The role of concrete resistivity in chloride-induced macro-cell corrosion of reinforcement

Abstract

Norwegian concrete structures are exposed to salt contamination from both sea water and de-icing salts. Chlorides penetrate the concrete cover and initiate corrosion of the reinforcement. The deterioration process can propagate at high rates and endanger serviceability and structural safety. Concrete resistivity is usually assumed to correlate closely with the corrosion process, and its measurement is in principle simple and costefficient, and widely recommended to support service life assessment. However, the extent to which concrete resistivity measurements can provide us with detailed information about the corrosion process needs critical examination. The main objective of this PhD project was to study the applicability of concrete resistivity measurement for predicting the rate of reinforcement corrosion induced by chlorides. A review of the literature led to the conclusion that a general correlation could be found between increasing concrete resistivity and decreasing corrosion rate, but a large scatter was observed within and between the studies reviewed. Various possible reasons for the discrepancies were identified: researchers used dissimilar procedures for determining both concrete resistivity and corrosion rate, and concrete composition was recognised as affecting the overall tendency. Interestingly, the literature contained very limited discussion on the mechanisms involved in providing the correlation assumed between concrete resistivity and corrosion rate. On the basis of the literature review, it seemed appropriate to investigate the relationship between concrete resistivity and corrosion rate further using specially designed laboratory experiments. In total, three experimental studies were conducted using mortar to investigate the subject. The first study focused on the limitations of measuring galvanic currents in macro-cells. It was confirmed that galvanic current measurements provide an indication of the actual corrosion current, and thereby rate, only if the actively corroding segment is sufficiently small. The second experimental study investigated the limitations of bulk resistivity measurements for predicting the ohmic resistance in a macro-cell between small anodes and a large cathode network. It was found that for small anodes the ohmic resistance between the anode and the cathode is not directly correlated with the bulk resistivity. Local inhomogeneities strongly influence the measurements. The ratio between bulk resistivity and the resistance between a specific anode and the cathode was not constant over time and varied in particular with changes in the moisture conditions. The third experimental study examined the influence of mortar resistivity on the ratelimiting steps of chloride-induced corrosion. It was shown that for small anodes the local conditions around the anode are more decisive for the anodic and ohmic partial processes than the bulk resistivity. The cathodic partial process was found to be practically independent of the mortar resistivity. For the setup and materials investigated, it was not possible to identify any one rate-limiting step as dominant. The controlling rate-limiting step varied for the different materials investigated. For mortar prepared with fly ash, and therefore possessing high mortar resistivity, the concept of anodic-resistance control suggested in the literature was found to fit well. Cathodic control was identified for mortars with a comparatively low bulk resistivity (prepared only with Portland cement). Furthermore, it was confirmed that the generally assumed inverse correlation between corrosion rate and concrete resistivity depends very much on the material composition. In conclusion, it was found that there can be no direct relationship between concrete resistivity and the corrosion rate. The dependency on material composition and in particular the fact that concrete resistivity is not directly correlated with the underlying partial processes that determine the corrosion rate must be considered as decisive on this question. It must therefore be assumed that the direct relationships between concrete resistivity and corrosion rate introduced in the literature, and any models and general threshold values based upon them, are not generally applicable.