Quantum Simulation of Carbon Capture in Fe-MOF-74 Reveals DFT Functional Variability and Enables More Accurate Predictions

Carbon capture represents a critical step towards decarbonising industries like steel and chemicals, and researchers are increasingly focused on materials that can efficiently trap carbon dioxide. Dario Rocca, Jerome F. Gonthier, and Joshua Levin, from QC Ware Corporation, alongside Tobias Schafer and Andreas Gruneis from the Institute for Theoretical Physics at TU Wien, and Byeol Kang from POSCO Holdings, present a new approach to simulating this process within metal-organic frameworks, materials known for their exceptional surface area and potential for CO2 capture. Their work addresses a key challenge in accurately modelling these materials, as conventional computational methods struggle with their complex electronic structure, often leading to unreliable predictions. By combining advanced quantum simulation techniques with a strategy to focus computational power on the critical adsorption sites, the team achieves more accurate and scalable modelling of carbon capture, paving the way for the design of improved materials and a more efficient path towards carbon reduction technologies.

Metal-organic frameworks (MOFs) possess high surface area and tunable structures, making them ideal candidates for capturing carbon dioxide. Accurately predicting their performance using conventional computational methods remains a significant challenge, particularly for complex, realistic systems. This work introduces a quantum simulation technique to overcome these limitations and gain deeper insights into carbon capture processes within MOFs, leveraging the principles of quantum mechanics to achieve greater accuracy and efficiency.

The method involves mapping the electronic structure of the MOF and carbon dioxide onto a quantum computer, enabling the simulation of complex interactions beyond the reach of classical computers. By focusing on periodic boundary conditions, the researchers accurately represent the infinite nature of the MOF material, crucial for realistic simulations. This research demonstrates the feasibility of using quantum computers to model carbon capture materials, achieving results that surpass the capabilities of conventional computational techniques. The team successfully simulated the adsorption of carbon dioxide within a representative MOF structure, calculating binding energies with high precision, paving the way for the rational design of novel MOFs with enhanced carbon capture capabilities.

Metal-organic frameworks (MOFs) represent promising materials for carbon dioxide capture. This study focuses on Fe-MOF-74, a magnetic material possessing exposed metal sites that enhance carbon dioxide adsorption. Its complex electronic structure presents challenges for standard density functional theory methods, which often yield inconsistent predictions. The team initially evaluated adsorption energies using various density functional theory approaches, revealing substantial variability and underscoring the need for more accurate methods, such as those provided by quantum computing.

Metal-Organic Frameworks Studied Via Quantum Methods

This research presents a new workflow for simulating carbon dioxide adsorption within metal-organic frameworks, specifically focusing on Fe-MOF-74. The team successfully combined traditional plane-wave density functional theory calculations with a strategy to reduce the computational space using localized orbitals derived from Wannier functions and MP2 natural orbitals. This reduction created compact orbital spaces suitable for implementation with emerging quantum algorithms, paving the way for more accurate simulations of complex materials. The researchers demonstrated that a quantum number-preserving ansatz, used within a variational quantum eigensolver framework, accurately reproduces full configuration interaction results in these reduced spaces.

Furthermore, they performed experiments using sample-based quantum diagonalization on current quantum hardware, achieving meaningful correlation energies for active spaces containing up to 28 qubits. These results demonstrate a scalable path towards applying quantum computing to realistic surface science and adsorption problems in periodic materials, offering a significant advancement in the field. The authors acknowledge limitations in the size of active spaces currently accessible and the need for improved error mitigation strategies as quantum hardware evolves. Future work will focus on expanding these active spaces to include more realistic spin multiplicities and achieve quantitatively accurate energy predictions, ultimately aiding in the design of improved materials for carbon capture and energy applications.