The role of subsurface nano-scale voids present near the metal-oxide film interface, on pit nucleation was investigated. Electrochemical processes during the initial stages of etching were thoroughly characterized to evaluate the hypothesis that voids are the primary pit initiation sites. Dissolution rate measurements made in etch pits revealed contradicting trends of constant and potential dependent dissolution current densities. The same trends were also exhibited by the measurements made in etch tunnels, which eliminated the possible role of geometric form of corrosion in such a contradiction. Finally it was concluded that the experimental time-scale during which the dissolution rate measurements were made, was the only factor responsible for such varying trends. A kinetic model similar to the Vetter-Gorn model for metals covered with oxide films was proposed. The model was validated by its ability to predict the observed constant and potential dependent dissolution rates under different experimental-time scales. A mathematical model for pit initiation during the initial stages of galvanostatic anodic etching of aluminum in acid-chloride solutions was developed. The model incorporated all of the electrochemical processes characterized during the initial stages of etching and was based on the interfacial void hypothesis which assumes voids to be the only pit initiation sites. The effect of various experimental conditions such as caustic pretreatment time, applied current density, etchant temperature, chloride concentration in etchant etc. were analyzed using the etching potential transient. The dependence of void concentration on the caustic pretreatment time was determined by fitting the potential transients and the observed trend had a fine agreement with previously established PAS results. Using the fit void concentrations, the model could successfully track the experimental transients for various etchant temperatures and applied current densities. However, the model failed to predict the temperature-dependent pit depths found experimentally, which might be due to the simplified assumptions made in its development. Also the role of chloride ion kinetics in pit initiation mechanism should be clearly understood and incorporated in the model for successfully predicting the experimental results obtained for varied chloride concentrations in the etchant.