Rationalizing Polysaccharide Extraction with Deep Eutectic Solvents: From Supramolecular Architecture to Emerging AI-Guided Solvent Design

Abstract

Deep eutectic solvents (DESs) have emerged as sustainable and tunable alternatives to conventional solvents for the extraction of polysaccharides. This review presents a structure-informed framework linking DES composition to polysaccharide solubility, emphasizing the differential responsiveness of amorphous, interfacial, and crystalline domains. Amorphous polysaccharides are efficiently extracted under mild DES conditions, while crystalline polymers often require stronger hydrogen bond acceptors or thermal/mechanical activation. Beyond dissolution, DESs modulate key properties of the extracted polysaccharides, including molecular weight, monomer composition, and bioactivity. Comparative analysis highlights how acidic, basic, or metal-coordinating DESs selectively target distinct polymer classes. Emerging innovations, such as in situ DES formation, mechanochemical systems, and switchable solvents, enhance efficiency and reduce downstream processing demands. Furthermore, the integration of machine learning and COSMO-RS modeling enables predictive solvent design, reducing reliance on empirical screening. By combining mechanistic insight, compositional tailoring, and computational tools, this review provides a scientifically grounded perspective for advancing DES-mediated extraction processes and enabling structure-preserving, application-oriented recovery of polysaccharides in food, pharmaceutical, and biorefinery domains.