Mechanisms of salt frost scaling on portland cement-bound materials: studies and hypothesis

Sammanfattning: A hypothesis regarding the mechanism causing salt frost scaling on Portland cement-bound materials is described. It is assumed that deterioration is due to osmotic micro ice body growth, as has been previously proposed for frost deterioration of moisture-isolated specimens of cement-bound materials. In moisture-isolated specimens, this ice body growth stops when the micro structure is drained to a certain extent. In a specimen ex-posed to salt frost attack, however, an outer liquid phase will be present on the specimen surface (as long as temperature is above the eutecticum for the actual de-icer solution). When micro ice bodies just beneath the specimen surface begin to grow, thereby draining the surrounding pore system, the outer liquid is sucked into the pore system, thereby eliminating the drainage required for micro ice body growth to stop. In a zone close to the material surface in which the liquid from outside may be transported to the growing ice bodies at a high enough rate, micro ice bodies will thus be able to grow much more than in a moisture-isolated specimen. The specimen surface will therefore be destroyed. Predictions are made concerning the dependence of scaling and moisture uptake on minimum temperature, cooling rate, presence of de-icing agent and microstructure, and these are compared to results reported in the literature and also to experiments carried out as part of this research. Since these predictions are fulfilled, it seems reasonable that the major mechanism causing salt frost scaling is the one described.

Results are reported from a study of chloride ion penetration into hardened Portland cement paste over short periods, 0.5-3 hours. A simple relation, based on the Kozeny-Carman equation, between the transport coefficients and the porosity and specific surface of the material is presented.

A study of the differential sorption enthalpy and differential sorption entropy was carried out using a recently developed sorption calorimeter. The results cover the range 0-95% relative humidity at +25°C. Desorption isotherms were determined at 5° and 18°C by drying samples over saturated salt solutions. Freezing calorimetry tests were conducted on the same materials. The results from the room temperature experiments were used to calculate what heat flow rates were to be expected during the melting phase of the freezing calorimetry tests. These calculated heat flow rates are lower than those actually obtained. It is hypothesised that this is due to unintentional carbonation of the samples used for determination of the desorption isotherms and thus the calculated amount of water melting at a certain temperature is too low. The results though indicate that ice formation takes place mainly through successive drying of the micro pores.