Stochastic optimization of a uranium oxide reaction mechanism using plasma flow reactor measurements.

In this work, a coupled Monte Carlo Genetic Algorithm (MCGA) approach is used to optimize a gas phase uranium oxide reaction mechanism based on plasma flow reactor (PFR) measurements. The PFR produces a steady Ar plasma containing U, O, H, and N species with high temperature regions (3000-5000 K) relevant to observing UO formation via optical emission spectroscopy. A global kinetic treatment is used to model the chemical evolution in the PFR and to produce synthetic emission signals for direct comparison with experiments. The parameter space of a uranium oxide reaction mechanism is then explored via Monte Carlo sampling using objective functions to quantify the model-experiment agreement. The Monte Carlo results are subsequently refined using a genetic algorithm to obtain an experimentally corroborated set of reaction pathways and rate coefficients. Out of 12 reaction channels targeted for optimization, four channels are found to be well constrained across all optimization runs while another three channels are constrained in select cases. The optimized channels highlight the importance of the OH radical in oxidizing uranium in the PFR. This study comprises a first step toward producing a comprehensive experimentally validated reaction mechanism for gas phase uranium molecular species formation.

Experimental Investigation of Uranium Volatility during Vapor Condensation.

The predictive models that describe the fate and transport of radioactive materials in the atmosphere following a nuclear incident (explosion or reactor accident) assume that uranium-bearing particulates would attain chemical equilibrium during vapor condensation. In this study, we show that kinetically driven processes in a system of rapidly decreasing temperature can result in substantial deviations from chemical equilibrium. This can cause uranium to condense out in oxidation states (e.g., UO3 vs UO2) that have different vapor pressures, significantly affecting uranium transport. To demonstrate this, we synthesized uranium oxide nanoparticles using a flow reactor under controlled conditions of temperature, pressure, and oxygen concentration. The atomized chemical reactants passing through an inductively coupled plasma cool from ∼5000 to 1000 K within milliseconds and form nanoparticles inside a flow reactor. The ex situ analysis of particulates by transmission electron microscopy revealed 2-10 nm crystallites of fcc-UO2 or α-UO3 depending on the amount of oxygen in the system. α-UO3 is the least thermodynamically preferred polymorph of UO3. The absence of stable uranium oxides with intermediate stoichiometries (e.g., U3O8) and sensitivity of the uranium oxidation states to local redox conditions highlight the importance of in situ measurements at high temperatures. Therefore, we developed a laser-based diagnostic to detect uranium oxide particles as they are formed inside the flow reactor. Our in situ measurements allowed us to quantify the changes in the number densities of the uranium oxide nanoparticles (e.g., UO3) as a function of oxygen gas concentration. Our results indicate that uranium can prefer to be in metastable crystal forms (i.e., α-UO3) that have higher vapor pressures than the refractory form (i.e., UO2) depending on the oxygen abundance in the surrounding environment. This demonstrates that the equilibrium processes may not dominate during rapid condensation processes, and thus kinetic models are required to fully describe uranium transport subsequent to nuclear incidents.

Effects of Plume Hydrodynamics and Oxidation on the Composition of a Condensing Laser-Induced Plasma.

High-temperature chemistry in laser ablation plumes leads to vapor-phase speciation, which can induce chemical fractionation during condensation. Using emission spectroscopy acquired after ablation of a SrZrO3 target, we have experimentally observed the formation of multiple molecular species (ZrO and SrO) as a function of time as the laser ablation plume evolves. Although the stable oxides SrO and ZrO2 are both refractory, we observed emission from the ZrO intermediate at earlier times than SrO. We deduced the time-scale of oxygen entrainment into the laser ablation plume using an 18O2 environment by observing the in-growth of Zr18O in the emission spectra relative to Zr16O, which was formed by reaction of Zr with 16O from the target itself. Using temporally resolved plume-imaging, we determined that ZrO formed more readily at early times, volumetrically in the plume, while SrO formed later in time, around the periphery. Using a simple temperature-dependent reaction model, we have illustrated that the formation sequence of these oxides subsequent to ablation is predictable to first order.