Freezing mechanisms in concrete

This study addresses the durability of concrete subjected to freezing conditions. The solidification of pore water solution generates different local pressures in the cement matrix. It is possible to distinguish between 'Hydraulic Pressures', 'Osmotic Pressures' and pressures resulting from undercooled fluid circulations according to Litvan's Theory. By overcoming the local tensile strength of the material, these pressures can cause the material to crack. In the pores, ice does not form at 0 deg C but at lower temperatures, as low as -40 deg C. In the literature, this mechanism is explained by the superimposition of different effects, i.e. size restrictions and the presence of solutes. An attempt was made to evaluate numerically the undercooling of pore water solution due to these effects, using the thermodynamic equilibrium between phases for this calculation. An experimental study was carried out using a year old saturated Ordinary Cement Paste with a water to cement mass ratio of 0.5. The cement paste was cured for three months in NaOH and KOH solutions at 3 and 10 g/l, respectively, and then stored in a plastic bag to prevent water release. Solute concentrations were selected from values given by Longuet, who made pore water solution chemistry analyses. The ice formation temperature was measured by Differential Scanning Calorimetry on the BT 2.15 Setaram Calorimeter, and the values are shown. Two different peaks were observed for ice formation, centered near -10 deg C and -40 deg C respectively. According to the classic porous description of the material with capillary and gel pores, peak 1 would correspond to ice formation in capillary pores, and peak 2 in the coarser hydrate pores. Almost all the ice melted near -5 deg C. Undercooling of pore water solution was evaluated using thermodynamic laws. Based on the work done by Radjy, complete equilibrium between phases is described by Equation 1. The effects evaluated were: (i) size restrictions, (ii) presence of solutes, and (iii) confinement pressure of solution in the pores. All these effects tend to lower the freezing point of the solution. Integrated equations corresponding to each of the effects and results shown as graphs are presented respectively in Equations 2, 3 and 4 and Figures 2, 3 and 4. In order to evaluate undercooling of pore water solution, a medium capillary pore and a medium hydrate pore were modeled with the characteristics given in Table 1, and the associated undercooling then calculated, as -2.8 and -6.4 deg C respectively. The calculated undercooling were significantly lower than experimentally observed undercooling. In fact, calculated undercooling correspond to the lowering of the pore water solution freezing points as well the stability temperatures of ice crystals in the medium pores. This explains why the calculated values show good agreement with measured ice melt temperatures. One possible explanation of the lowest ice formation temperatures observed can be found in the mechanisms of ice germination. These mechanisms are described by nucleation theory [8]. Ice forms from clusters or nuclei which are heaps of H2O molecules tied by hydrogen bonds. At any given temperature below 0 deg C, a nucleus must reach a critical size before nucleation can occur, followed by ice crystal growth. At the stability temperature, the concentration in a critical nucleus is very low, and the probability of nucleation is 0. With decreasing temperature, the critical nucleus concentration, increases exponentially up to a threshold value, where nucleation probability suddenly becomes 1 and ice forms. In a volume of pure H2O molecules, the spontaneous nucleation point is close to -40 deg C, as in the case of ice formation in clouds. For a volume of ordinary water, impurities and surfaces act as nucleation sites. They modify the surface energy of the clusters and permit nucleation at a temperature closer to the crystal stability temperature, which is generally 0 deg C. Considering all these factors, it appears possible to study the ice formation temperature in a cement paste, not in the conventional way invoking only size restrictions and presence of a solute, but also considering the mechanisms of ice germination described by nucleation theory. (authors)