Compartmentalized Water Warping Experiment


When a 3% NaCl solution is frozen on the surface of a cement paste plate, the mechanical response is similar to that realized with pure water, but the kinetics of the response and the damage morphology are much different. Figure 1 shows that after the pool solidifies, and the temperature is reduced to -18°C, the sample exhibits a constant concave down (positive) deflection rate. It is beleived that this deflection is due to ice wedging, a reduction in the effective sample thickness (Figure 1), and creep - owing to the fact that the entire 450 μm deflection occurs over a period of 10 hours. The damage morphology and crack trajectory is exactly what is expected based on the prediction of the fracture mechanics analysis.

Figure 1 - Mechanical response of cement plate when 3% NaCl solution is frozen on the surface. On the right, the damage to the plate is displayed. The damage morphology and crack trajectory are identical to that predicted by the fracture mechanics analysis.

The difference in damage morphology realized with compartmentalized water and a 3% NaCl solution is due to the effect of solute on the ability of ice to wedge open the glue spall cracks. In the presence of salt, the stiffness of the ice is reduced, and the ice is always in contact with a solution at its melting point (above the eutectic temperature (-22°C)) . Therefore, the driving force for growth is supressed, and the ice does not force the fracture surfaces apart. In this case the glue-spall stress dominates the fracture process, and the crack turns into a trajectory that is parallel to the ice cement interface, as predicted by the fracture mechanics analysis. In the case of pure water the driving force for ice growth increases with undercooling, and the ice is capable of exerting high pressure on the fracture surfaces. Consider Figure 2, part (b); The curvature of the ice bludging into the pore mouth at N, is dictated by the amount of undercooling, and is given by the Gibbs - Thompson equation. Thus the ice along this liquid/crystal interface is in equilibrium with the adjacent water. Along the shoulder, at S, the ice would like to grow to obtain the same curvature at that at N. When the ice fills the crack, this is not possible, and the fracture surface must impose a confining pressure on the ice to suppress growth. The pressure that the ice exerts on the fracture surface at S is on the order of 1.3 MPa per degree of undercooling. The stress intensity at the crack tip of a crack whose fracture surface is uniformly loaded is on the order of 2.6ΔT MPa m0.5. Considering the fracture toughness of cement paste, 0.1 - 0.5 MPa m0.5, this indicates that ice wedging will promote crack penetration at very small undercoolings (ΔT).

Figure 2 - Schematic of the ice wedging mechanism from a macroscopic (a) and microscopic viewpoint (b). In the case of pure water, ice forms on the fracture surface, and draws moisture from the microstructure, quickly filling the crack with ice. Subsequently, the ice presses the fracture surfaces apart forcing the crack to penetrate into the body. In the case of salt solutions, the crack fills with brine from the pockets breached by fracture in the ice layer. Above the eutectic temperature the ice is in contact with a solution at its melting point, so the driving force for crytsal growth, and thus crack penetration is supressed. As a result the glue spall stress dominates the fracture process, yielding a damage morphology that agrees with the prediction of the corresponding fracture mechanics analysis.