Geological storage

Geological storage puts captured CO2 back underground: into deep saline aquifers, depleted oil and gas reservoirs, and unmineable coal seams. It is the only disposal route currently operating at megatonne scale, and the only one with multi-decade field evidence behind it [2].

Below roughly 800 metres, pressure and temperature put CO2 into a supercritical state. It behaves as a dense fluid rather than a gas, occupying a small fraction of the volume it would at the surface, which is what makes the economics of underground disposal work at all. It remains buoyant relative to the brine around it, so everything in site selection comes down to whether that buoyancy can be contained.

The trapping hierarchy

Containment is not one mechanism but four, and they take over from one another on very different timescales [1].

• Structural and stratigraphic trapping is immediate. An impermeable caprock, typically shale or evaporite, physically blocks upward migration. This does the work in the first years and is what site characterisation mainly assesses.
• Residual trapping follows over years to decades. As the plume migrates, capillary forces strand disconnected bubbles of CO2 in the pore network, where they can no longer move as a continuous phase.
• Solubility trapping operates over decades to centuries. CO2 dissolves into the formation brine, and the resulting fluid is slightly denser than the original brine, so it sinks rather than rising. Buoyancy stops working against you.
• Mineral trapping is the endpoint, over centuries to millennia. Dissolved CO2 reacts with silicate minerals to precipitate stable carbonates, at which point the carbon is no longer mobile in any meaningful sense.

The practical consequence is that a well-chosen site gets safer with time, not riskier. Modelling of storage security suggests that for sites that are appropriately selected and regulated, well over 90 percent of injected CO2 is retained on a 10,000-year horizon, with the dominant residual risk coming from leakage along legacy wellbores rather than through intact caprock [1].

What the field record shows

Sleipner, in the Norwegian North Sea, has injected on the order of a million tonnes of CO2 per year into the Utsira sandstone since 1996, making it the longest-running dedicated storage operation in the world. Two decades of repeated seismic surveys tracked the plume as it spread beneath the caprock, and the monitoring record is the main empirical basis for confidence that conformance can be verified over time [2]. It also demonstrated something less comfortable: plume behaviour was not fully predicted by the pre-injection models, and history matching required revising the understanding of thin intra-reservoir shale layers.

Capacity, and why headline numbers mislead

Estimates of global storage capacity run to thousands of gigatonnes, which sounds like the problem is solved. It is not, for two reasons.

First, only a small fraction of pore volume is actually usable. Storage efficiency factors for deep saline aquifers are typically a few percent, governed by sweep efficiency, formation heterogeneity, and how much the reservoir pressure can be raised before fracturing the seal [3]. Theoretical pore volume and injectable volume are different quantities by an order of magnitude or more.

Second, capacity is not the binding constraint. Injectivity, well count, and project delivery rates are. Reaching the storage rates implied by mitigation scenarios means growing from today's tens of megatonnes per year to several gigatonnes per year, which is an infrastructure build-out problem on the scale of the existing oil and gas industry, not a geology problem [4][5]. Offshore continental margins are attractive here because they combine large sedimentary basins with fewer legacy wells and clearer regulatory ownership [4].

Open challenges

• Legacy wellbores. Old, poorly recorded, or badly abandoned wells are the most likely leakage pathway in mature basins, and locating them is often harder than characterising the reservoir [1].
• Pressure management. Injection raises pore pressure across a far larger area than the CO2 plume itself, which can induce seismicity or drive brine into shallower formations. Pressure interference between neighbouring projects in the same basin is an emerging regulatory question.
• Monitoring and verification. Demonstrating retention for accounting or credit purposes requires decades of measurement, on timescales longer than the commercial life of most operators.
• Liability after closure. Who owns the site once injection stops is a legal question with no settled international answer, and it bears directly on whether projects get financed [5].