Evaluation and Identification of Promising CO2 Photocapture Materials

Abstract

Anthropogenic CO₂ emissions demand rapid mitigation, yet prevailing capture technologies remain energy-intensive due to thermally driven regeneration. In direct air capture (DAC), dilute ambient CO₂ concentrations make regeneration the dominant cost driver. This thesis presents a materials-to-device framework for photo-assisted CO₂ capture, integrating density functional theory (DFT), fixed-bed breakthrough experiments, and machine-learning (ML) surrogates to link molecular mechanisms to reactor performance. DFT calculations on low-index transition-metal facets under controlled surface charge densities, a representation for photoexcited carriers, reveal that adsorption strength is strongly charge and composition-dependent, producing a photoswitch-like response where charge state governs adsorption versus desorption. A photo-assisted breakthrough reactor was subsequently developed using PEI-impregnated silica, with indirect solar-selective illumination yielding more uniform heating and faster desorption than direct in-bed lighting. Open-cell Cu, Al, and Ni foams were evaluated as conductive model substrates, achieving ~0.3–1.1 mmol m⁻² per cycle, with Ni exhibiting the highest gravimetric uptake. Illumination and applied DC bias produced reproducible desorption enhancements under isothermal conditions, supporting surface-charge modulation beyond bulk thermal effects. ML models trained on the DFT dataset achieved strong predictive accuracy (R² > 0.8), enabling scalable screening of alloy–facet–site candidates. These findings motivate hybrid reactor designs combining conductive metal foams with chemisorptive overlayers for energy-efficient, electroresponsive CO₂ regeneration.