热解
动力学
氮气
化学
环境化学
制浆造纸工业
有机化学
生物化学
生化工程
工程类
量子力学
物理
作者
Qixing Hu,Yibo Zhang,Chengyi Luo,Y. Mi,Mingming Chen,Yijie Zheng,Zhiquan Hu
标识
DOI:10.1016/j.biortech.2025.132899
摘要
Pyrolysis offers a promising pathway for converting microalgae into high-value products. However, the governing mechanisms, particularly those involving amino acids, remain inadequately understood. Clarifying the behavior of amino acids during pyrolysis is essential for optimizing conversion efficiency and controlling pollution from N-containing products. This study investigates the pyrolysis of N-rich microalgae using three kinetic approaches, alongside product characterization. The model fitting quality is ranked as follows: the machine learning (ML) model (0.999), the model-free methods (0.981), and the independent parallel reaction model combined with particle swarm optimization algorithms (IPR-PSO) (0.910). With respect to actual components specificity, the ranking is reversed: IPR-PSO (actual components), model-free methods (pseudo-components), and ML model (component-independent). Compared with purely data-driven approaches, the IPR-PSO model provides mechanistic insights by associating reaction kinetics with specific amino acid contributions, identifying Leucine, Tyrosine, and Aspartate as key contributors with weights of 0.188, 0.149, and 0.081, respectively. Pyrolysis products were characterized by X-ray photoelectron spectroscopy (XPS), gas chromatography-mass spectrometry (GC-MS), Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR-MS), and a gas analyzer. Results indicate that as temperature increases, the protein-N in microalgae transforms into quaternary-N in biochar; pyridines, amides, unsaturated amide with carbon chain in bio-oil; and NH3 in syngas-these products primarily originate from amino acid decomposition and reforming. This investigation elucidates the significant role that amino acids play in optimizing microalgae pyrolysis as a sustainable resource recovery technology.
科研通智能强力驱动
Strongly Powered by AbleSci AI