Depleted gas reservoirs offer promising options for large-scale CO 2 sequestration due to their pre-existing pore networks, low formation pressure at the start of injection, demonstrated caprock integrity, and substantial operational history. However, safe and efficient storage requires a detailed understanding of coupled thermo-hydro-mechanical-chemical (THMC) processes, impurity interactions, and geomechanical responses. This review provides the first integrated synthesis that connects these multiphysics mechanisms in depleted gas reservoirs, addressing a gap not covered in existing CCS reviews. The study consolidates experimental, numerical, and field-scale evidence to explain interactions governing CO 2 behavior in carbonate, sandstone, and shale systems. Supercritical CO 2 injection maximizes storage efficiency but remains sensitive to phase transitions and hydrate formation, which may impair injectivity and containment. A novel contribution of this review is the combined assessment of impurity effects and depletion history, demonstrating how species such as H 2 S, SO x , NO x , and N 2 modify CO 2 properties, drive mineral reactions, and influence plume migration pathways in ways distinct from saline aquifers. Geomechanical risks, including compaction, subsidence, and potential fault reactivation, are linked to pore-pressure evolution and stress changes. Integrated monitoring strategies (such as downhole gauges, fiber-optic sensing, microseismic surveillance, InSAR, and time-lapse seismic) together with advanced THMC modeling enhance anomaly detection and guide adaptive operations. Insights from long-term field projects (K12-B, Rousse, Otway) confirm that with careful pressure management, impurity awareness, and continuous monitoring, depleted gas reservoirs can safely store CO 2 for decades. By unifying these mechanisms into a coherent framework, this review provides a novel and practical roadmap for optimizing storage performance and mitigating operational and geomechanical risks in depleted gas reservoirs.
Jahan et al. (Mon,) studied this question.