Distribution of Steel Corrosion Products in a Cracked Concrete Beam Exposed to Coastal Atmosphere

Jianfeng Dong, Xuan Lin, Xueyuan Liu, Jing Guo, Yuxi ZhaoLecture Notes in Civil Engineering 2024 · DOI · OpenAlex

Abstract

This study investigated the distribution of steel corrosion products in a cracked concrete beam that was in service under coastal atmospheric conditions for more than 40 years. The research measured the thickness of the corrosion-product-filled paste (CP) and the dense corrosion layer (CL), while also investigating the migration behavior of Iron (Fe) ions within concrete cracks and air voids. The findings reveal that steel corrosion and CP propagation occur significantly faster in the tension region of the beam compared to the compression region. Notably, in natural atmospheric environments, corrosion products do not fill cracks or air voids, nor do they accumulate on crack edges—a finding that contradicts many laboratory-based simulations.

Introduction

Predicting the service life of reinforced concrete structures is a critical challenge in civil engineering. The primary driver of degradation is the corrosion of steel rebar, which produces rust occupying two to six times the volume of the original metal. This expansion generates internal pressure, leading to "corrosion-induced cracking." While numerous predictive models exist, most are derived from accelerated laboratory experiments using salt baths or electrical currents. This study provides a rare forensic analysis of a real-world structure—a concrete beam salvaged from a building roof on an island off the east coast of China—to determine how decades of natural coastal exposure affect rust distribution.

Methodology

The subject of the study was a concrete beam exposed to a coastal environment for over 40 years. The island environment featured an average temperature of 16.2°C and high annual rainfall (1624 mm), providing a rigorous natural testbed for corrosion.

To analyze the internal state of the beam, researchers employed a precise sectioning and microscopy workflow:

  1. Sectioning: The beam was cut into segments and embedded in low-viscosity epoxy resin to stabilize the internal structures. It was then sliced into 12mm thick "wafers."
  2. Microscopy and Mapping: Scanning Electron Microscopy (SEM) in Backscattered Electron (BSE) mode was used to observe the steel/concrete interface, while Energy Dispersive X-Ray Spectroscopy (EDS) tracked the migration of Iron (Fe) ions.
  3. Quantification: Researchers focused on two distinct layers: the CL (Corrosion Layer), which is the dense rust layer exerting pressure on the concrete, and the CP (Corrosion-product-filled Paste), representing the surrounding concrete pores saturated with migrating iron ions.
Figure 4: Schematic of the beam sectioning and rebar identification (S1-C, S2-C for compression; S1-T, S2-T for tension).

Key Insights and Results

Stress-Dependent Corrosion Rates

The study found that the mechanical stress state of the beam significantly dictates the speed of rust propagation. Corrosion was found to be most severe in the tension regions (T1 and T2). Tensile stress opens micro-pores and connects them, creating pathways for chloride and moisture ingress. In contrast, compressive stress (C) densifies the concrete matrix, slowing the chemical attack. Except for minor outliers, the thickness of the corrosion layer followed the gradient: T1 (highest tension) > T2 > C (compression).

The "Natural vs. Lab" Discrepancy

A pivotal finding of this research is the behavior of rust in voids and cracks. In laboratory settings, rust often fills cracks and air voids. However, in this 40-year-old beam, the voids remained largely empty. The authors hypothesize that in lab simulations, liquid solutions (like NaCl baths) physically carry dissolved rust into these spaces. In a natural atmospheric environment, there is insufficient liquid flow to transport the rust, leaving it confined to the cement paste immediately surrounding the rebar.

Figure 6: SEM/BSE image showing the distinct layers at the interface: Steel, Mill Scale (MS), Corrosion Layer (CL), and Corrosion-product-filled Paste (CP).

Environmental Impact: Windward vs. Leeward

The orientation of the beam also played a major role. The windward side (S2), which directly faced the sea and chloride-laden breezes, exhibited significantly higher corrosion ($T_{CL}$) than the leeward side (S1). This confirms that local micro-climates are as important as structural design in determining long-term durability.

Conclusion

This forensic analysis challenges several assumptions used in current corrosion modeling. By demonstrating that air voids do not act as "sinks" for rust in natural atmospheric conditions, the study suggests that pressure buildup may occur sooner than lab models predict. Furthermore, the clear acceleration of corrosion in tension zones emphasizes the need for stress-dependent durability parameters in engineering design. Future work will utilize remaining portions of the beam to track these degradation patterns as the structure reaches the 50-year mark.

Summary updated

Abstract

Abstract This study investigated the distribution of steel corrosion products in a cracked concrete beam, which was in service under coastal atmospheric conditions for more than 40 years. The thickness of corrosion-product-filled paste (CP) and the thickness of corrosion layer (CL) were measured. The migration behaviour of Fe in the concrete cracks and air voids are investigated. The mechanism of CP formation was revealed. The steel corrosion and corrosion-product-filled paste propagates faster in the tension region than in the compressive region of the beam. In the atmospheric environment, the corrosion products neither fill the cracks or air voids nor do they accumulate on the edge of the cracks or inside the air voids.
BibTeX
@inproceedings{dong2024distribution,
  author = {Jianfeng Dong and Xuan Lin and Xueyuan Liu and Jing Guo and Yuxi Zhao},
  title = {{Distribution of Steel Corrosion Products in a Cracked Concrete Beam Exposed to Coastal Atmosphere}},
  booktitle = {Lecture Notes in Civil Engineering},
  year = {2024},
  doi = {10.1007/978-981-97-6238-5\_16},
}
Figure