Optimally designed solvent system for lignocellulosic biomass conversion supported by property predictions

Abstract

The conversion of biomass with high sugar yields is enabled by a process using the solvent γ-valerolactone. There, the lactone dissolves the organic species, and a co-solvent is used to switch the solvent system's number of phases for efficient separation of the sugars in the aqueous phase. However, selecting the right co-solvent, a key economic driver for this process, currently involves several material-intensive and labor-intensive steps, from selecting candidates by experts to extensive experimental evaluation, and can lead to suboptimal choices. Here, we report a cost-optimal solvent-based biorefinery by combining process-based co-solvent screening and experimental validation of the best co-solvent candidate found. Assisted by property predictions, the solvent system we propose results from screening a broad range of molecules while reducing the manual effort compared to conventional solvent selection. The integration of reduced-order models embedded in process optimization allows identifying a cost-optimal co-solvent systematically. Additionally, environmental, health, and safety (EHS) evaluations assist in excluding hazardous co-solvents. The best candidate is validated experimentally inside the co-solvent hydrolysis reaction. Our findings show that through process optimization and the use of ethylbenzene as a co-solvent, we can enable 15% savings in operating costs and achieve a better EHS score than the reported benchmark toluene. Ethylbenzene shows lower performance than toluene in the analysis of phase partitioning and, therefore, would not be a leading co-solvent based on a laboratory-based evaluation alone. Here, we demonstrate that we can improve the final co-solvent choice, and a process-based co-solvent selection is needed.