文摘
Proton conduction in olivine is directly related to the diffusion rate of hydrogen by the Nernst–Einstein equation, but prior attempts to use this relationship have always invoked additional terms to try to reconcile laboratory measurements of proton conduction and hydrogen diffusion data. New diffusion experiments on olivine demonstrate that lattice diffusion associated with vacancies is indeed highly dependent on the defect site where hydrogen is bonded, but from none of the sites is diffusion fast enough to explain the observed laboratory proton conduction experiments. Hydrogen diffusion associated with polarons (redox-exchange) is significantly faster but still cannot explain the low activation energy typical of electrical conductivity measurements. A process of bulk diffusion, which combines lattice diffusion (either associated with redox-exchange or vacancies) with the far faster grain boundary diffusion, explains the laboratory results, but does not explain the field observations with an average grain size of 0.5–2 cm at 100 km below the Jagersfontein kimberlite field on the Kaapvaal craton. Either conduction is dominantly along well-interconnected grain boundaries of very fine-grained (0.01 mm) damp (80 wt ppm) olivine grains or fine-grained (0.05 mm), wet (400 wt ppm) pyroxene grains, or another conduction mechanism must be primarily responsible for the field observations. If diffusion is the correct explanation, the conductivity below the Gibeon kimberlite field in Namibia is too high to be explained by increased thermal state alone of a diffusion process, even for such fine-grained pyroxenes.