面向 CO₂捕集的吸附剂研究：从传统设计到机器学习

打开文本图片集

关键词： CO2 捕集；吸附剂；机器学习

中图分类号：X131.1文献标志码：A

DOI：10.7652/xjtuxb202511001 文章编号：0253-987X（2025）11-0001-18

Abstract： As global climate issues become increasingly severe， carbon capture technology has grown in importance for achieving the“dual carbon” goals. Among various carbon capture materials， adsorbents have emerged as a core material system for efficient CO2 capture due to their advantages such as excellent selectivity， high adsorption capacity，and long-term stability. Specifically， CO2 adsorbents for carbon capture applications primarily include carbon-based adsorbents，amine-based composite adsorbents， zeolite molecular sieves，mesoporous silica， metal-organic framework （MOFs）， covalent organic frameworks （COFs），and magnetic nanoparticles. First， this paper systematically reviews the latest research progress in these CO2 adsorbent materials，with a focus on comparative analysis of key performance indicators such as adsorption kinetics， adsorption mechanisms，and cycling stability. Second， it provides an in-depth analysis of the advancements in machine learning technology for CO2 adsorbent development， including the creation of new material screening systems based on active learning，the evaluation of performance metrics such as adsorption capacity using neural networks and high-throughput calculations，and mechanistic analysis linking adsorption parameters to pore structure through feature engineering. Finally， this paper highlights current bottlenecks in the field，such as the complexity of constructing standardized multi-indicator databases，challenges in cross-scale data integration，and delays in experimental validation. Future efforts should focus on the deep integration of materials genome engineering and machine learning，the development of next-generation algorithmic frameworks that balance interpretability and predictive power，and the establishment of closed-loop validation systems combining machine learning and experiments.

Keywords： carbon dioxide capture;adsorbents;machine learning

随着全球主要温室气体浓度的持续攀升，气候变暖、海平面上升与荒漠化等全球生态环境危机日趋严峻[1]。（剩余41879字）