Date of Award
Doctor of Philosophy (PhD)
Geologic CO2 sequestration (GCS) in subsurface saline aquifers is a promising strategy to mitigate climate change caused by increasing anthropogenic CO2 emissions from energy production. At GCS sites, interactions between fluids and geomedia are important because they can affect CO2 trapping efficiency and the safety of CO2 storage. These interactions include the dissolution and precipitation of minerals. One of the most important minerals is calcium carbonate, because it can permanently trap CO2.
In this work, Portland cement was used as a model geomedium to investigate the chemical reactions, mechanical alterations, transport of reactive fluids, and the interplay of all these aspects. Also, because Portland cement is used in building and decommissioning CO2 injection wells, its alteration is important for wellbore integrity. Wellbore cement can deteriorate as a result of extensive reactions with injected CO2. Typically, a carbonated layer forms, which can partially reduce CO2 attack by clogging pores in the cement. We conducted high temperature/pressure experiments using Portland cement paste samples, and after 10 days of reaction, quantified the chemical changes using scanning electron microscope backscattering electron imaging and X-ray diffraction. The mechanical changes were quantified as well using a three-point bending setup and nanoindentation. The experimental results showed that after CO2 attack, the cement samples decreased in strength by ~80%, and this decrease was closely related to the formation of a wide and weak portlandite-depleted zone in the cement matrix immediately inside of the carbonated layer. The effects of 0.05 M of sulfate ions were also examined. Interestingly, the additional sulfate ions were found to mitigate CO2 attack by forming a more protective and less soluble carbonated layer, and thus a thinner portlandite-depleted zone.
To further investigate the detailed mechanisms by which the wide and weak portlandite-depleted zone formed and the carbonated layer’s surface dissolved, we set up a one-dimensional continuum reactive transport model using the CrunchTope software. Two mechanisms were found to be critical in reproducing our main observations: First, the precipitated CaCO3 could not fill the entire pore spaces in the carbonated layer. The inefficiency of CaCO3 precipitation in filling all the pores might be due to fractures and defects in the carbonated layer, or due to the extent of pore-size-dependent precipitation. Second, nucleation kinetics had to be incorporated into the model to predict the mineral precipitation observed in the reaction solution and to capture the dissolution of the carbonated layer’s surface.
To acquire parameters for the incorporation of nucleation kinetics, CaCO3 nucleation experiments were conducted primarily using atomic force microscopy and synchrotron-based in situ grazing incidence small angle X-ray scattering. Newly obtained interfacial energies were compared for mica and quartz systems, and a slightly higher interfacial energy was found in the quartz system. The effects of salinity were investigated in the range of 0.15–0.85 M ionic strengths, and we found a decrease of interfacial energies at high salinity. The kinetic factors, including the apparent activation energy and the pre-exponential factor in the nucleation rate equation, were experimentally obtained for the first time by varying temperatures in the range of 12–31 oC. These parameters provided the key information for modeling nucleation in geomedia and synthesizing well controlled materials in materials science.
The CaCO3 nucleation studies advanced our current understanding of nucleation under various conditions, and the acquired parameters were indispensable for our numerical simulations of the cement deterioration. The reactive transport modeling work revealed the important mechanisms in the cement–CO2 reactions, and provided many insights for understanding the chemical and mechanical alterations of geomedia. The investigation of cement deterioration quantitatively coupled the chemical and mechanical changes of the cement samples, and proved that the molecular scale of water–rock reactions can have a substantial impact on the change of the bulk geomedia. Such information can be also be applied to shale/sandstone–CO2 interactions. Overall, this dissertation presents a platform to understand fluid–geomedia interactions, combining experimental and modeling approaches, and connecting basic sciences and real applications. The advanced understanding of fluid–geomedia interactions will help improve GCS operation and thus address the climate change challenge.
Daniel E. Giammar, John D. Fortner, Fuzhong Zhang, Katharine M. Flores, Carl I. Steefel