Mineral reactions in glauconitic sandstones induced by CO2 experiments: Project Greensand Phase 2 WP3

Glauconitic clasts constitute a significant proportion (25-35%) of the targeted sandstones for CO2 storage in the Nini West depleted oil field. Their stability plays a crucial role in maintaining reservoir integrity under changing environmental conditions, including burial diagenesis or CO2 injection. The glauconitic clasts, composed of mixed-layer glauconitic mica/Fe-smectite, undergo a gradual process of illitization during burial at depths ranging from 1700 to 2550 meters. This is evident from reduced expandability, increasing K-content, and decreasing Fe-content in the clasts with increasing burial depth, as observed in wells where the reservoir is present in different parts of the depth interval. The successful storage of CO2 in subsurface reservoirs requires a stable formation that exhibits minimal reactions with CO2 or predictable and favourable reactions. In the case of glauconitic sandstone, the stability of glauconitic clasts is of utmost importance. The reactivity of these clasts depends on various factors such as pore size distribution, smectitic content, and the presence of coatings or cement that may hinder interactions between glauconitic clasts and formation water.
To ascertain the stability of glauconite and other mineralogical reactions, we conducted comprehensive studies on 32 core samples and 6 sediment trap “fines” samples from the Frigg sand in the Nini-4 well's oil leg. These samples were provided as part of the experiment conducted on Project Greensand Phase 2. They included plug end trims, reacted plug ends, tested plug interiors, and particles collected from sediment traps in the flow rings, also known as "fines". Our study employed various techniques, including Thin Section Analysis, Scanning Electron Microscopy (SEM), X-Ray Diffraction (XRD), Coupled Emission Spectroscopy (CES), Automated Quantitative Mineralogy (AQM), and Micro Computed Tomography (micro CT), or combinations of these methods.
The experiments from Greensand Phase 2, Work Package 3, included in this petrography report are:
1. Three-phase flow experiments of varying flow rates (WP 3.1 Exp-1, -2, and -3; Mohammad-khani et al. 2023).
The objective was to test the response of glauconitic sandstone plugs to cyclic injection of scCO2 at reservoir conditions and in the presence of hydrocarbons and formation water, hence three-phase flow.
2. Long tube experiment (WP 3.4; Narayanan et al. 2023).
The objective was to test salt precipitation during scCO2 injection in crumbled glauconitic sandstones mounted in a long (1 m) coil.
3. Geochemical two-phase flow experiments (WP 3.5 Exp-1 and -2; Kazmierczak et al. 2023, Holmslykke et al. 2023).
The objective was to test the reactions in glauconitic sandstone plugs and formation water during injection of scCO2 with and without impurities (NOx, SOx) at reservoir conditions and in the presence of hydrocarbons and formation water.
4. Acid Test Batch 1, 2, and 3, in which reservoir rock is exposed to acid or acidified formation brine (pH less than 1 to 2) (WP 3.5).
G E U S 5
The objective was to mimic the acidic conditions expected to be generated when injecting scCO2 with impurities (NOx, SOx) and investigate the accompanying reactions in the glauconitic sandstones.
The primary findings indicate that most framework grains remain largely unaltered following laboratory experiments. However, there are exceptions: feldspar grains and glauconitic clasts may exhibit minor alterations, especially in the samples treated with acid at very low pH. The K-feldspar grains exhibit more extensive internal fracturing after certain experiments, which could possibly be due to mechanical testing or increased sleeve pressure. Pressure build-up was noted in some experiments, but it was generally considered insufficient to cause the fracturing. Either of these factors could be responsible for the fracturing of feldspar grains, while glauconitic clasts seem minimally affected. It is possible that glauconitic clasts can be compressed and expand without causing significant fracturing. In one instance of Acid Test Batch 1, severe mechanical damage was observed in glauconite clasts, possibly due to swelling or capillary forces acting within the grain.
Acid tests were conducted at a very low pH below 1 (Acid Batch Test 1 and 2) and at a pH of 1-2 (Acid Batch Test 3) using a highly potent mixed acid solution. These tests demonstrated that at very low pH, glauconite reacts, and the rock framework deteriorates. At a slightly higher pH, still considered lower than the expected operational window, the glauconite rock matrix remains stable. Only a limited chemical exchange reaction was observed over a week-long reaction time, with no apparent geomechanical weakening as a consequence.
