Bridging Morphological Transitions and Performance with Machine-Learning-Enhanced Property Analysis for Lignin-Containing Cellulose Nanofibrils

Abstract This study investigates the impacts of mechanical microfibrillation on the production evolution of Lignin-Containing Cellulose Nanofibrils (LCNFs) made from unbleached northern spruce kraft pulp at various stages of energy input, focusing on fibril properties, structural changes, and rheological properties with machine-learning-assisted analysis used to support structure–property interpretation. The energy input for microfibrillation applied up to 3433 kW hr/MT significantly affected the lignin content, crystallinity, surface charge, and particle morphology of LCNFs. The microfibrillation process enhanced colloidal stability by increasing zeta potential from −4.29 to −45.44 mV, enhanced crystalline ordering, and rendered reductions in lignin content. Chemical and structural analytical results confirmed that mechanical processing primarily induced physical rearrangement without major chemical modification with enhanced crystallinity. Morphological analysis revealed substantial fiber shortening and increased fines generation, which corresponded with improved suspension stability. Rheological measurements showed a systematic transition from particle-dominated behavior in the early stages to fiber-network-dominated viscoelasticity at advanced microfibrillation levels, accompanied by reduced thixotropic hysteresis, enhanced recovery, and improved yield strain. The machine-learning analysis supported the interpretation of the relationships between processing and morphological descriptors and the LCNF rheological properties. Collectively, these results highlight the critical role of controlled mechanical energy input in tuning LCNF performance and provide a comprehensive understanding of how microfibrillation governs structure–property relationships in lignocellulosic nanofibrillar systems.