Using stable carbon isotopes to track potential leakage of carbon dioxide: Example from an enhanced coal bed methane recovery site in West Virginia, USA

• CO2 injected into 396 m deep Upper Freeport coal seam.
• Wells drilled in overlying coal seam, aquifer and soil used to monitor CO2 leakage.
• Carbon isotope signature of different end members at site are distinct.
• δ13C signatures of injected/produced/ soil CO2, and DIC used to monitor leakage.

Produced natural gas, shallow groundwater, and soil vadose gas samples were collected at the CONSOL Energy Inc. CO2 Sequestration Pilot Test Site in Marshall County, WV to test the feasibility of using carbon isotope signatures to detect potential leakage of CO2. CO2 was injected into the Upper Freeport coal bed at a depth of ∼396 m intermittently from September 2009 to December 2013. Water and gas samples were collected over a 12-month period (August 2013–August 2014). The injected CO2 had δ13C values ranging from −12.0‰ to −11.0‰ V-PDB. The average δ13CCO2 values of the soil vadose gas, the groundwater, and the produced CO2 from overlying Pittsburgh coal bed were −25.0‰, −16.6‰ and +22.8‰ V-PDB, respectively. These distinct carbon isotope signatures allow us to use δ13C as a natural tracer to detect any potential leakage from the injected coal bed into the overlying monitored zones. In September 2013, the monitored downstream production well in Upper Freeport showed presence of injected CO2 indicating injected CO2 plume had migrated in the coal bed over the entire study area. However, the carbon isotope values observed in the overlying formations indicate no significant leakage of injected CO2 from the underlying Upper Freeport coal bed.