Selective Extraction of Furfural and 5-Hydroxymethylfurfural from Mixed Lignocellulosic Biomass-Derived Feedstocks in Biphasic Solvent Systems

Simultaneous reaction and product separation (i.e. reactive separation) is essential for process intensification to achieve high selectivities of desired reaction intermediates. However, biphasic systems for liquid-phase extraction of such products often utilize common solvents selected via heuristic methods, resulting in suboptimal yields. In this work, we provide a systematic framework for solvent selection in biphasic systems for two major platform chemicals in biorefineries, furfural and 5-hydroxymethylfurfural (HMF), which are reactive intermediates in the dehydration of pentose and hexose sugars.

In order to identify premier extractants for furfural and HMF extraction, the multiscale COSMO-RS predictive model was first used to predict solute distribution (partition coefficients) between an aqueous phase and an organic solvent. Interestingly, the model predicts higher furfural partitioning to the organic phase compared to HMF (i.e. P_furfural > P_HMF) for all 2500+ solvents in the database. 670 solvents are predicted to form biphasic systems that favorably extract both HMF and furfural at 298 K (i.e. P_furfural and P_HMF> 1), while solvents selective for furfural but not HMF (i.e. P_HMF <1< P_furfural) number 775 at 298 K. This in silico screening was followed by experimental validation for selected solvents, and solvents with order-of-magnitude increase in P_HMF were identified. Using a high-throughput in situ phase sampling and subsequent sample handling process, solute partitioning at the relevant reaction temperature of 423 K was experimentally determined. These results were supplemented by solvent stability and solute compatibility experiments.