Title: An Investigation of Laboratory-Grown "Ice Spikes"

Abstract: We have investigated the formation of 10-50 mm long ``ice spikes'' that sometimes appear on the free surface of water when it solidifies. By freezing water under different conditions, we measured the probability of ice spike formation as a function of: 1) the air temperature in the freezing chamber, 2) air motion in the freezing chamber (which promotes evaporative cooling), 3) the quantity of dissolved salts in the water, and 4) the size, shape, and composing material of the freezing vessel. We found that the probability of ice spike formation is greatest when the air temperature is near -7 C, the water is pure, and the air in the freezing chamber is moving. Even small quantities of dissolved solids greatly reduce the probability of ice spike formation. Under optimal conditions, approximately half the ice cubes in an ordinary ice cube tray will form ice spikes.
Guided by these observations, we have examined the Bally-Dorsey model for the formation of ice spikes. In this model, the density change during solidification forces supercooled water up through a hollow ice tube, where it freezes around the rim to lengthen the tube. We propose that any dissolved solids in the water will tend to concentrate at the tip of a growing ice spike and inhibit its growth. This can qualitatively explain the observation that ice spikes do not readily form using water containing even small quantities of dissolved solids.