Viewing frost heave on a microscopic scale.

In cold climates, spring often brings city streets strewn with potholes, highway surfaces punctuated by unexpected bumps, fields overgrown with freshly exposed boulders, and soils that feel spongy and soft.

These effects all result from a common phenomenon known as frost heave. It occurs when water-saturated soil freezes, pushing the ground up.

But this upward movement isn't due solely to the expansion that occurs when water turns to ice. A number of other factors come into play, including the existence of unfrozen water at temperatures below water's normal freezing point; this superchilled water tends to dribble through soil toward layers having a lower temperature, where it adds to the existing mass of ice.

Physicists have now studied this temperature-dependent flow at the flat surface of a single crystal of ice. These experiments confirm the presence of a thin layer of supercooled water alongside the ice crystal. The flow induced by a temperature gradient in the liquid layer causes the ice to grow at right angles to the water-ice interface.

Larry A. Wilen and J.G. Dash of the University of Washington in Seattle report their findings in the June 19 Physical Review Letters.

The research hinges on the idea that a solid in contact with its melted counterpart often develops a film of liquid along the interface; this film has a temperature lower than the solid's normal melting point. The effect is called surface melting.

In 1989, Dash proposed that modest temperature differences can drive flows in these liquid surface films and contribute to frost heave. In soil, this chilled but unfrozen water tends to migrate along temperature gradients toward lower temperatures, eventually accumulating and freezing into ice "lenses." It is these ice accumulations, created by water drawn to the freezing sites, that force the soil apart, both lifting and compacting it.

To test the idea on a microscopic scale, Wilen and Dash used an apparatus formed by punching a circular hole in a fiberglass wafer, then closing the hole with a glass plate on the bottom and a flexible, plastic sheet on the top (see diagram). They filled the resulting cavity with water, then cooled its center to a temperature below water's freezing point, while keeping its outer edge at a higher temperature.

Touching the membrane with a cotton tip chilled in liquid nitrogen initiated ice formation at the center. After the ice disk finished growing, the researchers used a microscope to examine changes taking place at an ice crystal's edge.

They observed that over several days, a pronounced ice ridge develops along the disk's rim, pushing up the flexible membrane. The formation of such a structure serves as evidence of surface melting and liquid flow along the thin film of water between the ice and the membrane.

"The ridge is due to water at the ice-membrane interface flowing radially inward toward the colder region," Wilen says. "This liquid film gets thinner and thinner as the interface gets colder and colder. Eventually, the melted layer becomes so thin that no more water can flow inward. The water freezes onto the ice to create the ridge."

Wilen and Dash also noted that the presence of ice crystal boundaries seems to greatly increase liquid movement. "You can see a distortion of the shape of the membrane that's much bigger than what you see just at the edge from a single grain [or crystal]," Wilen says. "Grain boundaries may play a very big role in enhancing this flow."

The researchers are now taking a closer look at the effects of impurities and crystal boundaries on liquid movement and freezing.