Article pubs.acs.org/est
Monitoring of a Simulated CO2 Leakage in a Shallow Aquifer Using Stable Carbon Isotopes Alexandra Schulz,*,† Carsten Vogt,† Hendrik Lamert,‡ Anita Peter,§ Ben Heinrich,⊥ Andreas Dahmke,§ and Hans-Hermann Richnow† †
UFZ - Helmholtz Centre for Environmental Research, Department of Isotope Biogeochemistry, Permoserstraße 15, 04318 Leipzig, Germany ‡ UFZ - Helmholtz Centre for Environmental Research, Department of Monitoring and Exploration Technologies, Permoserstraße 15, 04318 Leipzig, Germany § Institute for Geosciences, Christian-Albrechts-University Kiel, Ludewig-Meyn-Straße 10, 24118 Kiel, Germany ⊥ GICON - Tiergartenstraße 48, 01219 Dresden, Germany S Supporting Information *
ABSTRACT: Artificial carbon dioxide leakage into a shallow aquifer was monitored using stable carbon isotope measurements at a field site near the town of Wittstock, Brandenburg, Germany. Approximately 400 000 L of CO2 were injected into a shallow aquifer at 18 m depth over 10 days. The 13C/ 12C ratios of the CO2 were measured in both groundwater and soil gas samples to monitor the distribution of the injected CO2 plume and to evaluate the feasibility and reliability of this approach to detect potential CO2 leakage, for example from carbon capture and storage (CCS) sites. The isotopic composition of the injected CO2 (δ13C −30.5 ‰) was differentiable from the background CO2 (δ13C −21.9 ‰) and the artificial CO2 plume was monitored over a period spanning more than 204 days. The results demonstrate that this stable isotope monitoring approach can be used to identify CO2 sources and detect potential CO2 migration from CCS sites into overlying shallow aquifers or even into the upper subsurface. A significant difference between the isotope ratios of the natural background and the injected CO2 is required for this monitoring approach to be effective.
■
INTRODUCTION The development of the carbon capture and storage (CCS) technology, wherein carbon dioxide (CO2) is sequestered indefinitely in underground storage areas, has progressed rapidly in recent years. This technology is considered an option for the reduction of CO 2 emissions to the atmosphere.1,2 The potential for leakage of stored CO2 through faults, boreholes, or permeable rock formations into shallow aquifers is one concern associated with CCS.3−5 Until now, no direct observations of leakage from a CO2 storage site existed. Thus, only natural and/or industrial analogues were used to predict the effects of CO2 leakage in shallow groundwater. A review by Lemieux6 pointed out the necessity of further field studies to establish data for the behavior of CO2 in a shallow aquifer. Monitoring tools for detecting potential leakages of CO2 from CCS sites in deep geological formations are needed for the development of early warning systems for assessing the tightness of cap rocks, and for a rigorous evaluation of risks associated with CCS technology. Thus far, several methods have been developed for measuring CO2 concentrations associated with underground leakage. In the near-surface environment for example, the © 2012 American Chemical Society
chamber method is used, which is based on measurements of CO2 accumulations at the potential leakage site,7 whereas eddyflux methods analyze net CO2 leakage flux over a large area using a fast CO2 analyzer.8 Parameters like pH and/or electrical conductivity allow for the monitoring of overall CO2 flux in shallow aquifers, but source identification is not possible. On the one hand, the problem is insufficient spatial resolution for realistic monitoring tools; on the other hand, the discrimination between biogenic and fossil sources of CO2 is an unresolved disadvantage for both techniques. Stable carbon isotope analysis has the potential to address both of these shortcomings. CO2 occurs naturally in both aquifers and in the shallow soil zone. The challenge for a CCS monitor seeking to provide early warning or conduct risk assessment is to distinguish naturally present CO2 from potential underground CO2 leakage. The geochemistry of a shallow aquifer will be affected by the upward migration of CO2 from a storage site. Batch-reactor experiReceived: Revised: Accepted: Published: 11243
July 4, 2012 September 11, 2012 September 24, 2012 September 24, 2012 dx.doi.org/10.1021/es3026837 | Environ. Sci. Technol. 2012, 46, 11243−11250
Environmental Science & Technology
Article
Figure 1. Map of Germany with the field site in the northeast (left), and detailed plan for the monitoring field of the CO2 injection test site. The plan is showing the installed monitoring wells and the three CO2 injection lances (red) (modified after (Peter et al., 2012)). The control levels (B−F) are explained in more detail in the text.
an average value of −2.8 ‰.24 These results show that, even at field sites with naturally increased groundwater-CO2 concentrations, interference from the background isotope signatures with the injected CO2 isotope signatures cannot be expected. Therefore, the different isotope ratios of carbon dioxide in shallow confined aquifers can aid in discriminating between carbon sources. A CO2 leakage scenario was simulated at a field site in Germany by injecting carbon dioxide into a shallow aquifer.25,26 To test whether the CO2 plume of a leaking CCS site could be assessed, spatial and temporal development of the plume was monitored in several ways: the isotope composition of CO2 in soil gas and groundwater samples, geochemical parameters (pH, total inorganic carbon - TIC, electric fluid conductivity), and geoelectrics for the groundwater samples. CO2 (397 938 L, 787 kg) were injected 18 m below ground level (bgl) into a shallow aquifer over a period of ten days. Groundwater sampling and soil gas measurements were performed to determine whether the released CO2 migrated from the aquifer to the upper soil horizon and whether it was possible to detect migrating CO2 within a reasonable time frame. The overall goal was to determine whether stable isotope measurement of CO2 is a reliable monitoring tool for detecting potential leakage events from CCS sites. A further aim was to test and evaluate the application of the soil gas sampling system as a routine monitoring method.
ments,9 computational simulations,10−13 and a small-scale field study at the zero emission research and technology (ZERT) site14,15 all suggest the mobilization of cations from the host rock, pH decrease, and electrical conductivity increase following an underground CO2 release. These parameters have the potential to be used as indirect indicators of CO2 intrusion.10 However, measurements of CO2 concentration or indirect parameters can both be distorted by naturally occurring spatial and temporal variations in the CO2 inventory of soil−aquifer systems. This impedes upon the ability to reliably identify CO2 sources and suggests the necessity of isotope monitoring to allow the clear attribution of detected CO2 to its source. Several researchers have used and established a stable carbon isotope approach to determine changes in fluid isotope composition due to CO2 injection16 or to quantify ionic trapping of injected CO2 as HCO3−.17 Furthermore, the isotope signature of CO2 was used for baseline monitoring before CO2 injection to detect anomalous locations that could act as possible leakage pathways.18 Additionally, at the ZERT site a laser-based analyzer capable of measuring both CO2 concentrations and isotope ratios was used to monitor a controlled subsurface CO2 release event.19 In this study, the behavior of an artificially generated gaseous CO2 plume, caused by a controlled and temporally limited CO2 injection experiment, was monitored by stable carbon isotope analysis. The natural isotopic composition of CO2 in groundwater is determined by the isotopic composition of CO2 in the atmosphere, the composition of carbonate minerals, and by microbial recycled plant material. Conversely, the CO2 to be stored in the underground sites is produced by the combustion of fossil fuels and has the isotope composition of the carbon source (i.e., oil, coal, or gas). This source is generally more isotopically depleted in δ13C (
