Diffusion and mobility of electrically conducting defects in olivine

Abstract

Electrical conductivity of lherzolite (65% olivine), measured as a function of time after changes in the oxygen fugacity (fo2) of the surrounding CO2/CO atmosphere, is used to infer the diffusivity of the point defects responsible for conduction in olivine. A total of 63 equilibration runs at temperatures of 900, 1000, 1100, and 1200 ∘C were fit using nonlinear parameter estimation to recover time constants (directly related to diffusivity) and conductivity steps. An observed fo2 dependence in the time constants associated with re-equilibration implies two defect species of fixed diffusivity but with fo2-dependent concentrations. Although the rate-limiting step may not necessarily be associated with a conducting defect, when time constants are converted to diffusivities, the magnitudes and activation energies agree extremely well with the model for magnesium vacancies (the slower species) and small polarons (holes localized on Fe3+) derived by Constable and Roberts (1997). This earlier study used an independent method of simultaneous modeling of thermopower and electrical conductivity as a function of fo2 and temperature, on data from a different type of sample (a dunite). We observe that at high fo2 where polarons dominate over magnesium vacancies in the defect population, re-equilibration is dominated by magnesium vacancy diffusion, and vice versa (at low fo2 magnesium vacancies dominate and re-equilibration proceeds at the faster rate associated with polaron mobility). We interpret this to suggest association between the cation vacancies and polarons, as has been suggested by Tsai and Dieckmann (1997), making the concentration of the minority defect the rate-limiting step in the oxidation/reduction reactions.