Discovery of new formula accurately determining organic compounds aqueous diffusion coefficients by symbolic regression

Aqueous diffusion coefficients of organic compounds are important parameters for many studies in geosciences and relevant applications, in particular, for modelling solute transport in groundwater. Accurately determining organic compounds aqueous diffusion coefficients still remains challenging due to lack of high-quality experimental data as well as reliable predictive tools. Based on symbolic regression incorporating theoretical constraints derived from the Stokes-Einstein hydrodynamics and Wilke-Chang theory, our study discovered a symbolic formula, which is mathematically linking the diffusion coefficients with water temperature, molecular weight, molecular volume and viscosity of water. The proposed formula achieved a preferential accuracy for predicting different organic compounds diffusion coefficients in water comparing with the currently available empirical equations. The complex interplay of molecular weight, molecular volume and water temperatures on aqueous diffusion coefficients of organic compounds is substantiated by causal inference and controlled test. The insights from this study are shedding light on better understanding of organic compounds diffusion in water.