Service Life Prediction of Concrete
Considerations of Specimen Conditioning and Testing Methods on the Measurement of Porosity and Diffusion Coefficient of Chloride
Dr. Yiwen Bu recently joined Twining’s AER division as a Research Engineer. Dr. Bu is a recent graduate from Purdue University with her PhD in Materials Science and Engineering. Her work at Purdue University includes nanotechnology of cementitious materials, service life modeling and durability assessment of concrete to corrosive environment, as well as new technologies for quality control of concrete production. Dr. Bu co-authored more than 10 publications in proceedings of international conferences as well as peer-reviewed journals on cement and concrete materials. Using her knowledge and expertise to bring about technology innovation and upgrade to the construction industry is her greatest enthusiasm.
Service life prediction models have been developed and are becoming frequently used to predict the time of corrosion initiation for concrete materials. These models usually require the input of experimentally obtained material properties such as porosity and ionic diffusion coefficients. In this thesis, the influence of specimen conditioning and testing methods on the measured porosity and ionic diffusion coefficients is investigated.
It is preferred to measure the transport properties of concrete materials at a later age (typically 91 days) when the degree of hydration ceases to increase substantially and when the microstructure has matured. Accelerated curing (at 50 ± 2 °C) was evaluated in an attempt to shorten the time needed to acquire transport properties for use in a Nernst-Planck service life prediction model (Stadium®). However, when the samples were stored in lime water, the microstructure formed at the higher temperature showed an increase in total porosity, chloride diffusion coefficient, and permeability. Using these values as inputs for service life prediction for the example used here resulted in a prediction of corrosion initiation times that were approximately 30 % earlier than those predicted using samples moist cured at 23 ± 2 °C (NA).
While accelerated curing is an approach to accelerate the curing of concrete materials, electrical testing is being proposed as a rapid assessment of the transport properties of concrete materials, due to the relationship between electrical resistivity and diffusion coefficients in the Nernst-Einstein equation. The influence of the alkali content of cement and the leaching of alkalis during wet curing on the electrical measurements and bulk properties of concrete materials were investigated. It was found that during wet curing, a significant amount of alkalis leach from the pore solution to the storage solution, resulting in higher pore solution resistivity and higher bulk resistivity. Alkali leaching may also influence the bulk resistivity through the microstructure that forms, since higher alkali content in the pore solution results in a more dense microstructure that has reduced porosity and reduced ionic diffusion coefficients.
Porosity measurements using ASTM C642-13 and using vacuum saturation were investigated. ASTM C642-13 is being increasingly used in measurements of the total pore volume (porosity) for use in concrete transport and durability calculations. CHAPTER 4 demonstrated that ASTM C642-13 provides a measure of the porosity of concrete that does not appear to include a large portion of entrapped/entrained air void system. Rather, ASTM C642-13 provides a volume that is primarily the porosity of the matrix and aggregate. However, vacuum saturation under a sufficient pressure could saturate the entrained air void system. This chapter indicates that tests for the pore space obtained using vacuum pressures would be a valuable specification for the increasing need for consistent total pore volumes in durability and transport calculations.
The results of steady-state and non steady-state migration tests, as well as electrical resistivity measurements when entrained air voids are saturated and when they are primarily empty were compared. It is shown that specimens with entrained air voids empty showed reduced surface chloride concentration in non-steady state migration tests, reduced cumulative charges passed during steady state migration tests, and increased bulk electrical resistivity, compared to specimens with entrained air voids saturated.
There are two approaches for assessing chloride ingress: Fick’s second law and the Nernst-Planck equation were examined. Fick’s second law provides an empirical fit of data for the apparent diffusion coefficient DAPP and surface concentration CS, both of which were found to be dependent on the chloride concentrations, the duration of exposure, and the types of co-present ionic species. The Nernst-Planck approach used in this study is less empirical as a single diffusion coefficient (DCl–) is determined in a separate test and the physical/chemical reactions in the cementitious systems are independently addressed. It was found that, at lower concentrations, the Nernst-Einstein approach provides a reasonable prediction when compared with the experimentally (and independently) determined chloride profiles. However, at higher concentrations (>2mol/L), the simulated results deviate from the experiments. The authors suspect that this deviation may be due to changes in viscosity that occur at high salt concentrations that do not appear to be considered. Additionally, the chloride binding and porosity change at higher concentrations (>2mol/L) appear to be under-predicted in the current approach. Moreover, the possible expansion due to ettringite formation upon exposure to seawater was not predicted. The influence of CSH surface charges on chloride ingress was also not considered. The above reasons caused discrepancies between the experimental chloride profiles and the simulated chloride profiles and are important considerations for future improvement in the service life prediction of concrete.
