| Kinetic studies of the early stages of the oxidation of pure iron, initiated at a clean metal surface, have been carried out at a pressure of 76 mm. Hg absolute and isothermally in the temperature range 700-870°C by using a constant pressure differential gas apparatus. The duration of each isothermal study has been restricted to the first two minutes approximately and the thickness of the oxide formed is calculated to be of order of 150,000 A. During each isothermal study values of the instantaneous rate of reaction dW/[dt] , where W is the weight of oxygen consumed per unit area of the specimen and t is the time, have been obtained at intervals of a few seconds from the instant when it becomes practically possible to take the first reading. The relation between dW/[dt] and W is shown to be exponential. The plot of logarithm of reaction rate constant k against l/[T], where T is the absolute temperature, is continuous as a cusp which rises to a sharp peak at the Curie temperature (755°C). The activation energies for oxidation in the temperature-independent ranges above and below the Curie temperature are 33,100 and 21,700 calories/mole respectively. Hence it is deduced that the observed non-parabolic oxidation is controlled by the rate of electron transfer at the metal-oxide interface. For cobalt, using the same experimental technique, a logarithmic relationship between dW/dt and W has been confirmed up to oxide-film thickness of 220,000 A and the log[10]k/[1/T] plot gives a cusp peaking at 1121°C, The results for nickel, though not conclusive, show a cusp at the Curie point (353°C). A sharp discontinuity in the oxidation rate of cobalt observed at 1175°C is attributed to a suspected allotropic change in the metal. The values of the activation energy for oxidation of cobalt above and below the transformation temperature range (1055-1175°C) are 35,000 and 22,000 calories/mole respectively. Some interesting and striking correlations between the electronic structure of iron, cobalt and their initial oxidation behaviour have been pointed out and discussed in the light of certain features of Uhlig's theory which appear compatible with the present results. |
