Experimental constraints on the kinetics of ice lens initiation and growth

Abstract

Ice lenses are formed by the migration and solidification of unfrozen water during soil freezing, which can lead to the upwards displacement of the ground surface known as frost heave. The complicated interplay between heat and mass transport that causes ice lens formation has been addressed by several theoretical models, but uncertainties remain that require further experimental constraints. In particular, the initiation of ice lenses has long posed theoretical difficulties. We performed a series of stepwise freezing experiments in fine granular materials to observe the initiation and growth of ice lenses. Our experiments demonstrate clear and systematic relationships between the behavior of ice lenses, and the particle size and cooling temperature. Ice lenses are thicker when formed in sediments with smaller particle sizes and the initial formation position is further from the cooled boundary when it is set to lower temperatures. Our temperature measurements and photographic documentation demonstrate that ice lenses are formed below the nominal melting temperature, at a location that is sufficiently distant for the freezing velocity to have slowed below a threshold. We compared our experimental results to numerical predictions of ice lens formation that were applied to our experimental conditions. Our experimental trends are consistent with predictions of our simple, initial model. However, important quantitative differences motivate a refined treatment that emphasizes the kinetics of liquid supply from the pore space through the thin films that separate ice lenses from particle surfaces. We obtained good quantitative agreement between our experimental measurements and the refined model predictions, emphasizing the importance of kinetic effects as a control in ice lens initiation and growth.