Estimating the nuclear magnetic resonance surface relaxivity of Eocene sandstones: A comparison of different approaches

Abstract

Nuclear magnetic resonance (NMR) relaxometry typically involves the analysis of a relaxation time distribution. The surface relaxivity ([Formula: see text]) is the key parameter that relates the relaxation time to the pore radius. Only a good estimate of the surface relaxivity enables a reliable determination of the pore radius distribution in a rock or sediment sample. A wide variety of approaches for the estimation of [Formula: see text] has been proposed; however, the accuracy of [Formula: see text] determination approaches has rarely been checked. We have compared different approaches of [Formula: see text] determination for a set of Eocene sandstone samples. Most approaches based on a weighted logarithmic mean of relaxation times or the peak relaxation time result in significant underestimation of [Formula: see text]. However, the correct weighting of the set of relaxation times has proven to be a crucial approach in [Formula: see text] determination. The consequent application of geometric rules suggests the application of the weighted harmonic mean ([Formula: see text]). The specific surface area per unit pore volume ([Formula: see text]), which results from the gas adsorption method, is another crucial parameter in most approaches for [Formula: see text] estimation. The quantities [Formula: see text] and [Formula: see text] depend on the resolution of the used method. Applying the fractal theory, we adopt an approach that performs an upscaling of [Formula: see text] to the resolution of the NMR relaxometry. Using equal resolution for [Formula: see text] and [Formula: see text], we obtain more reliable [Formula: see text] estimates. The resulting [Formula: see text] values are comparable with the ones determined by using the median relaxation time from NMR and the median pore-throat radius from the mercury injection capillary pressure method.