1. Introduction Almost every code of concrete structural design relates permissible crack width to exposure conditions, indicating the general agreement about the influence of crack on corrosion. The relationship between crack width and corrosion, however, has not been clarified. The locations of steel corrosion normally coincided with tips of flexural cracks revealing the indisputable effect of crack on corrosion initiation. It also appeared that steel was not corroded at every crack. In addition, the effect of crack on corrosion propagation is remarkably perplexed. With the same surface crack width, the deterioration of steel corrosion can range from zero to high degree of corrosion. The study aims to clarify the mechanism of corrosion induced by crack and the relationship between crack width and corrosion by using an electrochemical technique. 2. Experimental Procedures The Specimens and loading arrangement are shown in Fig. 1. The designation and steel stress of the specimens are tabulated in Table 1. The steel reinforcement was one D 13 millscale bar. The mix proportion and strength properties of concrete are shown in Table 2. Two levels of flexural crack width were introduced and sustained under tensile steel stresses of 2000 and 3000 kg/cm^2. Steel corrosion was accelerated by subjecting the specimens to repeated cycles of one day wetting in 65℃, 3.1 % NaCl solution and one day drying under room environment. Exposure periods are also shown in Table 1. During the exposure test, Half cell potential (E_c) and Polarization resistance (R_p) of steel were periodically measured against a Ag/AgCl reference electrode and a platinum counter electrode. The specimens were split at the end of their exposure periods to observe corrosion situations such as corrosion form, corroded length and rusted area. 3. Results and Discussion 3.1 Locations of Crack and Rust Fig. 2 shows the locations of crack and rust. All rusted areas were associated with crack tips as they were reported. The aggressive substances essential for corrosion processes, such as Cl^-, H_2O, O_2 etc., can undoubtly penetrate through crack easier than through sound cover. However, it was also found that, steel did not corrode at every crack as presented in the literature 3.2 The electrochemical characteristics Fig. 3 shows the typical distribution of E_c and R_p, measured at various exposure ages, along four locations of four specimens. After a period of exposure, in each specimen, one of the four locations remarkably showed the higher reduction in Ec and Rp values comparing to other three locations. Corrosion was certainly found under these spots (see specimens S (3, 2) and L (1, 1)). Fig. 4 shows the typical E_c and R_p time curves. The Ec negative shift revealed the depassivating period after the amount of absorbed chloride ions had exceeded the critical value of C1^-/OH^<-20>). During the depassivating, although all points of the specimens altogether shifted to more negative potentials, only R_p of the corroding spots declined continuously. R_p of the uncorroded spots slightly increased with time. For specimens with more than one corroded spot, the corroding order of the corroded spots can be determined by their R_p characteristics (see specimens S (4, 1) and L (4, 1)). 3.3 The electrochemical characteristics and corrosion situations The comparison of average electrochemical values of corroding and uncorroded spots, measured at various locations, is shown in Fig. 5. Both corroding and uncorroded spots had the same trend of E_c with time. The actual E_c of the uncorroded spots might be defected by the presence of E_c of the corroding spots. In contrast with E_c time curve, the corroding and uncorroded spots had clearly different trend of R_p with time. The R_p of the corroding spots continuously shifted to more negative values, whilst the R_p of the uncorroded spots tended to increase gradually. The active

(View PDF for the rest of the abstract.)