Emulsification in Porous Media: From Static Phase Behavior to Dynamic Displacement Process

Emulsions play a crucial role in various situations and applications, from everyday tasks such as cleaning oily dishes to advanced processes like tertiary oil recovery and the remediation of soils from non-aqueous phase liquid (NAPL) contaminants. They facilitate effective oil removal and enhance hydrocarbon recovery. Traditional phase behavior experiments, with which emulsification is assessed by mixing oils and aqueous surfactant solutions in test tubes, help optimize injection water chemistry. In these experiments, the fluids separate after mixing by gravity, and the resulting phases are then investigated at equilibrium conditions. Nevertheless, recent studies have revealed that emulsification in porous media under flow conditions can vary considerably from the results obtained in traditional tests. In particular, the turbulent mixing that occurs in test-tube experiments differs from the laminar conditions that occur in pore space flow, where fluids interact with one another in a gradual and limited manner, with minimal contact between the phases. Despite extensive research, understanding of emulsification in porous media remains limited, primarily due to reliance on core flooding experiments that typically provide insufficient data for reconstructing complex displacement processes. Recent advancements in pore-scale physics with high-resolution imaging have offered deeper insights, but studies predominantly use generic fluids that do not accurately represent the behavior of natural NAPLs. The present thesis aims to bridge this gap by systematically investigating emulsification with synthetic and natural fluids. Emulsification is investigated for different fluid compositions, in various flow regimes and by using high-resolution imaging, such as optical imaging in a microfluidic 'lab-on-a-chip' and micro-computed tomography (¿CT)-based core flood experiments. The study integrates visual, statistical, and topological analyzes to characterize the oleic phase by examining different fluid pairs within a defined chemical parameter space. Advanced image segmentation and machine learning techniques provide in-situ characterization of emulsion phases. The integration of statistical, topological, and microstructural data allows for the qualitative and quantitative determination of the interplay of emulsification and displacement efficiency. Furthermore, it was possible to refine experimental and numerical workflows in a way to establish a highly efficient screening tool for designing and optimizing injection fluids for various applications.