The dynamic response of ionic iron-bearing multicomponent silicate melts to reducing environments: Kinetics and structure

We examine the kinetic responses of two multicomponent aluminosilicate melts (Fe2O3-FeO-MgO-Al2O3-SiO 2 "FeMAS" and Fe2O3-FeO-CaO-MgO-Al 2O3-SiO2 "FeCMAS"), subjected to extreme reduction potentials with a low- pO2 environment maintained by a CO-CO2 buffer at high temperatures. Both reactions were characterized by a reaction front sweeping into the melt (internal reduction). Despite being exposed to the same experimental conditions, the two melts exhibited differing dynamic responses with FeMAS precipitating metallic bcc-Fe crystals both at the surface and internally while FeCMAS only formed a molten alloy of Fe-Si-C at the surface with a reaction rate ∼100x slower than FeMAS. Driving the reaction in Fe-CMAS harder, through use of a lower pO2 , resulted in bubble formation in the quenched specimens.; These experiments demonstrate the significance of minor compositional/structural changes in the melt on the kinetic response. For Fe-MAS, the reaction is rate-limited by chemical diffusion of Mg2+ into the melt with electron holes (h•) counter-diffusing to provide charge compensation. In FeCMAS, however, the CaO component in increases the chemical solubility of carbonate. The carbonate-inclusion reaction, by consuming h• , shuts down the reactions seen in FeMAS, and the incorporation of carbonate polyanions props open the melt network increasing the physical solubility of CO in the melt. This cascades the system down a kinetic path that favors the diffusion of molecular CO. Upon quenching, the system becomes closed to chemical diffusion, but when multiple heterovalent cations are present, local redox adjustments can occur, accounting for the production of CO bubbles as the carbonate back-reacts with Fe2+ and Si2+.; The differing reaction dynamics between melts provokes thought concerning the molecular structure of the melts as affected by the on-going reduction reaction. There is no reason to expect that the local molecular structure of the reacting melt is equivalent to the equilibrium structure of that melt for the set thermodynamic conditions. Thus, use of "bulk" analytical techniques is insufficient for characterizing this structure. High-resolution electron energy-loss spectroscopy was pursued to characterize molecular structure as a function of location in the specimens. The approach proved unsuccessful: the specimens oxidized (via emission of Auger electrons) and so changed structure under the electron beam.