Geochemical interactions between supercritical CO2 and the Midale Formation. V : experiments investigating reactions of the Midale Vuggy

This report describes work undertaken at the British Geological Survey (BGS) that forms part of the international IEA Weyburn Carbon Dioxide (CO2) Monitoring and Storage Project. This project aims to monitor and predict the behaviour of injected CO2 into the Midale reservoir at the Weyburn oil field in southern Saskatchewan, Canada, using methods that include; time-lapse geophysics, modelling its subsurface distribution and migration, and simulating likely chemical interactions with the host rock. This report aims to provide a description of fluid chemical and mineralogical changes occurring in a series of experiments that have been conducted within the Hydrothermal Laboratory of the British Geological Survey. These experiments were undertaken to identify what geochemical changes would result from the injection of CO2 into the Midale Vuggy formation. The experiments utilised samples of Midale Vuggy core material from the Weyburn field, synthetic formation water based upon measured well fluid compositions, and either CO2 or N2 as a pressurising medium. The experiments were conducted at 60°C and pressurised to either 150 bar [15 MPa] or 250 bar [25 MPa], using either CO2 or N2. Experiment durations ranged from one week to 6 months. The evolution over time of a selection of solutes was followed. Relative to the N2 ‘baseline’ experiments, it was found that the impact of CO2 was to: - increase the concentrations of Ca, Si and HCO3 - - decrease the concentrations of total S and possibly Sr, and pH values - have little impact on the concentrations of Mg, Mn and Al It is noted that these fluid chemical changes are not dissimilar to those found in the Midale Marly experiments (Rochelle et al., 2003a) All monoliths reacted in CO2-rich synthetic pore waters showed clear evidence of ‘tidemarks’ on their external surfaces, with the area below the water-CO2 interface appearing bleached. After 4 weeks of reaction of the monoliths with CO2, euhedral prismatic gypsum crystals up to 500 µm in length formed below the water line in the CO2 experiment. By 8 weeks reaction the gypsum crystals were at least 2.5 mm long, and at 17 weeks reaction gypsum crystals up to 500 µm long also developed in the baseline N2 experiment. In addition, most calcite and anhydrite surfaces below the water line were corroded to a depth of 10-30 µm in both the CO2 and the baseline N2 experiments. This porosity was easily distinguishable from the vuggy porosity developed during diagenesis. Scanning electron microscopy also revealed that a fine coating of halite developed above the water-CO2 interface during the experiment. In the experiments containing crushed Midale Vuggy, euhedral tabular prismatic gypsum crystals up to 1.8 mm long developed after 2 weeks reaction. Only limited evidence for minor corrosion was tentatively observed. After 26 weeks of reaction, the only evidence for dissolution in the <250 µm crushed samples was slightly less ‘dust’ in the baseline N2 experiment relative to the CO2 experiment. It was noted that the CO2 experiments give lower S concentrations compared to the N2 experiments, with S (as SO4) removed from solution by gypsum precipitation. During the early parts of the experiments at least, this appears to be faster than the rate of SO4 addition from anhydrite dissolution. Later in the CO2 experiments steady-state concentrations appear to be reached, and it is likely that saturation with respect to gypsum balances lower S concentrations with higher Ca concentrations. The changes described above were interpreted as being due to some calcite dissolution (probably more than observed in the Midale Marly experiments), some anhydrite dissolution, a little aluminosilicate mineral dissolution and a fair amount of gypsum precipitation. It is still unclear if there is an overall net increase or decrease in porosity or permeability. However, if significant gypsum precipitation reduced the permeability of the Midale Vuggy unit, then this may be a beneficial reaction in terms of the EOR operation, as it might reduce the potential for the injected CO2 to ‘under-ride’ the target Marly unit.