Unraveling the coupling effect of micropore confinement and functional sites of carbon-based adsorbents on flue gas CO2 adsorption: A machine learning study based on multi-scale simulations

Abstract

Carbon material is a type of promising adsorbent for flue gas CO₂ capture, where micropore and dopants are key functional units and intertwined with each other. Due to the difficulty in detaching micropore and functional sites, their effects on CO₂ adsorption are still in debate. Herein, we unravel coupling effects of micropore confinement and functional sites combining machine learning (ML) and multi-scale simulations. High-throughput Grand Canonical Monte Carlo (GCMC) simulations in combination with density functional theory (DFT) calculations clarify that CO₂ adsorption mechanism under pore-dopant coupling is dependant on both micropore confinement environment and interaction type of CO₂ with functional sites. For basic dopants owning chemical interactions with CO₂, adsorption potential driven by Lewis acid-base interactions dominate CO₂ adsorption behavior and the optimal pore size is distributed at 7 Å. For dopants that predominantly adsorb CO₂ by physisorption interaction, steric effect becomes a key factor influencing CO₂ adsorption behavior, which will result in a shift in optimal pore size for CO₂ adsorption from 7 to 8-10 Å and alter adsorption selectivity. In this case, new descriptor free volume (V_f) was identified to describe coupling effects of micropore and functional sites. Guided by theoretical findings, we prepare carbon adsorbent with both heteroatom dopants and enlarged pore size, which exhibits leading-level CO₂ adsorption capacity of 4 mmol g⁻¹ at ambient condition, 130% higher than that without pore size optimization. This work demonstrates crucial role of micropore-dopant coupling mode on CO₂ adsorption, and provides new direction of developing high-performance carbon adsorbent beyond traditional standalone pore or doping engineering.