No abstract

No abstract

SCOPEThe proper use of liquid metals as solvents makes possible a wide range of new and advantageous chemical processes. For example the carbothermic reduction of reactive metal oxide ores in appropriate solvents could represent very substantial savings over the conventional, highly energy-intensive electrochemical production of aluminum, magnesium, and titantium. Traditionally the refining of pure reactor grade zirconium from ziconia by chemical separation of the hafnium impurities has been most difficult, but the Zn-Hf reduction and separation may be. accomplished directly and virtually quantitatively in a proper metal solvent. Solvent metal processes have even been shown to be successful in the separation of uranium from fission byproducts, thus leading to an environmentally acceptable and secure method for the reprocessing of spent nuclear fuel elements. All of these processes are currently at bench scale, and for implementation a mathematical model is required for scale-up. All depend on the unique thermodynamic properties of liquid metal solutions, and in this work a chemical theory of solutions is used to characterize the strong interactions of solutions forming intermetallic compounds, when activity coefficients may deviate from Raoult's law by many orders of magnitude. Such models should be instrumental in the investigation, development, and rational design of liquid metal solvent processes. CONCLUSION AND SIGNIFICANCEStrongly solvating (compound-forming) liquid metal mixtures can be well characterized thermodynamically by a chemical model of solution behavior. Using such a model, the liquid activities are well represented for mixtures forming either a single or multiple compounds. In addition the method is successfully extended to multicomponent systems and to the representation of the solid-liquid phase boundary. Since such systems always exhibit strong negative derivations from Raoult's Law, liquidliquid equilibria never occurs. The model presented is capable of accounting for both "physical" and "chemical" forces, and the general equations are developed. For the systems studied the compound formation is generally quite strong and the physical forces may be considered negligible by comparison. This assumption permits a direct linear least-squares solution of the equations to find the equilibrium constants for compound formation, which are the only parameters in the excess Gibbs energy equation. The chemical model is shown to be far superior for these systems to the traditional solutions to the Gibbs-Duham Equation; the Redlich-Kister, Van Laar, and Wilson equations are mathematically incapable of representing such systems.

No abstract

No abstract