Carbon capture sorbents and DAC
What and why
The primary job of CO2 capture sorbents and DAC materials is to remove carbon dioxide from the environment; DAC materials address more diluted CO2 and CO2 capture sorbents are used for industrial point sources. For both concentrations, metal-organic frameworks (MOFs) are the most promising candidate, since their cage-like structure can be precisely tuned to vary selectivity and capacity depending on where they must be deployed. MOFs are essentially organic linkers bonded to metal ions, and their appeal over other solid sorbents is that they have a more customisable porous structure and large chemical selectivity. The challenge is finding MOF candidates that combine high CO2 selectivity with low regeneration energy, and the large variety of possible frameworks makes selecting the best candidate for a given application a serious computational problem. The idea is that quantum computing might identify good MOF candidates and clarify, at the electronic-structure level, how the choice of metal governs the interaction with CO2 and with competing atmospheric gases like H2O and N2.
Four gates of quantum advantage
Theoretical quantum advantage
Quantum phase estimation might be used to estimate the ground-state energy of the system, and so predict adsorption energies more accurately than approximate classical methods. Unfortunately, CO2 uptake in a MOF like Mg-MOF-74 is dominated by physisorption, and the experimental binding energies are weak. Weak interactions, which are also affected
by dispersion effects and electrostatics, can already be modelled adequately by certain DFT techniques. There is a minority of open-shell transition-metal sites (an Fe centre, say) where multireference character makes the simulation harder, where quantum phase estimation might provide genuine advantage. Unfortunately even in this classically hard regime, there is the fact that exponential speedup is contingent on preparing an initial state with non-vanishing ground-state overlap, which is not guaranteed as the active space grows (Lee et al. 2023).
Practical quantum advantage
Every hardware result to date is a proof of concept on a drastically simplified model, not a MOF. Barroca et al. reduced Mg-MOF-74 to a single Mg²⁺ metal centre and computed the potential energy surface for CO2, N2, and H2O using qubit-ADAPT-VQE within a four-qubit active space on IBM hardware, with twirled readout-error mitigation and no explicit noise model in the simulation. The method reaches chemical accuracy against a UCCSD-VQE benchmark while staying hardware-efficient, which is a positive result about the ansatz, but the single-ion approximation overestimates the CO2 binding energy by roughly a factor of six against experiment. Rocca et al. go further, treating periodic Fe-MOF-74 with a 28-qubit symmetry-preserving ansatz executed via sample-based quantum diagonalisation to keep circuit depth down, using DFT to compare against. These are the state of the art, and they are still small active spaces validated against classical methods, with no sign of outdoing them at the moment.
Economic quantum advantage
The Open DAC 2023 effort (Meta FAIR and Georgia Tech), which serves as the best classical baseline, ran DFT across about 8,000 MOFs to generate roughly 176,000 adsorption-energy calculations from around 38 million single-point evaluations, then trained machine-learning interatomic potentials such as EquiformerV2 on the result (Sriram et al. 2024). That pipeline screens candidate materials at a marginal cost per structure that a quantum computation of a single metal centre does not yet compare with, and the machine-learned models for simulation are getting cheaper and more accurate each year. Against a cost-equivalent classical machine there is no economic case.
Ecosystem considerations
Open Quantum Institute at CERN, with QCentroid and ETH Zurich, is exploring quantum generative models (GANs) to propose MOF candidates across the vast configuration space, which is another approach different from ground-state estimation. But nothing yet is deployed, the demonstrations are single-site models. Nonetheless, given this is a usecase that contributes directly to clean energy, it has drawn considerable interest from diverse stakeholder groups.