Geochemically induced shear slip in artificially fractured dolomite- and clay-cemented sandstone

Highlights

• Sandstone fractures were exposed to acidic or reservoir-simulated brine flow

• Acidic brine flow induced greater shear slip than reservoir conditions

• Acidic brine dissolved cementing dolomite, reducing fracture toughness

• Fracture surface roughness and porosity increased with carbonate dissolution

Abstract

Geologic carbon sequestration in deep saline aquifers results in a low pH brine plume that pushes into subsurface storage reservoirs and can access pre-existing or induced microfractures. This work investigates the effect of acidic brine on displacement of an artificial fracture under stress in the dolomite- and clay-cemented Bandera Gray sandstone. Samples were held under stress in a custom flow cell housed within an industrial CT scanner, and either acidic (pH 4) or reservoir-simulated (pH 8.3) brine was flowed through the artificial fracture for seven days. CT imaging shows that acidic brine resulted in greater shear slip than reservoir-simulated brine, with 0.379 ± 0.022 mm shear slip after pH 4 flow and only 0.213 ± 0.011 mm slip after pH 8.3 flow. Fracture surfaces exposed to acidic brine had rougher surfaces and lower fracture toughness, respectively, than those exposed to reservoir-simulated brine. SEM images of fracture surfaces indicate a loss by area of Fe-dolomite cementing crystals (6.05 ± 1.37% to 3.78 ± 0.73%) after exposure with the acidic brine, as well as a corresponding porosity increase (24.6 ± 1.1% to 26.1 ± 1.1%). These results indicate dissolution and weakening of the dolomite cements by acidic brine frees individual grains at the fracture surface to move, resulting in geochemically-induced stress release. Acidic brine created from geological carbon sequestration can dissolve sandstone cements and lead to increased shear slip at fracture interfaces, but further work at larger scales and with more realistic fracture conditions is needed.

Introduction

Anthropogenic-driven climate change is an immensely important societal challenge with the potential to adversely affect all life on Earth. A promising economical and technically viable strategy is CO₂ capture coupled with subsurface geologic storage of carbon (Geological Storage of CO₂, GSCO2). A large body of work already exists on carbon capture and storage as a mitigation strategy for carbon dioxide emissions. CO₂ can be captured from power plant emissions at 85-90% efficiency (EPA 2019). The pure gas can be transported to and injected into subsurface geological formations, such as depleted oil and gas reservoirs, deep saline formations, and coal seams. Deep saline formations represent the greatest potential storage capacity, with the United States alone having a minimum storage estimate of 1610 Gt of CO₂, compared to just 120 Gt in oil and gas reservoirs (Wright et al. 2013). During injection, the CO₂ can reduce brine acidity to a pH of almost 4 (Wigand et al. 2008), potentially inducing vigorous geochemical reactions. Several pilot and industrial scale efforts have probed the viability of large scale GSCO2 in deep saline formations. The first such industrial offshore injection took place at the Sleipner site in the North Sea, where more than 16 million metric tons of CO₂ have been injected into the Utsira formation since 1996 (Furre et al. 2017). Extensive site characterization led to numerous storage and reactive transport models, resulting in the prediction that carbonate and eventual feldspar dissolution will take place within the subsurface sandstone reservoir over thousands of years (Gaus, Azaroual, and Czernichowski-Lauriol 2005; Audigane et al. 2007). More recently, the Illinois-Basin Decatur Project (IBDP) in central Illinois injected 1 million metric tons of CO₂ in the Mt. Simon sandstone reservoir from 2011-2014. This resulted in significant increases in microseismic events that were detected away from previously known faults both during and following injection (Bauer, Carney, and Finley 2016). Several contributing reasons were identified, including the geochemical weathering of clay-cements (Fuchs et al. 2019) and the impacts of injection pressure on the stress-field with anisotropy in the geologic fabric (Will et al. 2014). These projects highlight the pressing need to understand how mineral reactions induced by GSCO2 affect reservoir geomechanical integrity, especially the promotion of seismic events.

Acidic brines have been shown to dissolve numerous minerals, and in some cases result in secondary mineral precipitation. Both batch and core-flood experiments have shown that particularly reactive mineral end members such as calcite and dolomite readily dissolve in low pH solutions (Czernichowski-Lauriol et al. 2006; Deng et al. 2015; Wang et al. 2013). Less reactive minerals like hematite, K-feldspar, and clays (e.g., illite, chlorite) also dissolve, but at slower rates (Wigley et al. 2013; Rosenqvist et al. 2014; Amram and Ganor 2005; Brandt et al. 2003; Malmström et al. 1995). Secondary mineral precipitation has been observed in heterogeneous reservoir samples exposed to CO₂ under supercritical conditions to form multiple types of minerals. For example, phyllosilicate clay dissolution and formation of nanoparticulate amorphous silica and kaolinite were observed in a batch study by using phlogopite (a phyllosilicate mineral) as a clay surrogate under supercritical conditions in 1 M NaCl brine (Shao, H., Ray, J., Jun 2010). In this case, the precipitates formed at the location of dissolution before they mobilized and aggregated. Weathering of K-feldspar, albite, and dolomite that had precipitated within pore spaces (called cements) were observed in the European Bunter sandstone under supercritical conditions (60°C, 15 MPa pore pressure) in a tri-axial flow cell. This dissolution and ensuing recrystallization led to precipitation of montmorillonite clays, although dissolution was the dominant geochemical reaction observed in the short term (i.e., weeks) (Wigand et al. 2008). Secondary mineral formation of carbonates has been shown to contribute to sequestration of carbon in reservoirs based on numerical simulations performed over tens to thousands of years (Xu, Apps, and Pruess 2005).

