On the history of solid-state diffusion

The development of models for solid-state diffusion in the period 1900-1950 is described, Several models have been proposed, i.e. the moving column by von Hevesy, the interstitial and vacancy mechanism by Frenkel, the ring mechanism in different versions by Rosenhain and Langmuir, diffusion along inner surfaces of tiny channels in the crystals by Smekal. Also direct place exchange was considered as a possibility. Arguments in favour and against the models were based on many experiments on diffusion as well as on electrolytical conductivity in salts, and on diffusion in metals and alloys. The first measurements concerned the diffusion of ''impurities'' in metals; self-diffusion could only be measured in lead and bismuth, which have natural radioactive isotopes. Impurity diffusion appeared to be faster than self-diffusion. It is described how the thermodynamic basis was laid by Wagner and Schottky, who derived equations for the number of defects in compounds, which have a composition which deviates from the stoichiometric ratio. The first calculations of formation energy and migration energy for different mechanisms were performed by Jest. The decisive experiments and calculations in favour of the vacancy mechanism, which took place in the period 1940-1950, are treated in detail.