Mathematical model of the oxidation of a uranium carbide fuel pellet including an adherent product layer

Abstract

Highlights• Model of the oxidation of UC including an adherent U 3 O 8 product layer is presented.• Highly non-linear numerical solution to a double moving boundary problem.• Transient heat and mass transfer in expanding oxide and shrinking carbide phases• Temperature and oxygen concentration sensitivities tested and recommendations made. AbstractUranium carbide is a candidate fuel for Generation IV nuclear reactors. However, like any candidate fuel, a reprocessing route should be established before implementation. One proposed method involves a preoxidation step, where the carbide fuel is oxidised to an oxide and then reprocessed as normal. A mathematical model has been developed to simulate such an oxidation using finite difference approximations of the heat and mass transfer processes occurring. Available literature was consulted to provide coefficients for the reaction rates and importantly the diffusion of oxygen through the adherent oxide layer that forms on the carbide: the rate limiting step. The transient temperature, oxygen and carbon monoxide distributions through the system are modelled in order to predict oxidation completion times and the temperatures reached. It was found that for a spherical pellet of radius 0.935 cm, the oxidation can take between 1 h to 19 h depending on the oxidation conditions and reach temperatures of up to 1556 • C. A robust model results that offers increased understanding of a process crucial to the sustainable use of carbide fuels in energy generation.

Oxidation of UC: An in situ high temperature environmental scanning electron microscopy study

“…This oxide acts as a diffusion barrier for oxygen transport through the carbide matrix 28 . It is unclear whether the oxide product readily spalls off or stays adherent to the carbide core, as contradictory results have been reported 11,12,13 . Mukerjee et al 11 proposed that at low oxygen partial pressures (< 5kPa) the slow release of CO2 does not generate stresses in the oxide layer and therefore an intact oxide layer at the end of the reaction is formed.…”

“…It was found to be affected by several parameters such as temperature, oxygen partial pressure, morphology of the oxide layer formed, separation of the reaction products and stoichiometry of the initial material 10 . The oxidation mechanism of UC is complex as similar studies 5,11 performed over the same temperature range and under the same oxygen atmosphere, suggest different oxidation mechanisms which are affected by the formation of a protective oxide layer or a low density non-protective pulverised oxide layer 12,13 . Indeed U3O8 appears to be a detachable powder whilst the thin layer adherent at the surface of uranium carbide is usually characterised as UO2 14,15 .…”

Uranium carbide oxidation from 873 K to 1173 K

“…This work not only provided a possible solution to an industrial problem, it also provides data for improved modelling of the UC oxidation mechanism. As recently shown by the modelling efforts of Shepherd et al [21], depending on experimental conditions the oxide produced can be eitherprotective or non-protective [21]. Shepherd et al [21], [22] assessed the oxidation of UC pellets in air considering the presence of an adherent [21] or non-adherent oxide layer [22].…”

“…The presence of an adherent or nonadherent oxide layer was found to affect the reaction rate. The presence of an adherent oxide layer made of U3O8, for example, was found to slow down the completion of the reaction from 3.71 h to 3.92 h in a 0.935 cm radius spherical UC sample oxidised in air at 773 K [21]. Shepherd et al [21] recommended a temperature of 1023 K, in air, for an optimal conversion of UC to an oxide.…”

“…The presence of an adherent oxide layer made of U3O8, for example, was found to slow down the completion of the reaction from 3.71 h to 3.92 h in a 0.935 cm radius spherical UC sample oxidised in air at 773 K [21]. Shepherd et al [21] recommended a temperature of 1023 K, in air, for an optimal conversion of UC to an oxide. However, the model proposed [21], [22] did not rely on enough experimental data to justify the choice of considering an adherent or non-adherent oxide layer at the temperature considered.…”