The pH-dependent reaction of glauconite and carbonate minerals was expected and clearly indicates that the pH should be kept within a reasonable range in the reservoir to avoid or limit reservoir failure leading to injectivity losses.
Internal mobilisation of particles, i.e., fines relocated within the sandstone plugs, was observed in one flow experiment (WP 3.1 Exp-2) that subjected the plug to the highest maximum flow rate. Here, the first 160 pore volumes of injection in Exp-2 were conducted using 800 ml/h. An increase in pressure occurred after 110 pore volumes of injection. Thus, high flow rates may result in the mobilisation of relatively large particles, even though the high permeability counteracts the overall effect on the reservoir sandstone. Evidence of the formation of new minerals, including barite, cel-estite, and gypsum, in addition to halite, has also been documented after some experiments. These sulphates form due to ions liberated from K-feldspar reacting at elevated SO42- pore water concentrations from impurities such as SOx. Halite forms due to the drying out of pore water. The impact of this on permeability has not yet been documented.
Summary of risks identified in relation to the glauconitic nature of the Nini reservoir sand
In the Greensand Phase 1 project, four main risks were identified in relation to the glauconitic nature of the Nini reservoir sand (ID 204, 205, 207, 208 in the Greensand risk register). The risks were all operational and related to the inability to achieve the planned injection rates.
The risks included:
• Swelling of glauconite in the presence of CO2. This is considered plausible since glauconite is a clay mineral that could reduce permeability and injectivity.
• Glauconite mobilisation. With high injection rates, glauconite grains could be mobilised, leading to local reductions in permeability due to fines migration.
6 G E U S Report 24/2023
• Reservoir framework deterioration due to high flow rates. When subjected to high injection rates, the reservoir may lose structural strength, leading to compaction and reduced permeability. Glauconite could be the weak mineral in the framework, determining the overall strength.
• Reservoir framework deterioration due to geochemical reactions with CO2. Dissolution of framework minerals negatively affects the geomechanical stability of the reservoir, reducing permeability. Glauconite stability is pH-dependent, especially when impure CO2 streams containing impurities such as SOx and NOx generate acidic conditions that could make glauconite unstable.
Swelling of glauconite in the presence of CO2:
Our studies have identified a significant observation regarding glauconite's behaviour under CO2 exposure. We found that some glauconitic clasts could be encapsulated by a residue of solidified heavy hydrocarbons, such as solid bitumen/asphaltenes. This encapsulation appears to remain intact even during scCO2 flooding, potentially providing a protective shield against glauconite swelling. Interestingly, the presence and penetration of hydrocarbons into the clasts seemed to vary, which might affect the process. This variation is particularly notable as it relates to the reduction of iron, where hydrocarbons might facilitate the conversion of Fe3+ to the more soluble Fe2+ (Petersen et al., 2023; Weibel et al., submitted).
Glauconite mobilisation:
The composition of fines in both Greensand Phases 1 and 2 closely matches the glauconitic clast composition before the experiment. This suggests that the fines are mainly remnants of the clasts, likely due to the pre-treatment of samples, which removed a substantial amount of fines. Internal mobilisation of clays has been observed, and potential sources of fines from precipitating sulphate minerals have been documented. This suggests that glauconite may be mobilised at the highest flow rates. No effect on permeability has been observed.
Reservoir framework deterioration due to high flow rates:
In the experiment with the highest flow rate, we noted particle mobilisation both externally caught in traps and internally seen as clay drapes. Despite this, the high permeability was maintained in the sandstones, which suggests that this mobilisation may have only a minor impact on overall reservoir properties. This conclusion, however, presumes that not all glauconitic clasts begin to disintegrate, a scenario which would pose a severe threat to reservoir stability.
Reservoir framework deterioration due to geochemical reactions with CO2:
Changes in the formation water composition, particularly a reduction in salinity or Ca2+ content, could potentially initiate the swelling of glauconitic clasts. While this is unlikely to impact the static storage of CO2 in the glauconitic sandstones, it could cause issues during cyclic scCO2 injection due to the reduction in flow properties caused by disintegrated glauconitic clasts. However, salinity is expected to increase due to the drying effects. Only precipitation of gypsum could destabilise the clasts by reducing the Ca2+ level; however, this remains hypothetical.
Overall, while these risks are significant, our current geomechanical testing data does not suggest that any drastic revisions to our geomechanical model for the reservoir rock are required. Specifically, our findings do not demonstrate any substantial changes in rock strength due to exposure to scCO2.