Improved thermodynamic modeling of the Soret coefficient and validation with ISS microgravity experiments

This study aims to improve the theoretical prediction of the Soret coefficient by combining microgravity measurements from the Soret-Facet Mission on the International Space Station with newly acquired thermodynamic data on the ground. Interference fringes obtained in space were analyzed to determine the Soret coefficient of the salol/ t -butanol system. In addition to the space experiment, partial excess molar enthalpies required for the thermodynamic analysis were newly measured in this study using a high-sensitivity calorimeter. Using these experimental inputs, we reconstructed the Kempers-type thermodynamic framework. Two modified models were developed: the mass-based Kempers model (M-KEM), obtained by replacing the molar volume with molar mass, and the thermal-vibration-augmented model (M-KEM+TV), which incorporates an additional vibrational contribution to the Soret coefficient. The predictive performance of these models was evaluated using literature data for the water/ethanol benchmark system and experimental data for the salol/ t -butanol system. The M-KEM+TV model significantly reduced the deviation from experimental values—by approximately 55% compared with M-KEM—for the water/ethanol system and also improved accuracy for the salol/ t -butanol system near the reference temperature. The remaining discrepancies are mainly attributed to the neglected temperature dependence of the thermodynamic parameters. These results demonstrate that incorporating newly measured enthalpy data and thermal-vibration effects enhances the predictive capability of Kempers-type models for binary mixtures. • ISS experiments provided high-quality Soret data for salol/ t -butanol system. • Partial excess molar enthalpies were measured using a high-sensitivity calorimeter. • New thermodynamic data led to two modifications of the Kempers-type model. • Thermal-vibration-augmented model reduced deviation by 55% for water/ethanol system. • The model showed reasonable predictive capability for salol/ t -butanol system.