A new model for H permeation was proposed in order to evaluate H diffusivities, while allowing for absorption and desorption. An analytical solution, which was derived from the model, predicted the effects of sample thickness, absorption and desorption upon permeation. A drift velocity through the surface and a drift velocity in the bulk were defined. The ratio of the drift velocity through the surface, to that in the bulk, determined the suitability of the time-lag method. The method could be applied for any ratio of the drift velocities, while the time-lag method became appropriate only when the ratio approached infinity. In order to verify the theoretical predictions, electrochemical permeation experiments were performed at room temperature on fully-annealed samples of commercial-purity Fe. The samples were of various thicknesses, and were coated with Pd on 2 sides. Two groups of samples were used. Group-I samples were plasma-cleaned and pre-sputtered. Group-II samples were untreated. The diffusivity which was estimated by using the present model was independent of the sample thickness and surface treatment. The time-lag method led to a diffusivity that was 2 orders of magnitude smaller. The experimental results showed that the ratio of the drift velocities for the group-I samples was some 100 times higher than that for group-II samples. This confirmed that the difference, between diffusivity values obtained using the present model and those found by using the time-lag method, decreased upon increasing the ratio of the drift velocities.

T.Y.Zhang, Y.P.Zheng: Acta Materialia, 1998, 46[14], 5023-33