Pore network connectivity affecting the NMR relaxation in unsaturated porous media

Nuclear magnetic resonance (NMR) reveals pore water properties due to its unique sensitivity to water, making it a powerful tool in hydrogeological studies. By measuring the magnetization and relaxation time of hydrogen atoms, NMR enables estimation of water content, pore size distribution, irreducible and free water content, and hydraulic conductivity in geologic media. However, interpreting NMR data in the vadose zone remains challenging. While established relationships between NMR signals, pore structure, and physiochemical properties are reliable under saturated conditions, they often fail or yield significant errors in unsaturated environments due to the complex pore structure and solid-liquid-vapor interactions within vadose zone’s pore spaces. A key challenge in unsaturated NMR data interpretation is the pore coupling effect, where protons diffuse across multiple pore environments before relaxing. This phenomenon can distort NMR relaxation time distributions, resulting in averaged representations of pore networks rather than individual pore environments, leading to misinterpretation of NMR data. In this study, we investigate the impact of pore coupling on NMR signals using experimental and numerical methods. Using glass bead samples of different sizes (0.05-0.1 and 0.4-0.6 mm diameters) under different saturation states, we measure the NMR T₂ and T₂-store-T₂ measurements to quantify pore coupling phenomena. Our T₂-store-T₂ data show that the decreased saturation weakens the influence of pore coupling on NMR relaxation. We further scan our samples using micro X-ray computed tomography (µCT) to establish 3D structures with detailed structural characteristics and solid-water-air interfaces. We develop a numerical simulation framework incorporating geometric models derived from µCT scans, acquired using HECTOR at the Center of X-ray Tomography (UGCT) with the EXCITE network, to simulate the NMR T₂ and T₂-store-T₂ responses. This framework enables investigation of how various pore network structures and water distribution patterns influence NMR relaxation under different saturations, providing theoretical support for our experimental observations. Our findings enhance the understanding of NMR response in unsaturated porous media with the presence of pore coupling, providing improved interpretation strategies for NMR vadose zone characterization.