UKCCSRC - Carbon Capture & Storage (CCS) Determining thresholds for CO2 leakage detection (CO2DETECTIVE) (Flexible Funding 2024)

Ceri Vincent, British Geological Survey, was awarded funding in the UKCCSRC’s Flexible Funding 2024 call to look at “Determining thresholds for CO2 leakage detection (CO2DETECTIVE)”.

The CO₂Detective project compared the efficacy of established and newly developed sensors and techniques with fixed and mobile deployments.

Setting the scene

In the unlikely event that geological carbon dioxide (CO₂) storage sites leak, surface expression will vary widely both temporally and spatially, so mobile techniques that can quickly and reliably screen large areas for further focused investigation are needed. However, sensitivity is often sacrificed for mobility, reducing the chances of successfully detecting leaks presenting with low CO₂ concentrations. Thus, techniques for point measurement remain a key aspect of leakage detection.

A mobile form of detection would also benefit other near-surface gas monitoring, for example when checking the integrity of subterranean pipelines that transport CO₂ from onshore industrial clusters to storage sites.

Tools for demonstrating the absence of any detectable leakage are important for project operators and for providing assurance and evidence to regulators, local populations, investors, insurers and policymakers.

Objectives of the project

The two objectives of the CO₂Detective project field work were:

benchmark the efficacy of established and newly-developed near-surface gas sensors and techniques with fixed and mobile deployments (‘near-surface’ in this context is less than around two metres above and below the ground surface)
assess the scalability of research findings to understand the opportunities, issues and risks to commercial-scale CO₂ transport and storage

Method

The project team created an artificial leak of a known quantity of CO₂ into shallow soil, to benchmark different soil gas detection techniques’ deployability and ability to detect leakage.

Field work was divided between three teams:

team one set up the leakage experiment
team two discharged and metered the CO₂ release
team three consisted of soil-gas monitoring experts who investigated the ‘leak’

Information was not shared between the teams, to create a realistic scenario where leaking CO₂ needed to be found and quantified at a site of suspected of escaping CO₂. The ‘leak area’ (roughly triangular, with sides around 30m each) was determined partly by practical considerations, including enabling CO₂ cylinders to be secured for safe operation.

Figure 1: Team one after setting up the ‘leak’. BGS © UKRI 2026

Figure 2: Team three looking for the ‘leaks’ using the ECCSEL BGS soil-gas facility. An ultraportable off-axis integrated cavity output spectroscopy and soil gas analysers are in the cart. A hand-held gas analyser is being used to analyse a sample drawn via a hollow steel probe driven into the ground. A portable CO2 flux accumulation chamber (analyser in the ‘backpack’) is also being deployed. Soil sample collection for the CO2GeoNet SLINKI project is also being carried out. BGS © UKRI 2026

Field work

Team three deployed sensors on two days during one week in February 2025. On the first day, the weather was calm. Winds were stronger on the second day, making locating leaks with the mobile sensors more challenging as the CO₂ gas quickly dispersed. The team located two main leaks and an intermittent leak using the ECCSEL BGS soil-gas facility.

Benchmarking

The aim of this project was to benchmark techniques rather than specific models, based on the experience of the field team when deploying the techniques, so manufacturer details are not provided here. The following techniques were deployed:

portable chamber-based CO₂flux
ultraportable off-axis integrated cavity output spectroscopy (OA-ICOS)
portable IR- and laser-based field gas meters, drawing soil gas from about 0.5 m below ground level via hollow probes driven into the ground

Deployed sensors were assessed for:

ease of field deployment
detection range, reaction and stabilisation times
ease of use, data retrieval and interpretation
robustness and service requirements

The portable CO₂ flux analyser scored highly against these criteria. The analyser weighs around 10kg and is supplied in ‘backpack’ format for portability. Acquiring data was easy and rapid; however, good data depends on getting a good seal with the ground surface. There is a risk of missing a leak, since this technique delivers point rather than continuous measurements.

The OA-ICOS also scored highly against these criteria. This technique can cover large areas quickly, reducing the risk of missing a leak signal, but the equipment was the heaviest (about 30kg) of our tested techniques, making field deployment more difficult. It had good sensitivity at near-baseline CO₂ concentrations (for better baseline characterisation and source apportionment) and it coped well with extremely high CO₂ concentrations, even when they exceeded the certified upper measurement range. It also flushed and recovered to background quickly. The OA-ICOS detected CO₂ at the soil surface for all but the lowest rate release during these tests.

Field analysis of CO₂ in soil gas accessed through hollow steel probes scored reasonably against the criteria by being fast, robust and portable. Driving the probe into the ground avoided atmospheric air being measured instead of intended soil gas. Gas samples could also be collected ‘end of probe’ for laboratory analysis, for example isotopic compositions to distinguish leaking stored ‘fossil’ CO₂ from biogenic CO₂ produced within the soil. This method generally produces a point measurement, which means a small area leak could be missed, and sensitivity at low concentrations may be inadequate for attributing CO₂ sources. Additionally, these meters cannot easily be owner-serviced and factory servicing (involving international shipping) can be time-consuming and expensive.

Given that the commercially available techniques benchmarked here can cost tens of thousands of pounds and some have to be sent to the manufacturer for servicing, the project team wanted to develop a lower-cost (a few hundred pounds) sensor in-house. A low-cost sensor could mean that, for a moderate budget, multiple sensors could be deployed to examine a wider area, or be left unattended at a suspected leak to monitor over longer periods and be easily replaced if lost. Our lower-cost sensor was developed using a Raspberry Pi and COZIR CO₂ sensor housed in a pelicase and was tested in the laboratory; improvements are planned for future projects.

Interpretation and upscaling

One of the greatest challenges with potential leaks is finding them, as CO₂ storage sites may cover several square kilometres and buried pipelines could be many kilometres long. Another challenge in monitoring for CO₂ in the near-surface zone is that CO₂ quickly disperses in air, therefore sensors must usually be close to the leak to find and quantify it. Highly mobile and quick-to-deploy surface measurement techniques that draw samples very close to the ground are generally protected from wind turbulence, therefore they are extremely valuable for locating potential leaks.

Another significant challenge for investigating potential leaks is distinguishing leaking CO₂ from CO₂ that is naturally present in soil and air at highly variable concentrations, depending on the season and the time of day, land-use and vegetation, as well as meteorological (especially humidity) and other environmental conditions. Understanding the origin of detected CO₂ requires site-specific knowledge of geology, soil and moisture, weather systems and other natural biotic and abiotic processes; findings need to be set within or account for these influences. Thus, when a leak location is found, following up with a campaign of detailed point measurements and methods, including collecting gas samples for isotopic analysis, is a natural next step.

Each line of evidence helps the CO₂ Detectives build towards a complete picture of site behaviour and sources of CO₂ at the surface. Both mobile and point measurement sensors have a place in CO₂ leakage detection and quantification. Deployability of techniques and being able to rapidly collect reliable data are important monitoring considerations; benchmarking these techniques illustrates the role for each in leakage detection.

Next steps

The OA-ICOS is already becoming one of our default wide-area scan methods, especially where options for mounting lasers on off-road vehicles are limited. For the time being, the OA-ICOS will continue to be deployed alongside benchmarked instruments while we further explore its potential in monitoring CCS.

Following on from the CO₂Detective project, funding is being sought to further develop the low-cost sensor.

Determining thresholds for CO2 leakage detection (CO2DETECTIVE) (Flexible Funding 2024)