It was recalled that the reproducibility of liquid diffusion coefficients, when measured in a gravitational field of 1g, was poor. This was because the apparent diffusivity was increased due to uncontrolled convection; unless capillary tubes which were less than about 1mm in diameter were used. It was common therefore to treat the measured diffusivity as being the sum of components which were labelled: intrinsic, wall-effect, thermal and buoyancy. It was noted that, if a diffusion couple were to be used isothermally under micro-gravity conditions, then the thermal and buoyancy components would be absent. The use of capillary tubes with various diameters would furnish an

estimate of the wall-effect component, and thus permit an accurate value to be obtained for the intrinsic diffusivity. If diffusion couples were used at various temperatures, the temperature dependence of the transport process could be deduced. Such experiments were carried out here on a number of alloys. The most complete data set was for Au in Pb. Gravity-driven convection was successfully distinguished by performing experiments under free-fall (space) conditions. Concentration distributions were analyzed by means of atomic absorption spectrophotometry. It was found that, when compared with terrestrial results, the free-fall (micro-gravity) experiments indicated lower diffusion coefficients (table 217); thus confirming that gravity-driven convection enhanced mass transport. The temperature dependence of the diffusion coefficient in molten dilute Pb-Au alloys was best fitted by using an Arrhenius relationship; thus indicating that liquid diffusion in this system was thermally activated. Most of the concentration profiles were smooth. However, any vibration during experiments could affect the concentration distribution.

X.Zhu, R.W.Smith: Materials Science Forum, 1996, 215-216, 113-8

Table 216

Diffusivity of Ag in Pb-Ag