(364ab) Low-Field Magnetic Resonance Relaxation: Signals, Mechanisms, and Applications in Porous Media

Research Interests: My research focuses on sensor physics applications, particularly using electromagnetic resonance tools. I investigate magnetic resonance signal processing through data-driven algorithm designs and spin relaxation mechanisms through molecular dynamic simulations. Nuclear magnetic resonance (NMR) not only aids in understanding porous formations (e.g., pore-size distribution, permeability, tortuosity, wettability, fluid characterization) but also has applications in remote sensing, exploration and wellbore imaging. My aim is to explore subsurface dynamics using sensor tools and promote energy sustainability, carbon neutrality, and digitalization with generative AI as a cross-cutting tool in the energy sector.

We have explored the use of NMR in unconventional formations by examining NMR relaxation signals and mechanisms. We introduced a novel Spliced NMR inversion method to separate liquid-like components with an exponential decay ($T_{2e}$) in transverse magnetization from solid-like components with a Gaussian decay ($T_{2G}$). The $T_1–T_2$ maps distinctly differentiate liquid-like signals (e.g., micro/meso/macro pore fluids, heptane dissolved in bitumen, clay-bound water) from solid-like signals (e.g., kerogen, bitumen, clay hydroxyls) in organic-rich chalks. This method enhances fluid typing and saturation analysis from liquid components and aids in clay mineral identification and kerogen content determination from solid components.

Building on this Spliced NMR signal inversion method, we further investigated kerogen quantity using the “2D splice NMR” method, which combines $T_1$ with solid-echo ($T_{2G}^*$) and spin-echo train ($T_{2e}$). Integrating this 2D splice NMR with Rock-Eval analysis, we studied organic matter in Type II-S organic-rich chalk as a function of maturity (i.e., depth), from immature to oil-window. We distinguished readily extractable bitumen from residual bitumen after solvent extraction based on depth. Additionally, we examined the elemental $H/C$ ratio, kerogen swelling, kerogen nanopore size, and compaction effects on macropores as a function of depth.

One notable phenomenon is the narrowing $T_{1,2}$ distribution observed in unconventional formations due to the cross-relaxation effect (a.k.a. spin diffusion). We studied the effect of $^1$H NMR cross-relaxation $\sigma_1$ (a.k.a., spin diffusion), which appears as a narrowing in the $T_1$ distribution, using a proposed metric $|\sigma_1|/R_1$ for relative cross-relaxation strength. These insights into $^1$H NMR relaxation provide valuable information about the molecular dynamics of viscous fluids, benefiting both the medical and energy fields without relying on the physics of paramagnetism.

Another intriguing observation is the mild increase in $T_{1_Ker}$ with depth for kerogen in organic-rich chalk. The decrease in the second moment $\Delta\omega^2$ ($\propto H/C$) with maturity partly explains this. We employed molecular dynamic (MD) simulations of realistic kerogen models with varying maturity to compute the NMR $^1$H-$^1$H dipole-dipole autocorrelation function. MD simulations provided new insights into the intramolecular versus intermolecular NMR relaxation in bulk kerogen molecules with varying maturity. We combined MD simulations with the Plateau model to predict NMR relaxation of bulk kerogen molecules in the slow-motion regime. We found a consistent trend between the simulated $T_{2G}$ (and $T_1$) versus H/C and the trend from NMR measurements of Type II-S organic-rich chalk as a function of maturity using a solid-echo pulse sequence to detect solid kerogen. This finding is crucial during the energy transition, as kerogen can serve as a clean-burning energy source and a potential sink for geological carbon capture and storage (CCUS) in porous media.