A mechanistic model, based on a quasi-equilibrium analysis of oxidation reactions, is proposed for predicting high-temperature corium oxidation. The analysis suggests that oxide forming on the surface of corium containing uranium, zirconium, and iron is similar to the oxides formed on zirconium and uranium as long as there is a small percentage of unoxidized zirconium or uranium in the metallic phase. This is because of the higher affinity of zirconium and uranium for oxygen. Hence, oxidation rates and heat production rates are similar to (U,Zr) compounds until nearly all the uranium and zirconium in the corium oxidizes. Oxidation rates after this point are predicted to be similar to those implied by the oxide thickness present when the forming oxide ceases to be protective, and heat generation rates should be similar to those implied by iron oxidation, i.e., ~4% of the zirconium oxidation heating rate.

The maximum atomic ratio of unoxidized iron to unoxidized liquid zirconium plus uranium for the formation of a solid protective oxide below 2800 K is estimated for a temperature, T (in Kelvin), as follows:

(unoxidized iron)/(unoxidized zirconium + turanium) = (1/28){5.7/exp[-(147 061 + 12.08T log(T) - 61.03T - 0.000555T2/1.986T)]}1/2.

As long as this limit is not exceeded, either zirconium or uranium metal oxidation rates and heating describe the corium oxidation rate. If this limit is exceeded, diffusion of steam to the corium surface will limit the oxidation rate, and linear time-dependent growth of a nonprotective, mostly FeO, layer will occur below the protective (Zr,U) O2 scale. When this happens, the oxidation should be at the constant rate given by the thickness of the protective layer. Heat generation should be similar to that of iron oxidation.