Thermodynamic modeling of the LiCl-KCl-LaCl₃ system with Bayesian model selection and uncertainty quantification

Chloride molten salts are increasingly recognized for their applications in pyroprocessing techniques for the separation of lanthanides. Understanding the thermodynamic properties of these molten salts is essential to optimize the separation process. Several thermodynamic models, including the associate model, the two-sublattice ionic model, and the modified quasichemical model with quadruplet approximation (MQMQA), are utilized to capture the complexity of molten salts. In the present work, the Bayes factor was used to guide model selection process for the thermodynamic modeling of the KCl-LaCl₃ system and provide statistical comparisons of liquid models. The results indicate that the MQMQA model is the most favorable model based on the available thermochemical data. The LiCl-KCl-LaCl₃ system was further optimized with uncertainty quantification using MQMQA. The thermodynamic properties of compounds in the KCl-LaCl₃ system were obtained from DFT-based phonon calculations. The calculated phase stability shows excellent agreement with experimental data, indicating that an appropriate model is important for accurately predicting the behavior of complex molten salts.