Liquefied dimethyl ether-alcohol systems for enhanced saponins recovery from Gac seed kernels: Systematic HSP theory and experimental evaluation

• GSKs is an important source of saponins-rich oil production for various applications. • HSP theory served as powerful tool for solvent design for GSK saponin extraction. • LDME-alcohol mixtures enabled one-step recovery of both oil-bound and free saponins. • Up to 87% saponins recovery was achieved by LDME-n-butanol (75/25%wt). This study presents a systematic approach using Hansen Solubility Parameters (HSP) theory to predict a suitable green solvent system to efficiently extract saponin-rich oil from seed kernels of Momordica cochinchinesis Spreng (gac). The novelty lies in modeling saponins distributed in both oil (oil-bound saponins) and non-oil matrix (free saponins) of the seed kernels with unique HSP profiles, enabling rational design of suitable extraction systems. Initial HSP analysis revealed that single solvents were insufficient for complete extraction – polar solvents (such as methanol, ethanol, propanol, and n-butanol) only extract free saponins while non-polar solvents like hexane and liquified dimethyl ether (LDME), a greener alternative for hexane, primarily recovered oil-bound saponins. HSP modeling further predicted that LDME-alcohol mixtures could simultaneously access both saponins forms by creating a solvent system whose HSP values lie within the overlapping region of the solubility spheres of both target saponins. Experimental validation confirmed these predictions, with a mixture of LDME-n-butanol (75/25%wt) achieving 87% total saponins recovery in 90 minutes at room temperature, significantly outperforming single solvents (LDME: 54%, n-butanol: 38%). This work demonstrates how HSP theory can guide the development of efficient single-step processes for extraction of complex natural products with dual-phase distribution.