Compartmentalized Water Warping Experiment


In light of the empirical and theoretical evidence presented to this point, it can be concluded that glue-spalling is the primary cause of salt scaling. Since it is necessary for the ice layer to crack for the glue-spall mechanism to cause damage, the pessimum arises due to the effect of brine pockets on the mechanical properties of ice. It is also important to verify that the glue-spall mechanism can account for all of the characteristics of salt scaling:

Chapter 2 (Request this document) of my Ph. D thesis consists of a survey of the Salt Scaling studies performed over the past 60 years. This extensive review elucidated the following characteristics of salt scaling. We use these characteristics to judge the adequacy of hte glue-spall mechanism.

  • 1) Salt Scaling consists of the progressive removal of small flakes or chips of binder.

    The morphology of damage from glue-spalling is similar to that observed from salt scaling. The damage should be progressive. The fracture mechanics analysis indicates that the depth of scalloping depends only on the mechanical properties and dimensions of the constituents. If the properties of the surface do not vary significantly with depth, then with each freeze/thaw cycle a relatively consistent amount of material will be removed by this mechanism. Once the coarse aggregate is exposed, the ice layer will bind to the surface of the aggregates, and cracks in the ice above aggregate will not result in damage, so the rate of removal of material per unit area in each freeze/thaw cycles should eventually drop off. Accordingly, when sawn surfaces are tested in scaling experiments, the scaling rate is very similar to that realized on a finished surface, if the weight loss is calculated based on the fraction of the surface occupied by cementitious binder.
  • 2) A pessimum exists at a solute concentration of ~ 3%, independent of the solute used.

    The glue-spall mechanism requires that the ice layer crack to result in scaling damage. The viscoelastic stress calculation indicated that the pessimum concentration is a consequence of the effect of brine pockets on the mechanical properties of ice: i) due to creep, pure ice will not crack; ii) the stress rises faster than the strength in ice formed from a moderately concentrated brine; and iii) more highly concentrated solutions produce ice that has no strength in the temperature range of interest. The same behavior is observed with various solutes because the solutes exhibit similar melting point reductions as a function of solute concentration. Therefore, the microstructure of the brine ice is consistent, which suggests that the fracture and creep behavior are also consistent.
  • 3) No scaling occurs when the pool of solution is missing from the concrete surface.

    There is no composite, and no stress, without the pool.
  • 4) No damage occurs when the minimum temperature is held above -10°C; the amount of damage increases as the minimum temperature decreases below -10°C and with longer time at the minimum temperature.

    Ice formed from a 3% NaCl solution, which is used in all conventional tests, does not have strength above this temperature. The glue-spall stress is proportional to the undercooling, ΔT, so smaller and smaller flaws will propagate as the temperature drops.
  • 5) Air entrainment improves salt scaling resistance.

    Air entrainment reduces bleeding so it produces a stronger surface. The beneficial effect of AEA is expected for all test surfaces (ie: finished, molded, or sawn), because the effect on bleeding is realized throughout the concrete body. Furthermore, when an effective air-entraining agent is properly utilized, initial freezing in air voids imposes suction in the pore fluid that compresses the porous skeleton.
  • 6) The salt concentration of the pool on the surface is more important than the salt concentration in the pore solution.

    The pore liquid has no effect on the severity of the glue-spall stress, because internal ice formation does not play a role in salt scaling.
  • 7) Susceptibility to salt scaling is not correlated with susceptibility to internal frost action.

    The glue-spall mechanism is not related to internal crystallization.
  • 8) The strength of the surface governs the ability of a cementitious body to resist salt scaling.

    The strength of the finished surface is important, because the estimated glue-spall stress is marginal with respect to the tensile strength of the cementitious media (~3 MPa). Any treatment or handling expected to weaken the surface (e.g., bleeding) will therefore result in more scaling.