Practically achievable limits for pressure-vacuum swing adsorption (PVSA)-based post-combustion carbon capture are evaluated. The adsorption isotherms of CO2 and N2 are described by competitive Langmuir isotherms. Two low-energy process cycles are considered and a machine learning surrogate-model is trained with inputs from an experimentally-validated detailed PVSA model. Several case studies are considered to evaluate two critical performance indicators, namely, minimum energy and maximum productivity. For each case study, the genetic algorithm optimizer that is coupled to the machine learning surrogate model, searches tens of thousands of combinations of isotherms and process operating conditions. The framework pairs the optimum material properties with the optimum operating conditions, hence providing the limits of achievable performance. The results indicate that very low pressures ( <~0.2 bar) may be required to achieve process constraints for low feeds with low feed compositions ($<0.15$ mol fraction), indicating that PVSA may not be favourable. At higher CO2 feed compositions, PVSA can be attractive and can be operated at practically achievable vacuum levels. Further, the gap between the energy consumption of available adsorbents and the achievable limits with a hypothetical -best adsorbent varies between 20% to 2.5% as the CO2 feed composition changes between 0.05 to 0.4. This indicates a limited potential for development of new adsorbents of PVSA-based CO2 capture. Future work for PVSA should focus on flue gas streams with high CO2 compositions
Screening Supported Amine Sorbents in the Context of Post‐combustion Carbon Capture by Vacuum Swing Adsorption