THE MECHANISM OF OXIDATION OF A BASALTIC GLASS: CHEMICAL DIFFUSION OF NETWORK-MODIFYING CATIONS

Abstract

Rutherford backscattering spectroscopy, in conjunction with optical and scanning electron microscopy, has been used to characterize the oxidation process in a homogeneous, well-annealed glass prepared from a nepheline-normative olivine basalt. Initially melted and annealed at an oxygen fugacity substantially below the quartz-fayalite-magnetite (QFM) buffer, the glasses were oxidized in air under the time and temperature ranges 1-100 h and 550-600oC, respectively. Oxidation causes (1) formation of crystalline CaO and MgO that partially covers the free surface of the glass and (2) an internal reaction zone that is depleted of Ca2+ and Mg2+ but enriched in Na+. The reaction morphology is uniquely consistent with a model in which oxidation occurs by the outward diffusion (to the free surface) of Ca2+ and Mg2+ that is charge compensated by an inward flux of electron holes (polarons): oxidation of the glass occurs as the oxygen/cation ratio increases, not by addition of oxygen, but rather by removal of cations. The flux of Na+ from depth in the glass to the oxidizing region, which is also charge compensated by a counterflux of electron holes, is a response to the thermodynamic driving force seeking to stabilize Fe3+ as a network former, consistent with equilibrium thermodynamic and spectroscopic studies. Growth of the oxidized/transformed glass follows parabolic (chemical-diffusion-limited) kinetics. Using a first-order, Wagnerian approach, the diffusion coefficient and driving force terms of the parabolic reaction-rate constant are separated, giving an average divalent cation diffusion coefficient ofDA2+ (cm2.s-2) = 9.9 x 10-2 exp (-210 kJ.mol-1RT) .The oxidation mechanism seen for the glass, that is, one dominated by diffusion of network modifying cations and not an oxygen species, is anticipated to also occur in iron-bearing aluminosilicate melts: the discrepancy between the kinetics of redox reactions and of oxygen tracer diffusion noted in the literature for melts is most likely explained in this way.