Quantification and division of unfrozen water content during the freezing process and the influence of soil properties by low-field nuclear magnetic resonance

In this study, a low-field nuclear magnetic resonance (NMR) apparatus with the ability to adjust sample temperatures was used to measure the unfrozen water content of frozen soils sampled from the Qinghai-Tibet Plateau. Based on low-field NMR theory, two cutoff values were proposed to quantitatively identify three types of unfrozen water (bulk, capillary, and bound water) in the frozen soils. The influences of soil properties (the clay content, particle size distribution, and clay mineral content) on these three different types of unfrozen water during the soil freezing process were analyzed. The results showed that the soil-freezing characteristic curve of unfrozen water could be divided into three stages, namely, the unfrozen, the rapid-drop, and the residual stages. In the rapid-drop stage, the unfrozen bound water content increases gradually with the increasing clay or clay mineral content, and at a certain temperature, the unfrozen capillary water content increases as the sand and silt content increases. A published power function model is then used to fit the measured data in this study. We found that the single parameter related to soil properties in this model could not sufficiently reflect the influence of clay mineral content or clay content on unfrozen water content.