Effect of High Pressure on Self-Diffusion in Beta-Titanium

Abstract

Diffusion of ${\mathrm{Ti}}^{44}$ into $\ensuremath{\beta}$-titanium has been measured as a function of pressure at approximately 1000 \ifmmode^\circ\else\textdegree\fi{}C using standard lathe-sectioning techniques. The diffusion coefficient is found to decrease with pressure, indicating a positive activation volume. This result contradicts earlier reports of negative activation volumes for diffusion of Fe in $\ensuremath{\beta}\ensuremath{-}\mathrm{T}\mathrm{i}$ and U in $\ensuremath{\gamma}\ensuremath{-}\mathrm{U}$. The value obtained for the activation volume for self-diffusion in $\ensuremath{\beta}\ensuremath{-}\mathrm{T}\mathrm{i}$ is approximately 3.6 \ifmmode\pm\else\textpm\fi{} 1.0 ${\mathrm{cm}}^{3}$/mole, corresponding to $\frac{\ensuremath{\Delta}V}{{V}_{M}}=0.33\ifmmode\pm\else\textpm\fi{}0.1$, where ${V}_{M}$ is the molar volume. Comparison is made with atomic models and various proposed mechanisms for diffusion in the anomalous body-centered cubic (bcc) metals. The magnitude of $\ensuremath{\Delta}V$ appears to be too large to be consistent with the "extrinsic" vacancy mechanism of Kidson, but smaller than would normally be expected for simple vacancy diffusion in bcc metals. It is concluded that diffusion in $\ensuremath{\beta}$-titanium most probably proceeds via a combination of vacancy and short-dislocation-path diffusion, known as the Hart mechanism. Non-Gaussian penetration profiles were observed in some of the runs. This has been interpreted in terms of oxide holdup of the tracer at the surface and possible diffusion of the tracer as an oxide along dislocation pipes or grain boundaries near the surface prior to dissociation.