Computational modeling-guided design of deep eutectic solvents for tailoring lignin chemistry during lignocellulose pretreatment

Abstract

Lignocellulosic biorefineries offer a sustainable approach to decarbonization and biofuel production, but the full utilization of biomass components, particularly lignin, remains a challenge due to its complex structure. Deep eutectic solvents (DESs) have emerged as promising green solvents for lignin extraction and structure regulation, offering chemical tunability, recyclability, and environmental benefits. However, their potential to precisely tailor lignin linkages during biomass pretreatment has been underexplored. In this study, we integrated computational modeling with experimental validation to design DESs for lignin property regulation and efficient delignification. A total of 260 DES candidates, comprising 13 hydrogen bond acceptors (HBAs), 20 hydrogen bond donors (HBDs), and 4 lignin dimer and 4 lignin carbohydrate complex models, were screened to predict activity coefficients (γ), focusing on their effects on β-O-4 and β-5 linkages in using the Conductor-like Screening Model for Real Solvents (COSMO-RS). Nine representative DESs were synthesized and tested with hardwood pretreatment. The results showed that smaller γ values indicate stronger degradation of β-O-4 and β-5 linkages, with both the HBD and HBA playing a significant role in delignification. The β-O-4 linkage is a critical determinant of lignin's properties and applications in value-added biomaterials. Multivariate analysis reveals the overall impact of lignin structures on β-O-4 and β-5 by accounting for interactions between variables, highlighting the importance of a multivariate approach. Incorporating model compounds with etherified phenol structures and lignin–carbohydrate complexes provided a more comprehensive calculation representation of the delignification process. Experimental validation demonstrated that the 1,8-diazabicyclo[5.4.0]undec-7-ene: lactic acid DES extracted lignin with a high β-O-4 content (47%), suitable for producing carbon fibers with superior mechanical properties. In contrast, a choline chloride: lactic acid DES completely cleaved β-O-4 linkages (0%), yielding uniform lignin nanoparticles with an enhanced zeta potential. These DESs also achieved effective delignification, allowing carbohydrates to be used for biofuels. This research establishes a computational modeling-guided framework for designing DESs to achieve controllable lignin linkage profiles, optimizing both delignification efficiency and material properties. The findings provide a pathway for enhancing the economic and environmental sustainability of lignocellulosic biorefineries and expand the applications of lignin in diverse, high-value biomaterials.