Corrosion of Steel in Alkali-Activated Materials

The alkaline environment in reinforced concrete structures allows the embedded steel reinforcement to be passivated by a thin film, composed of a complex solid assemblage of iron oxides and hydroxides. When in service, the alkalinity of the pore solution in mild steel reinforced concretes needs to be sufficiently high (pH > 9-10) to ensure the stability of the passive film. Corrosion of the embedded steel reinforcement has been identified as one of the primary causes of premature degradation of reinforced concrete structures. Corrosion of steel in concrete environments is primarily driven by either the lowering of the pH of the pore solution (leaching of alkalis, or carbonation) which can lead to uniform corrosion; or due to the ingress of chloride in significant concentrations resulting in localised breakdown of the passive film. In developed parts of the world, where the use of de-icing salts and occurrence of marine sprays (in coastal regions) are frequent, chloride ingress is one of the major causes of premature failure of reinforced concrete structures. Therefore, the durability or the service life of a structure is often described by the rate of chloride ingress, or the time required for chloride ions to achieve a threshold concentration at the steel-concrete interface. Alkali-activated materials (AAMs) and are the products of the reaction between an aluminosilicate source and an aqueous alkaline 'activator' and have galvanised significant interest within industry and academia as a sustainable and technically sound alternative to Portland cement (PC). The pore structure and pore solution chemistry of AAMs have been found to be significantly different from those of PC-based concretes; and within the class of materials that can be described as AAMs, the composition of the pore solution and the pore network vary due to the differences in the precursor chemistry. The ingress of chloride into AAMs has also been found to differ from the behaviour of PC-based concretes. The development of AAMs as an alternative to PC has seen substantial progress in the past decades, however, there still remains significant uncertainty regarding their long-term performance when used in steel-reinforced structures. This study investigates the durability of steel-reinforced AAMs through a simulated pore solution approach, understanding the interaction of the steel reinforcement with the pore solution at the steel-concrete interface, and the initiation of chloride-induced corrosion. In the case of simulated pore solutions representative of low-Ca AAMs (such as alkali-activated fly ashes and metakaolin), the passive film on the steel rebars was complex in chemical makeup, composed of Fe-hydroxides, oxy-hydroxides and oxides. However, in the case of high-Ca AAMs, the surface of the steel reinforcements were found to be primarily composed of Fe2+ species, with the inner layer being Fe(OH)2 (that could potentially oxidise to a hydrated Fe3+ oxide if sufficient oxygen is available at later stages) and the outer layer being a Fe-S complex, possibly resembling disordered mackinawite. The onset of corrosion or the chloride 'threshold' value in both low-Ca and high-Ca AAMs were found to be very different from PC based binders. In the case of low-Ca AAMs, the chloride 'threshold' value ([Cl-]/[OH-] ratio) required to induce depassivation of steel was strongly dependent on the alkalinity of the pore solution, and a novel relationship to predict the onset of pitting, interlinking chloride concentration and the solubility of the passive film, given by [Cl-]/[OH-]3 = 1.25, was developed. In the case of high-Ca AAMs, chloride-induced corrosion was usually not evidenced. The high concentration of HS- in the pore solution of high-Ca AAMs not only creates a highly reducing environment around the SCI and hinders the development of passive iron oxide film (and creates a different 'surface' film consisting of mackinawite), but also restricts the cathodic reduction of oxygen and the formation of a macro-cell. Additionally, the influence of [HS-] and time of exposure on the chloride 'threshold' value were investigated, and for pore solutions with intermediate concentrations of HS- (0.01 M) were found to be the most susceptible to chloride-induced corrosion. Finally, a user-defined framework for predicting the service-life of steel reinforced AAMs (relevant to alkali-activated slags as of now) was developed that investigates diffusion of chloride in slag-based binders activated using various activators (NaOH, Na2SiO3, Na2Si2O5, Na2CO3 and Na2SO4). The model predicts the chemistry of the concrete cover, and the influence of the reaction products on the chloride binding capacity of the binder used, for any user defined composition of slag and choice of activator. Based on the calculated chloride binding capacity of the binder, the model uses an explicit numerical method to estimate the ingress of chloride through the concrete cover as a function of space and time.