Frost Damage in Unsaturated Porous Media

Frost damage in porous materials is a weathering mechanism that can cause dangerous rockfalls or damage to built cultural heritage. The volume expansion of 9% when water freezes is usually seen as the cause of frost damage. This does not, however, explain why partially saturated porous stones also show damage despite the fact that ice should have room to grow. By performing experiments both at the scale of a single pore and in a real stone, we propose an explanation for the mechanism of frost damage at low water saturations: the meniscus at an air-water interface confines the water in the pores. Because of this confinement, ice that forms will exert a pressure on the pore walls rather than growing into the pore. The amplitude of stress is found to be larger in small pores and when the meniscus has a larger contact angle with the walls. The contact angle is also observed to increase in the case of multiple freeze-thaw cycles, which increases the likelihood of damage. We find that cracks start first in the ice (being weaker than the confining material), followed by damage in the material itself. Remarkably, when multiple air-water interfaces are induced within limestone samples through a hydrophobic surface treatment, the stones are much more susceptible to frost damage than are uncoated stones, with cracks appearing preferentially at the hydrophilic-hydrophobic interface. This shows that indeed the meniscus confining the water during freezing and consequently the wetting properties are the relevant factors for frost damage in partially saturated porous stones.

Figure 1

(a) Setup of the experiment. (b) Cross section of a capillary.

Figure 2

Example of one measurement. Freezing pressure peak at point 1 and thawing pressure peak at point 2. The temperature of the Peltier element is shown by the dashed blue line.

Figure 3

Water droplet in the capillary over a single cycle: (a) before freezing, (b) when frozen, (c) just before thawing, and (d) thawed. The red arrow in (c),(d) indicates the presence of a fracture in the capillary after freezing.

Figure 4

Maximum pressures averaged over several freeze-thaw cycles measured for capillaries of different sizes (series A, B, and C) and water volumes (see Table 1 for the sizes of capillaries A, B, and C) . The error bars are given by the standard deviation in the measured pressures. The dotted line is a guide for the eye.

Figure 5

(a) Contact angles of the four corners of the water-air meniscus of an entrapped water droplet in a capillary over multiple freeze-thaw cycles. Measurements are done at the end of each cycle. (b) Contact-angle evolution over one freeze-thaw cycle. ${\theta }_1^w$ is the contact angle before freezing, ${\theta }_1^{\mathrm{ice}}$ is the contact angle when frozen, and ${\theta }_2^w$ is the contact angle after thawing.

Figure 6

Average contact angle versus peak pressure for the highest-volume droplets. On average, the pressure increases with the contact angle.

Figure 7

(a) Microscope pictures during the freezing of water in a partially hydrophobic capillary. The red lines show the interface between the hydrophilic side (1) and the hydrophobic side (2) of the capillary. (b) Enlargement of the red-rectangle region in (a), showing microcracks in the glass capillary close to the hydrophilic-hydrophobic interface.

Figure 8

Two cross sections of a drop in a cylindrical capillary. The radial stress from the crystallization pressure transforms into a tangential (hoop) stress in the wall of the capillary.

Figure 9

Three different ways water might be present in a porous medium. The grains are in gray, the water is in blue, and the hydrophobic coating is in red. (a) A fully saturated stone. (b) A Partially saturated hydrophilic stone, with the water forming a film around the grains and some capillary bridges. (c) A partially saturated stone with hydrophobic grains at the surface. (d) Limestone sample after treatment with a hydrophobic coating on its five outer surfaces: one can notice the wet, hydrophilic core (1) and the dry hydrophobic outer layer (2).

Figure 10

Untreated and treated limestone samples partially saturated (60%) with water and subjected to multiple freeze-thaw cycles. (a) An untreated stone, showing no damage. (b) Treated limestone samples with a hydrophobic coating at the surface: cracks and granular disintegration are indicated by white arrows. (c) A treated stone broken in two showing the depth of the cracks and the hole caused by granular disintegration. Water droplets are added as an indication for the hydrophobic character of the surface.

Figure 11

(a) Compression-test curves for untreated and hydrophobic-hydrophilic treated stones. (b) Under compression, the treated stone breaks at its weakest points, i.e., microcracks and holes develop after the freeze-thaw cycles.

Figure 12

comsol multiphysics model of the capillaries used in the experiment. An outwards pressure is applied from a section in the middle of the capillary, deforming the capillary.

Figure 13

Deformation constant of the $0.2\times 2{\mathrm{mm}}^2$ capillary type for different drop sizes as found by the comsol multiphysics modeling.

Figure 14

Stresses applied on the capillary wall during freezing.