• Developed a kinetic framework for glucomannan and arabinoxylan pyrolysis. • Achieved improved prediction agreement with experimental data. • Proposed a feedstock-specific hemicellulose characterization method. • Enhanced biomass pyrolysis modeling for optimized conversion system design. The inherent structural diversity of biomass across different feedstocks presents significant challenges in developing accurate pyrolysis kinetic models. Hemicellulose, a key biomass component comprising 15–35% of dry weight, exhibits compositional variability and exists in distinct forms across different feedstocks. While existing lumped models treat hemicellulose as a homogeneous polymer or differentiate it by source, predicting its pyrolysis behavior remains challenging due to variations in decomposition characteristics among hemicellulose forms: particularly glucomannan (dominant in softwoods) and arabinoxylan (prevalent in grasses). This study develops a comprehensive kinetic framework that resolves three hemicellulose forms: glucuronoxylan, glucomannan, and arabinoxylan. Reaction mechanisms specifically developed for each hemicellulose subtype, based on high-precision TGA-MS/GC data from analytical-grade hemicelluloses, demonstrate good agreement with experimental measurements, including TGA/DTG curves, product distributions, and validation against literature pyrolysis experiments. A hemicellulose characterization method is proposed based on an extensive investigation of a large dataset of extracted hemicelluloses from the literature, indicating representative ratios of individual chain types for hardwood, softwood, and grass biomass. Integration of the new hemicellulose pyrolysis model within the CRECK-S-B biomass pyrolysis model enabled validation against a large database of biomass TGA curves and pyrolysis experiments. Statistical analysis of model deviations for individual hemicellulose models demonstrates improved alignment between model predictions and experimental TGA curves. The new hemicellulose model results in an increased curve matching index and a decreased sum of squared errors compared to the previous model. This work delivers a significant enhancement to the overall CRECK-S-B framework and establishes an important tool for optimizing biomass pyrolysis systems.
Suleiman et al. (Sat,) studied this question.