Abstract
A series of classical G109 type concrete specimens was exposed to cyclic wet and dry ponding with 15 w/o NaCl solution for approximately five years. Mix design variables included 1) three cement alkalinities (EqA of 0.97, 0.52, and 0.36) and 2) three water-cement ratios (0.50, 0.41, and 0.37). To determine the corrosion initiation time, corrosion potential and macro-cell current between top and bottom bars were monitored. Subsequent to corrosion initiation, specimens were autopsied and visually inspected. Concrete powder samples were collected from top rebar trace and chloride concentration was measured. Also, time-to-corrosion, $T_i$, for specimens of the individual mix designs was represented using Weibull analysis. Time-to-corrosion was a distributed parameter; and because of this, corrosion initiation of four identical specimens for each mix varied, often over a relatively wide range. Specimens fabricated using the lowest water cement ratio and the highest alkalinity cement exhibited the longest time-to-corrosion initiation and the highest chloride threshold levels. Time-to-corrosion did not increase monotonically with cement alkalinity, however, presumably as a consequence of relatively high $Cl^-$ binding in the lower pore water pH range. The chloride threshold level, $Cl_{th}$, increased with increasing $T_i$ and, consequently, was greatest for the highest cement alkalinity specimens.