Pore-scale simulation of reactive transport processes at the steel–concrete interface

Corrosion of steel in concrete involves complex transport and (electro-) chemical processes at the steel–concrete interface (SCI). Reactive transport modeling is a powerful tool to investigate these mechanisms and has strong potential to provide critical insights into corrosion-related deterioration and thereby contribute to ensuring the durability of traditional and modern low-carbon concretes. Traditional reactive transport models use homogenization techniques that are computationally efficient but overlook the detailed pore structure at the SCI. Our study introduces an approach that leverages high-resolution FIB-SEM nanotomograms to create discretized finite element domains, enabling direct pore-scale reactive transport modeling. We showcase selected examples involving aqueous species relevant to corrosion, including corrosion product precipitation in the interfacial pore network. Simulation results demonstrate that pore structure gives rise to significant local variations in concentrations and precipitate formation, thereby explaining the heterogeneity of corrosion phenomena in cementitious materials. Notably, narrow pores acting as transport bottlenecks caused local aqueous accumulation, promoting precipitation in contrast to more open regions. For comparison, a homogenized model – using independently computed diffusion tortuosity and porosity from 3D pore structures – was implemented. Remarkably, without parameter fitting, this model captured average trends but failed to resolve key localized phenomena like early corrosion product precipitation. We conclude that explicitly modeling pore-scale structures is essential for accurately capturing reactive transport processes at the SCI. Our framework offers a path towards more accurate simulation of micron- and sub-micron scale processes, challenging simplifying assumptions in traditional models, and emphasizing the need for incorporating local structural features at the SCI in corrosion modeling.