This thesis investigates the transport mechanism of the Kroll Process, the commercially dominant method to produce high-purity titanium sponge. A bench-scale Kroll reactor was designed, built and used for a series of experiments. The results from these experiments, along with a series of numerical and physical models, facilitated the development of a novel description of the transport and reaction mechanisms controlling the Kroll process. A correlation that characterizes the interface exchange present in the Kroll process is presented. The correlation was developed from a computational fluid dynamics parametric study of dimensionless groups identified from an analysis of the Rayleigh-Taylor instability. The correlation was subsequently validated at the lab scale with analogous fluid systems. The validated correlation was then used to predict the maximum titanium tetrachloride feed rate for a given reactor size, which showed agreement with work presented in the literature. Four experiments with varying process conditions using a benchtop-scale Kroll reactor were performed. The results showed that titanium can grow on any substrate material irrespective of electrical conductivity, indicating that electron transport through crucible walls is not essential for titanium growth. Instead, it was found that the capillary action of magnesium (a reactant) through porous titanium sponge was the driving force behind the continuation of the reaction. Furthermore, the reaction rate changed throughout the titanium tetrachloride (a reactant) feed, suggesting a changing surface area where the heterogeneous reaction can occur. Following the series of experiments, a multi-phase, three-dimensional, thermal transient computational fluid dynamics model was developed and validated and was used to analyze the evaporation behaviour of titanium tetrachloride liquid and the resulting effect on the reactor fluid flow. The measured data, along with the predicted evaporation rate, were used to estimate the evolution of the reaction rate. A three-phase, isothermal, three-dimensional model of the settling behaviour of the reaction by-product magnesium chloride was also developed and showed to be qualitatively consistent with what was observed experimentally. The combination of the numerical models and detailed data analysis facilitated the proposal of a novel description of the transport mechanisms controlling the reaction within the Kroll process.
T. Ferguson (2026) studied this question.