Using laboratory methods, we investigate the effect of the presence and mineralogic form of iron on measured proton nuclear magnetic resonance (NMR) relaxation rates. Five samples of quartz sand were coated with ferrihydrite, goethite, hematite, lepidocrocite, and magnetite. The relaxation rates for these iron-oxide-coated sands saturated with water were measured and compared to the relaxation rate of quartz sand saturated with water. We found that the presence of the iron oxides led to increases in the relaxation rates by increasing the surface relaxation rate. The magnitude of the surface relaxation rate was different for the various iron-oxide minerals because of changes in both the surface-area-to-volume ratio of the pore space, and the surface relaxivity. The relaxation rate of the magnetite-coated sand was further increased because of internal magnetic field gradients caused by the presence of magnetite. We conclude that both the concentration and mineralogical form of iron can have a significant impact on NMR relaxation behavior.

A laboratory study was conducted to measure the effect of the mineralogic form and concentration of iron(II) [Fe(II)] minerals on nuclear magnetic resonance (NMR) relaxation rates of water-saturated sand mixtures. We measured mixtures of quartz sand and three common Fe(II)-bearing minerals in granular form: siderite (FeCO3), pyrite (FeS2), and pyrrhotite (Fe1−xS; 0<x<0.2) at two concentrations of iron by weight. The NMR response of these samples was used to calculate four transverse relaxation rates for each Fe(II) mineral mixture: total mean log, bulk fluid, diffusion, and surface relaxation rates. The surface area of the samples was used to calculate the surface relaxivity of the sample and the magnetically active surface. For each iron mineral, the mean log and surface relaxation rates were greater for samples with higher Fe(II) concentration. For the siderite, pyrrhotite, and high-concentration pyrite mixtures, surface relaxation was the dominant relaxation mechanism. Bulk fluid relaxation contributed significantly to the total relaxation for the siderite and pyrite mixtures; for the low-concentration pyrite mixtures, bulk fluid relaxation was the dominant relaxation mechanism. For the pyrrhotite mixtures, the diffusion relaxation rate was nonzero and slower than the surface relaxation rate; for the siderite and pyrite mixtures, the diffusion relaxation rate was zero. Surface relaxivity calculations revealed that, for the pyrite mixtures, relaxation occurred in the fast diffusion regime; for the siderite and pyrrhotite mixtures, relaxation did not occur in the fast diffusion regime. The range of surface relaxivity values calculated depends on mineralogic form. We conclude that Fe(II) concentration and mineralogic form are important factors in determining relaxation rate.

Quantifying the unfrozen water content of permafrost is critical for assessing impacts of surface warming on the reactivation of groundwater flow and release of greenhouse gasses from degrading permafrost. Unfrozen water content was determined along an ~12‐km transect in the Adventdalen valley in Svalbard, an area with continuous permafrost, using surface nuclear magnetic resonance and controlled source audio‐magnetotelluric data. This combination of measurements allowed for differentiation of saline from fresh pore water, and frozen from unfrozen pore water. Above the limit of Holocene marine transgression, no unfrozen water was detected, associated with high electrical resistivity. Below the marine limit, within several kilometers of the coast, up to ~10% unfrozen water content was detected, associated with low resistivity values indicating saline pore water. These results provide evidence for unfrozen water within continuous, thick permafrost in coastal settings, which has implications for groundwater flow and greenhouse gas release in similar Arctic environments.

Nuclear magnetic resonance (NMR) relaxation measurements are sensitive to the physiochemical environment of water in saturated porous media and can provide information about the properties of geologic material. Interpretation of NMR data typically relies on three assumptions: that pores within the geologic material are effectively isolated such that the diffusion of a proton between pores is limited (i.e., there is weak coupling); that relaxation occurs in the fast-diffusion regime; and that surface relaxivity ρ2 is uniform throughout the measured volume. We investigated the effect of spatial variation in ρ2 on the NMR relaxation measurement and evaluated two equations relating ρ2 to the NMR relaxation rate for samples containing two types of surfaces, each with a different surface relaxivity. One equation was valid when there is weak diffusional coupling between pores, the other is valid when there is strong diffusional coupling. We prepared a suite of samples composed of quartz sand and an iron-coated quartz sand. NMR relaxation occurred in two distinct regions: the weak- and strong-coupling regions. In the weak-coupling region, the equation did not accurately represent the relationship between the two ρ2 values and the NMR relaxation rate, suggesting that further research is required to understand the effect of spatially variable ρ2 in this relaxation region. In the strong-coupling region, the equation accurately represented the relationship between the two values of ρ2 and the NMR relaxation rate. The results from these laboratory experiments represented a first step towards accounting for spatial variability in ρ2 in the interpretation of NMR data.