| Thermal grooving at grain boundaries (GBs) is a capillary-driven evolution of surface topography in the region where the grain boundary emerges at a free surface. The study of these topographic changes can provide insight into surface energetics, and in our particular case, the measurement of surface diffusivity. We have measured the surface diffusion coefficient of 8mol% Y 2O3-ZrO2 by studying the formation of thermal grooves. We studied a total of five bicrystals, with well defined orientation relationships; random [110] -60°, random [001] -30°, Sigma13 [001]/{lcub}510{rcub}, Sigma13 [001]/{lcub}320{rcub}, Sigma5 [001]/{lcub}210{rcub}. Our calculations employed the Herring relation (1951), in which the variation in the chemical potential is related to changes in topography. The samples were annealed at 1300°C and 1400°C for various period of time. Atomic Force Microscopy was used to determine the exact geometry of the thermal grooves. A first approach consisted of estimating the diffusion coefficient by using Mullins' equation. yx=0= dsDs1/ 4gb2g s12G 5/4&parl0; WkTgs&parr0; 1/4t 1/4 Where y(x =0) is the groove depth at the GB triple junction, O is the atomic volume, gs is the surface tension, gb is the grain boundary surface energy, ds is the thickness of the diffusion layer, t is the annealing time, and Ds is the surface diffusion coefficient.; In Mullins' derivation, the atomic structure of the surface was ignored and it was assumed that the surface energy is independent of crystallographic orientation. In the case of zirconia, the surface energy is anisotropic. We will describe in this work a new approach to measuring surface diffusivity which accounts for the surface energy anisotropy. The study of these bicrystals will emphasize the effect of grain boundary structure on the surface diffusion coefficient, and it is for that purpose that we selected bicrystals with different tilt axes and angles. The results obtained using the equation set we have developed will be compared to those obtained by Mullins, and we show that the anisotropic groove evolution, even when perfectly symmetrical, is much slower than the corresponding isotropic case. |
