Geochemical tracers for monitoring fluid mixing during a CO2-water injection field test

Abstract:

A series of injection-withdrawl (push-pull) well tests were conducted at the Otway CO2CRC field site using the CRC-2 well to determine the impact of injecting impurities (54 ppm SO₂, 9 ppm NO₂, 1100 ppm N₂ and 6150 ppm O₂) with a CO₂ stream on mineral dissolution/precipitation processes in a siliciclastic reservoir.

Four geochemical tracers were added to the injection waters of two sequential tests to monitor for any fluid mixing in the reservoir during the tests. Bromide and strontium were added as tracers to the injection water of test 1, and fluoresceine and lithium were added as tracers to the injection water of test 2. Injection waters in both tests were allowed to soak in the reservoir before they were back-produced to monitor for any water-rock interactions that took place.

The results suggest mixing of injection and in situ formation waters as well as reactivity of some of the tracers. Bromide behaves as an inert tracer and the concentration decreases by 6 and 15% after 11 and 21 days of soaking, respectively, suggesting minor fluid mixing in the reservoir. Fluorescein drops by about 50% after two days of soaking, which may be due to adsorption onto minerals under acidic conditions. Strontium and lithium concentrations decrease over time in excess to the bromide concentration decrease, this may indicate the precipitation of Sr and Li bearing mineral phases that are calculated to be supersaturated with respect to the composition of back-produced waters.

The decrease in bromide tracer concentration over time can be explained by preferential flow paths and hydrodynamic mixing during the soak period. Importantly, ideal tracer behavior was observed during a subsequent experiment where water was continuously back-produced. Our results show that typical (shallow) groundwater tracers need to be applied with caution when studying the hydrodynamics in a CO₂ storage reservoir. Further geochemical and hydrodynamic modelling is underway to fully explain our observations.