NMR relaxometry has developed into a method for rapid pore-size determination of natural porous media. Nevertheless, it is prone to uncertainties because of unknown surface relaxivities which depend mainly on the chemical composition of the pore walls as well as on the interfacial dynamics of the pore fluid. The classical approach for the determination of surface relaxivities is the scaling of NMR relaxation times by surface to volume ratios measured by gas adsorption or mercury intrusion. However, it is preferable that a method for the determination of average pore sizes uses the same substance, water, as probe molecule for both relaxometry and surface to volume measurements. One should also ensure that in both experiments the dynamics of the probe molecule takes place on similar length scales, which are in the order of some microns. Therefore, we employed NMR diffusion measurements with different observation times using bipolar pulsed field gradients and applied them to unconsolidated sediments (two purified sands, two natural sands, and one soil). The evaluation by Mitra's short-time model for diffusion in restricted environments yielded information about the surface to volume ratios which is independent of relaxation mechanisms. We point out that methods based on NMR diffusometry yield pore dimensions and surface relaxivities consistent with a pore space as sampled by native pore fluids via the diffusion process. This opens a way to calibrate NMR relaxation measurements with other NMR techniques, providing information about the pore-size distribution of natural porous media directly from relaxometry.

Magnetic impurities are ubiquitous in natural porous media such as sand and soil. They generate internal magnetic field gradients because of increased magnetic susceptibility differences between solid and liquid phase in the pore space and because of the presence of magnetic centers. These internal gradients accelerate NMR relaxation rates and thus might limit the possibility of pore space characterization using NMR. In this study, we investigate the effects of coating the surface of natural sands by the antiferromagnetic iron oxyhydroxide goethite on NMR relaxation and diffusion properties. We found a non-quadratic dependence of the relaxation time distributions on the echo time indicating that the relaxation experiments were not performed in the fast diffusion limit, while the weak dependence on the external magnetic field strength is explained by the preponderance of the surface relaxation over the effect of diffusion in internal gradients. The surface to volume ratio of the pore space, determined by NMR diffusimetry ((S/V) ) remains approximately constant, whereas the same quantity, determined from gas adsorption ((S/V) ) increases proportional to the coating density. This is because gas adsorption measures surface roughness on sub-nanometer scale, whereas NMR diffusimetry averages over structures smaller than few microns. This has consequences for the calculation of the surface relaxivities. The usage of the (S/V) leads to constant values, whereas the usage of (S/V) leads to apparently decreasing relaxivities with increasing coating, which is unrealistic. Copyright © 2016 John Wiley & Sons, Ltd.

No abstract

No abstract