Kinetic modeling of cellulose-derived biogas formation with improved accuracy via mechanistic elucidation

Understanding cellulose pyrolysis is a critical step in converting lignocellulosic biomass into biogas, offering a sustainable pathway for renewable energy generation and carbon mitigation. However, existing methods for predicting the yield and composition of cellulose-derived biogas suffer from a trade-off between computational efficiency and mechanistic accuracy. In this study, we present an integrated approach combining density functional theory (DFT), molecular dynamics (MD) simulations based on a deep learning potential field, and experimental validation to construct an optimized kinetic model. Pyrolysis of cellulose at 1173 K with varying residence times shows that CO consistently dominates the biogas composition, with its yield increasing over time, while CO₂, CH₄, and H₂ decrease due to secondary reactions. MD simulations reveal consistent trends, and fourteen reaction pathways were extracted from MD trajectories to determine activation energies and reaction mechanisms. Incorporating inter-product reactions significantly improved the predictive accuracy in kinetic modeling, reducing deviation by up to 82 % compared to existing models. This DFT-MD-kinetic hybrid framework not only enhances the reliability of biogas yield forecasting but also offers mechanistic insights that are essential for optimizing pyrolysis conditions toward higher gas quality and energy efficiency.