Research on the production and characteristics of carbon materials with high CO2 adsorption performance: Based on machine learning and dung beetle optimizer methods

Abstract

Excessive CO₂ emissions have caused global climate change, leading to significant environmental impacts. Porous carbon materials are promising for CO₂ adsorption due to their high capacity and selectivity. Machine learning (ML) is a valuable tool in the search for ideal carbon materials for CO₂ adsorption. This study employs various machine learning (ML) methods to predict CO₂ adsorption capacity and examine the influence of carbon material characteristics. Results reveal that the pore structure, particularly micropore volume, significantly impacts CO₂ adsorption more than elemental composition. Subsequently, the impact of biomass feedstock properties and experimental conditions on the CO₂ adsorption capacity of the resulting carbon materials were explored. The content of hydrogen (H) in biomass feedstocks, the ratio of activating agents, and the activation temperature are the three most significant factors affecting the micropore volume of the carbon materials, which then influences the CO₂ adsorption performance. The optimal experimental conditions for the preparation of carbon materials with high CO₂ adsorption performance were determined by combining the Dung Beetle Optimizer(DBO) algorithm with machine learning methods. The integration of these approaches can effectively guide the synthesis of carbon materials for CO₂ adsorption.