Studies conducted over several decades have shown that the freeze-thaw cycles are a process of energy input and output in soil, which help drive the formation of soil structure, through water expansion by crystallization and the movement of water and salts by thermal gradients. However, most of these studies are published in Russian or Chinese and are less accessible to international researchers. This review brought together a wide range of studies on the effects of freezing and thawing on soil structure. The following findings are summarized: i) soil structure after freeze-thaw cycles changes considerably and the changes are due to the mechanical fragmentation of soil coarse mineral particles and the aggregation of soil fine particles; ii) the particle size of soil becomes homogeneous and the variation in soil structure weakens as the number of freeze-thaw cycles increases; iii) in the freezing process of soil, an important principle in the variation of soil particle bonding is presented as: condensation →aggregation→ crystallization; iv) the freeze-thaw cycling process has a strong effect on soil structure by changing the granulometric composition of mineral particles and structures within the soil. The freeze-thaw cycling process strengthens particle bonding, which causes an overall increase in aggregate stability of soil, showing a process from destruction to reconstruction.
The freeze-thaw cycling process considerably changes the composition, structure, and properties of soils. Since the grain size is the most important factor in determining soil characteristics, our current research primarily aims to investigate dynamic changes of the soil fraction when exposed to freeze-thaw conditions. We observed two series of Moscow morainic clayey specimens (gQⅡm): (Ⅰ) the original series, and (Ⅱ) the remolded series. We subjected each series of soil specimens to different frequencies of freeze-thaw cycles (3, 6, 20, and 40 cycles), and we used granulometric tests to analyze both series before and after exposure to freeze-thaw conditions. As a result of our experiments, the granulometric compositions tended to be distributed evenly after 40 freeze-thaw processes (i.e., content of fraction for 0.1-0.05 mm was increased after 40 freeze-thaw cycles) because the division of coarse grains and the aggregation of fine grains were synchronized during the freeze-thaw process. The soil grains in both series changed bi-directionally. In the original series, changes of the sand grains were conjugated with the clay grains, and in the remolded series, changes of the sand grains were conjugated with the silt grains, because potential energy difference caused the division and aggregation processes to relate to the counteraction process. The even distribution of soil grain size indicated the state of equilibrium or balance. The granulometric compositions were altered the most during the sixth freeze-thaw cycle, because the coefficient of the intensity variation of the grain fineness (Kvar) had its maximum value at that time.