Deep eutectic solvents for extracting nitrogen-containing heterocyclic neutral compounds from coal tar: Integrated physicochemical–thermodynamic screening and molecular insights

The large compositional space of deep eutectic solvents (DESs) complicates candidate selection for extracting nitrogen-containing heterocyclic neutral compounds (NHNCs) from coal-derived liquids. Existing computational screening for this separation has emphasized thermodynamic performance, with less attention to physicochemical operability. Here, we developed an integrated physicochemical–thermodynamic screening strategy to identify DESs for the simultaneous extraction of indole and pyrrole from coal-tar model oil. We first used extreme gradient boosting (XGBoost) models to assess the physicochemical operability of 492 DESs, narrowing the candidate pool to 78. We then used Conductor-like Screening Model for Real Solvents (COSMO-RS) to evaluate thermodynamic extraction performance and identified three tetraethylammonium chloride-based DESs containing acetamide, acetic acid, or formic acid for experimental validation. All three exhibited strong extraction performance, with the acetamide-based system (TEAC2AM) outperforming the other two for both solutes. Quantum chemical calculations identified cooperative N–H···Cl hydrogen bonds and C–H···π interactions that stabilize indole and pyrrole in the DES phase. Molecular dynamics simulations further revealed preferential solute enrichment in the DES phase and mobilities closer to those of the DES components than to those of the oil components. This screening–validation–mechanism workflow provides a rational basis for developing DES extractants for NHNCs in coal-derived liquids.