Computational Framework for Discovering Solvent Mixtures for Liquid–Liquid Separations

Abstract Solvent selection is a critical step in the design of liquid-phase chemical processes that impacts process sustainability and profitability. Solvents are also often incorporated as mixtures, substantially increasing the size of the solvent design space and motivating the development of computational methods to minimize experimental screening. In this work, we present a computational framework to address the challenge of selecting solvent mixtures for the liquid-phase separation of lignin-derived products, which is particularly challenging due to their similar physicochemical properties. We trained a machine-learning model to predict the partition coefficient, log10KP, of a solute as a thermodynamic descriptor of preferential partitioning between two immiscible liquid phases formed in a solvent mixture. By representing solvent mixture properties using Hansen solubility parameters and incorporating simple solute descriptors, the model can accurately predict log10KP values for a range of chemically similar lignin-derived solutes obtained from two distinct experimental datasets in the literature. We then combined log10KP predictions with the computational evaluation of phase behavior and individual solvent sustainability to screen solvent mixtures for two case studies: replacing mixtures containing halogenated solvents with more sustainable alternatives and identifying solvent mixtures suitable for separating lignin products via centrifugal partition chromatography. The proposed data-driven computational tools can (1) accelerate the discovery of solvent mixtures, including those with “green” solvents, and (2) promote sustainable lignin valorization by identifying solvent mixtures for separating lignin-derived products.