(383az) Enhanced Furfural Extraction from Aqueous Media Using Neoteric Hydrophobic Solvents for Sustainable Biomass Recovery

The recovery of furfural, a valuable platform chemical, from hemicellulosic biowastes is crucial for reducing dependence on fossil fuels and promoting sustainable biomass recovery. However, current methods for furfural extraction are often inefficient and environmentally questionable. This study introduces newly developed neoteric hydrophobic solvents, specifically deep eutectic solvents (DESs) and ionic liquids (ILs), as superior alternatives to conventional solvents for the extraction of furfural from aqueous media.

A systematic top-down approach was employed for solvent development, starting with the experimental screening of 32 DESs and ILs to evaluate their furfural extraction efficiency compared to the conventional solvent toluene. The selection criteria for the best solvents included not only extraction efficiency but also favorable physical properties, such as density differences, low viscosity, and minimal toxicity. Additionally, for DESs, natural origin, availability, and biodegradability were considered. Thymol:decanoic acid (Thy:DecA) 1:1 DES and trihexyltetradecyl phosphonium bis(trifluoro methylsulfonyl) imide [P14,6,6,6][NTf2] IL emerged as the top performers, achieving extraction efficiencies of 94.1% and 97.1%, respectively, surpassing toluene's 81.2% efficiency.

The selected DES and IL were then extensively characterized, focusing on their physical properties, thermal properties, critical properties, and cross-contamination solubility. The density values were 0.930 ±0.001 and 1.066 ±0.001 g/mL for the hydrophobic DES and the IL, respectively, at 25°C, with both solvents exhibiting a density difference of ≥0.050 g/mL from water, which is a recommended criterion for solvent suitability in aqueous extraction. The viscosity of Thy:DecA and [P14,6,6,6][NTf2] at 25°C was measured to be 11.9 ±0.1 mPa·s and 341.4 ±2.6 mPa·s, respectively, indicating their manageability in extraction processes. Remarkably, both solvents exhibited exceptional stability across a wide range of operating conditions, maintaining high efficiency at various temperatures (15-100°C), pH values (1-13), initial furfural concentrations (0.10-2.00 wt%), and sugar concentrations (0.50-2.00 wt%), with an equilibration time of only 2 minutes. Furthermore, the solvents demonstrated excellent durability through multiple extraction and regeneration cycles, with minimal cross-contamination of the aqueous media (~0.1%).

Machine learning techniques, specifically multiple non-linear regression (MNLR) and artificial neural networks (ANNs), were employed to model the extraction performance of the solvents. The MNLR model considered linear, squared, and interactive terms to capture the non-linear characteristics of the extraction process. The developed MNLR models exhibited robust predictive power, as evidenced by the high Fisher's F-ratio of 234.4 for Thy:DecA and 47.2 for [P14,6,6,6][NTf2], with absolute average relative deviation (AARD) percentages of 2.74% and 2.28%, respectively. On the other hand, the ANN model architecture was optimized through a grid search with leave-one-out cross-validation, resulting in a 2-3-2-2 structure. The ANN model demonstrated exceptional performance, with low cross-validation root mean square error (RMSE) values of 0.116 for Thy:DecA and 0.427 for [P14,6,6,6][NTf2], and AARD values of 0.10% and 0.41%, respectively, significantly outperforming the MNLR model.

To gain deeper insights into the solvation mechanisms of neoteric solvents, computational quantum chemistry modeling using density functional theory (DFT) and conductor-like screening model for real solvents (COSMO-RS) was conducted. DFT calculations revealed the importance of interaction energies in determining extraction efficiency, with higher interaction energies correlating with better performance. The interaction energies for the Furfural:Toluene, Furfural:Thy:DecA, and Furfural:[P14,6,6,6][NTf2] complexes were -0.380 eV, -1.116 eV, and -1.360 eV, respectively, mirroring the trend in the experimental extraction efficiencies. COSMO-RS analysis provided insights into the molecular interactions driving the extraction process, highlighting the role of hydrogen bonding (20%) and van der Waals forces (67%) in the DES system and the dominance of van der Waals interactions (77%) in the IL system. The COSMO-RS predicted extraction efficiencies closely aligned with the experimental data, with an overall AARD of 4.36%, validating its effectiveness as a tool for molecular insights and preliminary screening.

The exceptional effectiveness and stability of the newly developed neoteric solvents represent a significant advancement in sustainable furfural recovery from biowaste. These solvents offer an environmentally friendly and efficient alternative to conventional solvents, paving the way for more sustainable biomass recovery processes. The comprehensive experimental and computational approach employed in this study not only validates the superior performance of the selected DES and IL but also provides valuable insights into the underlying mechanisms governing their extraction efficiency. These findings contribute to the growing body of knowledge on neoteric solvents and their potential applications in green chemistry and sustainable technologies.

In conclusion, this study introduces highly effective, sustainable, and eco-friendly neoteric hydrophobic solvents for the extraction of furfural from aqueous media. The exceptional performance and stability of Thy:DecA (1:1) DES and [P14,6,6,6][NTf2] IL across various operational conditions and through multiple extraction and regeneration cycles highlight their potential as superior alternatives to conventional solvents. The insights gained from machine learning modeling, with the ANN model demonstrating remarkable accuracy (AARD ≤ 0.41%), and computational quantum chemistry further enhance our understanding of the solvation mechanisms and molecular interactions driving the extraction process. These findings contribute to the development of more sustainable and efficient biomass recovery processes, advancing the transition towards a greener and more environmentally friendly future.