Crevice corrosion occurs as a result of aggressive conditions inside small occluded regions on metal surfaces. One of the primary stumbling blocks to the comparison of modeling and experimental results in crevice corrosion has been the extraordinary difficulty involved in experimentally producing a crevice sample that is directly equivalent to a geometry that is amenable to computational modeling. In this work, both sides of the crevice are constructed on silicon wafer substrates. The metal side of the crevice is formed by patterning, etching, and oxidizing the silicon before deposition of the metal. A highly planar surface is produced. The metal surface is segmented to allow measurement of current distribution within the crevice. The crevice-former side of the crevice (formed by patterning and depositing SiO2) contains precisely grown spacers that control the gap between the crevice-former surface and the metal electrode to a uniform, preselected level. The crevice former contains microsensors for pH, chloride concentration, and electrochemical potential. Using these samples, measurements are made that directly relate to computational results from a recently developed two dimensional occluded region mass transport model. In this way, the phenomena controlling crevice corrosion initiation and propagation are probed. %%% The work investigates issues that have prevented the development of effective crevice corrosion mitigation strategies as well as life prediction methodologies. The research has impact on other forms of corrosion because the critical structures and processes that control many types of corrosion operate on the micron to sub-mm size scale. For example, the chemistry and electrochemistry associated with corrosion due to inclusions in stainless steel, constituent particles in aluminum alloys, droplets in atmospheric corrosion, and defects in organic coatings occur on the 1-micron to several mm size scales. The size scale and dimensionality of the controlling structures make experimental and computational studies very challenging, requiring compromises and approximations in order to make the problem tractable. This new approach permits improved quantitative assessments. ***
