(516e) Automated Reaction Network Generation and Kinetic Modeling for Fast Pyrolysis of Lignin with Model Compounds

Lignin is one of three major structural components of the lignocellulosic biomass and the second most abundant natural polymer. Due to its highly heterogenous molecular structure, lignin contributes most of the recalcitrance to biomass valorization. Fast pyrolysis is a promising thermochemical process for lignin valorization, allowing less char formation. However, the knowledge of the reaction mechanism for this process still needs to be completed, and many information gaps exist among the experimental and computational studies on this topic. To address it, we employ the concept of automated reaction network generation and kinetic modeling to study lignin pyrolysis using the model compounds. The computerized reaction network generation approach is a beneficial and realistic choice to do microkinetic modeling for such a complex and large reactive system. In this work, we constructed detailed reaction mechanisms for the fast pyrolysis process of anisole and guaiacol. These oxygenated aromatic compounds are the pyrolysate from lignin pyrolysis and serve as model lignin compounds. These compounds are also used as surrogates for lignin-derived primary tar. Therefore, a better understanding of gas-phase reactions among the species that evolved during the pyrolysis of these model compounds also helps better understand other thermal conversion processes, including gasification and combustion. These model compounds have been experimentally investigated, and specific product spectra have been reported in the literature. However, many missing links exist for the pyrolysis process for both model compounds. We employed the automated reaction network generation approach using these two model compounds and performed extensive reaction pathway analysis that would be insightful for bridging the existing information gaps. The resulting simulation from our work shows good agreement with the experimentally reported concentration profiles of pyrolysis products. We checked the effect of pyrolysis temperature on the generated reaction networks and analyzed the pathways for forming major products. We identified several areas of improvement so that this relatively less conventional approach to microkinetic modeling can be further used for models with similar complexity. Therefore, this work is the foundation for the successive model extension to automated reaction network generation and microkinetic modeling for lignin pyrolysis.