Principally due to doubt concerning the activation energy of interstitial formation, Efi, it was difficult to differentiate between an interstitial and vacancy mechanism for self-diffusion in graphite. However, the recent studies of Turnbull and Stagg, which were verified by Thrower, yield a value of 8.3eV for the activation energy of diffusion occurring both in pyrolytic and naturally occurring graphites. This value was identified with Ef + 1/2(Ema + Emc). By considering the structural perfection of the graphites employed by the above authors, it was concluded here that there was no necessity to assume a discrete value describing Emc. Rather the observed activation energy, E*, was interpreted as the summation of 1/2(Efi + Efv) + Emai, which implied that self diffusion occurred via a dynamic interchange mechanism. It was implicit for the operation of this mechanism that the matrix should contain sites of high vacancy concentrations, such as intercrystallite boundaries. If the value of Efv was assumed to be 7.0eV, the values of 8.8 and 9.4eV were derived for Efi, when the experimental and theoretical values of Emai were substituted, respectively. The latter number was in fairly close agreement with both Coulson's theoretical value and that determined experimentally by Murty (9.6eV).

Discussion of the Diffusion Mechanism in Graphite with Particular Reference to Irradiated Material. Roscoe, C.: Carbon, 1969, 7[1], 119-27