kirkendall effect as proposed in literature. The concentration dependence of Co diffusion in Co-Zr glasses at constant temperature shows a strong variation over more than three orders of magnitude with a pronounced minimum, whereas the isotope effect does not change. This concentration dependence reflects structural changes, which are consistent with recent MD simulations. Judging from the significant pressure dependence in combination with a vanishing isotope effect for Co-diffusion a diffusion mechanism via smeared out defects seems to apply for Co-diffusion in approximately equiatomic amorphous Co-Zr alloys. It is the aim of the present paper to review the available data and to give a consistent picture of diffusion in metallic glasses and some predictive criteria.