Mechanism-Informed Computational Entrainer Selection and Process Design for Biomolecule Extraction from Primary-Concentrated Microalgae in Subcritical Dimethyl Ether Systems

Efficient extraction of biomolecules from high-moisture microalgae is constrained by the polarity mismatch between solvents and water. This study introduces a COSMO-RS-assisted computational–experimental framework for entrainer screening and mechanistic elucidation in subcritical dimethyl ether (DME)–water systems. A database of 1982 solvents was sequentially filtered by melting/boiling points, environmental, health, and safety (EHS) criteria, and COSMO-RS-predicted solubilities, followed by ternary-miscibility evaluation and targeted experiments. The integrated results reveal that extraction efficiency is primarily governed by ternary miscibility, with class-specific entrainers optimizing recovery across lipids, proteins, carbohydrates, and pigments. Excess-enthalpy (H E ) decomposition identifies hydrogen bonding as the principal favorable contribution, opposed by electrostatic misfit. Electrostatic potential (ESP), independent gradient model based on Hirshfeld partition (IGMH), and interaction region indicator (IRI) analyses clarify how selected entrainers increase DME–water miscibility and enhance solvation of representative biomolecules. The framework offers predictive guidance for designing efficient, selective, and environmentally sustainable extraction processes for wet biomass valorization.