High-throughput screening of 2,500 ionic liquids for sustainable furfural recovery: Bridging quantum simulations, machine learning, and experimental validation

The recovery of furfural is crucial in reducing reliance on fossil fuels. However, current recovery techniques are inefficient and often use volatile organic compounds (VOCs). In this study, we present a novel approach to identify effective ionic liquids (ILs) as alternative solvents for furfural recovery. For the first time, a comprehensive evaluation of 2,500 ILs was conducted using the COSMO-RS approach, saving time and resources. The results showed that the most effective cations were trihexyltetradecyl phosphonium, tetrabutyl phosphonium, and methyltrioctyl ammonium. The leading anions included bis(pentafluoroethylsulfonyl) imide, bis(trifluoromethylsulfonyl) imide [ntf2], and hexafluorophosphate. This suggests that ILs with hydrophobic and fluorinated anions, combined with large non-polar cations, have the most potential for furfural recovery. Notably, machine learning methods including decision trees and multiple linear regression were innovatively applied to analyze the σ-profiles of the selected 2,500 ILs and assess their impact on extraction. The analysis revealed that cations primarily influenced the distribution ratio, while anions played a significant role in determining the selectivity. To validate the findings, the theoretical predictions were compared with the experimental efficiencies of 13 ntf2-based ILs with different cations. Their performance surpassed that of the benchmark solvent, toluene (81.2%), achieving efficiencies in the range of 91.3% to 94.8%. The comparison also showed a strong correlation between the experimental and theoretical results, with an R² value of 0.877 and a deviation of 0.24%. This validates the computational approach. These novel findings demonstrate that these ILs outperform toluene in furfural recovery, offering improved efficiency with no drawbacks. They also highlight the transformative potential of computational methods in streamlining the development of sustainable chemical processes.