Historical understanding of frost heaving Urban Hjärne described frost effects in soil in 1694. By 1930, Stephen Taber, head of the Department of Geology at the
University of South Carolina, had disproved the hypothesis that frost heaving results from molar volume expansion with freezing of
water already present in the soil prior to the onset of subzero temperatures, i.e. with little contribution from the migration of water within the soil. Since the
molar volume of water expands by about 9% as it
changes phase from water to ice at its bulk
freezing point, 9% would be the maximum expansion possible owing to molar volume expansion, and even then only if the ice were rigidly constrained laterally in the soil so that the entire volume expansion had to occur vertically. Ice is unusual among compounds because it increases in molar volume from its liquid state,
water. Most compounds decrease in volume when changing
phase from liquid to solid. Taber showed that the vertical displacement of soil in frost heaving could be significantly greater than that due to molar volume expansion. This excluded molar volume changes as the dominant mechanism for vertical displacement of freezing soil. His experiments further demonstrated the development of
ice lenses inside columns of soil that were frozen by cooling the upper surface only, thereby establishing a
temperature gradient.
Development of ice lenses road during spring thaw The dominant cause of soil displacement in frost heaving is the development of
ice lenses. During frost heave, one or more soil-free ice lenses grow, and their growth displaces the soil above them. These lenses grow by the continual addition of water from a groundwater source that is lower in the soil and below the freezing line in the soil. The presence of
frost-susceptible soil with a pore structure that allows
capillary flow is essential to supplying water to the ice lenses as they form. Owing to the
Gibbs–Thomson effect of the confinement of liquids in pores, water in soil can remain liquid at a temperature that is below the bulk freezing point of water. Very fine pores have a very high
curvature, and this results in the
liquid phase being
thermodynamically stable in such media at temperatures sometimes several tens of degrees below the bulk freezing point of the liquid. This effect allows water to percolate through the soil towards the ice lens, allowing the lens to grow. Another water-transport effect is the preservation of a few
molecular layers of liquid water on the surface of the ice lens, and between ice and soil particles. Faraday reported in 1860 on the unfrozen layer of
premelted water. Ice premelts against its own
vapor, and in contact with
silica.
Micro-scale processes The same intermolecular forces that cause premelting at surfaces contribute to frost heaving at the particle scale on the bottom side of the forming ice lens. When ice surrounds a fine soil particle as it premelts, the soil particle will be displaced downward towards the warm direction within the thermal gradient due to melting and refreezing of the thin film of water that surrounds the particle. The thickness of such a film is temperature dependent and is thinner on the colder side of the particle. Water has a lower
thermodynamic free energy when in bulk ice than when in the supercooled liquid state. Therefore, there is a continuous replenishment of water flowing from the warm side to the cold side of the particle, and continuous melting to re-establish the thicker film on the warm side. The particle migrates downwards toward the warmer soil in a process that Faraday called "thermal regelation." This effect purifies the ice lenses as they form by repelling fine soil particles. Thus a 10-
nanometer film of unfrozen water around each
micrometer-sized soil particle can move it 10 micrometers/day in a thermal gradient of as low as 1 °C m−1. As ice lenses grow, they lift the soil above, and segregate soil particles below, while drawing water to the freezing face of the ice lens via capillary action. ==Frost-susceptible soils==