Pore-scale modeling of coking during catalytic conversion of glucose to levulinic acid

Abstract

The growing demand for biomass-derived products has driven the wider adoption of packed-bed reactors in industrial processes. However, inevitable catalyst coking during operation impedes the efficiency and stability of continuous-flow systems, yet this issue has received limited attention. To address this industrial challenge, a 3D reactive-transport lattice Boltzmann model was developed, incorporating the coking effect to systematically simulate glucose catalytic conversion in fixed-bed reactors. The continuous influence of humins-derived coking on mass transfer was quantified, and the simulations successfully reproduced experimental trends. Based on these analyses, the study revealed the mechanisms by which key operating parameters and bed characteristics affect coking behavior in industrial production. Elevated initial glucose concentrations accelerated coking, while accumulated coke subsequently impeded both reaction and mass transfer, thereby self-limiting further coking. Packed beds of small particles, characterized by low porosity, exhibited pronounced inter-particle coking, whereas larger particles showed enhanced internal carbon deposition due to extended diffusion pathways. Targeted mitigation strategies, including feed concentration control, particle size optimization, and bed architecture design, were proposed. This work provides actionable insights into reactor design, process optimization, and coking control, thereby enabling efficient industrial-scale biomass catalytic conversion.