Role of mineralogy in controlling fracture formation

Abstract

Geologic CO2 storage systems rely on impermeable caprock layers to prevent fluid leakage and maintain system integrity. In these systems, mineral dissolution and precipitation reactions controlled by the reactivity of the accessible mineralogy can alter formation porosity and permeability over time. In the case of fractures, mineralogy that may otherwise be inaccessible can be exposed at the fracture surface. CO2-brine-mineral interactions at the fracture surface control the potential for CO2 leakage where reactions may increase or decrease fracture permeability, promoting or inhibiting leakage. Therefore, predicting the mineralogy that is likely to be present at the fracture surface will aid in understanding how the fracture will evolve. Here, shale samples taken from the Mancos formation (western United States) and the Marcellus formation (northeastern United States) are mechanically fractured via unconfined compression. The fractured surfaces are examined using scanning electron microscopy (SEM) energy dispersive spectroscopy (EDS) to create quantifiable mineral maps representative of fracture surfaces. This process is repeated on thin sections of the same cores to analyze near fracture rock matrices both parallel and perpendicular to the fracture. The remaining sample is used for x-ray powder diffraction (XRD) to determine the mineralogy of the bulk sample. After fracturing, the Mancos sample had two visually distinct lithofacies present at the fracture surface that were quantified using optical microscopy. Two mineral maps of each Mancos lithofacies showed the light layer is made up of quartz (25-40%), calcite (22%), and clay (27-48%), and the dark layer is rich in clay (71-79%). The matrices parallel and perpendicular to the fracture are comprised of 6 and 10% calcite, 39 and 44% quartz, and 45 and 27% clay, respectively. Compared to the XRD and matrix data, the dark lithofacies that makes up 71% of the fracture surface is high in clay content and low in quartz content. The light lithofacies that makes up 29% of the surface is slightly low in quartz and high in calcite content. As a result, for the Mancos sample, the fracture surface has relatively high clay content suggesting the fracture formed predominately through weak mineral phases. In contrast to this, the Marcellus results show the mineralogy of the fracture surface, matrix, and bulk sample as over 94% calcite. Because the Mancos sample showed interesting results, the mineral maps were further analyzed using autocorrelation to quantify the mineral distributions at the fracture surface and matrices. From this we have found that the induced fracture formed within a clay rich region about 400 microns thick. In this region, clay is more likely than quartz or carbonate to occur near the fracture, with the probability of occurrence decreasing as distance from the fracture increases. When comparing the fracture surface and matrix mineralogies, we see that the fracture surface mineralogy has a more uniform distribution at all distances.