Screening and Performance Investigation of a Superior Extractant with Low Neutral Oil Entrainment for Separating Phenolic Compounds from Real Coal Tar

Separating phenolic compounds from low-temperature coal tar (LTCT) contributes to the rational resource utilization. Solvent extraction is efficient and environmentally friendly, but severe neutral oil entrainment in phenolic products remains a formidable challenge. This study established a machine-learning-based approach to systematically evaluate potential extractants, leading to the identification of formamide as an optimal candidate with low neutral oil entrainment. Thermodynamic analysis substantiated its superior anti-entrainment performance and operational flexibility. Microscopic mechanism analysis revealed that its effectiveness was attributed to the presence of multiple electronegative atoms and the absence of nonpolar groups. Finally, we applied formamide to a real LTCT system. Under optimized conditions, formamide achieved the complete extraction of phenolic compounds with only 0.67% neutral oil entrainment. Comparative assessment confirmed its pronounced advantage over other common extractants in suppressing the entrainment of the neutral oil. Moreover, a back-extraction protocol enabled the recovery of high-purity formamide, which supported its reuse.