Rate of hydrogen–iron redox exchange in silicate melts and glasses

Abstract

A kinetic model for the rate of iron–hydrogen redox exchange in silicate glasses and melts has been derived from time-series experiments performed on natural rhyolitic obsidians. Cylinders of the starting glasses were exposed to reducing mixtures composed of H 2-Ar-CO 2-CO in 1-atm furnaces and H 2-Ar in a cold seal pressure vessel. Overall, runs covered the temperature range 300 to 1000°C. The progression of a front of ferric iron reduction within the quenched melt was observed optically through a change of color. For all run conditions, the advancement of the front (ξ) was proportional to the square root of time, revealing the reaction as a diffusion-limited process. Iso- fO 2 runs performed in CO 2-CO, H 2-Ar, and H 2-CO 2 gases have shown that fH 2 rather than fO 2 is the dominant parameter controlling the reaction rate. The fH 2 dependence of the rate constant was characterized in the range 0.02 to 70 bar. The growth of the reduced layer, which is accompanied by an increase in reaction-derived OH-group content, was fitted considering that the reaction rate is controlled by the migration of a free mobile species (H 2) immobilized in the form of OH subsequent to reaction with ferric iron. The reaction rate is thus a function of both solubility and diffusivity of H 2 weighted by the concentration of its sink (ferric iron). We extracted a single law for both solubility and diffusivity of H 2 in amorphous silicates that applies over a range of temperatures below and above the glass transition temperature. Melt/glass structure (degree of polymerization) does not seem to significantly affect both solubility and diffusivity of H 2. We therefore provide a model that allows the prediction of oxidation–reduction rates in the presence of hydrogen for a wide range of compositions of amorphous glasses and melts. Comparisons with previous work elucidating rate of redox exchange in dry systems allow us to anticipate the fH 2-T domains where different redox mechanisms may apply. We conclude that equilibration of redox potential in nature should be dominated by H 2 transfer at a rate controlled by both H 2 solubility and diffusion. Numerical applications of the model illustrate redox exchanges in natural magmas and in glasses exposed to weathering under near surface conditions. We show that crustal events such as magmas mixing should not modify the iron redox state of magmas. In the case of nuclear-waste-bearing glasses, the fH 2 conditions in the host terrain are clearly a parameter that must be taken into account to predict glass durability.