Internal Reduction of an Iron-Doped Magnesium Aluminosilicate Melt Rebecca L. A. Everman* and Reid F. Cooper* Department of Materials Science and Engineering, University of Wisconsin–Madison, Madison, Wisconsin 53706-1595 Optical and electron microscopies are used to analyze the mechanism and kinetics of internal reduction of an Fe 2 - doped magnesium aluminosilicate melt. Melt samples are heated to temperatures in the range of 1300°–1400°C under a flowing gas mixture of CO/CO 2 , which corresponds to a p O 2 range of 1  10 13 –4  10 13 atm. The melt experiences an internal reaction in which a dispersion of nanometer-scale iron-metal precipitates forms at an internal interface. The metal precipitates show no signs of coarsening within the samples; however, the crystals at the surface (which formed in the initial part of the reaction) do grow via vapor phase transport. The overall reaction is characterized by parabolic kinetics, which is indicative of chemical diffusion being the rate-limiting step. The diffusion of network-modifier divalent cations—particularly Mg 2 cations—is demonstrated to be the rate-limiting factor, and its diffusion coefficient is calculated to be 1  10 6 cm 2 /s within the temperature range of the experiments. I. Introduction T HE crystallization of fine-grained ceramics from a melt requires a spatially uniform distribution of nucleation sites throughout the precursor material. Optimal properties such as mechanical strength and electrical insulation are achieved when grain size is limited to 1 m or less. To achieve this goal, the concentration of crystal nuclei would need to be 10 12 –10 15 nuclei/cm 3 throughout the melt. 1 Unfortunately, in ionic melts, most crystallization or devitrification processes are dominated by surface nucleation, which occurs because heterogeneous nucleation has a significantly lower activation energy than does homogeneous nucleation. In the production of glass-ceramics, internal crystalline nucleation is promoted by providing sites for heterogeneous nucleation via a phase separation of the glass or via the initial homogeneous nucleation of a metal colloid within the glass. These internal heterogeneities compete successfully with the free surface as nucleation sites, which is due, in part, to the high viscosity of the glass. In both cases, however, current manufacturing techniques for glass-ceramics require the initial formation of a glass and, therefore, are not applicable to low-viscosity melts (e.g., intrinsi- cally inviscid alumina- or magnesia-rich compositions). This gap in current processing capabilities prompts the search for a tech- nique that provides for internal nucleation in an inviscid ionic melt. Here, we seek to develop such a method, using internal redox reactions that are driven by a change in the oxygen activity of the environment. The environmental change can affect the crystalline phase at the liquidus phase boundary, as well as the liquidus temperature. Thus, one can envision an “isothermal undercooling,” where a potential for crystalline nucleation is created, not by a change of temperature, but rather by a change of oxygen activity. Given the proper conditions, the redox reaction occurs at an internal “front” that sweeps through the material. An internal chemical/structural reaction of this nature results in the precipita- tion of a second phase throughout the reacted material. If this second phase is resistant to coarsening (Ostwald ripening), and if the reaction is rate-limited by bulk chemical diffusion, in theory, the reaction will precipitate an even distribution of those product phases as it works through the material. These primary phases could serve as heterogeneous crystalline nucleation sites for the majority (secondary) phases. II. Theoretical Framework (1) Internal Oxidation/Reduction Reactions In a multicomponent ionic system, internal redox reactions are possible because of kinetic restrictions. Dissipation of a redox potential gradient will occur by the most-rapid path—that is, the path that dissipates the energy gradient the fastest, which is not necessarily the path that leads to the lowest energy state of the system. The product of concentration and mobility (i.e., the transport coefficient) of the various species determines which path is the fastest. If the transport coefficients vary widely between species, e.g., for isothermal conditions, c O 2-D O 2- c M 2+D M 2+ (where c i is the concentration of species i and D i is its diffusion coefficient; M 2+ represents a divalent cation), then the reaction can occur internally, but only if the ion species can move independently. This independence is available in an ionic material that fulfills the “semiconductor condition,” † e.g., for a p-type material, c h •D h •c M 2+D M 2+ (where h • is a hole in the valence band); the (relatively) highly mobile electronic species decouple the motions of component ions by providing the local charge neutrality. 2 (See the work of Cook and Cooper 3 for a discussion in application to amorphous materials.) One additional require- ment for redox reactions to occur internally is a significant driving force for one cation to be reduced/oxidized over the others (e.g., |G° AO |  |G° BO |, for oxidation of an A-B metal alloy). In other words, one species must be more noble than the others. Following the philosophy given in the pioneering work of Wagner 4 on the oxidation of multicomponent metal solutions, Schmalzried 5,6 summarized the characteristics of, and require- ments for, internal oxidation and reduction reactions in crystalline oxide solid solutions. The typical reaction morphology has an internal reaction front, as well as a reaction that occurs at the surface with a region of chemical diffusion between the two reaction interfaces. This zone is a region where the gradient in oxygen activity is nonzero. Internal reactions in crystalline ionic solids proceed most rapidly via the diffusion of point defects. The J. Kieffer—contributing editor Manuscript No. 187385. Received October 23, 2001; approved November 26, 2002. Research supported, in part, by the National Aeronautics and Space Administra- tion, Office of Microgravity Science and Applications (under Grant No. NAG8-1460). Presented at the Glass and Optical Materials Division Meeting of the American Ceramic Society, Corning, NY, October 2, 2000 (Paper No. GP-003-00). *Member, American Ceramic Society. † Although, by bulk-conductivity considerations, the material would be classified as an insulator, what conduction does occur is wrought almost exclusively by the motion of electronic defects, i.e., by electrons or (electron) holes. J. Am. Ceram. Soc., 86 [3] 487–94 (2003) 487 journal