Reactive minerals can serve as cementation agents that bind larger, less-reactive grains together. Feldspars and clays often serve as primary binding agents in siliciclastic (sandstone) subsurface reservoirs, followed by carbonates, iron oxides, and quartz (Wigand et al. 2008). However, post-depositional cementation is a complex physical, chemical and biological (diagenetic) process in subsurface burial environments, including multiple cementing regimes, such as calcite and quartz overgrowths (Hangx et al. 2013; Walderhaug 1996; Dong et al. 2014). Digital image analyses via scanning electron microscopy (SEM) and petrographic thin section analysis have shown that dissolution of mineral cementation agents (e.g., clays, carbonates) can occur in reservoir samples exposed to CO₂-saturated brine (Yoksoulian et al. 2013). These dissolution reactions can increase porosity and permeability (Zemke, Liebscher, and Wandrey 2010; Yoksoulian et al. 2013), create preferential flow channels, and lead to further reactions, especially along these flow paths or faults. However, dissolution reactions can also result in mobilization of fines and pore blockage, which can reduce reservoir permeability (Egermann, Bemer, and Zinszner 2006).

The linking of geochemical reactions in subsurface reservoir sandstones to geomechanical properties has received less study compared to transport and hydrological properties (Rohmer, Pluymakers, and Renard 2016). Fundamentally, the injection of CO₂ changes the stress and strain field of the reservoir, which can induce fracturing or slip events in preexisting fractures or natural faults (Streit and Hillis 2004; Shukla et al. 2010; Huang, Yang, and Hall 2020; Rutqvist 2012; Jonny Rutqvist et al. 2016; Cappa and Rutqvist 2011), and possibly lead to the loss of CO₂ from storage formations (Rutqvist 2012; Shao et al. 2015). The preexisting fault orientation and geological structure determine how fractures propagate during and after injection (Huang, Yang, and Hall 2020), and as a result, the extent of induced seismicity (Goertz-Allmann et al. 2017; Jonny Rutqvist et al. 2016). Dissolution of cementing minerals is generally expected to increase porosity and decrease geomechanical integrity, leading to lower fracture toughness (Sun, Aman, and Espinoza 2016; Major et al. 2014) and lowered creep modulus (Akono et al. 2020). In natural CO₂ fields that serve as an analog for a faulted GSCO2 reservoir, such as Crystal Geyser, Utah, the seepage of CO₂-saturated brine into carbonate contained sandstone resulted in weakening as measured by strain softening and reduced shear strength of the sandstone; this was determined using studies comparing virgin rock and rock exposed to natural CO₂ (Major et al. 2014; Sun, Aman, and Espinoza 2016; Espinoza et al. 2018) as well as virgin rock exposed to CO₂ in the laboratory (Aman et al. 2018). This reduction in toughness is expected to promote stress-induced fractures and slippage along existing fractures or faults (Rutqvist 2012). However, compaction and strain localization with cements such as calcite can also result in higher compressive strength than pristine rock (Del Sole and Antonellini 2019). Alternatively, when secondary precipitation occurs in pore spaces, the rock can become more brittle with a reduced fracture toughness (Espinoza et al. 2018). All of these competing events make predicting the effects of CO₂-induced mineral reactions on geomechanical integrity challenging, especially the creation of fractures and slippage along existing fractures or faults (Yoksoulian et al. 2013).

The objectives of this study were to determine whether acidic brine promoted shear slip within a fractured sandstone reservoir, and if so, the mechanisms contributing to this process. An artificial fracture was created between two halves of a cylindrical core and put under stress in a modified Hassler-style, tri-axial core holder designed for flow and shear displacement. A sample of Pennsylvanian-age Bandera Gray sandstone was evaluated as a representative siliciclastic reservoir rock. It was flushed with either an acidic (pH 4) or reservoir-simulated (pH 8.3) brine, and experiments were monitored by X-ray computed tomography (CT) to identify changes in position and fracture dimensions in situ by direct image comparison (DIC). The measurement of shear slip in situ after a single applied stress is a novel use of dynamic X-ray tomography with sheared sandstone. The three-dimensional (3D) imagery was complemented by pre- and post-fracture surface analysis of physical and chemical properties using scanning electron (SEM) and optical microscopy, energy dispersive spectroscopy (EDS), cathodoluminescence (CL), optical profilometry, and scratch testing. Changes in fracture dimensions from CT images were used to evaluate changes in fracture flow using a 2D Stokes model. The experiments were designed to test the hypothesis that geochemical reaction of sandstone cements with acidic brine promotes dissolution and shear slip within an existing fracture. Because the relevance of geochemically-induced shear slip vs. pressure-induced shear slip in existing fractures at GSCO2 sites remains complex, these results provide valuable new insights to link these phenomena with microseismicity.